Substantial effort has gone into predicting the characteristics of subsurface fracture systems in fault-related folds. Curvature analysis is a common method used to predict the location and characteristics of fracture networks in folded rock layers. In curvature analysis, it is assumed that layers of rock deform like elastic plates so that layer-parallel strains are directly related to the curvature of the folded surface. This article tests the underlying assumption of all curvature analyses: that curvature is a direct proxy for strain in folded rock layers. We test the assumption by analyzing the curvature and strain in a series of scaled physical models of contractional, basement-involved, fault-related folds. Our objective is to constrain the conditions that lead to a strong positive correlation between curvature and extensional strain. Of particular interest is whether curvature and strain correlate over a wide range of fault throws and dips. The analysis of our folds demonstrates that both the distribution and magnitude of a fold-axis normal extension in the surface of the overlying layer appear to vary as a function of fault dip. Our results indicate that the correlation between strain and curvature generally becomes worse with decreasing fault dip. Fault throw also exhibits an effect on the curvature-strain relationship. However, this effect is dependent on the dip of the fault as well and only exhibits an effect on the curvature-strain relationship for moderately dipping faults. A direct correlation between curvature and strain at all stages of fold development is observed only for steeply dipping basement faults. These results suggest that curvature may not be a consistently reliable strain proxy in basement-involved fault-related folds and that the accuracy of curvature-related strain predictions will be strongly dependent on the dip and throw of the underlying basement fault.