This study used epidemiologic case-control methodology to examine habitat and environmental factors contributing to amphibian declines in Arizona. Risk factors were compared between sites where frogs disappeared (cases) and persisted (controls) using univariate and multivariable logistic regression analyses. Thirty-six percent (117/324) of all sites became cases during the study period. Elevation, non-native predators, hydrologic characteristics, aspect, and effects of nearby sites were significantly associated with frog persistence or disappearance. In the final multivariable model, risk for disappearance increased with increasing elevation (OR=2.7 for every 500 meters, P<0.01). Sites where disappearances occurred were 4.3 times more likely to have other nearby sites that also experienced disappearances (P<0.01), while having an extant population nearby decreased risk of disappearance by 85% (OR=0.15, P<0.01). Sites experiencing disappearances were 2.6 times more likely to have crayfish than control sites (P=0.04). Identification of risk factors associated with frog disappearances will guide future research and conservation efforts.