Laboratory specimen preparation methods can significantly influence the dynamic shear-induced volumetric response of sands. In this study, a medium dense to very dense Ottawa sand (relative density, D_R = ∼60–95 %) was systematically tested using a multidirectional cyclic direct simple shear device to evaluate the influence of specimen preparation on drained cyclic volumetric strain. Both unidirectional and bidirectional cyclic loads were applied to specimens created using four preparation techniques: dry funnel deposition, dry soil raining, dry funnel deposition followed by table vibrations, and dry funnel deposition followed by cyclic shearing. The results indicate that as D_R increases, the effect of specimen preparation on volumetric strain decreases. Specifically, the results illustrate that volumetric strain of medium dense sand (D_R = 57–74 %) is affected by specimen preparation, which is consistent with the literature. In contrast, dense to very dense specimens (D_R > 80 %) exhibited comparable volumetric strain independent of the specimen preparation method. This observation is consistent for both unidirectional and bidirectional cyclic loading.