Purpose: The purpose of this study was to examine the sensitivity of the microtensile bond strength (µTBS) to changes in the specimen design using three different dentin bonding systems, utilizing both experimental and finite element analysis (FEA). Materials and Methods: Freshly extracted bovine central incisors were randomly divided into 9 experimental groups. Flat dentin surfaces were exposed and polished with 600-grit SiC papers. The dentin bonding agents (DBA) Adper Prompt L-Pop, Adper Scotchbond 1XT, and Adper Scotchbond Multi-Purpose Plus were bonded to the dentin surface. A 5- to 6- mm-thick layer of resin composite Filtek Supreme was built up incrementally. For each DBA, after 24 h in water, the bonded specimens were sectioned vertically into slabs of stick-, dumbbell- and hourglass-shaped specimens (n = 30). The µTBS test was performed at a crosshead speed of 0.5 mm/min and the fracture surfaces were examined under optical microscopy and SEM. FEA models of the three designs were developed using the TOMECH program and viewed using FEMSYS visualization software. Results: One-way ANOVA, Tukey's and Kaplan-Meier tests revealed no significant differences between the bond strengths for the three designs for the same dentin bonding agent (p > 0.05). However, there were significant differences between the three adhesive systems when using one specimen design. FEA results showed no major differences between the three designs in the maximum stress within the adhesive layer, although the stress distributions for the hourglass shape followed a different pattern. Conclusions: The three different specimen designs had no influence on the bond strength to dentin within one adhesive system. FEA predictions are consistent with the results of the experiments. Keywords: microtensile bond strength, specimen design, FEA, dentin bonding agents