Omitting substitute prices from a travel cost model is shown to cause a significant bias in consumer surplus estimates. Three sets of travel cost models are developed from a common data base representing 60,000 day‐users of U.S. Army Corps of Engineer reservoirs in Kansas and Missouri. The firct set of models omitted substitute prices; the latter two sets included them. An analysis of variance test showed that consumer surplus estimates from the first set of models were significantly higher than the other two (F = 26.2 with 2, 20 degrees of freedom). The theoretical and practical implications of these findings are discussed.