Consistency of Self-Ratings Across Measurement Conditions and Test Administrations

Abstract

An experiment was conducted to examine the consistency of self-ratings of ability across four different measurement conditions and the relative leniency of retrospective pretest self-ratings compared to pretest self-ratings. Participants in a continuing education conference, randomly assigned to one of four measurement conditions, rated their knowledge of the conference content three times: pretest, retrospective pretest, and posttest. Of the four measurement conditions, two yielded significantly different mean self-rating scores (p < .05). Participants who were asked tojustify their self-ratings to the investigator rated themselves significantly lower than participants who were encouraged to keep their selfratings as evidence of their achievement for future career advancement. In addition, retrospective pretest self-ratings (p <. 001), and females rated themselves significantly higher than males rated themselves (p < .05). These findings have implications for evaluators and other users of self-ratings.