Jump to main content.

Valuing Environmental Damages with Stated Preference Methods: New Approaches that Yield Demonstrably Valid Values for Non-Priced Environmental Goods

Quick Links

Research Funding

Metadata


Stated preference or Contingent Valuation studies are often criticized on the grounds that responses to hypothetical willingness to pay questions may indicate a higher willingness to pay than would be observed if respondents were asked to make an actual cash payment (this difference is called hypothetical bias). The objective of this project is to develop new stated preference institutions by extending and testing the robustness of two new designs for Contingent Valuation surveys which are referred to as the "Cheap Talk" and "Learning" design. These two designs have demonstrated in preliminary trials that they are capable of eliciting responses to hypothetical valuation questions that are indistinguishable from parallel valuation questions requiring actual payment.

Development of these surveys draws from lessons learned in experimental economics and psychology concerning the design of valuation institutions. The Cheap Talk design introduces, as a part of the willingness to pay question, an in-depth and candid description of the hypothetical bias encountered in Contingent Valuation studies. In the Learning design, respondents participate in a series of willingness to pay surveys: a hypothetical, then real (requiring actual cash payments) survey for one good, and then a final hypothetical survey for a different good. This design explores the extent to which subjects, having gone through one hypothetical-then-real valuation series will learn to anticipate a real question in responding to the hypothetical question for the second good, and thereby respond to the second hypothetical question as they would to a question involving real cash payments.

Laboratory experiments are conducted wherein surveys are conducted that involve either real or hypothetical payments to non-profit organizations affecting non-priced environmental goods. Evidence consistent with hypothetical bias has been observed in previous experiments using this methodology. The Cheap Talk and Learning Design for the hypothetical survey instrument are then conducted in an effort to eliminate any differences found between responses to the surveys involving real or hypothetical payments. To the extent these new designs can close the gap between responses to hypothetical surveys and surveys involving cash payments, in a demonstrable manner, then the credibility and acceptability of environmental assessments are enhanced. The results of these experiments are expected to provide guidelines for the conduct of stated preference studies that produce valid responses to hypothetical valuation questions.

Metadata

EPA/NSF ID:
R824710-010
Principal Investigators:
Cummings, Ronald G.
Osborne, Laura L.
Pate, James
Brewer, Paul
Technical Liaison:
Research Organization:
Georgia State University
Funding Agency/Program:
EPA/ORD/Valuation
Grant Year:
1995
Project Period:
November 1995 - October 1996
Cost to Funding Agency:
$113,856
Project Status Reports:
      For the year 1997
      Stated preference or Contingent Valuation (CV) studies are often criticized on the grounds that responses to hypothetical willingness to pay questions may result in higher willingness to pay statements than would be observed if respondents were asked to make actual cash payments (this difference is called hypothetical bias). The objective of this project is to develop new methods for eliciting stated preferences that are demonstrably effective in eliciting unbiased responses to hypothetical questions. We developed two new designs for CV surveys, referred to as “Cheap Talk” and “Learning,” and tested them for robustness. Results from over 1,500 surveys indicate that these two designs are capable of eliciting responses to hypothetical valuation questions that are indistinguishable from parallel valuation questions requiring actual payment.

      Development and testing of these new survey designs drew from lessons learned in experimental economics and psychology concerning the design of valuation institutions. The Cheap Talk design introduces as part of the willingness to pay question an in-depth and candid description of the hypothetical bias encountered in CV studies. In the Learning design, respondents participate in a series of willingness to pay surveys: a hypothetical survey followed by a real survey (requiring actual cash payments) for one good, and then a final hypothetical survey for a different good. This design explores the extent to which subjects, having completed one hypothetical-then-real valuation series, will learn to anticipate a real question in responding to the hypothetical question for the second good. We might then expect responses to the second hypothetical question to be the same as if the question involved real cash payments.

      We conducted laboratory experiments with groups of individuals to obtain either real or hypothetical payments to non-profit organizations affecting non-priced environmental goods. Our experiments found evidence consistent with hypothetical bias, supporting previously published evidence using this methodology. The Cheap Talk and Learning Design for the hypothetical survey instrument were conducted in an effort to eliminate any differences found between responses to the surveys involving either real or hypothetical payments. In each case, we found both designs to be effective in eliminating hypothetical bias. Specifically, responses to the hypothetical questions (HQ) using either the Cheap Talk design (CTD) or Learning design (LD) were statistically indistinguishable from responses to questions involving real cash payments (RQ). These results were robust to changes in the good, changes in the survey design, and changes in the experimental design.

      To the extent that these new designs can close the gap between responses to hypothetical surveys and surveys involving cash payments in a demonstrable manner, then the credibility and acceptability of environmental assessments are enhanced. The results from these experiments are expected to provide guidelines for the conduct of stated preference studies that produce valid responses to hypothetical valuation questions.

      Figure 1. Responses to Referenda Questions Involving a $10 Donation to a Public Good

Project Reports:
      Ronald G. Cummings and Laura L. Osborne, Valuing Environmental Damages with Stated Preference Methods: New Approaches that Yield Demonstrably Valid Values for Non-priced, Environmental Goods, Final Report to EPA, October 25, 1996, 33 pp. Download report - Type: Acrobat, Size: 379KB Content: Entire document

      Additional papers:
      Ronald G. Cummings and Laura L. Osborne, "Unbiased Value estimates for Environmental Goods: A Cheap Talk Design for the Contingent Valuation Method," no date. 33 pp.

      David Bjornstad, Ronald Cummings, and Laura Osborne, "A Learning Design for Contingent Valuation," no date, 24 pp.

Local Navigation


Jump to main content.