Date: Tue, 9 Sep 2003 15:26:28 +1000
Reply-To: Frank Milthorpe <Frank.Milthorpe@TRANSPORT.NSW.GOV.AU>
Sender: "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From: Frank Milthorpe <Frank.Milthorpe@TRANSPORT.NSW.GOV.AU>
Subject: Re: Weighting Data
Content-Type: text/plain; charset=US-ASCII
If you have a large number of population segments you can automate the
process using SPSS syntax.
Essentially you need an external file with the population control
totals for each segment. Using the AGGREGATE COMMAND you determine the
number of survey observations in each segment. The aggregated file is
then merged with the population control totals. On this file you can
calculate the weight for each segment. This is then matched back onto
your original data file. This is very useful if you have 100 segments;
less useful if you have only a few categories. This avoids the need to
The approach I have outlined above is the one that would adopt for
weighting survey observations where there are 100 survey observations
for population of say 2000000. You appear to be undertaking a different
weighting process to adjust the proportions. The same approach can still
I hope this helps.
Transport Modelling Manager
Transport Data Centre
Department of Infrastructure, Planning and Natural Resources
GPO Box 1620, Sydney NSW 2001
Level 17, 227 Elizabeth St, Sydney
Tel: +61 2 9268 2937
Fax: +61 2 9268 2853
>>> Kathleen Tobin-Flusser <Kathleen.Tobin-Flusser@marist.edu> 09/09/03
I currently use this syntax to weight data to certain population
COMPUTE secwt =
COMPUTE genwt =
COMPUTE wtfactor = secwt*genwt.
weight by wtfactor.
I would like to know what other options there are for weighting 1)
SPSS and 2) solutions that export data to other software to assign
using more complex wieghting (ie RIM) and then pulling the data back
SPSS for analysis with those weights.
Kathleen Tobin Flusser
Senior Research Associate
Marist College Institute for Public Opinion