There is increasing recognition that statistical estimates of the aver
age willingness to pay for environmental programs may be biased by the
failure of respondents, or of the statistical methodology, to conside
r the effect of substitute programs and goods. We design and implement
a CVM survey that estimates the impact of substitution effects on the
willingness to pay for an environmental program. We find a statistica
lly significant impact and suggest that the set of potential substitut
es is very large.