Using theoretical derivations, it is shown that collecting data on individuals' visitation rates to a recreation site by each of these methods: (1) on-site sampling of visits; (2) sampling individuals surrounding the recreations site; and (3) sampling license holders, results in three unique heteroskedacity problems. A different weighted least squares approach is offered in each case when estimating the visits per capita-travel cost relationship in zonal travel cost models. Furthermore, to the extent that individuals within an origin zone face different prices, there is an inherent aggregation bias when estimating consumer surplus.