Most studies examining continental to global scale patterns of taxon richness, body size or range size rely on range maps overlaid onto an arbitrary grid system. It has long been known that a species does not occur at all locations throughout even the most detailed range map, and therefore range map-derived data necessarily represent patterns at some relatively coarse, but undefined, grain. With the increasing availability of high resolution climate and land cover data, however, broad-scale studies are increasingly likely to estimate richness (or other community parameters) at high resolutions. Due to the scale-dependence of most ecological phenomena, a significant mismatch between the presumed and actual scale of ecological data may call into question both the results of simple correlative analyses as well as the implications for conservation. Here, we examine avian range maps in conjunction with geographically extensive survey datasets on two continents to determine for the first time the spatial resolutions over which range map data can be used to accurately depict patterns in ecology and conservation. We found that at resolutions below 2°, range maps overestimate the area of occupancy of individual species and mischaracterize spatial patterns of species richness. Up to two-thirds of biodiversity hotspots are also misidentified at resolutions finer than 2°. Our results suggest that range map data do not contain as much information as biogeographers and conservation biologists implicitly assume, and that the analysis of spatial patterns at inappropriately fine resolutions can result in erroneous or misleading conclusions.