I have a question really close to the former one.
I have a full list of 22 items and 4 "target populations".
One of the groups has a design with the full list of items. For the 3 other ones, I have different designs, consisting of the same list with 2 to 3 items excluded.
How can I handle it in HB? (I use the version 5.2.8 and specified user specified and dummy coding).
I thought I could use the same cho file with all respondents - including choices and design - and then specify covariates (I would like to use the information common across my 4 groups), But the utilities computed don't seem to be as described above :
For the people not having seen the item "a draw near the mean of those having seen it".
And then, how should I rescale the scores , just excluding the items is not enough, is it?
Am I missing something?
FYI, the link provided doesn't work anymore.
Many thanks in advance for your answer.