Re: st: regression models with small number of clusters

[Roger Newson <roger.newson@kcl.ac.uk>]

Krishna doesn't tell us which "factors" are being assessed as predictors of 
patient satisfaction. However, I think I would probably use a fixed-effects 
model here. It seems more realistic to calculate clinic-adjusted mean 
satisfaction differences (or geometric mean satisfaction ratios) for the 
population of patients in these 12 clinics than to attempt to make 
inferences about the population of clinics at large, based on such a small 
sample of clinics.

Hope this helps.

Roger


At 22:06 05/01/2005, Krishna wrote:

> Dear All:
>
> I had previously posted this and would like to thank
> Roger Newson for his reply. Here is my problem in more
> detail.
>
> I am trying to fit a regression model to data which
> was generated from cluster sampling. There are 12
> clusters (clinics) and the  average cluster size is
> around 30. The outcome of interest is a continuous
> variable (patient satisfaction) and I am interested in
> the factors which affect it. Since the data is
> clustered and it is potentially correlated, I need to
> adjusted for this in my regression model. What would
> be the appropriate regression technique/model to do
> this ? I understand that models such as multiple
> regression with sandwich estimator (regress with
> cluster option), and GEE are not appropriate as they
> require a large number of clusters to provide
> consistent estimates. Would a Random Effects (xtreg,
> re) or Fixed Effects (xtreg, fe) model be appropriate
> ? Are there any others you would recomend ? Many
> thanks for all your help..
>
> Best regards
> Krishna
> Dept. of International Health
> Johns Hopkins University

--
Roger Newson
Lecturer in Medical Statistics
Department of Public Health Sciences
King's College London
5th Floor, Capital House
42 Weston Street
London SE1 3QD
United Kingdom

Tel: 020 7848 6648 International +44 20 7848 6648
Fax: 020 7848 6620 International +44 20 7848 6620
  or 020 7848 6605 International +44 20 7848 6605
Email: roger.newson@kcl.ac.uk
Website: http://phs.kcl.ac.uk/rogernewson/

Opinions expressed are those of the author, not the institution.

*
*   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/