Research Formats
Chapter 2
Theories in Research and Practice
Concepts and Operational Definitions among Minority Populations
Cause-and-Effect Relationships
Class Exercises for Competency Assessment
After dashing through the Looking-glass House to view its garden,
Alice says, I should see the garden far better … if I could get to the top of that hill: and here's a path that leads straight to it—at least, no, it doesn't do that … but I suppose it will at last. But how curiously it twists! It's more like a corkscrew than a path! Well this turn goes to the hill, I suppose—no it doesn't! This goes straight back to the house! Well then, I'll try it the other way. (Carroll, 1946, pp. 21–22)
Understanding the world—especially human behavior—sometimes bears a striking resemblance to Alice's convoluted and frustrating journey in Wonderland. People do what we least expect, and without any apparent rhyme or reason: A prisoner on parole who appeared to be “making it on the outside” suddenly commits another offense and goes back to jail; a marriage of 25 years that seemed to be quite solid suddenly ends in divorce; a respected and successful business executive commits suicide. Human service providers, in particular, are familiar with experiences such as these, and the path to understanding often mirrors Alice's corkscrew.
Science, however, provides a method for mapping and understanding that corkscrew. In this chapter, we discuss the basic logic underlying scientific research, beginning with an assessment of how science differs from other ways of gaining knowledge. Then, we analyze the importance of theories and their role in scientific research, drawing a parallel with the use of theories in human service practice. Following this, we discuss the role of concepts and hypotheses, showing how hypotheses serve to link theory and research. Finally, we analyze the nature of causality, because research is, at its core, a search for cause-and-effect relationships among phenomena.
2.1 SOURCES OF KNOWLEDGE
Human service practice is based on knowledge of human behavior and the social environment. There are numerous ways of gaining such knowledge, but all sources of knowledge have their pitfalls. We argued in Chapter 1 that practice knowledge should be grounded in scientific research. This does not mean that science is infallible, but science does have advantages as a source of knowledge that makes it superior to other ways of gaining knowledge.
To see why this is the case, we contrast science with four other common sources of knowledge: tradition, experience, common sense, and journalism. We then discuss how science can improve professional practice.
Traditional knowledge is knowledge based on custom, habit, and repetition. It is founded on a belief in the sanctity of ancient wisdom and the ways of our forebears. People familiar with the musical Fiddler on the Roof will recall how the delightful character Tevye, a dairyman in the village of Anatevka, sang the praises of tradition:
Because of our traditions, we've kept our balance for many, many years. Here in Anatevka we have traditions for everything—how to eat, how to sleep, how to wear clothes …. You may ask, how did this tradition start? I'll tell you—I don't know! But it's a tradition. Because of our traditions, everyone knows who he is and what God expects him to do. Tradition. Without our traditions, our lives would be as shaky as—as a fiddler on the roof! (Stein, 1964, pp. 1, 6)
For Tevye and the villagers of Anatevka, where traditions come from is unimportant. Traditions provide guidance; they offer “truth”; they are the final word. Traditions tell us that something is correct because it has always been done that way.
Traditional knowledge is widespread in all societies. Many people, for example, believe that a two-parent family is preferable to a single-parent family because the former provides a more stable and effective socializing experience for children and reduces the likelihood of maladjustment. In some cases these beliefs are grounded in religious traditions, whereas in other cases they are accepted because “everybody knows” how important two parents are to a child's development. In fact, some human service providers accept these beliefs about the traditional two-parent family despite the existence of considerable research suggesting that the two-parent family may not always be essential for high-quality adoption or foster care. For example, one review of research into this issue concluded, “In the studies reviewed here, single-parent families were found to be as nurturing and viable as dual-parent families. In fact, single-parent adoption emerged as a good plan for children” (Groze 1991, p. 326).
Human service providers can be affected in other ways by traditional beliefs. For example, the works of a Sigmund Freud or an Erik Erikson might be accepted without question, and emphasis might be placed on remaining true to their words rather than on assessing the accuracy or utility of their ideas.
Tradition can be an important source of knowledge, especially in such areas as moral judgments or value decisions, but it can have some major disadvantages. First, tradition is extremely resistant to change, even for those cases in which change might be necessary because new information surfaces or new developments occur. Second, traditional knowledge easily confuses knowledge (an understanding of what is) with values (a preference for what ought to be). For many people, the traditional emphasis on the two-parent family actually is based on a value regarding the preferred family form rather than on a knowledge of the effect that such a family has on child development.
Experience as a source of knowledge refers to firsthand, personal observations of events. Experiential knowledge is based on the assumption that truth and understanding can be achieved through personal experience, and that witnessing events will lead to an accurate comprehension of those events.
Experience is a common source of knowledge for human service workers, who have numerous opportunities to make firsthand observations of emotionally disturbed children, people with physical disabilities, foster children, and other service populations. From these contacts, practitioners can develop an understanding—not necessarily an accurate one—of what motivates their clients and what social or psychological processes have influenced them.
For example, a person working in a spousal abuse shelter will have considerable contact with women whose husbands have physically and psychologically abused them. Because of this, the worker likely is sensitive to the harm that can come to women from their husbands. After seeing women who have been so abused, this worker may conclude that marital counseling with such spouses cannot work in a climate of violence and anger, and may even be dangerous. In fact, social worker Liane Davis (1984) found that shelter workers were much less likely to recommend marital counseling than were family court judges. Family court judges did not have the powerful experience of seeing women when the effects of their abuse were most visible; moreover, family court judges have a mandate to maintain the integrity of the family. For them, marital counseling seems to be both a feasible and an appropriate way to keep the family intact. So we see that the experiences of shelter workers and judges in different settings can lead them to perceive problems and assess solutions differently.
This experiential knowledge about family dynamics and abuse may be reinforced by traditional knowledge about the importance of family life. Armed with this knowledge, a practitioner might shape an intervention effort that focuses on individual counseling or on marital counseling.
Experiential knowledge, however, has some severe limitations that can lead to erroneous conclusions. First, human perceptions are notoriously unreliable. Perception is affected by many factors, including the cultural background and the mood of the observer, the conditions under which something is observed, and the nature of what is being observed. Even under the best conditions, some misperception is likely; thus, knowledge based on experience often is inaccurate.
Second, human knowledge and understanding do not result from direct perception but, rather, from inferences that are made from those perceptions. The conclusion that marital counseling doesn't work is an inference—that is, it is not directly observed. All that has been observed is that these women have been battered by their husbands. There is no observation of the effectiveness of any type of counseling. (We discuss making inferences from observations in more detail when we address the issue of causality later in this chapter.)
Third, the very people in positions to experience something directly frequently have vested interests in perceiving that thing in a certain way. Teachers, for example, observe that the students who do poorly are the ones who do not pay strict attention during class. However, teachers have a vested interest in showing that their teaching techniques are not the reason for poor performance among students. Therefore, teachers probably would be inclined to attribute students' failings to the students' lack of effort and inattentiveness rather than to their own inadequacies as educators.
A final limitation on experiential knowledge is that it is difficult to know if the people directly available to you are accurate representatives of all the people about whom you wish to draw conclusions. If they are not, then any conclusions drawn from your observations may be in error. To use our earlier example, are the battered women who contact a spousal abuse shelter representative of all battered women? If the women who contact a shelter are different in some way, and if these differences influence the effectiveness of counseling in the shelter, then you cannot generalize conclusions from their outcomes in counseling to the experiences of all battered women. Battered women who go to a shelter may be more affluent or less isolated and, therefore, might demonstrate different outcomes in counseling than less affluent or more isolated women would.
The accumulation of knowledge from tradition and experience often blends to form what people call common sense: practical judgments based on the experiences, wisdom, and prejudices of a people. People with common sense are presumed to be able to make sound decisions even though they lack specialized training and knowledge. Yet, is common sense an accurate source of knowledge? Consider the following contradictory examples. Common sense tells us that people with similar interests and inclinations will be likely to associate with one another. When we see a youngster who smokes marijuana associating with others who do the same, we may sagely comment, “Birds of a feather flock together.” Then, however, we see an athletic woman become involved with a bookish, cerebral man, and we say, “Opposites attract.”
In other words, common sense often explains everything—even when those explanations contradict one another. This is not to say that common sense is unimportant or always useless. Common sense can be valuable and accurate, which is not surprising, because people need sound information as a basis for interacting with others and functioning in society. However, common sense does not normally involve a rigorous and systematic attempt to distinguish reality from fiction. Rather, it tends to accept what “everyone knows” to be true and to reject contradictory information. Furthermore, common sense often is considered to be something that people either have or don't have, because it is not teachable. In fact, it often is contrasted with “book learning.” This discourages people from critically assessing their commonsense knowledge and tempering it with knowledge acquired from other sources. For this reason, commonsense knowledge should be accepted and used cautiously. As a basis for human service practice, knowledge needs to be based on the rigorous and systematic methods used in scientific research. Common sense or a vague feeling of “helping” is not enough.
The materials prepared by journalists for newspapers, magazines, television, websites, or other media are another important source of knowledge about the world for most people. With the explosion of news sources available on cable television and the Internet, people now have access to vast amounts of journalistic information. Though some journalism consists of opinion pieces based on the speculations and inferences of the journalist, much of it, like science, is grounded in observation: Reporters interview people or observe events and write their reports based on those observations. In addition, with modern technology, journalists often are in a position to provide a video and/or audio record of what happened at a scene.
So it may seem, at first glance, that science and journalism have much in common as sources of knowledge, and significant similarities between the two endeavors can be identified. Both use observation to seek out accurate knowledge about the world. In fact, some journalism can, at times, take on many of the characteristics of social science research. Some journalistic output, for example, can look a lot like the in-depth interviews and case studies we will discuss in Chapter 9. However, although scientific standards require that scientists use the systematic procedures discussed in this book, journalism can—and often does—fall far short of meeting these standards.
