Psychometric modeling of computer abilities in first-year university students

Abstract

This article informs on the process and the results of an instrument aimed at measuring the level of computer competencies in first-year university students. The new instrument, which was constructed from an existing questionnaire (originally for an English-speaking population), measures six competencies using 45 items in which students express their familiarity with specific computer tasks. Data included responses from two cohorts of medical students (for a total of 2 389 participants): Based on the results for the first cohort, items were selected for the questionnaire’s final version, which subsequently was evaluated in a cross-validation study on the second cohort. The analyses were based on (a) a variant of the common factor model and a novel variant of the generalized partial credit model that accounts for local item dependencies, and (b) structural equation modeling which examines the relations among the six abilities measured by the instrument. The psychometric results showed good fit of the models to the data and adequate reliability and validity indices. The structural analysis revealed prerequisite relations among the abilities. The results lend support to the instrument as a functional tool in the students’ formative evaluation process.