A key prediction of uncertainty-identity theory is that under conditions of high self-uncertainty, people will identify more strongly with their group. This has been supported by numerous studies. To quantify this relationship, a meta-analysis was conducted on 35 studies from 30 papers (N = 4,657). The relationship between self-uncertainty and group identification varied significantly as a function of how psychologically real the uncertainty was, as reflected in how uncertainty was operationalized and how the study was conducted. Self-uncertainty operationalized as social identity uncertainty had the strongest relationship with identification (r = −.26, 6.8% variance accounted for), followed by indirect operationalization of self-uncertainty (r = .23, 5.3% variance accounted for), and direct operationalization of self-uncertainty (r = .14, 2.0% variance accounted for). The relationship did not differ between measured self-uncertainty (r = −.13, 1.7% variance accounted for) and manipulated self-uncertainty (r = .17, 2.9% variance accounted for). Implications and future directions are discussed.