Rodwan M Abouharb <[email protected]> is using the -cluster()- and
-vce(bootstrap)- options with -ologit- and is getting an error message stating
"repeated time values within panel":
> One final question. When trying to run the equation below STATA won't let me
> cluster by country 'cluster (ccode)' when bootstrapping. I have checked the
> integrity of the data-set and there are no repeated values. Is there some
> way to cluster by country?
> ologit tort ... , cluster (ccode) vce(boot)
> (running ologit on estimation sample)
>
> Bootstrap replications (50)
> ----+--- 1 ---+--- 2 ---+--- 3 ---+--- 4 ---+--- 5
> repeated time values within panel
> the most likely cause for this error is misspecifying the cluster(), idcluster(), or group() option
> r(451);
Rodwan's dataset has been -tsset- with a panel variable. When called with the
-cluster()- option, -bootstrap- attempts to make sure that the integrity of
the -tsset-ings are preserved within each bootstrap sample. Since -bootstrap-
samples clusters with replacement, Rodwan will need to specify some extra
-bootstrap- options in order to accomplish this.
If 'ccode' is the panel variable, then Rodwan needs to use -bootstrap-s
-idcluster()- option. For example,
. gen bootccode = ccode
. ologit tort ..., vce(bootstrap, cluster(bootccode) idcluster(ccode))
If 'ccode' is not the panel variable, then Rodwan will need to verify that the
panel variable is nested within 'ccode' and use -bootstrap-s -group()- and
-idcluster()- options. If we assume the panel variable is named
'panelid' then,
. gen bootccode = ccode
. ologit tort ..., vce(bootstrap, cluster(bootccode) idcluster(ccode) group(panelid))
--Jeff
[email protected]
*
* 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/
