Latent Partition Analysis of Attributions for Actual Achievement

Abstract

Two experiments were conducted to examine undergraduate students’ open-ended attributions for their performance on a midterm exam. In Experiment I, 53 different attribution statements were generated from 85 undergraduate students and were categorized by another group of students in a different section of the course. A latent partition analysis (LPA) of subjects’ categories of attribution was performed. Seven interpretable attribution clusters were derived from the LPA. A second-order factor analysis of the seven attribution clusters yielded two factors, success and failure. Experiment II was a cross-validation of Experiment I with different subjects. The findings in Experiment II were, in large part, similar to those reported for Experiment I. The combined attribution clusters reported in the two experiments reflect the locus of control and stability dimensions noted by Weiner. However, the second-order factor patterns in both experiments indicated that “naive psychologist” college students differentiated attribution clusters only into success and failure ascriptions.