Statalist The Stata Listserver


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: "testing" a cluster analysis


From   Ronán Conroy <rconroy@rcsi.ie>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: "testing" a cluster analysis
Date   Wed, 7 Feb 2007 14:56:41 +0000

On 7 Feabh 2007, at 14:45, Adam Seth Litwin wrote:

Hi, Ronán. Your point is well-taken, and conventional hypothesis test might not be the best tool. I have already analyzed the data more formally with OLS, but one of my advisors suggested I see how the observations cluster with respect to these seven binary indicators.
If they cluster well, then the clusters will make sense. Don't forget that clustering is an exploratory technique—it doesn't have a dependent variable. Try to characterise the people in your clusters and see does it make scientific sense. If the clusters are useful, you can come up with short descriptive labels for the type of person they contain. If not, I wonder if the process tells you anything.

Remember that cluster structures need not be simple. A cluster may be made up of component clusters. It's worth exploring solutions with smaller numbers of clusters to see if this happens. I'm working on a similar problem (though clustering variables, not observations) where there are clear clusters-within-clusters which make scientific sense.

And above all, harass your supervisor. These people can make airy, partly-thought out suggestions that cause their students endless work!

=========
Ronán Conroy
Royal College of Surgeons in Ireland
rconroy@rcsi.ie
+353 (0) 1 402 2431
+353 (0) 87 799 97 95
http://www.flickr.com/photos/ronanconroy




*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/




© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index