We used capture, harvest, and observation histories of Canada geese (Branta canadensis) banded in the Mississippi flyway, 1974-88, to examine the problem of neck-band retention. Methods for the analysis of survival data were used to estimate rates of neck-band retention and to evaluate factors associated with neck-band loss. Sex, age of bird at banding, rivet use, and neck-band type significantly influenced neck-band retention. For most of the resulting cohorts (e.g., sex, age, rivet, and neck-band type categories), neck-band retention rates decreased through time. We caution against using small samples or data collected during short-term studies to determine retention rates. We suggest that observation data be used in neck-band retention studies to increase the efficiency of estimating retention time.