Wechsler Adult Intelligence Scale - Third Edition

Just as the WIE replaced the HAWIE-R in 2006 in the German-speaking countries, the WAIS-IV has since 2012 replaced the WIE in turn. The “Wechsler Intelligenztest für Erwachsene (WIE)” described here is therefore now outdated as well. The documentation text remains only for historical reasons and for purposes of scientific comparison. We do not know whether the norms of the WIE and the German WAIS-IV differ (that is, whether examinees systematically obtain different subtest scaled scores and IQs on one procedure than on the other). There are no equivalence tables such as those available for the change from the HAWIE-R to the WIE.

The Wechsler-Intelligenztest für Erwachsene (WIE), is the German adaptation of the Wechsler Adult Intelligence Scale III (WAIS-III). The American WAIS-III was published in the USA in 1997,¹⁾ the WIE in 2006.²⁾ David Wechsler, an American psychologist, constructed the first intelligence battery of this series in the late 1930s with the Wechsler-Bellevue Scale,³⁾ which was intended mainly for clinical purposes. The WAIS-III and its German adaptation WIE belong to the fourth generation of the Wechsler intelligence test series.

The current test manual for the German adaptation contains only the material necessary for administration and scoring. Historical background information on the Wechsler tests can be found in the classic manual by Matarazzo⁴⁾. Recent application-oriented literature on the clinical interpretation of the WAIS-III (and thus of its German counterpart) is currently available only in English.^{5) 6)}

With their collection of 11 subtests, the predecessors HAWIE and HAWIE-R had already offered a good overview of the structure of cognitive performance, especially in cognitively impaired patients. With the WIE (WAIS-III), the composition of the subtests was changed more substantially for the first time in order to better meet the needs of neuropsychological examinations. In addition to the usual division of test performance into a verbal and a performance part, the 14 subtests of the WIE allow a division according to neuropsychological considerations, which connects the test to more modern, neuropsychologically oriented research on the structure of cognitive performance. Table 1 gives an overview of the subtests with their English names, their assignment to cognitive domains, the abbreviation customary in the English manual, the rank order of administration and the possible raw score range.

Table 1: Subtests of the WIE

Only the subtests LN, MR and SS are new, but all the others have also been changed compared with the HAWIE-R, in some cases considerably.

Two of the domain names are problematic. “Verbal comprehension” is too narrow as the domain also includes the Similarities subtest, which requires the active production of a superordinate concept, something that “comprehension” covers inadequately. In TDB2Online and in the clinical report of the TDB2Online app, the term “Verbal abilities” is used, which is on the one hand more neutral and on the other more comprehensive. The second unsuitable name is “Perceptual organization” for a cognitive domain that has mainly to do with active, hands-on problem solving. In TDB2Online we make this more concrete with the term “Visual problem solving”. For the sake of completeness, it should be said that the choice of terms does not go back to the editors of the German version, but to the editors of the original English version.

For the administration of the WIE one needs the complete test kit, which in addition to the manual and an initial supply of record forms and task booklets also contains the test material for the practical subtests. These include, for example, the templates for the new Matrix Reasoning subtest or the blocks for Block Design. This test material lasts for many testing sessions, but not forever. Templates that are handled frequently have to be replaced from time to time. Only the record forms and the special task booklets for certain subtests are continuously needed as consumables.

The test material for the Wechsler tests is usually available only until a new generation of the tests appears. This is now the case. Since 2012 the successor of the WIE, the Wechsler Adult Intelligence Scale - Fourth Edition (WAIS-IV)⁷⁾, has been available in German. The test material for the WIE is out of print.

Unfortunately, the manual contains many printing errors, in part also slips in the tables or numbers. To our knowledge they have never been corrected by the authors or the publisher.

The WIE is protected by copyright; the rights holder is Pearson Assessment, represented in Germany by Pearson Assessment & Information GmbH in Frankfurt/Main. Use is additionally controlled, and the material is supplied only to certain professional groups.

The copyright history of the Wechsler tests is not uninteresting. Wechsler's tests were long published by the Psychological Corporation, an organization founded in 1921 by James McKeen Cattell as a publication platform for applied psychological projects. In 1942 McKeen Cattell bequeathed most of his personal shares in the Psychological Corporation to a foundation that to this day supports applied psychological research.

The Psychological Corporation was bought by Harcourt Brace Jovanovich in 1970 and later merged with Harcourt's own test publisher. After several internal name changes, Harcourt Assessment was taken over by Reed Elsevier in 2001 and by Pearson Education in 2007, where today it belongs to Pearson's Assessment & Information Group.⁸⁾

The rights for the German versions of the Wechsler tests were held before 2011 by the publisher Huber in Bern.

All Wechsler intelligence test batteries are individually administered tests. The test material is varied and, through its diversity, helps to sustain the usually rather moderate motivation of examinees in clinical examination settings. Individual administration with only one patient also contributes to this.

