The Group Engagement Measure (GEM) assesses a commonly used, but rarely measured, process in group work. Earlier studies examined the reliability and validity of the GEM, but none empirically examined its factor structure. The authors examined the seven-factor, 37-item structure of the GEM, using confirmatory factor analysis involving a combined clinical and nonclinical sample of 207 adults. On the seven-factor, 37-item model, the fit indices standardized root mean square residual (SRMR) and root mean square error of approximation (RMSEA) were .07, considered a fair fit; the comparative fit index (CFI) was .91, considered acceptable. On the seven-factor, 27-item GEM the SRMR was .05, the RMSEA was .05, and the CFI was .95, indicating a good fit; however, neither model was invariant across the clinical and nonclinical groups. Tests for factor invariance indicated that a five-factor, 21-item model produced the best fit across groups (CFI = .96, SRMR = .04, RMSEA = .06). Testing of these models is needed with more people.