Wechsler Adult Intelligence Scale - Fourth Edition
Abbreviation: WAIS-IV
Background
The Wechsler Adult Intelligence Scale - Fourth Edition (WAIS-IV) has been available in a German adaptation since 2012. It has retained its English title.1) The original instrument was published in the USA in 2008.2) David Wechsler, an American psychologist, constructed the first intelligence test battery of this series in the late 1930s with the Wechsler-Bellevue Scale3); it was intended primarily for clinical purposes. Counting the Wechsler-Bellevue Scale, the WAIS-IV is the fifth generation of the Wechsler intelligence test series.
The predecessors from the second and third generations of the Wechsler tests, the WAIS4) and WAIS-R5), already offered, with their 11 subtests, a good overview of the structure of cognitive performance, especially in cognitively impaired patients. In the fourth generation, the WAIS-III6), the composition of the subtests was changed more substantially for the first time in order to better meet the needs of neuropsychological examinations. Besides the breakdown of test performance into a verbal and a performance portion that had been customary until then, the 14 subtests of the WAIS-III allowed a breakdown according to neuropsychological criteria, which connected the test to neuropsychologically oriented research on the structure of cognitive performance. The WAIS-IV essentially adopts the structure of the WAIS-III and summarizes the subtest performances in four index scores (Verbal Comprehension, Perceptual Reasoning, Working Memory, and Processing Speed) and a Full Scale IQ. The breakdown into verbal and performance portions is dropped.
The test kit of the German version of the WAIS-IV includes two test manuals. One contains the instructions necessary for administration, the other information on foundations, scoring, and interpretation. Further application-oriented literature on the clinical interpretation of the WAIS-IV is currently available only in English.7)8)9)
Structure and subtests
Table 1 gives an overview of the subtests, their assignment to cognitive domains, their abbreviation, the rank order in test administration, and the possible raw score range.
Table 1: Domains and subtests of the WAIS-IV (German language version)
| Domain | Subtest | Abbreviation | Order | Raw score range |
|---|---|---|---|---|
| Verbal Comprehension (VC): | Similarities | SI | 2 | 0-36 |
| Vocabulary | VC | 5 | 0-57 | |
| Information | IN | 9 | 0-26 | |
| Comprehension | CO | (13) | 0-36 | |
| Perceptual Reasoning (PR): | Block Design | BD | 1 | 0-66 |
| Matrix Reasoning | MR | 4 | 0-26 | |
| Visual Puzzles | VP | 8 | 0-26 | |
| Figure Weights | FW | (12) | 0-27 | |
| Picture Completion | PC | (15) | 0-24 | |
| Working Memory (WM): | Digit Span | DS | 3 | 0-30 |
| Arithmetic | AR | 6 | 0-22 | |
| Letter-Number Sequencing | LN | (11) | 0-30 | |
| Processing Speed (PS): | Symbol Search | SS | 7 | 0-60 |
| Digit Symbol Coding | DSC | 10 | 0-135 | |
| Cancellation | CA | (14) | 0-72 |
The subtests in order positions 1 to 10 are referred to as core subtests; those in positions 11 to 15 are optional. Among the core subtests, Visual Puzzles is new; among the optional ones, Figure Weights and Cancellation are new. Object Assembly and Picture Arrangement from the WAIS-III are no longer included in the WAIS-IV. In the WAIS-IV, all 15 subtests are assigned to an index, whereas in the WAIS-III some remained unassigned. Details can be explored by consulting the structure table in the section on the WAIS-III for comparison.
At this point, the German version includes some notes on domain naming and spelling. Neither is relevant to this translation.
Test material and copyright
Administering the WAIS-IV requires the complete test set, which, besides the two manuals and an initial supply of record forms, also contains task booklets and test material. Only the record forms and the special task booklets for certain subtests are needed on an ongoing basis as consumables. The material is protected by copyright and is, moreover, supplied only to certain professional groups. It is available via the publisher's German web store or via the Testzentrale (test distributor) in Germany or Switzerland. The TDB2Online article on the WAIS-III also contains some information on the copyright history of the Wechsler tests.
