A behavioral urban recreation site choice model

Abstract

Abstract A multinomial logit site choice model (MNL) derived by utility maximization and based on reported travel distance and perceived site attributes was estimated for 21 forest‐related urban recreation sites in Chicago. Reported travel distance was the dominant explanatory variable. Perceived site attributes included stimulating/educational, quiet, safe, and forested. When estimated for subsets of the 21 sites, the coefficient for travel distance remained stable in sign, magnitude, and significance, while coefficients for “quiet” and “forested” changed in sign and significance. This supports the application of travel cost models at the urban scale, but suggests that site choice criteria may be sensitive to variables not included in the study.