DEBATING ABILITY TESTING

5.1 DEFINITIONS OF INTELLIGENCE Before we discuss definitions of intelligence, we need to clarify the nature of definition itself. Sternberg (1986) makes a distinction between operational and “real” definitions that is important in this context. An operational definition defines a concept in terms of the way it is measured. Boring (1923) carried this viewpoint to its extreme when he defined intelligence as “what the tests test.” Believe it or not, this was a serious proposal, designed largely to short-circuit rampant and divisive disagreements about the definition of intelligence. Operational definitions of intelligence suffer from two dangerous shortcomings (Sternberg, 1986). First, they are circular. Intelligence tests were invented to measure intelligence, not to define it. The test designers never intended for their instruments to define intelligence. Second, operational definitions block further progress in understanding the nature of intelligence,

because they foreclose discussion on the adequacy of theories of intelligence. This second problem—the potentially stultifying effects of relying on operational definitions of intelligence—casts doubt on the common practice of affirming the concurrent validity of new tests by correlating them with old tests. If established tests serve as the principal criterion against which new tests are assessed, then the new tests will be viewed as valid only to the extent that they correlate with the old ones. Such a conservative practice drastically curtails innovation. The operational definition of intelligence does not allow for the possibility that new tests or conceptions of intelligence may be superior to the existing ones. We must conclude, then, that operational definitions of intelligence leave much to be desired. In contrast, a real definition is one that seeks to tell us the true nature of the thing being defined (Robinson, 1950; Sternberg, 1986). Perhaps the most common way—but by no means the only way—of producing real

definitions of intelligence is to ask experts in the field to define it. Expert Definitions of Intelligence Intelligence has been given many real definitions by prominent researchers in the field. In the following, we list several examples, paraphrased slightly for editorial consistency. The reader will note that many of these definitions appeared in an early but still influential symposium, “Intelligence and Its Measurement,” published in the Journal of Educational Psychology (Thorndike, 1921). Other definitions stem from a modern update of this early symposium, What Is Intelligence?, edited by Sternberg and Detterman (1986). Intelligence has been defined as the following: • Spearman (1904, 1923): a general ability

that involves mainly the eduction of relations and correlates.

• Binet and Simon (1905): the ability to judge well, to understand well, to reason well.

• Terman (1916): the capacity to form concepts and to grasp their significance.

• Pintner (1921): the ability of the individual to adapt adequately to relatively new situations in life.

• Thorndike (1921): the power of good responses from the point of view of truth or fact.

• Thurstone (1921): the capacity to inhibit instinctive adjustments, flexibly imagine different responses, and realize modified instinctive adjustments into overt behavior.

• Wechsler (1939): The aggregate or global capacity of the individual to act purposefully, to think rationally, and to deal effectively with the environment.

• Humphreys (1971): the entire repertoire of acquired skills, knowledge, learning sets, and generalization tendencies considered intellectual in nature that are available at any one period of time.

• Piaget (1972): a generic term to indicate the superior forms of organization or equilibrium of cognitive structuring used for

adaptation to the physical and social environment.

• Sternberg (1985a, 1986): the mental capacity to automatize information processing and to emit contextually appropriate behavior in response to novelty; intelligence also includes metacomponents, performance components, and knowledge- acquisition components (discussed later).

• Eysenck (1986): error-free transmission of information through the cortex.

• Gardner (1986): the ability or skill to solve problems or to fashion products that are valued within one or more cultural settings.

• Ceci (1994): multiple innate abilities that serve as a range of possibilities; these abilities develop (or fail to develop, or develop and later atrophy) depending upon motivation and exposure to relevant educational experiences.

• Sattler (2001): intelligent behavior reflects the survival skills of the species, beyond

beyond those associated with basic physiological processes.

The preceding list of definitions is representative although definitely not exhaustive. For one thing, the list is exclusively Western and omits several cross-cultural conceptions of intelligence. Eastern conceptions of intelligence, for example, emphasize benevolence, humility, freedom from conventional standards of judgment, and doing what is right as essential to intelligence. Many African conceptions of intelligence place heavy emphasis on social aspects of intelligence such as maintaining harmonious and stable intergroup relations (Sternberg & Kaufman, 1998). The reader can consult Bracken and Fagan (1990), Sternberg (1994), and Sternberg and Detterman (1986) for additional ideas. Certainly, this sampling of views is sufficient to demonstrate that there appear to be as many definitions of intelligence as there are experts willing to define it! In spite of this diversity of viewpoints, two themes recur again and again in expert

definitions of intelligence. Broadly speaking, the experts tend to agree that intelligence is (1) the capacity to learn from experience and (2) the capacity to adapt to one’s environment. That learning and adaptation are both crucial to intelligence stands out with poignancy in certain cases of mental disability in which persons fail to possess one or the other capacity in sufficient degree (Case Exhibit 5.1). CASE EXHIBIT 5.1 Learning and Adaptation as Core Functions of Intelligence Persons with mental disability often demonstrate the importance of experiential learning and environmental adaptation as key ingredients of intelligence. Consider the case history of a 61-year-old newspaper vendor with moderate mental retardation well known to local mental health specialists. He was an interesting if not eccentric gentleman who stored canned goods in his freezer and cursed at welfare workers who stopped by to see how he was doing. In spite of his need for financial support

from a state agency, he was fiercely independent and managed his own household with minimal supervision from case workers. Thus, in some respects he maintained a tenuous adaptation to his environment. To earn much-needed extra income, he sold a local 25-cent newspaper from a streetside newsstand. He recognized that a quarter was proper payment and had learned to give three quarters in change for a dollar bill. He refused all other forms of payment, an arrangement that his customers could accept. But one day the price of the newspaper was increased to 35 cents, and the newspaper vendor was forced to deal with nickels and dimes as well as quarters and dollar bills. The amount of learning required by this slight shift in environmental demands exceeded his intellectual abilities, and, sadly, he was soon out of business. His failed efforts highlight the essential ingredients of intelligence: learning from experience and adaptation to the environment. How well do intelligence tests capture the experts’ view that intelligence consists of

learning from experience and adaptation to the environment? The reader should keep this question in mind as we proceed to review major intelligence tests in the topics that follow. Certainly, there is cause for concern: Very few contemporary intelligence tests appear to require the examinee to learn something new or to adapt to a new situation as part and parcel of the examination process. At best, prominent modern tests provide indirect measures of the capacities to learn and adapt. How well they capture these dimensions is an empirical question that must be demonstrated through validational research. Layperson and Expert Conceptions of Intelligence Another approach to understanding a construct is to study its popular meaning. This method is more scientific than it may appear. Words have a common meaning to the extent that they help provide an effective portrayal of everyday transactions. If laypersons can agree on its meaning, a construct such as intelligence is in

some sense “real” and, therefore, potentially useful. Thus, asking persons on the street, “What does intelligence mean to you?” has much to recommend it. Sternberg, Conway, Ketron, and Bernstein (1981) conducted a series of studies to investigate conceptions of intelligence held by American adults. In the first study, people in a train station, entering a supermarket, and studying in a college library were asked to list behaviors characteristic of different kinds of intelligence. In a second study—the only one discussed here—both laypersons and experts (mainly academic psychologists) rated the importance of these behaviors to their concept of an “ideally intelligent” person. The behaviors central to expert and lay conceptions of intelligence turned out to be very similar, although not identical. In order of importance, experts saw verbal intelligence, problem-solving ability, and practical intelligence as crucial to intelligence. Laypersons regarded practical problemsolving ability, verbal ability, and social competence to

be the key ingredients in intelligence. Of course, opinions were not unanimous; these conceptions represent the consensus view of each group. In their conception of intelligence, experts place more emphasis on verbal ability than problem solving, whereas laypersons reverse these priorities. Nonetheless, experts and laypersons alike consider verbal ability and problem solving to be essential aspects of intelligence. As the reader will see, most intelligence tests also accent these two competencies. Prototypical examples would be vocabulary (verbal ability) and block design (problem solving) from the Wechsler scales, discussed later. We see then that everyday conceptions of intelligence are, in part, mirrored quite faithfully by the content of modern intelligence tests. Some disagreement between experts and laypersons is also evident. Experts consider practical intelligence (sizing up situations, determining how to achieve goals, awareness and interest in the world) an essential constituent of intelligence, whereas laypersons identify social competence (accepting others for

what they are, admitting mistakes, punctuality, and interest in the world) as a third component. Yet, these two nominations do share one property in common: Contemporary tests generally make no attempt to measure either practical intelligence or social competence. Partly, this reflects the psychometric difficulties encountered in devising test items relevant to these content areas. However, the more influential reason intelligence tests do not measure practical intelligence or social competence is inertia: Test developers have blindly accepted historically incomplete conceptions of intelligence. Until recently, the development of intelligence testing has been a conservative affair, little changed since the days of Binet and the Army Alpha and Beta tests for World War I recruits. There are some signs that testing practices may soon evolve, however, with the development of innovative instruments. For example, Sternberg and colleagues have proposed innovative tests based on his model of intelligence. Another interesting instrument based on a new model of intelligence is the

Everyday Problem Solving Inventory (Cornelius & Caspi, 1987). In this test, examinees must indicate their typical response to everyday problems such as failing to bring money, checkbook, or credit card when taking a friend to lunch. Many theorists in the field of intelligence have relied on factor analysis for the derivation or validation of their theories. In fact, it is not an overstatement to say that perhaps the majority of the theories in this area have been impacted by the statistical tools of factor analysis, which provide ways to portion intelligence into its subcomponents. One of the most compelling theories of intelligence, the Cattell-Horn-Carroll theory reviewed later, would not exist without factor analysis. Thus, before summarizing theories, we provide a brief review of this essential statistical tool.

5.2 A PRIMER OF FACTOR ANALYSIS

Broadly speaking, there are two forms of factor analysis: confirmatory and exploratory. In confirmatory factor analysis, the purpose is to confirm that test scores and variables fit a certain pattern predicted by a theory. For example, if the theory underlying a certain intelligence test prescribed that the subtests belong to three factors (e.g., verbal, performance, and attention factors), then a confirmatory factor analysis could be undertaken to evaluate the accuracy of this prediction. Confirmatory factor analysis is essential to the validation of many ability tests. The central purpose of exploratory factor analysis is to summarize the interrelationships among a large number of variables in a concise and accurate manner as an aid in conceptualization (Gorsuch, 1983). For instance, factor analysis may help a researcher discover that a battery of 20 tests represents only four underlying variables, called factors. The smaller set of derived factors can be used to represent the essential constructs that underlie the complete group of variables.

Perhaps a simple analogy will clarify the nature of factors and their relationship to the variables or tests from which they are derived. Consider the track-and-field decathlon, a mixture of 10 diverse events including sprints, hurdles, pole vault, shot put, and distance races, among others. In conceptualizing the capability of the individual decathlete, we do not think exclusively in terms of the participant’s skill in specific events. Instead, we think in terms of more basic attributes such as speed, strength, coordination, and endurance, each of which is reflected to a different extent in the individual events. For example, the pole vault requires speed and coordination, while hurdle events demand coordination and endurance. These inferred attributes are analogous to the underlying factors of factor analysis. Just as the results from the 10 events of a decathlon may boil down to a small number of underlying factors (e.g., speed, strength, coordination, and endurance), so too may the results from a battery of 10 or 20 ability tests reflect the operation of a small number of basic cognitive

attributes (e.g., verbal skill, visualization, calculation, and attention, to cite a hypothetical list). This example illustrates the goal of factor analysis: to help produce a parsimonious description of large, complex data sets. We will illustrate the essential concepts of factor analysis by pursuing a classic example concerned with the number and kind of factors that best describe student abilities. Holzinger and Swineford (1939) gave 24 ability-related psychological tests to 145 junior high school students from Forest Park, Illinois. The factor analysis described later was based on methods outlined in Kinnear and Gray (1997). It should be intuitively obvious to the reader that any large battery of ability tests will reflect a smaller number of basic, underlying abilities (factors). Consider the 24 tests depicted in Table 5.1. Surely some of these tests measure common underlying abilities. For example, we would expect Sentence Completion, Word Classification, and Word Meaning (variables 7, 8, and 9) to assess a factor of general language ability of some kind. In like manner, other

groups of tests seem likely to measure common underlying abilities—but how many abilities or factors? And what is the nature of these underlying abilities? Factor analysis is the ideal tool for answering these questions. We follow the factor analysis of the Holzinger and Swineford (1939) data from beginning to end. TABLE 5.1 The 24 Ability Tests Used by Holzinger and Swineford (1939)

1.Visual Perception 2.Cubes 3.Paper Form Board 4.Flags 5.General Information 6.Paragraph Comprehension 7.Sentence Completion 8.Word Classification 9.Word Meaning 10.Add Digits 11.Code (Perceptual Speed) 12.Count Groups of Dots 13.Straight and Curved Capitals 14.Word Recognition 15.Number Recognition 16.Figure Recognition 17.Object-Number 18.Number-Figure 19.Figure-Word 20.Deduction 21.Numerical Puzzles 22.Problem Reasoning 23.Series Completion 24.Arithmetic Problems

The Correlation Matrix The beginning point for every factor analysis is the correlation matrix, a complete table of intercor-relations among all the variables.1 The correlations between the 24 ability variables discussed here can be found in Table 5.2. The reader will notice that variables 7, 8, and 9 do, indeed, intercorrelate quite strongly (correlations of .62, .69, and .53), as we suspected earlier. This pattern of intercorrelations is presumptive evidence that these variables measure something in common; that is, it appears that these tests reflect a common underlying factor. However, this kind of intuitive factor analysis based on a visual inspection of the correlation matrix is hopelessly limited; there are just too many intercorrelations for the viewer to discern the underlying patterns for all the variables. Here is where factor analysis can be helpful. Although we cannot elucidate the mechanics of the procedure, factor analysis relies on modern high-speed computers to search the correlation matrix according to

objective statistical rules and determine the smallest number of factors needed to account for the observed pattern of intercorrelations. The analysis also produces the factor matrix, a table showing the extent to which each test loads on (correlates with) each of the derived factors, as discussed in the following section. The Factor Matrix and Factor Loadings The factor matrix consists of a table of correlations called factor loadings. The factor loadings (which can take on values from −1.00 to +1.00) indicate the weighting of each variable on each factor. For example, the factor matrix in Table 5.3 shows that five factors (labeled I, II, III, IV, and V) were derived from the analysis. Note that the first variable, Series Completion, has a strong positive loading of .71 on factor I, indicating that this test is a reasonably good index of factor I. Note also that Series Completion has a modest negative loading of −.11 on factor II, indicating that, to a slight extent, it measures the opposite of this factor;

that is, high scores on Series Completion tend to signify low scores on factor II, and vice versa. TABLE 5.2 The Correlation Matrix for 24 Ability Variables

1 2 3 4 5 6 7 8 9 1011121314151617181920212223 2 32 3 4032 4 472331 5 32292523 6 3423273362 7 301622346672 8 33173839585362 9 3320183372716953 10126 8 103120252917 1131159 11343523302848 123115141622101827115943 13492432333431354028415451 141310187 282924252617351320 1524137 1323251718251524171437 16412726321929183024123112284133 17181 1819212723262729362819343532 183726212526171625213235353221333445 19271131141925232727192911262119263236 2037293034404445434517202524302739263017

Note: Decimals omitted. Source: Reprinted with permission from Holzinger, K., & Harman, H. (1941). Factor analysis: A synthesis of factorial methods. Chicago: University of Chicago Press. Copyright © 1941 The University of Chicago Press. The factors may seem quite mysterious, but in reality they are conceptually quite simple. A factor is nothing more than a weighted linear sum of the variables; that is, each factor is a precise statistical combination of the tests used in the analysis. In a sense, a factor is produced by “adding in” carefully determined portions of some tests and perhaps “subtracting out” fractions of other tests. What makes the factors special is the elegant analytical methods used to derive them. Several different methods exist. These methods differ in subtle ways beyond the scope of this text; the reader can gather a sense of the differences by examining names of

213731173532263136274140364318233517363341 22412325384439403648163019282425282732344637 2347353834444341505026253538242636292730514550 242821202542434439425341413630172633413737453843

procedures: principal components factors, principal axis factors, method of unweighted least squares, maximum-likelihood method, image factoring, and alpha factoring (Tabachnick & Fidell, 1989). Most of the methods yield highly similar results. The factor loadings depicted in Table 5.3 are nothing more than correlation coefficients between variables and factors. These correlations can be interpreted as showing the weight or loading of each factor on each variable. For example, variable 9, the test of Word Meaning, has a very strong loading (.69) on factor I, modest negative loadings (−.45 and −.29) on factors II and III, and negligible loadings (.08 and .00) on factors IV and V. TABLE 5.3 The Principal Axes Factor Analysis for 24 Variables

Factors I II III IV V

23 Series Completion 0.71-0.110.140.110.07 8 Word Classification 0.70-0.24-0.15-0.11-0.13 5 General Information 0.70-0.32-0.34-0.040.08 9 Word Meaning 0.69-0.45-0.290.080.00

Geometric Representation of Factor Loadings

6 Paragraph Comprehension

0.69-0.42-0.260.08-0.01

7 Sentence Completion 0.68-0.42-0.36-0.05-0.05 24 Arithmetic Problems 0.670.20-0.23-0.04-0.11 20 Deduction 0.64-0.190.130.060.28 22 Problem Reasoning 0.64-0.150.110.05-0.04 21 Numerical Puzzles 0.620.240.10-0.210.16 13 Straight and Curved

Capitals 0.620.280.02-0.36-0.07

1 Visual Perception 0.62-0.010.42-0.21-0.01 11 Code (Perceptual Speed) 0.570.44-0.200.040.01 18 Number–Figure 0.550.390.200.15-0.11 16 Figure Recognition 0.530.080.400.310.19 4 Flags 0.51-0.180.32-0.23-0.02

17 Object–Number 0.490.27-0.030.47-0.24 2 Cubes 0.40-0.080.39-0.230.34

12 Count Groups of Dots 0.480.55-0.14-0.330.11 10 Add Digits 0.470.55-0.45-0.190.07 3 Paper Form Board 0.44-0.190.48-0.12-0.36

14 Word Recognition 0.450.09-0.030.550.16 15 Number Recognition 0.420.140.100.520.31 19 Figure–Word 0.470.140.130.20-0.61

It is customary to represent the first two or three factors as reference axes in two- or three- dimensional space.2 Within this framework the factor loadings for each variable can be plotted for examination. In our example, five factors were discovered, too many for simple visualization. Nonetheless, we can illustrate the value of geometric representation by oversimplifying somewhat and depicting just the first two factors (Figure 5.1). In this graph, each of the 24 tests has been plotted against the two factors that correspond to axes I and II. The reader will notice that the factor loadings on the first factor (I) are uniformly positive, whereas the factor loadings on the second factor (II) consist of a mixture of positive and negative.

