A direct observation of crack propagation in the microbond test was carried out for five different fiber/polymer matrix systems. This technique appeared to be a very effective tool for interface characterization. Experimental plots of the force required for further crack propagation as a function of debond length were analyzed using both energy-based and stress-based models of debonding. The fracture mechanics analysis was used to construct families of crack resistance or R-curves which showed the variation of energy release rate, G, with the debond length, and included the effect of interfacial friction in debonded regions. For the first time, analogs of the R-curves were created within the scope of the stress-based model to present the local shear stress near the crack tip, T, as a function of crack length. In both models, the behavior of the interfacial parameter (G or tau) strongly depends on the assumed value of the interfacial frictional stress (tau(f)). However, for each matrix/fiber system there exists such a tau(f) value for which the investigated parameter is nearly constant over the whole region of stable crack propagation (70-90% of the embedded length). Moreover, these best-fit tau(f) values for each specimen appeared to be practically the same for both energy-based and stress-based approaches. Thus, both interfacial toughness, G(ic), and local interfacial shear strength, tau(d), adequately characterize the strength of a fiber/matrix interface. Extrapolation of R-curves and their analogs to zero crack length allows measurement of the interfacial parameters with good accuracy.