Self-evaluation

The way we think about ourselves is reflected in, and shaped by, the way we think about our actions. The identification of action can vary from mechanistic, low level depictions to far more comprehensive, high level depictions. This study tests the supposition that comprehensive understanding of action is a precondition for coherence in self-concept and that failure to generate stable, high level action identifications interferes with the formation of a stable and personally certain self-concept. It was predicted (and the results confirmed) that persons who experience fragmented action understanding (i.e., relatively low level action identification) demonstrate less coherence in their overall self-evaluative stance and greater self-uncertainty and self-instability. The relationships between self-concept and coherence (i.e., self-stability and self-certainty) and action identification level was studied in both a context-free manner and with respect to specific social roles. The results expand our understanding of the nature and impact of the organizational processes at work within the self system.

This work argues that the self can be viewed as a dynamical system, with lower-level cognitive and affective elements that interact over time giving rise to global patterns of thought and behavior. The underlying structural features of the self-system promote temporal variation in self-evaluation in the stream of thought. To capture the dynamics of self-evaluation, a refinement of the "mouse paradigm" is employed. This procedure assesses the stream of self-evaluative thought and thus provides insight into the global structural features of the self. The goals of this research were to establish the validity of the "mouse paradigm" and to explore the relationship between the structure and dynamics of the self as expressed in self-reflection. The "mouse paradigm" is a procedure where individuals talk about themselves and subsequently indicate the valence of their self-description using a computer mouse. Participants move the mouse pointer (towards a target in the center of the screen to indicate a positive self-evaluation and away from the target to indicate a negative self-evaluation) while listening to the previously recorded version of their self-description. This technique makes it possible to examine self-evaluation as it changes over time. In study 1, participants described themselves after exposure to various self-relevant memory primes (positive, negative, mixed, and no prime). Traditional measures of the structural features of the self were found to be systematically related to the dynamic properties of participants' mouse movements (distance from target, variance, rate of change, and time at rest). Also, the priming manipulations were found to significantly alter the valence and dynamic properties of participants' self-evaluation. Study 2 explored the relationship between individual differences in self-structure and susceptibility to external influence. Participants were asked to describe themselves from three different perspectives (ideal self, actual self, and feared self). Perspective taking was found to systematically alter the nature of participants' self-evaluations. Furthermore, participants with relatively "weak" self-structures were found to be more influenced by the perspective manipulation. In general, results suggest that global properties of self-structure are related to the temporal flow of self-evaluation. Furthermore, individual differences in self-structure result in different levels of susceptibility to extrinsic influence.

Dynamical systems research has found evidence that global properties of self-system, i.e., self-esteem (SE), self-stability (SS), and self-certainty (SC), emerge from coherence of self-relevant information. Self-certainty may play a role in self-esteem and self-stability. Traits were deconstructed into levels from extremely to not at all. Ss indicated frequencies of trait levels, completed Rosenberg SE, SS; and self-evaluations of traits, roles, certainty, and importance. Entropy H identifies meaningful distributions of deconstructed traits. H may be useful alternative to variance from the mean in identifying individuals for whom a trait is meaningful despite the high variability of endorsements. Low H indicated order. H correlated negatively with SC, SE, SS, ratings of traits, roles, and importance as predicted. SC accounted for some of the variance of the relations; SS, however, was found to play a greater role in the relations of entropy and global properties.