FIGURE 5.1 Geometric Representation of the First Two Factors from 24 Ability Tests The Rotated Factor Matrix An important point in this context is that the position of the reference axes is arbitrary. There is nothing to prevent the researcher from

rotating the axes so that they produce a more sensible fit with the factor loadings. For example, the reader will notice in Figure 5.1 that tests 6, 7, and 9 (all language tests) cluster together. It would certainly clarify the interpretation of factor I if it were to be redirected near the center of this cluster (Figure 5.2). This manipulation would also bring factor II alongside interpretable tests 10, 11, and 12 (all number tests). Although rotation can be conducted manually by visual inspection, it is more typical for researchers to rely on one or more objective statistical criteria to produce the final rotated factor matrix. Thurstone’s (1947) criteria of positive manifold and simple structure are commonly applied. In a rotation to positive manifold, the computer program seeks to eliminate as many of the negative factor loadings as possible. Negative factor loadings make little sense in ability testing, because they imply that high scores on a factor are correlated with poor test performance. In a rotation to simple structure, the computer program seeks

to simplify the factor loadings so that each test has significant loadings on as few factors as possible. The goal of both criteria is to produce a rotated factor matrix that is as straightforward and unambiguous as possible.

FIGURE 5.2 Geometric Representation of the First Two Rotated Factors from 24 Ability Tests The rotated factor matrix for this problem is shown in Table 5.4. The particular method of rotation used here is called varimax rotation. Varimax should not be used if the theoretical expectation suggests that a general factor may occur. Should we expect a general factor in the analysis of ability tests? The answer is as much a matter of faith as of science. One researcher may conclude that a general factor is likely and, therefore, pursue a different type of rotation. A second researcher may be comfortable with a Thurstonian viewpoint and seek multiple ability factors using a varimax rotation. We will explore this issue in more detail later, but it is worth pointing out here that a researcher encounters many choice points in the process of conducting a factor analysis. It is not surprising, then, that different researchers may reach different conclusions from factor analysis, even when they are analyzing the same data set.

The Interpretation of Factors Table 5.4 indicates that five factors underlie the intercorrelations of the 24 ability tests. But what shall we call these factors? The reader may find the answer to this question disquieting, because at this juncture we leave the realm of cold, objective statistics and enter the arena of judgment, insight, and presumption. In order to interpret or name a factor, the researcher must make a reasoned judgment about the common processes and abilities shared by the tests with strong loadings on that factor. For example, in Table 5.4 it appears that factor I is verbal ability, because the variables with high loadings stress verbal skill (e.g., Sentence Completion loads .86, Word Meaning loads .84, and Paragraph Comprehension loads .81). The variables with low loadings also help sharpen the meaning of factor I. For example, factor I is not related to numerical skill (Numerical Puzzles loads .18) or spatial skill (Paper Form Board loads .16). Using a similar form of inference, it appears that factor II is mainly numerical ability (Add Digits

loads .85, Count Groups of Dots loads .80). Factor III is less certain but appears to be a visual-perceptual capacity, and factor IV appears to be a measure of recognition. We would need to analyze the single test on factor V (Figure-Word) to surmise the meaning of this factor. TABLE 5.4 The Rotated Varimax Factor Matrix for 24 Ability Variables

Factors I II III IV V

7 Sentence Completion 0.860.150.130.030.07 9 Word Meaning 0.840.060.150.180.08 6 Paragraph

Comprehension 0.810.070.160.180.10

5 General Information 0.790.220.160.12-0.02 8 Word Classification 0.650.220.280.030.21

22 Problem Reasoning 0.430.120.380.230.22 10 Add Digits 0.180.85-0.100.09-0.01 12 Count Groups of Dots 0.020.800.200.030.00 11 Code (Perceptual

Speed) 0.180.640.050.300.17

13 Straight and Curved Capitals

0.190.600.40-0.050.18

Note: Boldfaced entries signify subtests loading strongly on each factor. These results illustrate a major use of factor analysis, namely, the identification of a small number of marker tests from a large test battery. Rather than using a cumbersome battery of 24 tests, a researcher could gain nearly the same information by carefully selecting several tests with strong loadings on the five factors. For example, the first factor is well represented by

24 Arithmetic Problems 0.410.540.120.160.24 21 Numerical Puzzles 0.180.520.450.160.02 18 Number-Figure 0.000.400.280.380.36 1 Visual Perception 0.170.210.690.100.20 2 Cubes 0.090.090.650.12-0.18 4 Flags 0.260.070.60-0.010.15 3 Paper Form Board 0.16-0.090.57-0.050.49

23 Series Completion 0.420.240.520.180.11 20 Deduction 0.430.110.470.35-0.07 15 Number Recognition 0.110.090.120.74-0.02 14 Word Recognition 0.230.100.000.690.10 16 Figure Recognition 0.070.070.460.590.14 17 Object-Number 0.150.25-0.060.520.49 19 Figure-Word 0.160.160.110.140.77

test 7, Sentence Completion (.86) and test 9, Word Meaning (.84); the second factor is reflected in test 10, Add Digits (.85), while the third factor is best illustrated by test 1, Visual Perception (.69). The fourth factor is captured by test 15, Number Recognition (.74), and Word Recognition (.69). Of course, the last factor loads well on only test 19, Figure-Word (.77). Issues in Factor Analysis Unfortunately, factor analysis is frequently misunderstood and often misused. Some researchers appear to use factor analysis as a kind of divining rod, hoping to find gold hidden underneath tons of dirt. But there is nothing magical about the technique. No amount of statistical analysis can rescue data based on trivial, irrelevant, or haphazard measures. If there is no gold to be found, then none will be found; factor analysis is not alchemy. Factor analysis will yield meaningful results only when the research was meaningful to begin with. An important point is that a particular kind of factor can emerge from factor analysis only if

the tests and measures contain that factor in the first place. For example, a short-term memory factor cannot possibly emerge from a battery of ability tests if none of the tests requires short- term memory. In general, the quality of the output depends upon the quality of the input. We can restate this point as the acronym GIGO, or “garbage in, garbage out.” Sample size is crucial to a stable factor analysis. Comrey (1973) offers the following rough guide:

In general, it is comforting to have at least five subjects for each test or variable (Tabachnick & Fidell, 1989). Finally, we cannot overemphasize the extent to which factor analysis is guided by subjective choices and theoretical prejudices. A crucial

Sample Size Rating 50 Very poor

100 Poor 200 Fair 300 Good 500 Very good

1,000 Excellent

question in this regard is the choice between orthogonal axes and oblique axes. With orthogonal axes, the factors are at right angles to one another, which means that they are uncorrelated (Figures 5.1 and 5.2 both depict orthogonal axes). In many cases the clusters of factor loadings are situated such that oblique axes provide a better fit. With oblique axes, the factors are correlated among themselves. Some researchers contend that oblique axes should always be used, whereas others take a more experimental approach. Tabachnick and Fidell (1989) recommend an exploratory strategy based on repeated factor analyses. Their approach is unabashedly opportunistic: During the next few runs, researchers

experiment with different numbers of factors, different extraction techniques, and both orthogonal and oblique rotations. Some number of factors with some combination of extraction and rotation produces the solution with the greatest scientific utility, consistency, and meaning; this is the solution that is interpreted.

With oblique rotations it is also possible to factor analyze the factors themselves. Such a procedure may yield one or more second-order factors. Second-order factors can provide support for the hierarchical organization of traits and may offer a rapprochement between ability theorists who posit a single general factor (e.g., Spearman) and those who promote several group factors (e.g., Thurstone). Perhaps both camps are correct, with the group factors sitting underneath the second-order general factor. We turn now to a review of major theories of intelligence. A reminder: The justification for reviewing theories is to illustrate how they have influenced the structure and content of intelligence tests. In addition, the construct validity of IQ tests depends on the extent to which they embody specific theories of intelligence, so a review of theories is pertinent to test validation as well. 1In this example, the variables are tests that produce more or less continuous scores. But the variables in a factor analysis can take other forms, so long as they can be expressed as continuous scores. For example, all of the following could be variables in a

factor analysis: height, weight, income, social class, and rating- scale results. 2Technically, it is possible to represent all the factors as reference axes in n-dimensional space, where n is the number of factors. However, when working with more than two or three reference axes, visual representation is no longer feasible.

5.3 GALTON AND SENSORY KEENNESS

The first theories of intelligence were derived in the Brass Instruments era of psychology at the turn of the twentieth century. The reader will recall from Topic 2A that Sir Francis Galton and his disciple J. McKeen Cattell thought that intelligence was underwritten by keen sensory abilities. This incomplete and misleading assumption was based on a plausible premise: The only information that reaches us concerning

outward events appears to pass through the avenues of our senses; and the more perceptive the senses are of difference, the larger is the field upon which our judgment and intelligence can act. (Galton, 1883)

The sensory keenness theory of intelligence promoted by Galton and Cattell proved to be largely a psychometric dead end. However, we do see vestiges of this approach in modern chronometric analyses of intelligence such as the Reaction Time–Movement Time (RT-MT) apparatus, an experimental method favored by Jensen (1980) for the culture-reduced study of intelligence (Figure 5.3). In RT-MT studies, the subject is instructed to place the index finger of the preferred hand on the home button; then an auditory warning signal is sounded, followed (in 1 to 4 seconds) by one of the eight green lights going on, which the subject must turn off as quickly as possible by touching the microswitch button directly below it. RT is the time the subject takes to remove his or her finger from the home button after a green light goes on. MT is the interval between removing the finger from the home button and touching the button that turns off the green light. Jensen (1980) reported that indices of RT and MT correlated as high as .50 with traditional psychometric tests of intelligence.3 P. A. Vernon has also reported

substantial relationships—as high as .70 for multiple correlations—between speed-of- processing RT-type measures and traditional measures of intelligence (Vernon, 1994). These findings suggest that speed-of-processing measures such as RT might be a useful addition to standardized intelligence test batteries. In general, test developers have resisted the implications of this line of research.

FIGURE 5.3 Schematic Diagram of a Reaction Time—Movement Time Apparatus Note: The square box □ indicates the starting point; the open circles O indicate the signal lights; the dark circles ● indicate the push buttons.

One reason for the lack of ongoing progress in mental chronometry is the absence of standardization in measurement and data analysis. Not all devices for measuring reaction time are the same; consequently, the data from one laboratory cannot be compared to results from another setting. Making matters worse, many “reaction time” devices lump together RT (the time needed to lift the finger off the home button) and MT (the time in transit to the target button), which drastically obscures the relationship between chronometric data and intelligence (Jensen, 2006). The problem with combining the two is that RT is related to IQ, whereas MT is a motor measure uncorrelated with IQ. In addressing these issues, Jensen (2011) has commissioned a leading electronics company to create a standard apparatus for administering and recording reaction time and other indices of mental chronometry. Use of a single standard instrument would provide an vital foundation for progress in this area of assessment.

3Actually, the raw correlation coefficient is negative because faster reaction times (lower numerical scores) are associated with higher intelligence scores.

5.4 SPEARMAN AND THE g FACTOR Based on extensive study of the patterns of correlations between various tests of intellectual and sensory ability, Charles Spearman (1904, 1923, 1927) proposed that intelligence consisted of two kinds of factors: a single general factor g and numerous specific factors s1, s2, s3, and so on. As a necessary adjunct to his theory, Spearman helped invent factor analysis to aid his investigation of the nature of intelligence. Spearman used this statistical technique to discern the number of separate underlying factors that must exist to account for the observed correlations between a large number of tests. In Spearman’s view, an examinee’s performance on any homogeneous test or subtest of intellectual ability was determined mainly by two influences: g, the pervasive general factor,

and s, a factor specific to that test or subtest. (An error factor e could also sway scores, but Spearman sought to minimize this influence by using highly reliable instruments.) Because the specific factor s was different for each intellectual test or subtest and was usually less influential than g in determining performance level, Spearman expressed less interest in studying it. He concentrated mainly on defining the nature of g, which he likened to an “energy” or “power” that serves in common the whole cortex. In contrast, Spearman considered s, the specific factor, to have a physiological substrate localized in the group of neurons serving the particular kind of mental operation demanded by a test or subtest. Spearman (1923) wrote, “These neural groups would thus function as alternative ‘engines’ into which the common supply of ‘energy’ could be alternatively distributed.” Spearman reasoned that some tests were heavily loaded with the g factor, whereas other tests— especially purely sensory measures—were representative mainly of a specific factor. Two

tests each heavily loaded with g should correlate quite strongly. In contrast, psychological tests not saturated with g should show minimal correlation with one another. Much of Spearman’s research was aimed at demonstrating the truth of these basic propositions derived from his theory. We have illustrated these points graphically in Figure 5.4. In this figure, each circle represents an intelligence test, and the degree of overlap between circles indicates the strength of correlation. Notice that tests A and B, each heavily loaded on g, correlate quite strongly. Tests C and D have weak loadings on g and subsequently do not correlate well.

FIGURE 5.4 Spearman’s Two-Factor Theory of Intelligence

Note: Tests A and B correlate strongly, whereas C and D correlate weakly. See text. Spearman (1923) believed that individual differences in g were most directly reflected in the ability to use three principles of cognition: apprehension of experience, eduction of relations, and eduction of correlations. Incidentally, the little-used term eduction refers to the process of figuring things out. These three principles can be explained by examining how we solve analogies of the form A:B::C:? that is, A is to B as C is to? A simple example might be HAMMER:NAIL::SCREWDRIVER:? To solve this analogy, we must first perceive and understand each term based on past experience; that is, we must have apprehension of experience. If we have no idea what a hammer, nail, and screwdriver are, there is little chance we can complete the analogy correctly. Next, we must infer the relation between the first two analogy terms, in this case, HAMMER and NAIL. Using a somewhat stilted phrase, Spearman referred to the ability to infer the relation between two concepts as eduction of

relations. The final step, eduction of correlates, refers to the ability to apply the inferred principle to the new domain, in this case, applying the rule inferred to produce the correct response, namely, SCREWDRIVER:SCREW. Although Spearman’s physiological speculations have been largely dismissed, the idea of a general factor has been a central topic in research on intelligence and is still very much alive today (Jensen, 1979). The correctness of the g factor viewpoint is more than an academic issue. If it is true that a single, pervasive general factor is the essential wellspring of intelligence, then psychometric efforts to produce factorially pure subtests (e.g., measuring verbal comprehension, perceptual organization, short- term memory, and so on) are largely misguided. To the extent that Spearman is correct, test developers should forgo subtest derivation and concentrate on producing a test that best captures the general factor. The most difficult issue faced by Spearman’s two-factor theory is the existence of group factors. As early as 1906, Spearman and his

contemporaries noted that relatively dissimilar tests could have correlations higher than the values predicted from their respective g loadings (Brody & Brody, 1976). This finding raised the possibility that a group of diverse measures might share in common a unitary ability other than g. For example, several tests might share a common unitary memorization factor that was halfway between the g factor and the various s factors unique to each test. Of course, the existence of group factors is incompatible with Spearman’s meticulous two-factor theory.

5.5 THURSTONE AND THE PRIMARY MENTAL ABILITIES Thurstone (1931) developed factor-analysis procedures capable of searching correlation matrices for the existence of group factors. His methods permitted a researcher to discover empirically the number of factors present in a matrix and to define each factor in terms of the

tests that loaded on it. In his analysis of how scores on different kinds of intellectual tests correlated with each other, Thurstone concluded that several broad group factors—and not a single general factor—could best explain empirical results. At various points in his research career, he proposed approximately a dozen different factors. Only seven of these factors have been frequently corroborated (Thurstone, 1938; Thurstone & Thurstone, 1941) and they have been designated primary mental abilities (PMAs). They are as follows: • Verbal Comprehension: The best measure is

vocabulary, but this ability is also involved in reading comprehension and verbal analogies.

• Word Fluency: Measured by such tests as anagrams or quickly naming words in a given category (e.g., foods beginning with the letter S).

• Number: Virtually synonymous with the speed and accuracy of simple arithmetic computation.

• Space: Such as the ability to visualize how a three-dimensional object would appear if it was rotated or partially disassembled.

• Associative Memory: Skill at rote memory tasks such as learning to associate pairs of unrelated items.

• Perceptual Speed: Involved in simple clerical tasks such as checking for similarities and differences in visual details.

• Inductive Reasoning: The best measures of this factor involve finding a rule, as in a number series completion test.

Thurstone (1938) published the Primary Mental Abilities Test consisting of separate subtests, each designed to measure one PMA. However, he later acknowledged that his primary mental abilities correlated moderately with each other, proving the existence of one or more second- order factors. Ultimately, Thurstone acknowledged the existence of g as a higher- order factor. By this time, Spearman had admitted the existence of group factors representing special abilities, and it became apparent that the differences between Spearman

and Thurstone were largely a matter of emphasis (Brody & Brody, 1976). Spearman continued to believe that g was the major determinant of correlations between test scores and assigned a minor role to group factors. Thurstone reversed these priorities. P. E. Vernon (1950) provided a rapprochement between these two viewpoints by proposing a hierarchical group factor theory. In his view, g was the single factor at the top of a hierarchy that included two major group factors labeled verbal-educational (V:ed) and practical- mechanical-spatial-physical (k:m). Underneath these two major group factors were several minor group factors resembling the PMAs of Thurstone; specific factors occupied the bottom of the hierarchy. Thurstone’s analysis of PMAs continues to influence test development even today. Schaie (1985) has revised and modified the Primary Mental Abilities Test and used these measures in an enormously influential longitudinal study of adult intelligence. If intelligence were mainly a matter of g, then the group factors should

change at about the same rate with aging. In support of the group factor approach to intellectual testing, Schaie (1985) reports that some PMAs show little age-related decrement (Verbal Comprehension, Word Fluency, Inductive Reasoning), whereas other PMAs decline more rapidly in old age (Space, Number). Thus, there may be practical real- world reasons for reporting group factors and not condensing all of intelligence into a single general factor.

5.6 CATTELL-HORN- CARROLL (CHC) THEORY Raymond Cattell (1941, 1971) proposed an influential theory of the structure of intelligence that has been revised and extended by John Horn (1968, 1994) and John Carroll (1993). Based on the reanalysis of 461 data sets from hundreds of independent studies published by other researchers, Carroll’s contributions to the theory are especially vital. The ensuing theory, known as Cattell-Horn-Carroll (CHC) theory, is

a taxonomic tour de force that synthesizes the findings from almost a century of factor-analytic research on intelligence. Many psychometricians consider CHC theory to possess the strongest empirical foundation of any theory of intelligence and also to provide the most far-reaching implications for psychological testing (McGrew, 1997). Although the “big picture” of CHC theory is well established, researchers continue to refine the details. Under the direction of Kevin McGrew, the Institute for Applied Psychometrics manages an informative website dedicated to the advancement of CHC theory and applications (www.iapsych.com).