A key difference between science and journalism is that the observations of scientists are much more systematic in nature. This means that scientists utilize far more careful procedures than journalists to reduce the chances that their conclusions will be inaccurate. For example, a journalist interested in the experiences of prison inmates probably will interview a few inmates who are made available to him by prison authorities and then use these interviews to draw conclusions, at least implicitly, about the experiences of all prisoners. Social scientists would recognize that prisoners selected by the authorities are likely to differ from other prisoners in some important ways: They may have been selected because they committed less serious offenses or were model prisoners. Their experience of prison also is likely to be very different from that of a more serious offender or someone who has chronic confrontations with prison authorities. Recognizing this, social scientists would be very careful about how they selected inmates on whom to make observations, and usually would not accept a sample selected by prison authorities. The best sampling procedures, which will be discussed in Chapter 6, would be those that ensure that all types of prisoners have a chance to appear in the sample. This could be done, for example, by interviewing all the prisoners or, if that were not feasible, interviewing a randomly selected group of prisoners. If sampling procedures fall short of these standards, then social scientists have reduced confidence in the resulting conclusions. Journalists often do not use such rigorous sampling procedures.
A second key difference between science and journalism is that journalism is not concerned with theory building and verification as a way of developing an abstract explanation of people's behavior. Journalists are much more focused on, as the saying goes, “just the facts.” Scientists, on the other hand, recognize that facts often don't speak for themselves—they need to be interpreted in the context of a theoretical understanding to fully comprehend their meaning.
Winston Churchill, the prime minister of Britain during World War II, is reported to have said that democracy is an imperfect form of government, but that it is far superior to all other forms. Many scientists have a similar view of science: They realize it is imperfect and limited, but they also recognize that it is far superior to other sources of knowledge for gaining an understanding of the world. Science is a method of obtaining objective knowledge about the world through systematic observation. (The term science also sometimes is used to refer to the accumulated body of knowledge that results from scientific inquiry.) Science has five distinguishing characteristics that, taken together, set it apart from the other sources of knowledge.
First, science is empirical, which simply means that science is based on direct observation of the world. Science is not, as some people mistakenly believe, founded in theorizing, philosophizing, or speculating. At times, scientists do all these things; but eventually, they must observe the world to see whether their theories or speculations agree with the facts. Because of this, the topics that can be subjected to scientific scrutiny are limited; any issue that cannot be resolved through observation is not within the scope of science. For example, the questions of whether God exists or what values should underlie a human service profession are not scientific issues, because it is impossible to determine their truth (or lack thereof) through observation. These are matters of faith or preference, not of science.
Second, science is systematic, meaning that the procedures used by scientists are organized, methodical, public, and recognized by other scientists. One dimension of the systematic nature of science is that scientists report, in detail, all the procedures used in coming to a conclusion. This enables other scientists to assess whether the inferences and conclusions drawn are warranted given the observations made. A second dimension of the systematic nature of science is replication—that is, repeating studies numerous times to determine if the same results can be obtained. Scientists are very cautious about drawing hard-and-fast conclusions from a single observation or investigation. In fact, and quite at variance with experiential knowledge, scientists assume that a single direct observation is as likely to be incorrect as correct. Only repeated observations can reduce the chance of error and misinterpretation (Rosenthal, 1991).
Third, science is the search for causes. Scientists assume that there is order in the universe, that there are ascertainable reasons for the occurrence of all events, and that scientists can discover the orderly nature of the world. If we assumed there was no order, no pattern, then there would be no need to search for it. We could write off events as the result of chance or the intervention of some benevolent (or malevolent, or indifferent) otherworldly force that we can never understand.
Fourth, science is provisional, which means that scientific conclusions are always accepted as tentative, as subject to question and possible refutation. There are no ultimate, untouchable, irrevocable truths in science. There are no scientists whose work is held in such esteem that it cannot be criticized or rejected. As the philosopher Jacob Bronowski (1978, pp. 121–122) put it, “Science is not a finished enterprise…. The truth is [not] a thing that you could find … the way you could find your hat or your umbrella.” Science is a process of continuous movement toward a more accurate picture of the world, and scientists fully realize that we will never achieve the ultimate and final picture.
Finally, science strives for objectivity, which means that scientists try to avoid having their personal biases and values influence their scientific conclusions. This is a controversial and complicated characteristic of science, because many social scientists would argue that true objectivity is impossible for human beings to achieve. We will discuss this issue at a number of points in this book, but it is sufficient to say here that all scientists are concerned that their scientific conclusions are not solely (or merely) a product of their own personal biases and values. This doesn't mean that scientists should be devoid of values. Quite the contrary, they can be as passionate, concerned, and involved as any other group of citizens. However, they realize that their values and biases can—and probably will—lead to erroneous scientific conclusions. To address this problem, science incorporates mechanisms to reduce the likelihood of biased observations becoming an accepted part of the body of scientific knowledge. For example, publicizing all research procedures enables others to assess whether the research was conducted in a way that justifies the conclusions reached. Furthermore, such detailed reporting permits replication so that other researchers, with different values, can see if they come to the same conclusions regarding a set of observations.
Despite these checks, of course, values and biases will still be found in research. The very decision of what topics to investigate, for example, often is shaped by the researcher's values: One person studies family violence because a close friend was the victim of spousal abuse, and another studies factors contributing to job satisfaction because of a personal belief that work is central to identity. Values and biases also enter research through the interpretation of observations. For personal reasons, one researcher may desperately want to show that the criminal justice system rehabilitates (or does not rehabilitate). This may well influence how he or she goes about conducting research and interpreting the results. There are even a few cases (most commonly in biomedical research) of outright falsification of data to show a certain conclusion. The point is that values and biases commonly intrude on scientific research, but the overall scientific enterprise is organized to reduce their impact on the body of scientific knowledge.
The scientific method, then, with the characteristics just described, is viewed by scientists as preferable to other ways of gaining knowledge, because it is more likely to lead to accurate knowledge of the world. To return to our earlier example of single-parent families and adoption, science views all knowledge regarding the family as provisional and open to question, and there have been many scientific investigations of the role of the family in these matters. Child adjustment and development in two-parent and single-parent families have been compared in both adoptive and biological families as well as in stepfamilies.
The conclusion from these various studies is that the traditional, two-parent family does not seem to play the indispensable role that much commonsense knowledge would accord it—or at least that the role of parents in families is more complicated than was once thought. For example, the research shows that adoption by a single parent is not always detrimental when compared with two-parent adoptive settings; some adoptees with one parent do quite well (Groze & Rosenthal, 1991; Shireman & Johnson, 1986). A related finding is that nontraditional family structure (having one rather than two parents) probably is less critical to children's development than is family process (warm relationships and low conflict between parents and children). The negative consequences often associated with single-parent families may arise from the conflict that often accompanies divorce, because divorce is how many single-parent families are produced (Amato & Booth, 1997; Jekielek, 1998; Lansford et al., 2001). Also, single-parent families often experience certain negative factors: low income, inadequate parental guidance, and less access to community resources. Single-parent families that overcome these difficulties do as well as other family forms in raising children (McLanahan & Sandefur, 1994). So the commonsense or traditional view that the two-parent family is always superior to the single-parent setting is shown by research to be vastly oversimplified at best.
In this fashion, then, scientific knowledge overcomes many of the weaknesses of traditional, experiential, and commonsense knowledge. In particular, it enables us to accumulate accurate information despite the personal biases of individual researchers or practitioners. These positive attributes of science do not mean, however, that science is perfect. Scientists do make errors. But, as Jacob Bronowski (1978, p. 122) so aptly put it, “Science is essentially a self-correcting activity.” If proper scientific procedures are followed, today's errors will be corrected by researchers in the future, whose errors in turn will be corrected by still more research.
With this more detailed understanding of science and its characteristics, we can return to the concept of evidence-based practice introduced in Chapter 1 and elaborate on what this approach says about the parallels between science and practice. Basically, this approach argues that effective human service practice should have characteristics that parallel those of science (Hayes, Barlow, and Nelson-Gray, 1999; Rosen, 1996). Human service practice would be enhanced if it was shaped by a consideration of each of the five characteristics of science just discussed. First, practice, like science, should be empirical, stressing problem assessment involving direct observation of client problems, actual counts of behaviors, and independent observations from multiple data sources. Such data are less subject to distortion and bias than self-reports, speculations, and philosophizing. Second, practice, like research, should be systematic. To the extent that practice procedures are well organized, clearly specified, and made public, they can be replicated and tested by others. In this manner, ineffective procedures can be eliminated, and promising ones can be refined and improved. One of the recurring criticisms of many human service interventions is that the intervention itself is not well specified. Consequently, research evaluating the intervention cannot clearly indicate what did (or did not) work. Practice models also involve causality in terms of specifying a clear link between cause and effect or explaining why a proposed intervention should work with the particular identified problem. Again, a criticism of human service projects in the past has been that many of them have consisted of a conglomeration of intervention efforts without a clearly articulated linkage between cause and effect.
Practice theory, like science, should be provisional. All practice models and techniques should be viewed as fair game for criticism and refutation. Through such a process of testing and challenging existing practices, healthy growth can occur in practice methodology.
Finally, human service professionals must deal with the problem of professional objectivity. The determination of the utility and effectiveness of practice procedures needs to be done under objective conditions. Just as the researcher must attempt to safeguard against the intrusion of values into the conduct of research, so practitioners must guard against the intrusion of values into practice. This issue of values and objectivity is particularly difficult for human service practitioners, who often approach problems with a strong set of values, both personal and professional. Personally, practitioners may have strong feelings about such matters as abortion, alcohol use, or domestic violence that may clash with those of the groups with which they work. Furthermore, some human service providers are conventional and middle class in their personal lives, which may influence what they see as successful social functioning. In addition to these personal values, human service practice itself is heavily imbued with professional values. In fact, as one human service educator put it, “Social work is among the most value based of all professions.…. [It] is deeply rooted in a fundamental set of values that ultimately shapes the profession's mission and its practitioners' priorities” (Reamer, 2006, p. 3).
At times, these values may emphasize a conservatism or pressure to preserve the status quo; at other times, the values may reflect a commitment to support vulnerable or oppressed populations. In either event, it is challenging to satisfy the recommendation that human service providers not let their values intrude into the provision of services to clients. In fact, it probably is impossible to mount an effective change effort without some imposition of values, either implicit or explicit. Even more, some therapeutic approaches, such as those of Carl Rogers, Albert Ellis, and Hobart Mowrer, include as one of their goals the acceptance by the client of new—and more realistic—values.
