David L. Nash

Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)




This study investigated subjects' reports of the sources of self-knowlege in light of the growing literature about the shortcomings of social judgment. The hypothesis of this study was the subjects' tendency to report self-observation at the expense of feedback and social comparison as the best source of self-knowledge is based on an implicit evaluation bias similar to Schlenker's (1980) notion of "self-projection" and Jellison and Greens' (1981) "norm of internality". It was predicted that subjects' intuitive notions about the sources are such that self-observation is seen as a generally "better" method of self-knowledge than feedback which, in turn, is better than social comparison.

These hypotheses were tested by having subjects rate story characters who learned about themselves through either self-observation, unsolicited feedback, solicited feedback, or social comparison. Subjects rated the characters on semantic differential type items and rated the source itself for "accuracy", "reliability", and "believability". It was assumed that these character ratings would be influenced and reflect the implicit evaluative notions about the different sources. An additional measure used was a ranking procedure similar to those used by Schoeneman (1981). Several personality scales (Self-Consciousness Scale, State-Trait Anxiety Scale, Social Desirability Scale, and the Self-Monitoring Scale) were also given.

The ranking portion of this study replicated the results of Schoeneman (1981) quite closely. Several significant and theoretically consistent differences were found between the sources of self-knowledge groups on the character rating measures. The predictions that self-observation would be seen most favorably, social comparison would be reported least favorably and solicited feedback would be reported more neutrally were generally supported. Predictions regarding unsolicited feedback and the no source ending group were not supported; they were reported as much worse and better, respectively, than was expected. Examination of the independent variables trait valence, subject sex and character sex produced a variety of main effects and interactions which are typically consistent with findings already reported in the sex-role stereotype literature. The social anxiety subscale of the Self-Consciousness Scale was the only personality scale to produce significant results.