The current work proposed a new conceptualization and measurement of self-complexity based on Carlston’s (1994) Associated Systems Theory (AST). AST is a systematic approach to classifying the cognitive representations of social entities and as such, it provides a rich and theoretically-based framework for examining self-concept representation in ways that existent approaches cannot. In the current study, an AST-based approach to assessing self-complexity was related to other measures of self-concept representation complexity (i.e., number of self-aspects, H), providing evidence of convergent validity for this new approach. Moreover, the AST-based measure showed evidence of the buffering effect whereas traditional measures did not. That is, when facing greater stress in their lives, those reporting greater self-complexity (as assessed by our AST-based measure) reported greater well-being (i.e., fewer physical illnesses and less depression). Implications of these results for understanding self-concept representation, AST, and their consequences are discussed.