Ability Measure Equivalence of Computer Adaptive and Pencil and Paper Tests: A Research Synthesis.

This paper reports on existing studies and uses meta analysis to compare and synthesize the results of 20 studies from 8 research reports comparing the ability measure equivalence of computer adaptive tests (CAT) and conventional paper and pencil tests. Using the research synthesis techniques developed by Hedges and Olkin (1985), it is possible to estimate a scale-free index of effect magnitude for each study and then compare results on a common scale. Most studies, despite differences in test content, age of examinees, item response theory model used, and study design, show comparable mean ability measures on CAT and pencil and paper test versions. Where mean differences are statistically significant, mean measures are higher for a pre-existing pencil and paper test. Three tables and two figures summarize the study results. (Contains 21 references.) (SLD)