The purpose of this paper is to test for question-ordering effect with the contingent valuation method. We use data from a survey focusing on the conservation values of three waterfowl's wildlife sanctuaries in Taiwan. Respondents are separated into groups of wildlife sanctuary closely related and less related. Empirical evidence shows that, for less-related group or non-residents/non-participants, willingness to pay is sensitive to the sequencing of Lan-Yang waterfowl area. Similar result is found for the closely-related group when evaluating Wu-Wei area. No effect is identified for all other question ordering. Significant differences between the two groups, however, suggest that respondents' experience and residential district may affect their valuations toward wildlife sanctuaries. Dissemination of the importance of wildlife sanctuaries is therefore suggested as an essential instrument to promote conservation policy in Taiwan.