THE PROCESS OF LEARNING

Chapter 3 Behaviorism

It’s the end of the school day at Park Lane Elementary, and three teachers leave the building together: Leo Battaglia, Shayna Brown, and Emily Matsui. Their conversation as they walk to the parking lot is as follows:

Leo:	Boy, they were wild today. I don’t know what got into them. Hardly anyone earned any points today.
Emily:	What points, Leo?
Leo:	I give points for good behavior, which they then can exchange for privileges, such as extra free time. I take away points when they misbehave.
Emily:	And it works?
Leo:	Sure does. Keeps them in line most days. But not today. Maybe there was something in the water.
Shayna:	Or in their heads, most likely. What do you suppose they were thinking about? Maybe spring break next week?
Leo:	Perhaps. But it’s not my job to see into their heads. Lots of things can trigger wild behavior. How am I supposed to know what does? That’s why I focus on the behavior.
Shayna:	But sometimes we need to go beyond the behavior. For example, Sean’s been acting out lately. If I had just focused on his behavior, I would not have learned that his parents are getting divorced and he’s blaming himself for it.
Leo:	Isn’t that why we have a counselor? Isn’t that her job?
Shayna:	Yes, it is, but we have a role, too. I think you focus too much on what you see and not enough on what you don’t see.
Leo:	Perhaps, but at least I keep them under control with my system of rewards and punishments. I don’t waste a lot of time on classroom management issues.
Emily:	Or on personal issues, like their thoughts and emotions.

Against the background of structuralism and functionalism ( Chapter 1 ), behaviorism began its rise to become the leading psychological discipline of the first half of the 20th century. John B. Watson (1878–1958), generally considered to be the founder and champion of behaviorism (Heidbreder,  1933 ; Hunt,  1993 ), believed that theories and research methods that dealt with the mind were unscientific. If psychology were to become a science, it had to structure itself along the lines of the physical sciences, which examined observable and measurable phenomena. Behavior was the proper material for psychologists to study (Watson,  1924 ). Introspection ( Chapter 1 ) was unreliable; conscious experiences were not observable and people having such experiences could not be trusted to report them accurately (Murray, Kilgour, & Wasylkiw,  2000 ).

Watson ( 1916 ) thought that Pavlov’s conditioning model (discussed later in this chapter) was appropriate for building a science of human behavior. He was impressed with Pavlov’s precise measurement of observable behaviors. Watson believed that Pavlov’s model could account for diverse forms of learning and personality characteristics. For example, newborns are capable of displaying three emotions: love, fear, and rage (Watson,  1926a ). Through Pavlovian conditioning, these emotions could become attached to stimuli to produce a complex adult life. Watson expressed his belief in the power of conditioning in this famous pronouncement:

·  Give me a dozen healthy infants, well-formed, and my own specified world to bring them up in and I’ll guarantee to take any one at random and train him to become any type of specialist I might select—a doctor, lawyer, artist, merchant-chief and, yes, even into beggar-man and thief, regardless of his talents, penchants, tendencies, abilities, vocations and race of his ancestors. (Watson,  1926b , p. 10)

Although Watson’s research has little relevance for academic learning, he spoke and wrote with conviction, and his adamant views influenced psychology from around 1920 until the early 1960s (Hunt,  1993 ). His emphasis on the importance of the environment is readily seen in the ensuing work of Skinner (discussed later in this chapter; Horowitz,  1992 ).

This chapter covers behaviorism as expressed in conditioning theories of learning. The hallmark of conditioning theories is not that they deal with behavior (all theories do that), but rather that they explain learning in terms of environmental events. While not denying the existence of mental phenomena, these theories contend that such phenomena are not necessary to explain learning. In the opening scenario, Leo espouses a conditioning theory position.

The best-known conditioning theory is B. F. Skinner’s  operant conditioning . Before discussing this theory, historical work on conditioning is presented to set the backdrop for Skinner’s theory: Thorndike’s connectionism, Pavlov’s classical conditioning, and Guthrie’s contiguous conditioning.

· When you finish studying this chapter, you should be able to do the following:

· ■ Explain how behaviors are learned according to connectionism theory.

· ■ Discuss some of Thorndike’s contributions to educational practice.

· ■ Explain how responses become conditioned, extinguished, and generalized, according to classical conditioning theory.

· ■ Describe a process whereby an emotional response might become conditioned to an initially neutral object.

· ■ Explain, using contiguous conditioning principles, how movements are combined to become an act.

· ■ Describe Skinner’s three-term contingency model of operant conditioning, and provide examples.

· ■ Define and exemplify key operant conditioning concepts: positive and negative reinforcement, punishment, generalization, discrimination, shaping, and the Premack Principle.

· ■ Explain how operant principles are reflected in educational applications: behavioral objectives, learning time, mastery learning, programmed instruction, and contingency contracts.

CONNECTIONISM

Edward L. Thorndike (1874–1949) was a prominent U.S. psychologist whose  connectionism  theory of learning was dominant in the United States for a long time (Mayer,  2003 ). Unlike many early psychologists, he was interested in education and especially learning, transfer, individual differences, and intelligence (Hilgard,  1996 ; McKeachie,  1990 ). He applied an experimental approach when measuring students’ achievement outcomes. His impact on education is reflected in the Thorndike Award, the highest honor given by the Division of Educational Psychology of the American Psychological Association for distinguished contributions to educational psychology.

Trial-and-Error Learning

Thorndike’s major work is the three-volume series Educational Psychology (Thorndike,  1913a ,  1913b ,  1914 ). He postulated that the most fundamental type of learning involves the forming of associations ( connections ) between sensory experiences (perceptions of stimuli or events) and neural impulses (responses) that manifest themselves behaviorally. He believed that learning often occurs by  trial and error  (selecting and connecting).

Thorndike began studying learning with a series of experiments on animals (Thorndike,  1911 ). Animals in problem situations try to attain a goal (e.g., obtain food, reach a destination). From among the many responses they can perform, they select one, perform it, and experience the consequences. The more often they make a response to a stimulus, the more firmly that response becomes connected to that stimulus. For example, a cat in a cage can open an escape hatch by pushing a stick. After a series of random responses, the cat eventually escapes by pushing the stick. Over trials, the cat reaches the goal (escape) quicker and makes fewer errors prior to responding correctly. A typical plot of results is shown in  Figure 3.1 .

Trial-and-error learning occurs gradually ( incrementally ). Connections are formed through repetition; conscious awareness is not necessary. Animals do not “catch on” or “have insight.” Thorndike understood that human learning is more complex because people engage in learning involving connecting ideas, analyzing, and reasoning (Thorndike,  1913b ). Nonetheless, the similarity in research results from animal and human studies led Thorndike to explain complex learning with elementary learning principles. An educated adult possesses millions of  stimulus–response  connections.

Principles of Learning

Laws of Exercise and Effect.

Thorndike’s basic ideas about learning are embodied in the Laws of Exercise and Effect. The  Law of Exercise  has two parts: the  Law of Use —a response to a stimulus strengthens their connection; the  Law of Disuse —when a response is not made to a stimulus, the connection’s strength is weakened (forgotten). The longer the time interval before a response is made, the greater is the decline in the connection’s strength.

Figure 3.1 Incremental performance over trials exemplifying Thorndike’s trial-and-error learning.

The  Law of Effect  emphasizes the consequences of behavior: Responses resulting in satisfying (rewarding) consequences are learned; responses producing annoying (punishing) consequences are not learned (Thorndike,  1913b ). This is a functional account of learning because satisfiers (responses that produce desirable outcomes) allow individuals to adapt to their environments.

The following study illustrates application of the Law of Effect (Thorndike,  1927 ). Participants were shown 50 strips of paper, ranging in length from 3 to 27 cm, one at a time. Next to each strip was a second strip that participants knew was 10 cm long. They initially estimated the length of each strip without feedback. Following this pretest, the 50 strips were presented again, one at a time. After each estimate, they were told “right” or “wrong” by the experimenter. After the 50 strips were presented repeatedly over several days, they again were presented without feedback about accuracy of length judgments. Following training, participants’ length estimates more closely approximated the actual lengths of the strips than had their prior estimates. Thorndike concluded that these results, which were similar to those from experiments in which animals were rewarded with food or freedom, support the idea that satisfying (correct) stimulus–response connections are strengthened and annoying (incorrect) ones are weakened.

Other Principles.

Thorndike’s ( 1913b ) theory included other principles relevant to education. The  Law of Readiness  states that when one is prepared (ready) to act, to do so is rewarding and not to do so is punishing. If one is hungry, responses that lead to food are in a state of readiness, whereas other responses not leading to food are not in a state of readiness. If one is fatigued, it is punishing to be forced to exercise. Applying this idea to learning, we might say that when students are ready to learn a particular action (in terms of developmental level or prior skill acquisition), then behaviors that foster this learning will be rewarding. When students are not ready to learn or do not possess prerequisite skills, then attempting to learn is punishing and a waste of time.

The principle of  associative shifting  refers to a situation in which responses made to a particular stimulus eventually are made to an entirely different stimulus if, on repeated trials, there are small changes in the nature of the stimulus. For example, to teach students to divide a two-digit number into a four-digit number, we first teach them to divide a one-digit number into a one-digit number and then gradually add more digits to the divisor and dividend.

The principle of  identical elements  affects  transfer  ( generalization ), or the extent that strengthening or weakening of one connection produces a similar change in another connection (Hilgard,  1996 ; Thorndike,  1913b ; see  Chapter 7 ). Transfer occurs when situations have identical (highly similar) elements and call for similar responses. Thorndike and Woodworth ( 1901 ) found that practicing a skill in a specific context did not improve one’s ability to execute that skill generally. Thus, training on estimating the area of rectangles does not advance learners’ ability to estimate the areas of triangles, circles, and irregular figures. Skills should be taught with different types of educational content for students to understand how to apply them ( Application 3.1 ).

Revisions.

APPLICATION 3.1 Facilitating Transfer

Thorndike suggested that drilling students on a specific skill does not help them master it nor does it teach them how to apply the skill in different contexts.

When teachers instruct students how to use map scales, they also must teach them to calculate miles from inches. Students become more proficient if they actually apply the skill on various maps and create maps of their own surroundings than if they are just given problems to solve.

Elementary teachers work with students on liquid and dry measurement. Having the students use a recipe to measure ingredients and create a food item is more meaningful than using pictures, charts, or filling cups with water or sand.

In teacher education courses, having students observe and become involved in actual classrooms is more meaningful than reading about and watching videos on teaching and learning.

Thorndike revised the Laws of Exercise and Effect after other research evidence did not support them (Thorndike,  1932 ). Thorndike discarded the Law of Exercise when he found that simple repetition of a situation does not necessarily “stamp in” responses. In one experiment, for example, participants closed their eyes and drew lines they thought were 2, 4, 6, and 8 inches long, hundreds of times over several days, without feedback on accuracy of the lengths (Thorndike,  1932 ). If the Law of Exercise were correct, then the response performed most often during the first 100 or so drawings ought to become more frequent afterward; but Thorndike found no support for this idea. Rather, mean lengths changed over time; people apparently experimented with different lengths because they were unsure of the correct length. In the absence of feedback, people are unlikely to perform the same behavior.

With respect to the Law of Effect, Thorndike originally thought that the effects of satisfiers (rewards) and annoyers (punishments) were opposite but comparable, but research showed this was not the case. Rather, rewards strengthened connections, but punishment did not necessarily weaken them (Thorndike,  1932 ). Instead, connections are weakened when alternative connections are strengthened. In one study (Thorndike,  1932 ), participants were presented with uncommon English words (e.g., edacious, eidolon). Each word was followed by five common English words, one of which was a correct synonym. On each trial, participants chose a synonym and underlined it, after which the experimenter said “right” (reward) or “wrong” (punishment). Reward improved learning, but punishment did not diminish the probability of that response occurring to that stimulus word.

Punishment suppresses responses, but they are not forgotten. Punishment is not an effective means of altering behavior because it does not teach students correct behaviors but rather informs them of what not to do. This also is true with cognitive skills. Brown and Burton ( 1978 ) found that students learn  buggy algorithms  (incorrect rules) for solving problems (e.g., subtract the smaller number from the larger, column by column, 4371 − 2748 = 2437). When students are informed that this method is incorrect and are given corrective feedback and practice in solving problems correctly, they learn the correct method but do not forget the old way.

Thorndike and Education

As a professor of education at Teachers College, Columbia University, Thorndike wrote books that addressed topics such as educational goals, learning processes, teaching methods, curricular sequences, and techniques for assessing educational outcomes (Hilgard,  1996 ; Mayer,  2003 ; Thorndike,  1906 ,  1912 ; Thorndike & Gates,  1929 ). Some of Thorndike’s many contributions to education are the following.

Principles of Teaching.

Teachers should help students form good habits. As Thorndike ( 1912 ) noted:

· ■ Form habits. Do not expect them to create themselves.

· ■ Beware of forming a habit which must be broken later.

· ■ Do not form two or more habits when one will do as well.

· ■ Other things being equal, have a habit formed in the way in which it is to be used. (pp. 173–174)

The last principle cautions against teaching content that is removed from its applications: “Since the forms of adjectives in German or Latin are always to be used with nouns, they should be learned with nouns” (p. 174). Students need to understand how to apply knowledge and skills they acquire. Uses should be learned in conjunction with the content.

Sequence of Curricula.

· A skill should be introduced (Thorndike & Gates,  1929 ):

· ■ At the time or just before the time when it can be used in some serviceable way

· ■ At the time when the learner is conscious of the need for it as a means of satisfying some useful purpose

· ■ When it is most suited in difficulty to the ability of the learner

· ■ When it will harmonize most fully with the level and type of emotions, tastes, instinctive and volitional dispositions most active at the time

· ■ When it is most fully facilitated by immediately preceding learnings and when it will most fully facilitate learnings which are to follow shortly (pp. 209–210)

These principles conflict with typical content placement in schools, where content is segregated by subject (e.g., social studies, mathematics, science). But Thorndike and Gates ( 1929 ) urged that knowledge and skills be taught with different subjects ( Application 3.2 ). For example, forms of government are appropriate topics not only in civics and history, but also in English (how governments are reflected in literature) and foreign language (government structure in other countries).

Mental Discipline.

APPLICATION 3.2 Sequence of Curricula

Thorndike’s views on the sequence of curricula suggest that learning should be integrated across subjects. Mrs. Woleska prepared a unit on pumpkins for her second-grade class in the fall. The students studied the significance of pumpkins to the American colonists (history), where pumpkins currently are grown (geography), and the varieties of pumpkins grown (agriculture). They measured and charted the various sizes of pumpkins (mathematics), carved the pumpkins (art), planted pumpkin seeds and studied their growth (science), and read and wrote stories about pumpkins (language arts). This approach provides a meaningful experience for children and “real life” learning of various skills.

In developing a history unit on the Civil War, Ms. Parks went beyond covering factual material and incorporated comparisons from other wars, attitudes and feelings of the populace during that time period, biographies and personalities of individuals involved in the war, and the impact the war had on the United States and implications for the future. In addition, she worked with other teachers in the middle school to expand the unit by examining the terrain of major battlefields (geography), weather conditions during major battles (science), and the emergence of literature (language arts) and creative works (art, music, drama) during that time period.

Mental discipline  is the view that learning certain subjects (e.g., the classics, mathematics) enhances general mental functioning better than learning other subjects. Mental discipline was a popular view during Thorndike’s time. He tested this idea with 8,500 students in grades 9 to 11 (Thorndike,  1924 ). Students were given intelligence tests a year apart, and their programs of study that year were compared to determine whether certain courses were associated with greater intellectual gains. The results provided no support for mental discipline. Students who had greater ability to begin with made the best progress regardless of what they studied.

·  If our inquiry had been carried out by a psychologist from Mars, who knew nothing of theories of mental discipline, and simply tried to answer the question, “What are the amounts of influence of sex, race, age, amounts of ability, and studies taken, upon the gain made during the year in power to think, or intellect, or whatever our stock intelligence tests measure,” he might even dismiss “studies taken” with the comment, “The differences are so small and the unreliabilities are relatively so large that this factor seems unimportant.” The one causal factor which he would be sure was at work would be the intellect already existent. Those who have the most to begin with gain the most during the year. (Thorndike,  1924 , p. 95)

So rather than assuming that some subject areas improve students’ mental abilities better than others, we should assess how different subject areas affect students’ ability to think, as well as other outcomes (e.g., interests, goals). Thorndike’s influential research led educators to redesign curricula away from the mental discipline idea.

CLASSICAL CONDITIONING

We have seen that events in the United States in the early 20th century helped establish psychology as a science and learning as a legitimate field of study. At the same time, there were important developments in other countries. One of the most significant was the work of Ivan Pavlov (1849–1936), a Russian physiologist who won the Nobel Prize in 1904 for his work on digestion.

Pavlov’s legacy to learning theory was his work on  classical conditioning  (Cuny,  1965 ; Hunt,  1993 ; Pavlov,  1927 ,  1928 ; Windholz,  1997 ). While Pavlov was the director of the physiological laboratory at the Institute of Experimental Medicine in Petrograd, he noticed that dogs often would salivate at the sight of the attendant bringing them food or even at the sound of the attendant’s footsteps. Pavlov deduced that the attendant was not a natural stimulus for the reflex of salivating; rather, the attendant acquired this power by being associated with food.

Basic Processes

Classical conditioning is a multistep procedure that initially involves presenting an  unconditioned stimulus  ( UCS ), which elicits an  unconditioned response  ( UCR ). Pavlov presented a hungry dog with meat powder (UCS), which would cause the dog to salivate (UCR). To condition the animal requires repeatedly presenting an initially neutral stimulus immediately before presenting the UCS. Pavlov often used a ticking metronome as the neutral stimulus. In the early trials, the ticking of the metronome produced no salivation. Eventually, the dog salivated in response to the ticking metronome prior to the presentation of the meat powder. The metronome had become a  conditioned stimulus  ( CS ) that elicited a  conditioned response  ( CR ) similar to the original UCR ( Table 3.1 ). Repeated nonreinforced presentations of the CS (i.e., without the UCS) cause the CR to diminish in intensity and disappear, a phenomenon known as  extinction (Larrauri & Schmajuk,  2008 ; Pavlov,  1932b ).

Table 3.1 Classical conditioning procedure.

Phase	Stimulus	Response
1	UCS (food powder)	UCR (salivation)
2	CS (metronome), then UCS (food powder)	UCR (salivation)
3	CS (metronome)	CR (salivation)

Spontaneous recovery  may occur after a time lapse in which the CS is not presented and the CR presumably extinguishes. If the CS then is presented and the CR returns, we say that the CR spontaneously recovered from extinction. Pairings of the CS with the UCS can restore the CR to full strength. The fact that CS–CR pairings can be reinstated without great difficulty suggests that extinction does not involve unlearning of the associations (Redish, Jensen, Johnson, & Kurth-Nelson,  2007 ).

Generalization  means that the CR occurs to stimuli similar to the CS ( Figure 3.2 ). Once a dog is conditioned to salivate in response to a metronome ticking at 70 beats per minute, it also may salivate in response to a metronome ticking faster or slower, as well as to ticking clocks or timers. The more dissimilar the new stimulus is to the CS or the fewer elements that they share, the less generalization occurs (Harris,  2006 ).

Figure 3.2 Generalization curve showing decreased magnitude of conditioned response as a function of increased dissimilarity with the conditioned stimulus.

Discrimination  is the complementary process that occurs when the dog learns to respond to the CS but not to other, similar stimuli. To train discrimination, an experimenter might pair the CS with the UCS and also present other, similar stimuli without the UCS. If the CS is a metronome ticking at 70 beats per minute, it is presented with the UCS, whereas other cadences (e.g., 50 and 90 beats per minute) are presented but not paired with the UCS.

Once a stimulus becomes conditioned, it can function as a UCS and  higher-order conditioning  can occur (Pavlov,  1927 ). If a dog has been conditioned to salivate at the sound of a metronome ticking at 70 beats per minute, the ticking metronome can function as a UCS for higher-order conditioning. A new neutral stimulus (such as a buzzer) can be sounded for a few seconds, followed by the ticking metronome. If, after a few trials, the dog begins to salivate at the sound of the buzzer, the buzzer has become a second-order CS. Conditioning of the third order involves the second-order CS serving as the UCS and a new neutral stimulus being paired with it. Pavlov ( 1927 ) reported that conditioning beyond the third order is difficult.

Higher-order conditioning is a complex process that is not well understood (Rescorla,  1972 ). The concept is theoretically interesting and might help to explain why some social phenomena (e.g., test failure) can cause conditioned emotional reactions, such as stress and anxiety. Early in life, failure may be a neutral event, but it usually becomes associated with disapproval from parents and teachers that may be an UCS eliciting anxiety. Through conditioning, failure can elicit anxiety. Cues associated with the situation also can become conditioned stimuli. Students may feel anxious when they walk into a room where they will take a test or when a teacher passes out a test.

CSs capable of producing CRs are called  primary signals . People have a  second signal system —language—that greatly expands the potential for conditioning (Windholz,  1997 ). Words or thoughts are labels denoting events or objects and can become CSs. Thinking about a test or listening to a teacher discuss a forthcoming test may cause anxiety. Thus, tests do not make students anxious but rather their linguistic representations or meanings.

Informational Variables

Pavlov believed that conditioning is an automatic process that occurs with repeated CS–UCS pairings and that nonpairings extinguish the CR. In humans, however, conditioning can occur rapidly, sometimes after only a single CS–UCS pairing. Repeated nonpairings of the CS and UCS may not extinguish the CR. Extinction seems highly context dependent (Bouton, Nelson, & Rosas,  1999 ). Reponses stay extinguished in the same context, but when the setting is changed, CRs may recur. Further, conditioning cannot occur between any two variables. Within any species, responses can be conditioned to some stimuli but not to others. Conditioning depends on the compatibility of the stimulus and response with species-specific reactions (Hollis,  1997 ). These findings call into question Pavlov’s description of conditioning.

Research subsequent to Pavlov has shown that conditioning depends less on the CS–UCS pairing and more on the extent that the CS conveys information about the likelihood of the UCS occurring (Rescorla,  1972 ,  1976 ). As an illustration, assume that one stimulus always is followed by a UCS and another stimulus sometimes is followed by it. The first stimulus should result in conditioning because it reliably predicts the onset of the UCS. It even may not be necessary to pair the CS and UCS; conditioning can occur by simply telling people that they are related (Brewer,  1974 ). Likewise, repeated CS–UCS nonpairings may not be necessary for extinction; telling people the contingency is no longer in effect can reduce or extinguish the CR.

An explanation for these results is that people form expectations concerning the probability of the UCS occurring (Rescorla,  1987 ). For a stimulus to become a CS, it must convey information to the individual about the time, place, quantity, and quality of the UCS. Even when a stimulus is predictive, it may not become conditioned if another stimulus is a better predictor. Rather than conditioning being automatic, it appears to be mediated by cognitive processes. If people do not realize there is a CS–UCS link, conditioning does not occur. When no CS–UCS link exists, conditioning can occur if people believe it does. Although this contingency view of conditioning may not be entirely accurate (Papini & Bitterman,  1990 ), it provides a different explanation for conditioning than Pavlov’s and highlights its complexity.

Conditioned Emotional Reactions

Pavlov ( 1932a ,  1934 ) applied classical conditioning principles to abnormal behavior (e.g., neuroses). His views were speculative and unsubstantiated, but classical conditioning principles have been applied by others to condition emotional reactions.

Watson claimed to demonstrate the power of emotional conditioning in the well-known Little Albert experiment (Watson & Rayner,  1920 ). Albert was an infant who showed no fear of a white rat when he was tested between the ages of 8 and 11 months. The conditioning involved a hammer being struck against a steel bar behind Albert as he reached out for the rat. “The infant jumped violently and fell forward, burying his face in the mattress” (p. 4). This sequence was immediately repeated. One week later when the rat was presented, Albert began to reach out but then withdrew his hand. The previous week’s conditioning was apparent. Tests over the next few days showed that Albert reacted emotionally to the rat’s presence. There also was generalization of fear to a rabbit, dog, and fur coat. When Albert was retested a month later with the rat, he showed a mild emotional reaction.

This study is widely cited as showing how conditioning can produce emotional reactions, but there are questions about the study’s validity. Recent evidence suggests that Albert was neurologically impaired (Bartlett,  2012 ). With this deficit, his reactions to the white rat would not be typical of a healthy child. Albert died at the age of 6 years of hydrocephalus (Beck, Levinson, & Irons,  2009 ), a condition he apparently had from birth. He never learned to walk or talk and had vision problems. Drawing conclusions from this study and generalizing its results seem problematic.

Further, the influence of conditioning usually is not that powerful (Harris,  1979 ). Classical conditioning is a complex phenomenon; one cannot condition any response to any stimulus. Species have evolved mechanisms predisposing them to being conditioned in some ways and not in others (Hollis,  1997 ). Among humans, conditioning occurs when people are aware of the relation between the CS and the UCS, and information that the UCS may not follow the CS can produce extinction. Attempts to replicate Watson and Rayner’s findings were not uniformly successful (Valentine,  1930a ).

A more reliable means of producing emotional conditioning is with  systematic desensitization , which is often used with individuals who possess debilitating fears (Wolpe,  1958 ; see  Application 3.3 ). Desensitization comprises three phases. In the first phase, a therapist and client jointly develop an anxiety hierarchy of several situations graded from least-to-most anxiety-producing for the client. For a test-anxious student, low-anxiety situations might be hearing a test announcement in class and gathering together materials to study. Situations of moderate anxiety might be studying the night before the test and walking into class on the day of the test. High-anxiety situations could include receiving the test and not knowing the answer to a question.

In the second phase, the client learns to relax by imagining pleasant scenes (e.g., lying on a beach) and cuing relaxation (saying “relax”). In the third phase, the client, while relaxed, imagines the lowest (least-anxious) scene on the hierarchy. This may be repeated several times, after which the client imagines the next scene. Treatment proceeds up the hierarchy until the client can imagine the most anxiety-producing scene without feeling anxious. If the client reports anxiety while imagining a scene, the client drops back down the hierarchy to a scene that does not produce anxiety. Treatment may require several sessions.

APPLICATION 3.3 Emotional Conditioning

Principles of classical conditioning seem relevant to some emotions. Children entering kindergarten or first grade may be fearful. At the beginning of the school year, primary teachers might develop procedures to help desensitize the fears. Visitation sessions allow students to meet their teachers and other students and see their classrooms. On the first few days of school, teachers might plan fun but relatively calm activities involving students getting to know their teachers, classmates, rooms, and school buildings. Students could tour the buildings, return to their rooms, and draw pictures. They might talk about what they saw. Students can be taken to offices to meet the principal, assistant principal, nurse, and counselor. They also could play name games in which they introduce themselves and then try to recall names of classmates.

These activities represent an informal desensitization procedure. For some children, cues associated with the school serve as stimuli eliciting anxiety. The fun activities elicit pleasurable feelings, which are incompatible with anxiety. Pairing fun activities with cues associated with school may cause the latter to become less anxiety producing.

Education students may be anxious about teaching an entire class. Anxieties should be lessened when students spend time in classrooms and gradually assume more responsibility for instruction. Pairing classroom and teaching experiences with formal study can desensitize fears related to being responsible for children’s learning.

Some drama students have stage fright. Drama teachers may work with students to lessen these anxieties by practicing more on the actual stage and by opening up rehearsals to allow others to watch. Practice in performing in front of others should help diminish their fears.

Desensitization involves counterconditioning. The relaxing scenes that one imagines (UCS) produce relaxation (UCR). Anxiety-producing cues (CS) are paired with the relaxing scenes. Relaxation is incompatible with anxiety. By initially pairing a weak anxiety cue with relaxation and by slowly working up the hierarchy, all of the anxiety-producing cues eventually should elicit relaxation (CR).

Desensitization is an effective procedure that can be accomplished in a therapist’s or counselor’s office. It does not require the client to perform the activities on the hierarchy. A disadvantage is that the client must be able to imagine scenes. People differ in their ability to form mental images. Desensitization also requires the skill of a professional therapist or counselor and should not be attempted by anyone unskilled in its application.

CONTIGUOUS CONDITIONING

Acts and Movements

Edwin R. Guthrie (1886–1959) postulated behavioral learning principles based on associations (Guthrie,  1940 ). These principles reflect the idea of contiguity of stimuli and responses:

·  A combination of stimuli which has accomplished a movement will on its recurrence tend to be followed by that movement. (Guthrie,  1952 , p. 23)

Movements  are discrete behaviors, whereas  acts  are large-scale classes of movements that produce an outcome. Playing the piano and using a computer are acts that include many movements. A particular act may be accompanied by a variety of movements; the act may not specify the movements precisely. In basketball, shooting a basket (an act) can be accomplished with a variety of movements.

Contiguity learning implies that a behavior in a situation will be repeated when that situation recurs (Guthrie,  1959 ); however, contiguity learning is selective. At any given moment, a person is confronted with many stimuli, and associations cannot be made to all of them. Rather, only a few stimuli are selected, and associations are formed between them and responses. The contiguity principle also applies to memory. Verbal cues are associated with stimulus conditions or events at the time of learning (Guthrie,  1952 ).  Forgetting  involves new learning and is due to interference in which an alternative response is made to an old stimulus.

Guthrie’s theory contends that learning occurs through pairing of stimulus and response. Guthrie ( 1942 ) also discussed the strength of the pairing, or  associative strength :

·  A stimulus pattern gains its full associative strength on the occasion of its first pairing with a response. (p. 30)

This  all-or-none  principle of learning rejects the notion of frequency, as embodied in Thorndike’s original Law of Exercise (Guthrie,  1930 ). Although Guthrie did not suggest that people learn complex behaviors (e.g., solving equations, writing papers) by performing them once, he believed that initially one or more movements become associated. Repetition of a situation adds movements, combines movements into acts, and establishes the act under different environmental conditions.

Practice links the various movements involved in the acts of solving equations and writing papers. The acts themselves may have many variations (types of equations and papers) and ideally should transfer—students should be able to solve equations and write papers in different contexts. Guthrie accepted Thorndike’s notion of identical elements. Behaviors should be practiced in the exact situations in which they will be called for (e.g., in class).

Guthrie believed that responses do not need to be rewarded to be learned. Rather, learning requires close pairing in time between stimulus and response ( contiguity ). Guthrie ( 1952 ) disputed Thorndike’s Law of Effect because satisfiers and annoyers are effects of actions; therefore, they cannot influence learning of previous connections but only subsequent ones. Rewards might help to prevent  unlearning  (forgetting) because they prevent new responses from being associated with stimulus cues.

Contiguity is a central feature of school learning. Flashcards help students learn arithmetic facts. Students learn to associate a stimulus (e.g., 4 × 4) with a response (16). Foreign-language words are associated with their English equivalents, and chemical symbols are associated with their element names.

Habit Formation and Change

Guthrie’s ideas are relevant to habit formation and change.  Habits  are learned dispositions to repeat past responses (Wood & Neal,  2007 ). Because habits are behaviors established to many cues, teachers who want students to behave well in school should link school rules with many cues. “Treat others with respect,” needs to be linked with the classroom, computer lab, halls, cafeteria, gymnasium, auditorium, and playground. By applying this rule in each of these settings, students’ respectful behaviors toward others become habitual. If students believe they have to practice respect only in the classroom, respecting others will not become a habit.

The key to changing behavior is to “find the cues that initiate the action and to practice another response to these cues” (Guthrie,  1952 , p. 115). Guthrie identified three methods for altering habits: threshold, fatigue, and incompatible response ( Table 3.2  and  Application 3.4 ).

Table 3.2 Guthrie’s methods for breaking habits.

Method	Explanation	Example
Threshold	Introduce weak stimulus. Increase stimulus, but keep it below threshold value that will produce unwanted response.	Introduce academic content in short blocks of time for children. Gradually increase session length, but not to a point where students become frustrated or bored.
Fatigue	Force child to make unwanted response repeatedly in presence of stimulus.	Give child who makes paper airplanes a stack of paper and have child make each sheet into a plane.
Incompatible response	In presence of stimulus, have child make response incompatible with unwanted response.	Pair cues associated with media center with reading rather than talking.

APPLICATION 3.4 Breaking Habits

Guthrie’s contiguity principle offers practical suggestions for how to break habits. One application of the threshold method involves the time young children spend on academic activities. Many young children have short attention spans, which limit how long they can sustain work on one activity. Most class activities are scheduled to last no longer than 30–40 minutes. However, at the start of the school year, attention spans quickly wane for many children. To apply Guthrie’s theory, a teacher might, at the start of the year, limit activities to 15–20 minutes. Over the next few weeks the teacher could gradually increase the time students spend working on a single activity.

The threshold method also can be applied to teaching printing. When children first learn to form letters, their movements are awkward and they lack fine motor coordination. The distances between lines on a page are purposely wide so children can fit the letters into the space. If paper with narrower lines were initially introduced, students’ letters would spill over the borders and students might become frustrated. Once students can form letters within the wider lines, they can use paper with narrower lines to help them refine their skills.

Teachers need to be judicious when using the fatigue method because of potential negative consequences. Jason likes to make paper airplanes and sail them across the room. His teacher might remove him from the classroom, give him a large stack of paper, and tell him to start making paper airplanes. After Jason has made several airplanes, the activity should lose its attraction and paper will no longer be a cue for him to make airplanes.

Some students like to race around the gym when they first enter their physical education class. To employ the fatigue method, the physical education teacher might just let these students keep running after the class has begun. Soon they will tire and quit running.

The incompatible response method can be used with students who talk and misbehave in the media center. Reading is incompatible with talking. The media center teacher might ask the students to find interesting books and read them while in the center. Assuming that the students find the books enjoyable, the media center will, over time, become a cue for selecting and reading books rather than for talking with other students.

A social studies teacher has some students who regularly do not pay attention in class. The teacher realized that using many slides while lecturing was boring. Soon the teacher began to incorporate other elements into each lesson, such as experiments, video clips, and debates, in an attempt to involve students and raise their interest in the course.

In the  threshold  method, the cue (stimulus) for the habit to be changed (the undesired response) is introduced at such a weak level that it does not elicit the response; it is below the threshold level of the response. Gradually the stimulus is introduced at greater intensity until it is presented at full strength. Were the stimulus introduced at its greatest intensity, the response would be the behavior that is to be changed (the habit). For example, some children react to the taste of spinach by refusing to eat it. To alter this habit, parents might introduce spinach in small bites or mixed with a food that the child enjoys. Over time, the amount of spinach the child eats can be increased.

In the  fatigue  method, the cue for engaging in the behavior is transformed into a cue for avoiding it. Here the stimulus is introduced at full strength and the individual performs the undesired response until he or she becomes exhausted. The stimulus becomes a cue for not performing the response. To alter a child’s behavior of repeatedly throwing toys, parents might make the child throw toys until it is no longer fun (some limits are needed!).

In the  incompatible response  method, the cue for the undesired behavior is paired with a response incompatible with the undesired response; that is, the two responses cannot be performed simultaneously. The response to be paired with the cue must be more attractive to the individual than the undesired response. The stimulus becomes a cue for performing the alternate response. To stop snacking while watching TV, people should keep their hands busy (e.g., sew, paint, work puzzles). Over time, watching TV becomes a cue for engaging in an activity other than snacking. Systematic desensitization (described earlier) also makes use of incompatible responses.

Punishment is ineffective in altering habits (Guthrie,  1952 ). Punishment following a response cannot affect the stimulus–response association. Punishment given while a behavior is being performed may disrupt or suppress the habit but not change it. Punishment does not establish an alternate response to the stimulus. It is better to alter negative habits by replacing them with desirable ones (i.e., incompatible responses).

Guthrie’s theory does not include cognitive processes. Although it is not a viable learning theory today, its emphasis on contiguity is timely because current theories stress contiguity. Cognitive theories predict that learning requires understanding the relationship between a stimulus (situation, event) and the appropriate response. Guthrie’s ideas about changing habits provide general guidance for developing better habits.

OPERANT CONDITIONING

A well-known behavior theory is  operant conditioning , formulated by B. F. (Burrhus Frederic) Skinner (1904–1990). Beginning in the 1930s, Skinner published a series of papers on laboratory studies with animals in which he identified the components of operant conditioning. He summarized this early work in his influential book, The Behavior of Organisms (Skinner,  1938 ).

Skinner applied his ideas to human functioning. Early in his career, he became interested in education and developed teaching machines and programmed instruction. The Technology of Teaching (Skinner,  1968 ) addresses instruction, motivation, discipline, and creativity. In 1948 he published Walden Two, which describes how behavioral principles can be applied to create a utopian society. Skinner ( 1971 ) addressed the problems of modern life and advocated applying a behavioral technology to the design of cultures in Beyond Freedom and Dignity. Skinner and others have applied operant conditioning principles to school learning and discipline, child development, language acquisition, social behaviors, mental illness, medical problems, substance abuse, and vocational training (DeGrandpre,  2000 ; Karoly & Harris,  1986 ; Morris,  2003 ).

As a young man, Skinner aspired to be a writer (Skinner,  1970 ):

·  I built a small study in the attic and set to work. The results were disastrous. I frittered away my time. I read aimlessly, built model ships, played the piano, listened to the newly-invented radio, contributed to the humorous column of a local paper but wrote almost nothing else, and thought about seeing a psychiatrist. (p. 6)

He became interested in psychology after reading Pavlov’s ( 1927 ) Conditioned Reflexes and Watson’s ( 1924 )  Behaviorism . His subsequent career had a profound impact on the psychology of learning.

Despite his admission that “I had failed as a writer because I had had nothing important to say” (Skinner,  1970 , p. 7), he was a prolific writer who channeled his literary aspirations into scientific writing that spanned six decades (Lattal,  1992 ). His dedication to his profession is evident in his giving an invited address at the American Psychological Association convention eight days before he died (Holland,  1992 ; Skinner,  1990 ). The association honored him with a special issue of its monthly journal, American Psychologist (American Psychological Association,  1992 ). Although his theory has been discredited by current learning theorists because it cannot adequately explain higher-order and complex forms of learning (Bargh & Ferguson,  2000 ), his influence continues as operant conditioning principles are commonly applied to enhance student learning and behavior (Morris,  2003 ). In the opening scenario, for example, Leo employs operant conditioning principles to address student misbehavior. Emily and Shayna, on the other hand, argue for the importance of cognitive factors.

Conceptual Framework

This section discusses the assumptions underlying operant conditioning, how it reflects a functional analysis of behavior, and the implications of the theory for the prediction and control of behavior. Operant conditioning theory is complex (Dragoi & Staddon,  1999 ); its principles most relevant to human learning are covered in this chapter.

Scientific Assumptions.

Pavlov viewed behavior as a manifestation of neurological functioning. Skinner ( 1938 ) did not deny this but believed a psychology of behavior can be understood without reference to neurological or other internal events.

He raised similar objections to the unobservable processes and entities proposed by cognitive views of learning (Overskeid,  2007 ).  Private events  are internal responses accessible only to the individual and can be studied through people’s verbal reports, which are forms of behavior (Skinner,  1953 ). Skinner did not deny the existence of attitudes, beliefs, opinions, desires, and other forms of self-knowledge (he, after all, had them), but rather qualified their role.

People do not experience consciousness or emotions but rather their bodies, and internal reactions are responses to internal stimuli (Skinner,  1987 ). A further problem with internal processes is that translating them into language is difficult, because language does not completely capture the dimensions of an internal experience (e.g., pain). Much of what is called “knowing” involves using language ( verbal behavior ). Thoughts are types of behavior that are brought about by other stimuli (environmental or private) and that give rise to responses (overt or covert). When private events are expressed as overt behaviors, their role in a functional analysis can be determined.

Functional Analysis of Behavior.

Skinner ( 1953 ) referred to his theory as a  functional analysis :

·  The external variables of which behavior is a function provide for what may be called a causal or functional analysis. We undertake to predict and control the behavior of the individual organism. This is our “dependent variable”—the effect for which we are to find the cause. Our “independent variables”—the causes of behavior—are the external conditions of which behavior is a function. Relations between the two—the “cause-and-effect relationships” in behavior—are the laws of a science. A synthesis of these laws expressed in quantitative terms yields a comprehensive picture of the organism as a behaving system. (p. 35)

Learning  is “the reassortment of responses in a complex situation”; conditioning refers to “the strengthening of behavior which results from reinforcement” (Skinner,  1953 , p. 65). There are two types of conditioning: Type S and Type R.  Type S  is Pavlovian conditioning, characterized by the pairing of the reinforcing (unconditioned) stimulus with another (conditioned) stimulus. The S calls attention to the importance of the stimulus in eliciting a response from the organism. The response made to the eliciting stimulus is known as  respondent behavior .

Although Type S conditioning may explain conditioned emotional reactions, most human behaviors are emitted in the presence of stimuli rather than automatically elicited by them. Responses are controlled by their consequences, not by antecedent stimuli. This type of behavior, which Skinner termed  Type R  to emphasize the response aspect, is  operant behavior  because it operates on the environment to produce an effect.

·  If the occurrence of an operant is followed by presentation of a reinforcing stimulus, the strength is increased…. If the occurrence of an operant already strengthened through conditioning is not followed by the reinforcing stimulus, the strength is decreased. (Skinner,  1938 , p. 21)

We might think of operant behavior as “learning by doing,” and in fact much learning occurs when we perform behaviors (Lesgold,  2001 ). Unlike respondent behavior, which prior to conditioning does not occur, the probability of occurrence of an operant is never zero because the response must be made for reinforcement to be provided. Reinforcement changes the likelihood or rate of occurrence of the response. Operant behaviors act upon their environments and become more or less likely to occur because of reinforcement.

Basic Processes

This section examines the basic processes in operant conditioning: reinforcement, extinction, primary and secondary reinforcers, the Premack Principle, punishment, schedules of reinforcement, generalization, and discrimination.

Reinforcement.

Reinforcement  is responsible for response strengthening—increasing the rate of responding or making responses more likely to occur. A reinforcer (or  reinforcing stimulus ) is any stimulus or event following a response that leads to response strengthening. Reinforcers are defined based on their effects, which do not depend upon mental processes such as consciousness, intentions, or goals (Schultz,  2006 ). Because reinforcers are defined by their effects, they cannot be determined in advance.

