PURPOSE:To compare the mean shear bond strength (SBS) and failure mode of a resin-modified glass-ionomer luting cement (RM-GIC) to five different metals using unsupported and supported cement specimens with different placement of the shear load. MATERIALS AND METHODS:A RM-GIC was bonded to five metals using "unsupported" and "supported" techniques at a SBS-specimen diameter of 2.36 mm. The bond was stressed to failure using shear knife and wire loop debonding protocols. For the shear knife method, the distance of the shear force from the interface was 0 mm or 0.3 mm. Failure analysis was assessed by stereomicroscope and SEM. RESULTS:Two-way ANOVA and post-hoc Tukey's test revealed a significant difference between the unsupported and supported mean SBS. The SBS of supported specimens, where the shear force was applied to the mold that enclosed the specimens, were in most cases statistically significantly higher (p < 0.05) than specimens that were not supported. The mean bond strengths of RM-GIC ranged from 4.5 ± 2.3 MPa to 27.4 ± 3.7 MPa. Analysis of the failure mode showed significant differences (p < 0.001) for the test methods except for adhesion to gold-based metal. The adhesive failure mode was between 91% and 97% for supported specimens and between 47% and 63% for unsupported specimens. CONCLUSION:Within the limits of this study, supported specimens exhibited higher mean SBS than unsupported specimens. The method of debonding had a significant effect on the mean SBS for RM-GIC bonded to metal. Mold-supported specimens had a higher incidence of adhesive failure than unsupported cement specimens.