ABSTRACT
This paper aims to examine whether or not differing sample compositions influence conditional willingness-to-pay (WTP) estimates for the same respondents, using a Monte Carlo evaluation. The results in this paper challenge the use of individual-specific WTP values. Approximately at least one out of five respondents are likely to show frequent extreme WTP values. Interval variables might give more stable tastes across varying sample compositions than dummy variables. Nonmarket valuation researchers, exploring preference heterogeneity, should be more cautious about potentially unstable conditional WTP estimates, particularly when respondents are regrouped with differing memberships.
Disclosure statement
No potential conflict of interest was reported by the author(s).