So even though professional practice in the human services is clearly oriented toward the fulfillment of certain values, practicing in the profession requires that the worker establish checks on the intrusion of values into practice, much as the researcher does in the conduct of research. In later chapters, we will discuss research techniques that are less subject to biases in observation and measurement. Application of these principles in practice also can help restrict the unwanted intrusion of personal values into service delivery. Another way to control the influence of values is to do research on the role of values in practice and to design agency procedures that help provide services objectively. As Research in Practice 2.1 illustrates, the influence of values on the actual conduct of practice cannot be totally eliminated. However, by relying on a practice approach that is empirically based, employing procedures supported by research, and incorporating rigorous evaluation procedures, it is possible to sensitize professionals to the impact of their value positions and, thus, enhance the objectivity of service delivery.
2.2 THEORIES IN RESEARCH AND PRACTICE
Theory is a word that is misunderstood by many people. To the neophyte, theories often are associated with the abstract, the impractical, or the unreal. In actuality, nothing could be further from the truth. In both research and practice settings, theories play a critical role in our understanding of reality and our ability to cope with problems. In fact, people commonly use theories in their daily lives without recognizing that they do so.
A theory is a set of interrelated, abstract propositions or statements that offers an explanation of some phenomenon (Skidmore, 1979). Three key elements in this definition are important to understanding theories. First, theories are made up of propositions , which are statements about the relationship between some elements in the theory. For example, a proposition from the differential association theory of crime is that a person becomes criminal because of an excess of definitions favorable to the violation of the law over definitions unfavorable to the violation of the law. Elements in this proposition include “criminal” and “definitions favorable to the violation of the law” (Sutherland, 1939). Behavior modification theory also contains numerous propositions, such as: Behavior change can occur through a reorganization of the environmental cues that reward and punish behavior (Sulzer-Azaroff & Mayer, 1991). The elements in this proposition include “behavior change,” “environmental cues,” and “reward and punish behavior.”
RESEARCH IN PRACTICE 2.1 Practice Effectiveness: Providing Services to Gays and Lesbians
Human service practitioners often deal with clients or situations that involve value-laden controversies. One area where this clearly is the case is the issue of whether gays or lesbians should be permitted to have custody of their children or to adopt. Traditional—and, in some cases, religious—values lead some people to the conclusion that only intact, heterosexual, married couples provide a suitable environment for child rearing and that being raised by a gay or lesbian parent would be harmful to a child. The difficult problem for the human service provider is how to serve clients within the context of one's own personal and professional values.
The movement toward evidence-based practice in the human services suggests that it is possible to use scientific research to assess whether personal or professional values are unreasonably influencing the services provided to clients. One of the primary features of evidence-based practice involves identifying general trends gleaned from a review of studies that are applicable to practice. The provision of child welfare services to gay and lesbian clients offers an excellent example of this feature. A human service professional charged with making recommendations about child custody and adoption would approach this situation by systematically searching the literature and asking the question: How well do gays and lesbians perform in the role of parent?
Marcus Tye (2003) summarizes the theoretical and empirical evidence currently available to answer this question. Over the past quarter century, beginning with the work of Karen Lewis (1980), a considerable body of research has accumulated on the development and experiences of the children of lesbians and gay men (Stacey & Biblarz, 2001). For the most part, this research does not support the negative developmental outcomes for the children that would be predicted by much traditional and experiential knowledge or some people's personal values. Children raised by gay or lesbian parents do not differ in their sexual orientation or personal development from children raised by heterosexual parents (Goldberg, 2009; Gottman, 1990).
By the 1990s the research focus had shifted from looking at the quality of family life after a parent had acknowledged his or her homosexuality toward the examination of gay men and lesbians who, often in the context of a committed homosexual relationship, were opting to have children via the processes of adoption, artificial insemination, or surrogate mothers. This research also has found no negative effects for adopted children raised in gay- or lesbian-headed families. These adoptive families exhibited a level of family functioning that, by many measures, was at least as high as that found in other types of families (Erich et al. 2005; Leung, Erich, & Kanenberg, 2005).
So this evidence-based review suggests that children can be reared into healthy adults in families parented by homosexuals and sexual orientation should not preclude individuals from serving as adoptive or foster parents. Thus the general trends in the existing data help provide direction for shaping agency policy. Science cannot inform human service workers regarding what their personal values ought to be, but it can point out practice situations in which personal values seem to be unreasonably intruding on intervention decisions. Such research also safeguards against the danger that decisions based on personal values will masquerade as “in the client's best interest” by providing an empirical knowledge base for decision making.
A second important part of our definition of theory is that theories are abstract systems, meaning they link general and abstract propositions to particular, testable events or phenomena. In many cases, these abstract systems are deductive systems, a general set of propositions that can be used to deduce further, more concrete relationships between the elements of the theory. Differential association theory is again illustrative. As noted, this theory relates definitions favorable to the violation of the law with the greater likelihood of criminal behavior. This means that the theory is supposed to apply to all specific types of crimes, such as robbery, larceny, and auto theft. So it would be logical to deduce from the theory that greater exposure to definitions favorable to the violation of the law would be associated with higher incidences of robbery, larceny, and auto theft. Theories are abstract, because they have this deductive power: The broader and more abstract the propositions and their related concepts, the more numerous the specific relationships that can be deduced from them.
The third key aspect of theories is that they provide explanations for the phenomena they address. Indeed, the ultimate purpose of a theory is to explain why something occurred. In differential association theory, the phenomenon to be explained is criminal behavior, and the explanation is that criminality is learned through much the same process as noncriminal behavior is. The content of what is learned—namely, definitions favorable to violation of the law—makes the difference. Thus differential association provides an explanation for the development of criminal behavior.
In comprehending theories and the roles they play, it is helpful to realize that we all use theories in our everyday lives, although we may not call them theories or even be consciously aware of using them. Nonetheless, we base our decisions and behavior on our past experiences and what we have learned from others. From these experiences, we generalize that certain physical, psychological, and social processes are operative and will continue to be important in the future, with predictable consequences. This is our “commonsense theory” about how the world operates and forms the basis for our decisions. For example, most people have certain general notions—that is, personal theories—about what causes poverty. Some personal theories emphasize poverty as an individual problem: People are poor because of their individual characteristics, such as laziness, low intelligence, poor education, or lack of marketable skills. Others' theories of poverty emphasize structural features of the American economy dictating that, even in times of economic expansion, some people will be left impoverished through no fault of their own. Which of these theories people identify with most closely determines, in part, how they react to poor people and which public policy provisions toward poverty they support. Advocates of the individualistic theory might be hostile toward the poor and programs to aid them, because they believe the poor are undeserving people who suffer only from their own shortcomings. Supporters of the structural theory may view the poor as victims and tend to be more benevolent toward them.
Personal theories like these concerning poverty may be extreme and misleading, because they are based on casual observations, personal experience, or other information lacking the rigorous concern for accuracy of scientific investigations. Unlike commonsense theories, the theories in research and practice are precise, detailed, and explicit. It is important, however, to recognize that a theory is always tentative in nature—that is, any theory is best viewed as a possible explanation for the phenomenon under investigation. By conducting research, scientists gather evidence that either supports or fails to support a theoretical explanation or practice intervention. No theory stands or falls on the basis of one trial. Theories are tested over a long period of time by many investigations. Only with the accumulation of research outcomes can one begin to have confidence concerning the validity of a theory.
We have all heard the refrain “It's only a theory” or “That's your theory.” Such phrases often are used in the context of deflating an argument. Actually, these comments, though often intended in a disparaging sense, convey some truth regarding theories. In particular, they point out that theories are sometimes untested (but testable) assertions about reality and that theories are not the end product of scientific investigation but, rather, a part of the process of science. Theories have particular purposes in both research and practice settings. In fact, the same theories often are used in both research and practice, because both researchers and practitioners turn to them for similar reasons. We can identify three major functions of theories in research and practice.
Explanation of Phenomena. As we have seen, theories provide an explanation for phenomena. They say not only what will happen under certain conditions (which is what hypotheses also do, but more concretely) but also why it will happen. This provides a much more powerful understanding of human behavior. In differential association theory, for example, the phenomenon to be explained is criminal behavior, and the explanation is that criminal behavior is a product of learning appropriate behaviors from others who are important to us. Thus, differential association theory provides a broad, abstract explanation for the development of criminal behavior that links such behavior with general processes of conformity and group process. People learn to be criminals in the same way they learn to be doctors, nurses, or lawyers—namely, by learning through association with other people.
Guide for Research and Practice. Theories guide and direct research and practice. They focus attention on certain phenomena as relevant to the issues of concern. If we were to dispense with theories altogether, as some would suggest, then what would we study? What data would be collected? What intervention strategy would be adopted? Theories help us find answers to these questions.
Imagine that a counseling center wants to attack the problem of teenage alcohol consumption at a particular high school and the staff decides to study the problem. Where to begin? What variables are important? As a first step, it is essential to fall back on some theory related to these issues. We might, for example, use the theory of differential association, which posits that alcohol consumption results from attitudes and patterns of behavior that are learned in association with other people, particularly peers. To test this theory, we could determine whether alcohol consumption is more common when it is viewed as an acceptable form of behavior among peers. We are then in a position to collect data on attitudes toward alcohol and patterns of alcohol consumption in peer groups. If the theory is confirmed, then it supports the idea that effective intervention will need to focus on attitudes toward alcohol consumption in peer groups.
We could have selected a different theory regarding alcohol consumption. For example, some theories posit an inherited predisposition toward alcoholism. Other theories suggest that alcoholism results from a nutritional deficiency that is satiated by alcohol consumption. We do not presume to suggest which theory is more accurate—future research will settle that issue, one hopes. The same thing occurs in practice intervention. If a practitioner used crisis intervention theory to deal with the disruption caused by an alcoholic parent, then the theory would direct attention to such factors as family coping strengths and emotional adaptation. Community-organization practice theory, on the other hand, would focus on the community resources available to recovering alcoholics and community services for their families. The point is that the theories used by researchers and practitioners guide their approaches and focus their attention on particular phenomena.
