Travel cost (TC) data are often collected on-site, raising the question of on-site sampling bias. In addition, the motivations behind visiting a recreational area are diverse and differ across individuals, with the potential implications that preferences for trips are heterogenous. This is particularly true for recreational fishing areas that include multiple sites and within which multiple fishing activities are pursued. Comparing four single-site negative binomial models, this paper shows the importance of accounting for on-site sampling and heterogenous preferences simultaneously. The magnitude of the bias resulting from addressing these two issues separately is examined in terms of variation in consumer surplus, predicted mean trip and price elasticity. The paper further proposes an empirical strategy that addresses the issue of missing income bias. This bias occurs when respondents who do not report their income are significantly different in terms of their trip behavior from those with an average income level.

差旅成本 (TC) 数据通常是在现场收集的，这就提出了现场抽样偏差的问题。此外，访问休闲区背后的动机是多种多样的，并且因人而异，潜在的影响是对旅行的偏好是异质的。对于包括多个地点并在其中进行多种捕鱼活动的休闲捕鱼区尤其如此。比较四个单点负二项式模型，本文显示了同时考虑现场抽样和异质偏好的重要性。根据消费者剩余、预测平均出行和价格弹性的变化来检验因分别解决这两个问题而产生的偏差的大小。该论文进一步提出了一种解决收入偏差缺失问题的经验策略。当不报告收入的受访者在旅行行为方面与平均收入水平的受访者显着不同时，就会出现这种偏差。