Test administration
All intelligence test batteries by Wechsler are individually administered tests. The test material is still relatively varied in the WAIS-IV, even though in earlier versions, with the subtests Picture Arrangement and Object Assembly, it was more action-oriented for examinees. Variety in the material always helps to sustain the usually rather moderate motivation of examinees in clinical testing situations.
Administration is described well and in great detail in the administration manual (Manual 2). In the WAIS-IV it is more difficult than in many other ability tests and must be learned and practiced. All subtests from the Verbal Comprehension domain, for example, are answered freely. This is good for the examinee, but it means for the test administrators that the response must be compared with sample responses in the manual and scored. One must also know in which cases to ask for clarification. It is not sufficient to read afterwards that one should have asked a follow-up question for a particular response in order to arrive at a clear scoring. Many subtests have skip and discontinue rules with which one must be familiar.
Administering the 10 core subtests of the WAIS-IV takes a good 90 minutes in psychiatric patients; including further subtests, correspondingly longer. In addition there is the scoring time, which, however, depends strongly on the experience of the test administrator.
Test raw scores
The raw scores of the WAIS-IV are defined quite differently across the subtests:
- Number of correct solutions: DS, MR, IN, LN
- Number of correct solutions within a specified time: AR, VP, DSC, FW, PC
- Number of correct minus number of incorrect solutions within a specified time: SS, CA
- Sum of points for the quality of a solution (0, 1, or 2 points per item): SI, VC, CO
- Complex scoring according to correctness and speed of the solution: BD
The raw score ranges also differ in size: the smallest range is found in Arithmetic (0 to 22), the largest in Digit Symbol Coding (0 to 135).
Scaled score transformation
Due to the varying ranges of row scores, raw scores have been converted into so-called scaled scores using tables since the first generation of Wechsler's intelligence tests. In the Wechsler tests, scaled scores are standard scores with a mean of 10 and a standard deviation of 3. The methods for converting raw scores to scaled scores have changed over the generations of Wechsler tests and are described in detail in the corresponding section on the WAIS-III.
In TDB2Online, test results are always presented in the form of age-independent performance scores. For calculating the performance scores, the scaled score transformations of the age groups 20-24 and 25-29 were used for the subtests of the WAIS-IV. The details are explained below.
Norm sources
For the German language WAIS-IV, no normative data other than those reported in the manual exist to date in the German-speaking world. For many nonverbal neuropsychological tests, however, TDB2Online also draws on non-German-language norm sources. It would therefore be quite conceivable to consult other norms for the nonverbal subtests of the WAIS-IV as well, for example the American, British, or French norms, and to combine them meta-analytically with the German ones. A simple comparison of the German norms with other sources would also be quite interesting. This remains a task to be done but has not been carried out so far.
Processing of the normative data
The German manual of the WAIS-IV gives no information whatsoever on how the scaled scores were calculated from the raw score sums. However, the preface reveals that the calculations were carried out by the same Pearson staff member in San Antonio who also produced the norms for the American original version. One may therefore presumably assume that the procedures were identical. The American manual10) describes that “the method of inferential norming” was used. The associated reference refers to an unpublished conference presentation from 2005, but there is a later publication11) that describes the method well. It is a variant of “continuous norming”: first, means, standard deviations, and skewness of the raw scores were calculated separately for each age group and then smoothed across age groups using polynomial regressions. With the moments adjusted in this way, distribution functions were calculated and again smoothed, also across age groups. Their percentiles at the midpoints of the age ranges were ultimately used for converting raw score sums into scaled scores.