·  The only way to tell whether or not a given event is reinforcing to a given organism under given conditions is to make a direct test. We observe the frequency of a selected response, then make an event contingent upon it and observe any change in frequency. If there is a change, we classify the event as reinforcing to the organism under the existing conditions. (Skinner,  1953 , pp. 72–73)

Reinforcers are situationally specific: They apply to individuals at given times under given conditions. What is reinforcing to Maria during reading now may not be during mathematics now or reading later. Despite this specificity, stimuli or events that reinforce behavior can be predicted (Skinner,  1953 ). Students typically find reinforcing such events as teacher praise, free time, privileges, stickers, and high grades. Nonetheless, one never can know for certain whether a consequence is reinforcing until it is presented after a response and we see whether behavior changes.

The basic operant model of conditioning is the  three-term contingency :

SD → R → SR

A  discriminative stimulus  (SD) sets the occasion for a response (R) to be emitted, which is followed by a  reinforcing stimulus  (SR, or  reinforcement ). The reinforcing stimulus is any stimulus (event, consequence) that increases the probability the response will be emitted in the future when the discriminative stimulus is present. In more familiar terms, we might label this the A-B-C model:

A (Antecedent) → B (Behavior) → C (Consequence)

Positive reinforcement involves presenting a stimulus, or adding something to a situation, following a response, which increases the future likelihood of that response occurring in that situation. A  positive reinforcer  is a stimulus that, when presented following a response, increases the future likelihood of the response occurring in that situation. In the opening scenario, Leo uses points as positive reinforcers for good behavior ( Table 3.3 ).

Negative reinforcement involves removing a stimulus, or taking something away from a situation following a response, which increases the future likelihood that the response will occur in that situation. A  negative reinforcer  is a stimulus that, when removed by a response, increases the future likelihood of the response occurring in that situation. Some stimuli that often function as negative reinforcers are bright lights, loud noises, criticism, annoying people, and low grades, because behaviors that remove them tend to be reinforcing. Positive and negative reinforcement have the same effect: They increase the likelihood that the response will be made in the future in the presence of the stimulus.

To illustrate these processes ( Table 3.3 ), assume that a teacher is holding a question-and-answer session with the class. The teacher asks a question (SD or A), calls on a student volunteer who gives the correct answer (R or B), and says to the student “That’s good” (SR or C). If volunteering by this student increases, saying “That’s good” is a positive reinforcer and this is an example of positive reinforcement because volunteering increased. Now assume that after a student gives the correct answer the teacher tells the student he or she does not need to do the homework. If volunteering by this student increases, homework is a negative reinforcer and this is an example of negative reinforcement because removing the homework increased volunteering.  Application 3.5  gives other examples of positive and negative reinforcement.

Table 3.3 Reinforcement and punishment processes.

SD→ Discriminative Stimulus	R → Response	S R Reinforcing (Punishing) Stimulus
Positive Reinforcement (Present positive reinforcer)
T asks question	S volunteers *	T says to S, “That’s good”
Negative Reinforcement (Remove negative reinforcer)
T asks question	S volunteers	T says S does not have to do homework
Punishment (Present negative reinforcer)
T asks question	S misbehaves	T assigns S homework
Punishment (Remove positive reinforcer)
T asks question	S misbehaves	T says S will miss free time

* T refers to teacher and S to student.

A positive reinforcer commonly used is praise (e.g., “Good job!”). Indeed, as people typically enjoy receiving praise, it usually functions as a positive reinforcer. However, praise is not the same as task feedback informing students how they performed. When the two are combined (e.g., “Good job! Your answer is correct.”), it is difficult to know which had the stronger effect on subsequent behavior. A danger in combining the two is that with corrective feedback (e.g., “Good job! But you still need to work on this part.”) students may attend more to the praise and miss the part needing correction (Hattie,  2012 ). If praise is used in conjunction with correction, it is best to ensure that students understand what they need to improve.

Extinction.

Extinction  involves the decline of response strength due to nonreinforcement. Students who raise their hands in class but never get called on may stop raising their hands. People who send many e-mails to the same individual but never receive a reply eventually may quit sending e-mails to that person.

How rapidly extinction occurs depends on the  reinforcement history  (Skinner,  1953 ). Extinction occurs quickly if few preceding responses have been reinforced. Responding is much more durable with a lengthier history of reinforcement. Extinction is not the same as  forgetting . Responses that extinguish can be performed but are not because of lack of reinforcement. In the preceding examples, the students still know how to raise their hands and the people still know how to send e-mails. Forgetting involves a true loss of conditioning over time in which the opportunities for responding have not been present.

APPLICATION 3.5 Positive and Negative Reinforcement

Teachers can use positive and negative reinforcement to motivate students to master skills and spend more time on task. For example, while teaching concepts in a science unit, Mrs. Davos might ask students to complete questions at the end of the chapter. She also might set up activity centers around the room that involve hands-on experiments related to the lesson. Students would circulate and complete the experiments contingent on their successfully answering the chapter questions (positive reinforcement). This contingency reflects the Premack Principle of providing the opportunity to engage in a more-valued activity (experiments) as a reinforcer for engaging in a less-valued one (completing chapter questions). Students who complete 80% of the questions correctly and who participate in a minimum of two experiments do not have to complete homework. This would function as negative reinforcement to the extent that students perceive homework as a negative reinforcer.

A middle school counselor working with Penny to improve classroom behavior could have each of her teachers rate her class behavior for that day as acceptable or unacceptable. For each “acceptable” Penny receives 1 minute to work on a computer (positive reinforcement for Penny). At the end of the week Penny can use the earned computer time following lunch. Further, if she earns a minimum of 15 minutes in the lab, she does not have to take a behavior note home to be signed by parents (this assumes Penny perceives a behavior note as a negative reinforcer).

Primary and Secondary Reinforcers.

Stimuli such as food, water, and shelter are called  primary reinforcers  because they are necessary for survival.  Secondary reinforcers  are stimuli that become conditioned through their association with primary reinforcers. A child’s favorite milk glass becomes secondarily reinforcing through its association with milk (a primary reinforcer). A secondary reinforcer that becomes paired with more than one primary reinforcer is a  generalized reinforcer . People work long hours to earn money (a generalized reinforcer), which they use to buy many reinforcers (e.g., food, housing, TVs, vacations).

Operant conditioning explains the development and maintenance of much social behavior with generalized reinforcers. Children may behave in ways to draw adults’ attention. Attention is reinforcing because it is paired with primary reinforcers from adults (e.g., food, water, protection). Important educational generalized reinforcers are teachers’ praise, high grades, privileges, honors, and degrees. These reinforcers often are paired with other generalized reinforcers, such as approval (from parents and friends) and money (a college degree leads to a good job).

Premack Principle.

Recall that we label a behavioral consequence as reinforcing only after we apply it and see how it affects future behavior. It seems troubling that we must use common sense or trial and error in choosing reinforcers because we cannot know in advance whether a consequence will function as a reinforcer.

Premack ( 1962 ,  1971 ) described a means for ordering reinforcers that allows prediction. The  Premack Principle  says that the opportunity to engage in a more valued activity reinforces engaging in a less valued activity, where “value” is defined in terms of the amount of responding or time spent on the activity in the absence of reinforcement. If a contingency is arranged such that the value of the second (contingent) event is higher than the value of the first (instrumental) event, an increase will be expected in the probability of occurrence of the first event (the reward assumption). If the value of the second event is lower than that of the first event, the likelihood of occurrence of the first event ought to decrease (the punishment assumption).

Suppose that a child is allowed to choose between working on an art project, going to the media center, reading a book, or using a computer. Over the course of 10 such choices the child works on an art project once, goes to the media center 3 times, never reads a book, and uses a computer 6 times. For this child, the opportunity to use a computer is valued the most. To apply the Premack Principle, a teacher might say to the child, “After you finish reading this book, you can use a computer.” Considerable empirical evidence supports Premack’s ideas, especially with respect to the reward assumption (Dunham,  1977 ).

The Premack Principle offers guidance for selecting effective reinforcers: Observe what people do when they have a choice, and order those behaviors in terms of likelihood. The order is not permanent, since the value of reinforcers can change. Any reinforcer, when applied often, can result in  satiation  and lead to decreased responding. Teachers who employ the Premack Principle need to check students’ preferences periodically by observing them and asking what they like to do. Determining in advance which reinforcers are likely to be effective in a situation is critical in planning behavioral change (Timberlake & Farmer-Dougan,  1991 ).

Punishment.

Punishment  decreases the future likelihood of responding to a stimulus. Punishment may involve withdrawing a positive reinforcer or presenting a negative reinforcer following a response, as shown in  Table 3.3 . Assume that during a question-and-answer session a student misbehaves when the teacher asks a question (and maybe is not watching; teacher asks question = SD or A; misbehavior = R or B). The teacher spots the misbehavior and assigns the student homework (SR or C). If the student stops misbehaving, assigning homework operates as a negative reinforcer and this is an example of punishment because assigning the homework decreased misbehavior. But note that from the teacher’s perspective, this is an example of negative reinforcement (misbehavior = SD or A; assigning homework = R or B; end of misbehavior = SR or C). Since the teacher was negatively reinforced, the teacher may be likely to assign homework in response to student misbehavior.

Instead of assigning homework, assume that the teacher takes away the student’s free time. If the student’s misbehavior stops, free time operates as a positive reinforcer and this is an example of punishment because the loss of free time stops the misbehavior. As before, the cessation of student misbehavior is negatively reinforcing for the teacher.

Punishment suppresses a response but does not eliminate it; when the threat of punishment is removed, the punished response may return. The effects of punishment are complex (Skinner,  1953 ). Spanking a child for misbehaving may produce guilt and fear, which can suppress misbehavior. If the child misbehaves in the future, the conditioned guilt and fear may reappear and stop the child from misbehaving. Punishment also conditions responses that lead one to escape or avoid punishment. Students whose teacher criticizes incorrect answers soon learn to avoid volunteering answers. Punishment can condition maladaptive behaviors, because punishment does not teach how to behave more productively. Punishment can further hinder learning by creating a conflict such that the individual vacillates between responding one way or another. If a teacher sometimes criticizes students for incorrect answers and sometimes does not, students never know when criticism is forthcoming. Such variable behavior can have emotional byproducts (e.g., fear, anger) that interfere with learning.

Common school punishments are loss of privileges, removals from the classroom, in and out-of-school suspensions, and expulsions (Maag,  2001 ). Yet there are alternatives to punishment ( Table 3.4 ). One is to change the discriminative stimuli for negative behavior. For example, a student seated in the back of the room may misbehave. Teachers can change the discriminative stimuli by moving the disruptive student to the front of the class. Another alternative is to allow the unwanted behavior to continue until the perpetrator becomes satiated, which is similar to Guthrie’s fatigue method. A parent may allow a child throwing a tantrum to continue to throw it until he or she becomes fatigued. A third alternative is to extinguish an unwanted behavior by ignoring it. This may work well with minor misbehaviors (e.g., students whispering to one another), but when classrooms become disruptive, teachers need to act in other ways. A fourth alternative is to condition incompatible behavior with positive reinforcement. Teacher praise for productive work habits helps condition those habits. The primary advantage of this alternative over punishment is that it shows the student how to behave adaptively.

Schedules of Reinforcement.

Schedules refer to when reinforcement is applied (Ferster & Skinner,  1957 ; Skinner,  1938 ; Zeiler,  1977 ). A continuous schedule involves reinforcement for every correct response. This may be desirable for a short period while skills are being acquired.  Continuous reinforcement  helps to ensure that incorrect responses are not learned.

Table 3.4 Alternatives to punishment.

Alternative	Example
Change the discriminative stimuli	Move misbehaving student away from other misbehaving students.
Allow the unwanted behavior to continue	Have student who stands when he or she should be sitting continue to stand.
Extinguish the unwanted behavior	Ignore minor misbehavior so that it is not reinforced by teacher attention.
Condition an incompatible behavior	Reinforce learning progress, which occurs only when student is not misbehaving.

An intermittent schedule involves reinforcing some but not all correct responses.  Intermittent reinforcement  is common in classrooms, because usually it is not possible for teachers to reinforce each student for every correct or desirable response. Students are not called on every time they raise their hands, are not praised after working each problem, and are not constantly told they are behaving appropriately.

Intermittent schedules are defined in terms of time or number of responses. An  interval schedule  involves reinforcing the first correct response after a specific time period. In a fixed-interval (FI) schedule, the time interval is constant from one reinforcement to the next. An FI5 schedule means that reinforcement is delivered for the first response made after 5 minutes. Students who receive 30 minutes of free time every Friday (contingent on good behavior during the week) are operating under a fixed-interval schedule. In a variable-interval (VI) schedule, the time interval varies from occasion to occasion around some average value. A VI5 schedule means that on the average, the first correct response after 5 minutes is reinforced, but the time interval varies (e.g., 2, 3, 7, or 8 minutes). Students who receive 30 minutes of free time (contingent on good behavior) on an average of once a week, but not necessarily on the same day each week, are operating under a variable-interval schedule.

A  ratio schedule  depends on the number of correct responses or rate of responding. In a fixed-ratio (FR) schedule, every nth correct response is reinforced, where n is constant. An FR10 schedule means that every 10th correct response receives reinforcement. In a variable-ratio (VR) schedule, every nth correct response is reinforced, but the value varies around an average number n. A teacher may give free time after every fifth workbook assignment is completed (FR5) or periodically around an average of five completed assignments (VR5).

Figure 3.3 Patterns of responding under different reinforcement schedules.

Note: VR = variable ratio; FR = fixed ratio; FI = fixed interval; VI = variable interval.

Reinforcement schedules produce characteristic patterns of responding, as shown in  Figure 3.3 . Ratio schedules often produce higher response rates than interval schedules, but a limiting factor is fatigue due to rapid responding. Fixed-interval schedules produce a scalloped pattern. Responding drops off immediately after reinforcement but picks up toward the end of the interval between reinforcements. The variable-interval schedule produces a steady rate of responding. Unannounced quizzes operate on variable-interval schedules and help keep students studying regularly. Intermittent schedules are more resistant to extinction than continuous schedules: When reinforcement is discontinued, responding continues for a longer time if reinforcement has been intermittent rather than continuous. The durability of intermittent schedules can be seen in people’s persistence at such events as playing slot machines, fishing, and shopping for bargains.

Generalization.

Once a certain response occurs regularly to a given stimulus, the response also may occur to other stimuli.  Generalization  (Skinner,  1953 ) seems to be a problem for operant theory, because a response should not be made in a situation in which it never has been reinforced. Skinner explained generalization by noting that people perform many behaviors that lead to the final (reinforced) response. These component behaviors are often part of the chains of behavior of different tasks and therefore are reinforced in different contexts. When people are in a new situation, they are likely to perform the component behaviors, which produce an accurate response or rapid acquisition of the correct response.

For example, students with good academic habits typically come to class, attend to and participate in the activities, take notes, do the required reading, and keep up with the assignments. These component behaviors produce high achievement and grades. When such students begin a new class, it is not necessary that the content be similar to previous classes in which they have been enrolled. Rather, the component behaviors have received repeated reinforcement and thus are likely to generalize to the new setting.

Generalization, however, does not occur automatically. O’Leary and Drabman ( 1971 ) noted that generalization “must be programmed like any other behavioral change” (p. 393). One problem with many behavior modification programs is that they change behaviors but the new behaviors do not generalize outside the training context. O’Leary and Drabman ( 1971 ) offer suggestions on ways to facilitate generalization ( Table 3.5  and  Application 3.6 ).

Discrimination.

Discrimination , the complementary process to generalization, involves responding differently (in intensity or rate) depending on the stimulus or features of a situation (Rilling,  1977 ). Although teachers want students to generalize what they learn to other situations, they also want them to respond discriminately. In solving mathematical word problems, teachers might want students to adopt a general problem-solving approach comprising steps such as determining the given and the needed information, drawing a picture, and generating useful formulas. Teachers also want students to learn to discriminate problem types (e.g., area, time-rate-distance, interest rate). Being able to identify quickly the type of problem enhances students’ successes.

Teaching discrimination requires that desired responses be reinforced and unwanted responses extinguished by nonreinforcement. Teachers can highlight similarities and differences among similar content and provide for periodic reviews to ensure that students discriminate properly and apply correct problem–solution methods.

Table 3.5 Suggestions for facilitating generalization.

Parental Involvement:	Involve parents in behavioral change programs.
High Expectations:	Convey to students that they are capable of performing well.
Self-Evaluation:	Teach students to monitor and evaluate their behaviors.
Contingencies:	Withdraw artificial contingencies (e.g., points), and replace with natural ones (privileges).
Participation:	Allow students to participate in specifying behaviors to be reinforced and reinforcement contingencies.
Academics:	Provide a good academic program because many students with behavior problems have academic deficiencies.
Benefits:	Show students how behavioral changes will benefit them by linking changes to activities of interest.
Reinforcement:	Reinforce students in different settings to reduce discrimination between reinforced and nonreinforced situations.
Consistency:	Prepare teachers in regular classes to continue to reinforce behaviors of students in special classes after they are mainstreamed.

Errors can be disruptive and produce learning of incorrect responses, which suggests that students’ errors should be minimized. But eliminating all errors may not be desirable. Motivation research shows that students who learn to deal with errors adaptively subsequently persist longer on difficult tasks than do students who have experienced errorless learning (Dweck,  1975 ;  Chapter 8 ).

APPLICATION 3.6 Generalization

Generalization can advance skill development across subject areas. Finding main ideas in text is relevant to language arts, social studies, mathematics (word problems), and other content areas. A language arts teacher might provide students with a strategy for finding main ideas. Once students master this strategy, the teacher explains how to modify its use for other academic subjects and asks students to think of uses. By teaching the strategy well in one domain and facilitating potential applications in other domains, teachers save much time and effort because they do not have to teach the strategy in each content area.

Teaching expected behaviors (e.g., walking in the hall, raising a hand to speak) can also be generalized. For example, if all seventh-grade teachers decide to have students use the same academic planner, it could be explained in one class. Then students can be asked to use the same planner in each of their other classes, adapting entries as needed.

Behavior Change

Reinforcement can be given for correct responses only when people know what to do. Often, however, operant responses do not exist in final, polished form. If teachers wait to deliver reinforcement until learners emit the proper responses, many learners would never receive reinforcement because they do not acquire the responses. We now discuss how behavior change occurs in operant conditioning, which has important implications for learning.

Successive Approximations (Shaping).

· The basic operant conditioning method of behavioral change is  shaping , or differential reinforcement of  successive approximations  to the desired form or rate of behavior (Morse & Kelleher,  1977 ). To shape behavior, one follows this sequence:

· ■ Identify what the student can do now (initial behavior)

· ■ Identify the desired behavior

· ■ Identify potential reinforcers in the student’s environment

· ■ Break the desired behavior into small substeps to be mastered sequentially

· ■ Move the student from the initial behavior to the desired behavior by successively reinforcing each approximation to the desired behavior

Shaping is learning by doing with corrective feedback. A natural instance of shaping can be seen in a student attempting to shoot a basketball from a point on the court. The first shot falls short of the basket. The student shoots harder the second time, and the ball hits the backboard. The student does not shoot quite as hard the third time, and the ball hits the right rim and bounces off. On the fourth attempt, the student shoots as hard as the third attempt but aims left. The ball hits the left rim and bounces off. Finally, the student shoots just as hard but aims slightly to the right, and the ball goes into the basket. Gradually the shot is honed to an accurate form.

Shaping might be applied systematically with a student who can work on a task for only a few minutes before becoming distracted. The goal is to shape the student’s behavior so she can work uninterrupted for 30 minutes. Initially the teacher delivers a reinforcer when the student works productively for 2 minutes. After several successful 2-minute intervals, the criterion for reinforcement is raised to 3 minutes. Assuming that she works uninterrupted for several 3-minute periods, the criterion is raised to 4 minutes. This process continues to the goal of 30 minutes as long as the student reliably performs at the criterion level. If the student encounters difficulty at any point, the criterion for reinforcement decreases to a level at which she can perform successfully.

Chaining.

Most human actions are complex and include several three-term contingencies (A-B-C) linked successively. For example, shooting a basketball requires dribbling, turning, getting set in position, jumping, and releasing the ball. Each response alters the environment, and this altered condition serves as the stimulus for the next response.  Chaining  is the process of producing or altering some of the variables that serve as stimuli for future responses (Skinner,  1953 ). A chain consists of a series of operants, each of which sets the occasion for further responses.

Consider a student solving an algebraic equation (e.g., 2x − 10 = 4). The −10 serves as the SD, to which the student makes the appropriate response (R, add 10 to both sides of the equation). This product (2x = 14) is the SR and also the SD for the next response (divide both sides of the equation by 2) to solve the equation (x= 7). This stimulus serves as the SD to move to the next equation. Operations within each equation constitute a chain, and the entire problem set constitutes a chain.

Chains are similar to Guthrie’s acts, whereas individual three-term contingencies resemble movements. Many chains are integrated sequences such that successful implementation of the chain defines a skill. When skills are well honed, execution of the chain occurs automatically. Riding a bicycle consists of several discrete acts, yet an accomplished rider executes these with little or no conscious effort. Such automaticity is evident often in cognitive skills (e.g., reading, solving mathematical problems).

Behavior Modification

Behavior modification  (or  behavior therapy ) refers to the systematic application of behavior principles to facilitate adaptive behaviors (Ullmann & Krasner,  1965 ). Behavior modification has been employed with adults and children in such diverse contexts as classrooms, counseling settings, prisons, and mental hospitals. It has been used to treat phobias, dysfunctional language, disruptive behaviors, negative social interactions, poor child rearing, and low self-control (Ayllon & Azrin,  1968 ; Becker,  1971 ; Keller & Ribes-Inesta,  1974 ; Ulrich, Stachnik, & Mabry,  1966 ). Lovaas ( 1977 ) successfully employed behavior modification to teach language to children with autism. Classroom applications are given in  Application 3.7 .

APPLICATION 3.7 Behavior Modification

Behavior modification for disruptive students is difficult because such students may display few appropriate responses to be positively reinforced. Ms. Tiebout has been having problems with Erik, who pushes and shoves other students when the class gets in line to go somewhere. When the class is going only a short distance, Ms. Tiebout could inform Erik that if he stays in line without pushing and shoving, he will be the line leader on the way back to the class; however, if he pushes or shoves, he immediately will be removed from the line. This procedure can be repeated until Erik can handle short distances. Ms. Tiebout then can allow him to walk with the class for progressively longer distances until he can behave in line for any distance.

Sarah, another child in Ms. Tiebout’s class, frequently turns in messy work. Ms. Tiebout might use generalized reinforcers such as special stickers (exchangeable for various privileges) to help Sarah, whose work typically is dirty, torn, and barely readable. Sarah is told if she turns in a paper that is clean, she can earn one sticker; if it is not torn, another sticker; and if the writing is neat, a third sticker. Once Sarah begins to make improvements, Ms. Tiebout gradually can move the rewards to other areas for improvement (e.g., correct work, finishing work on time).

Techniques.

The basic techniques of behavior modification include reinforcement of desired behaviors and extinction of undesired ones. Punishment is rarely employed but, when used, more often involves removing a positive reinforcer rather than presenting a negative reinforcer.

In deciding on a program of change, behavior modifiers typically focus on the following three issues (Ullmann & Krasner,  1965 ):

· ■ Which of the individual’s behaviors are maladaptive, and which should be increased (decreased)?

· ■ What environmental contingencies currently support the individual’s behaviors (either to maintain undesirable behaviors or to reduce the likelihood of performing more adaptive responses)?

· ■ What environmental features can be altered to change the individual’s behavior?

Change is most likely when modifiers and clients agree that a change is needed and jointly decide on the desired goals. The first step in establishing a program is to define the problem in behavioral terms. For example, the statement, “Keith is out of his seat too often,” refers to overt behavior that can be measured: One can keep a record of the amount of time that Keith is out of his seat. General expressions referring to unobservables (“Keith has a bad attitude”) do not allow for objective problem definition.

The next step is to determine the reinforcers maintaining undesirable behavior. Perhaps Keith is getting teacher attention only when he gets out of his seat and not when he is seated. A simple plan is to have the teacher attend to Keith while he is seated and engaged in academic work and to ignore him when he gets out of his seat. If the amount of times that Keith is out of his seat decreases, teacher attention is a positive reinforcer.

A behavior modification program might employ such generalized reinforcers as points that students exchange for  backup reinforcers , such as tangible rewards, free time, or privileges. Having more than one backup ensures that at least one will be effective for each student at all times. A behavioral criterion must be established to earn reinforcement. The five-step shaping procedure (discussed previously) can be employed. The criterion is initially defined at the level of initial behavior and progresses in small increments toward the desired behavior. A point is given to the student each time the criterion is satisfied. To extinguish any undesirable behavior by Keith, the teacher should not give him too much attention if he gets out of his seat, but rather should inform him privately that because he does not satisfy the criterion, he does not earn a point.

Punishment is used infrequently but may be needed when behavior becomes so disruptive that it cannot be ignored (e.g., fighting). A common punishment is  time-out  (from reinforcement). During time-out, the student is removed from the class social context. There the student continues to engage in academic work without peer social interaction or the opportunity to earn reinforcement. Another punishment is to remove positive reinforcers (e.g., free time, recess, privileges) for misbehavior.

Critics have argued that behavior modification shapes quiet and docile behaviors (Winett & Winkler,  1972 ). Although too much noise may disrupt learning, it is not necessary to have a quiet classroom at all times. Some noise from social interactions can facilitate learning. The use of behavior modification is inherently neither good nor bad. It can produce overly quiet classrooms or promote social interactions (Strain, Kerr, & Ragland, 1981). Like the techniques themselves, the goals of behavior modification need to be thought out carefully by those implementing the procedures.

Cognitive Behavior Modification.

Researchers also have incorporated cognitive elements into behavior modification procedures. In  cognitive behavior modification , learners’ thoughts (when verbalized) function as discriminative and reinforcing stimuli. Thus, learners may verbally instruct themselves what to do and then perform the appropriate behavior. Cognitive behavior modification techniques often are applied with students with handicaps (Hallahan, Kneedler, & Lloyd,  1983 ), and used to reduce hyperactivity and aggression (Robinson, Smith, Miller, & Brownell,  1999 ). Meichenbaum’s ( 1977 )  self-instructional training  is an example of cognitive behavior modification (see  Chapter 4 ).

Contemporary Perspective

With their emphasis on observable events, behavior theories seem mechanistic: Set the proper stimuli in the environment and the responses will occur. Although people experience internal events (e.g., thoughts, feelings), these are not necessary to explain behavior.

These assumptions have been repeatedly challenged, especially by cognitive theorists. But it is not necessary to completely reject behavior theories in favor of cognitive ones. As noted in this chapter, behavior principles can be applied without wholly subscribing to conditioning theories. For example, establishing a conducive learning environment and reinforcing students for learning are desirable regardless of one’s theoretical perspective.

There is increasing evidence that principles of conditioning do not operate in a completely mechanistic fashion. Recall the research by Rescorla ( 1987 ) showing that for classical conditioning to occur people must form expectations—which are cognitive beliefs—about the likelihood of the UCS following the CS. CSs that are good predictors of UCSs are most apt to produce conditioning.

Recent research also has investigated the nature of voluntary acts. Skinner ( 1953 ) contended that operant behaviors were voluntary behaviors emitted in the presence of discriminative stimuli. When followed by reinforcement, the probability of occurrence of such operants increases in the future in the presence of these discriminative stimuli.

The notion of voluntary acts seems somewhat incompatible with operant conditioning because these acts imply a degree of learner choice and control. Another issue is that there can be variability among operants since presumably not all possible variations will have been reinforced for a given individual. For example, a student who seeks teacher attention (a positive reinforcer) could conceivably engage in different acts such as perform well, misbehave, get sick, fall on the floor, and the like—not all of which may been reinforced in the past.

Neuringer and Jensen ( 2010 ) addressed this concern by proposing that voluntary acts (operants) are intentional and goal-directed. In line with Skinner’s contention, they predict that reinforcers and discriminative stimuli affect the form and rate of operants, but they also contend that reinforcers and discriminative stimuli influence the variability of operants, which can range from patterned and repetitive (and therefore predictable) to random (and therefore unpredictable). These predictions, which are supported by research, mean that if variability in responding is reinforced, people may act one way now and differently later. By imposing a measure of volition onto operant conditioning, the analysis by Neuringer and Jensen gives operant conditioning a bit of a cognitive flavor.

This behavior theory interpretation of voluntary acts has implications for education since it suggests that variability in actions can be reinforced and made more likely to occur. There are many situations in teaching and learning where teachers desire variability in student responses; for example, in problem solving, creative thinking, and brainstorming. Teachers who reinforce students who demonstrate variability can encourage this type of thinking.

INSTRUCTIONAL APPLICATIONS

Skinner ( 1954 ,  1961 ,  1968 ,  1984 ) wrote extensively on how his ideas can be applied to education. He believed there was too much aversive control. Although students rarely receive corporal punishment, they often work on assignments not because they want to learn or because they enjoy them but rather to avoid punishments such as teacher criticism, loss of privileges, and trips to the principal’s office.

A second concern is that reinforcement occurs infrequently and often not at the proper time. Teachers attend to each student for only a few minutes each day. While students are engaged in learning, minutes can elapse between when they finish an assignment and when they receive teacher feedback. Consequently, students may learn incorrectly, which means that teachers must spend additional time giving corrective feedback.

A third point is that the scope and sequence of curricula do not ensure that all students acquire skills. Students do not learn at the same pace. To cover all the material, teachers may move to the next lesson before all students have mastered the previous one.

Skinner contended that these and other issues cannot be addressed by paying teachers more money (although they would like that!), lengthening the school day and year, raising standards, or toughening teacher certification requirements. Rather, he recommended better use of instructional time. Since it is unrealistic to expect students to move through the curriculum at the same rate, individualizing instruction would improve efficiency.

Skinner believed that teaching required properly arranging reinforcement contingencies. Instruction is more effective when (1) teachers present the material in small steps, (2) learners actively respond rather than passively listen, (3) teachers give feedback immediately following learners’ responses, and (4) learners move through the material at their own pace.

The basic process of instruction involves shaping. The goal of instruction (desired behavior) and the students’ initial behavior are identified. Substeps (behaviors) leading from the initial behavior to the desired behavior are formulated. Each substep represents a small modification of the preceding one. Students are moved through the sequence using various approaches including demonstrations, small-group work, and individual seat work. Students actively respond to the material and receive immediate feedback.

This instructional approach involves specifying learners’ present knowledge and desired objectives in terms of what learners do. Desired behaviors often are specified as behavioral objectives (discussed shortly). Individual differences are taken into account by beginning instruction at learners’ present performance levels and allowing them to progress at their own rates. Given the prevailing teaching methods in our educational system, these goals seem impractical: Teachers would have to begin instruction at different points and cover material at different rates for individual students. Programmed instruction circumvents these problems: Learners begin at the point corresponding to their performance levels and progress at their own rates.

The remainder of this section describes some instructional applications that incorporate behavioristic principles. Not all of these applications are derived from Skinner’s or other theories covered in this chapter, but they all reflect key ideas of behavior theory.

Behavioral Objectives

Behavioral objectives  are clear statements of the intended student outcomes of instruction. Objectives can range from general to specific. General or vague objectives such as “improve student awareness” can be satisfied by different kinds of student behaviors. Conversely, objectives that are too specific and document every minute change in student behavior are time consuming to write and can cause teachers to lose sight of the most important learning outcomes. Optimal objectives fall somewhere between these extremes ( Application 3.8 ).

A behavioral objective describes what students do when demonstrating their achievements and how teachers know what students are doing (Mager,  1962 ). Four parts of a good objective are:

· 1. The specific group of students

· 2. The actual behaviors students are to perform as a consequence of instructional activities

APPLICATION 3.8 Behavioral Objectives

As teachers prepare lesson plans, it is important that they decide on specific behavioral objectives and plan activities to assist students in mastering these objectives. Instead of an art teacher planning a lesson with the objective, “Have students complete a pen drawing of the front of the building,” the teacher should decide on the major objective for the students to master. Is it to use a pen or to draw the front of the school building? The objective may be better stated as follows: “Have the students draw the major lines of the front of the building in correct perspective (materials/medium: drawing paper, pens).”

A kindergarten teacher writes that she wants “Students to go to art, music, and physical education in an orderly fashion.” For that age child, it would be better if the teacher would spell out the objective in more specific terms; for example, “Students should move to other classrooms by walking in a line without talking and by keeping their hands to themselves.”

· 3. The conditions or contexts in which the students are to perform the behaviors

· 4. The criteria for assessing student behaviors to determine whether objectives have been met

A sample objective with the parts identified is:

·  Given eight addition problems with fractions of unlike denominators (3), fourth-grade students (1) will write the correct sums (2) for at least seven of them (4).

Behavioral objectives should specify the important learning outcomes, which aid in lesson planning and testing to assess learning. Formulating objectives also helps teachers decide what content students can master. Given unit-teaching objectives and a fixed amount of time to cover them, teachers can decide which objectives are important and focus on them. Although objectives for lower-level learning outcomes (knowledge, comprehension) are generally easier to specify, good behavioral objectives can be written to assess higher-order outcomes (application, analysis, synthesis, evaluation) as well.

Research shows that students given behavioral objectives have better verbatim recall of verbal information compared with students not provided with objectives (Faw & Waller,  1976 ; Hamilton,  1985 ). Objectives may cue students to process the information at the appropriate level; when students are given objectives requiring recall, they engage in rehearsal and other strategies that facilitate that type of recall. Research also shows that providing students with objectives does not enhance learning of material unrelated to the objectives (Duchastel & Brown,  1974 ), which suggests that students may concentrate on learning material relevant to the objectives and disregard other material.

The effect of objectives on learning depends on students’ prior experience with them and on how important they perceive the information to be. Training in using objectives or familiarity with criterion-based instruction leads to better learning compared to the absence of such training or familiarity. When students can determine on their own what material is important to learn, providing objectives does not facilitate learning. Informing students of the objectives seems to be more important when students do not know what material is important. Also, Muth, Glynn, Britton, and Graves ( 1988 ) found that text structure can moderate the effect of objectives on learning. Information in a prominent position (e.g., early in a text or highlighted) is recalled well, even when objectives are not provided.

Learning Time

Operant theory predicts that environmental variables affect students’ learning. One such variable is learning time.

Carroll ( 1963 ,  1965 ,  1989 ) formulated a model of school learning that places primary emphasis on time spent learning. Students successfully learn to the extent that they spend the amount of time they need to learn. Time means academically engaged time, or time spent paying attention and trying to learn. Although time is an environmental (observable) variable, this definition is cognitive because it goes beyond a simple behavioral indicator of clock time. Within this framework, Carroll postulated factors that influence how much time learning requires and how much time is actually spent learning.

Time Needed for Learning.

One influence on the time a student needs to learn is aptitude for learning the task. Learning aptitude depends on the amount of prior task-relevant learning and on personal characteristics such as abilities and attitudes. A second influence is ability to understand instruction. This variable relates to instructional method; for example, some learners comprehend verbal instruction well, whereas others benefit more from visual presentations.

A third influence is  instructional quality , or how well the task is organized and presented to learners. Quality includes what learners are told about what they will learn and how they will learn it, the extent to which they have adequate contact with the content to be learned, and how much prerequisite knowledge is acquired prior to learning the task. The lower the quality of instruction, the more time learners require to learn.

Time Spent in Learning

. How much time the student spends learning depends on the  time allowed for learning . School curricula include so much content that time allotted for a particular type of learning is less than optimal for some students. When teachers cover material with the entire class at once, some learners are more likely to experience difficulty grasping it and require additional instruction. When students are grouped by ability levels, the amount of time devoted to different content varies depending on the ease with which students learn.

A second influence is the time the learner is willing to spend learning. Even when learners are given ample time to learn, they may not spend that time working productively. Whether due to low interest, high perceived task difficulty, or other factors, students may not be motivated to persist at a task for the amount of time they require to learn it. Carroll incorporated these factors into a formula to estimate the degree of learning for any student on a given task:

degree of learning = time spent in learning/time needed for learning

Ideally, students spend as much time as they need to learn (degree of learning = 1.0), but learners typically spend either more time (degree of learning > 1.0) or less time (degree of learning < 1.0) than they require.

Carroll’s model highlights the importance of academic engaged time required for learning and the factors influencing time spent and time needed to learn. The model incorporates valid psychological principles, but only at a general level as instructional or motivational factors. It does not explore cognitive engagement in depth. Mastery learning researchers, by systematically investigating the time variable, have provided greater specificity (discussed in the next section).

Many educators have decried the way that learning time is misspent (Zepeda & Mayers,  2006 ). Time is central to current discussions on ways to maximize student achievement. For example, the No Child Left Behind Act of 2001 greatly expanded the role of the federal government in elementary and secondary education (Shaul & Ganson,  2005 ). Although the act did not specify how much time was to be devoted to instruction, its requirements for student achievement and its accountability standards, combined with critics calling for better use of time, have led school systems to reexamine their use of time to ensure better student learning.

One consequence is that many secondary schools have changed from the traditional six-hour schedule to block scheduling. Although there are variations, many use the A/B block, in which classes meet on alternate days for longer periods per day. Presumably block scheduling allows teachers and students to explore content in greater depth that often was not possible with the traditional shorter class periods (e.g., 50 minutes).

Given that block scheduling still is relatively new, there is not a lot of research assessing its effectiveness. In their review, Zepeda and Mayers ( 2006 ) found that block scheduling may improve school climate and students’ grade-point averages, but studies show inconsistent results for student attendance and scores on standardized tests. As block scheduling becomes more common, we can expect more research that may clarify these inconsistencies.

Another means for increasing time for learning is through out-of-school programs, such as after-school programs and summer school. Compared with research on block scheduling, research on the effects of out-of-school programs shows greater consistency. In their review, Lauer et al. ( 2006 ) found positive effects for such programs on students’ reading and mathematics achievement; effects were larger for programs with enhancements (e.g., tutoring). Mahoney, Lord, and Carryl ( 2005 ) found benefits of after-school programs on children’s academic performances and motivation; results were strongest for children rated as highly engaged in the after-school program’s activities. Consistent with Carroll’s model, we might conclude that out-of-school programs are successful to the extent that they focus on academic learning and provide supports to encourage it.

Mastery Learning

Carroll’s model predicts that if students vary in aptitude for learning a subject and if all receive the same amount and type of instruction, their achievement will differ. If the amount and type of instruction vary depending on individual differences among learners, then each student has the potential to demonstrate mastery.

These ideas form the basis of  mastery learning  (Anderson,  2003 ; Bloom,  1976 ; Bloom, Hastings, & Madaus,  1971 ). Mastery learning incorporates Carroll’s ideas into a systematic instructional plan that includes defining mastery, planning for mastery, teaching for mastery, and grading for mastery (Block & Burns,  1977 ). Mastery learning contains cognitive elements, although its formulation seems more behavioral in nature compared with many current cognitive theories.

To define mastery, teachers prepare a set of objectives and a final (summative) exam. Level of mastery is established (e.g., where A students typically perform under traditional instruction). Teachers break the course into learning units mapped against course objectives.

Planning for mastery means teachers plan instruction to include corrective feedback (formative evaluation). Such evaluation typically takes the form of unit mastery tests that set mastery at a given level (e.g., 90%). Corrective instruction, which is used with students who fail to master aspects of the unit’s objectives, is given in small-group study sessions, individual tutorials, and supplemental content.

At the outset of teaching for mastery, teachers orient students to the mastery procedures and provide instruction using the entire class, small groups, or individual activities. Teachers give the formative test and certify which students achieved mastery. Students who fall short might work in small groups reviewing troublesome content, often with the aid of peer tutors who have mastered it. Teachers allow students time to work on remedial content, along with homework. Grading for mastery includes a summative (end-of-course) test. Students who score at or above the course mastery performance level receive A grades; lower scores are graded accordingly.

The emphasis on student abilities as determinants of learning may seem uninteresting given that abilities generally do not change much as a result of instructional interventions. Bloom ( 1976 ) also stressed the importance of alterable variables of schooling: cognitive entry behaviors (e.g., student skills and cognitive processing strategies at the outset of instruction), affective characteristics (e.g., interest, motivation), and specific factors influencing the quality of instruction (e.g., student participation, type of corrective feedback). Instructional interventions can improve these variables.

Reviews of the effect of mastery learning on student achievement have yielded mixed results. Block and Burns ( 1977 ) found mastery learning more effective than traditional forms of instruction. With college students, Péladeau, Forget, and Gagné ( 2003 ) obtained results showing that mastery learning improved students’ achievement, long-term retention, and attitudes toward the course and subject matter. Kulik, Kulik, and Bangert-Drowns ( 1990 ) examined more than 100 evaluations of mastery learning programs and found positive effects on academic performances and course attitudes among college, high school, and upper-grade elementary school learners. They also found that mastery learning may increase the time students spend on instructional tasks. In contrast, Bangert, Kulik, and Kulik ( 1983 ) found weaker support for mastery learning programs. They noted that mastery-based instruction was more effective at the college level than at lower levels. Its effectiveness undoubtedly depends on the proper instructional conditions (e.g., planning, teaching, grading) being established (Kulik et al.,  1990 ).

Students participating in mastery instruction often spend more time in learning compared with learners in traditional classes (Block & Burns,  1977 ). Given that time is at a premium in schools, much mastery work must be accomplished outside of regular school hours. Most studies show smaller effects of mastery instruction on affective outcomes (e.g., interest in and attitudes toward the subject matter) than on academic outcomes.

Anderson ( 1976 ) found that when remedial students gained experience with mastery instruction, they gradually required less extra time to attain mastery because their entry-level skills improved. These results imply cumulative benefits of mastery learning. There remains, however, the question of how much practice is enough (Péladeau et al.,  2003 ). Too much repetitive practice might decrease motivation and thereby hinder learning. These points require further research, but have important instructional implications. Some examples of mastery learning are given in  Application 3.9 .

Programmed Instruction

Programmed instruction ( PI ) refers to instructional materials developed in accordance with operant conditioning principles of learning (O’Day, Kulhavy, Anderson, & Malczynski,  1971 ). In the 1920s, Sidney Pressey designed machines to use primarily for testing. Students were presented with multiple-choice questions, and they pressed a button corresponding to their choice. If students responded correctly, the machine presented the next choice; if they responded incorrectly, the error was recorded and they continued to respond to the item.