The administration is described well and in detail in the test manual. With the WIE it is more difficult than with many other ability tests and has to be learned and practiced. The verbal tests, for example, are answered freely. This is good for the examinee, but it means for the examiners that the answer has to be compared with example answers in the manual and scored. One also has to know in which cases a query is required. It is not sufficient to read later that one should have queried a particular answer in order to arrive at a clear scoring. Many subtests have discontinue rules with which one must be familiar.

The administration of the complete WIE, that is, of all 13 subtests required for determining the partial IQs and the index scores (without Object Assembly), usually takes about 2 hours in psychiatric patients. To this must be added the scoring time, which depends heavily on the examiner's experience.

The 14 subtests of the WIE yield raw scores of very different origin:

• Number of correct solutions: MR, IN, LN,
• Number of correct solutions within a certain time: PC, DSC
• Sum of points for the quality of a solution (0, 1 or 2 points per item): VC, SI, DS, CO
• Sum of points for the quality of a solution (0, 1 or 2 points per item) within a certain time per item: PA
• Complex scoring according to correctness and speed of the solution: BD, AR, OA
• Number of correct minus number of incorrect solutions within a certain time: SS

The complex scorings according to quality and time in particular are hardly found in other tests. In theory they have the disadvantage that the raw score distributions of the subtests scored in this way are not occupied at all at some points – especially in the lower range – and jump sharply at others. In the construction of the procedures an attempt was made to compensate for this fundamental disadvantage by choosing relatively high point totals. In addition, we will see that the calculation procedures for the raw scores were chosen wisely with regard to the shape of their distribution (see below).

The raw score range varies quite strongly from subtest to subtest. In the Letter-Number Sequencing subtest it can lie between 0 and 21, in Digit Symbol Coding between 0 and 133.

Because of the different raw score ranges and because of the deviations of the raw score distributions from a normal distribution, the raw scores have been converted into so-called scaled scores by means of tables since the first generation of the Wechsler intelligence tests. In the Wechsler tests, scaled scores are standard scores with a mean of 10 and a standard deviation of 3. Nevertheless, the exact definition of the scaled scores differs between the various versions of the Wechsler tests:

• In the Wechsler Bellevue Scale⁹⁾ the scaled scores were linearly scaled as standard scores so that the overall distribution of the standardization sample had a mean of 10 and a standard deviation of 3. This definition is problematic in terms of content because the subtests exhibit very different age-related trends. A mean score of 10 defined in this way characterizes a broad age range for some subtests, particularly the verbal ones. For subtests with a sharp decline in performance with age, however, it can only truly characterize a fairly narrow age range. Another drawback is that a change in the age range of the standardization sample leads to a change in the definition of the score points. This, too, has a particularly significant effect on subtests that show a sharp decline in performance with age.
• In the German HAWIE¹⁰⁾, which is still based on the Wechsler-Bellevue Scale in terms of content, the score points were also scaled linearly, but in such a way that a mean of 10 and a standard deviation of 3 were achieved in the age group of young adults between 20 and 34 years of age. In doing so, the German HAWIE anticipated a methodological change in the American WAIS¹¹⁾, which was published almost simultaneously. The score points could now be more precisely characterized in terms of content as a performance typical of young, healthy subjects at the peak of their cognitive abilities over the course of their lives. The age range of the standardization sample no longer has any influence on this. Age-related deficits are evident on both ends of the spectrum. For example, adolescents around 16 years of age score an average of only 8 points on the vocabulary test. In contrast, 70-year-old adults score only four to six points on some subtests of the performance section. Incidentally, the limited range of scores here limits the ability to scale poor performance among older individuals.
• Starting with the WAIS¹²⁾ and continuing with the WAIS-R¹³⁾ and its German version, the HAWIE-R¹⁴⁾, raw scores were no longer scaled linearly but rather through area transformations, which ensured that the raw score scales were linearized. The age group of young adults between 20 and 34 years of age was retained as the basis.
• In the WAIS-III¹⁵⁾, the WIE¹⁶⁾, the WAIS-IV¹⁷⁾ and its German version¹⁸⁾, this definition was unfortunately not continued. Here the scaled scores were instead defined by age group, so that a mean of 10 and a standard deviation of 3 results separately for each age group. (This variant was also already available for the WAIS-R as an alternative scoring.) In the WIE, therefore, each age group is transformed differently, and the scaled scores lose their ability to measure actual performance, including that of persons of different ages.

In TDB2Online, test results are presented in the form of age-independent performance scores. For the calculation of the performance scores for the subtests of the WAIS-III, the scaled score transformations of the age groups 20-24 and 25-29 were used. The details are explained in the next section.

For the WAIS-III there are to this day no practically usable normative data apart from those reported in the manual.