FIGURE 5.5 Outline of the CHC Three- Stratum Theory of Cognitive Abilities Source: Based on Carroll, J. B. (1993). Cognitive abilities: A survey of factor analytic studies. New York: Cambridge University Press, and table 3 from www.iapsych.com. According to CHC theory, intelligence consists of pervasive, broad, and narrow abilities that are hierarchically organized. These are known as Stratum III, II, and I, respectively (Figure 5.5). At the highest and most pervasive level called Stratum III, a single general factor known as little g oversees all cognitive activities. Stratum II capacities, which reside beneath general intelligence, include several prominent and well-established abilities. In Figure 5.5, we have depicted eight abilities originally identified by Carroll (1993), but other researchers have proposed a slightly larger list that includes additional tentative entries such as psychomotor, olfactory, and kinesthetic abilities. The precise name given to each broad factor differs slightly from one theorist to another, as well as the scale abbreviations. Even so, there is strong

consensus for the essential list. These broad factors include “basic constitutional and longstanding characteristics of individuals that can govern or influence a great variety of behaviors in a given domain” (Carroll, 1993, p. 634). The narrow abilities at Stratum I include approximately 70 abilities identified by Carroll (1993) in his comprehensive review of factor- analytic studies of intelligence. As might be expected, the list of narrow abilities is continually revised and expanded with ongoing research. These narrow abilities “represent greater specializations of abilities, often in quite specific ways that reflect the effects of experience and learning, or the adoption of particular strategies of performance” (Carroll, 1993, p. 634). Definitions of CHC Broad Ability Factors As noted, the broad factors of CHC are more firmly established than the narrow abilities, which continue to undergo revision and extension. We provide brief definitions of the

broad factors, based on Carroll (1993), McGrew (1997), and www.iapsych.com. • Fluid Intelligence/Reasoning (Gf): Fluid

intelligence encompasses high-level reasoning and is used for novel tasks that cannot be performed automatically. The mental operations of fluid intelligence may involve drawing inferences, forming concepts, generating and testing hypotheses, understanding implications, inductive reasoning, and deductive reasoning. The classic example of fluid intelligence is found in matrix reasoning tasks such as Raven’s Progressive Matrices (Raven, 2000). The abilities that make up fluid intelligence are largely nonverbal and not heavily dependent on exposure to a specific culture. For these reasons, Cattell (1940) believed that measures of fluid intelligence were culture- free. Based on this assumption, he devised the Culture Fair Intelligence Test (CFIT) in an attempt to eliminate cultural bias in testing. Of course, calling a test culture fair does not make it necessarily so. In fact, the

goal of a completely culture-free intelligence test has proved elusive. We discuss the CFIT in more detail in Topic 6A, Group Tests of Ability and Related Concepts.

• Crystallized Intelligence/Knowledge (Gc): This form of intelligence is typically defined as an individual’s breadth and depth of acquired cultural knowledge—knowledge of the language, information, and concepts of a person’s culture. The quintessential example is the extent of vocabulary that an individual understands. But crystallized intelligence also includes the application of verbal and cultural knowledge (e.g., oral production, verbal fluency, and communication ability). Because crystallized intelligence arises when fluid intelligence is applied to cultural products, we would expect these two kinds of cognitive ability to possess a strong correlation. In fact, it is commonly found that measures of crystallized and fluid intelligence possess a healthy relationship (r = .5).

• Domain-Specific Knowledge (Gkn): Domain-specific knowledge represents a person’s acquired knowledge in one or more specialized domains that do not represent the typical experiences of individuals in the culture. This might include, for example, knowledge of biology, skill in lip reading, or knowledge of how to use computers.

• Visual-Spatial Abilities (Gv): This ability has to do with imagining, retaining, and transforming mental representations of visual images. For example, visual-spatial ability involves the capacity to predict how a shape will appear when it is rotated, or to identify quickly a known object from a vague, incomplete picture, or to find an object hidden in a picture. This capacity includes visual memory.

• Auditory Processing (Ga): This is the ability to perceive auditory information accurately, which involves the capacity to analyze, comprehend, and synthesize patterns or groups of sounds. Auditory processing involves the ability to discriminate speech

sounds and to judge and discriminate tonal patterns in music. A key characteristic of Ga abilities is the cognitive talent needed to control the perception of auditory information (i.e., to filter signal from noise).

• Broad Retrieval [Memory] (Gr): Broad retrieval includes the ability to consolidate and store new information in long-term memory and then to retrieve the information later through association. Included in broad retrieval are such narrow abilities as associative memory (e.g., when provided the first part, recalling the second part of a previously learned but unrelated pair of items), ideational fluency (e.g., ability to call up ideas), and naming facility (e.g., rapidly providing the names of familiar faces). Some researchers further divide the broad memory factor into additional subtypes. In addition, some theorists propose a separate broad factor for short-term memory (Gsm), the ability to retain awareness of events that have occurred in the last minute or less (Horn & Masunaga, 2000).

• Cognitive Processing Speed (Gs): This ability refers to the speed of executing overlearned or automatized cognitive processes, especially when high levels of attention and focused concentration are required. For example, the ability to perform simple arithmetic calculations with lightning speed would indicate a high level of Gs ability.

• Decision/Reaction Time or Speed (Gt): This is the ability to make decisions quickly in response to simple stimuli, typically measured by reaction time. For example, the capacity to quickly press the space bar whenever the letter X appears on a computer screen would involve the use of Gt ability.

Utility of CHC Theory CHC theory is unusual in its detail, which permits robust theory testing. A number of lines of evidence support its validity. For example, the structure of intelligence as posited by CHC theory has been shown to be invariant across a number of key variables, including age,

ethnicity, and gender (Bickley, Keith, & Wolfe, 1995; Keith, 1999; Carroll, 1993). In empirical studies, the broad CHC abilities also reveal theory-confirming relationships with numerous academic and occupational variables (McGrew & Flanagan, 1998). In one study, for example, measures of CHC broad and narrow cognitive abilities were selectively and appropriately related to mathematics achievement in a representative sample of children and adolescents (Floyd, Evans, & McGrew, 2003). In general, practitioners praise the CHC approach to partitioning intelligence because the broad and narrow abilities are empirically verified and possess meaningful real-world implications (Fiorello & Primerano, 2005).

5.7 GUILFORD AND THE STRUCTURE-OF-INTELLECT MODEL After World War II, J. P. Guilford (1967, 1985) continued the search for the factors of

intelligence that had been initiated by Thurstone. Guilford soon concluded that the number of discernible mental abilities was far in excess of the seven proposed by Thurstone. For one thing, Thurstone had ignored the category of creative thinking entirely, an unwarranted oversight in Guilford’s view. Guilford also found that if innovative types of tests were included in the large batteries of tests he administered his subjects, then the pattern of correlations between these tests indicated the existence of literally dozens of new factors of intellect. Furthermore, Guilford noticed that some of these new factors had recurring similarities with respect to the kinds of mental processes involved, the kinds of information featured, or the form that the items of information took. As a result of these recurring similarities in the newly discovered factors of intellect, he became convinced that these multitudinous factors could be grouped along a small number of main dimensions. Guilford (1967) proposed an elegant structure-of-intellect (SOI) model to summarize his findings. Visually

conceived, Guilford’s SOI model classifies intellectual abilities along three dimensions called operations, contents, and products. By operations, Guilford has in mind the kind of intellectual operation required by the test. Most test items emphasize just one of the operations listed here:

Contents refers to the nature of the materials or information presented to the examinee. The five content categories are as follows:

Cogni tion

Discovering, knowing, or comprehending

Mem ory

Committing items of information to memory, such as a series of numbers

Diver gent

Retrieving from memory items of production a specific class, such as naming objects that are both hard and edible

Conv ergent

Retrieving from memory a correct production item, such as a crossword puzzle word

Evalu ation

Determining how well a certain item of information satisfies specific logical requirements

The third dimension in Guilford’s model, products, refers to the different kinds of mental structures that the brain must produce to derive a correct answer. The six kinds of products are as follows:

Visua l

Images presented to the eyes

Audit ory

Sounds presented to the ears

Symb olic

Such as mathematical symbols that stand for something

Sema ntic

Meanings, usually of word symbols

Beha vioral

The ability to comprehend the mental state and behavior of other persons

Unit A single entity having a unique combination of properties or attributes

Class What it is that similar units have in common, such as a set of triangles or high-pitched tones

Relation An observed connection between two items, such as two tones an octave apart

In total, then, Guilford (1985) identified five types of operations, five types of content, and six types of products, for a total of 5 × 5 × 6 or 150 factors of intellect. Each combination of an operation (e.g., memory), a content (e.g., symbolic), and a product (e.g., units) represents a different factor of intellect. Guilford claims to have verified over 100 of these factors in his research. The SOI model is often lauded on the grounds that it captures the complexities of intelligence. However, this is also a potential Achilles’ heel for the theory. Consider one factor of intellect, memory for symbolic units. A test that requires the examinee to recall a series of spoken digits (e.g., Digit Span on the WAIS-III) might capture

System Three or more items forming a recognizable whole, such as a melody or a plan for a sequence of actions

Transfor mation

A change in an item of information, such as a correction of a misspelling

Implicat ion

What an individual item implies, such as to expect thunder following lightning

this factor of intellect quite well. But so might a visual digit span test and perhaps even an analogous test with tactile presentation of symbols, such as vibrating rods applied to the skin. Perhaps we need a separate cube for hearing, vision, and touch; such an expanded model would incorporate 450 factors of intellect, surely an unwieldy number. Although it seems doubtful that intelligence could involve such a large number of unique abilities, Guilford’s atomistic view of intellect nonetheless has caused test developers to rethink and widen their understanding of intelligence. Prior to Guilford’s contributions, most tests of intelligence required mainly convergent production—the construction of a single correct answer to a stimulus situation. Guilford raised the intriguing possibility that divergent production—the creation of numerous appropriate responses to a single stimulus situation—is also an essential element of intelligent behavior. Thus, a question such as “List as many consequences as possible if clouds had strings hanging down from them”

(divergent production) might assess an aspect of intelligence not measured by traditional tests.

5.8 PLANNING, ATTENTION, SIMULTANEOUS, AND SUCCESSIVE (PASS) THEORY Some modern conceptions of intelligence owe a debt to the neuropsychological investigations of the Russian psychologist Aleksandr Luria (1902–1977). Luria (1966) relied primarily on individual case studies and clinical observations of brain-injured soldiers to arrive at a general theory of cognitive processing. The heart of his theory is as follows: Analysis shows that there is strong evidence for

distinguishing two basic forms of integrative activity of the cerebral cortex by which different aspects of the outside world may be reflected. . . . The first of these forms is the integration of the individual stimuli arriving in the brain into simultaneous, and primarily spatial groups,

and the second is the integration of individual stimuli arriving consecutively in the brain into temporally organized, successive series. (Luria, 1966)

Since this approach focuses upon the mechanics by which information is processed, it is often called an information processing theory. Luria (1970) proposed three functional units in the brain. Processing of information proceeds from lower units to higher units. The first unit is found in subcortical areas including the brain stem, midbrain, and thalamus. Attentional processes originate here, including selective attention and resistance to distraction. The second unit consists of the rearward sensory portions of the cerebral cortex (parietal, temporal, and occipital lobes). This large unit subserves the simultaneous and successive processes discussed later in this chapter. These processes are to some extent lateralized, with simultaneous processing engaged more with the right hemisphere, and successive processing connected more with the left hemisphere. However, lateralization is relative, not absolute

(Springer & Deutsch, 1997). The third unit is located in the frontal lobes. This is primarily where planning occurs and also where motor output initiates. Naglieri and Das (1990, 2005) have developed the Planning, Attention, Simultaneous, Successive (PASS) theory of intelligence as a modern extension of Luria’s work. Planning involves the selection, usage, and monitoring of effective solutions to problems. Anticipation of consequences and use of feedback are essential. Planning also entails impulse control. As noted, the frontal lobes are heavily engaged in this process. Even though it is listed first in the PASS acronym, Planning is actually the last stage of information processing. The first process is Attention, which requires selectively attending to some stimuli while ignoring others. In some cases, attention also entails vigilance over a period of time. Difficulties with this process underlie attention deficit/hyperactivity disorder. As noted, the brain stem and other midline subcortical structures are vital to attentional processes.

Simultaneous processing of information is characterized by the execution of several different mental operations simultaneously. Forms of thinking and perception that require spatial analysis, such as drawing a cube, require simultaneous information processing. In drawing, the examinee must simultaneously apprehend the overall shape and guide hand and fingers in the execution of the shape. A sequential approach to drawing a cube (if one were even possible) would be horrifically complex. In effect, the examinee would have to draw individual lines of highly specific lengths and angular orientations, and just hope that everything would line up. In the absence of a simultaneous mental gestalt to guide the drawing, a distorted production is almost guaranteed. Successive processing of information is needed for mental activities in which a proper sequence of operations must be followed. This is in sharp contrast to simultaneous processing (such as drawing), for which sequence is unimportant. Successive processing is needed in remembering a series of

digits, repeating a string of words (e.g., shoe, ball, egg), and imitating a series of hand movements (fist, palm, fist, fist, palm). Most forms of information processing require an interplay of simultaneous and successive mechanisms. Das (1994) cites the example of reading an unfamiliar word such as taciturn: The single letters are to be recognized, and that

involves simultaneous coding. The reader matches the visual shape of the letter with a mental dictionary and comes up with a name for it. The letter sequences, then, have to be formed (successive coding) and blended together as a syllable (simultaneous). Then the string of syllables has to be made into a word (successive), the word is recognized (simultaneous), and a pronunciation program is then assembled (successive), leading to oral reading (successive and simultaneous).

Das admits that this may be a simplified view of what occurs when a reader is confronted with a word. The essential point is that higher-level information processing relies upon an interplay

of specific, anatomically localizable forms of information processing. The challenge of a simultaneous-successive approach to the assessment of intelligence is to design tasks that tap relatively pure forms of each approach to information processing. Tests that use this strategy are the Kaufman Assessment Battery for Children II (K-ABC-II), discussed in the next topic, and the Das-Naglieri Cognitive Assessment System (Das & Naglieri, 2012). The Das-Naglieri battery includes successive tasks that involve rapid articulation (such as, “Say can, ball, hot as fast as you can 10 times”) and simultaneous measures of both verbal and nonverbal tasks. The battery also assesses planning and attention, so as to embody the PASS theory (Naglieri & Das, 2005).

5.9 INFORMATION PROCESSING THEORIES OF INTELLIGENCE

Information processing conceptions of intelligence propose models of how individuals mentally represent and process information. Borrowing from Campione and Brown (1978), Borkowski (1985) has put forward a comprehensive theory that bears a loose analogy to the functioning of a computer. The architectural system (hardware) refers to biologically based properties necessary for information processing, such as memory span and speed of encoding/decoding information. Properties of the architectural system include capacity (e.g., number of slots in short-term memory, capacity of long-term memory), durability (rate of information loss), and efficiency of operation (e.g., rate of memory search). The architectural system is considered to be relatively “hardwired” and impervious to change by the environment. In addition to the structural component of intelligence, there are various functional components (software). The executive system, which refers to environmentally learned components that steer problem solving, provides

overall guidance to the functional components. Elements of the executive system include the knowledge base (retrieval of knowledge from long-term memory), schemes (rules of thinking), control processes (rules and strategies such as self-checking and rehearsal), and metacognition (self-awareness of one’s own thought processes). Metacognition is the process of thinking about thinking. Flavell (1976), who pioneered research on this topic, explained it as follows: Metacognition refers to one’s knowledge

concerning one’s own cognitive processes or anything related to them, e.g., the learning-relevant properties of information or data. For example, I am engaging in metacognition if I notice that I am having more trouble learning A than B; if it strikes me that I should double check C before accepting it as fact. (p. 232)

The information processing approach to intelligence has generated a large body of research, especially on the concept of metacognition. A consistent finding in this

literature is that individuals who use meta- cognitive strategies perform at much higher levels than those who do not (Montague & Bos, 1990). For example, in a study of 32 Israeli kindergarten children who were taught metacognition related to mathematics, metacognitive skills explained more of the variance in mathematics performance than general ability (Mevarech, 1995). Metacognition is essential to intelligence and is one of the primary influences on student learning (Wang, Haertel, & Walberg, 1990).

5.10 GARDNER AND THE THEORY OF MULTIPLE INTELLIGENCES Howard Gardner (1983, 1993) has proposed a theory of multiple intelligences based loosely on the study of brain–behavior relationships. He argues for the existence of several relatively independent human intelligences, although he admits that the exact nature, extent, and number

of the intelligences have not yet been definitively established. Gardner (1983) outlines the criteria for an autonomous intelligence as follows: • Potential isolation by brain damage—the

faculty can be destroyed, or spared in isolation, by brain injury.

• Existence of exceptional individuals such as savants—the faculty is uniquely spared in the midst of general intellectual mediocrity.

• Identifiable core operations—the faculty relies upon one or more basic information processing operations.

• Distinctive developmental history—the faculty possesses an identifiable developmental history, perhaps including critical periods and milestones.

• Evolutionary plausibility—admittedly speculative, a faculty should have evolutionary antecedents shared with other organisms (e.g., primate social organization).

• Support from experimental psychology—the faculty emerges in laboratory studies in cognitive psychology.

• Support from psychometric findings—the faculty reveals itself in measurement studies and is susceptible to psychometric measurement.

• Susceptibility to symbol encoding—the faculty can be communicated via symbols including (but not limited to) language, picturing, and mathematics.

Based on these criteria, Gardner (1983, 1993) proposes that the following seven natural intelligences have been substantially confirmed. The seven intelligences are linguistic, logical- mathematical, spatial, musical, bodily- kinesthetic, interpersonal, and intrapersonal. Three of these seven types of intelligence are well known—linguistic (i.e., verbal) intelligence, logical-mathematical intelligence, spatial intelligence—and numerous formal tests have been devised to measure them, so we will not discuss them further here. The other four variations of intelligence are somewhat novel

and, therefore, require more detailed presentation. Bodily-kinesthetic intelligence includes the types of skills used by athletes, dancers, mime artists, typists, or “primitive” hunters. Although Western cultures are generally loath to consider the body as a form of intelligence, this is not the case in much of the rest of the world, nor was it true in our evolutionary history. Indeed, persons who could skillfully avoid predators, climb trees, hunt animals, and prepare tools were more likely to survive and pass on their genes to succeeding generations. The personal intelligences include the capacity to have access to one’s own feeling life (intrapersonal) as well as the ability to notice and make distinctions about the moods, temperaments, motivations, and intentions of others (interpersonal). Thus, personal intelligence encompasses both an intrapersonal and an interpersonal version. The former is found in great novelists who can write introspectively about their feelings, while the latter is often seen in religious and political

leaders (e.g., Mahatma Gandhi or Lyndon Johnson) who can fathom the intentions and desires of others and use this information to influence them and form useful alliances. Musical intelligence is perhaps the least understood of Gardner’s intelligences. Persons with good musical intelligence easily learn to perform an instrument or to write their own compositions. Although knowledge of the structural aspects of melody, rhythm, and timbre is important to musical intelligence, Gardner notes that many experts place the affective or feeling aspects of music at its core. He believes that when the neurological underpinnings of music are finally unraveled, we will have “an explanation of how emotional and motivational factors are intertwined with purely perceptual ones” (Gardner, 1983). The savant phenomenon provides strong support for the existence of separate intelligences, including musical intelligence.4 A savant is a mentally deficient individual who has a highly developed talent in a single area such as art, rapid calculation, memory, or music. An

example is the extraordinary case of Leslie Lemke, who was born blind and with mental retardation and cerebral palsy. He was not supposed to live. His adoptive mother had to coax him to suck milk from a bottle. Later, she strapped him to her back to help him learn to walk. In spite of his severe disabilities, Leslie became enamored of the piano and showed incredible precocity at picking out melodies on it. Within a few years, at the age of 18, he could listen to a piece of classical piano music a single time and then play it back flawlessly (Patton, Payne, & Beirne-Smith, 1986). The reader can find additional savant case studies in Miller (1989) and Treffert (1989). Recently, Gardner (1998) has added three tentative candidates to his list of intelligences. These are naturalistic, spiritual, and existential intelligences. Naturalistic intelligence is the kind shown by people who are able to discern patterns in nature. Charles Darwin would be a prime example of such a person. Gardner believes that the evidence for this kind of intelligence is relatively strong. In contrast,

spiritual intelligence (a concern with cosmic and spiritual issues in one’s development) and existential intelligence (a concern with ultimate issues, including the meaning of life) are less well proved as independent intelligences. In general, the theory of multiple intelligence is compelling in its simplicity, but there is little empirical investigation of its validity. 4Historically, savants have also been called idiot savants, which refers, literally, to a person who is both profoundly retarded and yet “wise” at the same time. For obvious reasons, the prefix has been dropped.

