i am currently trying to analyse my discrete choice data, and i was hoping to get some advise from you.
i created 16 holdout choice sets using kuhfeld's %mktex macro. They were tagged with the constant "w". Afterwards, transformed the linear design into a choice design, using %mktkey and %mktroll.
The problem I am facing now is the following:
Instead of my original72 choice sets, I have 88. However, i cannot identify which of the 88 choice sets is a holdout choice set, and which not, as data sets do not contain the indicating "w" constant anymore. I already collected my data.
I doubt that I could just randomly extract 16 choice sets and use them as holdouts. So - how can I find out which 16 of the 88 choice sets are the holdouts?