APPLICATION 3.9 Mastery Learning

A mastery learning approach can be beneficial in certain learning environments. In a remedial reading group for secondary students, a well-organized mastery learning program would allow students to progress at their own rates. Students motivated to make rapid progress are not slowed down by this type of instruction, as might happen if they are placed in a traditional learning format. A key requirement is to include a progression of activities from easier to more difficult. The program should have checkpoints at which the students interact with the teacher so that their progress is evaluated and reteaching provided if needed.

Young children enter school with a wide range of experiences and abilities. Mastery learning can help teachers deal effectively with the varying abilities and developmental levels. Mastery learning techniques can be implemented by using learning centers and small groups. Children can be placed in the different centers and groups according to their current levels. Then they can move through the various levels at their own rates.

Mastery learning also can build students’ self-efficacy for learning ( Chapter 4 ). As they perceive their progress in completing units, they are apt to believe they are capable of further learning. Enhancing self-efficacy is particularly important with learners who have encountered learning difficulties and doubt their capabilities to learn, as well as for young children with limited experiences and skills.

Skinner revived Pressey’s machines in the 1950s and modified them to incorporate instruction (Skinner,  1958 ). These teaching machines presented students with material in small steps (frames). Each frame required learners to make an overt response. Material was carefully sequenced and broken into small units to minimize errors. Students received immediate feedback on the accuracy of each response. They moved to the next frame when their answer was correct. When it was incorrect, supplementary material was provided. Although errors occurred, the programs were designed to minimize errors and ensure that learners typically succeeded (Benjamin,  1988 ).

There are many benefits when students generally perform well, but as noted earlier, research suggests that preventing errors may not be desirable. Dweck ( 1975 ) found that an occasional failure increased persistence on difficult tasks more than did constant success. Further, constant success is not as informative of one’s capabilities as is occasionally having difficulty because the latter highlights what one can and cannot do. This is not to suggest that teachers should let students fail, but rather that under the proper circumstances students can benefit from tasks structured so that they occasionally encounter difficulty.

PI does not require the use of a machine; a book by Holland and Skinner ( 1961 ) is an example of PI. Today, however, most PI is computerized and is a type of  computer-based instruction  (CBI). Current instructional programs are more elaborate than the early PI ones.

PI reflects several learning principles (O’Day et al.,  1971 ). Behavioral objectives specify what students should perform on completion of the instruction. Units are subdivided into sequenced frames, each of which presents a small bit of information and a test item to which learners respond. Although a lot of content may be included in the program, the frame-to-frame increments are small. Learners work at their own pace and respond to questions as they work through the program. Responses may require learners to supply words, provide numerical answers, or choose which of several statements best describes the idea being presented. Feedback depends on the learner’s response. If the learner is correct, the next item is given. If the learner answers incorrectly, additional information is presented and the item is tested in a slightly different form.

There are two types of programs—linear and branching—distinguished according to how they treat learner errors. In a  linear program , all students proceed in the same sequence but not necessarily at the same rate. Regardless of whether they respond correctly to a frame, they move to the next frame where they receive feedback on the accuracy of their answer. Programs minimize errors by covering the same material in more than one frame and by prompting student responses.

In a  branching program , students’ movement depends on how they answer the questions ( Figure 3.4 ). Those who learn quickly skip frames and bypass much of the repetition of linear programs, whereas slower learners receive additional instruction. A potential disadvantage is that branching programs may not provide sufficient repetition to ensure that all students learn concepts well.

Research suggests that linear and branching programs promote student learning equally well and that PI is as effective as conventional classroom teaching (Bangert et al.,  1983 ; Lange,  1972 ). PI seems especially useful with students who demonstrate skill deficiencies; working through programs provides remedial instruction and practice. PI also is useful for independent study on a topic.

Until it was supplanted by the Internet, CBI was the most common application of computer learning in schools (Jonassen,  1996 ). Studies investigating CBI in college courses show beneficial effects on students’ achievement and attitudes (Kulik, Kulik, & Cohen,  1980 ). Several CBI features are firmly grounded in learning theory and research. Computers command students’ attention and provide immediate feedback, which can be of a type typically not given in class (e.g., how present performances compare with prior performances to highlight progress). Computers can individualize content and rate of presentation. Information about students’ skills and prior responses can be stored. Through advances in technology, learning can be adapted to individual students’ needs, such that they move through learning units with personalized frames (Webley,  2013 ).

Even simple forms of personalization can be beneficial. Students can enter information about themselves, parents, and friends, which is then included in the instruction. Research shows that personalization can produce higher achievement (Anand & Ross,  1987 ; Ross, McCormick, Krisak, & Anand,  1985 ). Anand and Ross ( 1987 ) gave elementary children instruction in dividing fractions according to one of three problem formats (abstract, concrete, personalized):

Figure 3.4 Frames from a branching program.

·  (Abstract) There are three objects. Each is cut in half. In all, how many pieces would there be?

(Concrete) Billy had three candy bars. He cut each of them in half. In all, how many pieces of candy did Billy have?

(Personalized for Joseph) Joseph’s teacher, Mrs. Williams, surprised him on December 15 when she presented Joseph with three candy bars. Joseph cut each one of them in half so that he could share the birthday gift with his friends. In all, how many pieces of candy did Joseph have? (pp. 73–74)

The personalized format led to better learning and transfer than the abstract format and to more positive attitudes toward instruction than the concrete format.

Contingency Contracts

A  contingency contract  is an agreement between teacher and student specifying what work the student will accomplish and the expected outcome (reinforcement) for successful performance (Homme, Csanyi, Gonzales, & Rechs,  1970 ). A contract can be made verbally, although it usually is written. Teachers can devise the contract and ask if the student agrees with it, but it is customary for teacher and student to formulate it jointly. An advantage of joint participation is that students may feel more committed to fulfilling the contract’s terms. When people participate in goal selection, they tend to be more committed to attaining the goal than when they are excluded from the selection process (Locke & Latham,  1990 ).

Contracts specify goals or expected outcomes in terms of particular behaviors to be displayed. The “contingency” is the expected outcome, which often can be reduced to, “If you do this, then you will receive that.” The behaviors should be specific—for example, “I will complete problems 1–30 in my math book with at least 90% accuracy,” or “I will stay in my seat during reading period.” General behaviors (e.g., “I will work on my math” or “I will behave appropriately”) are unacceptable. With young children, time frames should be brief; however, objectives can cover more than one time, such as successive 30-minute periods or during each social studies period for one week. Contracts may include academic and nonacademic behaviors ( Application 3.10 ).

Developing contracts with students and monitoring progress is time consuming. Fortunately, most learners do not require contracts to behave appropriately or accomplish work. Contracts seem especially helpful as a means of assisting students to work on assignments more productively. A lengthy, long-term assignment can be subdivided into a series of short-term goals with due dates. This type of plan helps students keep up with the work and turn in material on time.

Contracts are based on the principle that goals that are specific, temporally close at hand, and difficult but attainable maximize performance (Schunk,  1995 ). Contracts also convey information to students about their progress in completing the task. Such information on progress raises student motivation and achievement (Locke & Latham,  1990 ). Contracts should promote achievement if they reinforce student progress in learning or in accomplishing more on-task behavior.

APPLICATION 3.10 Contingency Contracts

A contingency contract represents a systematic application of reinforcement principles to change behavior. It can be used to change any type of behavior, such as completing assignments, not disrupting the class, and participating in discussions. When developing a contract, a teacher should make sure that the reward is something that interests and motivates the students.

Assume that Mrs. Lauter has tried unsuccessfully to apply several motivational techniques to encourage James, a student in her class, to complete assignments in language arts. She and James might jointly develop a contract to address the inappropriate behaviors. They should discuss the problem, identify the desired behavior, and list the consequences and time frame for fulfilling the terms of the contract. A sample contract might be as follows:

Contract for the Week of January 9–13

I will complete my language arts assignments with 80% accuracy in the time allotted during class. If I complete my assignments, I will be allowed to participate in a learning center activity. If I do not complete my assignments, I will miss recess and complete them then.

·  Monday:

_____ Completed _____ Not completed

Tuesday:

_____ Completed _____ Not completed

Wednesday:

_____ Completed _____ Not completed

Thursday:

_____ Completed _____ Not completed

Friday:

_____ Completed _____ Not completed

Bonus: If I complete my assignments at least three of the five days, I will receive computer time for 30 minutes on Friday afternoon.

_____________	_____________
Student	Teacher
Signature/Date	Signature/Date

SUMMARY

Behaviorism—as reflected in conditioning theories—dominated the psychology of learning for the first half of the 20th century. These theories explain learning in terms of environmental events. Mental processes are not necessary to explain the acquisition, maintenance, and generalization of behavior.

The theories of Thorndike, Pavlov, and Guthrie helped to establish the psychology of learning as a legitimate area of study. These theories differ, but each views learning as a process of forming associations between stimuli and responses. Thorndike believed that responses to stimuli are strengthened when followed by satisfying consequences. Pavlov experimentally demonstrated how stimuli could be conditioned to elicit responses by being paired with other stimuli. Guthrie hypothesized that a contiguous relation between stimulus and response established their pairing. Although these theories are no longer viable in their original form, many of their principles are evident in current learning theories.

Operant conditioning—the learning theory formulated by B. F. Skinner—is based on the assumption that features of the environment (stimuli, situations, events) serve as cues for responding. Reinforcement strengthens responses and increases their future likelihood of occurring when the stimuli are present. It is not necessary to refer to underlying physiological or mental states to explain behavior.

The basic operant conditioning model is a three-term contingency involving a discriminative stimulus (antecedent), response (behavior), and reinforcing stimulus (consequence). The consequences of behaviors determine the likelihood that people will respond to environmental cues. Consequences that are reinforcing increase behavior; consequences that are punishing decrease behavior. Some other important operant conditioning concepts are extinction, generalization, discrimination, primary and secondary reinforcers, reinforcement schedules, and the Premack Principle.

Shaping—a process for altering behavior—involves reinforcing successive approximations of the desired behavior toward its desired form or frequency of occurrence. Complex behaviors are formed by chaining together simple behaviors in successive three-term contingencies. Behavior modification programs have been commonly applied in diverse contexts to promote adaptive behaviors.

The generality of operant conditioning principles has been challenged by cognitive theorists who contend that by ignoring mental processes, operant conditioning offers an incomplete account of human learning. Stimuli and reinforcement may explain some human learning, but much research shows that to explain learning—and especially higher-order and complex learning—we must take into account people’s thoughts, beliefs, and feelings. Newer behavior theory perspectives retain basic behavioral principles but interject some cognitive elements such as volition.

Operant principles have been applied to many aspects of teaching and learning. These principles can be seen in applications involving behavioral objectives, learning time, mastery learning, computer-based instruction, and contingency contracts. Research evidence often shows positive effects of these applications on student achievement. Regardless of theoretical orientation, one can apply behavioral principles to facilitate student learning and achievement.

A summary of the learning issues ( Chapter 1 ) for conditioning theories appears in  Table 3.6 .

Table 3.6 Summary of learning issues.

How Does Learning Occur? The basic model of operant learning is expressed by the three-term contingency: SD → R → SR. A response is performed in the presence of a discriminative stimulus and is followed by a reinforcing stimulus. The likelihood of the R being performed in the future in the presence of that SD is increased. To build complex behaviors requires shaping, which consists of chains of three-term contingencies, where gradual approximations to the desired form of behavior are successively reinforced. Factors affecting learning are developmental status and reinforcement history. For conditioning to occur, one must have the physical capabilities to perform the behaviors. The responses that one makes in given situations depend on what one has been reinforced for doing in the past. How Does Memory Function? Memory is not explicitly addressed by conditioning theories. These theories do not study internal processes. Responses to given stimuli are strengthened through repeated reinforcement. This response strengthening accounts for present behavior. What Is the Role of Motivation? Motivation is an increase in the quantity or rate of behavior. No internal processes are used to explain motivation. The increase in quantity or rate can be explained in terms of reinforcement history. Certain schedules of reinforcement produce higher rates of responding than others. How Does Transfer Occur? Transfer, or generalization, occurs when one responds in an identical or similar fashion to stimuli other than the ones that were used in conditioning. At least some of the elements in the transfer setting must be similar to those in the conditioning setting for transfer to occur. How Does Self-Regulated Learning Operate? As discussed in  Chapter 10 , operant conditioning construes self-regulated behavior as choosing among alternative actions, often by deferring an immediate reinforcer in favor of a different and usually greater future reinforcer. The key processes are self-monitoring, self-instruction, and self-reinforcement. One decides which behaviors to regulate, establishes discriminative stimuli for their occurrence, receives instruction, monitors performance and determines whether it matches the standard, and administers reinforcement. What Are the Implications for Instruction? Learning requires establishing responses to discriminative stimuli. Practice is needed to strengthen responses. Complex skills can be established by shaping progressive, small approximations to the desired behavior. Instruction should have clear, measurable objectives, proceed in small steps, and deliver reinforcement. Mastery learning, computer-based instruction, and contingency contracts are useful ways to promote learning.

Chapter 4 Social Cognitive Theory

The girls’ tennis team of Westbrook High School is practicing after school. The team has played a few matches; they are playing well, but some improvements are needed. Coach Sandra Martin is working with Donnetta Awalt, the number four singles player. Donnetta’s overall game is good, but lately she has been hitting many of her backhands into the net. Coach Martin asks Donnetta to hit backhands to her as she hits balls to Donnetta.

Donnetta:	This is impossible. I just can’t do it.
Coach Martin:	Sure you can. You’ve been able to hit backhands before, and you will again.
Donnetta:	Then what should I do?
Coach Martin:	You’re swinging downward during your backhand, which means you’ll hit the ball into the net a lot. We need for you to develop more of an upward swing. Come over here, please, and I’ll demonstrate (Coach Martin demonstrates Donnetta’s swing and then an upward swing and points out the differences). Now you try it, slowly at first. Do you feel the difference?
Donnetta:	Yes. But from where should I start my swing? How far back and how low down?
Coach Martin:	Watch me again. Adjust your grip like this before hitting a backhand (Coach Martin demonstrates grip). Get into position, about like this relative to the ball (Coach Martin demonstrates). Now start your backhand like this (Coach Martin demonstrates) and bring it through like this (Coach Martin demonstrates). You see you’re swinging upward, not downward.
Donnetta:	Okay, that feels better (practices). Can you hit some to me?
Coach Martin:	Sure. Let’s try it, slowly at first, then we’ll pick up speed (they practice for several minutes). That’s good. I’ve got a book I’ll give you. Look at the section on backhands. There are some good pictures with explanations of what I’ve been telling you.
Donnetta:	Thanks, I will. I really felt I couldn’t do this anymore, so I’ve been trying to avoid hitting backhands in matches. But now I’m feeling more confident.
Coach Martin:	That’s good. Keep thinking like that and practicing and you may move up to number three singles.

The preceding chapter focused on conditioning theories (behaviorism), which held sway in the field of learning for the first half of the 20th century. Beginning in the late 1950s and early 1960s, these theories were challenged on many fronts. Their influence waned to the point where today the prevailing theoretical perspectives are cognitive.

One of the major challenges to behaviorism came from studies on observational learning conducted by Albert Bandura and his colleagues. A central finding of this research was that people could learn new actions by observing others perform them. Observers did not have to perform the actions at the time of learning. Reinforcement was not necessary for learning to occur. These findings disputed central assumptions of conditioning theories.

The focus of this chapter is on Bandura’s ( 1986 ,  1997 ,  2001 )  social cognitive theory , which stresses the idea that much human learning occurs in a social environment. By observing others, people acquire knowledge, rules, skills, strategies, beliefs, and attitudes. Individuals also learn from models the usefulness and appropriateness of behaviors and the consequences of modeled behaviors, and they act in accordance with beliefs about their capabilities and the expected outcomes of their actions. The opening scenario portrays an instructional application of modeling.

The conceptual framework of social cognitive theory is discussed, along with its underlying assumptions about the nature of human learning and behavior. A significant portion of the chapter is devoted to modeling processes. The various influences on learning and performance are described, and motivational influences are discussed with special emphasis on the critical role of self-efficacy. Some instructional applications that reflect social cognitive learning principles are provided.

· When you finish studying this chapter, you should be able to do the following:

· ■ Describe and exemplify the process of triadic reciprocal causality.

· ■ Distinguish between enactive and vicarious learning and between learning and performance.

· ■ Explain the role of self-regulation in social cognitive theory.

· ■ Define and exemplify three functions of modeling.

· ■ Discuss the processes of observational learning.

· ■ Explain the various factors that affect observational learning and performance.

· ■ Discuss the motivational properties of goals, outcome expectations, and values.

· ■ Define self-efficacy and explain its causes and effects in learning settings.

· ■ Discuss how features of models (e.g., peers, multiple, coping) affect self-efficacy and learning.

· ■ Describe some educational applications that reflect social cognitive theoretical principles.

CONCEPTUAL FRAMEWORK FOR LEARNING

Albert Bandura was born in Alberta, Canada, in 1925. He received his doctorate in clinical psychology from the University of Iowa, where he was influenced by Miller and Dollard’s ( 1941 ) Social Learning and Imitation (discussed later in this chapter). After arriving at Stanford University in the 1950s, Bandura began a research program exploring the influences on social behavior. He believed that the conditioning theories in vogue at that time offered incomplete explanations of the acquisition and performance of prosocial and deviant behaviors:

·  Indeed, most prior applications of learning theory to issues concerning prosocial and deviant behavior … have suffered from the fact that they have relied heavily on a limited range of principles established on the basis of, and mainly supported by, studies of animal learning or human learning in one-person situations. (Bandura & Walters,  1963 , p. 1)

Bandura formulated a comprehensive theory of observational learning that he has expanded to encompass acquisition and performance of diverse skills, strategies, and behaviors. Social cognitive principles have been applied to the learning of cognitive, motor, social, and self-regulation skills, as well as to the topics of violence (live, filmed), moral development, education, health, and societal values (Zimmerman & Schunk,  2003 ).

Bandura is a prolific writer. Beginning with the book Social Learning and Personality Development, written in 1963 with Richard Walters, he has authored several other books, including Principles of Behavior Modification (1969), Aggression: A Social Learning Analysis (1973), Social Learning Theory(1977b), and Social Foundations of Thought and Action: A Social Cognitive Theory (1986). With the publication of Self-Efficacy: The Exercise of Control (1997), Bandura extended his theory to address ways people seek control over important events of their lives through self-regulation of their thoughts and actions. The basic processes involve setting goals, judging anticipated outcomes of actions, evaluating progress toward goals, and self-regulating thoughts, emotions, and actions. As Bandura ( 1986 ) explained:

·  Another distinctive feature of social cognitive theory is the central role it assigns to self-regulatory functions. People do not behave just to suit the preferences of others. Much of their behavior is motivated and regulated by internal standards and self-evaluative reactions to their own actions. After personal standards have been adopted, discrepancies between a performance and the standard against which it is measured activate evaluative self-reactions, which serve to influence subsequent behavior. An act, therefore, includes among its determinants self-produced influences. (Bandura,  1986 , p. 20)

Social cognitive theory makes assumptions about learning and the performance of behaviors (Schunk, 2012). These assumptions address the reciprocal interactions among persons, behaviors, and environments; enactive and vicarious learning (i.e., how learning occurs); the distinction between learning and performance; and the role of self-regulation.

Reciprocal Interactions

Bandura ( 1982a ,  1986 ,  2001 ) discussed human behavior within a framework of  triadic reciprocality , or reciprocal interactions among behaviors, environmental variables, and personal factors such as cognitions ( Figure 4.1 ). These interacting determinants can be illustrated using  perceived self-efficacy , or beliefs concerning one’s capabilities to organize and implement actions necessary to learn or perform behaviors at designated levels (Bandura,  1982b ,  1986 ,  1997 ). With respect to the interaction of self-efficacy (a personal factor) and behavior, researchers have shown that self-efficacy influences such achievement behaviors as choice of tasks, persistence, effort expenditure, and skill acquisition (person → behavior; Schunk,  2012 ; Schunk & Pajares,  2009 ). Notice in the opening scenario that Donnetta’s low self-efficacy led her to avoid hitting backhands in matches. In turn, students’ actions modify their self-efficacy. As students work on tasks, they note their progress toward their learning goals (e.g., completing assignments, finishing sections of a term paper). Such progress indicators convey to students that they are capable of performing well and enhance their self-efficacy for continued learning (behavior → person).

Research on students with learning disabilities has demonstrated the interaction between self-efficacy and environmental factors. Many such students hold a low sense of self-efficacy for performing well (Licht & Kistner,  1986 ). Individuals in students’ social environments may react to students based on attributes typically associated with students with learning disabilities (e.g., low self-efficacy) rather than on the individuals’ actual capabilities (person → environment). Some teachers, for example, judge such students less capable than students without disabilities and hold lower academic expectations for them, even in content areas where students with learning disabilities are performing adequately (Bryan & Bryan,  1983 ). In turn, teacher feedback can affect self-efficacy (environment → person). When a teacher tells a student, “I know you can do this,” the student likely will feel more confident about succeeding.

Students’ behaviors and classroom environments influence one another in many ways. Consider a typical instructional sequence in which the teacher presents information and asks students to direct their attention to a slide. Environmental influence on behavior occurs when students look at the slide without much conscious deliberation (environment → behavior). Students’ behaviors often alter the instructional environment. If the teacher asks questions and students give the wrong answers, the teacher may reteach some points rather than continue the lesson (behavior → environment).

The model portrayed in  Figure 4.1  does not imply that the directions of influence are always the same. At any given time, one factor may predominate. When environmental influences are weak, personal factors predominate. For instance, students allowed to write a report on a topic of their choosing will select one they enjoy. However, a person caught in a burning house is apt to evacuate quickly; the environment dictates the behavior.

Figure 4.1 Triadic reciprocality model of causality.

Source: Social Foundations of Thought and Action by A. Bandura, © 1986. Reprinted by permission of Pearson Education, Inc. Upper Saddle River, NJ.

Much of the time the three factors interact. As a teacher presents a lesson to the class, students think about what the teacher is saying (environment influences cognition—a personal factor). Students who do not understand raise their hands to ask a question (cognition influences behavior). The teacher reviews the content (behavior influences environment). Eventually the teacher gives students work to accomplish (environment influences cognition, which influences behavior). As students work on the task, they believe they are performing it well (behavior influences cognition). They decide they like the task, ask the teacher if they can continue to work on it, and are allowed to do so (cognition influences behavior, which influences environment).

Enactive and Vicarious Learning

In social cognitive theory:

·  Learning is largely an information processing activity in which information about the structure of behavior and about environmental events is transformed into symbolic representations that serve as guides for action. (Bandura,  1986 , p. 51)

Learning occurs either enactively through actual doing or vicariously by observing models (e.g., live, symbolic, electronic) perform (Schunk,  2012 ).

Enactive learning  involves learning from the consequences of one’s actions. Behaviors that result in successful consequences are retained; those that lead to failures are refined or discarded. Conditioning theories also say that people learn by doing, but social cognitive theory provides a different explanation. Skinner ( 1953 ) noted that cognitions may accompany behavioral change but do not influence it ( Chapter 3 ). Rather than strengthening behaviors as postulated by conditioning theories, social cognitive theory contends that behavioral consequences serve as sources of information and motivation. Consequences inform people of the accuracy or appropriateness of behavior. People who succeed at a task or are rewarded understand that they are performing well. When people fail or are punished, they know that they are doing something wrong and may try to correct the problem. Consequences also motivate people. People strive to learn behaviors they value and believe will have desirable consequences, whereas they avoid learning behaviors that are punished or otherwise not satisfying. People’s cognitions, rather than consequences, affect learning.

Much human learning occurs vicariously, or without overt performance by the learner, at the time of learning. Common sources of  vicarious learning  are observing or listening to models who are live (appear in person), symbolic or nonhuman (e.g., talking animals, cartoon characters), electronic (e.g., television, computer, DVD), or in print (e.g., books, magazines). Vicarious sources accelerate learning over what would be possible if people had to perform every behavior for learning to occur. Vicarious sources also save people from personally experiencing negative consequences. We learn that poisonous snakes are dangerous through teaching by others, reading books, watching films, and so forth, rather than by experiencing the unpleasant consequences of their bites!

Learning complex skills typically occurs through a combination of observation and performance (Schunk,  2012 ). Students observe models explain and demonstrate skills, then practice them. This sequence is evident in the opening scenario, where Coach Martin explains and demonstrates and Donnetta observes and practices. Through observation, students often learn some components of a complex skill and not others. Practice gives teachers and coaches opportunities to provide corrective feedback to help students perfect their skills. As with enactive learning, response consequences from vicarious sources inform and motivate observers. Observers are more apt to learn modeled behaviors leading to successes than those resulting in failures. When people believe that modeled behaviors are useful, they attend carefully to models and mentally rehearse the behaviors.

Learning and Performance

Social cognitive theory distinguishes between new learning and performance of previously learned behaviors. Although much learning occurs by doing, we learn a great deal by observing, often in the absence of a goal or reinforcement ( latent learning ). Whether we ever perform what we learn depends on factors such as our motivation, interest, incentives to perform, perceived need, physical state, social pressures, and type of competing activities. Reinforcement, or the belief that it will be forthcoming, affects performance rather than learning.

Some school activities (e.g., review sessions) involve performance of previously learned skills, but much time is spent on learning. By observing teacher and peer models, students acquire knowledge they may not demonstrate at the time of learning. For example, students might learn in school that skimming is a useful procedure for acquiring the gist of a passage and might learn a strategy for skimming, but may not employ that knowledge to promote learning until they read at home.

Self-Regulation

A key assumption of social cognitive theory is that people desire “to control the events that affect their lives” and to perceive themselves as agents (Bandura,  1997 , p. 1). This sense of  agency  manifests itself in intentional acts, cognitive processes, and affective processes (Bandura,  2006 ).  Perceived self-efficacy (discussed later in this chapter) is a central process affecting one’s sense of agency. Other key processes (also discussed in this chapter) are outcome expectations, values, goal setting, self-evaluation of goal progress, and cognitive modeling and self-instruction.

Central to this conception of personal agency is  self-regulation (self-regulated learning) , or the process whereby individuals activate and sustain behaviors, cognitions, and affects, which are systematically oriented toward the attainment of goals (Zimmerman,  2000 ,  2013 ). By striving to self-regulate important aspects of their lives, individuals attain a greater sense of personal agency. In learning situations, self-regulation requires that learners have choices; for example, in what they do and how they do it. Choices are not always available to learners, as when teachers control many aspects by giving students an assignment and spelling out the parameters. When all or most task aspects are controlled, it is accurate to speak of external regulation or regulation by others. The social cognitive theoretical perspective on self-regulation is covered in greater depth in  Chapter 10 .

MODELING PROCESSES

Modeling —a critical component in social cognitive theory—refers to behavioral, cognitive, and affective changes deriving from observing one or more models (Rosenthal & Bandura,  1978 ; Schunk,  1987 ,  2012 ; Zimmerman,  2013 ). Historically, modeling was equated with  imitation , but modeling is a more inclusive concept. Some historical work is covered next to provide a background against which the significance of modeling research by Bandura and others can be appreciated.

Theories of Imitation

Throughout history, people have viewed imitation as an important means of transmitting behaviors (Rosenthal & Zimmerman,  1978 ). The ancient Greeks used the term  mimesis  to refer to learning through observation of the actions of others and of abstract models exemplifying literary and moral styles. Other perspectives on imitation relate it to instinct, development, conditioning, and instrumental behavior ( Table 4.1 ).

Instinct.

At the beginning of the 20th century, the dominant scientific view was that people possessed a natural instinct to imitate the actions of others (James,  1890 ; Tarde,  1903 ). James believed that imitation was largely responsible for socialization, but he did not explain the process by which imitation occurs. McDougall ( 1926 ) restricted his definition of imitation to the instinctive copying by one person of the actions of another.

The instinct notion was discarded by behaviorists because it assumed the existence of an internal drive (and possibly a mental image) intervening between a stimulus (action of another person) and response (copying of that action). Watson ( 1924 ) believed that people’s behaviors labeled “instinctive” resulted largely from training and therefore were learned.

Development.

Table 4.1 Theories of imitation.

View	Assumptions
Instinct	Observed actions elicit an instinctive drive to copy those actions.
Development	Children imitate actions that fit with existing cognitive structures.
Conditioning	Behaviors are imitated and reinforced through shaping. Imitation becomes a generalized response class.
Instrumental behavior	Imitation becomes a secondary drive through repeated reinforcement of responses matching those of models. Imitation results in drive reduction.

Piaget ( 1962 ) offered a different view of imitation. He believed that human development involved the acquisition of  schemes (schemas) , or cognitive structures that underlie and make possible organized thought and action. Thoughts and actions are not synonymous with schemes; they are overt manifestations of schemes. Schemes available to individuals determine how they react to events. Schemes reflect prior experiences and comprise one’s knowledge at any given time.

Schemes presumably develop through maturation and experiences slightly more advanced than one’s existing cognitive structures. Imitation is restricted to activities corresponding to existing schemes. Children may imitate actions they understand, but they should not imitate actions incongruent with their cognitive structures. Development, therefore, must precede imitation.

This view limits the potential of imitation to create and modify cognitive structures. Further, there is little empirical support for this position (Rosenthal & Zimmerman,  1978 ). In an early study, Valentine ( 1930b ) found that infants could imitate actions within their capabilities that they had not previously performed. Infants showed a strong tendency to imitate unusual actions commanding attention. The imitation was not always immediate, and actions often had to be repeated before infants would imitate them. The individual performing the original actions was important: Infants were most likely to imitate their mothers. These and results from subsequent research show that imitation is not a simple reflection of developmental level but rather may serve an important role in promoting development (Rosenthal & Zimmerman,  1978 ).

Conditioning.

Conditioning theorists construe imitation in terms of associations. According to Humphrey ( 1921 ), imitation is a type of circular reaction in which each response serves as a stimulus for the next response. A baby may start crying (response) because of a pain (stimulus). The baby hears its own crying (auditory stimulus), which then becomes a stimulus for subsequent crying. Through conditioning, small reflex units form progressively more complex response chains.

Skinner’s ( 1953 ) operant conditioning theory treats imitation as a generalized response class ( Chapter 3 ). In the three-term contingency (SD → R → SR), a modeled act serves as the SD (discriminative stimulus). Imitation occurs when an observer performs the same response (R) and receives reinforcement (SR). This contingency becomes established early in life. For example, a parent makes a sound (“Dada”), the child imitates, and the parent delivers reinforcement (smile, hug). Once an imitative response class is established, it can be maintained on an intermittent reinforcement schedule. Children imitate the behaviors of models (e.g., parents, friends) as long as the models remain discriminative stimuli for reinforcement.

A limitation of this view is that one can imitate only those responses one can perform. In fact, much research shows that diverse types of behaviors can be acquired through observation (Rosenthal & Zimmerman,  1978 ). Another limitation concerns the need for reinforcement to produce and sustain imitation. Research by Bandura and others shows that observers learn from models in the absence of reinforcement to models or observers (Bandura,  1986 ). Reinforcement primarily affects learners’ performance of previously learned responses rather than new learning.

Instrumental Behavior.

Miller and Dollard ( 1941 ) proposed an elaborate theory of imitation, or  matched-dependent behavior , which contends that imitation is instrumental learned behavior because it leads to reinforcement. Matched-dependent behavior is matched to (the same as) that of the model and depends on, or is elicited by, the model’s action.

Miller and Dollard believed that initially the imitator responds to behavioral cues in trial-and-error fashion, but eventually the imitator performs the correct response and is reinforced. Responses performed by imitators previously were learned.

This conception of imitation as learned instrumental behavior was an important advance, but it has problems. Like other historical views, this theory postulates that new responses are not created through imitation; rather, imitation represents performance of learned behaviors. As such, it cannot account for learning through imitation, for delayed imitation (i.e., when imitators perform the matching responses some time after the actions are performed by the model), or for imitated behaviors that are not reinforced (Bandura & Walters,  1963 ). This narrow conception of imitation restricts its usefulness to imitative responses corresponding closely to those portrayed by models.

Functions of Modeling

Bandura ( 1986 ) distinguished three key functions of modeling: response facilitation, inhibition/disinhibition, and observational learning ( Table 4.2 ).

Response Facilitation.

People learn many skills and behaviors that they do not perform because they lack motivation to do so.  Response facilitation  refers to modeled actions that serve as social prompts for observers to behave accordingly. Consider an elementary teacher who has set up an attractive display in a corner of the classroom. When the first students enter in the morning, they spot the display and immediately go to look at it. When other students enter the room, they see a group in the corner, so they, too, move to the corner to see what everyone is looking at. Several students together serve as a social prompt for others to join them, even though the latter may not know why the others are gathered.

Response facilitation effects are common. Have you ever seen a group of people looking in one direction? This can serve as a cue for you to look in the same direction. Note that response facilitation does not reflect learning because people already know how to perform the behaviors. Rather, the models serve as cues for observers’ actions. Observers gain information about the appropriateness of behavior and may be motivated to perform the actions if models receive positive consequences.

Response facilitation modeling may occur without conscious awareness. Chartrand and Bargh ( 1999 ) found evidence for a  chameleon effect , whereby people nonconsciously mimic behaviors and mannerisms of people in their social environments. Simply perceiving behavior may trigger a response to act accordingly.

Table 4.2 Functions of modeling.

Function	Underlying Process
Response facilitation	Social prompts create motivational inducements for observers to model the actions (“going along with the crowd”).
Inhibition and disinhibition	Modeled behaviors create expectations in observers that they will experience similar consequences should they perform the actions.
Observational learning	Processes include attention, retention, production, and motivation.

Inhibition/Disinhibition.

Observing a model can strengthen or weaken inhibitions to perform behaviors previously learned.  Inhibition  occurs when models are punished for performing certain actions, which in turn stops or prevents observers from acting accordingly.  Disinhibition  occurs when models perform threatening or prohibited activities without experiencing negative consequences, which may lead observers to perform the same behaviors. Inhibitory and disinhibitory effects on behavior occur because the modeled displays convey to observers that similar consequences are probable if they perform the modeled behaviors. Such information also may affect emotions (e.g., increase or decrease anxiety) and motivation.

Teachers’ actions can inhibit or disinhibit classroom misbehavior. Unpunished student misbehavior may prove disinhibiting: Students who observe modeled misbehavior not punished might start misbehaving themselves. Conversely, misbehavior in other students may be inhibited when a teacher disciplines one student for misbehaving. Observers are more likely to believe that they, too, will be disciplined if they continue to misbehave and are spotted by the teacher.

Inhibition and disinhibition are similar to response facilitation in that behaviors reflect actions people already have learned. One difference is that response facilitation generally involves behaviors that are socially acceptable, whereas inhibited and disinhibited actions often have moral or legal overtones (i.e., involve breaking rules or laws) and have accompanying emotions (e.g., fears).

Observational Learning.

Observational learning  through modeling occurs when observers display new patterns of behavior that, prior to exposure to the modeled behaviors, have a zero probability of occurrence even when motivation is high (Bandura,  1969 ). A key mechanism is the information conveyed by models to observers of ways to produce new behaviors (Rosenthal & Zimmerman,  1978 ). In the opening scenario, Donnetta needed to learn (or relearn) the correct procedure for hitting a backhand. Observational learning comprises four processes: attention, retention, production, and motivation (Bandura,  1986 ; see  Table 4.3 ).

Table 4.3 Processes of observational learning.

Process	Activities
Attention	Student attention is directed by physically accentuating relevant task features, subdividing complex activities into parts, using competent models, and demonstrating usefulness of modeled behaviors.
Retention	Retention is increased by rehearsing information to be learned, coding in visual and symbolic form, and relating new material to information previously stored in memory.
Production	Behaviors produced are compared to one’s conceptual (mental) representation. Feedback helps to correct deficiencies.
Motivation	Consequences of modeled behaviors inform observers of functional value and appropriateness. Consequences motivate by creating outcome expectations and raising self-efficacy.

Observer  attention  is necessary so that relevant events are meaningfully perceived. At any given moment one can attend to many activities. Characteristics of the model and the observer influence one’s attention to models. Task features also command attention, especially unusual size, shape, color, or sound. Teachers often make modeling more distinctive with bright colors and oversized features. Attention also is influenced by perceived functional value of modeled activities. Modeled activities that observers believe are important and likely to lead to rewarding outcomes command greater attention. Students believe that most teacher activities are highly functional because they are intended to enhance student learning. Learners also are apt to believe that their teachers are highly competent, which enhances attention. Factors that promote the perception of model competence are modeled actions that lead to success and symbolic indicators of competence, such as one’s title or position.

Retention  requires cognitively organizing, rehearsing, coding, and transforming modeled information for storage in memory (see  Chapter 5 ). Social cognitive theory postulates that a modeled display can be stored as an image, in verbal form, or both (Bandura,  1977b ).  Rehearsal , or the mental review of information, serves a key role in retention. Bandura and Jeffery ( 1973 ) found benefits of coding and rehearsal. Adults were presented with complex-modeled movement configurations. Some participants coded these movements at the time of presentation by assigning to them numerical or verbal designators. Other participants were not given coding instructions but were told to subdivide the movements to remember them. In addition, participants either were or were not allowed to rehearse the codes or movements following presentation. Both coding and rehearsal enhanced retention of modeled events; individuals who coded and rehearsed showed the best recall.

Production  involves translating visual and symbolic conceptions of modeled events into overt behaviors. Many actions may be learned by simply observing them; subsequent production by observers indicates learning. Rarely, however, are complex behaviors learned solely through observation. Learners often will acquire a rough approximation of a complex skill by observing modeled demonstrations (Bandura,  1977b ). They refine skills with practice, corrective feedback, and reteaching.

Problems in producing modeled behaviors arise not only because information is inadequately coded but also because learners experience difficulty translating coded information in memory into overt action. For example, a child may have a basic understanding of how to tie shoelaces but not be able to translate that knowledge into behavior. Teachers who suspect that students are having trouble demonstrating what they have learned may need to test students in different ways.

Motivation  influences observational learning because people are more likely to engage in the preceding three processes (attention, retention, production) for modeled actions they believe are important. Individuals form expectations about anticipated outcomes of actions based on consequences experienced by them and models (Bandura,  1997 ). They perform those actions they believe will result in rewarding outcomes and avoid acting in ways they believe will be responded to negatively (Schunk,  1987 ). Persons also act based on their values, performing activities they value and avoiding those they find unsatisfying, regardless of the consequences to themselves or others. People forgo money, prestige, and power when they believe activities they must engage in to receive these rewards are unethical (e.g., questionable business practices). Teachers promote motivation in various ways, including making learning interesting, relating material to student interests, having students set goals and monitor goal progress, providing feedback indicating increasing competence, and stressing the value of learning ( Chapter 9 ).

Cognitive Skill Learning

Observational learning expands the range and rate of learning over what could occur through shaping ( Chapter 3 ), where each response must be performed and reinforced. Modeled portrayals of cognitive skills are standard features in classrooms. A teacher explains and demonstrates the skills to be acquired, after which students receive guided practice while the teacher checks for student understanding. Skills are retaught if students experience difficulty. When the teacher is satisfied that students have a basic understanding, they may engage in independent practice while the teacher periodically monitors their work ( Application 4.1 ).

APPLICATION 4.1 Teacher Modeling

Teachers often incorporate modeled demonstrations into lessons designed to teach students diverse skills such as solving mathematical problems, identifying main ideas in text, writing topic sentences, using power tools, and executing defensive basketball maneuvers. Modeled demonstrations can be used to teach elementary school children how to head their papers properly. Ms. Longanecker might draw on the board a sketch of the paper students are using. She then can review the heading procedure step by step, explaining and demonstrating how to complete it.

In a high school biology class, Mrs. Rollacci models how to study for a test. Working through several chapters, she explains and demonstrates how to locate and summarize the major terms and points for each section.

In a middle school life skills class, students can learn how to insert a sleeve into a garment through modeled demonstrations. The teacher might begin by describing the process and then use visual aids to portray the procedure. The teacher could conclude the presentation by demonstrating the process at a sewing machine.

Some students in Dr. Zicklin’s graduate methods course have been coming to his office after class with questions about how to present their findings from their research projects. During the next class, he uses a research project he completed to demonstrate how one might present findings to a group. He uses handouts and slides to illustrate ways to present data.

A drama teacher can model various performance skills while working with students as they practice a play. The teacher can demonstrate desired voice inflections, mood, volume, and body movements for each character in the play.

Many features of instruction incorporate models, and there is much research showing that students of various ages learn skills and strategies by observing models (Horner,  2004 ; Schunk,  2012 ). There even is evidence showing that students who collaboratively observe videos of tutoring sessions subsequently demonstrate greater engagement in learning and long-term retention compared with students who do not observe tutoring videos (Craig, Chi, & VanLehn,  2009 ). Two especially germane applications of modeling to instruction are cognitive modeling and self-instruction.

Cognitive Modeling.

Cognitive modeling  incorporates modeled explanation and demonstration with verbalization of the model’s thoughts and reasons for performing given actions (Meichenbaum,  1977 ; Zimmerman,  2013 ). Coach Martin used cognitive modeling with Donnetta. In teaching division skills, a teacher might verbalize the following in response to the problem 276 ÷ 4:

·  First, I have to decide what number to divide 4 into. I take 276, start on the left, and move toward the right until I have a number the same as or larger than 4. Is 2 larger than 4? No. Is 27 larger than 4? Yes. So my first division will be 4 into 27. Now I need to multiply 4 by a number that will give an answer the same as or slightly smaller than 27. How about 5? 5 × 4 = 20. No, too small. Let’s try 6. 6 × 4 = 24. Maybe. Let’s try 7. 7 × 4 = 28. No, too large. So 6 is correct.

Cognitive modeling can include other types of statements. Errors may be built into the modeled demonstration to show students how to recognize and cope with them. Self-reinforcing statements, such as “I’m doing well,” also are useful, especially with students who encounter difficulties learning and doubt their capabilities to perform well.

