Construction and validation of a behavioral measure of role-taking

Abstract

This study examines a new method for conceptualizing and measuring roletaking ability. Role-taking is defined in a manner that facilitates further theory building and testing. The task of designing and validating a measure of role-taking that departs from the self-evaluative measures currently used is undertaken and validated with an experimental design. A computer-based survey instrument is created consisting of video and written vignettes designed to test subjects’ ability to predict their study partner’s behavior. It is found that one type of vignette is more suitable for measuring role-taking accuracy than is the other. Females, regardless of experimental condition, record higher role-taking scores than do their male counterparts. Subjects’ self-reported role-taking accuracy is not correlated with their actual role-taking accuracy scores. Because this is the case, it leads to a re-thinking of the meaning of studies that use self-reported ability as the sole measure of role-taking ability. An additional finding is that participants seem to overestimate individual differences. Personality factors measured by the Big Five Inventory were not correlated with role-taking accuracy.