WTP for interaction terms

[connie]

Hi Experts,

I have a question about WTP for interaction terms. Suppose my function is:
V=a1 (constant1)+a2 (constant2)+a3 (attribute A)+a4 (attribute B)+a5 (price attribute)+a6 (cancer×attribute A)+a7 (cancer×attribute A)

a6 and a7 are coefficients of interaction terms between cancer indicator and attribute A/B.

If I want to know WTP for attribute A among those with cancer, will the function be: MWTP= -(a3+a6 )/a5

I read a paper which calculate WTP for those with interaction terms, but I am not sure whether it is right way to do so.

Paper title: The effects of attribute non-attendance, simple validation questions, and their interactions on willingness to pay estimates for meat choice experiments

Many thanks,

Connie

[Michiel Bliemer]

Assuming that all attributes are continuous variables and not dummy coded, you can compute WTP for any nonlinear function by taking the first derivatives with respect to the attriibute of interest, and dividing it by the first derivative of the price attribute.

Assuming that the last interaction effect is cancer * attributeB and not cancer * attributeA, you get the following:

First derivative of attribute AttributeA = a3 + a6*cancer
First derivative of attribute Price = a5

WTP = - (a3 + a6*cancer)/a5

So the WTP depends on the cancer attribute, i.e. you have multiple WTP values.

If dummy coded attributes are involved, then instead of derivatives you will need to work with finite differences, but that is more complicated to explain.

Michiel

[connie]

Many thanks again, Michiel!