Hi, Hannah,

There are several ways that folks calculate WTP and a lot of controversy about how best to do it and even about WHETHER it's a good idea at all.

The simplest (not the best, but the simplest) way to do it is to do as you suggest - with an aggregate logit calculate $ per utility point by dividing the range in dollars by the range in utilities. With this in hand you can see the $ value of your other attributes (e.g. subtract the utility of the worst level of an attribute from the utility of the best level to get the range in utility attributable to that attribute. Then multiply that utility range by the $/utility value from above and you have a measure of WTP. This is one you read about commonly in the literature but it's probably not the best way to do it.

Perhaps better is to do the above, but using respondent-level (i.e. HB) utilities and get the WTP at the respondent level. Then use the median WTP so as to avoid outliers.

I should add that both of these methods tend to inflate WTP and to inflate it to a greater extent if your experiment DOESN'T have a \"none\" alternative respondents can use to opt out altogether.

Or you could measure WTP with simulations. Simulate the product without the attribute benefit you want to measure (i.e. the attribute is absent or is present at its less desirable state) and at the lowest price in your range. Then have that product compete with a version of the product that HAS the benefit. Adjust the price on the second product upwards until you get a 50/50 share split. This is the share-equalizing price and is yet another measure of WTP.

We have included a chapter on WTP in our book \"Becoming an Expert in Conjoint Analysis.\"

As well, you can read more about our cautions and suggestions for measuring WTP here: http://www.sawtoothsoftware.com/support/technical-papers/general-conjoint-analysis/assessing-the-monetary-value-of-attribute-levels-with-conjoint-analysis-warnings-and-suggestions-2001

How do I calculate the willingness to pay for my attribute levels in a CBC?