Advances in biotelemetry have greatly increased the spatial and temporal resolution of data that can be obtained from wild mammals. These data have afforded opportunity to examine movement and associated distribution, dispersal and habitat selection. Most applications employ a use/available design using the logistic discriminant derived by G. A. F. Seber—I will explain how this is distinct from logistic regression analysis. Scale is a continuing challenge in habitat selection studies because of data constraints but also because mammals can select at a variety of spatial and temporal scales. We have proposed a new method for habitat selection, integrated step-selection analysis, that explicitly uses movement to constrain the available resource units focusing scale to the behavioural selection process. I will review case studies for ecology and conservation of elk (Cervus elaphus), cougars (Puma concolor), grizzly bears (Ursus arctos), and wolverines (Gulo gulo). Individual variability in habitat selection can be high for each of these species sometimes warranting the characterisation of a functional response shaped by learning.