Integration of Multiple Observations. Theories help integrate and explain the many observations made in diverse settings by researchers and practitioners. They tell us why something happened, and they enable us to link the outcomes of numerous studies and interventions made in a variety of settings. As long as the findings of these efforts remain individual and isolated, they are not particularly valuable to science. Recall that a single observation is viewed with considerable skepticism. Single research findings may be in error, may be passed over and forgotten, or their broader implications may be missed entirely. Theories enable us to organize these dispersed findings into a larger, explanatory scheme. For example, someone investigating problem pregnancies among teenagers might observe that the groups of teenagers among whom such pregnancies are common tend to view an out-of-wedlock parenthood in a positive fashion. A familiarity with differential association theory would suggest that the social learning processes important in teenage drinking also may be relevant in problem pregnancies among teenagers. If this is the case, practitioners working in one area may be able to borrow strategies for intervention from the other area. Thus, theories integrate the findings from independent research endeavors and provide implications for intervention strategies.
Theories, then, play an important part in both research and practice, but one point needs to be reiterated: The utility of theories must be based on their demonstrated effectiveness. Theories should never be allowed to become “sacred cows,” the use of which is based on tradition or custom. Most authorities would agree with this conclusion: “The most important criterion to consider is the extent to which a given theory has been supported by empirical research” (Hepworth & Larsen, 1990, p. 18). In other words, has the intervention been shown to produce the desired results? In scientific research, this is called the verification of theories. Researchers approach the problem of verification by developing and testing hypotheses. This process of verification is diagrammed in Figure 2.1, which also shows a parallel process as it occurs in human service practice.
2.3 CONCEPTS AND HYPOTHESES
An important part of theories is concepts: mental constructs or images developed to symbolize ideas, persons, things, or events. Concepts are the elements of theories discussed earlier; they are the building blocks that are interrelated in propositions to form the explanatory statements of a theory (Alford, 1998). Some of the concepts in behavior modification theory, for example, are reinforcement, conditioning, learning, and behavior change.
Concepts are similar in function to the words we use during everyday communication. The word automobile, for example, is the agreed on symbol for a particular object that is used as a mode of transportation. The symbol or word is not the object itself, but, rather, something that stands for or represents that object. Scientific concepts, like words in everyday language, also are symbols that can refer to an extremely broad range of referents. They may refer to something fairly concrete, like gender, or to something highly abstract, like reinforcement or cohesion.
Despite the similarities between scientific concepts and ordinary words, some differences are critical to the scientific endeavor. In particular, concepts used in scientific research must be defined very carefully. We can get along quite well with the words we use for everyday communication, having only a general idea of how these words are defined. In fact, it is doubtful whether most people could give a dictionary-perfect definition of even the most commonly used words. Such imprecision in the use of scientific concepts, however, is totally inadequate. Scientists, who are widely scattered both geographically and temporally, carry on research that tests various aspects of theories. For these disconnected research projects to produce information of maximum utility, all the bits of knowledge need to be integrated into an explanatory scheme—namely a theory. This accumulation of knowledge is severely hampered—in fact, it becomes practically impossible—if these isolated scientists use different definitions of the same concepts.
FIGURE 2.1 The Process of Theory Verification in Research Compared with Practice Intervention
For example, many studies of the relationship between reinforcement and learning have been conducted. If the results of different studies are to be comparable, then the concepts of reinforcement and learning should be defined the same way. Learning, for instance, can be defined in behavioral terms (as the performance of a new behavior) or in cognitive terms (as the understanding of how a particular behavior might be performed). When defined in these two different ways, the concept refers to something quite different in the world, and the results from two investigations using the different definitions would not be directly comparable. Perhaps, for example, behavioral learning occurs under quite different conditions than cognitive learning does.
Scientific analysis involves two types of definitions of concepts, each functioning at a different level of analysis and each serving a different purpose. At the theoretical or abstract level, concepts are given nominal definitions: verbal definitions in which scientists agree that one set of words or symbols will be used to stand for another set of words or symbols. Nominal definitions are directly analogous to the dictionary definitions of ordinary words in which a phrase is designed to give meaning to the word or concept being defined (Cohen & Nagel, 1934). For example, a nominal definition of “poverty” might be a deficiency in resources to the extent that people are not able to maintain a lifestyle considered to be minimally acceptable in a particular society (Sullivan, 2012).
An important step in moving from the abstract level of theory to the concrete level of research is to give concepts operational definitions: definitions that indicate the precise procedures or operations to be followed in measuring a concept. For example, Mollie Orshansky developed one of the most widely used operational definitions of poverty for the Social Security Administration (Ruggles, 1990). Her measure, still used by the government as a basis for policy decisions, is based on what it costs to purchase a low-budget, nutritious diet for a family. If we use U.S. Department of Agriculture figures, the poverty line is determined by the cost of food, the size of the family, and other factors. This operational definition of poverty yields a series of income cutoffs below which families are defined as poor. This is a precise definition that lists the exact operations—in this case, mathematical operations—to follow in defining poverty. Anyone using this definition measures the same thing in the same way.
The process of moving from nominal to operational definitions can be complex, because concepts are very general and abstract and controversy often arises over exactly what they refer to. Some concepts that have been a part of the literature for decades have yet to be operationalized in a way that is fully satisfactory. For example, “alcoholism” has proved to be extremely difficult to operationalize, especially in terms of establishing where social drinking leaves off and alcoholism begins (Schuckit, 2006). Because of substantial individual and cultural differences, simple measures relying on the amount and frequency of consumption are inadequate. Researchers have been forced to operationalize alcoholism on the basis of such symptoms as family or work problems, morning drinking, poor eating, and recurrent blackouts. Whereas symptom-based measures of alcoholism avoid the errors inherent in consumption measures, substantial controversy remains concerning which symptoms are the best indicators, how many symptoms must be evident, and how serious they must be before the label of alcoholic may be meaningfully applied.
Even the concept of poverty, which may seem straightforward and easy to operationalize, has proven to be controversial. There is, of course, the issue of where to set the income cutoffs. Orshansky's cutoffs are based on the assumption that the average American family spends one third of its income on food; some critics have argued that this results in poverty thresholds that are too low. Furthermore, Orshansky's definition sets a fixed income level as the poverty level; thus, it is unaffected by changing levels of affluence within society as a whole. Some have argued for a relative definition of poverty that defines as poor those who earn one third or one half of the median family income (Bell, 1987). With such a definition, the poverty thresholds would rise automatically if the affluence of society as a whole increased. So it should be evident that operationalizing concepts can be difficult, complex, and sometimes, controversial. The process of moving from the nominal to the operational level is called measurement, and it is treated extensively in Chapters 5 and 13.
A common strategy in scientific investigations is to move from a general theory to a specific, researchable problem. A part of this strategy is to develop hypotheses , which are testable statements of presumed relationships between two or more concepts. Hypotheses state what we expect to find rather than what has already been determined to exist. A major purpose of developing hypotheses in research is to test the accuracy of a theory (see Figure 2.1). The concepts and propositions of which theories are composed usually are too broad and abstract to be directly tested. Concepts such as reinforcement and learning, for example, need to be specified empirically through operational definitions before they are amenable to testing. Once operationally defined, these concepts generally are referred to as variables , or things that are capable of taking on more than one value. If hypotheses are supported, then this supplies evidence for the accuracy of the theory on which they are based.
In the construction of hypotheses, the relationship between variables is stated in one of two possible directions: a positive relationship, or a negative (also called an inverse) relationship. In a positive relationship, the values of the variables change in the same direction, such as both increasing or both decreasing. For example, we might hypothesize that the acceptance of the use of alcohol among an adolescent's peers will lead to an increased likelihood that the adolescent will consume alcohol. In other words, as acceptance of the use of alcohol by one's peers increases, so does the adolescent's own use of alcohol. In a negative relationship, or an inverse relationship, the values of variables change in opposite directions. We might hypothesize, for example, that, among adolescents, reduced parental supervision will lead to an increase in the likelihood of substance abuse. In this case, as the value of one variable (parental supervision) declines, the value of the other (substance abuse) is predicted to increase.
Useful guidelines to keep in mind for developing hypotheses include the following:
1. Hypotheses are linked to more abstract theories. Although generating hypotheses without deriving them from theories is possible, hypotheses are always linked to theories because the theories provide explanations for why things happen.
2. It is important that the independent and dependent variables in hypotheses be clearly specified. The independent variable is the presumed active or causal variable—the one believed to be producing changes in the dependent variable. The dependent variable is the passive variable, or the one that is affected. In the previous examples, peer acceptance of alcohol and parental supervision are the independent variables, and alcohol use and substance abuse are the dependent variables.
3. It is important that the precise nature and direction of the relationship between variables be specified in the hypothesis. Students sometimes are tempted to state hypotheses like this: “Parental supervision will have an effect on teenage alcohol use.” However, although this statement says that there is a relationship, it doesn't say what the nature or direction of the relationship is. A proper hypothesis, as above, would state how changes in one variable will be associated with particular changes in the other: “As parental supervision decreases, teenage alcohol use increases.”
4. Hypotheses should be stated in such a way that they can be verified or refuted. If this is not the case, then they are not hypotheses. Hypotheses, after all, are statements about which we can gather empirical evidence to determine whether they are correct or false. A common pitfall is to make statements that involve judgments or values rather than issues of empirical observation. For example, we might hypothesize that investigations should be increased to reduce the incidence of welfare fraud. On the surface, this statement might appear to be a hypothesis, because it relates to investigations and welfare fraud in a negative direction. As stated, however, it is not a testable hypothesis. The problem is the evaluative “should be.” What should or should not be social policy has no place in hypotheses. However, the statement can be modified so that it qualifies as a testable hypothesis: “Increased levels of investigation tend to reduce the incidence of welfare fraud.” The hypothesis now makes an empirical assertion that can be checked against fact.
5. All the concepts and comparisons in hypotheses must be clearly stated. For example, consider the following hypothesis: “Southern Baptists have superior moral standards.” The concept of “moral standards” is so abstract and vague that it is impossible to know what it means. This would have to be clearly specified in terms of what is considered to be a moral standard. In addition, to say that someone's standards are superior requires a referent for comparison: superior to whom or what? It could mean higher than some other religious group, or it could mean higher than some chosen, absolute standard.
Developing hypotheses from theories is a creative process that depends, in part, on the insight of the investigator. Because hypotheses link theories to particular, concrete settings, the researcher's insight often is the trigger for making such connections. In addition, researchers, at times, combine two or more theories to develop hypotheses that neither theory alone is capable of generating.
