362 December 1995
Journal OfAgricultural and Resource Economics
demonstrated willingness to pay for management, marketing, and weather information. For
this reason, these individuals might be expected to be more willing to pay for Mesonet. On
the other hand, these subscribers already have access to a wide variety of weather satellite
information and forecasts.
Farming experience and higher education levels are hypothesized to increase willingness
to pay for weather information since farmers who have been farming longer would be more
aware of weather risk and more highly educated farmers would be more capable of applying
additional information. The expected difference between full-time and part-time farmers
with respect to willingness to pay is ambiguous. Part-time farmers are obviously less reliant
on farming income and are therefore less susceptible to weather risk. On the other hand, they
may have less time available for management and, consequently, might have a higher
willingness to pay for decision aids. Farmers with heavier debt loads are expected to have
a higher willingness to pay for weather information since they have less capacity to absorb
weather-related losses.
Survey Design
Data to estimate the empirical model of willingness to pay for Mesonet access was obtained
from a mail survey of Oklahoma producers. The Division of Agricultural Sciences and
Natural Resources at Oklahoma State University maintains a list of 1371 producers who
have agreed to respond to surveys on particular topics.1 Although weather data collection
was not one of the topics agreed to by the participating producers, a sample of 508 was
selected from the list to obtain representation from cotton, peanut, alfalfa, wheat, and
diversified-crop and livestock producers. In addition, a sample of 137 irrigated-crop pro-
ducers was selected from the 5,959 producers holding current irrigation permits. Since the
survey population consisted of producers who had previous links with Oklahoma State
University or producers with irrigation permits, it is likely biased toward larger, higher
income operations. However, this sample is representative of types of producers who would
be targeted to subscribe to Mesonet.
Survey design followed guidelines set out by Dillman. The survey questions were
designed to be understandable to agricultural producers. Researchers and extension special-
ists who routinely survey producers on other topics reviewed the survey format and wording.
The survey was pretested by county agents and a small group of producers. The survey
instruments were constructed in booklet form and mailed to recipients with a cover letter
describing the purpose of the survey and a return postage-paid envelope. Because of time
and money constraints, follow-up mailings of the survey form were not made to nonrespon-
dents. Surveys were received by 623 producers and surveys were returned by 175 producers.
Thus, not counting undeliverable surveys, the response rate was 28%.2
lThe list was compiled in 1989 using names of farmers and ranchers submitted by county extension staff. A preliminary
survey of the producers revealed the areas about which they would be willing to respond. Subjects of surveys regularly
conducted in the state using this list include cropland- and pasture-leasing rates and custom hire rates. The most recent other
survey effort using this list (cropland- and pasture-leasing rates) obtained a 32% response rate.
2Mitchell and Carson have observed that response rates for mail surveys used in contingent value studies are generally quite
low, with some response rates below 20% reported in the literature. The low response rates raise questions about how the
results ofanalyses using the survey data can be inteɪpreted and expanded to aggregate values. The response rate for this study
could have been increased with the follow-up mailings recommended by Dillman. However, time and financial constraints
precluded that follow-up. Although the response rate for this study equals or exceeds that reported in a number of studies, the
implications of the low response rate for interpretation and expansion of results are addressed in a later section.