An Application of a Truncated Regression Model to Recreation Demand

Abstract

The problems associated with limited dependent variables in recreation demand models have been generally ignored in previous research efforts. A probable reason for not considering these problems is that practical procedures to test for truncation bias have only recently become available. A method suggested by Olsen can serve as an indicator regarding the severity of possible sample truncation bias inherent in OLS estimates of recreation demand parameters. This paper illustrates Olsen's method applied to wilderness recreation demand models. The results illustrate that a simple transformation of parameter estimates assocated with OLS can be used to adjust for a sample truncation problem.