Competitive and global models examined to explain variation in Cooper's hawk annual survival on two north-Florida study areas, 1995–2001. Sample size was 16 male and 26 female Cooper's hawks trapped at nest areas and fitted with backpack-mounted radio-transmitters; many individuals were monitored for >1 y, up to a maximum of 5 y. Model ranking was based on Akaike's Information Criterion (AIC) corrected for small sample size (AICc). All competitive models (within 2 AIC units of the top model) and the global model are shown. Survival was estimated in Program MARK using a sin-link function.