Two decades after their birth, the validity of fault slip inversion methods is still strongly debated. These methods are based upon a very simplified mechanical background, the Wallace-Bott hypothesis. Following previous studies, the 3D Distinct Element Method (3DEC software) is used to explore the effect of varying stress anisotropy (i.e. the "shape" ratio) on slip perturbation along pairs of faults. Two end-member configurations are modelled in taking into account fault friction and internal deformation of faulted blocks. The first model deals with a relatively simple case where two nonintersecting conjugate normal faults are reactivated in an oblique normal stress regime. The second one simulates an extreme situation where two perpendicular intersecting faults are submitted to oblique extension. The average direction of fault slip predicted by 3DEC models is compared to the corresponding slip predicted by the simplified Wallace-Bott model. For the two simulated cases, it is shown that results from 3DEC and Wallace-Bott models are mutually consistent and argue for the validity of fault slip data inversion methods. Consistency remains even if slip is significantly deviated near the intersection line of faults. These deviations depend on the degree of anisotropy of applied stresses in presence of fault friction. Furthermore, modelling results suggest that, for intersecting faults with convergent slip directions, consideration of fault friction in the models leads to reduction of slip perturbation. In other words, modelling results lead to the nonintuitive conclusion that the validity of the simplified Wallace-Bott model is strengthened when 3DEC model's complexity (i.e. the number of parameters incorporated) increases. -cr 2002 Elsevier Science B.V. All rights reserved.