The conversion from raw to scaled scores thus essentially represents a somewhat more formalized linearization by percentile-to-normal transformation and a transformation to the scaled score scale (10;3) for each age group. The “continuous norming” merely ensures that there are no gross sample-related jumps in the norms. Unfortunately, the tables in the manual give these conversion data only as whole numbers; many raw scores correspond to the same scaled score. To make the data usable for presentation in TDB2Online, the hard steps of this conversion were smoothed again by a moving average. The order of the moving average depended on the fineness of the raw score scale. At the ends of the distribution, the data were extrapolated so that the moving average could be retained. Subsequently, the data were also adjusted between the age groups. According to the manual, this had already been done when the norms were created. In the data, however, discontinuities in the age trend were occasionally observed that are the result of the limited sample size. They were eliminated by smoothing across the age groups. Here, third-order smoothing was always sufficient. Subsequently, the data for the two age groups 20-24 and 25-29 were averaged. The mean was the basis for calculating the performance scores.
Evaluation of the norming study
The results of 1425 persons from Germany aged between 16 and 90 years entered the norming of the WAIS-IV conducted in 2012. Unlike in the norming of the WAIS-III, persons from Austria or Switzerland were not included. The sample is of considerable size and covers a large age range. For each age group individually, a breakdown by sex, educational qualifications, migration background, and region is provided.
For sampling, sampling plans were developed on the basis of unspecified publications of the Federal Statistical Office on sex, education, migration background, and regional distribution. The extent to which the characteristics of the normative sample meet the targeted quotas is not described. Further information is missing, among other things on sample attrition (far more than 1425 persons were tested), on the type of sampling, on the type of contact, on the testing environment, and on the training of the roughly 120 test administrators who are thanked in the manual.
For every user switching from the older WAIS-III to the WAIS-IV, the question arises of the comparability of the subtest scaled scores, index scores, and IQs obtained with the two instruments. For the two previous version changes (from the WAIS to the WAIS-R in 1991 and from the WAIS-R to the WAIS-III in 2006), staff of the Psychological section of the Department of Psychiatry and Psychotherapy, University Hospital, LMU Munich presented results of equivalence studies a few years after the new versions appeared12) 13). According to these, there were large differences between the WAIS and the WAIS-R, and noticeable differences between the WAIS-R and the WAIS-III, in IQs and subtest results. The newer instrument in each case led to lower values, which was to be expected because of the Flynn effect14). Rather unexpectedly, strong qualitative differences were also found between the distribution shapes of the respective standardization samples, which would actually have made it necessary to provide an individual conversion for every raw score of every age group.
For the American original version of the WAIS-IV, the results of an equivalence study with the WAIS-III are available, published in the technical manual of the WAIS-IV15). For individual subtests, differences in the range of 1/3 standard deviation or 5 IQ points were found, generally higher in the verbal portion than in the performance portion.
For the German version of the WAIS-IV, to our knowledge there is currently no equivalence study. It would have made the most sense to carry it out together with the norming study. This was apparently neglected. Even if, according to some reports, the Flynn effect has meanwhile weakened16), certain differences are still likely to be expected. The user is left alone with this uncertainty, especially when older WAIS-III reports are to be compared with newer WAIS-IV reports.
Overview of the properties and TDB2 normative data of the subtests
In this section, organized by domain, the back-calculated raw score distributions for selected age groups are shown and discussed for each WAIS-IV subtest. Raw score distributions provide basic information on the discriminative power of a test. A second diagram shows the mapping of the raw scores onto the performance scores. In both figures, the discriminative power can be readily assessed at different points of the distribution. In the figure mapping raw scores onto performance scores, distribution parameters of all age groups (means plus/minus one and two sigma) are additionally plotted. In this diagram, the influence of age on subtest performance can be clearly seen. Peculiarities are commented on in each case.
Verbal Comprehension
The verbal abilities domain comprises the core subtests Similarities, Vocabulary, and Information. The composition of this domain is thus the same as in the WAIS-III. The subtest Comprehension is added as an optional subtest. All four subtests are old acquaintances and were included in all earlier German-language versions of the Wechsler tests since the WAIS.