Concepts and Operational Definitions among Minority Populations
When conducting research on minority populations, considerable opportunity for bias exists if concepts and operational definitions are not carefully developed. This has been a chronic problem with research on crime. For example, many people believe that nonwhites commit crimes at a higher rate than we would expect, given their numbers in the population. Although this is partly true, it greatly oversimplifies a complex reality, and it reflects how crime is typically operationalized. Official crime statistics from the Federal Bureau of Investigation (Federal Bureau of Investigation, 2012) are an important source of data on crime. The FBI operationalizes some crimes as “offenses cleared by arrest” and others as “offenses known to the police.” In other words, an occurrence is not officially considered to be a crime until it is “known to the police” or “cleared by arrest.” These official crime statistics show that nonwhites commit more crimes, proportionate to their numbers in the population, than whites do. However, this is a function, in part, of how the official statistics operationalize the concept of crime. We know that nonwhites are more likely to be arrested for a given offense, suggesting that it may be arrest that is more common among nonwhites rather than the actual commission of crimes. It has been proposed that nonwhites also are more likely to commit highly visible crimes, such as armed robbery or assault, that are more frequently reported to the police and result in an arrest. Some suggest that whites, on the other hand, commit more “hidden” crimes, such as embezzlement or fraud, that are less likely to come to the attention of the police. Research suggests that there may be no class difference in the amount of hidden crimes that are committed (Elliott & Huizinga, 1983). In addition, there are other ways to operationalize crime, such as through victimization studies (asking people if they have been a victim of a crime) and self-reports (having people anonymously report their own involvement in crime). Studies based on these operational definitions tend to show much smaller differences between white and nonwhite crime rates.
Another area in which poorly constructed operational definitions have produced misleading conclusions is that of intimate partner violence (Lockhart, 1991). Most studies have found rates of intimate partner violence to be considerably higher among African Americans than among whites. Typically, these studies have used one of the following as an operational definition for the occurrence of abusive violence: a homicide involving a domestic killing, a battered woman seeking care in an emergency room or social service setting, a wife-abuse claim handled by a domestic court, or a domestic dispute call to a police department. It is well known, however, that African Americans are overrepresented among people who come to the attention of police, emergency room personnel, or social service workers. Because they generally are overrepresented among these populations, they will appear to have higher rates of abuse than whites will when abuse is operationalized in this fashion. These problems can be reduced by selecting a sample of people from a community and having them answer questions about the amount of conflict and violence that occurs in their own families. This avoids the biased effect of looking only at certain locales. The National Family Violence Resurvey, for example, employed a sampling strategy that selected about 6,000 cases representing all racial and ethnic groups (Straus & Gelles 1988).
The Committee on the Status of Women in Sociology (1986) has indicated another area in which operational definitions have led to misleading results: studies of work and social contribution. Often, work is operationalized in terms of paid employment, but this excludes many other types of work from consideration, such as community service or home-based work. With this kind of operational definition, if an employee of a carpet-cleaning firm shampoos the carpets in a home for a fee, that counts as work, but if a woman does the same activity on her own time in her own home, that is not classified as work. Such an operationalization of work tends to underestimate the extent of productive activity engaged in by women, because women are less likely than men to be paid for their social contributions.
So in developing operational definitions, care must be taken to assess whether these definitions might lead to a distorted view of minorities. In some cases, this calls for careful consideration of what a concept is intended to mean. For example, is the focus of the research on paid employment, or is it on social contribution? In other cases, it calls for careful assessment of whether a definition will lead to an inaccurate, over- or underrepresentation of minorities.
2.4 PERSPECTIVES ON SCIENCE
Up to this point in these first two chapters, we have presented science as if it is a coherent, unified activity about which all scientists are in agreement. It isn't. Or, more accurately, we should say that some people believe it is a coherent and unified activity, whereas others are critical of that claim. Scientists vigorously debate a number of issues concerning the best ways to engage in scientific work. One such debate is whether science should be deductive or inductive in nature.
We mentioned earlier that theories often are deductive systems. This means that hypotheses can be logically derived from the propositions that make up a theory. So deductive reasoning involves deducing or inferring a conclusion from some premises or propositions. If the propositions—or the theory—are correct, then hypotheses logically derived from them also will be correct. In Figure 2.1, deduction involves moving from the level of theory to that of hypotheses or an intervention plan. Deductive reasoning is central to the scientific process.
Inductive reasoning, however, enables us to assess the validity of the hypotheses and the theory. Inductive reasoning involves inferring something about a whole group or class of objects from our knowledge of one or a few members of that group or class. We test one or a few hypotheses derived from a theory, and then we infer something about the validity of that theory as a whole. Thus, inductive reasoning carries us from the observations or interventions in Figure 2.1 to some assessment regarding the validity of the theory. The logic of scientific analysis involves an interplay between deduction, or deriving testable hypotheses, and induction, or assessing theories based on tests of hypotheses derived from the theories.
At times, inductive research is conducted without the benefit of previous deductive reasoning. This occurs in descriptive or exploratory research, where no theory exists from which to deduce hypotheses. In the absence of theory, we begin to make observations and then develop some theoretical propositions that would be plausible given those observations. For example, practitioners may observe that clients with problem pregnancies tend to come from families with low socioeconomic status. Based on the assumption that the parent-child bond is weaker in low socioeconomic families and, therefore, that such parents have less control over their children, the practitioners could inductively conclude that a weak parent-child bond leads to an increased risk of unwanted pregnancy. In other words, the observations are used to infer a proposition regarding the causes of unwanted pregnancies. In fact, as we will explore in more detail in Chapter 9, some researchers claim that such inductive approaches can be superior to deductive approaches, because the former can involve fewer hidden assumptions or preconceived notions on the part of the scientist. Some of these inductive approaches permit the data to shape the theory rather than having a preconceived theory impose meaning on the data. Inductive research of this sort can serve as a foundation for building a theory, and that theory, in turn, can serve as a source of testable hypotheses through deductive reasoning. Thus, induction and deduction are key links in the chain of scientific reasoning, and they parallel the reasoning process that is found in practice intervention.
Research in Practice 2.2 describes research projects that highlight many of the issues discussed in the previous two sections regarding the use of theories and hypotheses in research and the importance of inductive and deductive reasoning.
Beyond deciding whether to use deductive or inductive approaches, scientists also need to decide what type of explanations will be contained in the theory. Earlier in this chapter, we defined theories as involving explanations of some phenomena. An explanation is one way of gaining knowledge of something; it tells why something happens or specifies the conditions under which something occurs. Theories can focus on two different types of explanations (Miller, 1987; Nagel, 1961).
Nomothetic Explanations. Nomothetic explanations focus on a class of events and attempt to specify the conditions that seem common to all those events. We will use the social control theory of deviant behavior as an illustration. Social control theory argues, in part, that delinquent behaviors such as shoplifting are produced by weak attachments to parents. A nomothetic explanation, then, might attempt to prove that all juveniles who shoplift have weak attachments to their parents. The focus of the explanation is on understanding the entire category of youth who shoplift. These explanations do not focus on understanding all the causes of a phenomenon. In fact, control theory would recognize that a complex behavior such as shoplifting probably has many causes other than weak social bonds, and that other theories would be necessary to locate and identify those factors. For nomothetic explanations, knowledge results from an understanding of a particular cause in relation to a class of events.
Nomothetic explanations attempt to develop knowledge that can be generalized beyond a single study or set of circumstances. In a sense, a nomothetic explanation is designed to produce the conclusion that weak attachment to parents in all cases increases the likelihood that shoplifting will result. This doesn't mean that every person who experiences weak attachment will shoplift; however, it does mean that those people have a higher probability of engaging in shoplifting. To put it another way, a randomly selected group of teens with weak parental attachments will have a higher rate of shoplifting than a randomly selected group with strong attachments. The explanation or knowledge that is gained is probabilistic in nature: It tells us something about the probability of events occurring. The knowledge gained is about the aggregate, or the whole group, rather than about specific individuals in the group.
RESEARCH IN PRACTICE 2.2 Practice Effectiveness: Social Theory and Burnout among Social Workers
A social worker: I began to despise everyone and could not conceal my contempt.
A psychiatric nurse: Sometimes you can't help but feel “Damn it, they want to be there, and they're fuckers, so let them stay there.” You really put them down …
A social worker: I find myself caring less and possessing an extremely negative attitude.
[Maslach, 1979, p. 217]
These are hardly the caring, empathic reactions one would expect from human service workers, yet negative attitudes toward clients are expressed at some point by many social workers, nurses, psychologists, and others. The problem of burnout is of considerable concern to human service professionals, because it can impair their ability to deal with client problems. Burnout refers to a service worker's emotional disengagement from clients, dissatisfaction with his or her job, feelings of worthlessness, and physical and interpersonal problems (Arches, 1991). Commonsense approaches often focus on the personal abilities of human service workers to explain why they suffer burnout: They lack sufficient emotional strength or distance from clients, or they overidentify or overempathize with their clients. Rather than relying on such intuition, scientific researchers turn to theories for direction in identifying variables that might play a part.
When social work researcher W. David Harrison (1980) approached these issues, he turned to role theory, which views human behavior as resulting from conformity to expectations that are associated with particular roles. One of the tenets of role theory is that role expectations should be clear, unambiguous, and achievable. Furthermore, the various expectations associated with a role should not conflict with one another. Previous research suggested that situations in which role expectations are conflicting, incompatible, or unclear lead to personal stress and dissatisfaction. Role theory enabled Harrison to identify two different kinds of role difficulty: Role conflict refers to a situation in which conflicting and incompatible demands are placed on a person in a role, and role ambiguity refers to a lack of clarity in terms of what is expected of a person in a particular role. Harrison's research on child protective service workers showed that role difficulties, especially role ambiguity, produced job dissatisfaction and burnout among these social workers.
In contrast, Joan Arches (1991) turned to theories of organizational structure and change, reasoning that recent developments in social service organizations might have an impact on burnout. These theories suggest that increasing bureaucratization and centralization in organizations can reduce workers' feelings of autonomy, and, in turn, this can contribute to the job dissatisfaction that often is a part of burnout. Arches's research then provided evidence that this was the case, offering further verification for those organizational theories.
