Date: Tue, 22 Jan 2002 12:27:15 -0500
Reply-To: manon girard <mansof@VIDEOTRON.CA>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: manon girard <mansof@VIDEOTRON.CA>
Subject: weighted data and SAS (long)
Content-Type: text/plain; charset="us-ascii"; format=flowed
I encountered some problems with weighted data. I usually work with
populational data which contains weights and "design effects" (or variance
inflation due to sampling design - like stratification). In fact, each
sampled individual in the dataset is given a weight such that this guy
represents X persons in the total targeted population.
If I inflate the dataset (change WEIGHT for the FREQ subcommand) I will
produce statistical tests that will always reject the null hypothesis
(because the n is too large).
Usually I take sampling weights to readjust the sample size. By this I mean
to get n the same number of my sample size but individual may represents
less than 1 person (my weights would contain a fraction of individual).
Recently, I tried to make multiple comparisons with SAS-GLM (Anova) and all
the homogeneity of variance tests (underlying assumption of multiple
comparisons in ANOVA) (e.g. : HOVTEST=BF) could not be used with weights.
What can I do ? Bootstrapping ?
I have got the same problem with NPAR1WAY ! No weights !
I contacted SAS Institute and they could not help me.
And I did not test every procedures of SAS, yet. So what should I do when I
need weighted data (which is 100% of the data I work with) ? I thought of
re-program everything but I was not sure if it is ok to do so ?
Is there any reference that could help me ?
Is there someone who has the same problems as I have and how do you figure
it out ?