Objective: To quantify concordance between recommendations to repair a questionable restoration based on: (1) hypothetical clinical scenarios in a questionnaire; and (2) actual clinical treatment by the same dentists. Methods: The study group comprised 197 dentists in the National Dental Practice-Based Research Network (www.NationalDentalPBRN.org) who participated in two studies: “Study A”, a questionnaire about repair or replacement of questionable restorations based upon three hypothetical clinical scenarios, and “Study B”, a clinical study with the same dentists in which data were collected about repair or replacement of existing restorations during actual clinical treatment. In Study B, each dentist repaired or replaced approximately 50 existing restorations. We report the rates at which existing restorations were repaired (as compared to replacing the entire restoration) in the clinical study (Study B) based upon whether the dentist treatment planned a repair (instead of replacement) in the hypothetical scenarios (Study A) for restorations of similar type and clinical status. Wilcoxon’s Rank Sums Test was used at the 0.05 level of significance. A majority of dentists (61/[61+23]=73%) chose repair over replacement for the composite with enamel stain, as compared to the composite with dentinal caries (41/[41+68]=38%). Conclusion: Dentists who chose to repair in each of the three hypothetical scenarios had higher repair rates of similar restorations in actual clinical practice.