Psychological predictors of the propensity to omit short-response items on a high-stakes achievement test

Gabrielle Matters, Queensland University of Technology
Paul C. Burnett, Charles Sturt University

Abstract

This article presents the findings of a study of the psychological variables that discriminate between high and low omitters on a high-stakes achievement test using a shortresponse format. Data were obtained from a questionnaire administered to a random sample (N = 1,908) of students prior to sitting the 1997 Queensland Core Skills (QCS) Test (N = 29,273). Fourteen psychological variables were measured including test anxiety (four subscales), emotional stability, achievement motivation, self-esteem, academic self-concept, self-estimate of ability, locus of control (three subscales), and approaches to learning (two subscales). The results were analyzed using descriptive discriminant analysis and suggested that the psychological predictors of the propensity to omit shortresponse items include test-irrelevant thinking and academic self-concept, with sex of candidate being a mediating variable.