How to make cohort as the reference point

[Marcus Eklund]

Hi,

I'm using survey data and looking at the attitudes on a range of issues for different education groups. However, I would like to compare the respondent with others in his or her generation, not with the full sample. Do I then solely need to control for generation? Or how does it work?

e.g.;
reg attitudes education i.cohort

Best and happy for an answer,
Marcus

[Carlo Lazzaro]

Marcus:
1) if you're actually dealing with a survey, you should consider the -svy:- prefix (see -help svy estimation-) before -regress-;
2) your question is not that clear to me. An example of your dataset (with a more detailed description of what you're after) can help enormously. Thanks.