An Experimental Test of Response Consistency in Cvm

Using experimental methods, we study the reliability of responses received from CVM. We compare declared willingness to pay (WTP) for an improvement in the ecological performance of a piece of furniture with behavior in an experimental environment which replicates the hypothetical scenario of the questionnaire. Results show no signifficant median differences between declared (in the questionnaire) and revealed (in the experiment) WTP. However, evidence of overdeclaration (underdeclaration) is obtained for subjects whose declared WTP is high (low)