Comparability of scores from cat and paper and pencil implementations of student selection examination to higher education

Abstract

The purpose of this study was to investigate the possibility of computerized adaptive testing (CAT) format as an alternative to the paper and pencil (P&P) test of the student selection examination (SSE) in Turkey. The scores obtained from both P&P format of the SSE and CAT through post-hoc simulations were compared using science subtest items. Different test termination rules (fixed length and fixed standard error) and ability estimation methods (EAP and MLE) were used to operate the CAT version of the SSE P&P test. 10, 15 and 25 items were used as fixed length test and standard errors of 0.30, 0.20 and 0.10 were used as fixed standard error thresholds in terms of test termination rules. Results indicated significant correlations between scores from SSE and CAT. The comparisons between results obtained from CAT and P&P tests also revealed that there exists similar ability distributions and significant reduction in the number of items used through CAT. The findings from the research showed that CAT could calculate reliability using fewer items than P&P test. This study suggests that CAT can be an alternative to SSE with comparable scores to P&P format.