Animals select habitat resources at multiple spatial scales; therefore, explicit attention to scale-dependency when modeling habitat relations is critical to understanding how organisms select habitat in complex landscapes. Models that evaluate habitat variables calculated at a single spatial scale (e.g., patch, home range) fail to account for the effects of other scales on the probability of species occurrence. Although single-scale approaches are common, we hypothesize that such models will be less predictive and incur greater risk of false inferences than multiscale models that account for the effects of habitat variables that are individually scaled to reflect their optimum relation with species occurrence. This follows from the knowledge that an animal's location is not well defined by the effects of habitat conditions occurring at any 1 spatial scale; rather, animal locations represent the cumulative influence of habitat selection across a broad range of spatial scales, from the landscape to the microsite. In this chapter, we describe an approach that uses bivariate scaling, logistic regression, and information theory to derive a multiscale model combining habitat variables at the scales that optimally predict species occurrence. We demonstrate the utility of this approach by investigating habitat relations of the American marten (Martes americana) in northern Idaho using 2 modeling approaches. By comparing how a multiscale model differs in terms of fit, performance, variable use, and spatial prediction from a single-scale model constrained to the scale of habitat patches (i.e., within 90 m of the sampling location), we reveal the importance of considering all relevant spatial scales in habitat models. Our results show that both the strength and the nature of apparent habitat relations are highly sensitive to the scale of predictor variables, and that models that are naive to scale may easily misconstrue the nature of wildlife-habitat relations. We discuss this issue in the context of other marten habitat studies and conclude by describing the implications of our findings for managing marten habitat.