For the subtests of the WAIS-III, the conversion of raw scores into scaled scores involves a linearization by percentile-to-normal transformation and a transformation to the scaled score scale (10;3), separately for each age group. The tables in the manual give these conversion data only as whole numbers; many raw scores occupy the same scaled score. The hard steps of this conversion were first smoothed 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 indeed already been done when the norms were created. In the data, however, discontinuities in the age trajectory were occasionally observed that certainly do not reflect empirical reality but are the result of the limited sample size. They were eliminated by smoothing across the age groups. Third-order smoothing was always sufficient here. Subsequently the data for the two age groups 20-24 and 25-29 were averaged. The mean was the basis for the calculation of the performance scores.

In the Comprehension subtest, the conversion tables from raw scores to scaled scores contain a maximum of 33 raw points. In fact, however, 35 raw points can be achieved in this subtest, not only 33. An error in the design of the table is probably responsible here. Since a forty-year-old with 33 raw points reaches only 15 scaled scores, one must assume that persons with higher performance also appeared in the standardization sample. In any case there is still plenty of room at the top of the scaled score scale.

For TDB2Online the table was extended upward to 35 raw points by extrapolation. This extrapolation may not correspond exactly to the empirical data, but it is certainly more correct than the situation in the manual, where performance above 33 raw points is simply ignored. In TDB2Online, therefore, all raw scores that occur can also be entered for the Comprehension subtest.

A linearization of the raw scores is carried out implicitly via the scaled score transformation, namely via those for the age groups 20-24 and 25-29, which are used in combination for the calculation of the performance scores.

If several subtests are summed or averaged and the result is then newly normed as a new measured value, this gives rise to a new metric 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 intelligence quotients of the HAWIE-R as an example. We saw there that an examinee who reaches four scaled scores in every subtest does not obtain in the intelligence quotients the IQ of 70 that one would expect with a comparable metric (two standard deviations below the mean of his or her age group), but clearly lower values that depend on the correlation between the subtests and on the number of aggregated subtests.

In the WIE this concerns not only the three intelligence quotients but also the four indices. All seven measures were normed separately once more and therefore build up a metric of their own. We show the effect in Figure 6 using the example of a 29-year-old examinee who achieved 4 scaled scores in every subtest of the WIE.

Figure 6: WIE test profile of an examinee with 4 scaled scores in each subtest (see text)

In the TDB2Online profile, this examinee reaches a performance score (“Leistungswert”) of 70, or very close to 70, in every subtest; the small deviations from the line at 70 are due to the whole-number nature of the raw scores. Since the examinee is 29 years old, the performance scores are identical to the age-normed scores (“ASW” in the right number column). The mean performance scores in the seven domains (IQs and indices) are also all at or very close to 70, which corresponds to the aim of a uniform metric in TDB2Online.

The red circles drawn in by hand show the position of the IQ and index scores as calculated from the tables of the WIE manual. Without exception they lie below the subtest scores, with the correlation between the aggregated subtests and the number of aggregated subtests determining the extent of the metric spread. For the Processing speed index, in which only two subtests (DSC and SS) are combined, which in addition correlate only .68 (age group 24-29), the result is an index of 68, hardly lower than the two subtest scores of 70. For the Full Scale IQ, on the other hand, into which 11 subtests enter that correlate with one another on average at .54, the spread is much stronger and, with mean subtest scores of 70, yields a Full Scale IQ of only 60.

The change in the metric is not as strong in the WIE as in the HAWIE-R, because the mean intercorrelation of the subtests in the standardization sample is higher in the WIE than in the HAWIE-R.

In TDB2Online the separately standardized global measures are not taken into account any further. Instead we simply report the mean subtest performance scores or their age-normed scores, with exactly the same metric as for the subtests.

The sample for the norming of the WIE has a considerable size and a wide age range. For every age group separately there is a breakdown by educational level. These are, however, not easy to evaluate because of the different educational systems included in the norming (Switzerland, Austria, Germany, the latter with different systems in the 16 federal states), above all also when one considers the enormous span of time over which the persons in the normative sample attended school.

With regard to other details of the collection of the normative data, the manual is reticent. One learns nothing about the way the sample was drawn, nothing about the locations and nothing about agreed or fulfilled quotas (if there were any).

Adaptation and norming of the procedure were initially begun in Hannover under the responsibility of Uwe Tewes, the German editor of the predecessor version HAWIE-R¹⁹⁾. Numerous dissertations at the Medizinische Hochschule Hannover, whose authors are mentioned in the WIE manual as collaborators in data collection²⁰⁾, bear witness to this. Otherwise the WIE manual contains no indication at all of Tewes's role as editor. One of these dissertations ²¹⁾ shows that 99 of the 1,820 examinees tested in Germany under Tewes's responsibility came from the East German states, without any claim to representativeness. Another dissertation ²²⁾ notes that for the final publication of the WIE there were follow-up collections of individual examinees for the normative sample, by then under the responsibility of the new editors of the procedure. Presumably publishing contracts and/or agreements between the editors commissioned originally by Huber and later by Harcourt are responsible for the fact that the information on this in the manual is terse. On the other hand, it is unfortunately also a bad tradition in German test procedures that no great words are spent on sample selection.