The simulation of large-strain geotechnical laboratory tests with conventional Lagrangian finite element method (FEM) techniques is often problematic due to excessive mesh distortion. The multiple reversal direct shear (MRDS) test can be used to measure the residual shear strength of soils in a laboratory setting. However, modelling and simulation generally require advanced numerical methods to accom-modate the large shear strains concentrated in the shear plane. In reality, when the standard direct shear (DS) apparatus is used, the MRDS method is prone to two major sources of measurement error:load cap tilting and specimen loss. These sources of error make it difficult or even impossible to correctly determine the residual shear strength. This paper presents a modified DS apparatus and multi-reversal multi-stage test method, simulated using the coupled Eulerian-Lagrangian (CEL) method in a finite element environment. The method was successful in evaluating equipment and preventing both load cap tilting and specimen loss, while modelling large-deformation behaviour that is not readily simulated with the conventional FEM or arbitrary Lagrangian-Eulerian (ALE) analysis. Thereafter, a modified DS apparatus was created for the purpose of analysing mixtures of organic materials found in an Australian clay. The results obtained from the modified DS CEL model in combination with laboratory tests show a great improvement in the measured residual shear strength profiles compared to those from the stan-dard apparatus. The modified DS setup ensures that accurate material residual shear strengths are calculated, a factor that is vital to ensure appropriate soil behaviour is simulated for numerical analyses of large-scale geotechnical projects.