<p>This study evaluates the influence of three multistage direct shear methods on the mechanical response of five gravel-sand-fines mixtures with fines contents ranging from 5 to 40%. The mixtures were compacted to approximately 100% of the maximum dry unit weight to ensure a consistent initial state. The methods differ in the degree of shear displacement reset between stages (none, partial, or full). The mechanical response is characterised by void-ratio evolution, water-content changes, dilatancy, stress–dilatancy behaviour, and derived shear strength parameters. The Revised Soil Classification System was used to identify the dominant grain fraction. The results indicate that, although the intrinsic stress–dilatancy mechanism remains unchanged, the degree of peak-strength mobilisation is strongly influenced by the displacement-reset procedure. In coarse-grained mixtures, all methods produce strength values within a narrow range. In fine-dominated mixtures, full-reset mobilises higher peak shear stresses than no-reset, while method-dependent differences diminish due to the suppressed dilatant response. Among the mixtures, the transitional one is the most sensitive to the shear displacement reset procedure. No-reset underestimates peak shear stress, full-reset overestimates it, and partial-reset aligns closely with single-stage values. Multistage testing reduces material demand and testing time, but the obtained shear-strength parameters depend on the interaction between soil composition and the displacement-reset strategy. Therefore, reliable application requires explicitly accounting for the degree of fabric renewal between stages, rather than assuming method-independent results.</p>