182 July 1989
Western Journal of Agricultural Economics
Research Procedure
A systematic search of the available literature
was conducted in an effort to review as many
empirical studies as possible from 1968 to
1988. The selection process was designed to
fairly represent all the research on the topic in
the United States. Included were studies in
journals, chapters in books, unpublished re-
search reports, masters and doctoral theses,
research reports from private organizations and
government agencies, and conference papers.
In a number of cases, the authors were con-
tacted by telephone to clarify a methodological
question or to obtain the results of unpub-
lished studies. The overall effect of the selec-
tion process was to provide sufficient studies
to identify interesting trends and get a broad
flavor of the findings from both published and
unpublished studies.
The values reported here represent consum-
er surplus calculated by the authors of each
study from the demand functions they report-
ed. The net economic values are equivalent to
the dollar amount participants would be will-
ing to pay over and above their current ex-
penditures to ensure continued availability of
the opportunity to use recreation resources.
The review is limited to studies measuring the
on-site recreation-use benefits provided by a
natural resource of given quality. Many of the
studies also estimate the change in benefits with
changes in the quality of the resource, and in-
terested readers are referred to the detailed de-
scriptions of the original studies for estimates
(Walsh, Johnson, and McKean). Also, the val-
ues reported here do not include the public
benefits from preservation of resource quality
such as option values of future use and exis-
tence values to the general population of users
and nonusers (Walsh).
The standard unit of measurement is an ac-
tivity day, defined as one person on site for
any part of a calendar day. When values are
reported on any other basis than per activity
day, they are adjusted to this common unit.
For TCM demand functions, the appropriate
unit of analysis often is number of trips, but
most authors also report the results in terms
of value per activity day. If not, values per trip
are divided by the reported number of days
per trip. Similarly, annual values are divided
by the reported days of participation. House-
hold group values are divided by the number
of persons and days of participation per per-
son. Where the value of recreation activities
is reported for hypothetical quality changes,
the base value for current site quality is used.
There is a problem of defining recreation ac-
tivity days at some sites, notably reservoirs
with camping, swimming, boating, and fishing
on the same trip. In this case, the concept of
recreation use is based on the standard pro-
cedure of the U.S. Census in which an activity
is defined as primary use when it represents
over 5Oo∕o of total individual activity while at
the site.
Table 2 defines the explanatory variables in-
cluded in the equations. Most are conventional
measures and require little added explanation.
Nearly all of the variables are qualitative, in-
dicating that a particular treatment is either
present or absent. Of primary interest are the
three adjustments by Sorg and Loomis for
omission of travel time, the use of individual
observations, and in-state sample coverage
discussed earlier in this paper. Other impor-
tant determinants of demand are included to
hold constant their effects and to estimate the
partial effect of each of these variables and
other possible candidates for adjustment in
benefit estimates. The other variables are: rec-
reation activity; whether specialized or gen-
eral; site administration; quality; location; in-
flationary adjustment; method; open-ended,
iterative, or dichotomous choice question;
zonal, household production or hedonic price
approach. The variable list is constrained by
the availability of information, time, and bud-
get for this study. As a result, some potentially
important variables are omitted: direct travel
cost per mile, travel time cost per hour, income
and other specific socioeconomic variables,
sample size, functional form, and type of es-
timator used.
A quality variable is included to control for
specific characteristics of sites which vary
among recreation activities and expectations
of individual participants. Sufficient infor-
mation is available in the studies to apply a
rough index of site quality in three categories—
uniquely low, ordinary, and uniquely high—
based on a review of the physical and biolog-
ical information provided. A site administra-
tion variable is included to test the hypothesis
that Forest Service administered site benefits
are not significantly different from other public
and private sites. A mixed public-private site
variable tests the hypothesis that household
surveys are more effective than on-site studies,