Computer-based assessment of school readiness and early reasoning.

Abstract

This study explores the potential of using online tests for the assessment of school readiness and for monitoring early reasoning. Four tests of a face-to-face-administered school readiness test battery (speech sound discrimination, relational reasoning, counting and basic numeracy, and deductive reasoning) and a paper-and-pencil inductive reasoning test were transferred to an online platform and administered at the beginning of school to samples of first-grade children (the sample sizes were between 364 and 435). Results of the original and the computerized tests were analyzed to explore (a) whether the new scales were identical to the original ones; (b) how the change of media influenced the reliability of the tests; and (c) whether the migration into a new medium affected gender differences. Analyses indicated that measurement invariance held in a strict sense in the case of the inductive reasoning test (the migration did not change neither the general look of the test nor the item types) and only partially for the speech sound discrimination test (neither the item type nor the scoring principle was changed). Measurement invariance did not hold for the 3 remaining tests. In 3 tests—speech sound discrimination, relational reasoning, and deductive reasoning—the online versions demonstrated improved reliability. Only certain items of the numeracy test could be assessed on computer, and the reliability of the shortened test decreased. No differences were found between the 2 versions of the inductive reasoning test. Gender differences were explored for the speech sound discrimination test, and latent analyses indicated that measurement invariance did not hold. Girls’ performance was somewhat better, similarly to former face-to-face assessments, where girls performed slightly better than boys. These results encourage further research on the extension of computer-based assessment to the early childhood education.