Researchers have substantiated the useful role of cognitive modeling and shown that modeling combined with explanation is more effective in teaching skills than explanation alone (Rosenthal & Zimmerman,  1978 ). Schunk ( 1981 ) compared the effects of cognitive modeling with those of didactic instruction on children’s long-division self-efficacy and achievement. Children lacking division skills received instruction and practice. In the cognitive modeling condition, students observed an adult model explain and demonstrate division operations while applying them to sample problems. In the didactic instruction condition, students reviewed instructional material that explained and demonstrated the operations, but they were not exposed to models. Cognitive modeling enhanced children’s division achievement better than did didactic instruction.

Self-Instruction.

Self-instruction  has been used to teach students to regulate their activities during learning (Meichenbaum,  1977 ). In an early study, Meichenbaum and Goodman ( 1971 ) incorporated cognitive modeling into self-instructional training with impulsive second graders in a special-education class. The procedure included:

· ■  Cognitive modeling : Adult tells child what to do while adult performs the task.

· ■ Overt guidance: Child performs under direction of adult.

· ■ Overt self-guidance: Child performs while self-instructing aloud.

· ■ Faded overt self-guidance: Child whispers instructions while performing task.

· ■ Covert self-instruction: Child performs while guided by inner silent speech.

Self-instruction often is used to slow down children’s rate of performing. An adult model used the following statements during a line-drawing task:

·  Okay, what is it I have to do? You want me to copy the picture with the different lines. I have to go slow and be careful. Okay, draw the line down, down, good; then to the right, that’s it; now down some more and to the left. Good, I’m doing fine so far. Remember go slow. Now back up again. No, I was supposed to go down. That’s okay, just erase the line carefully…. Good. Even if I make an error I can go on slowly and carefully. Okay, I have to go down now. Finished. I did it. (Meichenbaum & Goodman,  1971 , p. 117)

Note that the model makes a mistake and shows how to deal with it. This is an important form of learning for students who may become frustrated and quit easily following errors. Meichenbaum and Goodman ( 1971 ) found that cognitive modeling slowed down response times, but that the self-instructions decreased errors.

Self-instruction has been used with a variety of tasks and types of students (Fish & Pervan,  1985 ). It is especially useful for students with learning disabilities (Wood, Rosenberg, & Carran,  1993 ) and for teaching students to work strategically. In teaching reading comprehension, the preceding instructions might be modified as follows: “What is it I have to do? I have to find the topic sentence of the paragraph. The topic sentence is what the paragraph is about. I start by looking for a sentence that sums up the details or tells what the paragraph is about” (McNeil,  1987 , p. 96). Statements for coping with difficulties (“I haven’t found it yet, but that’s all right”) can be built into the modeled demonstration.

Motor Skill Learning

Social cognitive theory postulates that motor skill learning involves constructing a mental model that provides the conceptual representation of the skill for response production and serves as the standard for correcting responses subsequent to receiving feedback (Bandura,  1986 ; McCullagh,  1993 ; Weiss, Ebbeck, & Wiese-Bjornstal,  1993 ). The conceptual representation is formed by transforming observed sequences of behaviors into visual and symbolic codes to be cognitively rehearsed. Individuals usually have a mental model of a skill before they attempt to perform it. For example, by observing tennis players, individuals construct a mental model of such activities as the serve, volley, and backhand. These mental models are rudimentary in that they require feedback and correction to be perfected, but they allow learners to perform approximations of the skills at the outset of training. We saw this in the opening scenario where Donnetta needed to construct a mental model of a backhand. In the case of novel or complex behaviors, learners may have no prior mental model and need to observe modeled demonstrations before attempting the behaviors.

This social cognitive approach differs from other motor learning explanations. Adams’s ( 1971 )  closed-loop theory  postulates that people develop perceptual (internal) traces of motor skill movements through practice and feedback. These traces serve as the reference for correct movements. As one performs a behavior, one receives internal (sensory) and external (knowledge of results) feedback and compares the feedback to the trace. The discrepancy serves to correct the trace. Learning is enhanced when feedback is accurate, and eventually the behavior can be performed without feedback. Adams distinguished two memory mechanisms, one that produces the response and one that evaluates its correctness.

Schema theory  (Schmidt,  1975 ) postulates that people store in memory information regarding motor skill movements, including the initial conditions, the characteristics of the generalized motor sequence, the results of the movement, knowledge of results, and sensory feedback. Learners store this information in two general schemas, or organized memory networks comprising related information. The recall schema deals with response production; the recognition schema is used to evaluate responses.

Social cognitive theory contends that by observing others, people form a cognitive representation that initiates subsequent responses and serves as a standard for evaluating the correctness of responses (Bandura,  1986 ). Motor learning theories differ from social cognitive theory in that the former place greater emphasis on error correction after acting and postulate two memory mechanisms to store information and evaluate accuracy (McCullagh,  1993 ). Social cognitive theory also highlights the role of personal cognitions (goals and expectations) in the development of motor skills ( Application 4.2 ).

A problem in motor skill learning is that learners cannot observe aspects of their performances that lie outside their field of vision. People who are swinging a golf club, hitting a tennis serve, kicking a football, throwing a baseball, or hurling a discus cannot observe many aspects of these sequences. Not being able to see what one is doing requires one to rely on kinesthetic feedback and compare it with one’s conceptual representation. The absence of visual feedback makes learning difficult.

Carroll and Bandura ( 1982 ) exposed learners to models performing a motor skill, and then asked them to reproduce the pattern. The experimenters gave some learners concurrent visual feedback of their performances by running a video camera and allowing them to observe their real-time performances on a monitor. Other learners did not receive visual feedback. When visual feedback was given before learners formed a mental model of the motor behavior, it had no effect on performance. Once learners had an adequate model in mind, visual feedback enhanced their accurate reproduction of the modeled behaviors. Visual feedback eliminated discrepancies between their conceptual models and their actions once the former were in place.

APPLICATION 4.2 Motor Skill Learning

Observational learning is useful for learning motor skills. To teach students to dribble a basketball, physical education teachers might begin with skill exercises, such as standing stationary and bouncing the ball and moving and bouncing the ball with each step. After introducing each skill leading to the final sequence, teachers can demonstrate slowly and precisely what the students are to perform. Students then should practice that skill. If they have difficulty on a particular step, teachers can repeat the modeled demonstration before the students continue practicing.

For high school students to successfully learn a dance to perform in the spring musical, the teacher needs to demonstrate and slowly progress toward putting the dance to music. The teacher may break up the dance, working on each step separately, gradually combining steps and eventually putting all the various steps together with the music.

Researchers also have examined the efficacy of using models to teach motor skills. Weiss ( 1983 ) compared the effects of a silent model (visual demonstration) with those of a verbal model (visual demonstration plus verbal explanation) on the learning of a six-part motor skill obstacle course. Older children (ages 7 through 9 years) learned equally well with either model; younger children (ages 4 through 6 years) learned better with the verbal model. Perhaps the addition of the verbalizations created a cognitive model that helped to maintain children’s attention and assisted with coding of information in memory. Weiss and Klint ( 1987 ) found that children in visual-model and no-model conditions who verbally rehearsed the sequence of actions learned the motor skills better than children who did not verbally rehearse. Collectively these results suggest that verbalization may facilitate motor skill learning.

INFLUENCES ON LEARNING AND PERFORMANCE

Table 4.4 Factors affecting observational learning and performance.

Characteristic	Effects on Modeling
Developmental status	Improvements with development include longer attention and increased capacity to process information, use strategies, compare performances with memorial representations, and adopt intrinsic motivators.
Model prestige and competence	Observers pay greater attention to competent, high-status models. Consequences of modeled behaviors convey information about functional value. Observers attempt to learn actions that they believe they will need to perform.
Vicarious consequences	Consequences to models convey information about behavioral appropriateness and probable outcomes of actions. Valued consequences motivate observers. Similarity in attributes or competence signals appropriateness and heightens motivation.
Outcome expectations	Observers are more likely to perform modeled actions that they believe are appropriate and will result in rewarding outcomes.
Goal setting	Observers are more likely to attend to models who demonstrate behaviors that help observers attain goals.
Values	Observers are more likely to attend to models who display behaviors that the observers believe are important and find self-satisfying.
Self-efficacy	Observers attend to models when they believe they are capable of learning or performing the modeled behavior. Observation of similar models affects self-efficacy (“If they can do it, I can too”).

Observing models does not guarantee that learning will occur or that learned behaviors will be performed later. Several factors influence vicarious learning and performance of learned behaviors ( Table 4.4 ). Developmental status, model prestige and competence, and vicarious consequences are discussed here; outcome expectations, goal setting, values, and self-efficacy are discussed in sections that follow.

Developmental Status of Learners

Learning depends heavily on developmental factors (Wigfield & Eccles,  2002 ), and these include students’ abilities to learn from models (Bandura,  1986 ). Research shows that children as young as 6–12 months can perform behaviors displayed by models (Nielsen,  2006 ), and peer modeling is effective with preschoolers (Ledford & Wolery,  2013 ); however, young children have difficulty attending to modeled events for long periods and distinguishing relevant from irrelevant cues. Information processing functions such as rehearsing, organizing, and elaborating (see  Chapter 5 ) improve with development. Older children have a more extensive knowledge base to help them comprehend new information and are more capable of using memory strategies. Young children may encode modeled events in terms of physical properties (e.g., a ball is round, it bounces, you throw it), whereas older children often represent information visually or symbolically.

With respect to production, information acquired through observation cannot be performed if children lack the requisite physical capabilities. Production also requires translating into action information stored in memory, comparing performance with memorial representation, and correcting performance as necessary. The ability to self-regulate one’s actions for longer periods increases with development. Motivational inducements for action also vary depending on development. Young children are motivated by the immediate consequences of their actions. As children mature, they are more likely to perform modeled actions consistent with their goals and values (Bandura,  1986 ).

Model Prestige and Competence

Modeled behaviors vary in usefulness. Behaviors that successfully deal with the environment command greater attention than those that do so less effectively. People attend to a model in part because they believe they might face the same situation themselves and they want to learn the necessary actions to succeed. Students attend to a teacher because the teacher prompts them but also because they believe they will have to demonstrate the same skills and behaviors. Donnetta attends to her coach because the coach is an expert tennis player and because Donnetta knows she needs to improve her game. When models compete for attention, people are more likely to attend to competent models.

Model competence is inferred from the outcomes of modeled actions (success, failure) and from symbols that denote competence. An important attribute is prestige. Models who have gained distinction are more apt to command attention than those of lower prestige. Attendance usually is higher at a talk given by a well-known person than by one who is less known. In most instances, high-status models have ascended to their positions because they are competent. Their actions have greater functional value for observers, who are apt to believe that rewards will be forthcoming if they act accordingly.

Parents and teachers are high-status models for most children. The scope of adult influence on children’s modeling can generalize to many domains. Although teachers are important models in the development of children’s intellect, their influence typically spreads to such other areas as social behaviors, educational attainments, dress, and mannerisms. The effects of model prestige often generalize to areas in which models have no particular competence, such as when adolescents adopt the dress and products touted by prominent entertainers in commercials (Schunk & Miller,  2002 ). Modeling becomes more prevalent with development, but young children are highly susceptible to adult influence ( Application 4.3 ).

APPLICATION 4.3 Model Attributes

People attend to models partly because they believe they may have to face the same situations themselves. Effective use of model prestige and competence can help motivate secondary students to attend to and learn from lessons.

If the use of alcohol is a problem in a high school, school personnel might deliver a program on alcohol education and abuse (prevention, treatment) to include speakers from outside the school. Influential speakers would be recent high school and college graduates, persons who have successfully overcome problems with alcohol, and those who work with alcohol abusers. The relative similarity in age of the models to the students, coupled with the models’ personal experiences, should make the models appear highly competent. Such individuals might have more impact on the students than literature or lessons taught by teachers and counselors.

At the elementary school level, using peers to help teach academic skills can promote learning and self-efficacy among the learners. Children may identify with other children who have had the same difficulties. A teacher has four students in her class who are having trouble learning to divide. She pairs these four students with students who have demonstrated that they understand how to perform division. A child explaining to a classmate how to solve a division problem will do so in a way that the classmate can understand.

Vicarious Consequences to Models

Vicarious consequences to models can affect observers’ learning and performance of modeled actions. Students who observe models rewarded for their actions are more likely to attend to the models, rehearse and code their actions for retention, and be motivated to perform the same actions. Thus, vicarious consequences serve to inform and  motivate  (Bandura,  1986 ).

Information.

The consequences experienced by models convey information to observers about the types of actions most likely to be effective. Observing competent models perform actions that result in success conveys information to observers about the sequence of actions one should use to succeed. By observing modeled behaviors and their consequences, people form beliefs concerning which behaviors will be rewarded and which will be punished.

In a classic demonstration, Bandura, Ross, and Ross ( 1963 ) exposed children to live aggressive models, filmed aggression, or aggression portrayed by cartoon characters. The models, who pummeled a Bobo doll by hitting, throwing, kicking, and sitting on it, were neither rewarded nor punished, which most likely conveyed to the observers that the modeled behaviors were acceptable. Children subsequently were allowed to play with a Bobo doll. Compared with youngsters not exposed to aggression, children who viewed aggressive models displayed significantly higher levels of aggression. The type of aggressive model (live, filmed, cartoon) made no difference in children’s level of aggression.

Similarity to models is important (Schunk,  1987 ,  2012 ). The more alike observers are to models, the greater is the probability that observers will consider similar actions socially appropriate for them to perform. Most social situations are structured so that behavioral appropriateness depends on factors such as age, gender, or status. Modeled tasks with which observers are unfamiliar or those that are not immediately followed by consequences may be highly influenced by model similarity (Akamatsu & Thelen,  1974 ).

Although some research shows that children are more likely to attend to and learn from same-sex models (Maccoby & Jacklin,  1974 ), other research suggests that model gender has a greater effect on performance than on learning (Bandura & Bussey,  2004 ; Perry & Bussey,  1979 ; Spence,  1984 ). Children learn from models of both sexes and categorize behaviors as appropriate for both sexes or as more appropriate for members of one sex. Model gender, therefore, seems important as a conveyor of information about task appropriateness (Zimmerman & Koussa,  1975 ). When children are uncertain about the gender appropriateness of a modeled behavior, they may model same-sex peers because they are more likely to think that those actions are socially acceptable.

Model–observer similarity in age is important when children perceive the actions of same-age peers to be more appropriate for themselves than the actions of younger or older models (Schunk,  1987 ). Brody and Stoneman ( 1985 ) found that in the absence of competence information, children were more likely to model the actions of same-age peers. When children were provided with competence information, modeling was enhanced by similar competence regardless of model age.

Although children learn from models of any age (Schunk,  1987 ), peers and adults use different teaching strategies. Peers often use nonverbal demonstrations and link instruction to specific items (e.g., how to do it); adults typically employ more verbal instruction stressing general principles and relate information to be learned to other material (Ellis & Rogoff,  1982 ). Peer instruction may be especially beneficial with students who have experienced learning problems and with those who have difficulty processing verbal information.

The highest degree of model–observer similarity occurs when one is one’s own model ( self-modeling ), which has been used to develop social, vocational, motor, cognitive, and instructional skills (Bellini & Akullian,  2007 ; Dowrick,  1983 ,  1999 ; Hartley, Bray, & Kehle,  1998 ; Hitchcock, Dowrick, & Prater,  2003 ). In a typical procedure, a person’s performance is recorded, and he or she subsequently views the recording. Observing a self-modeled performance is a form of review and is especially informative for skills one cannot watch while performing (e.g., gymnastics). Viewing video feedback of a skillful performance conveys that one is capable of learning and can continue to make progress with further work, which raises self-efficacy (Fukkink, Trienekens, & Kramer,  2011 ).

Schunk and Hanson ( 1989b ) found benefits of self-modeling during acquisition of arithmetic (fraction) skills. Children received instruction and problem-solving practice. Self-modeling students were videotaped while successfully solving problems and were shown their tapes, others were videotaped but not shown their tapes until after the study was completed (to control for effects of taping), and students in a third condition were not taped (to control for effects of participation). Self-modeling children scored higher on self-efficacy for learning, motivation, and posttest self-efficacy and achievement. Researchers found no differences between mastery self-model students who viewed tapes of their successful problem solving and self-model children whose tapes portrayed their gradual improvement as they acquired skills, which supports the point that the perception of progress or mastery can build efficacy (Schunk & Pajares,  2009 ).

Motivation.

Observers who see models rewarded become motivated to act accordingly. Perceived similarity enhances these motivational effects, which depend in part on self-efficacy (Bandura,  1982b ,  1997 ). By observing similar others succeed, students are apt to believe that if others can succeed, they can as well. Such motivational effects are common in classrooms. Learners who observe other students performing a task well may be motivated to try their best.

Of particular importance is the observation of effort that leads to success (Schunk,  1995 ). Seeing others succeed with effort and receiving praise from teachers may motivate observing peers to work harder. Students may become more motivated by watching similar others succeed than by those who they believe are superior in competence.

But vicarious success will not sustain behavior over long periods. Although motivation is boosted when students observe teachers giving praise and high grades to others for hard work and good performances, motivation is sustained over time when students believe their own efforts are leading to better performances.

MOTIVATIONAL PROCESSES

Among the important influences on enactive and vicarious learning and on performance of learned behaviors are observers’ goals, outcome expectations, values, and self-efficacy. This section covers the first three; self-efficacy is addressed in the next section.

Goals

Much human behavior sustained over long periods in the absence of immediate external incentives depends on goal setting and self-evaluations of progress. A  goal  reflects one’s purpose and refers to quantity, quality, or rate of performance (Locke & Latham,  1990 ,  2002 ; Locke, Shaw, Saari, & Latham,  1981 ).  Goal setting  involves establishing a standard or objective to serve as the aim of one’s actions. People can set their own goals or goals can be established by others (parents, teachers, supervisors).

Goals were a central feature of Tolman’s ( 1932 ,  1942 ,  1951 ,  1959 ) theory of  purposive behaviorism . Like most psychologists of his time, Tolman was trained in behaviorism. His experiments resembled those of Thorndike and Skinner ( Chapter 3 ) because they dealt with responses to stimuli under varying environmental conditions. But he disagreed with conditioning theorists over their view of behavior as a series of stimulus–response connections. He contended that learning is more than the strengthening of responses to stimuli, and he recommended a focus on  molar behavior —a large sequence of goal-directed behavior.

The “purposive” aspect of Tolman’s ( 1932 ) theory refers to his belief that behavior is goal directed. Stimuli in the environment (e.g., objects, paths) are means to goal attainment. They cannot be studied in isolation; rather, entire behavioral sequences must be studied to understand why people engage in particular actions. High school students whose goal is to attend a leading university study hard in their classes. By focusing only on the studying, researchers miss the purpose of the behavior. The students do not study because they have been reinforced for studying in the past (i.e., by getting good grades). Rather, studying is a means to intermediate goals (e.g., learning, high grades), which, in turn, enhance the likelihood of acceptance to the university.

Tolman qualified his use of “purposive” by noting that it is defined objectively. The behavior of people and animals is goal oriented. They act “as if” they are pursuing a goal and have chosen a means for attainment. Thus, Tolman went well beyond simple stimulus–response associations to discuss underlying cognitive mechanisms.

Social cognitive theory contends that goals enhance learning and performance through their effects on perceptions of progress, self-efficacy, and self-evaluations (Bandura,  1988 ,  1997 ; Locke & Latham,  1990 ,  2002 ; Schunk,  1990 ). Initially, people must make a commitment to attempt to attain their goals because goals do not affect performance without commitment. As they work on the task, they compare their current performances with their goals. Positive self-evaluations of progress raise self-efficacy and sustain motivation. A perceived discrepancy between present performance and the goal may create dissatisfaction, which can enhance effort.

Goals motivate people to exert effort necessary to meet task demands and persist over time (Locke & Latham,  1990 ,  2002 ). Goals also direct attention to relevant task features and behaviors to be performed, and they can affect how learners process information. Goals give people “tunnel vision” to focus on the task, select task-appropriate strategies, and decide on the effectiveness of their approach, all of which are likely to raise performance.

But goals, by themselves, do not automatically enhance learning and motivation. Rather, the properties of specificity, proximity, and difficulty enhance self-perceptions, motivation, and learning (Locke & Latham,  2002 ; Nussbaum & Kardash,  2005 ;  Application 4.4  and  Table 4.5 ).

Specificity.

Goals that incorporate specific standards of performance are more likely to enhance learning and activate self-evaluations than are general goals (e.g., “Do your best”; Locke & Latham,  2002 ). Specific goals boost task performance by providing more information about how much effort is required to succeed, and they promote self-efficacy because it is easy to evaluate progress toward an explicit goal.

Much research attests to the effectiveness of specific goals in raising performance (Bandura,  1988 ; Locke & Latham,  1990 ,  2002 ; Schunk,  2012 ). Schunk ( 1983b ) provided children with instruction and practice solving long-division problems. During the sessions, some children received a specific goal denoting the number of problems to complete; others had a general goal to work productively. Within each condition, half of the children received social comparative information on the number of problems that peers completed (which matched the session goal) to convey that goals were attainable. Goals raised self-efficacy; goals plus comparative information led to the highest self-efficacy and achievement.

APPLICATION 4.4 Goal Properties

Goal properties are easily incorporated into lessons. In his fourth-grade class, Mr. Zumbreski introduced a new spelling unit by stating the following goal:

·  Of our 20 words this week, I know that all of you will be able to learn to spell the first 15. We are going to work very diligently in class on these words, and I expect you to do the same at home. With our work at school and at home, I know that all of you will be able to spell these words correctly by Friday. The last 5 words are more difficult. These will be our bonus words.

This goal is specific, but for some children it is distant and might be viewed as too difficult. To ensure that all students achieve the overall goal, Mr. Zumbreski sets short-term goals each day: “Today we are going to work on these 5 words. By the end of class time I know that you will be able to spell these 5 words.” Children should view the daily goals as easier to attain than the weekly goal. To further ensure goal attainment, he will make sure that the 15 words selected for mastery by Friday challenge the students but are not overly difficult.

A teacher working with students on keyboarding might establish a words-per-minute goal for students to reach by the end of the semester:

·  Students, this semester I know that all of you will be able to learn to use the keyboard. Some of you, because of other experiences or talents, will be able to type faster, but I know that all of you will be able to enter at least 30 words per minute with no mistakes by the end of the semester.

To help students achieve this goal, the teacher might set weekly short-term goals. The first week the goal might be 10 words per minute with no mistakes, the second week 12 words per minute, and so forth, increasing the number each week.

Table 4.5 Goal properties and their effects.

Goal Property	Effects on Behavior
Specificity	Goals with specific standards of performance increase motivation and raise self-efficacy because goal progress is easy to gauge.
Proximity	Proximal goals increase motivation and self-efficacy and are especially important for young children who may not divide a long-term goal into a series of short-term goals.
Difficulty	Challenging but attainable goals raise motivation and self-efficacy better than easy or hard goals.

Schunk ( 1984a ) compared the effects of goals with those of rewards. Children received long-division instruction and practice over sessions. Some were offered rewards based on the number of problems completed, others pursued goals (number of problems to complete), and children in a third condition received rewards and goals. The three conditions promoted motivation during the sessions; rewards plus goals resulted in the highest division self-efficacy and achievement. Combining rewards with goals provided children with two sources of information to use in gauging learning progress.

Proximity.

Goals are distinguished by how far they project into the future. Proximal, short-term goals are closer at hand, are achieved quicker, and result in greater motivation than more temporally distant, long-term goals. Although benefits of proximal goals are found regardless of developmental status, short-term goals are needed with children because they have short time frames of reference and are not fully capable of representing distant outcomes in thought (Bandura & Schunk,  1981 ). Proximal goals fit well with lesson planning as teachers plan activities around blocks of time, such as when teachers ask children to complete 10 problems (specific) in 15 minutes (proximal).

Bandura and Schunk ( 1981 ) gave children subtraction instruction with practice opportunities over seven sessions. Children received seven packets of material. Some pursued a proximal goal of completing one packet each session; a second group received a distant goal of completing all packets by the end of the last session; a third group was given a general goal of working productively. Proximal goals led to the highest motivation during the sessions, as well as the highest subtraction self-efficacy, achievement, and intrinsic interest (based on the number of problems solved during a free-choice period). The distant goal resulted in no benefits compared with the general goal. Manderlink and Harackiewicz ( 1984 ) found that proximal and distant goals did not differentially affect adults’ performances on word puzzles, but proximal goals led to higher expectations of goal attainment and perceived competence.

Difficulty.

Goal difficulty refers to the level of task proficiency required as assessed against a standard. Individuals expend greater effort to attain a difficult goal than an easy one (Locke & Latham,  2002 ); however, difficulty level and performance do not bear an unlimited positive relationship to each other. Difficult goals do not enhance performance in the absence of needed skills. Self-efficacy also is important. Learners who think they cannot reach a goal hold low self-efficacy, do not commit to attempting the goal, and work halfheartedly.

Schunk ( 1983c ) gave children a difficult (but attainable) or an easier goal of completing a given number of long-division problems during each instructional session. To prevent students from believing goals were too difficult, the teacher gave half of each group attainment information (“You can work 25 problems”); the other half received social comparative information indicating that similar peers completed that many. Difficult goals enhanced motivation; children who received difficult goals and attainment information displayed the highest self-efficacy and achievement. Locke, Frederick, Lee, and Bobko ( 1984 ) found that assigning college students difficult goals led to better performance and to their subsequently setting higher goals for themselves compared with students who initially were allowed to set their own goals. When participants set their own goals, self-efficacy related positively to goal level and commitment.

Self-Set Goals.

Researchers have found that allowing students to set their goals enhances self-efficacy and learning, perhaps because self-set goals produce high goal commitment. Schunk ( 1985 ) provided subtraction instruction to sixth graders with learning disabilities. Some set daily  performance goals , others had comparable goals assigned, and a third group worked without goals. Self-set goals led to the highest judgments of confidence for attaining goals, self-efficacy for solving problems, and subtraction achievement. Children in the two goal groups demonstrated greater motivation during the instructional sessions than did those without goals.

Hom and Murphy ( 1985 ) put college students who were high or low in achievement motivation in self-set or assigned-goal conditions. Self-set participants decided how many anagrams they could solve; assigned-goal participants were given comparable goals. Students high in achievement motivation performed equally well under the two goal conditions; self-set goals enhanced the performances of students low in achievement motivation.

Goal Progress Feedback.

Goal progress feedback provides information about progress toward goals (Hattie & Timperley,  2007 ). Such feedback, which is especially valuable when people cannot derive reliable information on their own, should raise self-efficacy, motivation, and achievement when it informs people that they are competent and can continue to improve by working diligently. Higher self-efficacy sustains motivation when people believe that continued effort will allow them to attain their goals. Once individuals attain goals, they are likely to set new goals (Schunk,  2012 ).

Schunk and Rice ( 1991 ) taught students who had experienced reading difficulties a strategy to answer comprehension questions. Children were given a product goal of answering questions, a process goal of learning to use the strategy, or a process goal plus progress feedback that conveyed they were making progress toward their goal of learning to use the strategy to answer questions. Following the instruction, goal-plus-feedback children demonstrated higher reading self-efficacy and achievement than did learners assigned to the process and product goal conditions. Schunk and Swartz ( 1993a ,  1993b ) obtained comparable results in writing achievement with average-achieving and academically gifted elementary school children. Self-efficacy and achievement gains generalized across types of writing tasks and maintained themselves over time.

Outcome Expectations

Outcome expectations  are personal beliefs about the anticipated outcomes of actions (Schunk & Zimmerman,  2006 ). Outcome expectations were among the first cognitive variables to be included in explanations of learning. Tolman ( 1932 ,  1949 ) defined  field expectancies  as involving relations between stimuli (S1–S2) or among a stimulus, response, and stimulus (S1–R–S2). Relations between stimuli concern what stimulus is apt to follow what other stimulus; for example, thunder follows lightning. In three-term relations, people develop the belief that a certain response to a given stimulus produces a certain result. If one’s goal is to get to a roof (S2), the sight of the ladder (S1) could lead one to think, “If I place this ladder against the house (R), I can get to the roof.” This is similar to Skinner’s ( 1953 ;  Chapter 3 ) three-term contingency except that Tolman conceived of this type of relation as reflecting a cognitive expectancy.

Field expectancies helped people form  cognitive maps , or internal plans comprising expectancies of which actions are needed to attain goals. People follow signs to a goal; they learn meanings rather than discrete responses. People use their cognitive maps to determine the best course of action to attain a goal.

Figure 4.2 Experimental arrangement to study expectancy learning.

Source: Adapted from content of article, “Studies in Spatial Learning,” by E. C. Tolman, B. F. Ritchie, and D. Kalish, 1946, Journal of Experimental Psychology, 36, pp. 13–24.

Tolman tested his ideas in an ingenious series of experiments (Tolman, Ritchie, & Kalish,  1946a ,  1946b ). In one study, rats were trained to run an apparatus, shown in  Figure 4.2  (Maze 1). Subsequently, the apparatus was replaced with one in which the original path was blocked. Conditioning theories predict that animals will choose a path close to the original one, as shown in  Figure 4.2  (Maze 2a). In fact, rats most frequently chose a path following the direction in which they originally found food (Maze 2b). These results supported the idea that the animals formed a cognitive map of the location of the food and responded based on that map rather than on prior responses to stimuli.

Social cognitive theory contends that people form outcome expectations based on their personal experiences (Bandura,  1986 ,  1997 ). Individuals act in ways they believe will be successful and attend to models who teach them valued skills. Outcome expectations sustain behaviors over long periods when people believe their actions will eventually produce desired outcomes. They also figure prominently in transfer; people are apt to engage in actions in new situations that were successful in previous situations because they believe that similar consequences will follow.

Outcome expectations can refer to external outcomes (“If I try my best on this exam, I will make a good grade on it”) or to internal ones (“If I try my best on this exam, I will feel good about myself”). An important type of outcome expectation relates to progress in skill learning (“If I try my best, I will become a better reader”). Students who believe they are making little or no progress in learning may become demoralized and lackadaisical. Progress in academic learning often occurs slowly and students notice little day-to-day change. For example, learners may improve their skills in reading longer and more difficult passages, in finding main ideas, in drawing inferences, and in reading for details, but progress is slow. Teachers can inform students of their reading comprehension progress when it is not immediately apparent.

The influential role of outcome expectations was demonstrated by Shell, Murphy, and Bruning ( 1989 ). College students completed measures of reading and writing self-efficacy, outcome expectancies, and achievement. The self-efficacy assessment asked students to rate their competencies in performing various reading and writing tasks (e.g., letter from a friend, employment application, short fiction story). For the outcome expectancy measure, students judged the importance of reading and writing for achieving such life goals as getting a job, being financially secure, and being happy.

Self-efficacy and outcome expectancies related positively to achievement in reading and writing. In both domains, self-efficacy was more strongly related to achievement than outcome expectancies. This study also showed that the expectancy beliefs for each domain related significantly to achievement in the other domain, which suggests that teachers’ attempts to improve students’ self-efficacy and outcome expectations in one literacy area may generalize to others.

Values

Value  refers to the perceived importance or usefulness of learning. An important premise of social cognitive theory is that individuals’ actions reflect their value preferences (Bandura,  1986 ). Learners do things that bring about what they desire and work to avoid outcomes that are inconsistent with their values. Learners are motivated to learn and perform when they deem that learning or performance important.

Values can be assessed against external and internal standards. There are many reasons why students might value high grades. Making As and the honor roll may bring them recognition (i.e., from parents and teachers) and acceptances at universities. But high grades also can produce internal satisfaction, as when students feel proud of their work and a sense of accomplishment. Such internal satisfaction also occurs when learners act in accordance with their ethical beliefs.

Values are covered in more depth in  Chapter 9  because they figure prominently in theories of motivation. Values are intimately linked with the other motivational processes discussed here: goals, outcome expectations, and self-efficacy. For example, assume that Larissa’s family has moved and that Larissa (a fifth grader) is starting at a new school. One of her goals is to make new friends. She values friendships; she enjoys spending time with other children and sharing on a personal level with them (she has no brothers or sisters). She believes that if she is nice to other children that they will be nice to her and may become her friends (positive outcome expectations). Although she is somewhat shy initially in her new school, she has made new friends before and feels reasonably self-efficacious about doing so again. Larissa observes the actions of her new peers to learn what types of things they like to do. She interacts with her peers in ways that she believes will lead to friendships, and as she begins to develop new friends, her social self-efficacy becomes strengthened.

An important part of a teacher’s job is to determine students’ value preferences and especially if any of these reflect stereotypes or cultural differences. Research by Wigfield and Eccles ( 1992 ) showed some stereotypes among adolescents: Boys valued mathematics more, whereas girls placed greater emphasis on English. Mickelson ( 1990 ) contended that perceived racial inequalities can result in some minority students devaluing school achievement. Teachers have the responsibility of promoting achievement values in all students, which they can do by teaching students how to set goals and assess their progress, showing students how their achievement results in positive outcomes, and building learners’ self-efficacy.

SELF-EFFICACY

Conceptual Overview

Self-efficacy (efficacy expectations)  refers to personal beliefs about one’s capabilities to learn or perform actions at designated levels (Bandura,  1977a ,  1977b ,  1986 ,  1993 ,  1997 ). Self-efficacy is a belief about what one is capable of doing; it is not the same as knowing what to do. In gauging self-efficacy, individuals assess their skills and their capabilities to translate those skills into actions. Self-efficacy is a key to promoting a sense of  agency  in people that they can influence their lives (Bandura,  1997 ,  2001 ).

Self-efficacy and outcome expectations do not have the same meaning (Schunk & Zimmerman,  2006 ). Self-efficacy refers to perceptions of one’s capabilities to produce actions; outcome expectations involve beliefs about the anticipated outcomes of those actions. For example, Jeremy may believe that if he correctly answers the teacher’s questions, the teacher will praise him (positive outcome expectation). He also may value praise from the teacher. But he may not attempt to answer the teacher’s questions if he doubts his capabilities to answer them correctly (low self-efficacy).

Despite self-efficacy and outcome expectations being conceptually distinct, they often are related (Schunk,  2012 ). Students who typically perform well have confidence in their learning capabilities and expect (and usually receive) positive outcomes for their efforts. At the same time, there is no necessary relation between self-efficacy and outcome expectations. Even students with high self-efficacy for learning may expect a low grade as an outcome if they think that the teacher does not like them.

Although some evidence indicates that perceptions of self-efficacy generalize to different tasks (Smith,  1989 ), theory and research suggest that self-efficacy is primarily domain specific (Pajares,  1996 ,  1997 ; Schunk & Pajares,  2009 ). Thus, it is meaningful to speak of self-efficacy for drawing inferences from text, balancing chemical equations, solving fractions, running certain times at track events, performing computer operations, and so on. Smith and Fouad ( 1999 ) found that self-efficacy, goals, and outcome expectations are specific to subject areas and show little generalization across areas. Self-efficacy might transfer to new situations, however, when learners believe that the same skills will produce success. Learners who feel self-efficacious about outlining in English class also may feel confident about outlining in science class, and their self-efficacy may motivate them to construct an outline in science.

Self-efficacy is distinguished from  self-concept  (Pajares & Schunk,  2002 ; Schunk & Pajares,  2005 ,  2009 ), which refers to one’s collective self-perceptions formed through experiences with and interpretations of the environment and which depends heavily on reinforcements and evaluations by significant others (Shavelson & Bolus,  1982 ; Wylie,  1979 ). Self-efficacy refers to perceptions of specific capabilities; self-concept is one’s general self-perception that includes self-efficacy in different areas (Schunk & Zimmerman,  2006 ; see  Chapter 9 ).

Self-efficacy depends in part on student abilities. In general, high-ability students feel more efficacious about learning compared with low-ability students; however, self-efficacy is not another name for ability. Collins ( 1982 ) identified high-, average-, and low-ability students in mathematics. Within each level, she found students of high and low self-efficacy. She gave students problems to solve, and told them they could rework those they missed. Ability was positively related to achievement, but regardless of ability level, students with high self-efficacy solved more problems correctly and chose to rework more problems they missed than those with low self-efficacy.

Self-efficacy can have diverse effects in achievement settings (Bandura,  1993 ; Pajares,  1996 ,  1997 ; Schunk,  2012 ; Schunk & Pajares,  2009 ). Self-efficacy can influence choice of activities. Students with low self-efficacy for learning may avoid attempting tasks; those who judge themselves efficacious should participate more eagerly. Self-efficacy also can affect effort expenditure, persistence, and learning. Students who feel efficacious about learning generally expend greater effort and persist longer than students who doubt their capabilities, especially when they encounter difficulties. In turn, these behaviors promote learning.

People acquire information about their self-efficacy from their performances, observations of models (vicarious experiences), forms of social persuasion, and physiological indexes (e.g., heart rate, sweating). Actual performances offer the most valid information for assessing efficacy. Successes generally raise efficacy and failures lower it, although an occasional failure (success) after many successes (failures) should not have much effect.

Students acquire much self-efficacy information through knowledge of how others perform. Similarity to others is an important cue for gauging one’s self-efficacy (Brown & Inouye,  1978 ; Schunk & Pajares,  2009 ). Observing similar others succeed raises observers’ self-efficacy and motivates them to try the task because they believe that if others can succeed, they can as well. At the same time, a vicarious increase in self-efficacy can be negated by subsequent personal failures. Students who observe peers fail may believe they lack the competence to succeed, which can dissuade them from attempting the task. Donnetta experienced some increase in self-efficacy from watching her coach demonstrate the backhand, but her doing it without hitting into the net is a more potent influence.

Students often receive persuasive information from teachers that they possess the capability to perform well (e.g., “You can do it”). Although positive feedback enhances self-efficacy, this increase will not endure for long if students subsequently perform poorly. Learners also acquire some self-efficacy information from physiological symptoms they experience. They may interpret emotional symptoms (e.g., sweating, trembling) to mean that they are not capable of learning. When learners notice they are experiencing less stress in response to academic demands, they may feel more efficacious for mastering the task.

Information acquired from these sources does not influence self-efficacy automatically but is cognitively appraised (Bandura,  1982b ,  1993 ,  1997 ). Appraising self-efficacy is an inferential process in which persons weigh and combine the contributions of personal, behavioral, and environmental factors. In forming self-efficacy beliefs, students consider factors such as perceptions of their ability, effort expended, task difficulty, teacher assistance, and number and pattern of successes and failures (Bandura,  1981 ,  1997 ; Schunk,  2012 ).

Self-Efficacy in Achievement Situations

Self-efficacy is germane to academic learning. Researchers have found support for the hypothesized effects of self-efficacy on choice, effort, persistence, and achievement (Pajares,  1996 ,  1997 ; Schunk & Pajares,  2005 ,  2009 ). Self-efficacy is related as well to career choices. Betz and Hackett ( 1981 ,  1983 ; Hackett & Betz,  1981 ) found that although there are structural and social influences on career choices, self-efficacy is an important mediator of these external influences and has a direct bearing on career choices. In addition, gender differences that emerge in vocational choices are due to differences in self-efficacy. Women are more self-efficacious for careers traditionally held by women than for careers traditionally held by men, whereas men’s self-efficacy is less dependent on career gender typing.

Self-efficacy is strongly related to effort and task persistence (Bandura & Cervone,  1983 ,  1986 ; Schunk & Pajares,  2009 ). Individuals with high self-efficacy beliefs are likely to exert effort in the face of difficulty and persist at a task when they have the requisite skills. There is, however, some evidence that self-doubts may foster learning when students have not previously acquired the skills. As Bandura ( 1986 ) noted, “Self-doubt creates the impetus for learning but hinders adept use of previously established skills” (p. 394). Salomon ( 1984 ) found that students high in self-efficacy were more likely to be cognitively engaged in learning when the task was perceived as difficult but less likely to be effortful and less cognitively engaged when the task was deemed easy. Besides the quantity of effort, the quality of effort (deeper cognitive processing and general cognitive engagement) has been linked to self-efficacy (Graham & Golan,  1991 ; Pintrich & Schrauben,  1992 ). Pintrich and De Groot ( 1990 ) showed that junior high students high in self-efficacy were more likely to report using cognitive and self-regulatory learning strategies.

With respect to achievement and cognitive performance, Schunk ( 1982a ,  1982b ,  1983a ,  1983b ,  1983c ,  1983d ,  1984a ,  1984b ,  1996 ) found in a series of experimental studies that self-efficacious students mastered various academic tasks better than students with weaker self-efficacy. Students’ computer self-efficacy relates positively to their success in computer-based learning environments (Moos & Azevedo,  2009 ). Self-efficacy is a significant predictor of learning and achievement even after prior achievement and cognitive skills are taken into account (Schunk,  1981 ,  1982a ). Resuts of a meta-analysis by Beaudoin and Desrichard ( 2011 ) showed that memory self-efficacy related positively to memory performance.

In summary, self-efficacy is an important influence on motivation and achievement (Multon, Brown, & Lent,  1991 ; Pajares,  1996 ,  1997 ; Schunk & Pajares,  2005 ,  2009 ; Valentine, DuBois, & Cooper,  2004 ). Self-efficacy is assumed to be more situationally specific, dynamic, fluctuating, and changeable than the more static and stable measures of self-concept and general self-competence (Schunk & Pajares,  2002 ). One’s self-efficacy for a specific task might fluctuate due to one’s preparation, physical condition (sickness, fatigue), and affective mood, as well as external conditions such as the nature of the task (length, difficulty) and social milieu (general classroom conditions). In contrast, other views of self-competence view it more globally (e.g., mathematical competence) and are less concerned with instability of beliefs. In measuring self-efficacy, therefore, it is important to link it specifically with the key processes and capabilities underlying successful performance in the domain being assessed (Bruning, Dempsey, Kauffman, McKim, & Zumbrunn,  2013 ).

The reciprocal interaction between personal and environmental factors can be seen clearly with social and self variables. Social (environmental) factors can affect many self (personal) variables, such as learners’ goals, self-efficacy, outcome expectations, attributions, self-evaluations of learning progress, and self-regulatory processes. In turn, self influences can affect social environments, as when learners decide they need more instruction on a skill and seek out a qualified teacher (Schunk,  1999 ).

