With -tabulate-, weights are assumed to be frequency
weights unless otherwise indicated. Your weights
sound like analytic weights.
. by country: tab illness [aw=weight01]
With -summarize- weights are assumed to be analytic
weights unless otherwise indicated.
Also, for the purposes of the tab command, he might want [iw=weight01]. It
depends on what the goal is. This would give you the estimated number of
households in each country that have a particular characteristic, e.g.
there are are an estimated 400,000 people in the UK with illness 1. You
might get Ns in the millions, even though the sample is only a few
thousand. With aweights, the weights will be rescaled to reflect the
sample Ns. -help weights- provides more info, even better is section 23.16
of the User's Guide. Stata's weighting options are more extensive than
what I was used to so reading through that section was very helpful to me.
