In the United States, housing density has substantially increased in and adjacent to forests. Our goal in this study was to identify how housing density and human populations are associated with avian diversity. We compared these associations to those between landscape pattern and avian diversity, and we examined how these associations vary across the conterminous forested United States. Using data from the North American Breeding Bird Survey, the U.S. Census, and the National Land Cover Database, we focused on forest and woodland bird communities and conducted our analysis at multiple levels of model specificity, first using a coarse-thematic resolution (basic models), then using a larger number of fine-thematic resolution variables (refined models). We found that housing development was associated with forest bird species richness in all forested ecoregions of the conterminous United States. However, there were important differences among ecoregions. In the basic models, housing density accounted for < 5% of variance in avian species richness. In refined models, 85% of models included housing density and/or residential land cover as significant variables. The strongest guild response was demonstrated in the Adirondack-New England ecoregion, where 29% of variation in richness of the permanent resident guild was associated with housing density. Model improvements due to regional stratification were most pronounced for cavity nesters and short-distance migrants, suggesting that these guilds may be especially sensitive to regional processes. The varying patterns of association between avian richness and attributes associated with landscape structure suggested that landscape context was an important mediating factor affecting how biodiversity responds to landscape changes. Our analysis suggested that simple, broadly applicable, land use recommendations cannot be derived from our results. Rather, anticipating future avian response to land use intensification (or reversion to native vegetation) has to be conditioned on the current landscape context and the species group of interest. Our results show that housing density and residential land cover were significant predictors of forest bird species richness, and their prediction strengths are likely to increase as development continues.