The cluster option is used to correct for serial correlation within each cluster unit over time. The cluster option may only be necessary for panel data, and should give a more conservative of an estimate of your standard errors. If you're using cross-sectional data, robust standard errors should be fine.
Josh
----- Original Message -----
From: "Shang E. Ha" <[email protected]>
To: [email protected]
Sent: Monday, March 23, 2009 9:24:05 AM GMT -05:00 US/Canada Eastern
Subject: st: cluster robust SE options in 2SLS
Hi,
Can anyone look up the difference between these two standard error options; for
some reason they are producing slightly different standard errors.
ivregress 2sls Y (Z = X), vce(cluster unit)
ivreg Y (Z = X), robust cluster(unit)
Thanks!
Best,
Shang E. Ha
*
* 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/