5.11 STERNBERG AND THE TRIARCHIC THEORY OF SUCCESSFUL INTELLIGENCE Sternberg (1985b, 1986, 1996) takes a much wider view on the nature of intelligence than most previous theorists. In addition to proposing that certain mental mechanisms are required for intelligent behavior, he also emphasizes that intelligence involves adaptation to the real-

world environment. His theory emphasizes what he calls successful intelligence or “the ability to adapt to, shape, and select environments to accomplish one’s goals and those of one’s society and culture” (Sternberg & Kaufman, 1998, p. 494). Sternberg’s theory is called triarchic (ruled by three) because it deals with three aspects of intelligence: componential intelligence, experiential intelligence, and contextual intelligence. Each of these types of intelligence has two or more subcomponents. The entire theory is outlined in Table 5.5. Componential intelligence, also known as analytical intelligence, consists of the internal mental mechanisms that are responsible for intelligent behavior. The components of intelligence serve three different functions. Metacomponents are the executive processes that direct the activities of all the other components of intelligence. They are responsible for determining the nature of an intellectual problem, selecting a strategy for solving it, and making sure that the task is

completed. The metacomponents receive constant feedback as to how things are going in problem solving. Persons who are strong on the metacomponential aspect of intelligence are very good at allocating their intellectual resources. TABLE 5.5 An Outline of Sternberg’s Triarchic Theory of Intelligence

Source: Summarized from Sternberg, R. J. (1986). Intelligence applied: Understanding and increasing

Componential (Analytical) Intelligence Metacomponents or executive processes (e.g.,

planning) Performance components (e.g., syllogistic

reasoning) Knowledge acquisition components (e.g.,

ability to acquire vocabulary words) Experiential (Creative) Intelligence Ability to deal with novelty Ability to automatize information processing Contextual (Practical) Intelligence Adaptation to real-world environment Selection of a suitable environment Shaping of the environment

your intellectual skills. San Diego, CA: Harcourt Brace Jovanovich. In a problem-solving study using novel forms of analogies, Sternberg (1981) found that higher intelligence is associated with spending relatively more time on global or higher-order planning, and relatively less time on local or lower-order planning. For example, consider this analogy problem:

Man: Skin:: (Dog, Tree):(Bark, Cat) The examinee must choose the two correct terms on the right that will complete the analogy. (The correct choices are Tree and Bark.) Using reaction time measures for a series of such novel or nonentrenched problems, Sternberg (1981) found that persons of higher intelligence spend more time in global planning —forming a macrostrategy that applies to this and similar problems—than did persons of lower intelligence. Thus, a crucial aspect of intelligence is knowing when to step back and allocate intellectual effort instead of obtusely attacking a difficult problem.

Performance components are the well- entrenched mental processes that might be used to perform a task or solve a problem. These aspects of intelligence are the ones that are probably measured the best by existing intelligence tests. Examples of performance components include short-term memory and syllogistic reasoning. Knowledge acquisition components are the processes used in learning. Sternberg has emphasized that in order to understand what makes some people more skilled than others, we must understand their increased capacity to acquire those skills in the first place. A case in point is vocabulary knowledge, which is learned mainly in context rather than through direct instruction. More-intelligent persons are better able to use surrounding contexts to figure out what a word means; that is, they have greater knowledge-acquisition skills. Their increased vocabulary results, in large measure, from their increased ability to “soak up” the meanings of words they see and hear in their environment. Thus, vocabulary is an excellent measure of

intelligence because it reflects people’s ability to acquire information in context. The second aspect of Sternberg’s theory involves experiential intelligence. According to the theory, a person with good experiential intelligence is able to deal effectively with novel tasks. Experiential intelligence is also known as creative intelligence. This aspect of his theory explains why Sternberg is so critical of most intelligence tests. For the most part, the existing tests measure things already learned by presenting tasks that the subject has already encountered. According to Sternberg, intelligence also involves the capacity to learn and think within new conceptual systems, not just to deal with tasks already encountered. A second aspect of experiential intelligence is the ability to automatize or “make routine” tasks that are encountered repeatedly. An example of automatizing that applies to most of us is reading, which is carried out largely without conscious thought. But any task or mental skill can be automatized, if it is practiced enough. Playing music is an example of an extremely

high-level skill that can become automatized with enough practice. The third aspect of Sternberg’s theory involves contextual intelligence. Contextual intelligence, also known as practical intelligence, is defined as “mental activity involved in purposive adaptation to, shaping of, and selection of real-world environments relevant to one’s life” (Sternberg, 1986, p. 33). This aspect of Sternberg’s theory appears to acknowledge that human behavior has been shaped by selective pressures during our evolutionary history. Contextual intelligence has three parts: adaptation, selection, and shaping. Adaptation refers to developing skills required by one’s particular environment. Successful adaptation will differ from one culture to the next. In the pygmy cultures of Africa, adaptation might involve the ability to track elephants and kill them with poison-tipped spears. In the Western industrial nations, adaptation might involve presenting oneself favorably in a job interview.

Selection might be called niche finding. This aspect of contextual intelligence involves the ability to leave the environment we are in and to select a different environment more suitable to our talents and needs. Feldman (1982) has illustrated how selection can operate in the career choices of gifted children, thereby determining whether they are highly accomplished as adults. She followed up on the Quiz Kids who were featured in radio and television shows of the 1950s. These were extremely bright children by conventional standards, most with IQs of 140 and higher. A few became highly successful as adults. However, most of them led rather ordinary lives, devoid of the spectacular accomplishments that might have been predicted from their childhood precocity. Those who were most successful had found occupations highly suited to their abilities and interests. In sum, they had selected environmental niches that fitted them well. Sternberg would argue that the ability to select such environments is an important aspect of intelligence.

Shaping is another way to improve the fit between oneself and the environment, especially when selection of a new environment is not practical. In this application of contextual intelligence, we shape the environment itself so that it better fits our needs. An employee who convinces the boss to do things differently has used shaping to make the work environment more suited to his or her talents. Sternberg (1993) has developed a research instrument based on his theory and has used the test to examine the validity of the triarchic approach. The Sternberg Triarchic Abilities Test (STAT) is unique in going beyond the typical questions that invoke analytical intelligence; the test includes creative and practical questions as well. For example, in one subtest examinees are presented with a map of an area, such as an entertainment park, and then must answer questions about navigating effectively through the area shown in the map (practical intelligence). In another subtest examinees are presented with verbal analogies preceded by incorrect, counterfactual premises (e.g., money

falls off trees). Examinees must solve the analogies as though the counterfactual premises were true (creative intelligence). In factor- analytic studies of American, Finish, and Spanish samples, the triarchic model was a better fit to the data than the usual outcome of finding a single factor of general intelligence (Sternberg, Castejon, Prieto, Hautamaki, & Grigorenko, 2001). Although Sternberg’s triarchic theory is the most comprehensive and ambitious model yet proposed, not all psychometric researchers have rushed to embrace it. Detterman (1984) cautions that we should investigate the basic cognitive components of intelligence before introducing higher-order constructs that may be unnecessary. Rogoff (1984) questions whether the three subtheories (componential, experiential, contextual) are sufficiently linked. Other comments on the triarchic theory can be found in Behavioral and Brain Sciences (1984, pp. 287–304). Whatever the final verdict on the triarchic theory of intelligence, Sternberg’s insistence

that intelligence has several components not measured by traditional tests rings true to anyone who has studied or administered these tests. He cites the case of a colleague who was asked to test a number of residents at an institution for those with mental retardation. These residents had just planned and successfully executed an escape from the security-conscious school, a feat requiring high levels of practical intelligence. Yet, when administered the Porteus Maze Test (Porteus, 1965), a standardized test reputed to involve planning ability, they could not solve even the simplest maze correctly. Sternberg (1986) has made it clear that intelligence just has too many components to be measured by any single test. TOPIC 5B Individual Tests of Intelligence and Achievement 5.12 Orientation to Individual Intelligence Tests 5.13 The Wechsler Scales of Intelligence 5.14 The Wechsler Subtests: Description and Analysis 5.15 Wechsler Adult Intelligence Scale-IV

5.16 Wechsler Intelligence Scale for Children- IV 5.17 Stanford-Binet Intelligence Scales: Fifth Edition 5.18 Detroit Tests of Learning Aptitude-4 5.19 The Cognitive Assessment System-II 5.20 Kaufman Brief Intelligence Test-2 (KBIT-2) 5.21 Individual Tests of Achievement 5.22 Nature and Assessment of Learning Disabilities

Individual intelligence testing is one of the major achievements of psychology since the founding of the discipline. In response to the success of the Binet-Simon scales in the early 1900s, psychologists developed and refined dozens of individual tests of intelligence patterned after this pathbreaking instrument. The explosive growth in group tests of intelligence, fostered by the enthusiastic acceptance of the Army Alpha and Beta tests

during and after World War I, also provided impetus to the individual testing movement. Many contemporary individual tests of intelligence owe their lineage to Binet, Simon, and the Army testing programs. The successful application of intelligence tests inspired educators and psychologists to look for ways to appraise the academic progress of students with school-based achievement tests. In turn, this led to the puzzling discovery that many children of normal or even superior intelligence lagged far behind in school achievement. From this discovery, the concept of learning disability gradually developed, and a whole new field of assessment was born. The purpose of this topic is to provide an overview of noteworthy approaches to the testing of individual intelligence and achievement, and to introduce the reader to the essentials of learning disability assessment. However, an exhaustive survey of individual cognitive tests is simply beyond the scope of this or any other basic reference. New and revised tests appear practically every month,

and thousands of new research findings are published every year. We have chosen to review tests that are widely used or that illustrate interesting developments in theory or method. Readers can find information on additional tests in the Mental Measurements Yearbook series, now published every two or three years by the Buros Institute.

5.12 ORIENTATION TO INDIVIDUAL INTELLIGENCE TESTS The individual intelligence tests reviewed in this topic include the following: • Wechsler Adult Intelligence Scale-IV

(WAIS-IV) • Wechsler Intelligence Scale for Children-IV

(WISC-IV) • Stanford-Binet: Fifth Edition (SB5) • Detroit Tests of Learning Aptitude-4

(DTLA-4) • Cognitive Assessment System-II (CAS-II)

• Kaufman Brief Intelligence Test-2 (KBIT-2) Collectively, these instruments probably account for 95 percent of the intellectual assessments conducted in the United States. The Wechsler scales have dominated intelligence testing in recent years, but they are by no means the only viable choices for individual assessment. Many other instruments measure general intelligence just as well—some would say better. Consider the implications of a now familiar observation: For large, heterogeneous samples, scores on any two mainstream instruments (e.g., Wechsler, Stanford-Binet, McCarthy, Kaufman scales) typically correlate .80 to .90. Often the correlation between two mainstream instruments is nearly as high as the test–retest correlation for either instrument alone. For purposes of producing a global score, it would appear that any well-normed mainstream intelligence test will suffice. But producing an overall score is not the only goal of assessment. In addition, the examiner usually desires to gain an understanding of the

subject’s intellectual functioning. For this purpose, the overall IQ is important, but there are instances in which the global score may be irrelevant or even misleading. To understand a referral’s intellectual functioning, the examiner should also inspect the subtest scores in search of hypotheses that might explain the unique functioning of that individual. Of course, examiners need to undertake subtest analysis cautiously, armed with research-based findings on the nature and meaning of subtest scatter for the test in use (Gregory, 1994b). If the examiner’s goal is to understand intellectual functioning and not merely to determine an overall score, the differences between tests become quite real. Every instrument approaches the measurement of intelligence from a different perspective and yields a distinctive set of subtest scores. Furthermore, a test well suited for one referral issue might perform abysmally in another context. For example, the WAIS-IV performs admirably in the testing of mild mental retardation but contains too few simple items for

the effective assessment of persons with moderate or severe developmental disability. A central axiom of assessment is that the choice of a testing instrument should be based on knowledge of its strengths and weaknesses as they pertain to the referral question. Put simply, the skilled examiner does not blindly rely on a single test for every referral! Instead, the skilled examiner flexibly chooses one or more instruments in light of the perceived assessment needs of the examinee. Each of the tests discussed in this topic has its special merits and also its particular shortcomings. The test user must know these strong and weak facets in order to choose the instruments best suited for each unique referral.

5.13 THE WECHSLER SCALES OF INTELLIGENCE Beginning in the 1930s, David Wechsler, a psychologist at Bellevue Hospital in New York City, conceived a series of elegantly simple instruments that virtually defined intelligence

testing in the mid- to late twentieth century. His influence on intelligence testing is exceeded only by the pathbreaking contributions of Binet and Simon. It is fitting that we begin the survey of individual tests with a historical summary of the Wechsler tradition, followed by a discussion of individual instruments. Origins of the Wechsler Tests Wechsler began work on his first test in 1932, seeking to devise an instrument suitable for testing the diverse patients referred to the psychiatric section of Bellevue Hospital in New York (Wechsler, 1932). In describing the development of his first test, he later wrote, “Our aim was not to produce a set of brand new tests but to select, from whatever source available, such a combination of them as would meet the requirements of an effective adult scale” (Wechsler, 1939). In fact, the content of his scales was largely inspired by earlier efforts such as the Binet scales and the Army Alpha and Beta tests (Frank, 1983). Readers who peruse Psychological Examining in the United States

Army, a volume edited by Yerkes (1921) just after World War I, might be astonished to discover that Wechsler purloined dozens of test items from this source, many of which have survived to the present day in contemporary revisions of the Wechsler tests. Wechsler was not so much a creative talent as a pragmatist who fashioned a new and useful instrument from the spare parts of earlier, discontinued attempts at intelligence testing. The first of the Wechsler tests, named the Wechsler-Bellevue Intelligence Scales, was published in 1939. In discussing the rationale for his new test, Wechsler (1941) explained that existing instruments such as the Stanford-Binet were woefully inadequate for assessing adult intelligence. The Wechsler-Bellevue was designed to rectify several flaws noted in previous tests: • The test items possessed no appeal for

adults. • Too many questions emphasized mere

manipulation of words.

• The instructions emphasized speed at the expense of accuracy.

• The reliance on mental age was irrelevant to adult testing.

To correct these shortcomings, Wechsler designed his test specifically for adults, added performance items to balance verbal questions, reduced the emphasis upon speeded questions, and invented a new method for obtaining the IQ. Specifically, he replaced the usual formula

with a new age-relative formula

This new formula was based on the interesting presumption—stated in the form of an axiom— that IQ remains constant with normal aging, even though raw intellectual ability might shift or even decline. The assumption of IQ constancy is basic to the Wechsler scales. As Wechsler (1941) put it:

The constancy of the I.Q. is the basic assumption of all scales where relative degrees of intelligence are defined in terms of it. It is not only basic, but absolutely necessary that I.Q.’s be independent of the age at which they are calculated, because unless the assumption holds, no permanent scheme of intelligence classification is possible.

Although Wechsler’s view has been largely accepted by contemporary test developers, it is important to stress that the assumption of IQ invariance with age is really a statement of values, a philosophical choice, and not necessarily an inherent characteristic of human nature. Wechsler also hoped to use his test as an aid in psychiatric diagnosis. In pursuit of this goal, he divided his scale into separate verbal and performance sections. This division allowed the examiner to compare an examinee’s facility in using words and symbols (verbal subtests) versus the ability to manipulate objects and perceive visual patterns (performance subtests).

Large differences between verbal ability (V) and performance ability (P) were thought to be of diagnostic significance. Specifically, Wechsler believed that organic brain disease, psychoses, and emotional disorders gave rise to a marked V > P pattern, whereas adolescent psychopaths and persons with mild mental retardation yielded a strong P < V pattern. Subsequent research demonstrated many exceptions to these simple diagnostic rules, and also helped refine the nature of these two major elements of intelligence. For example, verbal intelligence is now better known as verbal comprehension, and performance intelligence is more commonly recognized as perceptual reasoning. Nonetheless, the distinction between verbal and performance skills has proved useful for many purposes, such as studying brain–behavior relationships, and examining age effects on intelligence. Wechsler’s armchair division of subtests into verbal and performance sections, even though refined and extended by others, continues to endure as a major contribution to

contemporary intelligence testing (Kaufman, Lichtenberger, & McLean, 2001). General Features of the Wechsler Tests Including revisions, David Wechsler and his followers have produced more than a dozen intelligence tests in a span of about 70 years. A major reason for the continued success of these instruments has been the faithful adherence to the familiar content and format first introduced in the Wechsler-Bellevue. By sticking with a single successful formula, Wechsler and company ensured that examiners could switch from Wechsler test to another with minimal retraining. This was not only good psychometrics but also shrewd marketing insofar as it guaranteed several generations of faithful test users. The latest editions of the Wechsler intelligence tests—the WPPSI-IV, WISC-IV, and WAIS-IV —possess the following common features: • Thirteen to fifteen subtests. The multisubtest

approach allows the examiner to analyze intra-individual strengths and weaknesses

rather than just to compute a single global score. In addition, it is possible to combine subtest scores in theoretical meaningful ways that provide useful information on the broad factors of intelligence. As the reader will learn subsequently, the pattern of subtest and factor scores may convey useful information that is hidden in the overall level of performance.

• An empirically based breakdown into composite scores and a full scale IQ. Whereas the original Wechsler intelligence scales provided only two composite scores —Verbal IQ and Performance IQ—the revisions have been moving toward a more sophisticated partitioning into composites confirmed from factor-analytic research. The WISC-IV and WAIS-IV now yield composite or index scores in the same four areas: ◦ Verbal Comprehension ◦ Perceptual Reasoning ◦ Working Memory ◦ Processing Speed

• The WPPSI-IV provides five index scores similar to the above (for ages 4:0 to 7:7) but also includes a Fluid Reasoning composite.

• A common metric for IQ and Index scores. The mean for IQ and Index scores is 100 and the standard deviation is 15 for all tests and all age groups. In addition, the scaled scores on each subtest have a mean of 10 and a standard deviation of approximately 3, which permits the examiner to analyze the subtest scores of the examinee for relative strengths and weaknesses.

• Common subtests for the different test versions. For example, the preschool, child, and adult Wechsler tests (WPPSI-IV, WISC- IV, and WAIS-IV) all share a common core of the same six subtests (Table 5.6). An examiner who masters the administration of a core subtest on any of the Wechsler tests (such as the Information subtest on the WAIS-IV) easily can transfer this skill within the Wechsler family of intellectual measures.

TABLE 5.6 Subtest Composition of the Wechsler Intelligence Tests

WPPSI- IV

WISC- IV

WAIS- IV

Similarities × × × Vocabulary × × × Comprehension × × × Information × × × Word Reasoning × Receptive Vocabulary

×

Picture Naming × Block Design × × × Picture Concepts × × Matrix Reasoning × × × Picture Completion × × Visual Puzzles × Figure Weights × Object Assembly × L-N Sequencinga × × Arithmetic × × Digit Span × × Coding × × Symbol Search × ×

aLetter–Number Sequencing. Note: The subtests common to all Wechsler intelligence tests are in boldface. Some subtests are optional or used as substitutions. See text for details.

