Abstract: Self-esteem, viewed for decades as psychology’s Holy Grail, has proved to be an elusive and surprisingly barren destination. One of the oldest concepts in psychology, self-esteem likely ranks among the top three covariates occurring in personality and social psychology research. Propelled by the self-esteem movement of the 1970s, it was popularly believed that self-esteem played a significant causal role in determining a wide range of both positive and negative social behaviors. However, the results of a 2003 large-scale literature review showed that it is actually not a major predictor of almost anything, with the exception of positive feelings (happiness) and possibly greater initiative. With over a decade passing since that publication, the current investigation sought to review, synthesize and critically analyze the more recent literature. Results confirmed a similar dearth of meaningful relationships connected to self-esteem, the only exceptions being some modest correlates to happiness, narcissism, and self-perceived attractiveness and intelligence. However, the literature continues to be plagued with myriad conceptual and methodological problems. This study utilizes specific critical thinking principles to advance understanding in this area, to address why the self-esteem obsession still persists, and to propose a new theoretical model, one that accounts for the construct’s heterogeneous and multidimensional nature. Self-esteem is defined as the appraisal of one’s own personal value, including both emotional components (self-worth) and cognitive components (self-efficacy). The multiple forms of self-esteem are a function of how accurately or closely it matches an individual’s measureable reality, composed of the objective outcome of one’s behavior (actual achievements, measurable capabilities) as well as one’s interpersonal interactions (i.e., the level of congruence between how one thinks he or she is perceived and how he or she is actually perceived). Self-esteem also varies in terms of its level of stability, or the degree to which it is influenced by evaluative events or the need to match external standards across time and situation. The permutations of these sorting variables yield eight types of self-esteem: Optimal High, Fragile High, Accurate Low, Fragile Low, Non-compensatory Narcissism, Compensatory Narcissism, Pessimal, and Disorganized. Specific recommendations for clinicians and researchers are provided.