N is always unweighted in SAS, as far as I know. SUMWGT would be the sum of the weights (effectively the weighted N). SUM would be the weighted sum [so if you want to base it on the value of Q01]. I don't know about COLPCTN unfortunately, whether you can get that weighted or not, it's not a statistic I've used; perhaps you can use COLPCTSUM in some manner? That should be weighted, I imagine.

-Joe

On Mon, Mar 2, 2009 at 8:10 AM, Erwan LE DU <dotker@gmail.com> wrote:

> Hi all, > > I have a continuous weight variable called wgt1. I want to use it in a > tabulate procedure to weigh categorical survey data. The problem is > that SAS doesn't take wgt1 into account when performing the > calculations. The output appears as unweighted...(I checked using > SPSS). I know I can do that with proc freq but I must keep the stub > and banner layout provided by proc tabulate. > > Here is a sample code : > > proc tabulate data=TEMPSAS (keep=b1 b2 b3 wgt1 Q01); > class b1 b2 b3 Q01 ; > weight wgt1; > table N='Base' (colpctn='Column Percent Total') (colpctn='Column > Percent')*(Q01) N*(Q01), all='Total' b1=b1 b2=b2 b2=b2*b3=b3; > run; > > Your help would be greatly appreciated. > > Regards, > > Erwan >