This is mik's Typepad Profile.
Join Typepad and start following mik's activity
Join Now!
Already a member? Sign In
Recent Activity
Hi John, Thanks for noting our paper. Actually, we say: "SP surveys apply various preference elicitation formats, which can be classified into one of the two categories: contingent valuation (CV) or discrete choice experiments (DCEs). In CV surveys, respondents indicate a specific number that expresses their valuation of the good or policy; respondents are usually asked to state their maximum WTP. An open-ended direct question and a payment card are the most commonly used CV formats. The former straightforwardly asks respondents about their WTP, whereas the latter provides respondents with a range of monetary values, from which they select the one representing their WTP. In contrast to CV, DCE surveys present respondents with a set of possible variants of a good or of a policy, and ask them to choose their most preferred option. Formats within this category differ with respect to the number of choice tasks and of possible response options. Table 1 briefly summarises commonly used DCE approaches." and we only talk about scope tests in relation to stated preference studies generally (not to differentiate between CVs and DCEs). I wish there was widely accepted "common nomenclature". Often CVs and DCEs are the same thing – like a single binary choice question that involves monetary payment. But not all CVs are DCEs (e.g., when they use non-discrete choice formats) and not all DCEs are CVs (e.g., when there is no valuation involved, like in some route choice studies). My impression is that the majority of people associate CVs with a single choice format (yes or no to a new policy at a cost, using payment card, etc.) and use DCEs to describe many choices or a choice from many alternatives for a single respondent. This is also the pressure I am getting from reviewers sometimes. So in the spirit of being understood better, I do not mind using CV or DCE (whichever is better received in a given context) when it is the same thing, really. The common nomenclature for modelling approaches is another thing to deal with. So that we do not have to say things like “discrete choice mixed (random parameters) conditional multinomial logit” to be clear what we mean when we use the MXL model. I wish guidelines dealt with things like that. Best, Mik
1 reply
You may want to compare with the graphs based on Econ Lit database or Google Scholar indexed papers: (pp 5-6)
1 reply
In discrete choice experiments respondents are sometimes asked to rank the alternatives, including the status quo / no choice / other (either by providing ranking numbers, consecutively choosing the best alternative of the remaining ones, or selecting best and worst). Interestingly, the preferences revealed at different stages are often not consistent (even if scale differences are controlled for).
1 reply
mik is now following The Typepad Team
Nov 15, 2016