Excess Zeros in Count Models for Recreational Trips

Latent class count models of total visitation demand: days out hiking in the eastern Alps

“…Count models allocate a probability to observed zero count, which is often insufficient to account for the vast numbers of zero demand points observed in empirical data (Gurmu andTrivedi 1996, Cameron andTrivedi 1998). So we inflate the zero count probability by a logit inflator, which is common across sites and across groups.…”

‘Adopt a Hypothetical Pup’: A Count Data Approach to the Valuation of Wildlife

“…The latter would be ÔtripsÕ or ÔvisitsÕ in the recreation demand literature and Ôproblem coyotes sparedÕ in the analysis below. Further details on this methodology can be found in Gurmu and Trivedi (1996) and Trivedi (1998 and.…”

“…Count data distributions typically exhibit a concentration of values on a few small discrete values, skewness to the left, and intrinsic heteroskedasticity with variance increasing with the mean (Cameron and Trivedi 2001). This methodology has been widely employed in the area of non-market valuation within the context of recreational demand analyses, since these often seek to place a value on natural resources such as national parks by modelling the number of trips to a recreational site (Englin and Shonkwiler 1995;Gurmu and Trivedi 1996). Hellerstein and Mendelsohn (1993) show that measures of consumersÕ surplus or total WTP can be calculated from the results of count data estimations.…”

“…The results of these hurdle specifications (not reported in Table II but available upon request) show no substantial differences relative to those reported in terms of the coefficient of spobid or goodness of fit. A hurdle model implies that once a respondent passes the hurdle sponsored will be strictly positive, hence the use of the truncated count models (Gurmu and Trivedi 1996). This is reasonable in some settings, 19 but not in the one considered here, since some of the potential supporters of coyote conservation might still choose a corner solution.…”

“…Again, no examples are available in the context of contingent valuation, but examples from the area of recreation demand analysis include Gurmu and Trivedi (1996) and Shonkwiler and Shaw (1996). The following subsections describe the theoretical features of these methods.…”