Try uninstalling and reinstalling the program. This sometimes fixes the problem…
Ariel
From: cem-bounces@lists.gking.harvard.edu [mailto:cem-bounces@lists.gking.harvard.edu] On Behalf Of Katelyn Sack
Sent: Friday, June 29, 2012 8:24 PM
To: Matt Blackwell
Cc: cem@lists.gking.harvard.edu
Subject: Re: [cem] Two different Stata versions, two different cem problems
Hi,
Thanks. The mata lines of code did indeed add the cem library and fix the first error in Stata 10, but the second error remains in both versions.
The error is "cem_weights not found r(111);" and it follows attempts to run difference-in-differences regression after pre-processing the data with cem. Maybe there is a better syntax to run DND accounting for cem? I'm telling Stata "reg lblackproportion POLYEXAM YP YEAR [iweight=cem_weights]."
Many thanks,
Katelyn
On Fri, Jun 29, 2012 at 7:41 PM, Matt Blackwell <blackwel@fas.harvard.edu> wrote:
For the stata 10 code, one problem might be that Stata is
not finding the library. When you open up Stata, try running the
following line of code:
mata: mata mlib query
(with the colon and double-mata and everything) This should list some
libraries all starting with the letter l ("el"). One of the should be
"lcem." If it isn't, try running this line of code (which may take a
bit):
mata: mata mlib index
Hopefully now you should see "lcem." The program should work after
this. If this does not work, I have a few more ideas, but I think this
should resolve the issue.
For Stata 12, I think you have combined two separate commands. You want to run everything before "reg". Should be the same command as in Stata 10.
Please let me know if this does not help.
cheers,
matt.
~~~~~~~~~~~
Matthew Blackwell
PhD Candidate
Institute for Quantitative Social Science
Department of Government
Harvard University
On Friday, June 29, 2012 at 3:55 PM, Katelyn Sack wrote:
Hi Matt & all,
Thanks. My bad - that was a "replicating my error" error. Fixing it generates the following errors:
cem AGCYTYPE POPULATION region lblackproportion, tr(POLYEXAM)
<istmt>: 3499 cemStata() not found
r(3499);
(Stata 10 still seems to get stuck trying to find cem.)
And the same error as before in Stata 12:
cem AGCYTYPE POPULATION region lblackproportion, tr(POLYEXAM)
reg lblackproportion POLYEXAM YP YEAR [iweight=cem_weights]
cem_weights not found
r(111);
Any other ideas how to troubleshoot?
Thanks and best,
Katelyn
On Fri, Jun 29, 2012 at 6:38 PM, Matt Blackwell <blackwel@fas.harvard.edu> wrote:
Hi Katelyn,
Without taking too close a look, I think that the problem with the Stata 10 code is the space between "tr" and "(". It should be "tr(POLYEXAM)". That might also be the problem in Stata 12.
Cheers,
Matt.
On Friday, June 29, 2012 at 1:39 PM, Katelyn Sack wrote:
Dear list,
When I try using cem in Stata 10, I get the error "no variables defined in option treatment(), r(111);" even though my syntax looks fine. (My command was "cem AGCYTYPE POPULATION region lblackproportion, tr (POLYEXAM).") The website indicates it works in version 10.
When I try using it in Stata 12, I think the program gets a little farther. But I still get the error "cem_weights not found" after the commands:
cem AGCYTYPE POPULATION region lblackproportion, tr (POLYEXAM)
reg lblackproportion POLYEXAM YP YEAR [iweight=cem_weights]
Any ideas how to troubleshoot? My best guess is that I skipped important pre-cem steps and the program is protesting that sort of human error. Perhaps the matched N (672) is too small and so I should simulate more data to produce more matches?
Many thanks,
Katelyn
--
Katelyn Sack, Ph.D. candidate
University of Virginia
-
--
cem Mailing List, served by HUIT
Send messages: cem@lists.gking.harvard.edu
[un]subscribe Options: http://lists.gking.harvard.edu/?info=cem
More information on cem: http://gking.harvard.edu/cem
Cem mailing list
--
Katelyn Sack, Ph.D. candidate
University of Virginia
--
Katelyn Sack, Ph.D. candidate
University of Virginia