For responses to a hypothetical question in which a respondent reports a high degree of
uncertainty, these studies generally find that recoding uncertain “yes” responses as “no”
yields WTP estimates that are a good approximation of actual payments. This recoding
scheme has effectively calibrated hypothetical responses in a number of studies, but the
cut-off point at which this technique work varies. Champ et al. (1997) found 10 as an
appropriate cut-off point; Champ and Bishop 2001 used 8, and Vossler et al. 2003 used 7
(Murphy et al. 2004). In addition, Samnaliev et al. (2003) found that certainty calibration
reduced the goodness of fit (of the Logit model) and increased the variance of responses.
Certainty scale calibration is another technique used to correct for hypothetical bias in
dichotomous choice (DC) CV questions where respondents may generally be less certain
of their answer to the offer price as compared with an open-ended or polychotomous
choice format (Ready et al. 1995). In the certainty scale calibration, the WTP question is
followed by a question that asks how sure respondents are about their response to the
valuation question. The effectiveness of this method has been established by comparing
hypothetical payments with actual donations (Champ et al. 2001; Polasky et al. 1996) as
cited by Samnaliev et al (2003).