Similarities
The raw score distribution of the subtest Similarities (SI, Figure 1) shows a certain ceiling effect (especially in young adults), which is due to the fact that the subtest contains relatively many easy items but only a few difficult ones. Such a raw score distribution means that precision is lower in the upper ability range than in the lower. We will see similar distributions in the three other subtests of the verbal abilities domain.
Figure 1: Reconstructed raw score distribution of the subtest Similarities for selected age groups
Figure 2: Performance scores in the subtest Similarities with raw scores and age norm limits (see text)
In an overview graphic, Figure 2 shows the relationship between raw scores and performance scores as well as the influence of age on the (average) performance for the subtest Similarities. In this figure (and all similar ones that follow), the vertical lines mark the performance scores from 40 to 145. On the black horizontal line, selected raw scores are plotted as diamonds in their correct positions. The alternately red and blue lines contain the norm limits for all age groups at a glance. The five markers on each line stand for the percentile ranks 2.5, 16, 50, 84, and 97.5. One such line, namely the one that corresponds to the examinee's age group, is drawn (in a different form) in the TDB2Online profile sheet to help with the individual interpretation of the test results. Because the performance scores are calculated from the results of the 20-29-year-old young adults, the 5 markers of the age groups 20-24 and 25-29 always enclose the performance scores 70, 85, 100, 115, and 130.
The position of the raw scores in the performance score grid of Figure 2 clearly shows the lower differentiation capacity in the upper ability range. Between the performance scores 85 and 100 there are about 8 raw score points (about 17 to 25), between 115 and 130 only 4 (31 to 35). One additional correct answer (scored with two points) leads to a performance that is 7 to 8 IQ points better in the upper ability range, whereas in the lower range it raises the subtest performance by only 4 points. The maximum possible raw score of 36 corresponds to a performance score of 137; better performance cannot be measured with the subtest. At the lower end, by contrast, the graduation of item difficulty is good. For a clinical test that is meant to detect performance deficits rather than performance peaks, such a characteristic is entirely acceptable.
The age-associated decline in performance is not strong in this subtest. Between the mean of the best age group (30-34) and that of the worst (85-89) there are only 10 IQ points (corresponds to 2/3 standard deviation).
Vocabulary
The raw score distribution of the subtest Vocabulary (VC, Figure 3) shows a ceiling effect from young adulthood up to the beginning of retirement age. In the upper range, the distribution curve for middle age cohorts is very steep; small raw score differences cause large performance score differences.
Figure 3: Reconstructed raw score distribution of the subtest Vocabulary for selected age groups
Figure 4: Performance scores in the subtest Vocabulary with raw scores and age norm limits (explanation of the figure in the text for Figure 2)
Accordingly, in Vocabulary the differentiation capacity in the upper ability range is clearly poorer than in the middle or lower range. The age-associated decline in performance is moderate. Between the mean of the best age group (35-44) and that of the worst (85-89) there is about one standard deviation (15 IQ points).
Information
Even the distribution curve of the raw scores of the subtest Information (IN, Figure 5) shows that the 65-69-year-olds know more than the young adults.
Figure 5: Reconstructed raw score distribution of the subtest Information for selected age groups
Figure 6: Performance scores in the subtest Information with raw scores and age norm limits (explanation of the figure in the text for Figure 2)
The more detailed age distribution in Figure 6 then shows that the 45-54-year-olds achieve the highest mean performance. The decline in performance from this group to the 85-89-year-olds amounts to about 10 IQ points on average.
Comprehension
Of the four subtests in the verbal abilities domain, the subtest Comprehension (CO, Figure 7) has the most skewed distribution. In the lower performance range, even large raw score differences have only a small influence, whereas in the upper ability range small changes have a large influence. The test ceiling is low.
