HAWIE-R with WIE norms

German version of the WAIS-R in the norms of the German WAIS-III

When this wiki page was added to TDB2Online, the German WAIS-IV did not yet exist and the WIE (the German version of the WAIS-III) was the current version of the Wechsler tests for adults. There is now no longer any reason to use the WIE instead of the WAIS-IV, and the WIE test material is also out of print. The wiki entry remains here only for historical reasons and in unchanged form. To our knowledge, no equivalence data between the WIE and the WAIS-IV are available.

The HAWIE-R is outdated and out of print. For initial examinations this is not a problem: the WIE is a more up-to-date and better replacement. In retesting, however, the problem can arise that newly obtained WIE results must be compared with older HAWIE-R results. This is not possible directly, either at the IQ level or at the subtest level. The two tests were normed 15 years apart and yield different results, which also differ from subtest to subtest.

Erzberger & Engel¹⁾ determined equivalent IQs and subtest raw scores for the two tests in 2010 by means of equipercentile linking²⁾ and published them in tabular form. The same data basis was used in TDB2Online to display HAWIE-R results as a profile in the norm grid of the WIE, thus making them comparable with new WIE results.

In the TDB2OnlineApp there is a “test” named “HAWIE-R with WIE norms”. If the raw scores of HAWIE-R subtests are entered there, the evaluation yields a test profile showing how the examinee would presumably have performed had the corresponding WIE subtests been administered at the initial examination instead of the HAWIE-R.

The conversion is based on the equipercentile equating of the raw scores. In the program this can be done somewhat more finely than the integer tables of the publication allow. The calculation of the performance scores and the age-dependent standard scores relies on the corresponding calculations for the WIE. In the profiles, only the equivalent raw scores of the HAWIE-R are underlaid beneath these. Figure 1 shows this by way of example for the Vocabulary subtest.

Figure 1: Performance scores in the HAWIE-R Vocabulary subtest with raw scores and age norm limits, displayed in the norm grid of the WIE

Equivalence calculations of this kind depend on how well the two subtests (in the example, the Vocabulary subtest in the HAWIE-R and the Vocabulary subtest in the WIE) actually correspond and to what extent they measure the same thing. In the sample of Erzberger & Engel (2010), these two subtests correlated at r=.91; in a similar study with the two American original tests (WAIS-R and WAIS-III), reported in the “Technical Manual” of the WAIS-III³⁾, the correlation between the Vocabulary tests of the two tests was r=.90. With such a high correlation, equivalence calculations work quite well, because there can hardly be any variance in these data that could systematically bias the equivalence calculations in one direction or the other. Pure error variance hardly interferes. However, the association is not high for all subtests. In Erzberger & Engel (2010), the correlations for the subtests of the Verbal part ranged from r=.73 (Digit Span) to r=.91 (Vocabulary), and for the subtests of the Performance part (which were changed more strongly in the WIE) from r=.56 (Picture Completion) to r=.87 (Block Design). For some of the only moderately high correlations of the Performance subtests, the equivalence values must therefore be accompanied by relatively wide confidence intervals (which cannot be calculated more precisely with the chosen method).

see the detailed test documentation