5) Socio-Demographic Questions: We end the survey with a series of socio-demographic questions
including age, gender, income, education, race, etc.
Table 2 presents the full set of variables, with definitions and descriptive statistics. We turn now to the
econometric analysis of the survey data.
5. Econometric Analysis
We break the econometric analysis into three sections. In section 5.1, we present the model to
estimate overall WTP for the NRP. Section 5.2 defines the model used to estimate the degree of
hypothetical bias in terms of increased probability that a stated-preference household will purchase the
pass at $65. Finally, in section 5.3, we describe how to use the two models to translate the probability
measure into a dollar-denominated measure of hypothetical bias. The hypothetical bias scaling factor can
be used to calibrate the WTP estimates to the actual purchasing decisions of households.
5.1 Overall WTP Model
Our first econometric model estimates WTP for the NRP, which in turn can be used to forecast pass
and gate revenue at various NRP fee levels. We use an interval regression model that follows directly
from the double-bounded dichotomous-choice (DBDC) survey design described above (Hanneman et al.,
1991; Herriges and Shogren, 1996). An interval regression is an ordered probit model with variable and
known cut points (Woolridge, 2002).
Start by writing the empirical model in terms of a household’s maximum WTP for the NRP, which is
indirectly derived from the utility of visiting federal recreation sites:
WTPi= exp( X'β + εi ), (3)
where WTPi is the latent willingness to pay for the NRP; Xi is a vector of explanatory variables; β is a
vector of coefficients; εi is a mean-zero Gaussian error term with variance σ2; and i = 1,..,N indexes
10