Beta diversity quantifies the spatial structuring of ecological communities and is a fundamental partition of biodiversity, central to understanding many macroecological phenomena in modern biology and paleobiology. Despite its common application in ecology, studies of beta diversity in the fossil record are relatively limited at regional spatial scales that are important for understanding macroevolutionary processes. The spatial scaling of beta diversity in the fossil record is poorly understood, but has significant implications due to temporal variation in the spatial distribution of fossil collections and the large spatiotemporal scales typically employed. Here we test the spatial scaling of several common measures of beta diversity using the Cenozoic shallow-marine molluscan fossil record of New Zealand and derive a spatially standardized time series of beta diversity. To measure spatial scaling, we use and compare grid-cell occupancy based on an equal-area grid and summed minimum spanning tree length, both based on reconstructed paleocoordinates of fossil collections. We find that beta diversity is spatially dependent at local to regional scales, regardless of the metric or spatial scaling utilized, and that spatial standardization significantly changes apparent temporal trends of beta diversity and, therefore, inferences about processes driving diversity change.