To understand changes in habitat selection in response to timber harvesting, we used radio-telemetry data from 82 adult wood frogs (Lithobates sylvaticus, formerly Rana sylvatica) and logistic regression modelling to assess habitat selection in response to an unharvested control and 3 forest management techniques: clearcutting (with removal of all merchantable timber > 10 cm diameter), clearcutting with coarse woody debris (CWD) retention, and partial harvesting with retention of ∼50% canopy cover. At the home range scale, frogs selected the partially harvested treatment in spring 2005 and avoided the CWD-retained treatment in fall 2006. Frogs spent 5 ± 2 d (mean ± SE) longer in forested treatments than in both clearcut treatments, but certain individuals were able to specialize on the clearcut treatments. At the weekly activity centre scale, the best-supported models indicated that frogs were more likely to occupy locations with more complex ground structure, especially coarse woody debris, warmer temperatures, moister substrates, and greater canopy cover than random. Resource use among frogs overlapped substantially at both the weekly activity centre and daily microhabitat scales. Frogs selected daily microhabitats with more complex ground structure, greater canopy cover, and moister substrates than random. Selection at coarser scales may be mechanistically linked to finer scale resource selection by the physiological processes of thermoand hydro-regulation. Our results support recommendations for minimizing the impact of logging by retaining coarse woody debris in clearcuts and partial harvesting with retention of ∼50% canopy cover.