5.14 THE WECHSLER SUBTESTS: DESCRIPTION

AND ANALYSIS Wechsler (1939) defined intelligence as “the aggregate or global capacity of the individual to act purposefully, to think rationally and to deal effectively with his environment.” He also believed that we can only know intelligence by what it enables a person to do. In designing his tests, then, Wechsler selected components to represent a wide array of underlying abilities so as to estimate the global capacity of intelligence. Furthermore, he asked his subjects to do things, not merely to answer questions. The Wechsler

Cancellation × × Picture Memory × Bug Search × Zoo Memory ×

subtests are quite diverse and often rely on what Wechsler referred to as “mental productions.” We present here a description of subtests from the WISC-IV and WAIS-IV. We also analyze the abilities tapped by each subtest and offer research-based comments. The reader is referred to Topic 7A, Infant and Preschool Assessment, for a description of the subtests unique to the WPPSI-IV. Information The Information subtest is found on all three Wechsler intelligence tests. Factual knowledge of persons, places, and common phenomena is tested here. Questions for children are like the following: • “How many eyes do you have?” • “Who invented the telephone?” • “What causes a solar eclipse?” • “Which is the largest planet?”

Questions for adults are similar but progress to higher levels of difficulty. Difficult questions on the adult Information subtest resemble:

• “Which is the most common element in air?”

• “What is the population of the world?” • “How does fruit juice get converted to

wine?” • “Who wrote Madame Bovary?”

Information items test general knowledge normally available to most persons raised in the cultural institutions and educational systems of Western industrialized nations. Indirectly, this subtest measures learning and memory skills insofar as subjects must retain knowledge gained from formal and informal educational opportunities in order to answer the Information items. Information is usually regarded as one of the best measures of general ability among the Wechsler subtests (Kaufman, McLean, & Reynolds, 1988). For example, the WAIS-IV manual reveals that Information typically has the second or third highest correlation with Full Scale IQ across the 13 age groups (Wechsler, Coalson, & Raiford, 2008). Information consistently loads strongly on the first factor

identified in factor analyses of the WAIS-IV subtest correlations (see the following). The first factor is labeled Verbal Comprehension. However, Information tends to reflect formal education and motivation for academic achievement and may therefore yield spuriously high ability estimates for perpetual students and avid readers. Digit Span Digit Span consists of two separate sections, Digits Forward and Digits Backward. In Digits Forward, the examiner reads a series of digits at one per second, then asks the subject to repeat them. If the subject answers correctly on two consecutive trials of the same length, the examiner proceeds to the next series, which is one digit longer, up to a maximum length of nine digits. For Digits Backward, a similar procedure is used, except the examinee must repeat the digits in reverse order, up to a maximum length of eight digits. For example, the examiner reads: “6–1–3–4–2–8–5”

and the subject tries to repeat the numbers in the reverse order: “5–8–2–4–3–1–6.” On the WAIS-IV only, the Digit Span subtest also includes a third section called Digit Sequencing. For this part, the examinee is asked to sort the series of digits into their correct order. For example, if the examiner says: “1–7–4–9–2” the examinee should respond: “1–2–4–7–9.” Digit Span is a measure of immediate auditory

recall for numbers. Facility with numbers, good attention, and freedom from distractibility are required. Performance on this subtest may be affected by anxiety or fatigue, and many clinicians have noted that patients hospitalized for medical or psychiatric reasons frequently perform poorly on Digit Span.

Digits Forward and Digits Backward may assess fundamentally different abilities. Digits Forward seems to require the examinee to access an auditory code in sequential fashion. In contrast,

to perform Digits Backward, the examinee must form an internal visual memory trace from the orally presented numerical sequences and then visually scan from end to beginning. Digits Backward is clearly the more complex test; not surprisingly, it loads higher on general intelligence than does Digits Forward (Jensen & Osborne, 1979). Gardner (1981) argues that examiners should supplement standard reporting procedures and list separate subscores for Digit Span. He presents separate means, standard deviations, and percentile ranks on Digits Forward and Backward for children ages 5 to 15. Vocabulary The Vocabulary subtest is found on all three Wechsler intelligence tests. The examinee is asked to define up to several dozen words of increasing difficulty while the examiner writes down each response verbatim. For example, on an easy item the examiner might ask, “What is a cup?” and the examinee would get partial credit for answering, “You drink with it” and full

credit for answering, “It has a handle, holds liquids, and you drink from it.” For adults and bright children, the advanced items on the Wechsler Vocabulary subtests can be very challenging, on a par with tincture, obstreperous, and egregious. Vocabulary is learned largely in context from reading books and listening to others. It is a rare individual who picks up vocabulary by reading the dictionary or memorizing word lists from the “Building Your Wordpower” section of popular magazines. In the main, a person’s vocabulary is a measure of sensitivity to new information and the ability to decipher meanings based on the context in which words are encountered. Precisely because the acquisition of word meaning depends on contextual inference, the Vocabulary subtest turns out to be the single best measure of overall intelligence on the Wechsler scales (Gregory, 1999). This is a surprise to many laypersons who regard vocabulary as merely synonymous with educational exposure and, therefore, a mediocre index of general intelligence. However, there is

simply no denying the empirical evidence: Vocabulary has among the highest subtest correlations with Full Scale IQ on both the WISC-IV and also the WAIS-IV. Arithmetic Except for the very easiest items for young people or persons who have mental retardation, the Arithmetic subtest consists of orally presented mathematics problems. The examinee must solve the problems without paper or pencil within a time limit (usually 30 to 60 seconds). The simple items stress fundamental operations of addition or subtraction, for example: “If you have fifteen apples and give seven away,

how many are left?” The more difficult items require proper conceptualization of the problem and the application of two arithmetic operations, for example: “John bought a stereo that was marked down 15

percent from the original sales price of $600. How much did John pay for the stereo?”

Although the mathematical requirements of the Arithmetic items are not excessively demanding, the necessity of solving the problems mentally within a time limit makes this subtest quite challenging for most examinees. In addition to rudimentary arithmetic skills, successful performance on Arithmetic requires high levels of concentration and the ability to maintain intermediate calculations in short-term memory. In factor analyses of the WISC-IV and WAIS-IV, Arithmetic often loads on a third factor interpreted as Working Memory. Comprehension Found on all three Wechsler intelligence tests, the Comprehension subtest is an eclectic collection of items that require explanation rather than mere factual knowledge. The easy questions stress common sense, whereas the more difficult questions require an understanding of social and cultural conventions. On the WAIS-IV, several of the

most difficult questions require the examinee to interpret proverbs. An easy item on Comprehension is of the form “Why do people wear clothes?” Difficult items resemble the following: • “What does this saying mean: ‘A bird in the

hand is worth two in the bush.’” • “Why are Supreme Court Judges appointed

for life?” Comprehension would appear to be, in part, a measure of “social intelligence” in that many items tap the examinee’s understanding of social and cultural conventions. Sipps, Berry, and Lynch (1987) found that Comprehension scores were moderately related to measures of social intelligence on the California Psychological Inventory. Of course, a high score signifies only that the examinee is knowledgeable about social and cultural conventions; choosing the right action may or may not flow from this knowledge. However, studies by Campbell and McCord (1996) and Lipsitz, Dworkin, and Erlenmeyer-Kimling (1993) provide no support for the commonly accepted clinical lore that

Comprehension scores are sensitive to social functioning. Similarities In this subtest, the examinee is asked questions of the type, “In what way are shirts and socks alike?” The Similarities subtest evaluates the examinee’s ability to distinguish important from unimportant resemblances in objects, facts, and ideas. Indirectly, these questions assess the assimilation of the concept of likeness. The examinee must also possess the ability to judge when a likeness is important rather than trivial. For example, “shirts” and “socks” are alike in that both begin with the letter s, but this is not the essential similarity between these two items. The important similarity is that shirts and socks are both exemplars of a concept, namely, “clothes.” As this example illustrates, Similarities can be thought of as a test of verbal concept formation and is found on all three Wechsler intelligence tests. Letter–Number Sequencing

The examiner orally presents a series of letters and numbers that are in random order. The examinee must reorder and repeat the list by saying the numbers in ascending order and then the letters in alphabetical order. For example, if the examiner says “R-3-B-5-Z-1-C,” the examinee should respond “1-3-5-B-C-R-Z.” This test measures attention, concentration, and freedom from distractibility. Together with Arithmetic and Digit Span, this subtest contributes to the Working Memory Index score on the WAIS-IV (see the following). Donders, Tulsky, and Zhu (2001) found the Letter– Number Sequencing subtest to be highly sensitive to the effects of moderate and severe traumatic brain injury. Picture Completion For this subtest, the examiner asks the examinee to identify the “important part” that is missing from a picture. For example, a simple item might be of this type: a picture of a table with one leg missing. The items get harder and harder; testing continues until the examinee

misses several in a row. Figure 5.6 depicts an item similar to those found on the WAIS-IV. The Picture Completion subtest presupposes that the examinee has been exposed to the object or situation represented. For this reason, Picture Completion may be inappropriate for culturally disadvantaged persons.

FIGURE 5.6 Picture Completion Item Similar to Those Found on the WAIS-IV Picture Concepts This subtest is found on the WPPSI-IV and the WISC-IV. For each item, the child is shown a

card with two or three rows of pictures and instructed to choose one picture from each row to form a group with a common characteristic. This is a recent subtest designed to measure abstract, categorical reasoning. The 28 items reflect increasingly more difficult levels of abstraction. For example, for an easy item the commonality might be that a fruit is found in each row, whereas for a more difficult item the commonality might be that a device used for signaling (bell, flashlight, flags) is found in each row. Block Design On the Block Design subtest, the examinee must reproduce two-dimensional geometric designs by proper rotation and placement of three- dimensional colored blocks. For all of the Wechsler scales, the first few Block Design items can be solved through trial and error. However, the more difficult items require the analysis of spatial relations, visual-motor coordination, and the rigid application of logic. Block Design demands much more problem-

solving and reasoning ability than most of the Performance subtests in which memory and prior experience are more heavily weighted. Block Design is a strongly speeded test. Consider the WAIS-IV version, which consists of 14 designs of increasing difficulty. To obtain a high score on this subtest, adults must not only reproduce each of the designs correctly, but they must also earn bonus points on the last six designs by completing them quickly. An examinee who solves all the designs within the time limit but who fails to garner any bonus points will test out at just slightly above average on this subtest. Block Design scores may be misleading for examinees who do not value speeded performance. Matrix Reasoning

FIGURE 5.7 Matrix Reasoning Item Similar to Those Found on the WAIS-IV Matrix Reasoning is included on all of the Wechsler intelligence tests. The subtest consists of figural reasoning problems arranged in increasing order of difficulty (Figure 5.7). Finding the correct answer requires the examinee to identify a recurring pattern or

relationship between figural stimuli drawn along a straight line (simple items) or in a 3 × 3 grid (hard items) in which the last item is missing. Based on nonverbal reasoning about the patterns and relationships, the examinee must infer the missing stimulus and select it from five choices provided at the bottom of the card. Matrix Reasoning was designed to be a measure of fluid intelligence, which is the capacity to perform mental operations such as manipulation of abstract symbols. The items tap pattern completion, reasoning by analogy, and serial reasoning. Overall, the subtest is an excellent measure of inductive reasoning based on figural stimuli. Matrix Reasoning is not timed. Interestingly, Donders et al. (2001) report that the Matrix Reasoning subtest is relatively unaffected by moderate and severe traumatic brain injury. Object Assembly This subtest is found only on the WPPSI-III. For each item, the examinee must assemble the pieces of a jigsaw puzzle to form a common

object (Figure 5.8). The examiner does not identify the items, so the examinee must first discern the identity of each item from its disarranged parts. Success on this subtest requires high levels of perceptual organization; that is, the examinee must grasp a larger pattern or gestalt based on perception of the relationships among the individual parts. Object Assembly is one of the least reliable of the Wechsler subtests. The modest reliability of Object Assembly may reflect, in part, the small number of items as well as the role of chance factors in solving jigsaw puzzles.

FIGURE 5.8 Object Assembly Item Similar to Those Found on the WPPSI-III Coding The WISC-IV version consists of two separate and distinct parts, one for examinees under age 8 (Coding A) and another for those 8 years of age and over (Coding B). In Coding A, the child must draw the correct symbol inside a series of randomly sequenced shapes. The task utilizes five shapes (star, circle, triangle, cross, and square), and each shape is assigned a unique symbol (vertical line, two horizontal lines, single horizontal line, circle, and two vertical lines, respectively). After a brief practice session, the child is told to draw the correct symbol inside 43 of the randomly sequenced shapes. However, since there is a two-minute time limit, high scores require rapid performance. Coding B on the WISC-IV and Coding on the WAIS-IV are identical in format (Figure 5.9). For both subtests, the examinee must associate one symbol with each of the digits 0 through 9

and quickly draw the appropriate symbol underneath a long series of random digits. The time limit for both versions is two minutes. Very few examinees manage to code all the stimuli in this amount of time.

FIGURE 5.9 Digit Symbol Items Similar to Those Found on the WAIS-IV Estes (1974) analyzed the Coding subtest from the standpoint of learning theory and concluded that efficient performance requires the ability to quickly produce distinctive verbal codes to represent each of the symbols in memory. For example, in Figure 5.9, the examinee might code the symbol underneath the number 2 as an “inverted T.” Verbal coding mediates quick performance by simplifying a difficult task. Efficient performance also demands immediate

learning of the digit-symbol pairings so that the examinee need not look from each digit to the reference table to determine the correct response. In this regard, Coding is unique: It is the only Wechsler subtest that necessitates on- the-spot learning of an unfamiliar task. Coding scores show a steep decrement with advancing age. In cross-sectional studies, raw scores on Coding decline by as much as 50 percent from age 20 to age 70 (Wechsler, 1981). The decrement is approximately linear and not easily explained by superficial references to motivational differences or motor slowing. Of course, cross-sectional results are not necessarily synonymous with longitudinal trends. However, the age decrement on Coding is so steep that it must indicate, in part, a real age change in the speed of basic information processing skills. Coding is one of the most sensitive subtests to the effects of organic impairment (Donders et al., 2001; Lezak, 1995). Symbol Search

This is a highly speeded subtest in which the examinee looks at a target group of symbols, then quickly examines a search group of symbols, and finally marks a “YES” or “NO” box to indicate whether one or more of the symbols in the target group occurred within the search group. A Symbol Search item is depicted in Figure 5.10. This subtest would appear to be a measure of processing speed. Symbol Search is highly sensitive to the impact of traumatic brain injury (Donders et al., 2001). Cancellation

FIGURE 5.10 Symbol Search Item Similar to Those Found on the WISC-IV Note: The examinee’s task is to determine whether either shape at the left occurs among the five shapes to the right. On the WISC-IV, this is a timed subtest in which the child is instructed to draw a line through or “cancel” drawings of animals placed randomly among drawings of inanimate objects (e.g., umbrella, car, hydrant, lightbulb). For

example, on a standard-sized sheet of paper, about 160 stimuli are pictured, including 30 animals (horse, bear, seal, fish, chicken). Cancellation consists of two trials: one with a random arrangement of visual stimuli, and one with clearly structured rows and columns of stimuli. In addition to a total subtest score, separate process scores for the random and the structured trials are available for comparison. This subtest is similar to existing cancellation tasks designed to measure processing speed, vigilance, and visual attention. It is well established that examinees with neuropsychological impairments perform poorly, especially on the random trial (e.g., Bate, Mathias, & Crawford, 2001; Geldmacher, 1996). On the WAIS-IV, Cancellation is somewhat more complex, involving two target stimuli consisting of geometric shapes. The examinee is told, for example, to cancel “red squares and yellow triangles” among an array of red and yellow squares and red and yellow triangles. A second trial involves stars and circles in orange and blue. This timed task (45

seconds per trial) is much more difficult than it seems. Visual Puzzles Visual Puzzles is found only on the WAIS-IV. The examinee is shown a picture of a completed shape such as a rectangle, and asked to select from six smaller shapes the three that could be used to assemble the larger completed shape. Successful performance requires visual-spatial analysis and the mental rotation of shapes. According to the WAIS-IV Technical Manual, this subtest taps for “visual perception, broad visual intelligence, fluid intelligence, simultaneous processing, spatial visualization and manipulation, and the ability to anticipate relationships among parts (Wechsler, 2008b, p. 14). The 26 items have strict time limits of 20 seconds for the initial easy items and 30 seconds for the remaining items. Visual Puzzles is a core subtest that contributes to the Perceptual Reasoning Index of the WAIS-IV. Figure Weights

Figure Weights is found only on the WAIS-IV. It is a supplemental subtest that contributes to the Perceptual Reasoning Index. The examinee is shown a picture of an old-fashioned fulcrum scale that is missing weight(s) on one side. The task is to select from six options the response that would bring the scale into balance. This subtest is a measure of quantitative and analogical reasoning; inductive and deductive logic are essential for success. Easy items provide a time limit of 20 seconds, hard items allow 40 seconds.

