The Akaike Information Criterion, or AIC (Akaike 1973, Sakamoto et al. 1986), has received much recent attention in release-recapture literature (e.g., Burnham and Anderson 1998). The details of the AIC are provided in the SURPH.1 manual. The smaller the value of the AIC, the better the fit of the model to the observed data. The AIC, unlike the Likelihood Ratio Test, has been used to select among non-nested models. In the example below, the model "SPer" has the smallest AIC (12335.5) and would be selected as the better fit based on the AIC.