Burnout continues to be a topic of interest to social scientists and human service practitioners throughout the world. Peter Janssen and his colleagues at Utrecht University in the Netherlands reviewed the extensive body of research on burnout that has accumulated over the past few decades and designed a study of Dutch nurses (Janssen, Schaufeli, & Houkes, 1999). Their research examined work-related and individual determinants of burnout by using conservation of resources theory as a framework. This theory focuses on the impact of work-related demands and resources on different dimensions of burnout. They found that a scarcity of resources, in the form of excess job demands and work overload, increased emotional exhaustion (one dimension of burnout) but not depersonalization (another dimension of burnout).
Thus, the theoretical considerations of role theory, organizational theory, and conservation of resources theory do not point toward excessive empathy or emotional weakness as the culprits in burnout among human service workers. Rather, the organizational and role structures that surround them are important. These investigations illustrate the importance of grounding research in theory, because it is theory that suggests which variables might be important and how they might relate to one another. Theory also shows how hypotheses can be developed through deductive reasoning. Once confirmed, the hypotheses of these researchers provide support, through inductive reasoning, for the interpretations of these theories in regard to the causes of burnout in the human services. Based on this slow, methodical accumulation of knowledge, we should eventually establish a solid foundation from which to develop programs to alleviate the problem of burnout among human service workers.
Once you understand what nomothetic explanations consist of, you can begin to see their weaknesses. One weakness is that you cannot say for sure what will happen in any particular case or to any specific person. You cannot say whether Joe Smith, who has experienced weak parental attachment, will become a shoplifter. A second weakness is that you cannot make any claims of knowing the totality of causes that produced some event or phenomenon. So the knowledge, though valuable, is incomplete. There may be, for example, some key factors that must occur in combination with weak parental attachments to produce shoplifting.
Idiographic Explanations. Idiographic explanations focus on a single person, event, or situation and attempt to specify all the conditions that helped produce it. An idiographic explanation of shoplifting, for example, might focus on one juvenile who shoplifts and attempt to understand the multiple factors that contributed to bringing about the shoplifting behavior in that person. The focus of the explanation is on a particular, unique individual or situation. These explanations do not attempt to understand all instances of shoplifting; in fact, they recognize that other shoplifters may be propelled by a different combination of causes. For idiographic explanations, knowledge results from a thorough understanding of the particular.
Idiographic explanations see causality in terms of a complex pattern of factors that combine over a period of time to produce an outcome. To truly understand something, researchers need to comprehend that whole patterned sequence, the whole complex context in which something occurs. When the nomothetic approach isolates particular variables for study, knowledge is incomplete for two reasons: First, some factors or variables have not been included in the investigation. Second, the isolating approach cannot see how the combination of, or the interaction among, the various elements plays a critical role in producing an outcome. It may be, for example, that weak parental bonds produce shoplifting only when they combine or interact with a host of other factors. In fact, it may be that the particular combination of factors that produces shoplifting in one person is unique and does not occur in other cases. It also may be that each distinct case of shoplifting is produced by a unique combination of factors. In other words, the explanation or knowledge that we gain is idiosyncratic.
Nomothetic explanations are probabilistic in nature, but idiographic explanations are deterministic in that the event being studied, such as shoplifting, actually did occur in the case being studied. In addition, the idiographic explanation identifies the causes that determined that outcome.
As with nomothetic explanations, idiographic explanations have weaknesses. One major fault is their limited generalizability. With such explanations, it is difficult to determine whether knowledge can be extended beyond the particular case or situation being studied.
Combining Explanations. Because each type of explanation has its strengths and its limitations, you might have guessed that our conclusion is going to be that neither type is inherently better than the other. As we alluded to in the beginning of this chapter with the excerpt from Through the Looking Glass, numerous routes to gaining knowledge about the world exist, and each type of explanation provides us with a valuable, though incomplete, route. In later chapters, we will see that some research methodologies, such as surveys and experiments, tend to be used to develop nomothetic explanations and that other methodologies, such as field research, in-depth interviewing, and historical comparative research, often are used to develop idiographic explanations. The point is to understand the logic of each type of explanation and be aware that conclusions supported by research using both types of explanations are more complete than if the research only uses one type.
Over the centuries, philosophers and scientists have debated the nature of reality and how people can know that reality (Couvalis, 1997; Miller, 1987). These have been controversial issues for scientists who study the physical world, but they are even more contentious among social scientists, who study human beings and their psychological and social reality. Part of the reason for this heightened contention is the belief that human beings are different from the natural world of physical objects and events. People emote, remember, speculate, love, and hate—they think about what is happening to them and have feelings about it. People refuse to behave the way a scientist hypothesizes that they might. People do the unexpected or the unpredictable. Atoms, molecules, and chemical compounds do not have these elusive properties, and this is one of the reasons why natural scientists often can make certain nonprobabilistic predictions about what will happen: Under a certain set of conditions, all water molecules, for example, will freeze when the temperature drops below zero degrees centigrade. Thus far, however, social scientists have been unable to make such statements about social reality.
Another reason that the issue of how we know the world has been controversial among social scientists is that the scientists who study social reality are people themselves, with personal values, goals, desires, and reactions to what they observe. These personal matters may interfere with their ability to comprehend the world accurately. Going a step further, the scientific endeavor is itself a social process, part of the social world that social scientists attempt to understand. After all, scientific work can advance one's career, help one make a living, and move one up (or down) in the stratification system. In doing their scientific work, scientists may be influenced by a variety of social and psychological factors that routinely influence other human beings in their social endeavors.
What does all this mean? For one thing, science is a much more complicated—and, in many respects, a much messier—enterprise than many people recognize. For another, a number of competing perspectives exist concerning the issues of how society works and what implications this has for how the scientific endeavor works. In fact, historian Thomas Kuhn (1970), in a groundbreaking study of scientific work over many centuries, concluded that scientific activity is shaped by paradigms , which are general ways of thinking about how the world works and how we gain knowledge about the world. Paradigms are fundamental orientations, perspectives, or world views that often are not questioned or subjected to empirical tests. People may not even be aware that their thinking about the world is shaped by an orientation or world view. In his study of the history of science, Kuhn discovered that, although paradigms change over time, scientific research at any given moment was shaped by the paradigm that was dominant at the time. Research that fell outside that paradigm was considered to be inappropriate, irrelevant, oddball, or just plain wrong. In a sense, the world of paradigms falls outside the scientific realm in that issues are not accepted or rejected on the basis of empirical evidence; instead, some things are considered true—and others false—because it is obvious that that is how things work. Evidence supporting the paradigm will be accepted and competing evidence either ignored or rejected.
At the risk of oversimplification, we can classify the paradigms in the social sciences into two general categories: positivist approaches, and a number of different approaches that we will call nonpositivist approaches (Alford, 1998; Benton, 1977; Smart, 1976). Keep in mind that these viewpoints are not necessarily mutually exclusive; people may adopt ideas from more than one of them at the same time. In addition, one could agree with some parts of a paradigm but disagree with other parts of the same paradigm. We address this issue early in the book because it is a debate that arises repeatedly as we discuss different research methodologies.
Positivist Approaches. Positivism (sometimes also called logical empiricism ) argues that the world exists independently of people's perceptions of it and that science uses objective techniques to discover what exists in the world (Blaikie, 1993; Durkheim, 1938; Halfpenny, 1982). Astronomers, for example, use telescopes to discover stars and galaxies, which exist regardless of whether we are aware of them. So, too, scientists can study human beings in terms of observable behaviors that can be recorded using objective techniques. Recording people's gender, age, height, weight, or socioeconomic position are legitimate and objective measurement techniques—the equivalent of the physicist measuring the temperature, volume, or mass of some liquid or solid. For the positivist, quantifying these measurements—for example, assessing the average age of a group or looking at the percentage of a group that is male—is merely a precise way of describing and summarizing an objective reality. Such measurement provides a solid, objective foundation for understanding human social behavior. Limiting study to observable behaviors and using objective techniques, positivists argue, is most likely to produce systematic and repeatable research results that are open to refutation by other scientists.
The natural and social world is governed by natural and social rules and regularities that give it pattern, order, and predictability. The goal of research in the natural and social sciences is to discover laws about how the world works and to express those discovered regularities in the deductive theories and propositions that are discussed in this chapter. As scientists conduct research, they move progressively closer to the truth, which involves uncovering the laws and patterns that underlie objective reality. So, at least in its ideal form, science is an objective search for the truth in which human values are a hindrance whose impact should be limited if not eliminated. Values can only interfere with the objective search for truth. For example, Emile Durkheim, an early sociologist, was a strong believer that sociologists could study the social world in much the same way that physical scientists could study the physical world. Durkheim believed that there were “social facts” that social scientists could observe and then use those observations to discover the social laws that govern the social world. He believed that once we discover these social laws, we will be able to both explain and predict human social behavior.
Of the various paradigms that we will review, positivism clearly is the most widely held view among natural scientists and, to a lesser degree, among social scientists. Among social scientists, those who adopt the positivist stance often tend to use certain kinds of research methodologies. For example, they tend toward quantitative research , which involves measurement of phenomena using numbers and counts. They also tend to use deductive and nomothetic explanations, experimental designs, and survey research. It is important not to oversimplify the link between a paradigm and the preferred research methodology, however, because positivists at times use qualitative research , which involves data in the form of words, pictures, descriptions, or narratives rather than numbers and counts. They also use inductive or idiographic explanations and field observations when these are appropriate to a research question.
Despite the popularity and dominance of the positivist paradigm, it has been subject to considerable criticism over the years. Some of this criticism arises out of empirical studies by social scientists of exactly how science operates (Galison & Stump, 1996; Lynch & Bogen 1997; Shapm 1995). What many of these researchers find is that what scientists actually do looks quite different from what the positivist paradigm says science should look like. This has led some critics to conclude that the positivist model is an idealized conception of science rather than an accurate description of it. Based on these and other concerns, alternative paradigms have emerged.