5.15 WECHSLER ADULT INTELLIGENCE SCALE-IV The WAIS-IV is a significant revision of the WAIS-III, even though many of the previous items were retained (Wechsler, 2008). The most significant changes include the addition of two subtests, a simplified test structure, and an emphasis on index scores that provide a sharper demarcation of discrete domains of cognitive functions. In addition, the WAIS-IV abandons

the familiar (but psychometrically indefensible) bifurcation of intelligence into Verbal IQ and Performance IQ, preferring instead the fourfold breakdown discussed below. In addition to traditional approaches to scoring the WAIS-IV subtests, the new edition also provides neuropsychologically relevant process scores for four of the subtests. These scores are useful mainly for advanced forms of test interpretation in the context of a comprehensive test battery. We do not discuss process scores in this section. Because of improvements in the WAIS-IV protocol forms (e.g., prominent display of discontinue rules), this test is somewhat easier to administer than its predecessor. Lichtenberger and Kaufman (2009) provide an outstanding overview of the WAIS-IV in clinical practice. The WAIS-IV is comprised of 15 subtests, but only 10 of the subtests, known as core subtests, are needed to obtain the traditional IQ score and the component index scores. The other five subtests are deemed supplemental. These are often used to provide additional clinical information; in specific instances, supplemental

subtests may be used as acceptable substitutes for core subtests. In addition to the traditional Full Scale IQ score, normed to a mean of 100 and standard deviation of 15, the WAIS-IV is scored for four index scores, each based on 2 or 3 of the 10 core subtests. These are derived from factor analysis of the subtests, which revealed four domains: Verbal Comprehension, Perceptual Reasoning, Working Memory, and Processing Speed. The index scores are also based on the familiar mean of 100 and standard deviation of 15. The breakdown of subtests for the four index scores is as follows: • Verbal Comprehension Index

• Similarities • Vocabulary • Information

• Perceptual Reasoning Index • Block Design • Matrix Reasoning • Visual Puzzles

• Working Memory Index • Digit Span

• Arithmetic • Processing Speed Index

• Symbol Search • Coding

The Verbal Comprehension Index (VCI) is similar to the outdated notion (used on the WAIS-III) of Verbal IQ or VIQ. However, from a psychometric standpoint, VCI is a cleaner and more direct measure of verbal comprehension than VIQ, hence it is now the preferred index. Likewise, the Perceptual Reasoning Index (PRI) is similar to the former notion (from the WAIS- III) of Performance IQ or PIQ. Yet, as a more refined measure of perceptual reasoning, PRI is therefore the preferred index. Put simply, VCI and PRI fit the factor analytic data better. Long- held conventions tend to persist, but it is time to let the outdated notions of Verbal IQ and Performance IQ fade into oblivion. The Working Memory Index (WMI) is comprised of subtests sensitive to attention and immediate memory (Digit Span and Arithmetic). A relatively low score on this index may signify that the examinee has an attentional or memory

problem, especially with orally presented materials. The Processing Speed Index (PSI) comprises subtest that require the highly speeded process of visual information (Symbol Search and Coding). The PSI is sensitive to a wide variety of neurological and neuropsychological conditions (Tulsky, Zhu, & Ledbetter, 1997). WAIS-IV Standardization The standardization of the WAIS-IV was undertaken with great care and based on data gathered by the U.S. Bureau of the Census in 2005. The total sample of 2,200 adults (ages 16 to 91) was carefully stratified on these variables: gender, race/ethnicity, education level, and geographic region. Census figures from 2005 were used as the target values for the stratification variables. For example, of persons in the 55- to 64-year-old range, the Census Bureau found that 3.35 percent are African Americans with high school education. In like manner, 3.00 percent of the standardization

participants were African Americans with high school education. The standardization sample was divided into 13 age bands: 16–17, 18–19, 20–24, 25–29, 30–34, 35–44, 45–54, 55–64, 65–69, 70–74, 75–79, 80– 84, 85–90. Except for the four oldest age groups, each sample included 200 participants carefully stratified on the demographic variables noted earlier; the last four age groups included 100 participants each. The resulting sample bears a very close correspondence to the U.S. Census proportions. However, persons suspected of even mild cognitive impairment were excluded, so that the standardizations sample likely is healthier than its census counterparts. Specifically, several exclusionary criteria were used in the standardization sample, including: uncorrected visual or hearing impairment, current hospitalization, evidence of drug/alcohol problems, upper extremity impairment, use of certain prescription drugs such as anticonvulsants, and a variety of potentially brain-impairing conditions (e.g., head injury, stroke, epilepsy, dementia, and

mood disorder). Uncooperative participants and those for whom English was a second language also were excluded. In sum, the standardization sample was restricted to cooperative, reasonably healthy, English-speaking individuals who did not manifest significant brain-impairing conditions. Although the WAIS-IV is similar to the WAIS- III and has a substantial item overlap, the two tests do not yield analogous IQs. In counterbalanced studies comparing scores of 240 adults on the two tests, WAIS-IV IQ scores are lower by 3 points. In sum, the WAIS-IV is a harder test than the WAIS-III. There is a troubling enigma here: Why does the normative sample for the WAIS-IV appear to be smarter than the normative sample for the WAIS-III? We take up this point in more detail in Topic 6B, Test Bias and Other Controversies. Reliability The reliability of the WAIS-IV is exceptionally good. Composite split-half reliabilities averaged across all age groups for the Index scores and

IQ are: VCI .96, PCI .95, WMI .94, PSI .90, and Full Scale IQ 98. Further supporting the reliability of the WAIS-IV, reliability estimates for subtest scores of special groups (e.g., persons with intellectual disability, probable Alzheimer’s disease, traumatic brain injury, major depression, autism) are equal to or higher than reliability estimates found in the general population (Wechsler, 2008b). This suggests that the WAIS-IV is a reliable tool not just with the general population but also with the special populations who are more likely to be the focus of assessment. For Full Scale IQ, the standard error of measurement is 2.6 points for the youngest examinees (ages 16 and 17), but even smaller at 2.1 points for all other age groups. Consider what this means: 95 percent of the time, an examinee’s true Full Scale IQ will be with ±4 points (2 standard errors of measure) of the obtained value. In common parlance, psychometrists would say that WAIS-IV IQ has an 8-point band of error, that is, IQ scores are accurate within about ±4 points. In contrast to

the strong reliabilities found for IQ and Index scores, the reliabilities of the 15 individual subtests are generally much weaker. The only subtests with stability coefficients in excess of .90 are Information (.90) and Vocabulary (.91). For the remaining subtests, reliability values range from the low .70s to the mid .80s. The most important implication of these weaker reliability findings is that examiners should approach subtest profile analysis with extreme caution. Subtest scores that appear discrepantly high (or low) for an individual examinee might be a consequence of the generally weak reliability of certain subtests rather than indicating true cognitive strengths or weaknesses. Some reviewers conclude that profile analysis (the identification of specific cognitive strengths and weaknesses based on analysis of peaks and valleys in the subtest scores) is not justified by the evidence. Validity The developers of the WAIS-IV provide a number of different lines of evidence to support

the validity of this instrument (Wechsler, 2008b). Good content validity was built in from the beginning through comprehensive literature review and consultation with experts to assure that items and subtests tap the relevant range of cognitive processes. Good criterion-related validity was demonstrated in several studies correlating the WAIS-IV with mainstream intelligence tests and other measures. For example, WAIS-IV Full Scale IQ correlates strongly with global scores on other mainstream measures: .94 with the WAIS-III, .91 with the WISC-IV (for 16-year-olds in the overlapping age group), and .88 with the Wechsler Individual Achievement Test-II. The WAIS-IV also reveals appropriate convergent and discriminant validity in the patternings of strong and weak correlations with a wide variety of other instruments, including measures of attention deficit disorder, executive functions, and memory. As a generalization, correlations are appropriately strong among similar subtests and constructs from the WAIS-IV and other

tests, and appropriately weak among dissimilar subtests and constructs. Studies with special groups also provide theory- confirming results that speak to the validity of the WAIS-IV. The multiplicity of these studies is such that we can only provide a few examples here. Specifically, when 41 young adults with diagnosed Mathematics Disorder were compared to matched controls on WAIS-IV subtests, the most substantial difference by far was found on the Arithmetic subtest, where the clinical group averaged 6.6 compared to 8.8 for the matched controls (a subtest score of 10 is average in the general population). This corroborates the sensitivity of the instrument to the elements of one specific learning disability. In like manner, when 22 individuals with a history of moderate or severe brain injury were compared to matched controls, the largest difference among the four index scores was found on the Processing Speed Index (mean of 80.5 versus mean of 97.6), whereas the smallest difference among the four index scores was found on the Verbal Comprehension Index

(mean of 92.1 versus mean of 100.8). These findings are exactly what would be predicted from a wide body of research on the impact of traumatic brain injury (e.g., Lezak, Howieson, & Loring, 2004). The construct validity of the WAIS-IV is also supported by confirmatory factor analyses of the subtest scores from the standardization sample, as detailed in the technical manual (Wechsler, 2008b). These complex analyses were designed to determine if the relations among observed subtest scores support the existence of the hypothesized factors of intelligence measured by the four index scores of VCI, PRI, WMI, and PSI. The goodness-of-fit of the four factor hierarchical model of intelligence (Full Scale IQ at the top, sitting above the four index scores, each sitting above two or three constituent subtest scores) turns out to be exceptionally strong, although difficult to summarize in visual form. A simple way to depict the strong confirmatory fit is through a 4 × 10 table that shows the correlations among the four index scores and the 10 core subtest scores (Table

5.7). Where appropriate, these correlations are corrected for overlap between the subtest scores and the index scores. For example, Similarities is a component of VCI, so the simple correlation between these two variables is artificially inflated. The values shown in Table 5.7 are corrected for this kind of overlap. The reader will notice that with only a single exception, the subtests that compose each index score reveal their highest correlations with that index score. The only exception is the Arithmetic subtest, which is factorially more complex than other subtests, showing an almost identical relationship with VCI, PRI, and WMI. Finally, the validity of the WAIS-IV is also buttressed by its strong overlap with the previous three editions of the test, for which there is an impressive array of validity data. For a full review of these findings the reader can consult Matarazzo (1972) and Kaufman (1990). TABLE 5.7 Correlations Among WAIS-IV Subtests and Index Scores

Source: Based on data in Wechsler, D. (2008). WAIS-IV technical and interpretive manual. San Antonio, TX: Pearson. Note: Decimals have been omitted. Where appropriate, these correlations are corrected for overlap. For example, because Similarities is a component of VCI, the simple uncorrected correlation between these two variables would be artificially inflated. The values

VC I

PR I

WM I

PS I

Verbal comprehension Subtests Similarities 74 57 57 42 Vocabulary 81 55 60 41 Information 63 54 56 37 Perceptual Reasoning Subtests Block Design 51 67 53 45 Matrix Reasoning 56 59 55 46 Visual Puzzles 48 66 49 41 Working Memory Subtests Digit Span 53 52 60 47 Arithmetic 63 59 60 44 Processing Speed Subtests Symbol Search 38 47 43 65 Coding 43 48 49 65

above are corrected for any componential overlap between subtests and index scores.

5.16 WECHSLER INTELLIGENCE SCALE FOR CHILDREN-IV The Wechsler Intelligence Scale for Children (WISC) was published in 1949 as a downward extension of the original Wechsler-Bellevue. Although used widely in the next two decades, psychometricians perceived a number of flaws in the WISC: absence of nonwhites in the standardization sample, ambiguities of scoring, inappropriate items for children (e.g., reference to “cigars”), and absence of females and African Americans in the pictorial content of items. The WISC-R, WISC-III, and WISC-IV corrected these flaws. The WISC-IV consists of 15 subtests, 10 of which are designated as core subtests used in the computation of composite scores and Full Scale IQ, and five of which are designated as supplemental:

Although the supplemental subtests are not required for the computation of Full Scale IQ and composite scores (discussed later), careful examiners nonetheless may choose to administer them because of the important diagnostic information they often provide. For example, the Cancellation subtest is supplemental but affords important information about vigilance and visual attention; hence, many examiners use it. The Arithmetic subtest also is supplemental but often chosen by examiners because it is helpful in the

Core Subtests Block Design Vocabulary Similarities Letter–Number Sequencing Digit Span Matrix Reasoning Picture Concepts Comprehension Coding Symbol Search Supplemental Subtests Picture Completion Cancellation Information Arithmetic Word Reasoning

assessment of auditory attention (the questions are presented orally). Another function of the supplementary subtests is suitable substitution for a core subtest. In well-defined circumstances, an examiner may elect to give a supplemental subtest in place of a core subtest. For example, when testing a child with fine motor problems—such as might be observed in a child with cerebral palsy—an examiner would be well advised to use Cancellation in place of Coding, and Picture Completion in place of Block Design. Both of these supplementary tests (Cancellation and Picture Completion) are relatively unaffected by fine motor difficulties. In contrast, the core subtests (Coding and Block Design) would be severely impacted by fine motor difficulties and, therefore, could yield unfair assessments of cognitive functioning. Substitutions also are allowed when a core subtest accidentally is invalidated. However, an examiner may not elect to substitute a supplemental subtest merely because a child has performed poorly on a core subtest.

The standardization of the WISC-IV is first class, based on 100 boys and girls at each year of age from 6½ through 16½ (total N = 2,200). These cases were carefully selected and stratified on the basis of the 2000 U.S. Census with respect to gender, race/ethnicity (white, African American, Hispanic, and Asian), geographic region, and parent educational level. A desirable feature of the standardization sample is that 5.7 percent of the sample consisted of children with defined characteristics such as giftedness, learning disability, expressive language disorder, head injury, autism, and motor impairment. The purpose of adding these children was to ensure that the normative sample accurately represented the population of children attending school. The correspondence between the standardization sample and the U.S. Census data on essential stratification variables was nearly perfect (Wechsler, 2003, p. 40). The reliability of the WISC-IV is strong and comparable to previous editions of the test. For example, the IQ and composite scores show

split-half and test–retest reliabilities in the .90s, whereas the individual subtests possess somewhat lower reliability coefficients, ranging from .79 (Cancellation and Symbol Search) to .90 (Letter–Number Sequencing). Most reliabilities are in the high .80s, for example, Block Design and Similarities at .86, and Vocabulary and Matrix Reasoning at .89. Test– retest reliabilities tend to be slightly lower. The validity of the WISC-IV rests, in part, on its overlap with the WISC-III, for which dozens of supportive studies could be cited. We do not want to overwhelm with excessive detail, so we refer the interested reader to Sattler (2001) for a good review of earlier studies. The WISC-IV manual cites an impressive array of validity studies, which we summarize here. First, we discuss correlations of WISC-IV test scores with its predecessor and with other Wechsler intelligence tests. The preliminary findings indicate strong correlations with comparable WISC-III subtests, most in the high .70s or low .80s. The correlation for Full Scale IQ is much higher, r = .89. Likewise, correlations with the

WPPSI-III are strong for comparable subtests, and, again, exceptionally strong for Full Scale IQ, r = .89. A similar pattern is found with 16- year-old examinees, who can be tested legitimately with both the WISC-IV and the WAIS-III. In a sample of 198 children tested in counterbalanced order over a period of about three weeks, correlations were strong for comparable subtests and exceptionally strong for composite and Full Scale IQ scores (r = .89). Overall, these are remarkable correlations, nearly as strong as the reliabilities of the respective scales would allow. An interesting finding is that WISC-IV IQs are an average of 2.5 points lower than WISC-III IQs and 3 points lower than WAIS-III IQs. This is a consistent finding in the history of individual intelligence tests; namely, newer tests almost invariably yield lower Q scores in comparison to older tests. We discuss this intriguing result, called the Flynn effect, in the next chapter. Factor-analytic studies of the standardization sample provided additional evidence for the utility of the WISC-IV in the diagnostic

assessment of children. The results of numerous factor analyses, including separate analyses for four age groups (6–7, 8–10, 11–13, 14–16) strongly confirmed a four-factor solution that was used to define the composite scores, called Index scores, for the test (Wechsler, 2003). The factors and the core subtests assigned to them were as follows: • Verbal Comprehension Index

• Similarities • Vocabulary • Comprehension

• Perceptual Reasoning Index • Block Design • Picture Concepts • Matrix Reasoning

• Working Memory Index • Digit Span • Letter–Number Sequencing

• Processing Speed Index • Coding • Symbol Search

The four Index scores are based on the familiar mean of 100 and standard deviation of 15. Thus,

the WISC-IV provides substantial detail about the nuances of intellectual functioning—up to 15 subtest scores, four Index scores, and the Full Scale IQ. The robust findings of the four-factor solution to the WISC-IV provided the rationale for abandoning Wechsler’s original two-factor division of Verbal IQ and Performance IQ. In fact, there is no longer any method on the WISC-IV to obtain a Verbal IQ or a Performance IQ—precisely because these partitions no longer fit with the emerging consensus about the nature of intelligence. The WISC-IV also revealed theory-confirming correlations with a variety of cognitive, ability, and achievement tests (Wechsler, 2003). In general, correlations with other measures were appropriately high for similar constructs and predictably low for dissimilar constructs—these are the prerequisites for convergent validity and discriminant validity, respectively. For example, in a sample of 550 children aged 6–16, reading achievement subtest scores from the Wechsler Individual Achievement Test-II correlated more strongly with Verbal Comprehension Index

scores from the WISC-IV than with the other Index scores. Likewise, in a sample of 126 children aged 6–16, the Attention/Concentration subtest from the Children’s Memory Scale (Cohen, 1997) correlated substantially (r = .74) with Working Memory Index scores from the WISC-IV but less robustly with the other Index scores. These and other findings indicate general support for the convergent validity of the WISC-IV Index scores. Discriminant validity was confirmed by the negligible relationships among WISC-IV Index scores and measures of emotional intelligence from the BarOn Emotional Quotient Inventory (BarOn EQI, Bar-On & Parker, 2000). For the most part, research has shown that emotional intelligence is independent of cognitive intelligence. Thus, relationships among Index scores from the WISC-IV and subtest scores from the BarOn EQI should bear out as insignificant. In fact, the correlations were negligible, in the range of .06 to .20. The only exceptions were sensible ones. For example, scores on the Adaptability subscale from the BarOn EQI correlated .34

with WISC-IV Full Scale IQ. Certainly, it is plausible that adaptability as measured by the BarOn EQI is rooted, in part, in a foundation of cognitive skills, as mirrored in IQ, thus illuminating the modest correlation between these two measures.

5.17 STANFORD-BINET INTELLIGENCE SCALES: FIFTH EDITION With a lineage that goes back to the Binet- Simon scale of 1905, the Stanford-Binet: Fifth Edition (SB5) has the oldest and perhaps the most prestigious pedigree of any individual intelligence test. In Table 5.8, we outline some important milestones in the development of the SB5 and its predecessors. Released in 2003, the SB5 is a very new test (Roid, 2002, 2003). For this reason, evaluation of this instrument is based, in part, on its resemblance in content and subtests to the SB4, for which a large body of

independent research literature has been amassed. The SB5 Model of Intelligence In early editions of the Stanford-Binet, the examiner obtained only a composite IQ. Although the pattern of right and wrong answers could be analyzed qualitatively, the earlier Stanford-Binet tests (prior to the fourth edition) did not provide a basis for quantitative analysis of the subcomponents of the entire scale. The fourth and fifth editions corrected this shortcoming. TABLE 5.8 Milestones in the Development of the Stanford-Binet and Predecessor Tests Ye ar

Test/Authors Comment

1905Binet and Simon Simple 30-item test 1908Binet and Simon Introduced the mental

age concept 1911Binet and Simon Expanded to include

adults

The organization of the SB5 was guided by the principle that each of five factors of intelligence can be assessed in two distinct domains— nonverbal and verbal. The five factors—derived from modern cognitive theories such as Carroll (1993) and Baddeley (1986)—are fluid reasoning, knowledge, quantitative reasoning,

1916Stanford-Binet Terman and Merrill

Introduced the IQ concept

1937Stanford-Binet-2 Terman and Merrill

First use of parallel forms (L and M)

1960Stanford-Binet-3 Terman and Merrill

Modern item-analysis methods used

1972Stanford-Binet-3 Terman and Merrill

SB-3 restandardized on 2,100 persons

1986Stanford-Binet-4 Thorndike, Hagen, and Sattler

Complete restructuring into 15 subsets

2003Stanford-Binet-5 Roid

Five factors of intelligence

visual-spatial processing, and working memory. When these five factors of intelligence are “crossed” with the two domains (nonverbal and verbal), the result is an instrument with 10 subtests (Figure 5.11). Thus, the SB5 provides a number of different perspectives on the cognitive functioning of an examinee: 10 subtest scores (mean of 10, SD of 3), three IQ scores (the familiar Full Scale IQ, Verbal IQ, and Nonverbal IQ), as well as five factor scores (Fluid Reasoning, Knowledge, Quantitative Reasoning, Visual-Spatial Processing, and Working Memory). The IQ and factor scores are normed to a mean of 100 and SD of 15. Routing Procedure and Tailored Testing The SB5 maintains the historical tradition of this instrument by using a routing procedure to estimate the general cognitive ability of the examinee before proceeding to the remainder of the test. The purpose of the routing procedure is to identify the appropriate starting points for subsequent subtests. The routing items are both nonverbal (object series and matrices) and

verbal (vocabulary). These items also provide the Abbreviated IQ, sometimes used for screening purposes. Roid (2002) describes the advantages of using a routing procedure: This tailored approach to assessment provides

greater richness of factor measurement within a shorter, efficient test administration. The use of modern item response theory in the design of SB5 allows for greater precision of measurement due to the adaption of the test to the functional level of the examinee in an efficient time frame.

