Researchers have come to different conclusions about the usefulness of habitat-relationship models for predicting species presence or absence. This difference frequently stems from a failure to recognize the effects of spatial scales at which the models are applied. We examined the effects of model complexity, spatial data resolution, and scale of application on the performance of bird habitat relationship (BHR) models on the Craig Mountain Wildlife Management Area and on the Idaho portion of the U.S. Forest Service's Northern Region. We constructed and tested BHR models for 60 bird species detected on the study areas. The models varied by three levels of complexity (amount of habitat information) and three spatial data resolutions (0.09 ha, 4 ha, 10 ha). We tested these models at two levels of analysis: the site level (a homogeneous area <0.5 ha) and cover-type level (an aggregation of many similar sites of a similar land-cover type), using correspondence between model predictions and species detections to calculate kappa coefficients of agreement. Model performance initially increased as models became more complex until a point was reached where omission errors increased at a rate greater than the rate at which commission errors were decreasing. Heterogeneity of the study areas appeared to influence the effect of model complexity. Changes in model complexity resulted in a greater decrease in commission error than increase in omission error. The effect of Spatial data resolution on the performance of BHR models was influenced by the variability of the study area. BHR models performed better at cover-type levels of analysis than at the site level for both study areas. Correct-presence estimates (1 - minus percentage omission error) decreased slightly as number of species detections increased on each study area. Correct-absence estimates (1 - percentage commission error) increased as number of species detections increased on each study area. This suggests that a large number of detections may be necessary to achieve reliable estimates of model accuracy.