Achievement variables such as perceived goal progress, motivation (e.g., choice of activities, effort, persistence), and learning are affected by social and self influences. In turn, learner actions affect these factors. As students work on tasks, they evaluate their learning progress. Perceptions of progress, which can be facilitated by feedback about progress, substantiate their self-efficacy for learning, which sustains motivation and learning (Hattie & Timperley,  2007 ; Schunk & Pajares,  2009 ).

A key process is the  internalization  of social variables to self influences. Learners transform information acquired from the social environment into mechanisms of self-regulation ( Chapter 10 ). With increased skill acquisition, this social-to-self transformation process becomes a bidirectional interactive process as learners alter and adjust their social environments to further enhance their achievement (Schunk,  1999 ).

Models and Self-Efficacy

The models in students’ environments (e.g., parents, teachers, coaches, peers) provide important sources of information for gauging self-efficacy.

Adult Models.

Research shows that exposing students to adult models influences their self-efficacy for learning and performing well. Zimmerman and Ringle ( 1981 ) had children observe a model unsuccessfully attempt to solve a puzzle for a long or short time and verbalize statements of confidence or pessimism, after which children attempted to solve the puzzle. Observing a confident but nonpersistent model raised self-efficacy; children who observed a pessimistic but persistent model lowered their self-efficacy. Relich, Debus, and Walker ( 1986 ) found that exposing low-achieving children to models explaining mathematical division and providing them with feedback stressing the importance of ability and effort positively affected self-efficacy.

Schunk ( 1981 ) showed that both cognitive modeling and didactic instruction by adults raised self-efficacy; however, cognitive modeling led to greater gains in division skill and to more accurate perceptions of capabilities as these children’s self-efficacy judgments corresponded more closely to their actual performances. Students who received only didactic instruction overestimated what they could do. Regardless of treatment condition, self-efficacy related positively to persistence and achievement. Bandura, Barbaranelli, Caprara, and Pastorelli ( 1996 ) found that parents’ academic aspirations for their children affected both children’s academic achievements and their self-efficacy.

Peer Models.

One way to raise self-efficacy is to use  coping models , who initially demonstrate fears and skill deficiencies but gradually improve their performance and self-efficacy. Coping models illustrate how determined effort and positive self-thoughts overcome difficulties. In contrast,  mastery models demonstrate faultless performance and high confidence from the outset (Thelen, Fry, Fehrenbach, & Frautschi,  1979 ). Coping models may enhance perceived similarity and self-efficacy for learning better than mastery models among students who are more likely to view the initial difficulties and gradual progress of coping models as more similar to their typical performances than the rapid learning of mastery models.

Children who had experienced difficulties learning subtraction with regrouping watched videos portraying a peer mastery model, a peer coping model, a teacher model, or no model (Schunk & Hanson,  1985 ). In the peer-model conditions, an adult teacher provided instruction, after which the peer solved problems. The peer mastery model easily grasped operations and verbalized positive achievement beliefs reflecting high self-efficacy and ability, low task difficulty, and positive attitudes. The peer coping model initially made errors and verbalized negative achievement beliefs but gradually performed better and verbalized coping statements (e.g., “I need to pay attention to what I’m doing”). Eventually, the coping model’s problem-solving behaviors and verbalizations matched those of the mastery model. Teacher-model children observed videos portraying only the teacher providing instruction; no-model children did not view videos. All children judged self-efficacy for learning to subtract and received instruction and practice over sessions.

Observing a peer model raised self-efficacy and achievement more than observing a teacher model or no model; the teacher-model condition promoted these outcomes better than no model. The mastery and coping conditions led to similar outcomes. Possibly children focused more on what the models had in common (task success) than on their differences. Children may have drawn on their prior successes in subtraction without regrouping and concluded that if the model could learn, they could as well.

Another important variable is number of models. Compared with a single model, multiple models increase the probability that observers will perceive themselves as similar to at least one of the models (Thelen et al.,  1979 ). Students who might easily discount the successes of a single model may be swayed by observing several successful peers and think that if all these models can learn, they can as well. Notice in the opening scenario that Donnetta’s coach served as a model, and she gave Donnetta materials portraying backhands demonstrated by other models.

Schunk, Hanson, and Cox ( 1987 ) investigated the effects of single and multiple coping and mastery models with a task (fractions) on which children had experienced few prior successes. Viewing a single coping model or multiple coping or mastery models enhanced children’s self-efficacy and achievement better than viewing a single mastery model.

Schunk and Hanson ( 1989a ) further explored variations in perceived similarity by having average-achieving children view one of three types of peer models. Mastery models easily grasped arithmetic operations and verbalized positive beliefs (e.g., “I know I can do this one”). Coping-emotive models initially experienced difficulties and verbalized negative statements (e.g., “I’m not very good at this”), after which they verbalized coping statements (e.g., “I’ll have to work hard on this one”) and displayed coping behaviors; eventually they performed as well as mastery models. Coping-alone models performed in identical fashion to coping-emotive models but never verbalized negative beliefs.

Coping-emotive models led to the highest self-efficacy for learning. Mastery and coping-alone children perceived themselves as equal in competence to the model; coping-emotive children viewed themselves as more competent than the model. The belief that one is more talented than an unsuccessful model can raise self-efficacy and motivation. The three conditions promoted self-efficacy and achievement equally well, which shows that actual task experience outweighed initial effects due to watching models.

Peer models can increase prosocial behaviors. Strain et al. ( 1981 ) taught peers to initiate social play with withdrawn children by using verbal signals (e.g., “Let’s play blocks”) and motor responses (handing child a toy). Training peer initiators is time consuming but effective because methods of remedying social withdrawal (prompting, reinforcement) require nearly continuous teacher involvement.  Application 4.5 discusses some additional uses of peer models.

Motor Skills

Self-efficacy has been shown to predict the acquisition and performance of motor skills (Bandura,  1997 ; Poag-DuCharme & Brawley,  1993 ; Wurtele,  1986 ). Gould and Weiss ( 1981 ) found benefits due to model similarity. College women viewed a similar model (female student with no athletic background) or dissimilar model (male physical education professor) perform a muscular endurance task. Students who viewed the similar model performed the task better and judged self-efficacy higher than those who observed the dissimilar model. Regardless of treatment condition, self-efficacy related positively to performance.

George, Feltz, and Chase ( 1992 ) replicated these results using female college students and models performing a leg-extension endurance task. Students who observed nonathletic male or female models extended their legs longer and judged self-efficacy higher than those who observed an athletic model. Among these unskilled observers, model ability was a more important similarity cue than model gender.

APPLICATION 4.5 Building Self-Efficacy with Peer Models

Observing similar peers performing a task increases students’ self-efficacy for learning. This idea is applied when a teacher selects certain students to solve mathematics problems while class members observe. By demonstrating success, the peer models help raise observers’ self-efficacy for performing well. If ability levels in a class vary considerably, the teacher might pick peer models at different levels of ability. Students in the class are more likely to perceive themselves as similar in competence to at least one of the models.

Peers who readily master skills may help teach skills to observing students but may not have much impact on the self-efficacy of those students who experience learning difficulties. For the latter, students who learn more slowly may be excellent models. Mr. Riordian’s history class has been learning the Civil War battles. Because so many battles occurred, learning all of them has been difficult for some of the students. He places students into three groups: Group 1—students who mastered the material easily; Group 2—students who have been working hard and are gradually developing mastery; and Group 3—students who still are having difficulty. He pairs Groups 2 and 3 for  peer tutoring , figuring that Group 2 students will be good models for Group 3 students.

Teachers can point out the concentration and hard work of peer models. For instance, as an elementary teacher moves about the room monitoring students’ work she provides learners with social comparative information (e.g., “See how well Kevin is working? I’m sure that you can work just as well”). Teachers need to ensure that learners view the comparative performance level as one they can attain; judicious selection of referent students is necessary.

Peers also can enhance students’ self-efficacy during small-group work. Successful groups are those in which each member has some responsibility and members share rewards based on their collective performance. The use of such groups helps to reduce negative ability-related social comparisons by students experiencing learning difficulties. Teachers need to select tasks carefully because unsuccessful groups do not raise self-efficacy.

In selecting students for working on group projects, Gina Brown might assess students’ abilities for skills needed (e.g., writing, analyzing, interpreting, researching, organizing) and then form groups by assigning students with different strengths to each group.

Lirgg and Feltz ( 1991 ) exposed sixth-grade girls to a skilled or unskilled teacher or peer model demonstrating a ladder-climbing task; girls in a control group observed no model. Girls then judged self-efficacy for climbing successively higher levels on the ladder and performed the task over trials. Control students demonstrated poorer performance than those exposed to models; among the latter, children who viewed a skilled model (adult or peer) performed better than those who observed an unskilled model. Skilled-model girls judged self-efficacy higher.

Bandura and Cervone ( 1983 ) showed how feedback was important during motor skill acquisition. College students operated an ergometer by alternatively pushing and pulling arm levers that resisted their efforts. Some participants pursued a goal of increasing performance by 40% over the baseline, others were told they had increased their performance by 24%, those in a third condition received goals and feedback, and control-group participants received neither goals nor feedback. Goals combined with feedback improved performance most and instilled self-efficacy for goal attainment, which predicted subsequent effort.

In follow-up research (Bandura & Cervone,  1986 ), participants received a goal of 50% improvement over baseline. Following their performance, they received false feedback indicating they achieved an increase of 24%, 36%, 46%, or 54%. Self-efficacy was lowest for the 24% group and highest for the 54% condition. After students set goals for the next session and performed the task again, effort expenditure related positively to goals and self-efficacy across all conditions.

Poag-DuCharme and Brawley ( 1993 ) found that self-efficacy predicted individuals’ involvement in community-based exercise programs. Self-efficacy was assessed for performing in-class activities and for overcoming barriers to exercising and scheduling problems. Self-efficacy related positively to the initiation and maintenance of regular exercise. In similar fashion, Motl and colleagues (Motl, Dishman, Saunders, Dowda, & Pate,  2007 ; Motl et al.,  2005 ) have shown that self-efficacy for overcoming barriers to exercise predicted physical exercise by adolescent girls. These results suggest that promoting exercise requires attention to developing individuals’ self-efficacy for coping with potential problems in scheduling and engagement.

Instructional Self-Efficacy

Self-efficacy is relevant to teachers as well as students (Pajares,  1996 ; Tschannen-Moran, Woolfolk Hoy, & Hoy,  1998 ; Woolfolk Hoy, Hoy, & Davis,  2009 ).  Instructional self-efficacy  refers to personal beliefs about one’s capabilities to help students learn. Instructional self-efficacy should influence teachers’ activities, effort, and persistence with students (Ashton,  1985 ; Ashton & Webb,  1986 ). Teachers with low self-efficacy may avoid planning activities they believe exceed their capabilities, not persist with students having difficulties, expend little effort to find materials, and not reteach content in ways students might understand better. Teachers with higher self-efficacy are more apt to develop challenging activities, help students succeed, and persevere with students who have problems learning. These motivational effects on teachers enhance student achievement. Teachers with higher self-efficacy also show stronger commitment to their work (Chan, Lau, Nie, Lim, & Hogan,  2008 ). Ashton and Webb ( 1986 ) found that teachers with higher self-efficacy were likely to have a positive classroom environment, support students’ ideas, and address students’ needs. Teacher self-efficacy was a significant predictor of student achievement. Woolfolk and Hoy ( 1990 ) obtained comparable results with pre-service teachers.

Not surprisingly, more experienced teachers tend to hold a higher sense of self-efficacy (Wolters & Daugherty,  2007 ). These researchers also found that teacher self-efficacy related positively to their efforts to create classroom mastery goal structures emphasizing learning progress and overcoming challenges (see  Chapter 9 ). Teacher self-efficacy has been shown to positively predict job satisfaction (Collie, Shapka, & Perry,  2012 ). Feltz, Chase, Moritz, and Sullivan ( 1999 ) showed that the same predictions for teacher self-efficacy also applied to coaches.

Researchers have investigated the dimensions of instructional efficacy that relate best to student learning (Gibson & Dembo,  1984 ; Woolfolk & Hoy,  1990 ). Ashton and Webb ( 1986 ) distinguished teaching efficacy, or outcome expectations about the consequences of teaching in general, from personal efficacy, defined as self-efficacy to perform particular behaviors to bring about given outcomes. As noted earlier, self-efficacy and outcome expectations often are related but need not be. A teacher might have a high sense of personal efficacy but lower teaching efficacy if he or she believes that most student learning is due to home and environmental factors outside of the teacher’s control. Other research suggests that instructional self-efficacy reflects an internal–external distinction: internal factors represent perceptions of personal influence and power and external factors relate to perceptions of influence and power of elements that lie outside the classroom (Guskey & Passaro,  1994 ).

Goddard, Hoy, and Woolfolk Hoy ( 2000 ) discussed  collective teacher efficacy , or perceptions of a group of teachers in a school that their efforts as a whole will positively affect students. Collective teacher efficacy requires support from administrators who facilitate improvement by creating an environment free of roadblocks, and it seems critical for effective school reform.

The role of collective teacher efficacy may depend on the level of organizational coupling (Henson,  2002 ). Collective teacher efficacy may not predict outcomes in loosely knit schools; individual self-efficacy may be a better predictor. This situation may occur in some secondary schools where coupling, if present, resides at the departmental level rather than at the whole-school level. Conversely, elementary schools typically are more closely coupled, and the collective efficacy of the school’s teachers may predict student outcomes.

Goddard et al. ( 2000 ) discussed the process whereby collective teacher efficacy can affect student learning. The same four sources of self-efficacy information affect collective efficacy: performance attainments, vicarious experiences, social persuasion, and physiological indicators. Collective efficacy is apt to be strengthened when teachers successfully work together to implement changes, learn from one another and from other successful schools, receive encouragement for change from administrators and professional development sources, and work together to cope with difficulties and alleviate stress (Goddard, Hoy, & Woolfolk Hoy,  2004 ). As collective teacher efficacy is strengthened, teachers continue to improve opportunities for students.

Caprara, Barbaranelli, Borgogni, and Steca ( 2003 ) found that teachers’ collective efficacy beliefs bore a positive relation to their job satisfaction. Further, collective efficacy depends on teachers believing that other constituencies (e.g., principals, staff, parents, students) are working diligently to fulfill their obligations. Consistent with Bandura’s ( 1997 ) position, even high self-efficacy will not lead to beneficial changes unless the environment is responsive to change. Retaining teachers in the profession—a critical priority given the teacher shortage in many areas—will be aided by creating an environment in which teachers’ sense of agency is fostered and their efforts lead to positive changes.

An important challenge for pre- and in-service teacher education programs is to develop methods for increasing teachers’ self-efficacy by incorporating efficacy-building sources (actual performances, vicarious experiences, persuasion, physiological indexes). Internships where students work with teacher mentors provide actual performance success plus expert modeling. Teacher models not only teach observers skills but also build their self-efficacy for succeeding in the classroom ( Application 4.6 ).

Health and Therapeutic Activities

APPLICATION 4.6 Instructional Self-Efficacy

Self-efficacy among teachers is developed in the same ways as among students. An effective means of building self-efficacy is to observe someone model specific teaching behaviors. A new elementary teacher might observe his or her mentor teacher implement the use of learning centers before the new teacher introduces the same activity. By observing the mentor, the new teacher acquires skill and self-efficacy for being able to implement the centers.

Self-efficacy in beginning teachers also may be aided by observing second- and third-year teachers successfully perform actions; new teachers may perceive greater similarity between themselves and other relatively new teachers than between themselves and those teachers with more experience.

Practicing helps to develop skills and also builds self-efficacy. Music teachers will increase their self-efficacy for teaching pieces to the class by practicing those pieces until they know them well and feel confident about working with students. Teachers should learn to use a new computer application well before they introduce it to their classes so they will feel self-efficacious about teaching their students to use it.

Becoming more knowledgeable about a particular subject increases self-efficacy for discussing the subject more accurately and completely. College instructors should review the work of significant researchers for each major topic area included in the course discussions. Such reviews help instructors provide students with information beyond what is in the text and builds instructors’ self-efficacy for effectively teaching the content.

Researchers have shown that self-efficacy predicts health and therapeutic behaviors (Bandura,  1997 ; Maddux,  1993 ; Maddux, Brawley, & Boykin,  1995 ). The Health Belief Model has been commonly applied to explain health behavior change (Rosenstock,  1974 ). This model assigns a prominent role to individuals’ perceptions of four factors that affect health behaviors: susceptibility (personal assessment of risk for a given health threat), severity of the health threat, benefits of the behavior recommended to reduce the threat, and barriers to action (personal belief of possible undesirable consequences that could result from performing the recommended preventive behavior). The barriers factor has the strongest research support; it relates closely to self-efficacy (i.e., self-efficacy for overcoming barriers; Maddux,  1993 ). A newer health behavior goal model (Maes & Gebhardt,  2000 ) includes perceived competence (analogous to self-efficacy) as a key influence.

The importance of self-efficacy as a predictor of health behaviors is evident in many studies (DiClemente,  1986 ; Strecher, DeVellis, Becker, & Rosenstock,  1986 ). Self-efficacy correlates positively with controlled smoking (Godding & Glasgow,  1985 ), positively with longest period of smoking cessation (DiClemente, Prochaska, & Gilbertini,  1985 ), negatively with temptation to smoke (DiClemente et al.,  1985 ), and positively with weight loss (Bernier & Avard,  1986 ). Love ( 1983 ) found that self-efficacy to resist bulimic behaviors correlated negatively with binging and purging. Bandura ( 1994 ) discussed the role of self-efficacy in the control of HIV infection.

In DiClemente’s ( 1981 ) study, individuals who had recently quit smoking judged their self-efficacy to avoid smoking in situations of varying stress levels; they were surveyed months later to determine maintenance. Maintainers judged self-efficacy higher than those who relapsed. Self-efficacy was a better predictor of future smoking than was smoking history or demographic variables. People tended to relapse in situations where they had judged their self-efficacy low for avoiding smoking.

Researchers have investigated how well self-efficacy predicts therapeutic behavioral changes (Bandura,  1991 ). In one study (Bandura, Adams, & Beyer,  1977 ), adult snake phobics received a  participant modeling treatment in which a therapist initially modeled a series of progressively more threatening encounters with a snake. After phobics jointly performed the various activities with the therapist, they were allowed to perform on their own. Compared with phobics who only observed the therapist model the activities and with those who received no training, participant-modeling clients demonstrated the greatest increases in self-efficacy and approach behaviors toward the snake. Regardless of treatment, self-efficacy for performing tasks was highly related to clients’ actual behaviors. In a related study, Bandura and Adams ( 1977 ) found participant modeling superior to systematic desensitization ( Chapter 3 ). These results support Bandura’s ( 1982b ,  1997 ) contention that performance-based treatments combining modeling with practice produce higher self-efficacy and greater behavioral change.

The development and maintenance of healthy lifestyles often have been explained in terms of prescriptive medical management, but increasingly researchers and practitioners are emphasizing collaborative self-management (Bandura,  2005 ). The latter includes many of the social cognitive processes described in this chapter: self-monitoring of health-related behaviors, goals and self-efficacy for attaining them, self-evaluation of progress, and self-motivating incentives and social supports for healthy lifestyles (Maes & Karoly,  2005 ).

This view of health and wellness reflects Bandura’s ( 2005 ) agentic perspective on human functioning described at the start of this chapter. Successful lifestyle change that is maintained over time requires that people feel self-efficacious for managing their own activities and controlling events that affect their lives. Self-efficacy affects actions through cognitive, motivational, affective, and self-regulatory processes. Self-efficacy affects whether people think in positive or negative ways, how they motivate themselves and persist during difficulties, how they handle their emotions and especially during periods of stress, how resilient they are to setbacks, and what choices they make at critical times (Benight & Bandura,  2004 ).

In summary, research evidence shows that self-efficacy predicts diverse outcomes such as smoking cessation, pain tolerance, athletic performance, assertiveness, coping with feared events, recovery from heart attack, and sales performance (Bandura,  1986 ,  1997 ). Self-efficacy is a key variable influencing career choices (Lent, Brown, & Hackett,  2000 ), and children’s self-efficacy affects the types of occupations in which they believe they can succeed (Bandura, Barbaranelli, Caprara, & Pastorelli,  2001 ). Self-efficacy researchers have employed diverse settings, participants, measures, treatments, tasks, and time spans to explore the generality of self-efficacy.

INSTRUCTIONAL APPLICATIONS

Many ideas in social cognitive theory lend themselves well to instruction and student learning. Instructional applications involving models and self-efficacy, worked examples, and tutoring and mentoring reflect social cognitive principles.

Models and Self-Efficacy

Teacher models facilitate learning and provide self-efficacy information. Students who observe teachers explain and demonstrate concepts and skills are apt to learn and believe that they are capable of further learning. Teachers also provide persuasive self-efficacy information to students. Teachers who introduce lessons by stating that all students can learn and that by working diligently they will master the new skills instill in students self-efficacy for learning, which is substantiated when students successfully work on the task. Teachers should ensure that their instructions to students (e.g., “keep your desk tidy”) are consistent with their own actions (teacher’s desk is tidy).

In similar fashion, peer models can promote student motivation and learning. Relative to teachers, peers may be more focused on “how to do it,” which improves learning in observers. Further, observing a similar peer succeed instills a vicarious sense of self-efficacy for learning in observers, which is validated when they perform well (Schunk,  1987 ). When using peers, it helps to choose models such that each student can relate to at least one. This may mean using multiple peer models, where the peers represent varying levels of skill.

To determine their instructional methods, teachers should gauge their effects on students’ self-efficacy as well as on their learning. It may be that a method that produces learning does not enhance self-efficacy. For example, providing students with extensive assistance is apt to aid their learning, but it will not do much for students’ self-efficacy for learning or performing well on their own. As Bandura ( 1986 ,  1997 ) recommended, periods of self-directed mastery, where students practice skills independently, are needed.

Competent models teach skills, but similar models are best for self-efficacy. Having the best mathematics student in the class demonstrate operations may teach skills to the observers, but many of the latter students may not feel efficacious because they may believe that they never will be as good as the model. Often top students serve as tutors for less-capable students, which may improve learning but should be accompanied by periods of independent practice to build self-efficacy (see section, Tutoring and Mentoring, below).

Pre-service teachers’ self-efficacy can be developed through teacher preparation that includes internships with master teachers where the pre-service teachers can observe and practice teaching skills. For in-service teachers, continuing professional development can help them learn new strategies to use in challenging situations, such as how to foster learning in students with varying abilities, how to work with students with limited English proficiency, and how to involve parents in their children’s learning. By removing impairments to teaching (e.g., excess paperwork), administrators allow teachers to focus on curricular improvement and student learning (see  Application 4.6 ).

Worked Examples

Worked examples  are graphic portrayals of problem solutions (Atkinson, Derry, Renkl, & Wortham,  2000 ). Worked examples present step-by-step problem solutions, often with accompanying diagrams or sound (narration). A worked example provides a model—with accompanying explanation—that illustrates how a proficient problem solver would proceed. Learners study worked examples before they attempt to solve problems themselves. Worked examples are often used in instruction in mathematics and science, although their use need not be confined to these disciplines.

The theoretical underpinnings for worked examples derive from information processing theory and are discussed in  Chapter 7 . But worked examples also reflect many principles of social cognitive theory (van Gog & Rummel,  2010 ). Worked examples incorporate cognitive models and demonstration plus explanation. As with other complex forms of observational learning, students do not learn how to solve a particular problem but rather general skills and strategies that they can use to solve a wider class of problems. Worked examples also have motivational benefits. They may help to raise self-efficacy in learners who, after reviewing worked examples, believe they understand the model and can apply the skills and strategies (Schunk,  1995 ).

Certain principles should be kept in mind when using worked examples. It is better to use more than one mode of presentation than a single mode. Thus, a worked example might include textual (words, numbers), graphical (arrows, charts), and aural (sounds) information. But too much complexity can overload learners’ attention and working memory. Research also shows that two examples are better than a single one, two varied examples are better than two examples of the same type, and intermixing practice with worked examples produces better learning than if all examples are presented first followed by practice (Atkinson et al.,  2000 ). Thus, an algebra teacher teaching a lesson on solving equations in one unknown might present two worked examples of the form 4x + 2 = 10, after which students solve problems. Then the teacher might present two worked examples of the form x ÷ 2 + 1 = 5, after which students solve problems of this type. The worked examples could be accompanied by graphics and sound, as in interactive computer-based learning environments.

Tutoring and Mentoring

Tutoring  refers to a situation in which one or more persons serve as the instructional agents for another, usually in a specific subject or for a particular purpose (Stenhoff & Lignugaris/Kraft,  2007 ). When peers are the instructional agents, tutoring is a form of peer-assisted learning (Rohrbeck, Ginsburg-Block, Fantuzzo, & Miller,  2003 ;  Chapter 8 ).

Tutors serve as instructional models for tutees by explaining and demonstrating skills, operations, and strategies that tutees are to learn. Both adults and children can be effective tutors for children. As noted earlier, however, there may be some motivational benefits that result from peer tutors. Effective peer tutors are those whom tutees perceive as similar to themselves except that tutors are farther along in their skill acquisition. The perception of similarity may lead tutees to believe that if the tutors could learn, they can as well, which can raise tutees’ self-efficacy and motivation.

Researchers also have examined the effects of tutoring on tutors. Similar to the results of instructional self-efficacy, tutors with higher self-efficacy for tutoring are more apt to exert effort, tackle difficult material, and persist longer with tutees than are tutors with lower self-efficacy (Roscoe & Chi,  2007 ). There also is some evidence that tutoring can enhance tutors’ motivation and self-efficacy (Roscoe & Chi,  2007 ).

Mentoring  refers to interactions between more-experienced mentors and less-experienced mentees (or protégés), where mentors provide career (instrumental) and psychosocial (relational) knowledge, advice, and support (Eby, Rhodes, & Allen,  2007 ; Fletcher & Mullen,  2012 ). The overall goal of mentoring is to help people function effectively in their professional and personal lives. Ideally mentoring incorporates mutual learning and engagement between the mentor and protégé. Thus, mentoring is a fuller and deeper educational experience than tutoring, which is more apprenticeship oriented. While tutoring emphasizes content instruction within a short time period, mentoring typically involves modeled counsel and guidance over a longer time (Johnson,  2007 ).

Mentoring is common at various levels of education, such as in learning communities, inquiry and writing groups, university–school partnerships, staff development, higher education, and peer coaching (Mullen,  2005 ). In higher education, mentoring often occurs between senior and junior faculty members or between professors and students. In this context, mentoring ideally becomes a developmental relationship where more-experienced professors share their expertise with and invest time in less-experienced professors or students to nurture their achievement and self-efficacy (Johnson,  2007 ; Mullen,  2011 ).

Mentoring reflects many social cognitive principles and can have instructional and motivational benefits (Schunk & Mullen,  2013 ). Protégés learn skills and strategies that help them succeed in their environments from mentors who model, explain, and demonstrate these skills and strategies. Protégés who perceive themselves as similar in important ways to mentors may develop higher self-efficacy for being successful through their interactions with mentors. Similar to motivation and self-regulated learning, mentoring emphasizes goal-directed activity over time (Schunk & Mullen,  2013 ). Mentoring of doctoral students has been shown to improve their self-regulation, self-efficacy, motivation, and achievement (Mullen,  2011 ). Mentors also can learn and refine their skills through their interactions with their protégés, which could raise their self-efficacy for continuing to succeed. Consistent with social cognitive theory, the mentoring relationship can result in reciprocal benefits for both parties (Schunk & Mullen,  2013 ).

SUMMARY

Social cognitive learning theory contends that people learn from their social environments. In Bandura’s theory, human functioning is viewed as a series of reciprocal interactions among personal, behavioral, and environmental factors. Learning is an information processing activity in which knowledge is cognitively represented as symbolic representations serving as guides for action. Learning occurs enactively through actual performances and vicariously by observing models, listening to instructions, and engaging with print or electronic content. The consequences of behavior are especially important. Behaviors that result in successful consequences are retained; those that lead to failures are discarded. Social cognitive theory presents an agentic perspective of human behavior in that persons can learn to set goals and self-regulate their cognitions, emotions, behaviors, and environments in ways to facilitate attainment of those goals.

Historical work perspectives on imitation do not fully capture the range and influence of modeling. Bandura and colleagues have shown how modeling greatly expands the range and rate of learning. Various modeling effects are distinguished: inhibition and disinhibition, response facilitation, and observational learning. Observational learning through modeling expands the learning rate, as well as the amount of knowledge acquired. Subprocesses of observational learning are attention, retention, production, and motivation.

According to social cognitive theory, observing a model does not guarantee learning or later ability to perform the behaviors. Rather, models provide information about probable consequences of actions and motivate observers to act accordingly. Factors influencing learning and performance are developmental status of learners, prestige and competence of models, and vicarious consequences to models.

Among the important motivational influences on learning are goals, outcome expectations, values, and self-efficacy. Goals enhance learning through their effects on perceived progress, self-efficacy, and self-evaluations. As people work on a task, they compare their progress with their goal. The perception of progress raises self-efficacy and sustains motivation. Goal properties of specificity, proximity, and difficulty enhance self-perceptions and motivation, as do self-set goals and goals that people make a commitment to attain.

Outcome expectations (perceived consequences of behavior) affect learning and motivation because people strive to attain desired outcomes and shun undesirable ones. People also act in concert with their values, working towards outcomes that they find self-satisfying.

Self-efficacy, or one’s perceived capabilities of learning or performing behaviors at designated levels, is not the same as knowing what to do. People gauge their self-efficacy based on their performance attainments, vicarious consequences to models, forms of persuasion, and physiological indicators. Actual performances provide the most reliable information to use in assessing self-efficacy. Self-efficacy can affect choice of activities, effort, persistence, and achievement. Instructional self-efficacy and collective self-efficacy, which have been studied with teachers, bear a positive relation to student learning and achievement.

Researchers have found support for Bandura’s theory in a variety of contexts involving cognitive, social, motor, health, instructional, and self-regulatory skills. Self-efficacy has been shown to predict behavioral change with different types of participants (e.g., adults, children) in various settings. This research also has shown that learning of complex skills occurs through a combination of enactive and vicarious learning. Observers acquire an approximation of the skill by observing models. Subsequent practice of the skill allows teachers to provide corrective feedback to learners. With additional practice, learners refine and internalize self-regulatory skills and strategies. Important instructional applications of social cognitive theory involve models and self-efficacy, worked examples, and tutoring and mentoring.

A summary of learning issues appears in  Table 4.6 .

Table 4.6 Summary of learning issues.

How Does Learning Occur? Learning occurs enactively (by doing) and vicariously (by observing, reading, and listening). Much school learning requires a combination of vicarious and enactive experiences. Observational learning greatly expands the scope of human learning possible. Observational learning consists of four processes: attention, retention, production, and motivation. A major contribution of social cognitive theory is its emphasis on learning from the social environment. How Does Memory Function? Social cognitive researchers have not investigated in depth the role of human memory. Social cognitive theory predicts that memory includes information stored as images or symbols. What Is the Role of Motivation? Key motivational processes are goals, values, and expectations. People set goals for learning and assess progress against goals. Values reflect what persons find self-satisfying and believe are important. Expectations are of two types. Outcome expectations refer to the expected outcomes of actions. Efficacy expectations, or self-efficacy, refer to one’s perceived capabilities for learning or performing tasks at designated levels. The belief that one is making goal progress substantiates self-efficacy and motivates one to continue learning. How Does Transfer Occur? Transfer is a cognitive phenomenon. It depends on people believing that certain actions in new or different situations are socially acceptable and will be met with favorable outcomes. Learners’ self-efficacy also can facilitate transfer. How Does Self-Regulated Learning Operate? In the classical view, self-regulation consists of three processes: self-observation, self-judgment, and self-reaction. This view has been broadened to include activities before and after task engagement. Social cognitive theory stresses goals, self-efficacy, attributions, learning strategies, and self-evaluations. These processes reciprocally interact with one another such that goal attainment can lead to the adoption of new goals. What Are the Implications for Instruction? The use of modeling is highly recommended. Effective instruction begins with social influences, such as models, and gradually shifts to self-influences as learners internalize skills and strategies. It is important to determine how instruction affects not only learning but also learners’ self-efficacy. Students should be encouraged to set goals and assess goal progress. Teachers’ self-efficacy affects instruction because efficacious teachers help promote student learning better. Social cognitive principles also are reflected in worked examples, tutoring, and mentoring.

FURTHER READING

Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice Hall.

Bandura, A. (1997). Self-efficacy: The exercise of control. New York, NY: Freeman.

Goddard, R. D., Hoy, W. K., & Woolfolk Hoy, A. (2004). Collective efficacy beliefs: Theoretical developments, empirical evidence, and future directions. Educational Researcher, 33(3), 3–13.

Locke, E. A., & Latham, G. P. (2002). Building a practically useful theory of goal setting and task motivation: A 35-year odyssey. American Psychologist, 57, 705–717.

Schunk, D. H. (2012). Social cognitive theory. In K. R. Harris, S. Graham, & T. Urdan (Eds.), APA educational psychology handbook. Vol 1: Theories, constructs, and critical issues (pp. 101–123). Washington, DC: American Psychological Association.

Schunk, D. H., & Pajares, F. (2009). Self-efficacy theory. In K. R. Wentzel & A. Wigfield (Eds.), Handbook of motivation at school (pp. 35–53). New York, NY: Routledge.

Zimmerman, B. J., & Schunk, D. H. (2003). Albert Bandura: The scholar and his contributions to educational psychology. In B. J. Zimmerman & D. H. Schunk (Eds.), Educational psychology: A century of contributions (pp. 431–457). Mahwah, NJ: Erlbaum.

Chapter 7 Cognitive Learning Processes

Meg LaMann, the principal of Franklin U. Nikowsky Middle School, was holding a faculty meeting. The school’s teachers had participated recently in a professional development session on helping students learn problem solving and critical thinking skills. Meg asked the teachers for feedback on the session.

“Tiny” Lawrance, one of the more outspoken teachers in the school, spoke first. “Well, Meg, I thought the presenters had lots of good things to say and suggestions for developing skills in the students. But, you know what the problem is. We don’t have time to do any of this. We’re too crunched with covering what we need to so that the kids are ready for the state tests. And besides, those tests cover mostly low-level factual information, not what you need problem solving for. So realistically I don’t see how I’ll use much of what I learned yesterday.”

Piper Rowland spoke up next. “That’s right, Meg. I thought it was wonderful information. And surely our kids would benefit from learning some of these strategies. But if we neglect the basic skills to teach this stuff and our test results fall, we’ll hear about it. I don’t know what to do.”

Meg replied, “I hear you and have the same concern. But I don’t think we need to work on problem solving and critical thinking in everything we teach. There are facts and basic skills to be learned. But sometimes we don’t think enough about how we might incorporate problem solving into our instruction. I think we all can do that.”

Tiny said, “I agree, Meg. What about some time set aside periodically to work on problem-solving skills?”

“You heard what the presenters said,” replied Meg. “Problem solving and critical thinking are best taught in the context of regular learning. That way the kids see how they can apply these skills as they’re learning math, English, science, social studies, and so on. The stand-alone thinking skills programs are less effective and the kids usually don’t apply any of those skills outside of the training setting.”

“I’m willing to work on this more in social studies,” said Tiny. “And I will in math,” replied Piper. “I just hope the test scores don’t fall.”

“Don’t worry about the test scores,” said Meg. “I’ll address that if it becomes an issue.”

The teachers made a concerted effort to incorporate suggestions they learned from the session into their teaching for the rest of the school year. The end-of-grade test scores for the school actually rose by a small amount.

At the start of the next academic year the school held a “walk the schedule” night for parents and students. Several parents told Meg how much they appreciated the teachers working more on problem solving. One parent remarked, “Those strategies are great, not just for school but for other things. I’m working with my son now, having him set goals for what he needs to do, check his progress, and so on.” Another parent told Meg, “My daughter loves the new emphasis on problem solving. She says that school now isn’t so boring and is more like its initials—FUN!”

Previous chapters covered cognitive theories of learning: social cognitive ( Chapter 4 ) and information processing ( Chapters 5  and  6 ). This chapter extends this perspective to the operation of key cognitive processes during learning. Following a discussion of skill acquisition, metacognition is covered, which is central to learning. Subsequent sections address concept learning, problem solving, critical thinking and creativity, cognition and technology, and instructional applications.

There is debate among professionals on the extent that the cognitive processes discussed in this chapter are involved in most, if not all, learning. Problem solving, for example, is thought by some to be the central process in learning (Anderson,  1993 ), whereas others limit its application to settings where specific conditions prevail (Chi & Glaser,  1985 ). Teachers generally agree on the importance of concept learning, problem solving, critical thinking, creativity, and metacognition, and educators recommend that these topics be incorporated into instruction (Pressley & McCormick,  1995 ). The opening vignette describes a schoolwide effort to integrate problem solving and critical thinking into the curriculum. The processes discussed in this chapter are integral components of complex types of learning that occur in school subjects such as reading, writing, mathematics, and science.

When you finish studying this chapter you should be able to do the following:

· ■ Distinguish between general and specific skills, and discuss how they work together in the acquisition of competence.

· ■ Describe the novice-to-expert research methodology.

· ■ Explain why metacognition is important for learning, and discuss variables affecting it.

· ■ Distinguish properties of concepts, and explain models of concept learning.

· ■ Explain the differences between various methods for solving problems.

· ■ Describe problem solving from an information processing perspective.

· ■ Explain the differences between critical thinking, reasoning, and creativity, and describe ways to help develop these cognitive processes in students.

· ■ Discuss key learning features of computer-based environments, online social media, and distance learning, and how these technologies may affect learning.

· ■ Describe some instructional applications involving worked examples, problem solving, and mathematics.

SKILL ACQUISITION

Developing competence in any domain represents a process of skill acquisition. We begin by examining issues relevant to the acquisition of general and specific skills.

General and Specific Skills

Skills may be differentiated according to degree of specificity.  General skills  apply to a wide variety of disciplines;  specific skills  are useful only in certain domains. For example, problem solving and critical thinking are general skills because they are useful in acquiring a range of cognitive, motor, and social skills, whereas factoring polynomials and solving square-root problems involve specific skills because they have limited mathematical applications.

Acquisition of general skills facilitates learning in many ways. Bruner ( 1985 ) noted that tasks such as “learning how to play chess, learning how to play the flute, learning mathematics, and learning to read the sprung rhymes in the verse of Gerard Manley Hopkins” (pp. 5–6) are similar in that they involve attention, memory, and persistence.

But each type of skill has unique features. Bruner ( 1985 ) contended that views of learning are not unambiguously right or wrong; rather, they can be evaluated only in light of such conditions as the nature of the task to be learned, the type of learning to be accomplished, and the characteristics that learners bring to the situation. The many differences between tasks, such as learning to balance equations in chemistry and learning to balance on a beam in gymnastics, require different processes to explain learning.

Domain specificity  is defined in various ways. Ceci ( 1989 ) used the term to refer to discrete declarative knowledge structures ( Chapter 5 ). Other researchers include procedural knowledge and view specificity as pertaining to the usefulness of knowledge (Perkins & Salomon,  1989 ). The issue really is not one of proving or disproving one position because we know that both general strategies and domain-specific knowledge are involved in learning (Nandagopal & Ericsson,  2012 ; Voss, Wiley, & Carretero,  1995 ). Rather, the issue is one of specifying the extent to which any type of learning involves general and specific skills, what those skills are, and what course their acquisition follows.

Thinking of skill specificity ranging along a continuum is preferable, as Perkins & Salomon ( 1989 ) explained:

·  General knowledge includes widely applicable strategies for problem solving, inventive thinking, decision making, learning, and good mental management, sometimes called autocontrol, autoregulation, or metacognition. In chess, for example, very specific knowledge (often called local knowledge) includes the rules of the game as well as lore about how to handle innumerable specific situations, such as different openings and ways of achieving checkmate. Of intermediate generality are strategic concepts, like control of the center, that are somewhat specific to chess but that also invite far-reaching application by analogy. (p. 17)

We then can ask: What counts most for ensuring success in learning? Some local knowledge is needed—one cannot become skilled at fractions without learning the rules governing fraction operations (e.g., adding, subtracting). As Perkins and Salomon ( 1989 ) noted, however, the more important questions are: Where are the bottlenecks in developing mastery? Can one become an expert with only domain-specific knowledge? If not, at what point do general competencies become important?

Ohlsson ( 1993 ) advanced a model of skill acquisition through practice that comprises three subfunctions: generate task-relevant behaviors, identify errors, and correct errors. This model includes both general and task-specific processes. As learners practice, they monitor their progress by comparing their current state to their prior knowledge. This is a general strategy, but as learning occurs, it becomes increasingly adapted to specific task conditions. Errors often are caused by applying general procedures inappropriately (Ohlsson,  1996 ), but prior domain-specific knowledge helps learners detect errors and identify the conditions that caused them. With practice and learning, therefore, general methods become more specialized.

Problem solving is useful for learning skills in many content areas, but task conditions often require specific skills for the development of expert performance. In many cases a merging of the two types of skills is needed. Research shows that expert problem solvers often use general strategies when they encounter unfamiliar problems and that asking general metacognitive questions (e.g., “What am I doing now?” “Is it getting me anywhere?”) facilitates problem solving (Perkins & Salomon,  1989 ). Despite these positive results, general principles often do not transfer (Pressley et al.,  1990 ; Schunk & Rice,  1993 ). Transfer requires combining general strategies with factors such as instruction on self-monitoring and practice in specific contexts. Motivation also is important (Nandagopal & Ericsson,  2012 ). The goal in the opening vignette is that once students learn general strategies, they will be able to adapt them to specific settings and will be motivated to do so.

In short, expert performance requires much domain knowledge (Lajoie,  2003 ; Nandagopal & Ericsson,  2012 ). It requires a rich knowledge base that includes the facts, concepts, and principles of the domain, coupled with general learning strategies that can be applied to different domains and that may have to be tailored to each domain. One would not expect strategies such as seeking help and monitoring goal progress to operate in the same fashion in disparate domains (e.g., calculus and pole vaulting). At the same time, Perkins and Salomon ( 1989 ) pointed out that general strategies are useful for coping with atypical problems in different domains regardless of one’s overall level of competence in the domain. These findings imply that students need to be well grounded in basic content-area (domain) knowledge (Ohlsson,  1993 ), as well as in general problem-solving and self-regulatory strategies ( Chapter 10 ).  Application 7.1  provides suggestions for integrating the teaching of general and specific skills.

