Measuring Self-Sentiments

Abstract

Chapter 4 compares semantic differential scales employed by ACT-Self to measure self-esteem, self-efficacy, and self-activation with Likert-scale measures of the same constructs. A multitrait-multimethod matrix analysis in conjunction with exploratory and confirmatory factor analysis demonstrates a dramatic lack of convergence between these two methods of measuring self-sentiments. Following a discussion of why this occurs, the chapter concludes with an extensive discussion of the relative advantages of semantic differential over Likert scales as measures of self-sentiments. Among the advantages identified in this discussion, the semantic differential approach to measuring self-sentiments is more consistent with the bipolar nature of affect, avoids the linguistic and cognitive complexity of Likert scales, and enters readily into mathematical models of the self-process.