Figure 7: Reconstructed raw score distribution of the subtest Comprehension for selected age groups
Figure 8: Performance scores in the subtest Comprehension with raw scores and age norm limits (explanation of the figure in the text for Figure 2)
The acquisition of the verbal ability measured by this subtest continues into middle adulthood; the maximum lies between 35 and 54 (Figure 8). Even more than in SI, CO shows a reduction in variance among the older persons in the sample. Actually, the percentile-to-normal transformation should stabilize the variance even across groups differing in performance. Possibly, other cognitive abilities enter the measurement in SI and CO at advanced age.
Perceptual Reasoning
The cognitive domain Visual problem solving comprises a total of five subtests: the three core subtests Block Design, Matrix Reasoning, and Visual Puzzles as well as the two optional subtests Figure Weights and Picture Completion. In the WAIS-III, this domain included the subtests Block Design, Matrix Reasoning, and Picture Completion. The new subtest Visual Puzzles is a very good addition to this domain, to which Picture Completion never fitted anyway. The subtest Figure Weights is also new. Block Design and Picture Completion have been old acquaintances since the WAIS, Matrix Reasoning since the WAIS-III. For Figure Weights, age norms are available only up to the group of 65-69-year-olds.
Block Design
Block Design (BD) contains many easy tasks. In the higher-performing age groups there is a negatively skewed distribution with poor differentiation in the upper ability range.
Figure 9: Reconstructed raw score distribution of the subtest Block Design for selected age groups
Figure 10: Performance scores in the subtest Block Design with raw scores and age norm limits (explanation of the figure in the text for Figure 2)
The performance optimum lies in the 18–29-year-olds. It is hard to believe that the decline in performance from 30 to about 60 should be steeper than from 60 to 90, which, however, is the case according to the data. The same effect was already seen in the WAIS-III. Since in the WAIS-IV, after all, 4 of 14 tasks were changed compared with the WAIS-III, a scaling artifact or a floor effect (which we considered likely for the WAIS-III at that time) is now less probable. Perhaps Block Design actually requires abilities that do not decline as strongly in old age as in middle age. Overall, the age difference is considerable: the mean performance score drops from about 100 in young adults to 76 in the 85–89-year-olds.
Matrix Reasoning
Matrix Reasoning (MR) is similar to Block Design in many properties. Here, too, we find poorer discriminative power in the upper ability range and a rapid age-related decline.
Figure 11: Reconstructed raw score distribution of the subtest Matrix Reasoning for selected age groups
Figure 12: Performance scores in the subtest Matrix Reasoning with raw scores and age norm limits (explanation of the figure in the text for Figure 2)
The 85–89-year-olds achieve on average a performance score of only 68, which means a loss of more than two standard deviations compared with young adults (PS 100).
Visual Puzzles
The raw score distributions of the subtest Visual Puzzles (VP, Figure 13) show large mean differences between young and old, but also large differences in spread, with a narrow distribution in the oldest and broader ones in younger age groups. Both can be seen even better in the diagram with the age norms (Figure 14).
Figure 13: Reconstructed raw score distribution of the subtest Visual Puzzles for selected age groups
Figure 14: Performance scores in the subtest Visual Puzzles with raw scores and age norm limits (explanation of the figure in the text for Figure 2)
At the level of performance scores, the standard deviation of the 85–89-year-olds is only about half as large as that of the 20–24-year-olds. Actually, the percentile-to-normal transformation should ensure that there is no direct dependence between performance level and variance. It is peculiar that this occurs here to such an extent. There is much to suggest that the old perceive or approach the tasks differently from the young. Possibly the time limit of 20 or 30 seconds per task plays a role here. This limit (if the processing time is longer, the task counts as not solved) is far less often relevant for the young than for the old. It could cause all old persons to achieve fewer points, even those who would be able to solve the task with more time. Unfortunately, the Wechsler manuals contain no item statistics with which one could trace this.
The average performance of the 85–89-year-olds, with a performance score of 72, is just under two standard deviations below the performance of young adults.
