The incomplete equivalence of the paper-and-pencil and computerized versions of the General Aptitude Test Battery.

Abstract

The performance of 163 applicants for the Dutch Royal Military Academy on the computerized version of the General Aptitude Test Battery (GATB) was compared with the performance of 163 matched applicants on the paper-and-pencil version. There was a modest but clearly discernible influence of computerization. A LISREL analysis showed a reasonable fit for a model postulating two factors that were equally patterned for both test versions. A model postulating equal factor loadings had to be rejected. Individual differences in both the computerized and conventional GATB were strongly related to intelligence. The computerized subtests produced faster and more inaccurate responses than the conventional subtests. Both in terms of number of solved items and correlations with other cognitive measures, the cognitively simple, clerical tests were more affected by computerization than the more complex tasks.