The LibQUAL + instrument has been widely adopted by libraries to evaluate user perceptions of library service quality. Studies combining groups (e.g., Lane et al., 2012) have shown high correlations between two factors, suggesting the possibility that a two-factor model may fit as well as the three-factor model theorized by the developers. Also, previous studies have not closely examined residuals to analyze local misfit in the context of theory but instead have often correlated error terms to improve model fit. This study uses LibQUAL + responses from undergraduates at a public, comprehensive university to test three-factor, two-factor, and one-factor models of user perceptions of library service quality. Global fit indices indicated that both two-factor and three-factor models were empirically supported, but the three-factor model had better theoretical support. Furthermore, this article adds to the literature the unique perspective of residual analysis and builds theoretical arguments in the interpretation of the final model. Areas of local misfit suggest the need for independent studies to examine residuals. If areas of misfit repeat across institutional populations, that could suggest the potential for further instrument development, while if areas of misfit are unique to institutional populations, this could target areas for institutions to investigate more closely.