Discussion of Morey and Waldmanʹs “Measurement Error in Recreation Demand Models”

Morey and Waldman have proposed a method for handling measurement error in site attributes, such as catch rates in models of anglersʹ choice of fishing sites. We discuss the properties of this method and compare these properties with those of the standard estimation procedure. We show that (i) both methods are biased and (ii) the standard procedure is consistent whenever Morey and Waldmanʹs procedure is consistent, as well as under more general conditions. In particular, Morey and Waldmanʹs procedure is inconsistent if any site attributes are omitted from the model, whereas the standard procedure is consistent in the presence of omitted site attributes, as long as the impact of these attributes on the dependent variable is uncorrelated with included variables (a condition shared by all regression models).