AbstractThis article compares five methods for estimating travel cost recreation demand models with microdata. The models are distinguished by their treatment of selection effects that arise with on‐site surveys. The comparison considers adjusting for selection effects in a variety of ways, including single and double selection rule models. Both parameter and consumer surplus estimates were evaluated. The findings indicate that the treatment of selection effects alone was not important for this case. However, the choice of an estimator did lead to large variations in per trip consumer surplus estimates.