On the use of the Causal Dimension Scale in a field setting: A test with confirmatory factor analysis in success and failure situations.

Abstract

We tested the reliability and validity ofthe Causal Dimension Scale (CDS) within success and failure conditions in a real-life achievement situation. A total of 260 undergraduate students who had just received their marks on a midterm exam were asked to fill out a questionnaire that included questions dealing with their perceptions of success and failure on the exam and on the CDS. Confirmatory factor and internal consistency analyses were performed on the overall sample, as well as on the data from the success and failure conditions. Results showed that internal consistency was found to be adequate for the Locus and Stability subscales but not for the Control subscale. Results of the confirmatory factor analyses provided partial support for the validity of the CDS factor structure in that three-factor solutions were obtained in all three conditions. However, further analyses revealed that an acceptable fit for the data was obtained only when obliqueness, involving the Locus factor with those of Stability and Control, and cross loadings were incorporated in the factor structures. Of particular interest were the findings ofa multigroup confirmatory factor analysis that revealed that solutions for the success and failure conditions were significantly different. The present set of findings provide some support for the reliability and validity of the CDS. However, we suggested that more work be conducted on the scale--especially the Control subscale--before the CDS can be used with full confidence in research.