We studied a fossorial endemic lizard (the sand skink, Plestiodon reynoldsi) to determine habitat structural and vegetation associations in altered habitats at micro-, local, and macroscales. Lizards were present in each of the structural categories and vegetation types we studied, although active orange groves negatively influenced distribution and relative density (determined as the proportion of coverboards within a sampling plot containing signs of sand skink presence). Conversely, relative densities were highest in forested habitats, which contain structural features similar to natural habitats. Sand skinks readily used the two soil types at our site, although the greatest densities were in sampling plots containing both types. We conclude that conditions are suitable for sand skink habitation within a variety of habitats with different structures, including those altered by humans. However, in all cases the underlying soil was also loose and dry, making fossorial locomotion possible for this species. Three lines of evidence suggest that populations in altered habitats were resident, rather than transient: (1) sand skinks are still present in altered habitats more than 35 skink generations after alteration occurred; (2) there was no relationship between sand skink density and distance to natural habitat; and (3) a very small portion of our site (<2%) consisted of natural habitat, which is most likely too small to be a long-term source population to nearby altered habitats. Our results indicate that although P. reynoldsi is often considered habitat-specific, this notion may be due to focusing sampling efforts on natural rather than disturbed habitats. Therefore, relying on preconceived notions of habitat associations may not be sufficient to understand the ecological relationships and local-scale distribution of this threatened species. Using such misinformation may lead to the design and implementation of inadequate conservation plans that ignore altered habitats in which focal species occur.