Sampling procedures are described for assessing bird damage to sprouting rice. Optimum sampling schemes are derived for both large-scale surveys and controlled experimental research from data collected from a 1980 controlled experiment and a 1981 survey. A basic recommendation is that each field be sampled by using a stratified or systematic design, depending on the objective of the specific study. Sampling points consist of paired plots, one of which is protected from bird damage by a wire exclosure, that are located on transects placed randomly across the field. Methods are derived for calculating the number of plots per transect and the number of transects necessary to achieve a given level of precision or to maximize precision for a given cost. A stratified design is also proposed for conducting large-scale surveys, and recommendations are given concerning the required number of fields for a desired precision.