What is "case" in -cmlogit-?

[Hadji Jalotjot]

My choice experiment data looks like this...The Stata manual on -cmlogit- states

The command requires multiple observations for each case (representing one individual or decision maker), where each observation represents an alternative that may be chosen.

I basically satisfied the requirement except I am not sure which should I consider the "case"...Each respondent was presented 6 choice sets...Which should I use as the "case" - the decision make/respondent or the choice set?

[Dimitriy V. Masterov]

It sounds like you may have panel choice data here. Does a choice set correspond to an occasion to make a choice? That is, you observe each person choose one out of three things on six occasions. If that is correct, then you have six cases for each respondent.

[Hadji Jalotjot]

Exactly Dimitriy. So this is considered panel data.

There is a lot of mixed logit commands in Stata...I might try Arne Hole's mixlogit as it does not require to specify an alternative variable (altvar)