Mixed logit with characteristic variable


I'm trying to run mixed logit on my data.
A total of 400 individuals were asked to choose option in a stated preference experiment. Each respondent recieved 5 menues. From each menu, the respondent had to choose one out of three alternatives. The alternatives involved 4 attributes: (1) Supply - factor variable with 3 levels; (2) Dirt - factor variable with 3 levels; (3) Density - factor variable with 3 levels; (4) Tax - numeric variable with range 0-40, but actually with 5 optional levels: 0, 10, 20, 25 and 40.

The choice of an alternative depands on the 4 attributes and I also want to add characteristics of the respondent such as region (factor variable with 3 levels).

So I tried to arrange the data with the following columns: Option (A, B or C), Supply.A, Supply.B, Supply.C, Dirt.A, Dirt.B, Dirt.C, Density.A, Density.B, Density.C, Tax.A, Tax.B, Tax.c.

Then, I run:

Result <- mlogit.data(df, shape = "wide", choice = "Option", varying = c(14:25))
Result.mxl <- mlogit(Option ~ Supply + Dirt + Density + Tax | 0, Result)

and basically it works.

Now, I want to add the characteristic variable - Region.
I couldn't make it work, not with dummy variables or with "Region" variable.
My teacher said to me that the right thing to do is to use dummy, but I don't know how to make it work.

I would kindly use any help.

Thanks in advance,

I want to emphasize: The 'letter' of the option (A, B or C) is not important at all, it was randomly selected. The important things for me is the "willingness to pay" for the whole alternative.

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.