Abstract: | Application of forecasting models reveals that parameter estimates vary substantially among regions reflecting spatial variation in consumer preferences and behavior. We present a simulation methodology to evaluate the stability in performance of recreation participation models between different regions. The sample data consisted of 632 household interviews obtained via random digit dialing techniques as part of Oklahoma's 1980 statewide survey of outdoor recreation demand. Three variables, the age of the head of the household, the number of household members, and distance to the facility were consistently accurate predictors of participation; the performance of other variables depended directly upon unknown attributes of the specific region. |