In a split sample design, we examine the impact of federal funding availability on Willingness to Pay (WTP) for watershed management program attributes and tradeoffs in a choice experiment. We also evaluate how presenting respondents with different sets of choice attributes, in alternative survey designs, affects the estimation of preference functions. We also compared preferences for watershed management attributes across sub-watersheds. These issues were evaluated using the Blackstone River Watershed Public Preference Survey, in Rhode Island, USA. Our results indicate that neither federal support nor geographically distinct sub-watersheds had significant impact on tradeoffs elicited among management attributes. However, survey design may induce respondents to show distinct preferences for watershed management. We examined these issues using a multinomial logit model in comparison with a Latent Class Model (LCM) to account for heterogeneity in preferences.