Self-other agreement and self-other differences on multidimensional self-concept ratings

The present investigation examines self-other agreement in relation to an intervention previously shown by Marsh, Richards, and Barnes (1986a, 1986b) to enhance self-concept. Subjects completed multidimensional self-concept instruments before and after the completion of Outward Bound courses (self-responses), and made ratings of how other participants “would” respond and “should” respond (observer responses). A multitrait-multimethod analysis demonstrated the convergent and discriminant validity of both self-and observer-responses. Whereas there was little support for the ability of observers to meaningfully differentiate between “would” and “should” responses, “would” responses were somewhat better predictors of self-responses. Consistent with previous research pre/post-intervention differences for self-responses were statistically significant. Observer ratings were more positive than self-responses at either time, and observers indicated that subjects “should” respond more favourably than they “would” respond. These results indicate a modesty bias in the self-responses that runs counter to the frequently posited self-favourability bias in self-concept responses and support previous interpretations of the Outward Bound intervention effect.