Testing for differences in Willingness To Pay (WTP) values

[Hector Tavarez]

Hello everyone,

I am new in the forum. Hope one of you with expertise in choice experiments method can help me today.

My name is Hector Tavarez. I am using Stata 14 for data analysis. I developed a questionnaire using choice experiments to estimate how much respondents are willing to pay for water service improvements. I am interested in comparing wilingness to pay (WTP) values from three different survey versions, which differ only in the way attributes in choice experiment are framed. I ran a conditional logit model (and other models such as the random parameter logit model) for estimation purposes. I used the "wtp" command in Stata developed by Arne Risa Hole to obtain WTP values and confidence intervals using the Krisnky and Robb procedure for each attribute level in my choice experiment. Does anyone know how to run the complete combinatorial approach suggested by poe et al. (2005)?

I used these commands:

clogit depvar indepvar1 indepvar2 indepvar3 cost, group(choicesets) ..... (I did this for the three survey versions)
wtp cost indepvar1 indepvar2 indepvar3

where depvar is the dependent variable (choice), indepvars are the independent variables (attributes used in the choice experiment), cost is the cost of each option presented in the choice experiment, and choicesets is a variable distinguishing each choice table presented to participants.

I would like to test for differences in WTP values for indepvar1 indepvar2 indepvar3 across survey versions.

Maybe if I can save the simulation results from the "wtp" command I will be able to do the rest, but the command does not have this option.

Reference:

Poe, G.L., K.L. Giraud, and J.B. Loomis. 2005. "Computational Methods for Measuring the Difference of Empirical Distributions."
American Journal of Agricultural Economics 87:353-365.

[Hilary Polis]

Hi Hector,

My name is Hilary Polis and I am having the exact same issue as you. I developed a survey to understand public willingness to pay for tidal energy R&D. I was asked to conduct a Poe (2005) test on my WTP values and I can't figure out how to conduct the Poe test. I see that the save command was added to the WTPKR function, but I was wondering if you figured out how to run the complete combinatorial approach and if so, what code did you use?

Thanks,

-Hilary

[Ersilia Schuster]

Dear all,

I am having exactly the same problem. I want to compare WTP values from a Stated and Revealed preference model. Did you manage to find the codes in the meantime?
Thank you very much!

[Carlo Lazzaro]

Dear All,
to the the best of my knowledge, Arne Risa Hole devoted most part of his research on WTP, creating user-written programmes for Stata (https://www.sheffield.ac.uk/economics/people/hole/stata).

[Hannah Sack]

Hi all did any of you figure out the solution? I want to test for differences in WTP for independent variables from two surveys (all independent variables are the same across treatments) but I have the data for the two surveys in two different datasets.
Will i have to merge the two datasets, run one clogit model across all respondents in total to see the difference in WTP ??

Regards
Hannah Sack