[Nick Cox]

Posts are crossing here but the main point in #13 still applies to posts that I can see at the moment. If you have a two-way breakdown tsset needs to be modified accordingly. Also if the tsset fails, then there is no point in trying to carry on. I can't see any reference to wave in your code; possibly that refers to a previous tsset of your dataset;

[Claire Wright]

Yes, since I collapsed my dataset there is no wave variable anymore and I think that may be a possible problem. tsset agegroup year fails as agegroup is a string variable. When I do carry over the wave varaible to use tsset wave year I get something that looks like this:

[Claire Wright]

Just recoded my age variable and solved it! Thank you for you help...