Practitioners often choose resin materials and temporary cements with little understanding about their effect on provisional crown retention.This study evaluated the retention of provisional restorations made with 2 materials and cemented with 4 temporary cements.Recently extracted molars were prepared with a flat occlusal surface, 4-mm axial length and 20-degree angle of convergence. Specimens were distributed into equivalent groups. Provisional crowns were constructed for each preparation with polymethyl methacrylate (Temporary Bridge Resin) or bis-acrylic composite (Protemp Garant) and later cemented with Temp-Bond, Temp-Bond NE, Temrex, and an experimental calcium hydroxide temporary cement. A second group with Temrex was evaluated using half the recommended liquid. A cementing force of 2.5 kg for 5 minutes was used. After initial bench set followed by 24 hours in room temperature water, the crowns were removed with an Instron mechanical testing machine at 0.5 mm/min. A 2-factor ANOVA was used with alpha=.05 (n = 10). Mode of debonding was analyzed with a nonparametric chi-square test of association.Mean dislodgment stresses ranged from 670 to 1072 kPa for polymethyl methacrylate crowns and 554 to 884 kPa for those made of composite. Differences were nearly significant for the type of provisional material (P =.061) and the cross-product interaction (P =.376) was not significant, whereas there were significant differences among the cements (P =.002) and the mode of debonding (P =.0034).Excluding Temp-Bond to eliminate a cross-product interaction demonstrated that the polymethyl methacrylate crowns were 19.3% more retentive than the composite crowns (P =.015). There was no statistically significant difference among the 4 temporary cements when the manufacturer's mixing instructions were followed (P =.186). However, the thicker consistency Temrex was more retentive than the recommended Temrex mix and Temp-Bond.