Novice-to-Expert Research Methodology

With the growth of cognitive ( Chapters 4 – 6 ) and constructivist ( Chapter 8 ) views of learning, researchers moved away from viewing learning as changes in responses due to differential reinforcement ( Chapter 3 ) and became interested in exploring students’ beliefs and thought processes during learning. The focus of learning research has shifted accordingly.

APPLICATION 7.1 Integrating the Teaching of General and Specific Skills

Teachers can help students learn general skills and strategies to increase students’ success in various domains, while also stressing the skills that are needed for learning within a specific domain.

Mr. Thomson might work with his fifth-grade students on the general strategy of goal setting to complete assignments. In reading, he might help students determine how to finish reading two chapters in a book by the end of the week. The students might establish a goal to read a certain number of pages or a subsection each day of the week. Because the goal comprises more than just reading pages, he also must teach specific comprehension skills, such as locating main ideas and reading for details. He can have students use goal setting in mathematics; they might decide how many problems or activities to do each day to complete a particular unit by the end of the week. Specific skills that come into play in this context are determining what the problem is asking for, representing the problem, and knowing how to perform the computations.

In physical education, students may use goal setting to master skills, such as working toward running a mile in 6 minutes. The students begin by running the mile and recording their times, after which they set a goal to decrease the running time by a certain amount every week. Motor and endurance skills must be developed to successfully meet the goal. Such skills are most likely to be specific to the context of running a mile in a good time.

· To investigate academic learning, some researchers have used a  novice-to-expert methodology  with the following steps:

· ■ Identify the skill to be learned.

· ■ Find an  expert  (i.e., one who performs the skill well) and a  novice  (one who knows something about the task but performs it poorly).

· ■ Determine how the novice can be moved to the expert level as efficiently as possible.

This methodology is intuitively plausible. The basic idea is that if you want to understand how to become more skillful in an area, closely study someone who performs that skill well (Bruner,  1985 ). In so doing you can learn what knowledge is needed, what procedures and strategies are useful, how to handle difficult situations, and how to correct mistakes. The model has many real-world counterparts and is reflected in mentoring, apprenticeships, and on-the-job training (Fletcher & Mullen,  2012 ).

Much of the knowledge on how more and less competent persons differ in a domain comes from research based in part on assumptions of this methodology (VanLehn,  1996 ). Compared with novices, expert performers have more extensive domain knowledge, have better understanding of what they do not know, spend more time initially analyzing problems, and solve them quicker and more accurately (Lajoie,  2003 ). Research also has identified differences in the stages of skill acquisition. Conducting such research is labor intensive and time consuming because it requires studying learners over time, but it yields rich results.

Bear in mind, however, that this model is descriptive rather than explanatory: It describes what learners do but does not explain why they do it. The model also tacitly assumes that a fixed constellation of skills exists that constitutes expertise in a given domain, but this is not always the case. With respect to teaching, Sternberg and Horvath ( 1995 ) argued that no one standard exists; rather, expert teachers resemble one another in prototypical fashion. This makes sense given our experiences with master teachers who typically differ in several ways.

Finally, the model does not automatically suggest teaching methods. As such, it may have limited usefulness for classroom teaching and learning. Explanations for learning and corresponding teaching suggestions should be firmly grounded in theories and identify important personal and environmental factors, which are emphasized in this and other chapters in this text.

Expert–Novice Differences in Science

A good domain to explore expert–novice differences is in science because researchers have compared novices with experts to identify the components of expertise. Researchers also have investigated students’ construction of scientific knowledge and the implicit theories and reasoning processes that they use during problem solving and learning (Linn & Eylon,  2006 ; Voss et al.,  1995 ; White,  2001 ; C. Zimmerman,  2000 ;  Chapter 8 ).

Expert performers in science differ from novices in quantity and organization of knowledge. Experts possess more domain-specific knowledge and are more likely to organize it in hierarchies, whereas novices often demonstrate little overlap between scientific concepts.

Chi, Feltovich, and Glaser ( 1981 ) had expert and novice problem solvers sort physics textbook problems on any basis they wanted. Novices classified problems based on superficial features (e.g., apparatus); experts classified the problems based on the principle needed to solve the problem. Experts and novices also differed in declarative knowledge memory networks. “Inclined plane,” for example, was related in novices’ memories with descriptive terms such as “mass,” “friction,” and “length.” Experts had these descriptors in their memories, but in addition had stored principles of mechanics (e.g., conservation of energy, Newton’s force laws). The experts’ greater knowledge of principles was organized with descriptors subordinate to principles.

Novices often use principles erroneously to solve problems. McCloskey and Kaiser ( 1984 ) posed the following question to college students:

·  A train is speeding over a bridge that spans a valley. As the train rolls along, a passenger leans out of a window and drops a rock. Where will it land?

About one-third of the students said the rock would fall straight down ( Figure 7.1 ). They believed that an object pushed or thrown acquires a force but that an object being carried by a moving vehicle does not acquire a force, so it drops straight down. The analogy the students made was with a person standing still who drops an object, which falls straight down. The path of descent of the rock from the moving train is, however, parabolic. The idea that objects acquire force is erroneous because objects move in the same direction and at the same speed as their moving carriers. When the rock is dropped, it continues to move forward with the train until the force of gravity pulls it down. Novices generalized their basic knowledge and arrived at an erroneous solution.

Figure 7.1 Possible answers to falling rock problem.

As discussed later in this chapter, another difference between novice and expert performers concerns the use of problem-solving strategies (Larkin, McDermott, Simon, & Simon,  1980 ; White & Tisher,  1986 ). When confronted with scientific problems, novices often use a  means–ends analysis , determining the goal of the problem and deciding which formulas might be useful to reach that goal. They work backward and recall formulas containing quantities in the target formula. If they become uncertain how to proceed, they may abandon the problem or attempt to solve it based on their current knowledge.

Experts quickly recognize the problem format, work forward toward intermediate subgoals, and use that information to reach the ultimate goal. Experience in working scientific problems builds knowledge of problem types. Experts often automatically recognize familiar problem features and carry out necessary productions. Even when they are less certain how to solve a problem, experts begin with some information given in the problem and work forward to the solution. Notice that the last step experts take is often novices’ first step. Klahr and Simon ( 1999 ) contended that the process of scientific discovery is a form of problem solving and that the general heuristic approach is much the same across domains.

METACOGNITION

Information processing theories primarily describe learning rather than explain it. These theories contend that inputs are received into  working memory  (WM), rehearsed, organized and elaborated, linked with relevant information in  long-term memory  (LTM), and stored in LTM. But we might ask why any of these activities happen. Especially during learning—when processing is not automatic—we need an explanation for why the system processes information. For example, what determines how much rehearsal takes place? How is relevant information selected in LTM? How do people know what knowledge is required in different situations?

Metacognition provides answers to some of these questions.  Metacognition  refers to the deliberate conscious control of cognitive activity (Brown,  1980 ; Matlin,  2009 ). Metacognition is, essentially, people’s awareness of their own cognitive processes (Rhodes & Tauber,  2011 ). Before discussing how metacognitive processes help to integrate information processing, the topic of conditional knowledge will be addressed, because this type of knowledge is part of metacognition.

Conditional Knowledge

Declarative and procedural knowledge refer to knowledge of facts and procedures, respectively ( Chapter 5 ).  Conditional knowledge  is knowledge about when and why to employ forms of declarative and procedural knowledge (Paris Lipson, & Wixson,  1983 ). Possessing requisite declarative and procedural knowledge to perform a task does not guarantee students will perform it well. Students reading a social studies text may know what to do (read a chapter), know the meanings of vocabulary words (declarative knowledge), and know how to read for understanding (procedural knowledge). But they might skim the chapter. As a consequence, they perform poorly on a comprehension test.

This type of situation is common. In this example, conditional knowledge includes knowing when skimming is appropriate. One might skim a newspaper or a Web page for the gist, but skimming should not be used to comprehend textual content.

Conditional knowledge helps students select and employ declarative and procedural knowledge to fit goals. To decide to read a chapter carefully and then do it, students should believe that careful reading is appropriate for the task at hand; that is, this strategy has functional value because it will allow them to comprehend the material.

Learners who do not possess conditional knowledge about when and why skimming is valuable will employ it at inappropriate times. If they believe it is valuable for all reading tasks, they may indiscriminately employ it unless otherwise directed. If they believe it has no value, they may never use it unless directed.

Conditional knowledge may be represented in LTM in propositional networks and linked with the declarative and procedural knowledge to which it applies. Conditional knowledge actually is a form of declarative knowledge because it is “knowledge that”—for example, knowledge that skimming is valuable to get the gist of a passage and knowledge that summarizing text is valuable to derive greater understanding. Conditional knowledge also is included in procedures: Skimming is valuable as long as I can get the gist; but if I find that I am not getting the gist, I should abandon skimming and read more carefully. The three types of knowledge are summarized in  Table 7.1 .

Table 7.1 Comparison of types of knowledge.

Type	Knowing	Examples
Declarative	That	Historical dates, number facts, episodes (what happened when), task features (stories have a plot and setting), beliefs (“I am good at math”)
Procedural	How	Mathematical algorithms, reading strategies (e.g., skimming, scanning, summarizing), goal setting (e.g., breaking long-term goals into subgoals)
Conditional	When, Why	Skim a Web page to get the gist in little time; read texts carefully for understanding

Conditional knowledge is an integral part of self-regulated learning (Zimmerman & Schunk,  2011 ;  Chapter 10 ). Self-regulated learning requires that students decide which learning strategy to use prior to engaging in a task (B. Zimmerman,  2000 ). While students are engaged in a task, they assess task progress (e.g., their level of comprehension) using metacognitive processes. When comprehension problems are detected, students alter their strategy based on conditional knowledge of what might prove more effective. It also has been suggested that computer-based learning environments can serve as metacognitive tools to foster students’ self-regulated learning (Azevedo,  2005a ,  2005b ).

Metacognition and Learning

Flavell ( 1985 ) construed metacognition as follows:

·  What is metacognition? It has usually been broadly and rather loosely defined as any knowledge or cognitive activity that takes as its object, or regulates, any aspect of any cognitive enterprise…. It is called metacognition because its core meaning is “cognition about cognition.” Metacognitive skills are believed to play an important role in many types of cognitive activity, including oral communication of information, oral persuasion, oral comprehension, reading comprehension, writing, language acquisition, perception, attention, memory, problem solving, social cognition, and various forms of self-instruction and self-control. (p. 104)

Metacognition comprises related sets of skills (Dimmitt & McCormick,  2012 ). One must understand what skills, strategies, and resources a task requires. Included in this cluster are finding main ideas, rehearsing information, forming associations or images, using memory techniques, organizing material, taking notes or underlining, and using test-taking techniques. One also must know how and when to use these skills and strategies and apply them to ensure the task is completed successfully. These monitoring activities include checking level of understanding, predicting outcomes, evaluating the effectiveness of efforts, planning activities, deciding how to budget time, and revising or switching to other activities to overcome difficulties (Baker & Brown,  1984 ). Collectively, metacognitive activities reflect the planned and strategic application of declarative, procedural, and conditional knowledge to tasks (Schraw & Moshman,  1995 ). Metacognitive skills contribute to the development of critical thinking and problem solving (Dimmitt & McCormick,  2012 ; Kuhn,  1999 ; discussed later in this chapter), and metacognition is a central aspect of self-regulated learning (Azevedo,  2009 ; Efklides,  2006 ; see  Chapter 10 ).

Metacognition is involved during all phases of learning (Efklides,  2006 ). Before learning students may experience feelings of familiarity, difficulty, knowing, interest, and liking, as well as judgments about the best strategy to use and time needed for learning. While engaged in the task, learners’ metacognitive processes may comprise feelings of difficulty, estimated effort and time needed for task completion, and judgments about strategy effectiveness. During pauses or when the learning task is complete students may experience feelings of confidence, satisfaction, and liking of the task, as well as cognitive estimates of solution accuracy.

Metacognitive skills develop slowly (Dimmitt & McCormick,  2012 ). Young children are not fully aware of which cognitive processes various tasks involve. For example, they typically are poor at recognizing that they have been thinking and then recalling what they were thinking about (Flavell, Green, & Flavell,  1995 ). They may not understand that disorganized passages are harder to comprehend than organized ones or that passages containing unfamiliar material are more difficult than those composed of familiar material (Baker & Brown,  1984 ). Dermitzaki ( 2005 ) found that second graders used metacognitive strategies, but that their use bore little relation to children’s actual self-regulatory activities. Monitoring activities are employed more often by older children and adults than by young children; however, older children and adults do not always monitor their comprehension and often are poor judges of how well they have comprehended text (Baker,  1989 ). Relative to regular learners, gifted students tend to show enhanced metacognitive capabilities (Snyder, Nietfeld, & Linnenbrink-Garcia,  2011 ).

At the same time, young children are cognitively capable of monitoring their activities on simple tasks (Kuhn,  1999 ). Learners are more likely to monitor their activities on tasks of intermediate difficulty as opposed to easy tasks (where monitoring may not be necessary) or on very difficult tasks (where one may not know what to do or may quit working).

Metacognitive abilities begin to develop around ages 5 to 7 and continue throughout the time children are in school, although within any age group there is much variability (Flavell,  1985 ; Flavell et al.,  1995 ). Preschool children are capable of learning some strategic behaviors (Kail & Hagen,  1982 ), but as a result of schooling, children develop the awareness that they can control what they learn by the strategies they use (Duell,  1986 ). Flavell and Wellman ( 1977 ) hypothesized that children form generalizations concerning how their actions influence the environment; for example, they learn “what works” for them to promote school achievement. This is especially true with memory strategies, perhaps because much school success depends on memorizing information ( Application 7.2 ).

Variables Influencing Metacognition

Metacognitive awareness is influenced by variables associated with learners, tasks, and strategies (Duell,  1986 ; Flavell & Wellman,  1977 ).

Learner Variables.

Learners’ levels of development influence their metacognition (Alexander, Carr, & Schwanenflugel,  1995 ). Older children understand their own memory abilities and limitations better than younger children do (Flavell, Friedrichs, & Hoyt,  1970 ; Flavell et al.,  1995 ). With development, children can more accurately gauge when they have learned material well enough to recall it.

APPLICATION 7.2 Metacognition

Teachers can help students develop meta-cognitive skills. A teacher working with students on listening comprehension might include situations such as listening to an enjoyable story, a set of explicit directions, and a social studies lecture. For each situation, the teacher could ask students why they would listen in that setting; for example, enjoyment and general theme (stories), specific elements (directions), facts and concepts (social studies). Then the teacher could ask the students to retell the story in their own words, visualize the directions, and take notes. To foster conditional knowledge, the teacher can discuss with students the various listening techniques that seem most appropriate for each situation.

A teacher helping students improve their memory skills might give them a list of items to memorize. The teacher could teach them to reconstruct the list of items given partial cues. The students might be encouraged to explore various memorization techniques such as putting the items into categories, visualizing a picture that contains the items, associating the items with a familiar setting or task, using acronyms that include the first letter of each item, and creating a jingle, poem, or song that incorporates the items. Then the teacher could help the students determine which technique works best for each of them.

Learners’ abilities to monitor how well they have done on a memory task also vary. Older children are more accurate in judging whether they have recalled all items they were to recall. Wellman ( 1977 ) presented children with pictures of objects and asked them to name the objects. If children could not name them, they were asked whether they would recognize the name. Compared with kindergartners, third graders were more accurate at predicting which object names they would be able to recognize.

Task Variables.

Knowing the relative difficulty of different forms of learning and retrieving from memory various types of information are parts of metacognitive awareness. Although kindergartners and first graders believe that familiar or easily named items are easier to remember, older children are better at predicting that categorized items are easier to recall than conceptually unrelated items (Duell,  1986 ). Older children are more likely to believe that organized stories are easier to remember than disorganized pieces of information. With respect to the goal of learning, sixth graders know better than second graders that students should use different reading strategies depending on whether the goal is to recall a story word for word or in their own words (Myers & Paris,  1978 ).

Some school tasks do not require much metacognition because they can be handled routinely. Part of the issue in the opening vignette is to use more tasks that require metacognition, with a corresponding decrease in low-level learning that can be accomplished easily.

Strategy Variables.

Metacognition depends on the strategies learners employ. Children as young as ages 3 and 4 can use memory strategies to remember information, but their ability to use strategies improves with development. Older children are able to state more ways that help them remember things. Regardless of age, children are more likely to think of external things (e.g., write a note) than internal ones (e.g., think about doing something). Students’ use of memory strategies such as rehearsal and elaboration also improves with development (Duell,  1986 ).

Although many students are capable of using metacognitive strategies, they may not know which strategies aid learning and LTM retrieval, and they may not employ those that are helpful (Flavell,  1985 ; Zimmerman & Martinez-Pons,  1990 ). Simply generating a strategy does not guarantee its use. This  utilization deficiency  is more common in younger children (Justice, Baker-Ward, Gupta, & Jannings,  1997 ) and appears to stem from children’s understanding of how a strategy works. Older learners understand that the intention to use a strategy leads to strategy use, which produces an outcome. Younger children typically have only partial understanding of the links between intentions, actions, and outcomes. Such understanding develops between the ages of 3 and 6 (Wellman,  1990 ).

Task, strategy, and learner variables typically interact when students engage in meta-cognitive activities. Learners consider the type and length of material to be learned (task), the potential strategies to be used (strategy), and their skill at using the various strategies (learner). If learners think that note taking and underlining are good strategies for identifying the main points of a technical article and if they believe they are good at underlining but poor at taking notes, they likely will decide to underline. As Schraw and Moshman ( 1995 ) noted, learners construct metacognitive theories that include knowledge and strategies that they believe will be effective in a given situation. Such metacognitive knowledge is critical for self-regulated learning (Dinsmore, Alexander, & Loughlin,  2008 ;  Chapter 10 ).

Metacognition and Behavior

Understanding which skills and strategies help us learn and remember information is necessary but not sufficient to enhance our achievement. Even students who are aware of what helps them learn do not consistently engage in metacognitive activities for various reasons. In some cases, metacognition may be unnecessary because the material is easily learned. Learners also might be unwilling to invest the effort to employ metacognitive activities. The latter are tasks in their own right; they take time and effort. Learners may not understand fully that metacognitive strategies improve their performances, or they may believe they do but that other factors, such as time spent or effort expended, are more important for learning (Borkowski & Cavanaugh,  1979 ; Flavell & Wellman,  1977 ; Schunk & Rice,  1993 ).

Metacognitive activities improve achievement, but the fact that students often do not use them presents a quandary for educators. Students need to be taught a menu of activities ranging from those applying to learning in general (e.g., determining the purpose in learning) to those applying to specific situations (e.g., underlining important points in text), and they need to be encouraged to use them in various contexts. Although the what component of learning is important, so are the when, where, and why of strategy use. Teaching the what without the latter will only confuse students and could prove demoralizing; students who know what to do but not when, where, or why to do it might hold low self-efficacy for performing well ( Chapter 4 ).

Learners often need to be taught basic declarative or procedural knowledge along with metacognitive skills (Duell,  1986 ). Students need to monitor their understanding of main ideas, but the monitoring is pointless if they do not understand what a main idea is or how to find one. Students must be encouraged to employ metacognitive strategies—this is one of the implications of the discussion at the Nikowsky Middle School—and given opportunities to apply what they have learned outside of the instructional context. Students also need feedback on how well they are applying a strategy and how strategy use improves their performance (Schunk & Rice,  1993 ; Schunk & Swartz,  1993a ). A danger of teaching a metacognitive strategy in conjunction with only a single task is that students will see the strategy as applying only to that task or to highly similar tasks, which does not foster transfer. It is desirable to use multiple tasks to teach strategies.

Metacognition During Reading

Metacognition is relevant to reading because it is involved in understanding and monitoring reading purposes and strategies (Dimmitt & McCormick,  2012 ). Beginning readers often do not understand the conventions of printed material: In the English language, one reads words from left to right and top to bottom. Beginning and poorer readers typically do not monitor their comprehension or adjust their strategies accordingly (Baker & Brown,  1984 ). Older and skilled readers are better at comprehension monitoring than are younger and less skilled readers, respectively (Alexander et al.,  1995 ).

Metacognition comes into play when learners set goals, evaluate goal progress, and make necessary corrections (McNeil,  1987 ). Skilled readers do not approach all reading tasks identically. They determine their goal: find main ideas, read for details, skim, get the gist, and so on. They then use a strategy they believe will accomplish the goal. When reading skills are highly developed, these processes may occur automatically.

While reading, skilled readers check their progress. If their goal is to locate important ideas, and if after reading a few pages they have not located any important ideas, they are apt to reread those pages. If they encounter a word they do not understand, they try to determine its meaning from context or consult a dictionary rather than continue reading.

Developmental evidence indicates a trend toward greater recognition and correction of comprehension deficiencies (Alexander et al.,  1995 ). Younger children recognize comprehension failures less often than do older children. Younger children who are good comprehenders may recognize a problem but may not employ a strategy to solve it (e.g., rereading). Older children who are good comprehenders recognize problems and employ correction strategies.

Children develop metacognitive abilities through interactions with others ( Chapter 8 ). Adults (e.g., parents, teachers) help guide children through solution steps, reminding them of their goal and assisting them to plan how to reach their goal. An effective teaching procedure includes informing children of the goal, making them aware of information relevant to the task, arranging a situation conducive to problem solving, and reminding them of their goal progress.

Strategy instruction programs generally have been successful in helping students learn strategies and maintain their use over time (Pressley & Harris,  2006 ). Brown and her colleagues advocated that strategy training incorporate practice in use of skills, instruction in how to monitor outcomes of one’s efforts, and feedback on when and where a strategy may be useful (Brown,  1980 ; Brown, Palincsar, & Armbruster,  1984 ).

Palincsar and Brown ( 1984 ) identified seventh graders with poor comprehension skills. They trained students in self-directed summarizing (review), questioning, clarifying, and predicting. Summarizing included stating what had happened in the text and also served as a self-test on the content. Questioning was directed at determining what main idea question a teacher or test might ask about that material. Clarifying was used when portions of the text were unclear and students could not adequately summarize. Predicting was used when text cues signaled forthcoming information.

Researchers taught these activities as part of an interactive dialogue between teacher and student known as  reciprocal teaching . During the lessons, an adult teacher met with students. Initially the teacher modeled the activities. The teacher and students silently read a passage, after which the teacher asked a question that a teacher or test might ask, summarized the content, clarified troublesome points, and predicted future content. Following the teacher’s modeled demonstration, the teacher and students took turns being the teacher. At first, students had difficulty assuming the role of teacher; the teacher often had to construct paraphrases and questions for students. Eventually students became more capable of following the procedure and implementing the four activities.

Compared with a condition in which students received instruction on locating information in text, reciprocal teaching led to greater comprehension gains, better maintenance over time, and better generalization to classroom comprehension tests. Students exposed to reciprocal teaching also showed greater improvements in quality of summaries and questions asked. The maintenance and generalization results are important because changes brought about by strategy training programs may not maintain themselves or generalize to other tasks (Phye,  2001 ).

The following dialogue occurred between teacher (T) and student (S) early in the training program (Palincsar & Brown,  1984 ).

·  The snake’s skeleton and parts of its body are very flexible—something like a rubber hose with bones. A snake’s backbone can have as many as 300 vertebrae, almost 10 times as many as a human’s. These vertebrae are connected by loose and rubbery tissues that allow easy movement. Because of this bendable, twistable spinal construction, a snake can turn its body in almost any direction at almost any point.

S:	Like, if a snake is turning around, he wouldn’t break any bones because he is flexible.
T:	And the reason he is so flexible is…
S:	If someone stepped on his tail, he wouldn’t be able to move unless he was flexible.
T:	Okay. That explains why it would be helpful to be so flexible, but what makes the snake so flexible? I think that is an important part of the summary.
S:	So he can move faster?
T:	That is another reason for being flexible. Let me do a summary for you. Sara [the student] is right. The most important thing we have learned is that snake[s are] very flexible. The reason they can bend so easily is that they have so many little bones in their backbones and each of them can move separately, making the snake very flexible. Does that make sense? (p. 142)

The last statement by the teacher is a modeled demonstration of summarization.

There are other reading programs that incorporate strategy instruction and teach students to self-regulate. For example, the Self-Regulated Strategy Development program combines strategy instruction with instruction in self-regulation (e.g., self-monitoring, self-instructions, goal setting, self-reinforcement; Graham & Harris,  2003 ; Harris, Graham, & Santangelo,  2013 ; Mason,  2004 ). This program has proven to be effective with children with learning disabilities and reading problems.

Concept-Oriented Reading Instruction (CORI) incorporates cognitive strategy instruction on the strategies of activating background knowledge, questioning, searching for information, summarizing, organizing graphically, and identifying story structure (Guthrie et al.,  2004 ; Guthrie, Wigfield, & Perencevich,  2004 ; Wigfield, Tonks, & Klauda,  2009 ). CORI has shown to be effective in raising students’ reading comprehension.

Motivation plays a critical role in reading comprehension. Guthrie, Wigfield, and VonSecker ( 2000 ) integrated reading strategy instruction with science content and found significant benefits on students’ motivation compared with traditional instruction emphasizing coverage of material. Student interest presumably was heightened with the real-world use of effective reading strategies. The CORI program also incorporates motivational practices such as goal setting and giving students choices. Compared with strategy instruction alone, Guthrie et al. ( 2004 ) found that CORI led to greater benefits in comprehension, motivation, and use of strategies.

Other research shows that motivational factors affect reading outcomes. Meece and Miller ( 2001 ) found that task-mastery goals predicted students’ use of learning strategies in reading instruction. After reviewing a large number of studies, Blok, Oostdam, Otter, and Overmaat ( 2002 ) concluded that computer-assisted instruction was effective in beginning reading instruction. It is possible that the motivational benefits of computers may aid in the development of early reading skill. Morgan and Fuchs ( 2007 ) examined 15 studies and obtained a positive correlation between children’s reading skills and motivation and also obtained evidence suggesting that skills and motivation can affect one another.

The increase in the number of English language learners in U.S. schools has necessitated expansion of programs for them. For English instruction students often are placed in immersion or second language programs. In immersion programs students learn English in an all-English-speaking classroom with formal or informal support when they have difficulties. In second language programs students receive instruction in reading and possibly other subjects in their native languages. Students often transition to English instruction around grade 2 or 3. Slavin and Cheung ( 2005 ) compared immersion with second language programs and found an advantage of second language programs on students’ reading competencies; however, the number of studies in their review was small, and longitudinal studies are needed to determine long-term effects.

CONCEPT LEARNING

The Nature of Concepts

Cognitive learning processes are involved in concept learning.  Concepts  are labeled sets of objects, symbols, or events that share common characteristics, or critical attributes. A concept is a mental construct or representation of a category that allows one to identify examples and nonexamples of the category. Concepts may involve concrete objects (e.g., “table,” “chair,” “cat”) or abstract ideas (e.g., “love,” “democracy,” “wholeness”). In fact, there are many types of concepts (for a detailed review, see Medin, Lynch, & Solomon,  2000 ).  Concept learning  refers to forming representations to identify attributes, generalize them to new examples, and discriminate examples from nonexamples.

Early studies by Bruner, Goodnow, and Austin ( 1956 ) explored the nature of concepts. Learners were presented with boxes portraying geometrical patterns. Each pattern could be classified using four different attributes: number of stimuli (one, two, three); shape (circle, square, cross); color (red, green, black); and number of borders on the box (one, two, three). The task was to identify the concept represented in different subsets of the boxes.

The configuration of features in a concept-learning task can be varied to yield different concepts. A conjunctive concept is represented by two or more features (e.g., two red circles). Other features (number of borders) are not relevant. A disjunctive concept is represented by one of two or more features; for example, two circles of any color or one red circle. A relational concept specifies a relationship between features that must be present, such as the number of objects in the figure must outnumber the number of borders (type of object and color are unimportant).

Bruner et al. ( 1956 ) found that learners formulated a hypothesis about the rule underlying the concept. Rules can be expressed in if-then form. A rule classifying a pet cat might be: “If it is domesticated, has four legs, fur, whiskers, a tail, is relatively small, purrs, and vocalizes ‘meow,’ then it is a cat.” Although exceptions exist, this rule will accurately classify cats most of the time. Generalization occurs when the rule is applied to a variety of cats.

People tend to form rules quickly (Bruner et al.,  1956 ). For any given concept, they retain the rule as long as it correctly identifies instances and noninstances of the concept and they modify it when it fails to do so. Learners acquire concepts better when they are presented with positive instances, or examples of the concept. Learning is much slower with negative (non-) instances. When trying to confirm the rule underlying the concept, people prefer to receive positive rather than negative instances.

The  features analysis theory  of concept learning derives from the work of Bruner and others and postulates that concepts involve rules that define the critical features, or the intrinsic (necessary) attributes, of the concept (Gagné,  1985 ; Smith & Medin,  1981 ). Through experiences with the concept, one formulates a rule that satisfies the conditions and retains the rule as long as it functions effectively.

This view predicts that different instances of a concept should be recognized equally quickly because each instance is judged against critical features; but this is not always the case. Most people find some instances of a category (e.g., a dolphin is a mammal) more difficult to verify than others (e.g., a dog is a mammal). This highlights the problem that many concepts cannot be defined precisely in terms of a set of critical attributes.

A second perspective is prototype theory (Rosch,  1973 ,  1975 ,  1978 ). A  prototype  is a generalized image of the concept, which may include only some of the concept’s defining attributes. When confronted with an instance, one recalls the most likely prototype from LTM and compares it to the instance to see if they match. Prototypes may include some nondefining (optional) attributes. In information processing terms, prototypes may be thought of as  schemas  (Andre,  1986 ), or organized forms for the knowledge we have about a particular concept ( Chapter 5 ).

Research supports the prototype theory prediction that instances closer to the prototype (e.g., prototype = “bird”; instances = “robin,” “sparrow”) are recognized quicker than those less typical (e.g., “owl,” “ostrich”; Rosch,  1973 ). One concern is that prototype theory implies that people would store thousands of prototypes in LTM, which would consume much more space than rules. A second concern is that learners easily could construct incorrect prototypes if they are allowed to include some nondefining characteristics and not all necessary ones.

Combining the features-analysis and prototype positions is possible. Given that prototypes include critical features, we might employ prototypes to classify instances of concepts that are typical (Andre,  1986 ). For instances that are ambiguous, we may employ critical feature analysis, which might modify the list of critical features to incorporate the new features.

Children’s understandings of concepts change with development and experience. There are multiple ways that conceptual change can occur (Chinn & Samarapungavan,  2009 ), including when learners recognize that rules pertinent in one domain also seem applicable to another domain (Ohlsson,  2009 ). Children in transition about the meaning of a concept may simultaneously keep a prior hypothesis in mind as they are developing a revised one (Goldin-Meadow, Alibali, & Church,  1993 ). This interpretation is consistent with Klausmeier’s position, which is discussed next.

Concept Attainment

Research indicates that there are multiple ways to learn and modify concepts (Chinn & Samarapungavan,  2009 ). One way to develop prototypes is to be exposed to a typical instance of the concept that reflects the classic attributes (Klausmeier,  1992 ). A second way is by abstracting features from two or more examples; for birds, features might be “feathers,” “two legs,” “beak,” and “flies,” although not every feature applies to every member of the class. Prototypes are refined and expanded when one is exposed to new examples of the concept; thus, “lives in the jungle” (parrot) and “lives by the ocean” (seagull).

Gagné’s ( 1985 ) theory ( Chapter 5 ) includes concepts as a central form of learning. Learners initially must have basic prerequisite capabilities to discriminate among stimulus features (i.e., distinguish relevant from irrelevant features).

In Gagné’s ( 1985 ) view, concept learning involves a multistage sequence. First, the stimulus feature is presented as an instance of the concept along with a noninstance. The learner confirms the ability to make the discrimination. In the next (generalization) stage, the learner identifies instances and noninstances. Third, the stimulus feature—which is to become the concept—is varied and presented along with noninstances. Concept attainment is verified by asking for identification of several instances of the class using stimuli not previously employed in learning. Throughout the process, correct responses are reinforced, and contiguity learning occurs ( Chapter 3 ) by presenting several instances of the concept in close association.

Klausmeier ( 1990 ,  1992 ) developed and tested a model of concept attainment. This model postulates a four-stage sequence: concrete, identity, classificatory, and formal. Competence at each level is necessary for attainment at the next level. The process of concept attainment represents an interaction of development, informal experience, and formal education.

At the concrete level, learners can recognize an item as the same one previously encountered when the context or spatial orientation in which it was originally encountered remains the same. This level requires learners to attend to the item, discriminate it as different from its surroundings on the basis of one or more defining attributes, represent it in memory as a visual image, and retrieve it from LTM to compare it with a new image and determine that it is the same item. Thus, a learner might learn to recognize an equilateral triangle and discriminate it from a right or isosceles triangle.

The identity level is characterized by recognizing an item as the same one previously encountered when the item is observed from a different perspective or in a different modality. This stage involves the same processes as at the concrete level as well as the process of generalization. Thus, the learner will be able to recognize equilateral triangles in different orientations or positions on a page.

The classificatory level requires that learners recognize at least two items as being equivalent. Additional generalization is involved; in the case of equilateral triangles, this involves recognizing a smaller and larger equilateral triangle as equivalent. The process continues until the learner can recognize examples and nonexamples; at this stage, however, the learner may not understand the basis for classification (e.g., equality of side length and angles). Being able to name the concept is not necessary at this level, but, as in the preceding stages, it can facilitate concept acquisition.

The final formal level requires the learner to identify examples and nonexamples of the concept, name the concept and its defining attributes, define the concept, and specify the attributes that distinguish it from closely related concepts (i.e., three equal sides and angles). Mastery of this stage requires the learner to implement classificatory-level cognitive processes and a set of higher-order thinking processes involving hypothesizing, evaluating, and inferring.

This stage model has instructional implications for learners at various points in development. Instruction can be spread over several grades in which concepts are periodically revisited at higher levels of attainment. Young children initially are provided with concrete referents and, with development, become able to operate at more abstract cognitive levels. For example, young children may learn the concept of “honesty” by seeing specific examples (e.g., not stealing, giving back something that is not yours), but with development they can understand the concept in more abstract and complex terms (e.g., recognize honest feedback by a supervisor of a worker’s performance; discuss benefits of honesty).

Teaching of Concepts

· Tennyson ( 1980 ,  1981 ; Tennyson, Steve, & Boutwell,  1975 ) developed a model of concept teaching based on empirical research. This model includes the following steps (Tennyson & Park,  1980 ):

· ■ Determine the structure of the concept to include superordinate, coordinate, and subordinate concepts, and identify the critical and variable attributes (e.g., features that can legitimately vary and not affect the concept).

· ■ Define the concept in terms of the critical attributes, and prepare several examples with the critical and variable attributes.

· ■ Arrange the examples in sets based on the attributes, and ensure that the examples have similar variable attributes within any set containing examples from each coordinate concept.

· ■ Order and present the sets in terms of the divergence and difficulty of the examples, and order the examples within any set according to the learner’s current knowledge.

Most concepts can be represented in a hierarchy with superordinate (higher) and subordinate (lower) concepts. For any given concept, similar concepts may be at roughly the same level in the hierarchy; these are known as coordinate concepts. For example, the concept “domestic cat” has “cat family” and “mammal” as superordinate concepts, the various breeds (short hair, Siamese) as subordinate concepts, and other members of the cat family (lion, jaguar) as coordinate concepts. The concept has critical attributes (e.g., paws, teeth) and variable attributes (e.g., hair length, eye color). A set comprises examples and nonexamples (e.g., dog, squirrel) of the concept.

Although the concept should be defined with its critical attributes before examples and nonexamples are given, presenting a definition does not ensure that students will learn the concept. Examples should differ widely in variable attributes, and nonexamples should differ from examples in a small number of critical attributes at once. This mode of presentation prevents students from overgeneralizing (classifying nonexamples as examples) and undergeneralizing (classifying examples as nonexamples).

Pointing out relationships among examples is an effective way to foster generalization. One means is by using  concept (knowledge) maps , or diagrams that represent ideas as node-link assemblies (Nesbit & Adesope,  2006 ). O’Donnell et al. ( 2002 ) showed that learning is facilitated with knowledge maps where ideas are interlinked. Nesbit and Adesope found that concept maps improved students’ knowledge retention.  Application 7.3  contains suggestions for teaching concepts.

The optimal number of examples to present depends on such concept characteristics as number of attributes and degree of abstractness of the concept. Abstract concepts usually have fewer tangible examples than concrete concepts, and examples of the former may be difficult for learners to grasp. Concept learning also depends on learner attributes such as age and prior knowledge (Tennyson & Park,  1980 ). Older students learn better than younger ones, and students with more relevant knowledge outperform those lacking such knowledge.

In teaching concepts, it is helpful to present examples that differ in optional attributes but have relevant attributes in common so that the latter can be clearly pointed out, along with the irrelevant dimensions. In teaching the concept “right triangle,” for example, the size is irrelevant, as is the direction it faces. One might present right triangles of various sizes pointing in different directions. Using worked examples is an effective cognitive instructional strategy (Atkinson, Derry, Renkl, & Wortham,  2000 ).

APPLICATION 7.3 Teaching of Concepts

Concept learning involves identifying attributes, generalizing them to new examples, and discriminating examples from nonexamples. Using superordinate, coordinate, and subordinate concepts and critical and variable attributes to present the concept to be learned should help students clearly define its structure.

A kindergarten teacher presenting a unit to teach students to identify and distinguish shapes (circle, square, rectangle, oval, triangle, diamond) might initially have children group objects alike in shape and identify critical attributes (e.g., a square has four straight sides, the sides are the same length) and variable attributes (squares, rectangles, triangles, and diamonds have straight sides but a different number of sides of different lengths and arranged in different ways). The teacher might then focus on a particular shape by presenting different examples representing each shape so children can compare attributes with those of other shapes. As for content progression, the teacher might introduce shapes familiar to students (e.g., circle and square) before moving to less common ones (e.g., parallelogram).

Ms. Lautter introduced a unit on mammals by having her elementary students sort a list of various animals into the major animal groups. Then the students discussed the major differences between the animal groups. After reviewing these facts, she focused on the amphibian group by expanding the knowledge about the physical characteristics and by reviewing other attributes such as eating habits and the ideal environment and climate.

A history teacher showed a slide portraying the various immigrant groups that settled in the United States. After reviewing the time periods when each group came to America, he and the students discussed the reasons why each group came, where they predominantly settled in the country, and what types of trades they practiced. Then they described the impact of each group separately and collectively on the growth and progress of the United States.

Not only must students learn to generalize right triangles, they also must learn to distinguish them from other triangles. To foster concept discrimination, teachers should present negative instances that clearly differ from positive instances. As students’ skills develop, they can be taught to make finer discriminations. The suggestions shown in  Table 7.2  are helpful in teaching students to generalize and discriminate among concepts.

This model requires a careful analysis of the taxonomic structure of a concept. Structure is well specified for many concepts (e.g., the animal kingdom), but for many others—especially abstract concepts—the links with higher- and lower-order concepts, as well as with coordinate concepts, are problematic.

Table 7.2 Generalizing and discriminating concepts.

Step	Examples
Name concept	Chair
Define concept	Seat with a back for one person
Give relevant attributes	Seat, back
Give irrelevant attributes	Legs, size, color, material
Give examples	Easy chair, high chair, beanbag chair
Give nonexamples	Bench, table, stool

Motivational Processes

Pintrich, Marx, and Boyle ( 1993 ) contended that conceptual change also involves motivational processes(e.g., goals, expectations, needs). These authors argued that four conditions are necessary for conceptual change to occur. First, dissatisfaction with one’s current conceptions is needed; change is unlikely if people feel their conceptions are accurate or useful. Second, the new conception must be intelligible—people must understand a conception in order to adopt it. Third, the new conception must be plausible—learners must understand how it fits with other understandings of how it might be applied. Finally, they must perceive the new conception as fruitful—being able to explain phenomena and suggesting new areas of investigation or application.

Motivational processes enter at several places in this model. For example, research shows that students’ goals direct their attention and effort, and their self-efficacy relates positively to motivation, use of effective task strategies, and skill acquisition (Schunk,  2012 ). Furthermore, students who believe that learning is useful and that task strategies are effective display higher motivation and learning (Pressley et al.,  1990 ; Schunk & Rice,  1993 ). Goals, self-efficacy, and self-evaluations of competence have been shown to promote learning and self-regulation in such domains as reading comprehension, writing, mathematics, and decision making (Pajares,  1996 ; Schunk & Pajares,  2009 ; Schunk & Swartz,  1993a ; Wood & Bandura,  1989 ; Zimmerman & Bandura,  1994 ). We see in the opening vignette that the shift toward more problem solving actually improved students’ motivation for learning.

In short, the literature suggests that conceptual change involves an interaction of students’ cognitions and motivational beliefs (Pintrich et al.,  1993 ), which has implications for teaching. Rather than simply help learners construct knowledge, teachers must take students’ preexisting ideas into account when planning instruction and ensure that instruction includes motivation for learning.

These ideas are highly applicable to science, where knowledge is constructed by learners rather than simply transmitted (Driver, Asoko, Leach, Mortimer, & Scott,  1994 ; Linn & Eylon,  2006 ). An interesting issue is how students develop scientific misconceptions and simplistic scientific models (Windschitl & Thompson,  2006 ). An important task is to help students challenge and correct misconceptions (Sandoval,  1995 ). Experiences that produce cognitive conflict can be helpful (Mayer,  1999 ; Sandoval,  1995 ; Williams & Tolmie,  2000 ). This might entail having students engage in hands-on activities and work with others (e.g., in discussions) to interpret their experiences through selective questioning (e.g., “Why do you think that?” “How did you figure that?”). This approach fits well with the Vygotskian emphasis on social influences on knowledge construction ( Chapter 8 ).

