Recreation planners and managers realize the importance of different tastes and preferences among recreationists. This fact is explicit in recreational planning efforts and in the theories developed to understand recreational behaviors. While market segmentation approaches steeped in behavioral theory provide a rich source of information to planners and managers, we argue that such exogenous market segmentation approaches alone are insufficient to understand the degree of variability in tastes and preferences among recreationists. In this paper we employ a random parameters logit model to account for both the extent and sources of heterogeneity in preferences among recreationists. We illustrate the random parameters logit choice model with an application to the stated choices for hunting sites by moose hunters from northwestern Ontario. The application shows that the random parameters logit represents a considerable improvement over multinomial logit models based on market segmentation approaches.
Acknowledgements
We are grateful for a reviewer's suggestion to estimate individual level parameters and to examine reasons for the source of variability among these parameter estimates. We also thank the Ontario Ministry of Natural Resources for funding the research.
Notes
1The scaling is relative since one systematic parameter estimate must be fixed for identification purposes.
2We used the mixed logit estimation for panel data using maximum simulated likelihood available on Train's web site at http://elsa.berkeley.edu/∼train/software.html.
* 0.10 > p ≥ 0.05.
** 0.05 > p ≥ 0.01
*** p < 0.01
* 0.10 > p ≥ 0.05.
** 0.05 > p ≥ 0.01
*** p < 0.01
*0.10 > p ≥ 0.05.
**0.05 > p ≥ 0.01
***p < 0.01