Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: st: Estimating the (possibly negative) intracluster correlation
From
Steve Samuels <[email protected]>
To
[email protected]
Subject
Re: st: Estimating the (possibly negative) intracluster correlation
Date
Sun, 5 Sep 2010 22:20:34 -0400
Here's a do-file to estimate a possibly negative ICC; it's based on
the definition of the ICC as the correlation of all pairs of
observations within groups (Cochran, Sampling Techniques, Third Ed,
1977, Wiley page 209)
Steve
Steven J. Samuels
[email protected]
18 Cantine's Island
Saugerties NY 12477
USA
Voice: 845-246-0774
Fax: 206-202-4783
**************************CODE BEGINS**************************
sysuse auto, clear
replace rep78=1 in 1
tab rep78
keep if rep78 <3
sort rep78
gen id1 = _n
loneway mpg rep78
keep id1 mpg rep78
tempfile t1
save `t1'
rename mpg mpgx
rename id1 id2
joinby rep78 using `t1'
sort id1 id2 // use for check
drop if id1 == id2
list id1 id2 if rep78==1 //check
corr mpg mpgx
***************************CODE ENDS***************************
On Sun, Sep 5, 2010 at 8:44 PM, Bert Jung <[email protected]> wrote:
> Dear Statalisters,
>
> I am interested in estimating an intracluster correlation, if possible
> conditional on several covariates, that could be negative. I wondered
> if anyone knows a command or strategy to do this?
>
> As background: I am estimating a simple OLS -regress- model and find
> that the default (unclustered, not robust) standard errors are
> *higher* than the clustered s.e. A potential cause is model
> misspecification that can induce negative intracluster correlation, as
> discussed in http://www.stata.com/support/faqs/stat/cluster.html. I
> hope to diagnose the problem, starting with a closer look at the
> intracluster correlation. (I will also work on a better
> specification, of course.)
>
> -loneway- calculates the intraclass correlation as ratio of two
> variances, hence constraining the correlation to be >=0. I am looking
> for an alternative way that also allows me to control for covariates.
>
> Thanks in advance for any pointers,
> Bert
> *
> * 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/
>
*
* 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/