Abstract
A logit regression model of participation in fishing, hunting, and noncon-sumptive wildlife recreation is estimated based on national survey data. The model is specified both with and without price and cross-price variables. Otherwise, the nonprice variables included are the same in both equations. The two equations are used in long-run forecasts of participation and the effects are noted. The null hypothesis of no price effect is rejected in two of three cases. Apparently, including price and cross-price variables can improve both the accuracy and usefulness of long-run forecasts of participation in outdoor recreation.