The Influence of Response Distortion in Assessing Self-Perceptions of Physical Ability and Attitude Toward Physical Activity

Publisher Website
Google Scholar
Abstract
Response distortion is an acknowledged concomitant variable in psychometric research. Although the influence of response bias on the Estimation scale of the Physical Estimation and Attraction Scales (PEAS) has previously been considered for adolescent males, experimental evidence concerning the sensitivity of the adult versions of both the Attraction and Estimation scales to “faking” has not been provided. To determine this sensitivity, adult versions of the PEAS were administered to male and female undergraduates (N = 122). Subjects were randomly assigned to one of three groups and received either “fake bad,” “fake good,” or standard instructional sets. Results generally supported the view that the total score on each of the Attraction and Estimation scales is relatively insensitive to attempts to “fake good.” However, attempts to “fake bad” consistently resulted in a significant decrease (p < .01) in both Attraction and Estimation scores as compared to a standard control group. These findings suggest the need for a criterion to detect suspected “fake bad” Attraction and Estimation responses and possibly a “fake good” Attraction response. Statistical procedures for determining test bias indicated that specific PEAS items discriminated between the experimental and control groups, and these items were incorporated into embedded scales for the detection of a distorted PEAS response. Evidence for the predictive validity of these “fake bad” and “fake good” scales was provided.