Welfare change estimates obtained from discrete-response contingent valuati
on experiments normally assume a particular distribution of,willingness-to-
pay (WTP). Using conventional microeconomic theory, we derive upper and low
er bounds on such estimates. These bounds are interpreted statistically in
terms of non-parametric estimators of mean WTP and its variance. Two contin
gent valuation surveys illustrate the proposed bounds, and the performances
of the variance estimators are scrutinized with a Monte Carlo simulation.
All calculations can be made by hand, simplifying communication among those
involved in interpreting results from contingent valuation studies using d
iscrete-response data. (JEL Q26).