FIGURE 5.11 Structure of the Stanford- Binet: Fifth Edition Thus, the purpose of the routing procedure is not just to reduce the number of items administered (and, therefore, save time), but to do so without loss of measurement precision. This is possible because the SB5 was constructed according to the principles of item response theory (Embretson, 1996). When a test is constructed within the framework of item response theory, item difficulty levels and other

parameters are precisely calibrated during the development phase. Special Features of the SB5 In addition to providing a more familiar partition of intelligence into Full Scale IQ, Verbal IQ, and Nonverbal IQ, the SB5 also features a number of other improvements over its predecessor, the SB4. The test now includes extensive high-end items, designed to assess the highest level of gifted performance. Many of these items are updates from very early editions of the Stanford-Binet, when the instrument was renowned for its very high ceiling. At the other extreme, improved low-end items provide better assessment for very young children (as young as age 2) and adults with mental retardation. In addition, the items and subtests that contribute to the Nonverbal IQ do not require expressive language, which makes this part of the test ideal for assessing individuals with limited English, deafness, or communication disorders. The developers of the SB5 also screened test items for fairness based on religious as well as

traditional concerns. Expert panels examined the entire test on fairness issues related to the standard variables (gender, race, ethnicity, and disability) and religious tradition (Christian, Jewish, Muslim, Hindu, and Buddhist backgrounds). This is the first time in the history of intelligence testing that religious tradition has been considered in test development. Finally, the Working Memory factor, consisting of both verbal and nonverbal subtests, shows promise in helping to assess and understand children with attention-deficit/ hyperactivity disorder. Standardization and Psychometric Properties of the SB5 The SB5 is suitable for children age 2 through adults age 85 and older, and the standardization sample consists of 4,800 individuals stratified by gender, ethnic, regional, and educational levels in the United States, based on the year 2000 census. In part because item selection was determined by modern item response theory, the reliability of subtests, indices, and IQ scores is

very strong and comparable to other mainstream individual intelligence tests. For example, the Verbal IQ, Nonverbal IQ, and Full Scale IQ each have reliabilities in the .90s, and the individual subtests are in the range of .70 to .85 (Roid, 2002). As is typical in the release of a new test, the manual for the SB5 (Roid, 2003) reports on numerous affirming correlational studies (e.g., with the Wechsler scales, the SB4, the UNIT) that provide strong support for criterion-related validity. The validity of the test as a measure of general intelligence is also supported by its resemblance to the SB4, about which a large body of research can be cited. For example, Lamp and Krohn (2001) studied the longitudinal predictive validity of the SB4 in a sample of 89 Head Start children (39 African American and 50 white) from impoverished backgrounds who ranged in age from about 4 to 6½. These children were retested several times over an eight-year period on both the SB4 and the Metropolitan Achievement Test. The correlations between the initial SB4 score and

the subsequent achievement scores were very strong (mainly in the .50s), and the test was equally good at predicting outcome for African American and white children. In another study (Atkinson, Bevc, Dickens, & Blackwell, 1992), the concurrent validity of the SB4 was tested against the Leiter International Performance Scale and the Vineland Adaptive Behavior Scales in a sample of 24 children with developmental delays. The correlations were very robust (.78 and .70, respectively). These and many other studies strongly support the validity of the SB4 as a measure of general intelligence. As new research is reported on the SB5, it is likely that this recent edition also will prove to be highly valid and even more useful than its predecessor as a measure of intelligence. In summary, the SB5 is a very promising new test that is especially useful at both ends of the cognitive spectrum—the very young or those with developmental delays, and very gifted persons. Based on the care with which the instrument was constructed, the test is likely to

become a mainstay of individual intelligence testing in a wide variety of settings.

5.18 DETROIT TESTS OF LEARNING APTITUDE-4 The Detroit Tests of Learning Aptitude-4 (DTLA-4; Hammill, 1999) is a recent revision of an instrument first published in 1935. The test is individually administered and designed for schoolchildren from 6 through 17 years of age. The DTLA-4 consists of 10 subtests that form the basis for computing 16 composites, including general intelligence, optimal level, and 14 ability areas. The subtests are largely within the Binet-Wechsler tradition, although there are a few surprises such as the inclusion of Story Construction, a measure of storytelling ability (Table 5.9). The General Mental Ability composite is formed by combining standard scores for all 10 subtests in the battery. The Optimal Level composite is based on the highest four standard scores earned by the examinee and is thought to represent how

well the examinee might perform under optimal circumstances. Each of the remaining 14 composite scores is derived from a combination of several subtests thought to measure a common attribute. For example, subtests that involve knowledge of words and their use are combined to form the Verbal Composite, whereas subtests that do not involve reading, writing, or speech comprise the Nonverbal Composite. Several of the composite scores are designed to represent major constructs within contemporary theories of intelligence. In addition to the General Mental Ability composite and the Optimal Level composite, the remaining 14 DTLA-4 composite scores are as follows: Verbal Nonverbal (Linguistic) Attention- enhanced

Attention- reduced

(Attentional)

Motor-enhanced Motor-reduced (Motoric) Fluid Crystallized (Horn &

Cattell) Simultaneous Successive (Das) Associative Cognitive (Jensen)

The 16 composite scores are based on the familiar mean of 100 and standard deviation of 15. The 10 subtests are normed for a mean of 10 and standard deviation of 3. The composites were designed to offer contrasting assessments such that a difference between scores may be of diagnostic significance. For example, an examinee who scored well on Attention-Reduced aptitude but poorly on Attention-Enhanced aptitude (in the Attentional domain) presumably experiences difficulty with immediate recall, short-term memory, or focused concentration. The DTLA-4 was standardized on 1,350 students whose backgrounds closely matched census data for sex, race, urban/rural residence, family income, educational attainment of parents, and geographic area. The reliability of this instrument is similar to other individual tests of intelligence, with internal consistency coefficients generally exceeding .80 for the subtests and .90 for the composites, and test- retest coefficients for the subtests and the

Verbal Performance (Wechsler)

composites in the .80s and .90s. Criterion- related validity is well established through correlational studies with other mainstream instruments such as the WISC-III, K-ABC, and Woodcock-Johnson. TABLE 5.9 Brief Description of the DTLA-4 Subtests Subtest Task Word Opposite s

Provide antonyms—word opposites.

Design Sequence s

Discriminate and remember nonsensical graphic material.

Sentence Imitation

Repeat orally presented sentences.

Reversed Letters

Short-term visual memory and attention.

Story Construct ion

Create a logical story from several pictures.

Design Reproduc tion

Copy designs from memory.

A concern with the DTLA-4 is that the conceptual breakdown into composites is not sufficiently supported by empirical evidence. For example, while it may be true that the Simultaneous composite does measure the simultaneous cognitive processes proposed by Das, Kirby, and Jarman (1979), there is scant empirical support to buttress this claim. Another problem with this instrument is that there are more composites than there are subtests! Inevitably, the composites will be highly intercorrelated, because each subtest occurs in

Basic Informati on

Knowledge of everyday facts and information.

Symbolic Relations

Select from a series of designs the part that was missing from a previous design.

Word Sequence s

Repeat a series of unrelated words.

Story Sequence s

Organize pictorial material into meaningful sequences.

several composites. In sum, DTLA-4 may be a good measure of general intelligence, but the use of composite scores for purposes of psychoeducational planning requires additional empirical study. Smith (2001) provides a thorough review of the DTLA-4.

5.19 THE COGNITIVE ASSESSMENT SYSTEM-II The Cognitive Assessment System-II (CAS-II) is an individually administered test of cognitive abilities designed for children and adolescents ages 5 through 17 (Naglieri, Das, & Goldstein, 2012). The CAS-II was explicitly constructed to embody the Planning, Attention, Simultaneous, and Successive (PASS) theory of intelligence discussed at the beginning of the chapter (Das, Kirby, & Jarman, 1979; Das, Naglieri, & Kirby, 1994). The Standard Battery consists of 12 subtests and takes about 60 minutes to complete (Figure 5.12). A shorter version of eight subtests is available, but most practitioners recommend

the full battery because it provides a better picture for diagnosis and intervention. The CAS-II provides a standard score (mean of 100, SD of 15) for each of the four process scales (Planning, Attention, Simultaneous, and Successive), as well as a Full Scale standard score. The 12 subtests are normed to a mean of 10 and SD of 3. The Planning Scale is a measure of the ability to develop strategies for task completion. For example, in the Matched Numbers subtest, the child views rows of six numbers and is instructed to underline the two numbers in each row that are identical. The numbers increase in length from one digit to seven digits. The subtest score is based on a combination of time to completion and number correct. In the Planned Codes subtest, the task is to learn a code depicted at the top of the page (such as A goes with X-O, B goes with O-O, C goes with X-X, D goes with O-X) and then fill in missing codes in the remainder of the page (for example, A _ _, C_ _, B_ _, A_ _, D_ _, etc.). In the Planned Connections subtest (a variation of the Trail Making Test, part B,

Reitan & Wolfson, 1993), the child draws a pencil line to connect randomly placed numbers and letters in sequential order, alternating between numbers and letters (1-A-2-B-3-C, etc). The Planning subtests involve cognitive control and self-regulation.

FIGURE 5.12 Cognitive Assessment System- II Scales and Subtests

The Attention Scale is a measure of the mental processes involved in resistance to distraction and focused attention over time. For example, in the Expressive Attention subtest, a variation of the Stroop procedure (Stroop, 1935), the child first reads a long list of color words (Blue, Yellow, Red, Green) repeated in random order, then quickly names blocks of color printed in these four colors. These tasks are preamble to the final task, the only part that is scored. In the final section of the Expressive Attention subtest, a lengthy list of the color words (Blue, Yellow, Red, Green) is presented, each word printed in a competing color (e.g., the word Blue printed in red ink), with instructions to name the colors, not read the words. The raw score is the ratio of the total number correct to the time needed for completion of the last section. In the Number Detection subtest, the child is required to underline specific digits in particular fonts, for example, the task might be to detect the numbers 1, 2, and 3 among random digits, but only when printed in bold font. In the Receptive Attention subtest, the child first underlines letter

pairs that are physically the same (e.g., TT but not Tt) and then underlines letter pairs that are the same name (e.g., Bb but not Ba). The score is based on accuracy and total time. The Simultaneous Scale is a measure of the ability to organize information into coherent wholes. Both nonverbal and verbal processes are utilized to analyze and synthesize spatial and verbal relationships. Nonverbal Matrices is a variation on the familiar matrix reasoning task first employed in the Raven Progressive Matrices (Raven, 1938) and found in many intelligence tests. A 3 × 3 matrix of geometric shapes is shown, with a missing shape in the lower right-hand corner. Below the matrix are six shapes, one of which completes the rules of progression in the matrix from left to right and top to bottom. Based on inference, the task is to choose the correct shape. In the Verbal Spatial Relations subtest, the child views six drawings, each depicting a particular spatial relationship between shapes, and then encounters a series of printed question such as Show me the square to the right of the circle. The task is to choose the

one drawing among six that depicts the relationship. In the Figure Memory subtest, the child views a two- or three-dimensional drawing for five seconds, and then must correctly locate the original drawing embedded within a larger, more complex drawing. The Simultaneous subtests involve the perception of stimuli as a whole, in contrast to what is needed in successive processing. The Successive Scale involves mental processes needed to remember and complete a task in a specific order or sequence. In Word Series, the task is to recall in correct order a series of two to nine words orally presented at one word per second. This task is similar to measures of digit span, except words are used instead of digits. The same nine words (one-syllable, high- frequency words such as Car, Dog, Shoe) are used. In the Sentence Repetition subtest, the child reads 20 sentences aloud, one by one. After each sentence is read, the child is asked to repeat it exactly, word for word, after the sentence is withdrawn from view. Color words are used so as to minimize meaning (e.g., The

green is yellowing). The sentences are of varying lengths. The raw score is the number of words correctly recalled. For younger children (ages 5 to 7), the child repeats a specific three- word combination (like cat-book-ball) 10 times in quick succession. The raw score is the total time required. In the Sentence Questions subtest (ages 8 to 17), the child answers questions about orally presented sentences similar to those used in Sentence Repetition (e.g., The green is yellowing. Who is yellowing?). For younger children (ages 5 to 7), Speech Rates is administered instead. This subtest requires the repetition of a one-syllable and two-syllable word combination 10 times as quickly as possible. The raw score is the total time needed to complete the repetitions. Correct sequencing of stimuli or activities is essential to the Successive subtests. In addition to 12 subtest scores and 4 process scores, The CAS also yields a Full Scale score based on the familiar mean of 100 and SD of 15. Psychometric properties of the test are excellent. The average internal consistency reliabilities

are: Planning (.88), Attention (.88), Simultaneous (.93), Successive (.93), and Full Scale (.96). The standardization sample consisted of 2,200 children and adolescents, stratified on demographic variables to closely match the U.S. population (Naglieri, Das, & Goldstein, 2012). The validity of the CAS-II rests in large measure on its similarity to the first edition, the CAS, which stands up well in factor analytic studies and yields meaningful results for special groups. For example, using multigroup confirmatory factor analysis, Naglieri, Taddei, and Williams (2012) found that the factorial structure of the CAS was highly similar in two cross-cultural samples, one comprised of 1,174 U.S. children and the other consisting of 809 Italian children. Further, results for both samples were broadly supportive of the four factors of the PASS theory embodied in the CAS. In a study of 60 children meeting the criteria for Attention-Deficit Hyperactivity Disorder (ADHD), Naglieri and colleagues found that subtest and process scores were theoretically

consistent with current understandings of ADHD. Specifically, average scores on the four process scales were: Planning 89.1, Attention 92.3, Simultaneous 101.2, and Successive 101.7 (Naglieri & Paolitto, n.d.). These findings fit well with the hypothesis that children with ADHD manifest problems with goal-directed planning and show difficulties with attention due to distractibility (Barkley, 1996). An intriguing result with the CAS is that differences between Black and White children on the Full Scale score are minimal when key demographic variables such as socioeconomic status are controlled. Naglieri, Rojahn, Matto, and Aquilino (2005) found an estimated CAS Full Scale mean score difference of 4.8 points between Black (N = 298) and White (N = 1,691) children, smaller than typically reported with traditional IQ tests. The relationships between CAS scores and school achievement were strongly positive and highly similar for both groups as well. Overall, these results indicate that the CAS is useful for assessment in special education. On a similar note, Naglieri and

Rojahn (2001) found that CAS scores classified a smaller proportion of Blacks as having intellectual disability than did WISC-III scores. They argued that the problem of disproportionate representation of Blacks in special education classes might be mitigated if the CAS were used for this assessment purpose. The CAS-II is a promising test that deserves to see wider use in assessment and research.

5.20 KAUFMAN BRIEF INTELLIGENCE TEST-2

(KBIT-2) The individual intelligence tests previously discussed in this and the preceding topic are excellent measures of intellectual ability, but they are not without their drawbacks. One problem is the time required to administer them. Testing sessions with the Wechsler scales, Kaufman Assessment Battery for Children, and the Stanford-Binet easily can last one hour, and two hours is not unusual if the examinee is

bright and highly verbal. A second disadvantage to these mainstream tests is the amount of training required to administer them. Proper administration of most individual intelligence tests is based upon the assumption that the examiner has an advanced degree in psychology or a related field and has received extensive supervised experience with the instruments in question. Alan Kaufman responded to the need for a brief, easily administered screening measure of intelligence by developing the Kaufman Brief Intelligence Test (K-BIT), recently released in a second edition, the KBIT-2 (Kaufman & Kaufman, 2004). The KBIT-2 consists of a Verbal or Crystallized scale that includes two types of items (Verbal Knowledge and Riddles) and a Nonverbal or Fluid Scale that consists of Matrices items (2 × 2 and 3 × 3 figural analogies). The KBIT-2 is normed for examinees ages 4 to 90 and can be administered in approximately 20 minutes. The test yields standard scores with means of 100 and standard deviation of 15 for

Verbal, Nonverbal, and combined scores. In spite of the comparability of these scoring dimensions with well-known intelligence tests, the KBIT-2 authors make it clear that their instrument is not intended as a substitute for traditional approaches (e.g., WPPSI-III, KABC-2, WISC-IV, or SB5). The KBIT-2 is mainly a screening test useful in signaling the need for more extensive assessment. The brevity of this test makes it a natural choice for research on intelligence. The test authors suggest a number of uses for the instrument, including the following: • Provide a quick estimate of intelligence

where accuracy is not essential • Estimate verbal versus nonverbal

intelligence in children or adults • Reevaluate intellectual status of previously

tested examinees • Screen students who may benefit from

placement in gifted programs • Screen high-risk students who may need

further assessment

• Obtain a quick estimate of intelligence in adult treatment or institutional settings

The KBIT-2 manual reports highly supportive validity data from numerous correlational studies. However, the most compelling evidence for the validity of the instrument is its strong resemblance to the K-BIT, for which a substantial body of research has been published. For example, Naugle, Chelune, and Tucker (1993) compared K-BIT results and WAIS-R scores for 200 referrals to a neuropsychological assessment center. The patient sample included persons with seizure disorders, head injuries, substance abuse, psychiatric disturbance, stroke, dementia, and other neurological conditions. The heterogeneity of the referral sample guaranteed a wide range of functional ability, a desirable feature in a validation study. Although the K- BIT scores tended to be about 5 points higher than their WAIS-R counterparts, the correlations between these two instruments were extremely high and theory-confirming. Vocabulary IQ (K- BIT) and Verbal IQ (WAIS-R) correlated .83;

Matrices IQ (K-BIT) and Performance IQ (WAIS-R) correlated .77; and overall IQs from the two instruments correlated an amazing .88. In a study comparing the K-BIT and the WISC- III scores for 50 referred students, Prewett (1995) also reported strong correlations (r = .78 for overall scores) and also discovered that the K-BIT scores tended to be about 5 points higher than their WISC-III counterparts. In a sample of 65 children with reading disability, Chin, Ledesma, Cirino, and others (2001) also found that the K-BIT overestimated WISC-III IQs by 1.2 to 5.0 points, on average. However, their study also showed that, in individual cases, K- BIT scores can underestimate or overestimate WISC-III scores by as much as 25 points, reaffirming that the K-BIT is not appropriate for placement and diagnostic purposes. Canivez (1995) found comparable scores between the K- BIT and the WISC-III for 137 elementary and middle school children and also reported very strong correlations between the two tests, especially for overall scores (r = .87). Eisenstein and Engelhart (1997) found that the K-BIT

performed well in estimating IQs in adult neuropsychology referrals, but Donders (1995) recommends caution when using the test with brain-injured children. The reason for caution is that K-BIT scores show a negligible relationship with length of coma; that is, the test is not a good index of neuropsychological status in children. In spite of these cautions about its predecessor, the KBIT-2 is an outstanding screening measure of general intelligence for use in research or in those situations listed earlier in which time constraints preclude use of a longer instrument.

