A debate exists concerning whether exclusion harms self-esteem. We hypothesized that social exclusion does harm self-esteem, but that this effect is evident only when self-presentational concerns to "appear fine" are minimal or people are unable to alter their report of self-esteem. In the first three studies, participants’ explicit and implicit self-esteem were measured following an exclusion or comparison condition where self-presentational pressures were likely high. Because respondents can easily control their reports on explicit measures, but not on implicit ones, we hypothesized that exclusion would result in lower self-esteem only when implicit measures were used. Results confirmed this hypothesis. In the final study, self-presentational concerns were directly manipulated. When self-presentational concerns were high, only implicit self-esteem was lowered by exclusion. But, when such concerns were low, this impact on self-esteem was seen on implicit and explicit measures. Implications for the sociometer hypothesis and the recent self-esteem debate are discussed.