Figure Weights
For Figure Weights (FW), standardization data are available only up to the age group 65-69. The distributions (see Figure 15) are relatively symmetric; deviations from linearity occur only in the extremely high ability range.
Figure 15: Reconstructed raw score distribution of the subtest Figure Weights for selected age groups
Figure 16: Performance scores in the subtest Figure Weights with raw scores and age norm limits (explanation of the figure in the text for Figure 2)
The average performance of the 65–69-year-olds, with a performance score of 82, is a good standard deviation lower than that of the young. This corresponds roughly to the decline in performance in VP or MR.
Picture Completion
The distributions of the raw scores in the subtest Picture Completion (PC, Figure 17) are, apart from the floor effect in the very old, actually quite exemplary and indicate good differentiation capacity across all performance groups.
Figure 17: Reconstructed raw score distribution of the subtest Picture Completion for selected age groups
Figure 18: Performance scores in the subtest Picture Completion with raw scores and age norm limits (explanation of the figure in the text for Figure 2)
This is also confirmed in the age norms (Figure 18). The average decline in performance of the 85–89-year-olds amounts to just under two standard deviations.
Working memory
The cognitive domain Working memory comprises three subtests: the two core subtests Digit Span (DS) and Arithmetic (AR) as well as the optional subtest Letter-Number Sequencing (LN). All three were already included in the WAIS-III, and in the WAIS-III all three subtests also formed the index score for the domain. In the WAIS-IV, however, age norms for LN are available only up to the group of 65-69-year-olds. There is one more difference: in the WAIS-III, DS consisted only of the subtasks “forward” and “backward”. In the WAIS-IV, the condition “sequencing” is added. This increases the working memory load of this subtest.
Digit Span
The WAIS-IV subtest Digit Span (DS) shows a fairly ideal raw score distribution (Figure 19): there is enough room on both sides of the distribution; both the raw scores 0 to 7 and those from 45 to 48 are reached only very rarely. There are thus neither floor nor ceiling effects. In addition, the distributions are fairly symmetric and have a relatively uniform variance across the entire age range.
Figure 19: Reconstructed raw score distribution of the subtest Digit Span for selected age groups
Figure 20: Performance scores in the subtest Digit Span with raw scores and age norm limits (explanation of the figure in the text for Figure 2)
The age-associated decline in performance (Figure 20) is moderate in this test: the performance of the 85–89-year-olds drops to a performance score of 81, that is, somewhat more than one standard deviation compared with the performance of young adults (PS 100).
Arithmetic
The WAIS-IV subtest Arithmetic (AR) is too easy. As can be seen from Figure 22, 95 percent of all persons in the normative sample (across all age groups) achieve scores between 5 and 22, where 22 is the maximum raw score. It corresponds to a performance score of only just under 130. In the lower part of the distribution (Figure 21), by contrast, the differentiation capacity is good.
Figure 21: Reconstructed raw score distribution of the subtest Arithmetic for selected age groups
Figure 22: Performance scores in the subtest Arithmetic with raw scores and age norm limits (explanation of the figure in the text for Figure 2)
The age-associated decline in performance is not very large: from young adulthood (PS 100) the performance scores drop to 88 in the 85–89-year-olds, which is somewhat less than one standard deviation.
Letter-Number Sequencing
The subtest Letter-Number Sequencing (LN) has a relatively symmetric raw score distribution (Figure 23). In the lower range there are many easy items that are answered correctly by the vast majority of persons. Hardly anyone achieves fewer than 10 raw score points.
Figure 23: Reconstructed raw score distribution of the subtest Letter-Number Sequencing for selected age groups
Figure 24: Performance scores in the subtest Letter-Number Sequencing with raw scores and age norm limits (explanation of the figure in the text for Figure 2)
The age-associated decline in performance (Figure 24) can be captured only partially in the subtest Letter-Number Sequencing, because it is normed only up to the age group of 65-69 years. Up to this age the decline is only 8 IQ points.