The role of motivation is critical. Although science has many themes that ought to be interesting, studying science holds little interest for many students. Learning benefits from hands-on instruction and links to aspects of students’ lives. For example, motion can be linked to the path of soccer balls, electricity to DVD players, and ecology to community recycling programs. Enhancing interest in topics also can improve the quality of student learning (Sandoval,  1995 ). Thus, using illustrations and diagrams helps students to understand scientific concepts (Carlson, Chandler, & Sweller,  2003 ; Hannus & Hyönä,  1999 ), although some students may need to be taught how to study illustrations as part of text learning.

PROBLEM SOLVING

One of the most important types of information processing that occurs during learning is problem solving. Although problem solving has been studied for a long time, interest in the topic has grown in recent years with the ascendance of cognitive learning theories. Some theorists consider problem solving to be the key process in learning, especially in domains such as science and mathematics (Anderson,  1993 ). “Problem solving” and “learning” are not synonymous, but the former often is involved in the latter—especially when learners engage in self-regulated learning ( Chapter 10 ) and when learning involves challenges and nonobvious solutions. In the opening vignette, Meg recommends more emphasis on problem solving.

A  problem  exists when there is a “situation in which you are trying to reach some goal, and must find a means for getting there” (Chi & Glaser,  1985 , p. 229). The problem may be to answer a question, compute a solution, find information using the Internet, locate an object, secure a job, teach a student, and so on.  Problem solving  refers to people’s efforts to achieve a goal for which they do not have an automatic solution.

Regardless of content area and complexity, all problems have certain commonalities. A problem has an initial state—the problem solver’s current status or level of knowledge, as well as a goal, or what the problem solver is attempting to attain. Most problems also require the solver to break the goal into subgoals that, when mastered (usually sequentially), result in goal attainment. Finally, a problem requires performing operations (cognitive and behavioral activities) on the initial state and the subgoals, which alter the nature of those states (Anderson,  1990 ; Chi & Glaser,  1985 ).

Given this definition, not all learning activities include problem solving. Problem solving likely is not involved when students’ skills become so well established that they automatically execute actions to attain goals, which happens with many skills in different domains. Problem solving also may not occur in low-level (possibly trivial) learning, where students know what to do to learn. This seems to be an issue at Nikowsky Middle School, as teachers are focusing on basic skills needed for the tests. At the same time, students learn new skills and new uses for previously learned skills, so most school learning might involve problem solving at some point.

Problem solving skills can be developed. Encouraging young children’s tool use (i.e., using a tool such as a rake to obtain objects) facilitates their problem solving (Keen,  2011 ). With development, students’ problem solving benefits more from concrete visual representations, or illustrations of real-life elements, than from abstract representations during instruction (Moreno, Ozogul, & Reisslein,  2011 ).

Historical Perspectives

Two historical perspectives on problem solving are examined as a backdrop to current cognitive views: trial and error, and insight.

Trial and Error.

Thorndike’s ( 1913b ) research with cats ( Chapter 3 ) required problem solving; the problem was how to escape from the cage. Thorndike conceived of problem solving as  trial and error . The animal was capable of performing certain behaviors in the cage. From this behavioral repertoire, the animal performed one behavior and experienced the consequences. After a series of random behaviors, the cat made the response that opened the hatch leading to escape. With repeated trials, the cat made fewer errors before performing the escape behavior, and the time required to solve the problem diminished. The escape behavior (response) became connected to cues (stimuli) in the cage.

We occasionally use trial and error to solve problems; we simply perform actions until one works. But trial and error is not reliable and often not effective. It can waste time, may never result in a solution, may lead to a less-than-ideal solution, and can have negative effects. In desperation, a teacher might use a trial-and-error approach by trying different reading materials with Kayla until she begins to read better. This approach might be effective but also might expose her to materials that prove frustrating and thereby retard her reading progress.

Insight.

· Problem solving often is thought to involve  insight , or the sudden awareness of a likely solution. Wallas ( 1921 ) studied great problem solvers and formulated a four-stage model as follows:

· ■ Preparation: A time to learn about the problem and gather information that might be relevant to its solution.

· ■ Incubation: A period of thinking about the problem, which may also include putting the problem aside for a time.

· ■ Illumination: A period of insight when a potential solution suddenly comes into awareness.

· ■ Verification: A time to test the proposed solution to ascertain whether it is correct.

Wallas’s stages were descriptive and not subjected to empirical verification. Hélie and Sun ( 2010 ) presented a more detailed, process-oriented conceptualization of the incubation and illumination stages. Gestalt psychologists ( Chapter 5 ) also postulated that much human learning was insightful and involved a change in perception. Learners initially thought about the ingredients necessary to solve a problem. They integrated these in various ways until the problem was solved. When learners arrived at a solution, they did so suddenly and with insight.

Many problem solvers report having moments of insight; Watson and Crick had insightful moments in discovering the structure of DNA (Lemonick,  2003 ). An important educational application of Gestalt theory was in the area of problem solving, or  productive thinking  (Duncker,  1945 ; Luchins,  1942 ; Wertheimer,  1945 ). The Gestalt view stressed the role of understanding—comprehending the meaning of some event or grasping the principle or rule underlying performance. In contrast, rote memorization—although used often by students—was inefficient and rarely used in life outside of school ( Application 7.4 ).

Research by Katona ( 1940 ) demonstrated the utility of rule learning compared with memorization. In one study, participants were asked to learn number sequences (e.g., 816449362516941). Some learned the sequences by rote, whereas others were given clues to aid learning (e.g., “Think of squared numbers”). Learners who determined the rule for generating the sequences retained them better than those who memorized.

APPLICATION 7.4 Role of Understanding in Learning

Teachers want students to understand concepts rather than simply memorize how to complete tasks. Gestalt psychologists believed that an emphasis on drill and practice, memorization, and reinforcement resulted in trivial learning and that understanding was achieved by grasping rules and principles underlying concepts and skills.

Teachers often use hands-on experiences to help students understand the structure and principles involved in learning. In biology, students might memorize what a cross section of a bean stem looks like under a microscope, but they may have difficulty conceptualizing the structures in the living organism. Mock-ups assist student learning. A large, hands-on model of a bean stem that can be taken apart to illustrate the internal structures should enhance student understanding of the stem’s composition and how the parts function.

Talking about child care in a high school family studies class is not nearly as beneficial as the hour students spend each week with children in a day care center and applying what they have been studying.

In discussing the applications of learning theories, it is preferable that students see firsthand the utilization of techniques that enhance student learning. To gain understanding when students in an educational psychology course observe in classrooms, they list examples of situations where various learning principles are evident.

Rules lead to better learning and retention than memorization because rules give a simpler description of the phenomenon so less information must be learned. In addition, rules help organize material. To recall information, one recalls the rule and then fills in the details. In contrast, memorization entails recalling more information. Memorization generally is inefficient because most situations have some organization (Wertheimer,  1945 ). Problems are solved by discovering the organization of the situation and the relationship of the elements to the problem solution. By arranging and rearranging elements, learners eventually gain insight into the solution.

Köhler ( 1925 ,  1926 ) did well-known work on problem solving with apes on the island of Tenerife during World War I. In one experiment, Köhler put a banana just out of reach of an ape in a cage; the ape could fetch the banana by using a long stick or by putting two sticks together. Köhler concluded that problem solving was insightful: Animals surveyed the situation, suddenly “saw” the means for attaining the goal, and tested the solution. The apes’ first problem-solving attempts failed as they tried different ineffective strategies (e.g., throwing a stick at the banana). Eventually they saw the stick as an extension of their arms and used it accordingly.

A barrier to problem solving is  functional fixedness , or the inability to perceive different uses for objects or new configurations of elements in a situation (Duncker,  1945 ). In a classic study, Luchins ( 1942 ) gave individuals problems that required them to obtain a given amount of water using three jars of different sizes. Persons from ages 9 to adult easily learned the formula that always produced the correct amount. Intermixed in the problem set were some problems that could be solved using a simpler formula. Persons generally continued to apply the original formula. Cuing them that there might be an easier solution led some to discover the simpler methods, although many persisted with the original formula. This research shows that when students do not understand a phenomenon, they may blindly apply a known algorithm and fail to understand that easier methods exist. This procedure-bound nature of problem solving can be overcome when different procedures are emphasized during instruction (Chen,  1999 ).

Gestalt theory had little to say about how problem-solving strategies are learned or how learners could be taught to be more insightful. Wertheimer ( 1945 ) believed that teachers could aid problem solving by arranging elements of a situation so that students would be more likely to perceive how the parts relate to the whole. Such general advice may not be helpful for teachers.

Heuristics

· Another way to solve problems is to use  heuristics , which are general methods for solving problems that employ principles (rules of thumb) that usually lead to a solution (Anderson,  1990 ). Polya’s ( 1945/1957 ) list of mental operations involved in problem solving is as follows:

· ■ Understand the problem.

· ■ Devise a plan.

· ■ Carry out the plan.

· ■ Look back.

Understanding the problem involves asking such questions as “What is the unknown?” and “What are the data?” It often helps to draw a diagram representing the problem and the given information. In devising a plan, one tries to find a connection between the data and the unknown. Breaking the problem into subgoals is useful, as is thinking of a similar problem and how that was solved (i.e., use analogies). The problem may need to be restated. While carrying out the plan, checking each step to ensure it is being properly implemented is important. Looking back means examining the solution: Is it correct? Is there another means of attaining it?

Bransford and Stein ( 1984 ) formulated a similar heuristic known as IDEAL:

· ■ Identify the problem.

· ■ Define and represent the problem.

· ■ Explore possible strategies.

· ■ Act on the strategies.

· ■ Look back and evaluate the effects of your activities.

General heuristics are most useful when one is working with unfamiliar content (Andre,  1986 ). They are less effective in a familiar domain, because as domain-specific knowledge develops, students increasingly use it. General heuristics have an instructional advantage: They can help students become systematic problem solvers. Although the heuristic approach may appear to be inflexible, there actually is flexibility in how steps are carried out. For many students, a heuristic will be more systematic than their current problem-solving approaches and will lead to better solutions.

Problem-Solving Strategies

Newell and Simon ( 1972 ) proposed an information processing model of problem solving that included a  problem space  with a beginning state, a goal state, and possible solution paths leading through subgoals and requiring application of operations. The problem solver forms a mental representation of the problem and performs operations to reduce the discrepancy between the beginning and goal states. The process of operating on the representation to find a solution is known as the search (Andre,  1986 ).

The first step in problem solving is to form a mental representation. Similar to Polya’s first step (understand the problem), representation requires  translating  known information into a model in memory. The internal representation consists of propositions, and possibly images, in WM. The problem also can be represented externally (e.g., on paper or computer). Information in WM activates related knowledge in LTM, and the solver eventually selects a problem-solving strategy. As people solve problems, they often alter their initial representation and activate new knowledge, especially if their problem solving does not succeed. Thus, problem solving includes evaluating goal progress.

The problem representation determines what knowledge is activated in memory and, consequently, how easy the problem is to solve (Holyoak,  1984 ). If solvers incorrectly represent the problem by not considering all aspects or by adding too many constraints, the search process is unlikely to identify a correct solution path (Chi & Glaser,  1985 ). No matter how clearly solvers subsequently reason, they will not reach a correct solution unless they form a new representation. Not surprisingly, problem-solving training programs typically devote a lot of time to the representation phase (Andre,  1986 ).

Like skills (discussed earlier), problem-solving strategies can be general or specific. General strategies can be applied to problems in several domains regardless of content; specific strategies are useful only in a particular domain. For example, breaking a complex problem into subproblems (subgoal analysis) is a general strategy applicable to problems such as writing a term paper, choosing an academic major, and deciding where to live. Conversely, tests that one might perform to classify laboratory specimens are task specific. The professional development given to Nikowsky’s teachers probably included general and specific strategies.

General strategies are useful when one is working on problems where solutions are not immediately obvious. Useful general strategies are generate-and-test, means–ends analysis, analogical reasoning, and brainstorming. The first three are discussed here; brainstorming is covered later in this chapter. These general strategies are less useful than domain-specific strategies when working with highly familiar content. Some examples of problem solving in learning contexts are given in  Application 7.5 .

Generate-and-Test.

Generate-and-test  is a useful strategy when a limited number of problem solutions can be tested to see if they attain the goal (Resnick,  1985 ). This strategy works best when multiple solutions can be ordered in terms of likelihood and at least one solution is apt to solve the problem.

As an example, assume that you walk into a room, flip the light switch, but the light does not come on. Possible causes include: the bulb is burned out; the electricity is turned off; the switch is broken; the lamp socket is faulty; the circuit breaker is tripped; the fuse is blown; or the wiring has a short. You will probably generate and test the most likely solution (replace the bulb); if this does not solve the problem, you may generate and test other likely solutions. Although content does not need to be highly familiar, some knowledge is needed to use this method effectively. Prior knowledge establishes the hierarchy of possible solutions; current knowledge influences solution selection. Thus, if you notice no lights are on in your neighborhood you would suspect that the power is off.

Means–Ends Analysis.

To use  means–ends analysis , one compares the current situation with the goal and identifies the differences between them (Resnick,  1985 ). Subgoals are set to reduce the differences. One performs operations to accomplish the subgoal, at which point the process is repeated until the goal is attained.

Newell and Simon ( 1972 ) studied means–ends analysis and formulated the General Problem Solver (GPS)—a computer simulation program. GPS breaks a problem into subgoals, each representing a difference from the current state. GPS starts with the most important difference and uses operations to eliminate that difference. In some cases, the operations must first eliminate another difference that is prerequisite to the more important one.

Means–ends analysis is a powerful problem-solving heuristic. When subgoals are properly identified, means–ends analysis is most likely to solve the problem. One drawback is that with complex problems means–ends analysis can create a heavy cognitive load for WM because one may have to keep track of several subgoals. Forgetting a subgoal thwarts problem solution.

Means–ends analysis can proceed from the goal to the initial state ( working backward ) or from the initial state to the goal ( working forward ). In working backward, one starts with the goal and asks what subgoals are necessary to accomplish it. One then asks what is necessary to attain these subgoals and so forth, until the initial state is reached. To work backward, therefore, one plans a series of moves, each designed to attain a subgoal. Successfully working backward requires a fair amount of knowledge in the problem domain to determine goal and subgoal prerequisites.

APPLICATION 7.5 Problem Solving

· Various ways exist to help students improve their problem-solving skills. When middle school students solve mathematical word problems, Mr. Quinn encourages them to state each problem in their own words, draw a sketch, decide what information is relevant, and state the ways they might solve the problem. These and other similar questions help focus students’ attention on important task aspects and guide their thinking:

· ■ What information is important?

· ■ What information is missing?

· ■ Which formulas are necessary?

· ■ What is the first thing to do?

Another way to assist students is to encourage them to view a problem from varying perspectives. In a world history class, high school students discussed how to categorize major wartime figures (e.g., Churchill, Hitler). They determined different ways these individuals could be categorized, such as by personality type, political makeup of their countries, goals of the war, and the effects that their leadership and goals had. This exercise illustrates different ways to organize information, which aids problem solving.

Teachers also can teach strategies. In a geography lesson, students might be given the following problem: “Pick a state (not your own) that you believe could attract new residents, and create a poster depicting the most important attributes of that state.” A working backward strategy could be taught as follows:

·   Goal : Create a poster depicting the state’s important attributes.

Subgoal: Decide how to portray the attributes in a poster.

Subgoal: Decide which attributes to portray.

Subgoal: Decide which state to pick.

Initial Subgoal: Decide which attributes attract new residents.

To attain the initial subgoal, students could collaborate in small groups to determine which factors attract people to a state. They then could conduct library research to check on which states possess these attributes. Students could reconvene to discuss the attributes of different states and decide on one. They then would decide which attributes to portray in the poster and how to portray them, after which they would create their poster and present it to the class.

When students are developing problem-solving skills, teachers might want to give clues rather than answers. A teacher working with younger children on categorizing might give the children a word list of names of animals, colors, and places to live. Children are most likely to experience some difficulty categorizing the names. Rather than telling them the answers, the teacher could provide clues such as, “Think of how the words go together. How are horse and lion alike? How are pink and house different?”

Figure 7.2 Means–ends analysis applied to a geometry problem.

Working backward may be used to prove geometric theorems. One starts by assuming that the theorem is true and then works backward until the postulates are reached. A geometric example is shown in  Figure 7.2 . The problem is to solve for angle m. Working backward, students realize that they need to determine angle n, because angle m = 180° − angle n (straight line = 180°). Continuing to work backward, students understand that because the parallel lines intersect, the corresponding angle d on line q equals angle n. Drawing on their geometric knowledge, students determine that angle d = angle a, which is 30°. Thus, angle n = 30°, and angle m = 180° − 30° = 150°.

As another example of working backward, suppose one has a term paper due in 3 weeks. The last step before turning it in is to proofread it (to do the day before the paper is due). The step before that is to type and print the final copy (allow 1 day). Before that, one makes final revisions (1 day), revises the paper (3 days), and types and prints the draft copy (1 day). Continuing to work backward, we might allow 5 days to write the draft, 1 day to outline, 3 days for library research, and 1 day to decide on a topic. We allow a total of 17 days to spend in part working on the paper. So we need to begin 4 days from today.

A second type of means–ends analysis is working forward, sometimes referred to as  hill climbing  (Matlin,  2009 ; Mayer,  1992 ). The problem solver starts with the current situation and alters it in the hope of moving closer to the goal. Several alterations usually are necessary to attain the goal. One danger is that working forward sometimes proceeds based on superficial problem analysis. Although each step represents an attempt to attain a necessary subgoal, one can easily veer off on a tangent or arrive at a dead end because typically one cannot see many alternatives ahead but rather only the next step (Matlin,  2009 ).

As an example of a working forward strategy, consider students in a laboratory who have various substances in jars. Their goal is to label the substances in their jars. To do so, they perform a series of tests on the substances which, if correctly done, will result in a solution. This represents a working forward strategy because each test moves students closer to their goal of classifying their substances. The tests are ordered, and the results show what the substances are not, as well as what they might be. To prevent students from going off on the wrong track, the teacher sets up the procedure carefully and ensures that students understand how to perform the tests.

Analogical Reasoning.

Another general problem-solving strategy is to use  analogical reasoning , which involves drawing an analogy between the problem situation (the target) and a situation with which one is familiar (the base or source; Anderson,  1990 ; Chen,  1999 ; Hunt,  1989 ). One works the problem through the familiar domain and then relates the solution to the problem situation (Holyoak & Thagard,  1997 ). Analogical reasoning involves accessing the familiar domain’s network in LTM and mapping it onto (relating it to) the problem situation in WM (Halpern, Hansen, & Riefer,  1990 ). Successful application requires that the familiar situation be structurally similar to the problem situation, although the situations may differ in surface features (e.g., one might involve the solar system and the other molecular structures). The subgoals in this approach are relating the steps in the original (familiar) domain to those in the transfer (problem) area. Students often use the analogy method to solve problems in textbooks. Examples are worked in the text (familiar domain), then students relate these steps to the problems they must solve.

Gick and Holyoak ( 1980 ,  1983 ) demonstrated the power of analogical problem solving. They presented learners with a difficult medical problem and, as an analogy, a solved military problem. Simply giving them the analogical problem did not automatically prompt them to use it. However, giving them a hint to use the military problem to solve the medical problem improved problem solving. Gick and Holyoak also found that giving students two analog stories led to better problem solving than giving one story, but what did not enhance problem solving was having them summarize the analog story, giving them the principle underlying the story while they read it, or providing them with a diagram illustrating the problem-solution principle. These results suggest that in an unfamiliar domain, students need guidance for using analogies and that multiple examples increase the likelihood of students’ linking at least one example to the problem to be solved.

To be most effective, analogical problem solving requires good knowledge of the familiar and problem domains. Students often have enough difficulty using analogies to solve problems even when the solution strategy is highlighted. With inadequate knowledge, students are unlikely to see the relation between the problem and the analogue. Even assuming good knowledge, the analogy is most likely to fail when the familiar and problem domains are conceptually dissimilar. Learners may understand how fighting a battle (the military problem) is similar to fighting a disease (the medical problem), but they may not grasp other analogies (e.g., fighting a corporate takeover attempt).

Developmental evidence indicates that, despite its difficulties, children can employ analogical reasoning (Siegler,  1989 ). Teaching analogies to children—including those with learning disabilities—can improve their subsequent problem solving (Grossen, 1991). The use of case studies and case-based reasoning can help develop analogical thinking (Kolodner,  1997 ). Effective techniques for using analogies include having the teacher and child verbalize the solution principle that underlies the original and transfer problems, prompting children to recall elements of the original problem’s causal structure, and presenting the two problems such that the causal structures proceed from most to least obvious (Crisafi & Brown,  1986 ). Other suggestions include using similar original and transfer problems, presenting several similar problems, and using pictures to portray causal relations.

This is not to suggest that all children can learn to use analogies well. The task is difficult, and children often draw inappropriate analogies. Compared with older students, younger ones require more hints, are more apt to be distracted by irrelevant perceptual features, and process information less efficiently (Crisafi & Brown,  1986 ). Children’s success depends on their knowledge about the original problem and their skill at encoding and making mental comparisons, which show wide individual differences (Richland, Morrison, & Holyoak,  2006 ; Siegler,  1989 ). Children learn problem-solving strategies better when they observe and explain them than when they merely observe (Crowley & Siegler,  1999 ).

Analogical problem solving is useful in teaching. Teachers often have students in their classes whose native language is not English. Teaching students in their native language is impossible. Teachers might relate this problem to teaching students who have difficulty learning. With the latter students, teachers would proceed slowly, use concrete experiences whenever possible, and provide much individual instruction. They might try the same tactics with English language learners while simultaneously teaching them English words and phrases so they can keep up with the other students in class.

This analogy is appropriate because students with learning problems and students who speak little English have difficulties in the classroom. Other analogies might be inappropriate. Unmotivated students also have learning difficulties. Using them for the analogy, the teacher might offer the English language learners rewards for learning. This solution may not be effective because the issue with English language learners is instructional rather than motivational.

Problem Solving and Learning

According to a contemporary information processing view (Anderson,  1990 ,  1993 ,  2000 ), problem solving involves the acquisition, retention, and use of  production systems , which are networks of condition–action sequences (rules) in which the conditions are the sets of circumstances that activate the system and the actions are the sets of activities that occur (Anderson,  1990 ; Andre,  1986 ;  Chapter 5 ). A production system consists of if-then statements. If statements (the condition) include the goal and test statements, then statements are the actions.

Productions are forms of procedural knowledge that include declarative knowledge and the conditions under which these forms are applicable. Productions are represented in LTM as propositional networks and are acquired in the same fashion as other procedural knowledge. Productions also are organized hierarchically with subordinate and superordinate productions. To solve two equations with two unknowns, one first represents one unknown in terms of the second unknown (subordinate production), after which one solves for the second unknown (production) and uses that value to solve for the first unknown (superordinate production).

Productions can be general or specific. Specific productions apply to content in well-defined areas. In contrast, heuristics are general productions because they apply to diverse content. A means–ends analysis might be represented as follows (Anderson,  1990 ):

·  IF the goal is to transform the current state into the goal state and D is the largest difference between the states

THEN set as subgoals

· 1. To eliminate the difference D

· 2. To convert the resulting state into the goal state. (p. 243)

A second production will then need to be employed with the if-then statement, “If the goal is to eliminate the difference D.” This sequence continues until the subgoals have been identified at a specific level; then domain-specific rules are applied. In short, general productions are broken down until the level at which domain-specific knowledge is applied. Production systems offer a means of linking general with specific problem-solving procedures. Other problem-solving strategies (e.g., analogical reasoning) also can be represented as productions.

School learning that is highly regulated may not require problem solving. Problem solving is not applicable when students have a goal and a clear means for attaining it. Problem solving becomes more important when teachers move away from regimented instruction and encourage more original and critical thinking by students. This is what the teachers at Nikowsky worked on after their meeting with Meg.

Experts and Novices

As with skill acquisition, researchers have identified differences between novice and expert problem solvers (Anderson,  1990 ,  1993 ; Bruning, Schraw, & Norby,  2011 ; Resnick,  1985 ). One difference involves the demands made on WM. Expert problem solvers do not activate large amounts of potentially relevant information; they identify key features of the problem, relate them to background knowledge, and generate one or a small number of potential solutions (Mayer,  1992 ). Experts reduce complex problems to manageable size by separating the problem space from the larger task environment, which includes the domain of facts and knowledge within which the problem is embedded (Newell & Simon,  1972 ). Coupled with the fact that experts can hold more information in WM (Chi, Glaser, & Farr,  1988 ), this reduction process retains relevant information, discards irrelevant information, fits within the limits of WM and does not cause excessive cognitive load, and is accurate enough to allow a solution.

Experts often employ a working forward strategy by identifying the problem format and generating an approach to fit it (Mayer,  1992 ). This typically entails breaking the problem into parts and solving the parts sequentially (Bruning et al.,  2011 ). Novice problem solvers, however, often attempt problem solving in piecemeal fashion, in part because of the poorer organization in their memories. They may use trial and error or try to work backward from what they are trying to find to the problem givens—an ineffective strategy if they are unaware of the substeps needed (Mayer,  1992 ). Their means–ends analyses often are based on surface features of problems. In mathematics, novices generate formulas from memory when confronted with word problems. Trying to store too much information in WM causes excessive cognitive load (Kalyuga, Renkl, & Paas,  2010 ).

Experts and novices appear to be comparably versed in knowledge of general problem-solving strategies (Elstein, Shulman, & Sprafka,  1978 ; Simon,  1979 ). Such generalized knowledge structures are essential for problem solving (Kalyuga et al.,  2010 ). But expert performers have more extensive and better organized LTM domain-specific knowledge (Chi et al.,  1981 ). The greater amount of knowledge experts can use in solving problems, the more likely they are to solve them and the better their memory organization facilitates efficiency.

Qualitative differences are evident in how knowledge is structured in memory (Chi, Glaser, & Rees,  1982 ). Experts’ knowledge is more hierarchically organized. Experts tend to classify problems according to “deep structure,” whereas novices rely more on surface features (Hardiman, Dufresne, & Mestre,  1989 ). Teaching novices to recognize deep features improves their performances.

Novices typically respond to problems in terms of how they are presented; experts reinterpret problems to reveal an underlying structure, one that most likely matches their own LTM network (Resnick,  1985 ). Novices attempt to translate the given information directly into formulas and solve for the missing quantities. Rather than generate formulas, experts may initially draw diagrams to clarify the relations among problem aspects. They often construct a new version of the problem. By the time they are ready to perform calculations, they usually have simplified the problem and perform fewer calculations than novices. While working, experts monitor their performances better to assess goal progress and the value of the strategy they are using.

Finally, experts spend more time planning and analyzing. They are more thoughtful and do not proceed until they have a strategy in mind. Experienced teachers spend more time planning than do less experienced teachers, as well as more time exploring new classrooms (Moore,  1990 ). Such planning makes strategy implementation easier.

· In summary, the differences between novice and expert problem solvers are many. Compared with novices, experts:

· ■ Possess more declarative knowledge

· ■ Have better hierarchical organization of knowledge

· ■ Spend more time planning and analyzing

· ■ Recognize problem formats more easily

· ■ Represent problems at a deeper level

· ■ Monitor their performances more carefully

· ■ Understand better the value of strategy use

CRITICAL THINKING, REASONING, AND CREATIVITY

In addition to metacognition, concept learning, and problem solving, complex cognitive processes include critical thinking, reasoning, and creativity.

Critical Thinking

The educators in the opening vignette struggled with how to incorporate more critical thinking into the curriculum.  Critical thinking  is reflective cognitive activity focused on deciding what to do or what to believe (Ennis,  1987 ). Critical thinking involves how to think rather than what to think. It is, essentially, better or deeper thinking.

Unlike problem solving that is focused on obtaining a solution to a problem, critical thinking is focused on understanding the nature of the problem. Problem solving also tends to be focused on specific domains (e.g., science, mathematics), whereas critical thinking may occur at a more general level (e.g., the effects of pollution) and may cut across multiple domains (Bruning et al.,  2011 ).

Critical thinking may, of course, include aspects of problem solving. We may want not only to understand the effects of pollution but also to generate some solutions to the problems it creates. However, as generally construed, critical thinking does not require decisions or solutions, rather only more complete understanding.

Investigators have proposed various components of critical thinking. Four that seem important are knowledge, inference, evaluation, and metacognition (Bruning et al.,  2011 ; Halpern,  1998 ). Some knowledge of the issue being considered helps individuals ask questions and judge new information or perspectives. Knowledge of strategies as discussed in this and other chapters can help focus the direction critical thinking takes. As a result of engaging in critical thinking, people acquire new knowledge.

Inference refers to making connections between two or more units of knowledge (Bruning et al.,  2011 ). Making inferences helps people understand issues better and at deeper levels. Later in this chapter two types of inference processes are discussed: deductive and inductive reasoning.

Evaluation refers to processes such as analyzing, judging, and weighing evidence. By analyzing we identify and select information that seems relevant to the issue at hand. Judging serves to assess the credibility of information or evidence and can help eliminate bias. Weighing means that we compare information we have and organize it into a fashion that makes sense to us.

Metacognition is a key aspect of critical thinking. Earlier we saw that metacognition was “thinking about thinking.” Metacognitive activities help us monitor our thought processes and reflect on the adequacy of the conclusions we draw. Through metacognitive activities, we may decide that we have thought about an issue sufficiently or conversely that we are not ready to make a decision because we need more information.

Reasoning

Reasoning  refers to the mental processes involved in generating and evaluating logical arguments (Anderson,  1990 ). Reasoning yields a conclusion from thoughts, percepts, and assertions (Johnson-Laird,  1999 ) and involves working through problems or issues to explain why something happened or what will happen (Hunt,  1989 ). Reasoning skills include clarification, basis, inference, and evaluation (Ennis,  1987 ; Quellmalz,  1987 ;  Table 7.3  and  Application 7.6 ). Note the partial overlap of some of these with critical thinking skills.

Table 7.3 Reasoning skills.

Skill	Definition	Sample Questions
Clarification	Identifying and formulating questions, analyzing elements, defining terms	“What do I know?” “What do I need to figure out?”
Basis	Determining source(s) of support for conclusions about a problem	“Is this a fact or opinion?” “What is the source of this information?”
Inference	Reasoning inductively from specific cases to general principles or deductively from general principles to specific cases	“What do these diverse examples have in common?” (induction) “How can I apply these general rules to this example?” (deduction)
Evaluation	Using criteria to judge adequacy of a problem solution	“Do I need more information?” “Is my conclusion reasonable?”

APPLICATION 7.6 Reasoning

· Students can learn how to ask questions to produce an accurate mental representation of a problem. A teacher might give elementary students objects to classify according to shape. To help students identify and clarify the problem, the teacher could ask questions such as:

· ■ What have you been asked to do?

· ■ What items do you have?

· ■ What are some of the shapes you know?

· ■ Does it matter if the items are different colors?

· ■ Does it matter if some of the items are little and some are big?

· ■ Does it matter if some of the items are soft and some are hard?

· ■ What do you think you will do with the items you have?

Students verbalize what information they need to use and what they are supposed to do with that information. Each time the teacher works with students in solving a problem, the teacher can help them generate questions to determine what information is important for solving the problem.

· A medical researcher working with a group of interns gives them information about a virus, and their task is to identify the virus. To assist the students in the identification process, the instructor might generate a list of questions similar to the following:

· ■ What effect does the virus have on blood cells?

· ■ What effect does the virus have on human tissue?

· ■ How quickly does the virus appear to grow, and under what conditions does it grow?

· ■ What does the virus do when exposed to warmth?

· ■ What does the virus do when exposed to cold?

· ■ What does the virus do when exposed to moisture?

· ■ What does the virus do in an airtight environment?

· ■ What reaction does the virus have when exposed to various drugs?

Clarification.

Clarification requires identifying and formulating questions, analyzing elements, and defining terms. These skills involve determining which elements in a situation are important, what they mean, and how they are related. At times, scientific questions are posed, but at other times students must develop questions such as “What is the problem, hypothesis, or thesis?” Clarification corresponds to the representation phase of problem solving; students define the problem to obtain a clear mental representation. Little productive reasoning occurs without a clear problem statement.

Basis.

People’s conclusions about a problem are supported by information from personal observations, statements by others, and previous inferences. Judging the credibility of a source is important. In so doing, one must distinguish between fact, opinion, and reasoned judgment. Assume that a suspect armed with a gun is apprehended near the scene of a murder. That the suspect had a gun when arrested is a fact. Laboratory tests on the gun, the bullets, and the victim lead to the reasoned judgment that the gun was used in the crime. Someone investigating the case might be of the opinion that the suspect is the murderer.

Inference.

Scientific reasoning proceeds inductively or deductively.  Inductive reasoning  refers to developing general rules, principles, and concepts from observations and knowledge of specific examples (Pellegrino,  1985 ). It requires determination of a model and its associated rules of inference (Hunt,  1989 ). People reason inductively when they extract similarities and differences among specific objects and events and arrive at generalizations, which are tested by applying them to new experiences. Individuals retain their generalizations as long as they are effective, and they modify them when they experience conflicting evidence.

Some of the more common types of tasks used to assess inductive reasoning are classification, concept, and analogy problems. Consider the following analogy (Pellegrino,  1985 ):

·  sugar : sweet :: lemon : ______

yellow sour fruit squeeze tea

The appropriate mental operations may represent a type of production system. Initially, the learner mentally represents critical attributes of each term in the analogy. She activates networks in LTM involving each term, which contain critical attributes of the terms to include subordinate and superordinate concepts. Next, she compares the features of the first pair to determine the link. “Sweet” is a property of sugar that involves taste. She then searches the “lemon” network to determine which of the five features listed corresponds in meaning to “lemon” as “sweet” does to “sugar.” Although all five terms are most likely stored in her “lemon” network, only “sour” directly involves taste.

Children begin to display basic inductive reasoning around age 8. With development, children can reason faster and with more complex material. This occurs because their LTM networks become more complex and better linked, which in turn reduces the burden on the WM. To help foster inductive thinking, teachers might use a guided discovery approach ( Chapter 8 ) in which children learn different examples and try to formulate a general rule. For example, children may collect leaves and formulate some general principles involving stems, veins, sizes, and shapes of leaves from different trees. Or teachers might pose a problem such as “Why does metal sink in water but metal ships float?” Rather than tell students how to solve the problem, the teacher might provide materials and encourage them to formulate and test hypotheses as they work on the task. Phye ( 1997 ; Klauer & Phye,  2008 ) discussed effective teaching methods and programs that have been used to teach inductive reasoning to students.

Deductive reasoning  refers to applying inference rules to a formal model of a problem to decide whether specific instances logically follow. When individuals reason deductively, they proceed from general concepts (premises) to specific instances (conclusions) to determine whether the latter follow from the former. A deduction is valid if the premises are true and if the conclusion follows logically from the premises (Johnson-Laird,  1985 ,  1999 ).

Linguistic and deductive reasoning processes are intimately linked (Falmagne & Gonsalves,  1995 ; Polk & Newell,  1995 ). One type of deduction problem is the three-term series (Johnson-Laird,  1972 ). For example,

·  If Karen is taller than Tina, and

If Mary Beth is not as tall as Tina, then

Who is the tallest?

The problem-solving processes employed with this problem are similar to those discussed previously. Initially one forms a mental representation of the problem, such as K > T, MB < T. One then works forward by combining the propositions (K > T > MB) to solve the problem. Developmental factors limit children’s proficiency in solving such problems. Children may have difficulty keeping relevant problem information in WM and may not understand the language used to express the relationships.

Another type of deductive reasoning problem is the  syllogism . Syllogisms are characterized by two premises and a conclusion containing the words all, no, or some (e.g., All As are Bs; Some As are not Bs; Khemlani & Johnson-Laird,  2012 ). The following are sample premises:

·  All university professors are teachers.

Some graduate students are not teachers.

No undergraduate student is a teacher.

A sample syllogism is as follows:

·  All the students in Ken’s class are good in math.

All students who are good in math will attend college.

(Therefore) All the students in Ken’s class will attend college.

Researchers debate what mental processes people use to solve syllogisms, including whether they use heuristics, rules of inference, or diagrams (e.g., Venn; Khemlani & Johnson-Laird,  2012 ). For example, using inference rules we might believe that a syllogism is true only if there is no way to interpret the premises to imply the opposite of the conclusion; that is, a syllogism is true unless an exception to the conclusion can be found.

In information processing terms, people may learn the rules (e.g., the modus ponens rule governs “if pthen q” statements) and then match instances to the rules. Or individuals may use content-specific rules, which can be expressed as productions such that specific instances trigger the production rules. Thus, a production may involve all cars and may be triggered when a specific car (“my brand X”) is encountered.

Solving syllogisms also might depend on semantic procedures that search for interpretations of the premises that are counterexamples to conclusions. According to this view, people construct one or more mental models for the assertions they encounter (interpretations of the premises); the models differ in structure and are used to test the logic of the situation. Students may repeatedly re-encode the problem based on information; thus, deduction largely is a form of verbal reasoning (Polk & Newell,  1995 ). Johnson-Laird and colleagues (Johnson-Laird,  1999 ; Johnson-Laird, Byrne, & Schaeken,  1992 ; Johnson-Laird, Byrne, & Tabossi,  1989 ) have extended this semantic analysis to various classes of inferences (e.g., those involving if, or, and, not, and multiple quantifiers). Further research will help clarify these processes and determine instructional implications.

Evaluation.

Evaluation involves using criteria to judge the adequacy of a problem solution. In evaluating, students address questions such as, “Are the data sufficient to solve the problem?” “Do I need more information?” and “Are my conclusions based on facts, opinions, or reasoned judgments?” Evaluation also involves deciding what ought to happen next—that is, formulating hypotheses about future events assuming that one’s analysis is correct so far.

Deductive reasoning also can be affected by content apart from the logic. Wason ( 1966 ) put four cards (showing A B 2 3) in front of participants. They were told that each card contained a letter on one side and a number on the other, and they were given a conditional rule: “If a card has A on one side, then it has 2on the other.” Their task was to select the cards that needed to be turned over to determine whether the rule was true. Although most participants picked the A card and many also chose the 2, few picked the 3; however, it must be turned over because if there is an A on the other side, then the rule is false. When the content was changed to an everyday generalization (e.g., letter = hair color, number = eye color, A = blond hair, 2 = blue eyes), most people made the correct selections (Wason & Johnson-Laird,  1972 ). These results speak to the importance of not assuming generalization in reasoning but rather giving students experience working on different types of content.

Metacognition is a central component of reasoning (Thompson, Turner, & Pennycook,  2011 ). Learners monitor their efforts to ensure that questions are properly posed, that data from adequate sources are available and used to draw inferences, and that relevant criteria are employed in evaluation. Teaching reasoning requires instruction in skills and in metacognitive strategies. Cognitive load also seems important ( Chapter 5 ). Reasoning is difficult if multiple sources of information must be processed simultaneously, which taxes WM. Carlson et al. ( 2003 ) found that students’ science performance benefited from two procedures designed to reduce cognitive load: diagrams and instructions that minimized the amount of information to be processed at the same time.

Creativity

Creativity  (or  creative thinking ) is closely aligned with other topics covered in this chapter. The features that distinguish creativity from other cognitive processes involve novelty and value (or appropriateness). Creative thinking involves the development of a novel idea, problem solution, or product that is of value and appropriate for the individual or larger social group (Hennessey & Amabile,  2010 ). Beyond these two criteria, researchers disagree about the necessary or desirable components of creativity.

Like problem solving, creativity deals with generating solutions; however, problem solving does not require that solutions be novel. They may be tried and true methods, just not those thought of previously by the problem solver. Critical thinking deals with outcomes that are valued and appropriate, but critical thinking does not require that one generate solutions, only that one consider an issue more thoroughly.

Creativity is not a single phenomenon; there are different forms. One distinction is between Big C creativity and little c creativity (Hennessey & Amabile,  2010 ). Big C creativity is eminent creativity, or a rare type that produces major breakthroughs and products and that has significant effects on others. While that type is newsworthy and often garners awards for the creator, it is much less common than little c creativity or that which occurs in everyday life and involves problem solving and ways to adapt to situations (e.g., creative ways to plan activities). In line with information processing theory, creativity also is at work in the construction of knowledge and linking knowledge with other knowledge in LTM networks. Regardless of the type, creativity seems dependent on the combining of concepts in new or unusual ways.

A key question is whether students can learn to be more creative. Like other cognitive processes, creativity capabilities can be improved. Teaching divergent thinking (or spontaneous thinking with the goal of generating many different ideas) seems to benefit creativity as opposed to convergent thinking (i.e., more disciplined thinking focused on narrowing possible solutions), and there is some evidence that creativity is enhanced when learners work in groups rather than individually (Hennessey & Amabile,  2010 ).

Creativity also can be affected by motivational factors. Intrinsic motivation (see  Chapter 9 ) facilitates creativity, whereas extrinsic motivation may not. Some research has explored whether rewarding students for creative thinking leads to increases in it. The research literature is not consistent on this issue (Joussemet & Koestner,  1999 ), but what does seem to help is providing students with instructions to think creatively (Hennessey & Amabile,  2010 ).

The Creative Problem Solving (CPS) model is a generic framework (Treffinger,  1985 ; Treffinger & Isaksen,  2005 ). This model comprises three major components: understanding the challenge, generating ideas, and preparing for action (Treffinger,  1995 ; Treffinger & Isaksen,  2005 ). Metacognitive components (e.g., planning, monitoring, modifying behavior) are present throughout the process.

Understanding the challenge begins with a general goal or direction for problem solving. After important data (e.g., facts, opinions, concerns) are obtained, a specific goal or question is formulated. The hallmark of generating ideas is divergent thinking to produce options for attaining the goal. Preparing for action includes examining promising options and searching for sources of assistance and ways to overcome resistance.

Brainstorming  is a general problem-solving strategy that is useful for formulating possible problem solutions (Isaksen & Gaulin,  2005 ; Mayer,  1992 ; Osborn,  1963 ). The steps in brainstorming are as follows:

· ■ Define the problem.

· ■ Generate as many solutions as possible without evaluating them.

· ■ Decide on criteria for judging potential solutions.

· ■ Use these criteria to select the best solution.

