Land-cover maps are used in numerous natural resource applications to describe the spatial distribution and pattern of land-cover, to estimate areal extent of various cover classes, or as input into habitat suitability models, land-cover change analyses, hydrological models, and risk analyses. Accuracy assessment quantifies data quality so that map users may evaluate the utility of a thematic map for their intended applications. Despite the widespread acceptance of accuracy assessment and the numerous articles published on this topic, the basic structures of a statistically rigorous accuracy assessment have not been fully described. The objective of this article is to elucidate these fundamental structures. We describe the three basic components of an accuracy assessment, the sampling design, the response design, and the estimation and analysis pro tocol, and we provide recommendations and general guidelines for a statistically rigorous assessment.