Processing speed
The cognitive domain processing speed comprises three subtests: the two core subtests Symbol Search (SS) and Digit Symbol Coding (DSC) as well as the optional subtest Cancellation (CA). In the WAIS-III, the domain was likewise measured by SS and DSC. For Cancellation, age norms are available only up to the group of 65-69-year-olds.
Symbol Search
All subtests that measure processing speed have good distributional properties. This also applies to the subtest Symbol Search (SS, Figure 25).
Figure 25: Reconstructed raw score distribution of the subtest Symbol Search for selected age groups
Figure 26: Performance scores in the subtest Symbol Search with raw scores and age norm limits (explanation of the figure in the text for Figure 2)
In all subtests that measure processing speed, an early-onset and relatively linear age-associated decline in performance can be observed. As Figure 26 shows, in Symbol Search it amounts to about two standard deviations or 30 IQ points from the 20s to the age of 85-89 years.
Digit Symbol Coding
What was said about the subtest Symbol Search also applies largely to Digit Symbol Coding (DSC, Figure 27). In addition, here the raw score range from 0 to 135 is practically never fully used in the upper ability range at first testings. The performance score of 145, the maximum performance that can be displayed graphically in TDB2Online, is already reached with a raw score of 112 (see Figure 28). The wide measurement range nevertheless makes sense, because the subtest is occasionally also used in repeated testing.
Figure 27: Reconstructed raw score distribution of the subtest Digit Symbol Coding for selected age groups
Figure 28: Performance scores in the subtest Digit Symbol Coding with raw scores and age norm limits (explanation of the figure in the text for Figure 2)
In this subtest, it is actually the 16–17-year-olds who perform best (mean performance score 101.2), although the difference from the young adults on whom the performance score is normed (20–29-year-olds, PS 100) is only minimal. The age-associated decline in performance between young adulthood and the group of 85–89-year-olds amounts to two standard deviations.
Cancellation
Cancellation (CA) has a good raw score distribution and a reasonable measurement range, as is evident from Figures 29 and 30. The only shortcoming is that it is normed only up to the age group 65-69.
Figure 29: Reconstructed raw score distribution of the subtest Cancellation for selected age groups
Figure 30: Performance scores in the subtest Cancellation with raw scores and age norm limits (explanation of the figure in the text for Figure 2)
As far as the age-associated decline in performance can be judged within the restricted age range, it lies in the range of that of the two other subtests of this domain.
Treatment of the global scores (total and domain scores)
When several subtests are added or averaged and the result of this operation is then newly normed as a measured value, a new metric arises that is no longer compatible with that of the subtests. This matter is explained in the general documentation in the section Multiple standardization at different levels and was made transparent using the example of the intelligence quotients of the WAIS-R. We saw there that an examinee who achieves four scaled scores in every subtest does not receive an IQ of 70 in the intelligence quotients, which one would expect with a comparable metric (two standard deviations below the mean of his or her age group), but a markedly lower one, whose exact value depends on the correlation between the subtests and on the number of aggregated subtests.
In the WAIS-IV, this concerns the four index scores and the Full Scale IQ. These five measures were normed separately once more and therefore build up a metric of their own. We show the effect in Figure 31 using the example of a 25-year-old examinee who achieved exactly 4 scaled scores in each core subtest.
Figure 31: WAIS-IV test profile of an examinee with exactly 4 scaled scores in each of the 10 core subtests (see text)
In the TDB2Online profile, this examinee achieves a performance score of 70 or very close to 70 in every subtest; the small deviations from the 70 line are due to the whole-number nature of the raw scores. Since this is a 25-year-old examinee, the performance scores that can be read off in the profile are practically identical to the age-normed scores (ANS) printed as numbers at the right margin. The mean performance scores in the four domains and in the overall score are also all at or very close to 70, which corresponds to the desire for a uniform metric in TDB2Online.
