A model of recreation user decision-making: Predicting recreation campsite choice at state campgrounds in New York's Adirondack Park

This study confirmed a proposed site choice process by using conjoint analysis and then showed different site choice processes among campers of different characteristics by segmenting the campers encountered. An expanded Elimination-By-Attributes (EBA) model of campsite choice decision-making, including the relationships between site choice processes and personal or situational variables, was constructed for New York State Department of Environmental Conservation (NYSDEC) campground users in the Adirondack Park.; This study administered two years of consecutive mail surveys. The first pilot study mail survey was implemented during the year 2000, and compared the important attributes affecting campers' decisions in selecting their preferred campsites at two different types of the NYSDEC campgrounds. Surveys were sent to campers using six NYSDEC campgrounds (three less-developed campgrounds and three more-developed campgrounds). Of the 240 surveys mailed, 13 were undeliverable and 116 were returned (51.1% response rate). A weighted attribute approach (WAA) was conducted to better understand the relative importance of notable campsite attributes for NYSDEC campground users in the Adirondack Park when selecting their preferred campsites. Among the 17 campsite attributes, four attributes demonstrated statistically significant differences in terms of importance between the respondents from the less- and more-developed campgrounds. The weighted importance of these fundamental attributes was evaluated using the importance-performance analysis (IPA) technique.; The second mail survey was administered to develop the expanded EBA campsite choice model using the conjoint analysis technique during the year 2001. Of the 1,734 surveys mailed, 67 were undeliverable and 888 were returned (53.2% response rate). Additionally, in the second mail survey, preferences for the site attributes were investigated to compare potential differences between two representation modes in conjoint analysis. Results suggest that photo-realistic pictorial stimuli generated by a computer-aided design technology produced more accurate results than did verbal stimuli, with respect to the four issues: (1) information accuracy of the preference for attributes, (2) heterogeneity of respondents in their preference for attribute levels, (3) effectiveness of segmentation, and finally (4) fit of the model. However, the expected significant difference between two printing modes (i.e., color vs. black and white) in the photo-realistic pictorial stimuli was not found.