I have a willingness to pay (wtp) situation at my hands and I am wondering if I can use the Sawtooth tools to model my idea. Here briefly the problem:
The product manager is wondering a) what the overall wtp for the current product in the market is and b) which features contribute how much to the wtp so she can evaluate the effect of scratching some of these features. I picture a linear price function for the whole product (i.e. one parameter). I could then implicitely find out the wtp associated with one unit of utility and via the rule of proportion find wtp´s for each feature. This reasoning might be controversial as units of utility are rather a measure of change in probability but I've seen this approach in publications before so I assume it is legitimate. (any comments on this?)
Here is my actual question:
To estimate the utility of a level being included in a product I need to include the base level (not available) for comparision. Does CBC treat the (not available) level as any other level and estimates some level of utility for it? If it did then - according to the above idea with the rule of proportion - I would automatically end up with a certain wtp for something not being there. So I would like to surpress or restrict certain parameters to be 0.
Is that possible with the regular CBC/HB estimation tool? If not, what would be feasible alternatives without having to buy more software?