The red markings drawn into the profile by hand afterwards show the position of the IQ and index scores as they result from a manual scoring using the tables of the WAIS-IV manual. Without exception, they lie below the subtest values, with the correlation between the aggregated subtests and the number of aggregated subtests determining the extent of the spread of the metric: the lower the intercorrelations of the subtests and the more subtests enter the global score, the greater the spread of the metric. For the domain scores, where only two or three relatively highly intercorrelated subtests are combined in each case, the resulting indices are 67 (Verbal abilities), 65 (Visual problem solving), 66 (Working memory), and 68 (Processing speed), thus only slightly lower than the respective subtest values. For the Full Scale IQ, by contrast, which comprises 10 subtests that measure different things and correlate less strongly with one another, the spread of the metric is much greater. The manually scored Full Scale IQ is only 60.
In TDB2Online, no separately standardized global measures are taken into account. Instead, we simply report the mean subtest performance scores or their age-normed scores.
Age norms of the global measures
The figures in this section show the age norms of the four domains of the WAIS-IV as well as of the overall performance in all 10 core subtests. In TDB2Online, these mean performance scores are produced in place of the index scores and the Full Scale IQ. For each cognitive domain and for the total test, the mean performance score is calculated and displayed graphically in TDB2Online below the subtests. Averaging is done only over the core subtests belonging to a domain. If raw scores were also entered for optional subtests, these are shown in the subtest results but not taken into account in the global measures. The corresponding age norm limits were likewise averaged only over the core subtests, because for three of the five optional subtests no norms are available for persons over 69.
Verbal Comprehension
Figure 32: Age norm limits of the verbal abilities domain
In verbal abilities, the highest performance in the WAIS-IV is achieved by the age groups between about 30 and 60. The two age groups that define the performance scores (20-24 and 25-29) lie two or three IQ points below. Overall, the age-associated fluctuations of the mean performance scores in this domain lie below one standard deviation.
Perceptual Reasoning
Figure 33: Age norm limits of the visual problem solving domain
The measurement ranges of the three subtests belonging to visual problem solving are all somewhat limited at the top, which means that mean performance scores above 130 hardly occur in this domain. The age-associated decline in performance begins at about 30 years and is steeper up to the age of about 70 than afterwards. An explanation for this is not easy to find, but the same could be observed in the WAIS-III with an entirely different normative sample.
Working memory
Figure 34: Age norm limits of the working memory domain
Working memory (Figure 34) and processing speed (Figure 35) differ considerably in their age trajectories. In working memory, maximum performance is reached only at an age of about 27 years. Until the end of the 80s, performance then declines by about one standard deviation. Working speed, by contrast, is already highest in adolescence (16-17) and declines fairly linearly from the mid-20s by about two standard deviations. This different age trajectory was already found in the WAIS-III, although there (in all domains) the age-associated decline was even larger.
Processing speed
Figure 35: Age norm limits of the processing speed domain
Total test
Figure 32: Age norm limits of the overall performance in the core subtests
A global score over 10 subtests can actually be meaningfully interpreted only if the subtest results hardly differ. Then it simply measures the central tendency of the cognitive abilities measured by the test. In clinical questions, deficits in individual cognitive domains are usually the focus.
Discussion
Despite all criticism in detail, it must be stated that the German version of the WAIS-IV is a successful adaptation of an indispensable diagnostic instrument. The subtests measure ability areas that are important for the clinical description of cognitive disorders, and the domains are, on the whole, sensibly defined. In the tension between differential-psychological and neuropsychological models for representing cognitive abilities, the Wechsler test batteries have since the WAIS-III taken a middle course that, in terms of vocabulary and subtest composition, is well suited to communicating clinical results for patients with mental disorders. Compared with many other test procedures, the WAIS-IV scores with good test material, good norming, and a wide age range. There is an extensive scientific and application-specific literature on the WAIS-IV, but almost exclusively in English.