Nonpositivist Approaches. One prominent nonpositivist approach to science is what is called the interpretive approach. Interpretive approaches (also called interactionist or verstehen approaches) posit that social reality has a subjective component that arises out of the creation and exchange of social meanings during the process of social interaction. Social science must have ways to understand this subjective reality (Holstein & Gubrium, 1994; Smith, 1989; Wilson, 1970). Interpretivists argue that the objective, quantitative approaches of positivism miss this very important part of the human experience: the subjective and personal meanings that people attach to themselves and what they do. Reality is seen as something emergent and in constant flux that arises out of the creation and exchange of social meanings during the process of social interaction. Rather than seeing reality as something apart from human perceptions, interpretive social science sees reality—or, at least, social reality—as created out of human perception and the interpretation of meaning. These kinds of ideas led many nineteenth-century and early twentieth-century theorists, such as Wilhelm Dilthey, Ernst Troeltsch, and Max Weber, to conclude that social life cannot be understood by the same method that is used to study the natural world (Barnes, 1948).
Weber, for example, argued that we need to look not only at what people do but also at what they think and feel about what is happening to them (Weber, 1957, orig. pub. 1925). This “meaning” or “feeling” or “interpretive” dimension cannot be adequately captured through objective, quantitative measurement techniques. Researchers need to gain what Weber called verstehen , or a subjective understanding. They need to view and experience the situation from the perspective of the people themselves, “to walk a mile in their shoes.” They need to talk to the people at length and immerse themselves in their lives so they can experience the highs and lows, the joys and sorrows, the triumphs and tragedies as seen from the perspective of the people being studied. Researchers need to see how individuals experience and give meaning to what is happening to them. Interpretive research methods provide an understanding through empathy or fellow feeling, whereas positivist methods provide understanding through abstract explanation. Yet, the important point is that both methods provide an understanding of the world, and both are a part of the scientific enterprise.
Qualitative research methods attempt to gain access to that personal, subjective experience; for interpretivists, quantitative research by its very nature misses this important dimension of social reality. Positivists, for their part, do not necessarily deny the existence or importance of subjective experiences, but they do question whether the subjective interpretations of the verstehen method have scientific validity.
According to the interpretivist approach, regularity and pattern in social life does not result from objective social laws that exist apart from the human experience and are discovered by scientists. Instead, pattern and predictability arise out of mutually created systems of meaning that emerge from social interaction (Rabinow & Sullivan, 1987; Roscoe, 1995). Regularity and pattern are created and maintained by people; they are not imposed by external force. Proponents of interpretive approaches argue that qualitative research methods enable the researcher to approximate verstehen, an understanding of the subjective experiences of people. Of course, actual access to such experience is impossible; thoughts and feelings, by their very nature, are private. Even when someone says how he or she feels, the speaker has objectified that subjective experience into words and, thus, changed it. Researchers, however, can gain some insight into subjective experiences by immersing themselves in the lives and daily experiences of the people they study. By experiencing the same culture, the same values, the same hopes and fears, researchers are in a better position to take on the point of view of these people. Despite its focus on subjective experiences, however, such research is still empirical in the sense that it is grounded in observation. Qualitative researchers consider their qualitative observations and conclusions to be no less systematic or scientific than the more positivistic quantitative research techniques. Although positivists would argue that subjective meaning is difficult to quantify and study objectively, interpretive researchers would argue that it is, nonetheless, a key part of human social reality.
Another important difference between positivists and interpretivists has to do with the role of science: Positivists argue that scientists merely discover what exists in the world, but some interpretivists claim that scientists actually help create social reality through their scientific work (Knorr, 1981). As researchers make observations, gather data, and draw conclusions, their activities contribute to the construction of patterns of meaning. Scientific principles and laws about social behavior become another aspect of reality that can influence people's behavior. Even something as simple as computing the average age of a group creates a new reality: Instead of recognizing that some people in the group are 22 years old, others 34 years old, and still others 43 years old, we now say that the “average age of the group is 36.7 years.” This summary statement gives the impression—and creates the reality—that the group members share something in common in terms of age and that we know something very precise about their ages. That sense of commonality or precision, however, comes from the numbers created by the scientist, not from reality. In addition, though the average appears to be very precise, it actually is less precise than listing all the ages of the group members.
The interpretive approach focuses more on inductive and idiographic theory construction than on deductive and nomothetic approaches, considering the theories to emerge out of people's experiences rather than viewing them as abstractions developed by scientists. Understanding and truth come from an empathic grasp of the social meanings of a setting rather than from statistical analysis and abstract generalization to large numbers of cases. Once again, however, the link between paradigms and research approaches is not mutually exclusive. At times, interpretive social scientists do deductive and nomothetic theory construction, and they have even been known to use quantitative methods when appropriate.
Other nonpositivist characterizations of science exist as well. For example, critical and feminist approaches to research argue that science is inevitably linked to inequitable distributions of power and resources. These approaches posit that groups can and do use science to enhance their position in society, and that patterns of dominance and subordination may exist between researchers and those on whom they conduct research. Other nonpositivist critiques will be addressed in later chapters. At this point, we simply want to raise the controversy regarding positivist and nonpositivist views of science to stress that science and scientific research are more complicated than you might have originally thought. The goal for the student should not be to attempt to resolve these disputes or choose among the paradigms. Instead, the goal should be to understand the dimensions of the debate, recognize how the paradigms are similar to or different from one another, and comprehend the implications of each paradigm for the research process. In addition, the paradigms are not completely exclusionary of one another. All the paradigms agree with much of what will be covered in this book. For example, all the paradigms base their search for knowledge on systematic observation, and all agree that scientific work should be open and public. Of course, they may not always agree on what makes observations systematic, but there is not total agreement within each paradigm about that issue, either.
Another reason why the student need not adopt a preferred paradigm is that many researchers do not choose a particular perspective to follow exclusively (Alford, 1998). Many researchers find that each of the approaches offers some insights into social life and the scientific process that the others ignore. They move back and forth among the paradigms, using the best that each has to offer in understanding a particular aspect of human social life.
2.5 CAUSE-AND-EFFECT RELATIONSHIPS
One of the more important yet difficult tasks in scientific research is the search for causes—that is, the reasons why particular forms of behavior occur. Why do child abuse and spousal abuse occur? Why do some juveniles become delinquent, whereas others present no behavior problems? Why do some people exhibit symptoms of mental illness and others appear to be psychologically stable?
Discovering causal relationships is a difficult task, because causality cannot be directly observed. Rather, it must be inferred from the observation of other factors. Because of this, the philosopher John Kemeny has labeled causality “the mysterious force” (1959, p. 49). We cannot see it, feel it, or hear it, but we often assume that it is there—and many scientists search for causality with hopefulness and tenacity. This search is a controversial task, because some philosophers, notably Bertrand Russell (1953), have argued for excluding the notion of causality from scientific investigation altogether. These people opt for restricting ourselves to description and analysis of “associations” without the implication that a “mysterious force” called causality lurks behind the scenes and orchestrates the actions of people and things. This controversy is longstanding; we do not presume to resolve it here. Nonetheless, it is important to understand the criteria that need to be satisfied if one wants to infer that one event caused another.
By causality , we mean that some independent variable (X) is the factor, or one of several factors, whose change produces variation in a dependent variable (Y). As noted, causality can only be inferred. We can observe the relationships among things in the world, and from that we can infer or deduce that changes in one factor are causing changes in another. However, it is always an inference. To infer the existence of a causal relationship, one must demonstrate the following:
1. A statistical association between the independent and dependent variables must exist.
2. The independent variable must occur prior in time to the dependent variable.
3. The relationship between independent and dependent variables must not be spurious; that is, the relationship must not disappear when the effects of other variables are taken into account.
TABLE 2.1 Effectiveness of Reading Media Reports on Smoking Cessation
|
|
|
Person Reads Report |
|
|
|
|
Yes |
No |
|
Person Quits |
Yes |
200 (50%) |
135 (27%) |
|
Smoking |
No |
200 (50%) |
365 (73%) |
|
|
Totals |
400 (100%) |
500 (100%) |
We will consider each requirement of causal inference in the context of an issue that is much in the news today—the campaign to reduce cigarette smoking. Over the years, there have been reports in the media about the negative impact of cigarette smoking on people's health. Some argue that making these reports public as part of a health campaign can motivate people to quit smoking. Table 2.1 presents hypothetical data that seems to show a link between reading such reports about smoking and actually quitting smoking: 50 percent of those who read the reports quit smoking, compared to only 27 percent of those who do not read the reports. Finding such a statistical relationship satisfies the first criterion for establishing a causal relationship.
The second requirement, that the independent variable occurs prior in time to the dependent, often is not as easy to establish. A major factor in this is the nature of the study. Some research techniques, such as the experiment or participant observation, are inherently longitudinal, which means that the researcher is in a position to trace the development of behavior as it unfolds over time. In these cases, establishing the time sequence of events generally is simple. Questions of temporal order are more difficult to resolve when dealing with cross-sectional data, such as surveys, in which measurements of the independent and dependent variables occur at the same time. This is especially true if the question of temporal sequence is not addressed until after the data have been collected. It is sometimes possible to sort out the time sequence of variables in survey data by asking additional questions. If the researcher does not gather the necessary information at the time of the survey, however, then establishing the appropriate time order of the variables may be impossible—hence the emphasis on the importance of carefully considering issues of data analysis when originally developing a research design.
The data in our illustration may suffer from this problem. One interpretation is that reading reports is the independent variable that influences whether people quit smoking, the dependent variable. For this interpretation to be correct, the reports would have to have been publicized before the people quit smoking. If the respondents were not asked when they quit smoking, then it would be impossible to say whether they quit smoking before or after reading the reports. Obviously, if they quit smoking before reading the reports, then such health campaigns could not have caused their change in behavior. In our example, without knowing the temporal sequence, one could argue logically for either factor being the cause of the other. Obviously the health campaign could encourage people to quit smoking if they become frightened by learning the dire consequences of their habit. However, those who quit smoking also could be happy with and proud of their victory and might enjoy reading reports on what could have happened to them had they not quit smoking. In this second scenario, quitting smoking would be the independent variable that increases the likelihood that people will read reports about the health threat of smoking, the dependent variable.
The final criterion necessary for inferring causality is that the relationship between the independent and dependent variables not be spurious, or disappear when the effects of other variables are considered. The logic of causal and spurious relationships is compared in Figure 2.2. This often is the most difficult of the three criteria to satisfy. In fact, one is never totally sure that some other variable—one you have not even considered—might not confound an apparent causal relationship. All that can be accomplished is to rule out as many extraneous variables as we can to the point where it is unlikely a variable exists that could render a given relationship spurious or noncausal.