Successful brainstorming requires that participants withhold criticism of ideas until after all ideas are generated. In addition, participants may generate ideas that build onto one another. Thus, “wild” and unusual ideas should be encouraged (Mayer,  1992 ).

The amount of knowledge one has about the problem domain affects the success of brainstorming because better domain knowledge allows one to generate more potential solutions and criteria for judging their feasibility. Brainstorming can be used individually, although the group interaction usually leads to more solutions.

Brainstorming lends itself well to many instructional and administrative decisions made in schools. It is most useful for generating many varied—and possibly some unique—ideas (Isaksen & Gaulin,  2005 ). Assume that a new school principal finds low staff morale. Staff members agree that better communication is needed. The grade-level leaders meet with the principal, and the group arrives at the following potential solutions: Hold a weekly meeting with staff, send out a weekly (electronic) bulletin, post notices on a bulletin board, hold weekly meetings with grade-level leaders (after which they meet with teachers), send e-mail informational messages frequently, make announcements over the public address system. The group formulates two criteria: (a) minimally time-consuming for teachers and (b) minimally disrupting to classes. With the criteria in mind, they decide that the principal should send out a weekly bulletin and frequent e-mail messages and meet with grade-level leaders as a group. Although they will take time, meetings between the principal and grade-level leaders will be more focused than those between the principal and the entire staff.

COGNITION AND TECHNOLOGY

The last few years have witnessed a rapid explosion of technology in instruction through electronic and distance learning (Bernard et al.,  2009 ; Brown,  2006 ; Campbell,  2006 ; Clark,  2008 ; Jonassen,  1996 ; Jonassen, Peck, & Wilson,  1999 ; Larreamendy-Joerns & Leinhardt,  2006 ; Roblyer,  2006 ; Winn,  2002 ). Technology often is equated with equipment (e.g., computers), but its meaning is much broader.  Technology  refers to the designs and environments that engage learners (Jonassen et al.,  1999 ). Research on the effects of technology on learning is increasing, as are efforts to remove barriers to infusing technology into instruction (Ertmer,  1999 ).

Technology has the potential to facilitate instruction in ways that formerly were unimaginable. Today’s students can experience simulations of environments and events that they never could in regular classes, receive instruction from and communicate with others at long distances, and interact with large knowledge bases and expert tutoring systems.

A challenge for researchers is to determine how technology affects learners’ cognitive processes during encoding, retention, transfer, problem solving, and so forth. This section focuses on the role that technology plays in learning. This material is not a practical guide on how to use technology in education. Readers interested in in-depth applications of technology should consult other sources (Brown,  2006 ; Kovalchick & Dawson,  2004a ,  2004b ; Roblyer,  2006 ; Seo, Pellegrino, & Engelhard,  2012 ).

Computer-Based Learning Environments

Computer-based learning environments  are becoming increasingly common. Researchers are interested in the roles that computer technologies play in teaching and learning. Although computer-based learning is not a theory, it is important to know whether computers improve learning and help develop complex cognitive processing.

It is tempting to evaluate computer-based learning by comparing it to learning not involving computers, but such comparisons can be misleading because other factors (e.g., authenticity of the content, teacher–student/student–student interactions) also may differ. Rather than focusing on this issue, it is better to examine the types of cognitive processing that can occur in computer-based environments and from other technological applications.

Jonassen et al. ( 1999 ) presented a dynamic perspective on the role of technology in learning. The maximum benefits of technology derive when it energizes and facilitates thinking and knowledge construction. In this conceptualization, technology can serve the functions shown in  Table 7.4 . The technological applications relevant to learning described in this section are differentially effective in accomplishing these functions.

Technology should be used in support of instructional goals, not because it is available and educators believe they should use it. The effectiveness of technology depends on how well it complements instructional goals and practices. Results of a meta-analysis showed that students in classes where technology was used scored about 12 percentage points higher in achievement compared with students in classes where technology was not used to enhance learning (Tamim, Bernard, Borokhovski, Abrami, & Schmid,  2011 ). However, there was much variability between classes, most likely due to how well technology was integrated with instruction.

Computer-Based Instruction.

Until a few years ago when it was supplanted by the Internet,  computer-based instruction  (CBI) (or CAI— computer-assisted instruction ) was the most common application of  computer learning  in schools (Jonassen,  1996 ). CBI is often used for drills and tutorials ( Chapter 3 ), which present information and feedback to students and respond based on students’ answers.

Table 7.4 Functions of technology.

· • Tool to support knowledge construction · • Information vehicle for exploring knowledge to support learning by constructing · • Context to support learning by doing · • Social medium to support learning by conversing · • Intellectual partner to support learning by reflecting (Jonassen et al.,  1999 )

Several CBI features are firmly grounded in learning theory and research. The material can command students’ attention and provide immediate response feedback. Feedback can be of a type not often given in the classroom, such as how students’ present performances compare with their prior performances (to show progress in learning). Computers individualize content and rate of presentation.

Another advantage of CBI is that many programs allow personalization; students enter information about themselves, parents, and friends, which is then included in the instructional presentation. Some evidence suggests that personalization can produce higher achievement than other formats (Anand & Ross,  1987 ). Personalizing instruction may improve meaningfulness and facilitate integration of content into LTM networks. Knowledge construction should be aided with familiar referents.

CBI also can be used for more complex learning through tutoring by  expert systems , or large computer programs that contain the knowledge and cognitive (thinking) processes of experts (Graesser, Conley, & Olney,  2012 ). Expert systems represent an application of  artificial intelligence , which refers to computer programs that simulate human cognitive processes and learning. Such systems can, for example, help students become better self-regulated learners by teaching them how to plan and monitor learning and use effective learning strategies (Schraw,  2010 ), and they also can be used for collaborative problem solving (Järvelä & Hadwin,  2013 ; see  Chapter 10 ). Unlike classic CBI that is answer based (i.e., student enters an answer and computer gives feedback on its correctness), intelligent tutoring systems are process based. Thus, the system may prompt for what method the student wants to use to solve a problem and then may engage in a dialogue with the student regarding the method to use. The system gives hints and feedback on each step in the process. Based on a research review, VanLehn ( 2011 ) found that intelligent tutoring systems were comparable to human tutoring in their effects on student learning. Adding worked examples to intelligent tutors helps to decrease instructional time and promote learning compared with the tutor alone, perhaps because worked examples help to decrease extraneous cognitive load (Salden, Koedinger, Renkl, Aleven, & McLaren,  2010 ).

One common problem is that students may use ineffective methods in CBI with the result being piecemeal learning. Such learning violates the idea that learning should be meaningful and linked with knowledge in LTM. Students who are taught effective study strategies (e.g., organizing, summarizing) show corresponding achievement gains (Jairam & Kiewra,  2010 ).

Simulations and Games.

Simulations  represent real or imaginary situations that cannot be brought into the learning setting. Examples are programs simulating the flights of aircraft, underwater expeditions, and life in a fictional city. Learners can build memory networks better when they have tangible referents during learning.

As a type of computer-based environment, simulations seem well suited for discovery and inquiry learning ( Chapter 8 ). In their review of studies using computer simulations in discovery learning, de Jong and van Joolingen ( 1998 ) concluded that simulations were more effective than traditional instruction in inculcating students’ “deep” (intuitive) cognitive processing.

For simulations to be effective it is essential that they not create excessive cognitive load for learners (see  Chapter 5 ). Splitting the content onto two successive screens rather than displaying all of it on one screen benefits learning and transfer (Lee, Plass, & Homer,  2006 ). Mayrath, Nihalani, and Robinson ( 2011 ) found that a voice tutorial reduces extraneous cognitive load better than a text tutorial and leads to greater transfer. The higher cognitive load of the text may result from learners splitting their visual attention between two sources of information.

Simulations also may be beneficial for developing problem-solving skills. Similar to the results for CBI, Moreno and Mayer ( 2004 ) found that personalized messages from an on-screen agent during simulations improved retention and problem solving better than did nonpersonalized messages. Woodward, Carnine, and Gersten ( 1988 ) found that the addition of computer simulations to structured teaching produced problem-solving gains for special-education high school students compared with traditional instruction alone. The authors noted, however, that the mechanism producing these results was unclear, and the results may not generalize to stand-alone computer simulations.

Games  are designed to create an enjoyable learning context by linking material with sport, adventure, or fantasy. Games can emphasize thinking skills and problem solving but also can be used to teach content (e.g., a basketball game to teach fractions).

Games also may influence learning by increasing motivation. Motivation is greater when an  endogenous (natural) relationship exists between the content and the means (“special effects”) by which the game or simulation presents the content (Lepper & Hodell,  1989 ). Fractions are endogenously related to a basketball game, for example, when students are asked to determine how much of the court is covered by players dribbling down the floor. Such an endogenous relationship enhances meaningfulness and LTM coding and storage. In many games and simulations, however, the relation between content and means is arbitrary, such as when a student’s correct response to a question produces fantasy elements (e.g., cartoon characters). When the relation is arbitrary, the game does not produce better learning than traditional instruction, although the former may be more interesting.

Another issue is that games have many interesting features that potentially can overload learners’ WMs and distract them from learning content. Focusing learners’ attention on the relevant content can promote learning and transfer beyond the learning context. Fiorella and Mayer ( 2012 ) found benefits on learning and transfer from providing students worksheets that directed their attention to relevant features of the game and summarized its underlying principles.

Multimedia.

Multimedia  refers to technology that combines the capabilities of various media such as computers, film, video, sound, music, and text (Roblyer,  2006 ). Multimedia learning occurs when students interact with information presented in more than one mode (e.g., words, pictures, video streaming).

The effectiveness of multimedia for learning depends on learners’ WMs. As information is presented in multiple modalities, the phonological loop (verbal information), visuo-spatial sketchpad (visual and spatial information), episodic buffer (temporary storage of multimodal information), and central executive (monitoring and coordinating functions and interface with LTM) are engaged (Baddeley,  1998 ; Schüler, Scheiter, & van Genuchten,  2011 ; see  Chapter 5 ). As WM can handle only so much information at once, instruction should optimize cognitive demands so that load is not excessive. Effective multimedia learning requires that students select the relevant information, integrate and organize it into a coherent representation in WM, and integrate this representation with existing LTM knowledge (Lee et al.,  2006 ).

Multimedia learning has important implications for teaching because it offers many possibilities for infusing technology into instruction (Roblyer,  2006 ). Research evidence provides some support for the benefits of multimedia for learning. In his review of research studies, Mayer ( 1997 ) found that multimedia enhanced students’ problem solving and transfer; however, effects were strongest for students with little prior knowledge and high spatial ability. Dillon and Gabbard ( 1998 ) also concluded from their review that effects depended in part on ability: Students with lower general ability had the greatest difficulty with multimedia. Learning style was important: Students willing to explore obtained the greatest benefits. Multimedia seems especially advantageous on specific tasks requiring rapid searching through information.

Researchers have investigated the conditions favoring learning from multimedia. When verbal and visual (e.g., narration and animation) information are combined during instruction, students benefit from dual coding (Adesope & Nesbit,  2012 ; Mayer & Johnson,  2008 ). The simultaneous presentation helps learners form connections between words and pictures because they are in WM at the same time (Mayer, Moreno, Boire, & Vagge,  1999 ), although as noted previously the dual modalities have the potential to create excessive cognitive load. Multimedia may facilitate learning better than tailoring media to individual student differences (Reed,  2006 ). By using different media, teachers increase the likelihood that at least one type will be effective for every student, but it is important that the media do not add interesting but irrelevant information (Mayer, Heiser, & Lonn,  2001 ). Some instructional devices that assist multimedia learning are text signals that emphasize the structure of the content and its relationship to other material (Mautone & Mayer,  2001 ); personalized messages (i.e., informal, conversational) that address students and make them feel like participants in the lesson (Kartal,  2010 ; Mayer, Fennell, Farmer, & Campbell,  2004 ; Moreno & Mayer,  2000 ); generating self-explanations for the phenomena presented (Eysink et al.,  2009 ); allowing learners to exercise control over the pace of instruction (Mayer & Chandler,  2001 ); animations that include movement and simulations (Mayer & Moreno,  2003 ); being able to interact with an on-screen speaker (Mayer, Dow, & Mayer,  2003 ); taking a practice test on the material (Johnson & Mayer,  2009 ); and being exposed to a human rather than a machine-generated speaker (Mayer, Sobko, & Mautone,  2003 ).

Maximal benefits of multimedia require that some logistical and administrative issues be addressed. Interactive capabilities are expensive to develop and produce, although they are very effective (Moreno & Mayer,  2007 ). Costs may prohibit many school systems from purchasing components. Interactive video may require additional instruction time because it presents more material and requires greater student time. But interactive multimodal learning environments provide great potential for increasing students’ motivation (Scheiter & Gerjets,  2007 ). The greater amount of learner control that is possible yields better benefits on learning and can foster self-regulation (Azevedo,  2005b ;  Chapter 10 ).

Despite potential issues involving costs and technological skills needed, multimedia and  hypermedia  seem to benefit student learning, and research increasingly is showing that this technology can help to develop students’ self-regulated learning (Azevedo,  2005a ,  2005b ; Azevedo & Cromley,  2004 ; Azevedo, Guthrie, & Seibert,  2004 ). Applications will continue to be developed as the technology advances (Roblyer,  2006 ). Further research is needed on multi-media’s effects on motivation and how to link it with a sequence of acquiring self-regulatory skills (e.g., social influence to self-influence; Zimmerman & Tsikalas,  2005 ;  Chapter 10 ).

E-learning.

E-learning  refers to learning through electronically delivered means. The term often is used to refer to any type of electronic communication (e.g., videoconferencing, e-mail); however, here it is used in the narrower sense of Internet (Web-based) instruction.

The  Internet  (an international collection of computer networks) is a system of shared resources that no one owns. The Internet provides access to other people (users) through e-mail and conferences (chat rooms), files, and the World Wide Web (WWW)—a multicomputer interactive multimedia resource. It also stores information that can be copied for personal use.

The Internet is a wonderful resource for information, but the relevant issue here is its role in learning. On the surface, the Internet has advantages. Web-based instruction provides students with access to more resources in less time than is possible in traditional ways; however, more resources do not automatically mean better learning. The latter is accomplished only if students acquire new skills, such as methods for conducting research on a topic or critical thinking about the accuracy of material on the Web. Building automated prompts into Web-based instruction (e.g., “Now would be a good time to ask yourself if you have collected all the important information;” Kauffman,  2004 , p. 149) increases students’ metacognitive activity and leads to higher achievement (Kauffman, Ge, Xie, & Chen,  2008 ). Further, guiding learners’ Web-based searches results in higher self-efficacy, performance, and satisfaction than does allowing learners to self-guide their searches (Debowski, Wood, & Bandura,  2001 ), which can prove frustrating for some students. Useful are virtual pedagogical agents (e.g., tutors with human-like bodies), which can focus students’ learning and enhance their motivation for learning (Krämer & Bente,  2010 ). Web resources also can promote learning when students take information from the Web and incorporate it into classroom activities (e.g., discovery learning;  Chapter 8 ).

Teachers can assist the development of students’ Internet skills with scaffolding ( Chapter 8 ). Students must be taught search strategies (e.g., ways to use browsers), but teachers also might conduct the initial Web search and provide students with names of helpful websites. Grabe and Grabe ( 1998 ) offer other suggestions. Applications involving technology in classroom instruction are given in  Application 7.7 .

A danger in students using the Internet is that the large array of information available could increase cognitive load on WM, thereby hindering students’ searches. Providing instructional guidance helps to minimize extraneous load (Kalyuga,  2007 ). So much information also can inculcate the belief among learners that everything is important and reliable. Students then may engage in “associative writing” by trying to include too much information in reports and papers. To the extent that e-learning helps teach students the higher-level skills of analysis and synthesis, they will acquire strategies for determining what is important and merging information into a coherent product.

Online Social Media

Online social media  are Internet tools used to collaborate, communicate, and distribute information. Four categories of online social media that have relevance to education are communication, collaboration, multimedia, and virtual world (Seo et al.,  2012 ).

The primary purpose of communication tools (e.g., Facebook, LinkedIn) is to facilitate communication among users. Instructional examples include instructors posting class notes, students completing group homework, and students posting and reviewing online reflections (Seo et al.,  2012 ). The collaboration category includes wikis, blogs, and social bookmarking. A wiki is a platform for group work; students can collaborate on a project and compose and edit it. Blogs include running dialogue between instructors and students on issues or questions. With social bookmarking, students bookmark selected Web pages to create a collection of related pages, or a resource for a specific topic area (Seo et al.,  2012 ). Multimedia tools (e.g., YouTube, Skype) provide materials for students to study before and after classes, tutorials and educational videos, and interactive group projects (Seo et al.,  2012 ). Finally, virtual world ( virtual reality ) media (e.g., Second Life) provide a platform for synchronous learning. Instructors and students can interact with one another without being physically present, and instructors can meet with students during virtual office hours (Seo et al.,  2012 ).

APPLICATION 7.7 Technology and Learning

Technological applications can be applied effectively to help improve student learning. Two high school classes worked together to develop a Civil War computer simulation. The classes drew straws to determine which class would be the Union and which the Confederacy. The students in each class then studied the battles of the Civil War and looked for information about the terrain, the weather at the time of each battle, the number of soldiers involved, and the leadership abilities of the individuals in charge. The students in both classes then simulated the battles on the computer, interacting with each other, using the data, trying to see if they might change the outcome of the original battle. When students made a strategic move, they had to defend and support their move with historical data.

A college professor uses streaming video and the Web to have her students study and reflect on educational psychology principles applied in classrooms. As students observe the video of an elementary class lesson, they stop the video and enter responses to relate educational practices to psychological principles they have been discussing in class. Then students are able to interact with other students and with her to share thoughts on the lesson observed. She also has a fictional classroom set up on a website. She poses questions to her students (e.g., “How might the teacher use authentic assessment in science?”), after which they go to the website, read and reflect, and construct a response that is distributed to her and all other students. Thus, everyone can respond and interact with others.

Ms. Tarkinton uses technology for creative writing with her elementary students. She starts a story on the computer entitled, “The Adventures of Ms. Tarkinton’s Class.” Children have the opportunity to add to the story as often as they wish. At the end of the month, they print the story and read it aloud in class. The computer-based environment provides a unique means for constructing a story collaboratively.

These and other online social media have revolutionized the way that people interact with one another. Our primary interest here is how they might affect learning. Online social media tools have several features that are positively associated with learning. They greatly facilitate the dissemination of information. By presenting knowledge in multiple modes (e.g., in verbal and visual forms) they allow encoding of knowledge in dual formats, which can enhance development of memory networks and subsequent retrieval from LTM ( Chapters 5  and  6 ). By allowing multiple users simultaneously, online social media can promote collaboration, which is emphasized by social cognitive and constructivist theories (see  Chapters 4  and  8 ). Collaboration is essential for peer-assisted learning ( Chapter 8 ). Further, students perceive online social media positively, which means that they have motivation for using them ( Chapter 9 ).

Research on online social media is in its infancy, so at this point it is difficult to validly assess their effects on learning. Kirschner and Karpinski ( 2010 ) found that college students who had Facebook accounts had lower grade point averages and spent fewer hours studying than students who did not have accounts. However, additional research is needed because Facebook primarily is used by students as a social medium rather than for course learning. As social media become better established in educational programs, researchers should be able to reliably evaluate their effects on student learning.

Today’s students use online social media frequently. They tend to be comfortable using these tools and learn new technological applications easily. On this count, therefore, teachers would be wise to tap into this student resource. We must keep in mind that online social media, like other forms of technology, should not be the center of learning but rather complement instructional objectives (Seo et al.,  2012 ). If course objectives suggest that an online social media tool can create an excellent learning environment for students, then teachers have an opportunity to try this application. If nothing else, the introduction of some variety into teaching may prove motivating for students ( Chapter 9 ).

Distance Learning

Distance learning (distance education)  occurs when instruction that originates in one location is transmitted to students at one or more remote sites. Interactive capabilities allow two-way feedback and discussions to become part of the learning experience. Distance learning saves time, effort, and money because instructors and students do not have to make long journeys to classes. Universities, for example, can recruit students from a wide geographical area. There is less concern about the students traveling great distances to attend classes. School districts can conduct in-service programs by transmitting from a central site to all of the schools. Distance learning sacrifices face-to-face contact with instructors, although if two-way interactive video is used the interactions are real-time ( synchronous learning ). In their review of distance education programs, Bernard et al. ( 2004 ) found their effects on student learning and retention comparable to those of traditional instruction. Effects for synchronous instruction favored classroom instruction, whereas distance education was more effective for  asynchronous learning applications (involving lag time).

Another  networking  application is the  electronic bulletin board (conference) . People networked with computers can post messages, but more important for learning can be part of a discussion (chat) group. Participants ask questions and raise issues, as well as respond to the comments of others. A fair amount of research has examined whether such exchanges facilitate writing skill acquisition (Fabos & Young,  1999 ). Whether this asynchronous means of telecommunication exchange promotes learning any better than face-to-face interaction is problematic because much of the research is conflicting or inconclusive (Fabos & Young,  1999 ); however, the review by Bernard et al. ( 2004 ) suggests that distance education may be more effective with asynchronous learning. Telecommunication has the benefit of convenience in that people can respond at any time, not just when they are gathered together. The receptive learning environment may indirectly promote learning.

Being forms of  computer-mediated communication (CMC) , distance learning and computer conferencing greatly expand the possibilities for learning through social interaction. Further research is needed to determine whether personal characteristics of learners and types of instructional content can affect students’ learning and motivation.

Web-based (online) learning is commonly incorporated into traditional instruction as a  blended model  of instruction (i.e., some face-to-face instruction and the rest online). Web-based learning also is useful in conjunction with multimedia projects. In many teacher preparation programs, pre-service teachers use the Web to obtain resources and then selectively incorporate these into multimedia projects as part of lesson designs.

In their review of online courses, Tallent-Runnels et al. ( 2006 ) found that students liked moving at their own pace, students with more computer experience expressed greater satisfaction, and asynchronous communication facilitated in-depth discussions. Distance education that incorporates interactions (student–student, student–teacher, student–content) helps to increase student achievement (Bernard et al.,  2009 ). Other types of interactions (e.g., wikis, blogs) also may be useful. Infusing multimedia presentations into distance education increases its personalization and thus makes it more akin to face-to-face instruction (Larreamendy-Joerns & Leinhardt,  2006 ), which may increase student motivation.

Attempting to compare online with traditional courses is difficult because there are so many differences, one of which is that, to date, online courses tend to enroll more non-traditional and White American students. This demographic will change as online courses become more prevalent, which will permit better assessment of online learning outcomes and environmental characteristics that facilitate learning.

INSTRUCTIONAL APPLICATIONS

Several instructional applications have been given in this chapter for the principles covered. This section describes three additional applications that reflect many of the principles discussed: worked examples, problem solving, and mathematics.

Worked Examples

Worked examples  (see  Chapter 4 ) present step-by-step problem solutions and often include accompanying diagrams. They portray an expert’s problem-solving model for learners to study before they begin to emulate it. Researchers have shown that studying worked examples promotes learning better than does simply solving problems (Atkinson et al.,  2000 ; Wittwer & Renkl,  2010 ).

Worked examples reflect Anderson’s ACT-R theory (Lee & Anderson,  2001 ) and are especially appropriate for complex forms of learning, such as algebra, physics, and geometry (Atkinson et al.,  2000 ; Atkinson, Renkl, & Merrill,  2003 ). Applying the novice–expert model, researchers have found that experts typically focus on deeper (structural) aspects of problems, whereas novices more often deal with surface features. Worked examples seem most beneficial with students in the early stages of skill acquisition; as learners become more proficient, solving problems enhances skills better (Salden et al.,  2010 ).

The applicability of worked examples is seen in the four-stage model of skill acquisition within the ACT-R framework (Anderson, Fincham, & Douglass,  1997 ;  Chapter 5 ). In stage 1 learners use analogies to relate examples to problems to be solved. In stage 2 they develop abstract declarative rules through practice. During stage 3, performance becomes quicker and smoother as aspects of problem solution become automatized. By stage 4 learners have in memory many types of problems and can retrieve the appropriate solution strategy quickly when confronted with a problem. Use of worked examples is best suited for stage 1 and early stage 2 learners. During later stages, people benefit from practice to hone their strategies, although even at advanced stages, studying solutions of experts can be helpful.

A key instructional issue is how to integrate the components of an example, such as diagram, text, and aural information. It is imperative that a worked example not overload the learner’s WM (create excessive cognitive load), which multiple sources of information presented simultaneously can do. Stull and Mayer ( 2007 ) found that providing graphic organizers (similar to worked examples) produced better problem-solving transfer than did allowing learners to construct their own. The latter task may have produced high cognitive load ( Chapter 5 ). Other evidence shows that worked examples can reduce cognitive load (Renkl, Hilbert, & Schworm,  2009 ).

Research supports the prediction that dual presentation facilitates learning better than single-mode presentation (Atkinson et al.,  2000 ; Mayer,  1997 ). This result is consistent with dual-coding theory (Paivio,  1986 ;  Chapter 6 ), with the caveat that too much complexity is not desirable. Similarly, examples intermixed with subgoals help create deep structures and facilitate learning.

A key point is that examples that include multiple presentation modes should be unified so that learners’ attention is not split across nonintegrated sources. Aural and verbal explanations should indicate to which aspect of the example they refer, so learners do not have to search on their own. Subgoals should be clearly labeled and visually isolated in the overall display.

A second instructional issue concerns how examples should be sequenced. Research supports the conclusions that two examples are superior to a single one, that varied examples are better than two of the same type, and that intermixing examples and practice is more effective than a lesson that presents examples followed by practice problems (Atkinson et al.,  2000 ). Gradually fading out worked examples in an instructional sequence is associated with better student transfer of learning (Atkinson et al.,  2003 ).

Chi, Bassok, Lewis, Reimann, and Glaser ( 1989 ) found that students who provided self-explanations while studying examples subsequently achieved at higher levels compared with students who did not self-explain. Presumably the self-explanations helped students understand the deep structure of the problems and thereby encode it more meaningfully. Self-explanation also is a type of rehearsal, and the benefit of rehearsal on learning is well established. Thus, students should be encouraged to self-explain while studying worked examples, such as by verbalizing subgoals.

Table 7.5 Using worked examples in instruction.

· • Present examples in close proximity to problems students will solve. · • Present multiple examples showing different types of problems. · • Present information in different modalities (aural, visual). · • Indicate subgoals in examples. · • Ensure that examples present all information needed to solve problems. · • Teach students to self-explain examples, and encourage self-explanations. · • Allow sufficient practice on problem types so students refine skills.

Another issue is that worked examples can produce passive learning since learners may process them superficially. Including interactive elements, such as by providing prompts or leaving gaps that learners must complete, leads to more active cognitive processing and learning (Atkinson & Renkl,  2007 ). Animations also are helpful (Wouters, Paas, & van Merriënboer,  2008 ).

In summary, there are several features that when incorporated with worked examples help learners create cognitive schemas to facilitate subsequent achievement ( Table 7.5 ). These instructional strategies are best employed during the early stages of skill learning. Through practice, the initial cognitive representations should evolve into the refined schemas that experts employ.

Problem Solving

The links between learning and problem solving suggest that students can learn heuristics and strategies and become better problem solvers (Bruning et al.,  2011 ). In addition, for information to be linked in memory, it is best to integrate problem solving with academic content (as Meg recommended in the opening vignette) rather than to teach problem solving with stand-alone programs. Nokes, Dole, and Hacker ( 2007 ) showed that heuristics instruction can be infused into classroom teaching without sacrificing students’ content learning.

· Andre ( 1986 ) listed several suggestions that are derived from theory and research and that are useful for training students in problem-solving skills, especially as they represent productions in memory.

· ■ Provide students with metaphorical representations. A concrete analogical passage given to students prior to an instructional passage facilitates learning from the target passage.

· ■ Have students verbalize during problem solving. Verbalization of thoughts during problem solving can facilitate problem solutions and learning.

· ■ Use questions. Ask students questions that require them to practice concepts they have learned; many such questions may be necessary.

· ■ Provide examples. Give students worked examples showing application of problem-solving strategies. Students may have difficulty seeing on their own how strategies apply to situations.

· ■ Coordinate ideas. Show how productions and knowledge relate to one another and in what sequence they might need to be applied.

· ■ Use discovery learning. Discovery learning often facilitates transfer and problem solving better than  expository teaching . Discovery may force students to generate rules from examples. The same can be accomplished through expository teaching, but discovery may lend itself better to certain content (e.g., science experiments).

· ■ Give a verbal description. Providing students with a verbal description of the strategy and its rules for application can be helpful.

· ■ Teach learning strategies. Learners may need assistance in using effective learning strategies. As discussed in  Chapter 10 , strategies help learning and problem solving.

· ■ Use small groups. A number of studies have found that small-group learning helps develop students’ problem-solving skills. Group members must be held accountable for their learning, and all students must share in the work.

· ■ Maintain a positive psychological climate. Psychological factors are important to effective problem solving. Minimize excessive anxiety among students and help to create a sense of self-efficacy among students for improving their skills ( Chapter 4 ).

Another instructional suggestion is to phase in problem solving, which may be especially helpful with students who have little experience with it. This can be done by using worked examples (Atkinson et al.,  2003 ; Renkl & Atkinson,  2003 ; discussed in this section). Mathematics texts, for example, often state a rule or theorem, followed by one or more worked examples. Students then solve comparable problems by applying the steps from the worked examples (a type of analogical reasoning). Renkl and Atkinson recommended reliance on examples in the early stages of learning, followed by a transition to problem solving as students develop skills. This process also helps to minimize demands on WM, or the cognitive load that learners experience ( Chapter 5 ). Thus, the transition might proceed as follows. Initially a complete example is given, then an example where one step is omitted. With each succeeding example an additional step is omitted until the learners reach independent problem solving.

Problem-based learning (PBL; Hmelo-Silver,  2004 ) offers another instructional application. In this approach, students work in groups on a problem that does not have one correct answer. Students identify what they need to know to solve the problem. Teachers act as facilitators by providing assistance but not answers. PBL has been shown to be effective in teaching problem-solving and self-regulation skills, but most research has been conducted in medical and gifted education (Evenson, Salisbury-Glennon, & Glenn,  2001 ; Hmelo-Silver,  2004 ). PBL is useful for the exploration of meaningful problems. Because it is time consuming, teachers need to consider its appropriateness given the instructional goals.

Mathematics

Mathematics has been a fertile area of cognitive and constructivist research (Ball, Lubienski, & Mewborn,  2001 ; Carr,  2012 ; National Research Council,  2000 ; Newcombe et al.,  2009 ; Schoenfeld,  2006 ). Researchers have explored how learners construct knowledge, how experts and novices differ, the role of motivation, and which methods of instruction are most effective (Mayer,  1999 ; Schoenfeld,  2006 ). Growth in mathematics achievement depends on not only cognitive but also motivational variables such as perceived control, self-efficacy, and intrinsic motivation (Murayama, Pekrun, Lichtenfeld, & vom Hofe,  2013 ; Schunk & Richardson,  2011 ).

A distinction typically is made between mathematical computation (use of rules, procedures, and algorithms) and  concepts  (problem solving and use of strategies). Computational and conceptual problems require students to implement productions involving rules and algorithms. The difference between these two categories lies in how explicitly the problem tells students which operations to perform. The following are computational problems.

· ■ 26 + 42 = ?

· ■ 5x + 3y = 19

· ■ 7x − y = 11

Solve for x and y.

· ■ What is the length of the hypotenuse of a right triangle with sides equal to 3 and 4 inches?

Although students are not explicitly told what to do in problems 2 and 3, recognition of the problem format and knowledge of procedures lead them to perform the correct operations.

Now contrast those problems with the following:

· ■ Alex has 20 coins composed of dimes and quarters. If the quarters were dimes and the dimes were quarters, he would have 90 cents more than he has now. How much money does Alex have?

· ■ If a passenger train takes twice as long to pass a freight train, after it first overtakes the freight train, as it takes the two trains to pass when going in opposite directions, how many times faster than the freight train is the passenger train?

· ■ When she hikes, Shana can average 2 mph going uphill and 6 mph going downhill. If she goes uphill and downhill and spends no time at the summit, what will be her average speed for an entire trip?

These word problems do not explicitly tell students what to do, but they require computations no more difficult than those needed in the first set. Solving word problems involves recognizing their problem formats, generating appropriate productions, and performing the computations.

This is not to suggest that conceptual expertise is better than computational proficiency, although Rittle-Johnson and Alibali ( 1999 ) found that conceptual understanding had a greater influence on procedural knowledge than did the reverse. Deficiencies in either area cause problems. Understanding how to solve a problem but not being able to perform the computations results in incorrect answers, as does being computationally proficient but not being able to conceptualize problems.

Computational Problems.

The earliest computational skill children use is counting (Resnick,  1985 ). Children count objects on their fingers and in their heads using a strategy. The  sum model  involves setting a hypothetical counter at zero, counting in the first addend in increments of one, and then counting in the second addend to arrive at the answer. For the problem “2 + 4 = ?” children might count from 0 to 2 and then count out 4 more. A more efficient strategy is to set the counter at the first addend (2) and then count in the second addend (4) in increments of one. Still more efficient is the  min model : Set the counter at the larger of the two addends (4) and then count in the smaller addend (2) in increments of one (Romberg & Carpenter,  1986 ).

These types of invented procedures are successful. Children and adults often construct procedures to solve mathematical problems. Errors generally are not random but rather reflect  buggy algorithms , or systematic mistakes in thinking and reasoning (Brown & Burton,  1978 ). Buggy algorithms reflect the constructivist assumption that students form procedures based on their interpretation of experiences ( Chapter 8 ). A common mistake in subtraction is to subtract the smaller number from the larger number in each column, regardless of direction, as follows:

53	602
−27	−274
34	472

Mathematical bugs probably develop when students encounter new problems and incorrectly generalize productions. In subtraction without regrouping, for example, students subtract the smaller number from the larger one column by column. It is easy to see how they could generalize this procedure to problems requiring regrouping. Buggy algorithms are durable and can instill in students a false sense of self-efficacy ( Chapter 4 ), perhaps because their computations produce answers.

Another source of computational difficulties is poor declarative knowledge of number facts. Many children do not know basic facts, have difficulty retrieving facts, and show deficiencies in numerical processing (Geary,  2011 ; Geary, Hoard, Byrd-Craven, Nugent, & Numtee,  2007 ). Until facts become established in LTM through practice, children count or compute answers. Speed of fact retrieval from memory relates directly to overall mathematical achievement in students from elementary school through college (Royer, Tronsky, Chan, Jackson, & Marchant,  1999 ). Computational skill improves with development, along with WM and LTM capabilities (Mabbott & Bisanz,  2003 ). Effective functioning of the central executive of WM (see  Chapter 5 ) predicts mathematical achievement (Geary,  2011 ). Computational problem solving also improves when students use written rather than mental calculations, especially on complex problems (Hickendorff, van Putten, Verhelst, & Heiser,  2010 ).

Many difficulties in computation result from using overly complex but technically correct productions to solve problems. Such procedures produce correct answers, but because they are complex, the risk of computational errors is high. The problem 256 divided by 5 can be solved by the division algorithm or by successively subtracting 5 from 256 and counting the number of subtractions. The latter procedure is technically correct but inefficient and has a high probability of error.

Learners initially represent computational skill as declarative knowledge in a propositional network. Facts concerning the different steps (e.g., in the algorithm) are committed to memory through mental rehearsal and overt practice. The production that guides performance at this stage is general; for example: “If the goal is to solve this division problem, then apply the method the teacher taught us.” With added practice, the declarative representation changes into a domain-specific procedural representation and eventually becomes automated. Early counting strategies are replaced with more efficient rule-based strategies (Hopkins & Lawson,  2002 ). At the automatic stage, learners quickly recognize the problem pattern (e.g., division problem, square root problem) and implement the procedure without much conscious deliberation.

Conceptual Problem Solving.

Conceptual problem solving requires students to accurately represent the problem including the given information and the goal and then select and apply a strategy (Mayer,  1985 ,  1999 ). Translating a problem from its linguistic representation to a mental representation is often difficult (Bruning et al.,  2011 ). The more abstract the language, the more difficult the text comprehension and the lower the likelihood of solution (Cummins, Kintsch, Reusser, & Weimer,  1988 ). Students who have difficulty comprehending show poorer recall of information and lower performance. Younger children often have difficulty translating abstract linguistic representations.

Translation also requires good declarative and procedural knowledge. Solving the earlier problem about Alex with 20 coins requires knowledge that dimes and quarters are coins, that a dime is one-tenth ($0.10) of $1, and that a quarter is one-fourth ($0.25) of $1. This declarative knowledge needs to be coupled with procedural understanding that dimes and quarters are variables such that the number of dimes plus the number of quarters equals 20.

One reason experts translate problems better is that their knowledge is better organized in LTM; the organization reflects the underlying structure of the subject matter (Romberg & Carpenter,  1986 ). Experts overlook surface features of a problem and analyze it in terms of the operations required for solution. Novices are swayed more by surface features. Silver ( 1981 ) found that good problem solvers organized problems according to the process required for solution, whereas poor problem solvers were more likely to group problems with similar content (e.g., money, trains).

Novices often adopt a working backward strategy, beginning with the goal and working their way back to the givens. This is a good heuristic useful in the early stages of learning when learners have acquired some domain knowledge but are not competent enough to recognize problem formats quickly, but experts often work forward. They identify the problem type and select the appropriate production to solve the problem. Hegarty, Mayer, and Monk ( 1995 ) found that successful problem solvers translated the problem into a mental model in which the numbers in the problem statement were tied to their variable names. In contrast, less successful solvers were more likely to combine the numbers in the problem with the arithmetic operations primed by the key words (e.g., addition is the operation linked with the key word “more”). The latter strategy is based on surface features, whereas the former strategy is linked better with meanings.

Experts develop sophisticated procedural knowledge for classifying mathematical problems according to type. High school algebra problems fall into roughly 20 general categories, such as motion, current, coins, and interest/investment (Mayer,  1992 ). These categories can be aggregated into six major groups. For example, the amount-per-time group includes motion, current, and work problems. These problems are solvable with the general formula: amount = rate × time. The development of mathematical problem-solving expertise depends on classifying a problem into the correct group and then applying the strategy.

APPLICATION 7.8 Mathematical Problem Solving

· Teachers use various ways to help students improve their skills on conceptual problems. As students solve mathematical word problems, they can state each problem in their own words, draw a sketch, decide what information is relevant, and state the ways they might solve the problem. A middle school teacher could use these and other similar questions to help focus her students’ attention on important task aspects and guide their thinking:

· ■ What information is important?

· ■ What information is missing?

· ■ Which formulas are necessary?

· ■ What is the first thing to do?

Verbalizing steps in problem solving aids the development of proficiency (Gersten et al.,  2009 ). Fyfe, Rittle-Johnson, and DeCaro ( 2012 ) found that providing feedback on strategies and outcomes to learners as they engaged in exploratory problem solving prior to instruction promoted achievement but only for students who had low knowledge of domain-specific strategies. The feedback may have helped them identify errors and look for alternative strategies to use.  Application 7.8  discusses teaching problem solving.

SUMMARY

Complex cognitive processes are involved in much human learning. Developing competence in an academic domain requires knowledge of the facts, principles, and concepts of that domain, coupled with general strategies that can be applied across domains and specific strategies that pertain to each domain. Research has identified many differences between expert and novice performers in a given domain.

Metacognition is the deliberate, conscious control of mental activities. Metacognition includes knowledge and monitoring activities designed to ensure that tasks are completed successfully. Conditional knowledge, or knowing when and why to employ declarative and procedural knowledge, is part of metacognitive activity. Metacognition begins to develop around ages 5 to 7 and continues throughout schooling. One’s meta-cognitive awareness depends on task, strategy, and learner variables. Learners benefit from instruction on metacognitive activities.

Concept learning involves higher-order processes of forming mental representations of critical attributes of categories. Current theories emphasize analyzing features and forming hypotheses about concepts (feature analysis), as well as forming generalized images of concepts that include only some defining features (prototypes). Prototypes may be used to classify typical instances of concepts, and feature analysis may be used for less typical ones. Models of concept acquisition and teaching have been proposed, and motivational processes also are involved in conceptual change.

Problem solving consists of an initial state, a goal, subgoals, and operations performed to attain the goal and subgoals. Researchers have examined the mental processes of learners engaged in problem solving and the differences between expert and novice performers. Problem solving may occur through trial and error, insight, and using heuristics. These general approaches can be applied to academic content. From an information processing perspective, problem solving requires forming a mental representation of the problem and applying sets of rules (production systems) to solve it. With well-defined problems where potential solutions can be ordered in likelihood, a generate-and-test strategy is useful. For difficult or less well-defined problems, means–ends analysis (working backward or forward) may be used, or analogical reasoning.

Critical thinking, reasoning, and creativity are related but distinct cognitive processes. Critical thinking is used to develop better understanding of problems or issues. Reasoning involves generating and evaluating logical arguments. It requires that learners work through problems to determine why something happened or what might happen. Creativity yields products or outcomes that are novel and valued by the larger community. Brainstorming—especially in groups—may help foster creative thinking.

Technology continues to increase in importance in learning and instruction. Three areas that have seen rapid growth are computer-based learning environments, online social media, and distance learning. Applications involving computer-based environments include computer-based instruction, games and simulations, multimedia, and e-learning. Online social media facilitate communication and collaboration among learners, as well as other benefits that may enhance student learning. Distance learning may involve two-way feedback and synchronous discussions, or online (Web-based) asynchronous instruction. Many courses utilize a blended model (some face-to-face and some online instruction). Research shows the benefits of technology on metacognition, deep processing, and problem solving.

Applications include worked examples, problem solving, and mathematics. Worked examples present problem solutions in step-by-step fashion and often include accompanying diagrams. Worked examples incorporate many features that facilitate learners’ problem solving. Problem-solving instruction is more effective when it is clearly linked with academic content. Other suggestions include giving learners worked examples, providing verbal descriptions, and using small group learning. Children display early mathematical competence with counting. Computational skills require algorithms and declarative knowledge. Students often overgeneralize procedures (buggy algorithms). With conceptual problems, students acquire knowledge of problem types through experience and apply increasingly more effective strategies.