Michael I. Lichter <[email protected]>:
There is a df adjustment (see below) I neglected on the grounds that
it should make no visible difference assuming you have a reasonable
sample size--which you don't. With 4 clusters, I cannot see how you
can use the cluster-robust estimator--its good properties are
increasing in the number of clusters and 4 is not near infinity. See
Rogers 1993 (http://www.stata.com/support/faqs/stat/stb13_rogers.pdf)
on how the cluster-robust estimator is biased downward and the bias is
larger for smaller numbers of clusters or unbalanced clusters. You
cannot even use the bootstrap, since you suspect clustering, and you
cannot really resample effectively from a sample of 4. I think you
need at least 20 clusters to use the cluster-robust estimator with any
confidence.
clear
webuse nhanes2
egen c=group(strata psu)
keep if inrange(c,1,4)
svyset c
qui svy: reg bpsystol height weight
mat svy=e(V)
qui reg bpsystol height weight, cl(c)
mat ols=e(V)
mat c_ols=ols*(564/565)^2
mat li ols
mat li svy
mat li c_ols
On Tue, Sep 9, 2008 at 4:32 PM, Michael I. Lichter <[email protected]> wrote:
> This is a small trial of an intervention designed to promote guideline-based
> diagnosis and treatment of patients with chronic kidney disease (CKD). Four
> medical practices were selected and two each were randomly assigned to
> control and intervention. (Yes, I know that it is not recommended to do CRT
> with fewer than 5 clusters per arm.) Primary indicators include glomerular
> filtration rate (GFR) and whether or not patients with substandard GFR were
> diagnosed during the trial period has having CKD. We predict stable or
> rising GFR in intervention practices compared to falling GFR in control
> practices, and higher rates of physician-diagnosed CKD in intervention
> practices compared to control practices. The universe of patients is those
> with substandard GFR levels prior to the intervention.
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
