Contingent valuation (CV) techniques are commonly used to value the potential effects of a policy change when market-based valuation of those effects is not possible.  In order to elicit a valid valuation of the effects of a policy change, studies employing CV techniques require analysts to provide respondents with a clear and thorough description of all likely effects.  Such a task can be quite difficult when a policy change is likely to have effects that span multiple dimensions and the magnitude—or even direction—of some or all of the effects is uncertain.  In this paper we analyze how the expression of uncertainty in the portrayal of the potential effects of a policy change can affect the valuations elicited from survey respondents in CV studies.

Using operations of the Glen Canyon Dam as our empirical application, we fielded an original survey of over 2000 respondents and describe how changing the flow of water through the dam is projected to affect electricity production at the dam, as well as natural and cultural resources in the surrounding area.  We then use standard CV procedures to elicit respondents’ valuations of the projected effects of the policy change.  Following this elicitation, we randomly assign respondents to one of three groups: 1) A control group; 2) A moderate uncertainty treatment group; and 3) A high uncertainty treatment group.  Members of the control group receive a reminder that the additional money they would pay in order to obtain the projected effects of the policy change would be rendered unavailable for other purposes, and are then asked if they would like to change their expressed valuation.  Members of the moderate uncertainty group receive the same budget reminder as the control group as well as an additional statement that there is some disagreement among experts regarding the projected effects and that the realized effects could be somewhat larger or smaller than the projections described when eliciting their initial valuation.  The treatment for the high uncertainty group is similar, but members are told that there is substantial disagreement among experts regarding the projected effects and that the realized effects could be a great deal larger or smaller than the projections presented initially.  Both the moderate and high uncertainty groups are then asked if they would like to change their expressed valuation.

By comparing the proportion of each group that reports wishing to change their valuation—and the manner in which they do so—our analysis will provide insight into how uncertainty in the projected effects of a policy change relate to the valuations that individuals place on them.  The results have several implications for both research and policy and we close the paper with a discussion of these issues.