FIGURE 2.2 Causal and Spurious Relationships
Considerable effort is expended during the design stage of research to control as many potentially troublesome extraneous variables as possible. Experiments, for example, are particularly good for avoiding spurious relationships because of the high degree of control that the experimental situation affords the researcher. Surveys, on the other hand, provide far less control, such that several variables capable of producing spuriousness typically have to be considered during data analysis. Several statistical techniques exist to control extraneous variables when the data are analyzed.
Returning to our example of smoking cessation, suppose we had solved the time sequence problem and, thus, had satisfied the first two requirements for establishing a causal relationship. We would now begin to consider variables that might render the relationship spurious. One variable that might do this is the level of education of the people studied. (The logic of this is outlined in Figure 2.3.) Considerable research links education with health behavior. Generally, people with higher levels of education engage in more health-promoting activities, such as quitting smoking or getting regular exercise. How do we determine whether the link between report reading and smoking cessation is spurious? We introduce the level of education as a control variable, which is illustrated with our hypothetical data in Table 2.2, in which we have divided the respondents in Table 2.1 into those with at least a high school education and those with less than a high school education. First, we can see by examining the row totals in each table that education is related to health behavior: 60 percent of the better-educated group have quit smoking, compared to only 19 percent of the less-educated group. However, we are really interested in what happens to the link between report reading and smoking cessation, and Table 2.2 shows that the relationship largely disappears: Within each educational group, the same percentage of people quit smoking among those who read the report as among those who did not. So educational level, not whether one has read the report, influences a person's likelihood of quitting smoking. Furthermore, in our hypothetical example, educational level also influences whether one reads the report: 300 out of 400, or 75 percent, of those with a high school education read the report, compared with only 100 of 500, or 20 percent, of those with less than a high school education. So in our example, the link between reading the report and quitting smoking is spurious; it occurs only because each of those two variables is affected by the same third variable.
FIGURE 2.3 Causal and Spurious Relationships between Reading a Report and Quitting Smoking
If we had found the link between report reading and smoking cessation to be nonspurious when we controlled for education, could we conclude that the relationship was causal? The answer is no. We could not come to that conclusion—at least not yet. All that we would know was that the relationship remained when one alternative explanation was ruled out. Any other variables that could render the relationship spurious also would have to be investigated—and the relationship still hold—before we could argue with any confidence that it was, in fact, causal. (More intricacies of this sort of analysis are addressed in Chapter 15.)
TABLE 2.2 Effectiveness of Reading Media Reports on Smoking Cessation, Controlling for Education
We said at the outset that establishing the existence of causal relationships is difficult. Statistical relationships are easy to find but, on further investigation, all too frequently turn out to be spurious. The appropriate time sequence also can be problematic, especially with survey data. All in all, establishing causal relationships is a difficult but an important and challenging task.
2.6 REVIEW AND CRITICAL THINKING
This chapter covers some of the building blocks of the scientific method: theories, propositions, hypotheses, concepts, variables, and so on. In using these building blocks, scientists try to be very careful when describing and analyzing the world in order to avoid misunderstanding. You can utilize some of these same building blocks in trying to critically analyze information for policy or practice purposes or for your everyday life.
Consider the following when reading or hearing something.
1. What sources of knowledge serve as the basis for a claim or assertion? What are the strengths or weaknesses of those sources of knowledge?
2. Specify any theories, concepts, or hypotheses that are contained in the information. Keep in mind that, in practice, policy, or everyday realms, these elements may be implicit in what people say and you may have to figure out what they are.
3. Is there any evidence that theories have been verified? How much verification?
4. Are there claims to objectivity (positivism)? Is there any indication that subjectivity, values, or interests play a part in shaping the knowledge?
5. Are there any causal statements being made? Does the evidence presented satisfy the criteria that warrant a causal inference?
One of the central themes of this book is how research can inform practice. For Internet resources on applying research to human service practice, we suggest you use a search engine and search for such phrases as “children's mental health” or “evaluation outcomes.” A site we found from such a search was the home page of the Office of Alcoholism and Substance Abuse Services (OASAS) of New York state (http://www.oasas.ny.gov). In this chapter, we discussed the development of hypotheses and introduced the ideas of operational definition, dependent variable, and independent variable. As a way of applying these concepts to actual research, select several of the research studies reported at the OASAS site. For each study you find, identify one or more hypotheses that the study addressed. Determine if the review identifies how the dependent and independent variables were operationally defined. Finally, if the review describes how the independent and dependent variables are measured, state whether success in the program would be expected to show an inverse or direct relationship between the dependent and independent variables.
We covered some fairly core and controversial issues in social research in this chapter, especially in relation to the positivist and nonpositivist paradigms. Suppose that you work for a research organization and your boss wants you to create an online discussion of these paradigms. You need to locate people, organizations, or resources on the web that espouse one position or another on these issues. Use a search engine and enter one or more of the terms used to identify the paradigms or that have some relationship to the paradigms: “positivism,” “logical positivism,” “subjectivism,” “relativism,” “postmodernism,” “feminism,” and so on. From your search, discuss how many different social sciences address issues having to do with these paradigms. Do some social sciences address different issues than other social sciences? Do the same kind of analysis for the natural sciences. Also address whether fields that are nonscientific, such as the humanities, address these issues. Can you put together the online discussion your boss has requested?
Bengtson, V. L., Acock, A. C., Allen, K. R., Dilworth-Anderson, P., & Klein, D. M. (2005) Sourcebook of Family Theory and Research. Thousand Oaks, Calif.: Sage. This reference work on theory and methods for family scholars demonstrates how the development of theory is crucial to the future of family research. The work focuses on the process of theory building and designing research.
Glaser, B.G., & Strauss, A. L. (1967). The Discovery of Grounded Theory. New York: Aldine. An excellent book about the virtues and procedures of developing theoretical propositions from data. This approach emphasizes qualitative research and induction.
Hoover, K. R., & Donovan, T. (2011). The Elements of Social Scientific Thinking. 10th ed. Belmont, Calif.: Wadsworth/Cengage Learning. A brief and readable initiation into social science thinking and research. It is intended for those who use the results of research and for those just getting into the field.
Merton, R. K. (1968). Social Theory and Social Structure. 2nd ed. New York: Free Press. A classic statement by a sociologist on the relationship between theory and research.
Shoemaker, P. J., Tankard, J. W. Jr., & Lasorsa, D. L. (2004). How to Build Social Science Theories. Thousand Oaks, Calif.: Sage. This book traces theories from their most rudimentary building blocks (terminology and definitions) through
multivariable theoretical statements, models, the role of creativity in theory building, and how theories are used and evaluated. The book includes a discussion of concepts and their theoretical and operational definitions.
Turner, J. H., ed. (1989). Theory Building in Sociology: Assessing Theoretical Accumulation. Newbury Park, Calif.: Sage. In this collection of essays, one of the foremost U.S. theoreticians in sociology addresses a key assertion of the positivist approach: Does knowledge accumulate through the deductive approach of theory building and hypothesis testing?
Van de Ven, A. H. (2007). Engaged Scholarship: A Guide for Organizational and Social Research. Oxford: Oxford University Press. The author is a leading management researcher and writes from that perspective, but the issues he addresses are relevant to human service research as well. He especially addresses the issue of whether research findings and knowledge should be useful for science, practice, and policy. He also discusses how such research should be designed, carried out, and disseminated to achieve the twin goals of rigor and relevance.
Watts, D. J. (2011). Everything Is Obvious⋆ (⋆Once You Know the Answer). New York: Crown. This is an entertaining and insightful exploration of common sense as a source of knowledge—pointing to its weaknesses and failures as a foundation for thinking, reasoning, and planning.
Class Exercises for Competency Assessment
2.1. [Critical Thinking] Many commonsense beliefs relate to child development, such as “spare the rod and spoil the child,” or the belief that age-graded schools enhance learning. Through class discussion, develop a list of such “known” principles of child development. For each such principle, decide whether it is based on traditional knowledge, experiential knowledge, or a combination of both. How might you conduct systematic observations to determine the worth of these statements as scientific knowledge?
2.2. A mental health worker assigned to a large residential facility for senior citizens receives a request from staff members to “do something” about a new resident, a 72-year-old woman. From the information provided, the woman has apparently been assigned to an eighth-floor room, but she refuses to take the elevator alone or if it is crowded. The woman becomes terrified of the enclosed space and uses the stairs unless she can ride the elevator with a staff member. The woman's husband died about six months ago, and she is now living alone for the first time.
a. [Critical Thinking] [Evidence-Based Evaluation] Consider this case from the alternative theoretical positions of behavior modification versus traditional Freudian psychology. (You may substitute some other relevant theories of human behavior with which you are familiar.) What are some major theoretical concepts from each theory that apply to this case?
b. [Research & Practice] For the concepts you identified in part (a), use the illustrative case to develop operational definitions for each concept.
2.3. [Evidence-Based Evaluation] [Research & Practice] Using each theory from Exercise 2.2, construct a possible explanation for the woman's behavior. Now try to state your explanations in terms of testable hypotheses. Can you foresee any problems in assessing causality when testing these hypotheses? Compare the hypotheses you developed with those of other students in terms of the theory used, the concepts selected, and the variables identified. How are the concepts and variables that are derived from the same theory alike? How do they differ from those derived from the other theory?
2.4. [Research & Practice] We have made the point that the same theories can be useful to both practitioners and researchers. Using the hypotheses developed in Exercise 2.3, explain how they could be used either to help the worker change the woman's behavior or to conduct a study. How might the hypotheses need to be changed to be useful in both practice and research?
2.5. [Critical Thinking] [Diversity] A clinician in a treatment program for woman batterers notices that about 75 percent of the individuals who are mandated by the court to participate in group therapy also have a history of substance abuse problems, such as arrests for drunk driving. Furthermore, a large proportion of the clients had been drinking prior to committing the assault. Can the clinician conclude that substance abuse causes woman battering? Identify what conditions would need to be met to support this contention. Prepare a diagram involving a third variable that might show that the relation between substance abuse and woman battering is spurious.