5.21 INDIVIDUAL TESTS OF ACHIEVEMENT Whereas intelligence tests are designed to measure the broad mental abilities of the individual, achievement tests are intended to appraise what a person has learned in school or some other course of study. Group achievement tests are paper-and-pencil measures given to dozens of students at a time. These kinds of

measures are discussed in Topic 6A, Group Tests of Ability and Related Concepts. Our focus here is on individual achievement tests administered one-on-one and, therefore, better suited for the appraisal of learning problems. Of course, scores on intelligence and achievement tests should bear a strong relationship to one another—brighter children likely are capable of higher achievement. In fact, as we shall see, the notion that intelligence and achievement typically parallel one another is at the very heart of the concept of learning disability—which commonly involves a discrepancy between the two. We introduce the reader here to the makeup of individual achievement tests as a backdrop to the final topic in this chapter, the assessment of learning disabilities. More than a dozen individually administered intelligence tests exist, but only a few are widely used in clinical and educational assessment. A number of prominent individual achievement tests are summarized in Table 5.10. Owing to limitations of space, we have selected

one test, the Kaufman Test of Educational Achievement-II (KTEA-II), for more detailed presentation (Kaufman & Kaufman, 2004b). Readers who seek further information are encouraged to consult Sattler (2001, Chapter 17) or the Mental Measurements Yearbook series. Kaufman Test of Educational Achievement-II (KTEA-II) The KTEA-II is an untimed test of educational achievement for children ages 4½ through 25. A brief, three-subtest version exists and extends the age range to 90+, but for diagnostic assessment of learning difficulties the Comprehensive Form is preferred. The core of the KTEA-II Comprehensive Form consists of eight subtests in four areas: • Reading

• Letter and Word Recognition • Reading Comprehension

• Mathematics • Math Concepts and Applications • Math Computation

• Written Language

• Written Expression • Spelling

• Oral Language • Listening Comprehension • Oral Expression

In addition to yielding scores on each subtest, the battery provides three composite scores (Reading, Mathematics, and Written Language) and a Total Battery Composite. For diagnostic purposes, a number of supplemental subtests designed to evaluate reading skills (e.g., Phonological Awareness) are also available. For older children, the test takes about 80 minutes to administer; for younger children about 30 minutes are needed. The KTEA-II is co-normed with the KABC-II. Brief examples of KTEA-II-like items are shown in Table 5.11. These examples would be at the upper end of the subtests, suitable for high school students. The KTEA-II utilizes entry and exit rules for each subtest to ensure that students only encounter items of appropriate difficulty. Scoring is objective and highly reliable. Raw scores are converted to standard scores (mean of

100, SD of 15) for each subtest, the composite scores, and the Total Battery Composite. TABLE 5.10 Survey of Widely Used Individual Achievement Tests

Diagnostic Achievement Battery-3 (DAB-3) (Newcomer, 2001) Suitable for ages 6 through 14, the DAB-3 consists of 14 subtests used to compute eight diagnostic composites. The composite scores include Listening, Speaking, Reading, Writing, Mathematics, Spoken Language, Written Language, and Total Achievement. More comprehensive than most achievement tests, the DAB-3 takes up to two hours to administer. The test was carefully normed on 1,534 children nationwide. Kaufman Test of Educational Achievement (KTEA-II) (Kaufman & Kaufman, 2004b) A well-normed individual test of educational achievement, a special feature of the KTEA-II is the detailed error analysis (see text). Currently, norms extend from age 4½ through age 25. A separate brief form that can be administered in 30 minutes or less is useful for screening purposes. Mini-Battery of Achievement (MBA) (Woodcock, McGrew, & Werder, 1994) Assesses four broad achievement areas— reading, writing, mathematics, and factual

In addition to formal scoring, the KTEA-II provides a systematic method for evaluating the qualitative nature of subtest errors. For example, on the Spelling subtest, errors can be classified according to whether they involve prefixes, suffixes, vowel digraphs (such as ue in blue) and diphthongs, consonant clusters (such as scr in unscrupulous), r-controlled patterns (such as er in inferior), and several other patterns. TABLE 5.11 Examples of Characteristic KTEA-II Items Applicable to Older Children

Kaufman and Kaufman (2004b) stress that the error analysis provides the diagnostician with a source of information from which instructional objectives can be developed. For example, a weakness in vowel digraphs and diphthongs on the Spelling subtest translates directly to classroom objectives: practice in the spelling and reading of these elements in isolation, progressing to spelling and pronouncing words containing digraphs and diphthongs, and ending in writing and reading sentences containing words with vowel digraphs and diphthongs. The KTEA-II manual contains many useful clinical insights with educational ramifications. The content validity of the KTEA-II appears to be very strong, but this point may vary from one school system to another. After all, individual school systems may choose to emphasize different domains of achievement. Salvia and Ysseldyke (1991) warn that users must be sensitive to the correspondence of test content with the students’ curriculum. As with any achievement test, the user should verify that the content of the KTEA-II is appropriate within the

curricular setting. Nonetheless, Kaufman and Kaufman (2004b) offer sufficient evidence for the validity of the test to make a case for general adequacy.

5.22 NATURE AND ASSESSMENT OF LEARNING DISABILITIES Because individual intelligence and achievement tests are foundational to the assessment of learning disabilities, we close this chapter with brief review of this topic. The learning disability (LD) field is one of the fastest growing areas within assessment. Paradoxically, it is also one of the most controversial and perplexing domains of psychological testing. Some background is needed to understand the role of intelligence and achievement tests in the evaluation of learning disabilities. We begin by asking a seemingly simple question that turns out to have a

complicated answer: What is a learning disability? Definitions of learning disability have gone through at least three phases in the last several decades. Early views were influenced heavily by federal legislation and relied on a discrepancy between intelligence and achievement as the defining characteristic. These ideas were followed by a model that featured intra-individual weakness in one or more core psychological processes as the essential attribute. Most recently, responsiveness to intervention has been featured as the prevailing quality. We turn now to a survey of these shifting paradigms in the history of LD assessment. The Federal Definition of Learning Disabilities For decades the essential nature of learning disabilities was understood in terms of a definition embedded in federal law. In 1975, Congress passed Public Law 94-142, the Education for All Handicapped Children Act.

One of the provisions of this act was a definition of learning disabilities as follows: The term “specific learning disability” means a

disorder in one or more of the basic psychological processes involved in understanding or in using language, spoken or written, which may manifest itself in imperfect ability to listen, speak, read, write, spell, or to do mathematical calculations. The term includes such conditions as perceptual handicaps, brain injury, minimal brain dysfunction, dyslexia, and developmental aphasia. The term does not include children who have learning disabilities which are primarily the result of visual, hearing, or motor handicaps, of mental retardation, or emotional disturbance, or of environmental, cultural, or economic disadvantage. (USDE, 1977, p. 65083)

The commitment to a federally mandated definition was reaffirmed in 1990 by passage of Public Law 101-476, the Individuals with Disabilities Education Act (IDEA).

The federal definition embodied in IDEA also stipulated an operational approach to the identification of children with learning disabilities. Specifically, candidates for an LD diagnosis had to demonstrate a severe discrepancy between general ability (intelligence) and specific achievement in one or more of these seven areas: • Oral expression • Listening comprehension • Written expression • Basic reading skill • Reading comprehension • Mathematics calculation • Mathematics reasoning

The discrepancy model for the identification of LD children functioned as a directive for school psychologists. In effect, the model mandated that psychologists should administer an individual intelligence test (general ability measure) and an individual achievement test (specific achievement measure) and then look for a discrepancy between Full Scale IQ and one

or more areas of school achievement (e.g., reading, mathematics, written expression). In practical terms, a severe discrepancy was defined as a difference of one standard deviation or more between general intelligence and specific achievement. A common practice in identification of LD children was to compare Full Scale IQ on an individual intelligence test such as the WISC-III with specific achievement scores on an individual achievement test such as the WIAT (Wechsler Individual Achievement Test) or similar instrument that has subtests normed with a mean of 100 and a standard deviation of 15. A difference of 15 points or more between Full Scale IQ and specific achievement in any of the previously listed areas would then raise the suspicion of learning disability. Unfortunately, the federal definition did not serve its intended purposes, and, increasingly, school psychologists and other professionals looked to other approaches for understanding and assessing learning disabilities in children. The fundamental problem was that many, many

children who exhibit serious learning problems in school and who would benefit from services for LD simply did not meet the psychometric criteria of a severe discrepancy. The National Joint Committee on Learning Disabilities Definition After a lengthy period of confusion and struggle over the definition of learning disabilities, specialists and educators began to rally around a consensus view in the early 1990s. The new definition was proposed by the National Joint Committee on Learning Disabilities (NJCLD), a group of representatives from eight national organizations with a special interest in learning disabilities. Although similar to the federal definition, the new approach contains important contrasts: Learning disabilities is a general term that refers to a heterogeneous group of disorders manifested by significant difficulties in the acquisition and use of listening, speaking, reading, writing, reasoning, or mathematical abilities. These disorders are intrinsic to the

individual, presumed to be due to central nervous system dysfunction, and may occur across the life span. Problems in self-regulatory behaviors, social perception and social interaction may exist with learning disabilities but do not by themselves constitute a learning disability. Although learning disabilities may occur concomitantly with other handicapping conditions (for example, sensory impairment, mental retardation [MR], serious emotional disturbance [ED]) or with extrinsic influences (such as cultural differences, insufficient or inappropriate instruction), they are not the result of those conditions or influences. (NJCLD, 1988, p. 1) The new definition avoided vague reference to “basic psychological processes,” specifies that the disorder is intrinsic to the individual, identifies central nervous system dysfunction as the origin of LD problems, and states explicitly that learning disabilities may extend into adulthood. Perhaps most important of all, the NJCLD approach abandoned the excessive reliance upon

discrepancy between ability and achievement as the hallmark of LD. Instead, the new model specified that the necessary (but not sufficient) condition of LD was that the individual (child or adult) exhibit an intraindividual weakness in one or more of the core areas of academic functioning (listening, speaking, reading, writing, reasoning, or mathematical abilities). Shaw et al. (1995) described how the NJCLD model might look in practice. In this approach, the first task is to identify one or more intraindividual weaknesses in the core areas. These are always relative to strengths in several other core areas. In other words, persons who are slow learners in all areas do not meet the criteria of LD. The second step is to trace the learning difficulties to central nervous system dysfunction, which may manifest as problems with information processing. For example, a young adult with a severe weakness in listening (as judged by her inability to learn from the traditional lecture approach to teaching) might exhibit a deficit on a test of verbal memory— confirming that an information processing

problem was at the heart of her disability. The purpose of the third step (examining psychosocial skills, physical and sensory abilities) is to specify additional problems that may need to be addressed for program-planning purposes. Finally, in the fourth step the examiner rules out non-LD explanations for the learning difficulties (since these explanations would mandate a different strategy for remediation). The New Face of Learning Disabilities: Response to Intervention In 2004, Congress reauthorized the Individuals with Disabilities Education Act (IDEA), which is the ongoing legislation governing special services, including the assessment of LD, in school systems that receive federal funding. IDEA 2004 changed the law about how to identify children with specific learning disabilities by moving away from the discrepancy model that had reigned supreme since the 1970s. Instead, the new law recommended response to intervention (RTI)

as the preferred method for identifying children with learning disabilities. In particular, IDEA 2004 says that a school “may use a process that determines if the child responds to scientific, research-based intervention as part of the evaluation procedures . . . ” in its evaluation for LD. RTI is a broader concept than LD and refers both to (1) methods for increasing the capacity of school systems to respond effectively to the diverse academic needs of students and (2) approaches for identifying LD children who need special education services. The RTI approach specifically deemphasizes cognitive discrepancies in the diagnostic process, focusing instead on low age-based achievement levels and failure to respond to evidence-based instructional approaches (Fletcher & Vaughn, 2009; Torgerson, 2009). The implementation of RTI is complicated and multifaceted. The process involves multiple feedback loops and decision points. Yet, proponents of RTI view it as an improvement because it provides for early, preventive

intervention in contrast to the “wait to fail” approach of the discrepancy model. Fuchs and Fuchs (2005) describe a systematic approach to using RTI in a school system. The first step is school-wide screening in the first weeks of the school year to identify children “at risk” for school failure. Those scoring below a certain prescribed cut-off (perhaps the 25th percentile in reading or math) would be noted. Teachers would then implement empirically validated curricular interventions for these children, who would be monitored for progress after eight weeks. Those who do not respond would receive another interval of supplementary instruction for an additional eight weeks. Those who still do not respond would receive a comprehensive, individualized evaluation to rule out sources of underachievement such as intellectual disability, visual problems, or emotional disturbance. Finally, with the involvement of parents, the child would receive a designation of LD and become eligible for special education placement.

In sum, RTI is a shift in perspective that focuses on early results and outcomes with at-risk children instead of later spending excessive time and resources on questions of discrepancy-based eligibility after children are already failing because of their LD. The hope is that the RTI perspective will catch at-risk children earlier and thereby reduce the number of children needing special education services. Essential Features of Learning Disabilities Even though the definition of LD remains a point of contention, we can cite several features of these disorders that are less controversial. As the reader will discover, the features discussed in the following dictate, to some extent, the nature of testing practices in the assessment of learning disabilities. There is general agreement —with occasional dissenting votes—on five features of learning disabilities. First, a learning disability involves an intraindividual discrepancy in cognitive functioning. The child (or adult) with LD

reveals a relative weakness in one area compared to strengths in most other areas. According to the federal definition followed within many school systems, the discrepancy is between general ability (intelligence) and specific achievement. We have described previously some of the pitfalls of this definition and prefer the NJCLD approach in which the discrepancy is not rigidly tied to a difference between IQ and achievement test scores. Second, an exclusionary clause is included in most definitions of learning disability. If the academic difficulties are primarily caused by other disabling conditions (mental retardation, emotional disturbance, visual or hearing impairment, cultural or social disadvantage), then a diagnosis of learning disability is typically ruled out. This clause is often misinterpreted. A person can be both learning disabled and impaired in other ways (e.g., have mental retardation). The important point is that the coexisting condition must not be the primary cause of the learning difficulties.

Third, Learning disabilities are heterogeneous; that is, there are many different varieties. Research on the identification of subtypes is still in its infancy, but most researchers express optimism that meaningful subgroups of persons with learning disabilities can be identified. Pending further research and refinement, only two broad categories of learning disability are recognized currently. These two types are dyslexia or verbal learning disability, and right hemisphere or nonverbal learning disability. Our coverage here is based on Forster (1994). The primary manifestation of dyslexia is an unexpected difficulty in learning to read or spell. The fundamental deficiency is thought to be a problem with phonological coding, which is the ability to automatically associate sounds with specific letter combinations. Verbal learning disability constitutes about 90 percent of all LD cases, and is much more common in boys than girls. In contrast, right hemisphere or nonverbal learning disability manifests as poor skills in mathematics, handwriting, and, often, social cognition. The fundamental problem is

thought to be a problem in spatial cognition, which is the visuospatial perception of relationships. The problem likely originates in right cerebral hemisphere dysfunction, and constitutes about 10 percent of all LD cases. Boys and girls are equally affected. Fourth, a learning disability is a developmental phenomenon that is usually evident in early childhood that may persist into adulthood. Even though remediation efforts should be based upon optimism—so as to avoid self-fulfilling prophecies—a dose of realism is needed, too. Longitudinal studies of children with severe learning disabilities suggest that marked improvement in academic achievement is the exception, not the rule, even when these subjects receive intensive educational intervention. For example, Frauenheim and Heckerl (1983) re-tested 11 adults diagnosed as having learning disabilities in childhood. All the participants had received special help for reading; nine had graduated from high school, and two completed the 10th grade. Full Scale IQs were typically in the low 90s, with Verbal

IQ below average (mean of 85) and Performance IQ above average (mean of 104). In spite of the remedial intervention, when retested as adults on exactly the same achievement test (Wide Range Achievement Test), these examinees were scarcely improved from their elementary school results. These findings are corroborated by several other follow-up studies (see Kolb & Whishaw, 1990, chap. 29, for a review). Such results indicate that specialists who work with children with learning disabilities should not become fixated solely on academic concerns. Social and emotional problems—which may be more amenable to intervention—also cry out for notice. Fifth, individuals with learning disabilities frequently experience social and emotional difficulties that are as pervasive and consequential as the deficits in academic achievement. These problems may persist into adolescence and adulthood. In fact, the socioemotional sequelae often become the primary presenting complaint, which can

complicate the testing process and obscure the diagnosis. For example, in a needs assessment study of 381 adults with learning disabilities, Hoffman, Sheldon, Minskoff, and others (1987) identified several crucial nonacademic areas meriting intervention by service providers. These adults self-endorsed several social and emotional problems with high frequency: feeling frustrated (40 percent), talking or acting before thinking (33 percent), being shy (31 percent), no self-confidence (28 percent), controlling emotions and temper (28 percent), and dating (27 percent). Many other problems were also endorsed, but by less than 25 percent of the sample. These findings indicate that learning disability assessments should incorporate measures of social and emotional functioning. Vaughn and Haager (1994) provide an excellent overview on the measurement of social skills in persons with learning disability. Causes and Correlates of Learning Disabilities

Approximately 4 to 5 percent of all school-aged children receive a diagnosis of LD, so this is not a rare problem (Lyon, 1996). The most common form of LD is dyslexia, and boys outnumber girls by about 3:1 or 4:1 (Forster, 1994). In a minority of cases, the etiology is clear and can be attributed to a specific cause such as a known brain injury. Left hemisphere impairment is especially likely to result in verbal difficulties, whereas right hemisphere impairment may lead to problems with spatial thinking or other nonverbal skills. Thus, head injury or other neurological problems can be the proximate cause of a child receiving an LD diagnosis. However, in the majority of cases the direct etiology of LD problems is unclear. A number of possibilities have been proposed and these may explain some but not all cases of LD. For example, pathological neurodevelopmental processes have been identified in some persons with severe dyslexia (Culbertson & Edmonds, 1996). Individuals with this disorder appear to have alterations in brain structures such as the planum temporale (the flat surface on the top of

the temporal lobes) known to be important for language processing. Whereas in normal individuals the planum temporale is much larger in the left temporal lobe than in the right, persons with severe dyslexia do not show this pattern of asymmetry (tending toward symmetry instead). Moreover, researchers have identified microscopic cortical malformations called polymicrogyria (numerous small convolutions) that parallel these structural differences. Several postmortem studies of persons with severe dyslexia have revealed these deviations at the cellular level. Spreen (2001) provides an outstanding review of the possible neurological substrates of learning disabilities. Dyslexia also appears to show a significant genetic component for some persons such that the idea of familial dyslexia needs to be taken seriously. However, what must be emphasized is that for most individuals the etiology of LD (whether dyslexia or other forms) remains a mystery. Achievement Tests in LD Assessment: A Final Word

Learning disabilities manifest primarily as academic problems; that is, a child with LD is typically unable to master skills important for school success such as reading, mathematics, or written communication. Because school-based accomplishment is at the heart of the problem, an evaluation for LD must include relevant measures of academic achievement. Furthermore, the evaluation of school achievement—one small part of an LD assessment—must be based on an individual test of achievement. Even though a group achievement test might raise the suspicion of a learning disability, practitioners must rely on individual achievement tests for definitive assessment. Individual achievement tests typically are administered one-on-one with the examiner sitting across from the respondent and posing structured questions and problems. Of course, any well-standardized achievement test will yield normative data about the functioning of a schoolchild. But the special virtue of individual achievement tests is that the examiner can

observe the clinical details of deficient (or superior) performance and form hypotheses about the cognitive capacities of the examinee. Consider the problem of poor spelling, widely observed in children and adults with verbal LD. Any good spelling achievement test will document the disability; however, little insight is gained from mere scores. What the examiner should seek to know is the qualitative nature of the problem, not just its quantitative dimensions. Individual achievement tests are invaluable in this regard. By observing the details of deficient performance, an astute examiner can form hypotheses about the origin of an achievement problem. For example, a child whose spelling is phonetically correct is at least hearing the words correctly, whereas a child with nonphonetic spelling might very well display a problem with auditory processing of speech sounds.