I have choice experiment data for estimating farmer's WTP for irrigation water. I asked farmers to choose from a set of 5 choice sets, each choice set containing 3 alternatives and a current status scenario. From each choice set they chose only 1 alternative. I am planning to use clogit to analyze my data. My dependent variable is a binary choice of whether they chose the alternative or not. If i do this each farmer will have 20 mutually exclusive observations from which 5 are selected as yes and the rest as no. Is this approach correct? or there is a better way of defining my dependent choice variable?
Hi Unity
Read here for a choice experiement example: http://www.sciencedirect.com/science/article/pii/S0261517711000604
You will have to (re)structure your dataset in a particular format and you will have to create some additional variabeles, such as for examplea variable that indicates the number of the choice set presented to a farmer, the number of the alternative in a particular choice set... I do not know what software you are using? In Stata, most of the (re)structuring can be done with the 'reshape' command.
A really simple introduction showing how to arrange the data, estimate the clogit model and estimate WTP is provided by:
Aizaki, H. and Nishimura, K. Design and analysis of choice experiments using R: A brief introduction
- Badges
- Science topic