Background: In this article, the authors present clinical factors associated with the type of cement practitioners use for restoration of single-unit crowns. Methods: A total of 202 dentists in The National Dental Practice-Based Research Network recorded clinical details (including cement type) used for 3,468 single-unit crowns. The authors classified crowns as bonded if the dentist used a resin cement. The authors used mixed-model logistic regression to assess the associations between various clinical factors and the dentist's decision to bond. Results: A total of 38.1% of crowns were bonded, and 61.9% were nonbonded; 39.1% (79 of 202) of dentists never bonded a crown, and 20.3% (41 of 202) of dentists bonded every crown in the study. Crowns with excessive occlusal reduction (as judged by laboratory technicians) were more likely to be bonded (P = .02); however, there was no association with bonding and excessive taper (P = .15) or axial reduction (P = .08). Crowns were more likely to be bonded if they were fabricated from leucite-reinforced glass ceramic (76.5%) or lithium disilicate (70.8%) than if they were fabricated from layered zirconia (38.8%), full-contour zirconia (30.1%), full metal (14.7%), or porcelain-fused-to-metal (13.8%) (P < .01) restorative material. There was no significant association between choice to bond and crown margin location (P = .35). Crowns in the anterior maxilla were more likely to be bonded (P < .01). Conclusions: Excessive occlusal tooth preparation, anterior location of a crown, and the use of glass ceramic crowns were associated significantly with the decision to bond. Practical Implications: In this study, the authors identified factors significantly associated with the clinical decision made by practicing dentists when selecting a cement for restoration of single-unit crowns.