30 Mar
2009
30 Mar
'09
10:37 a.m.
Hello,
I want to analyse the effect of several independent variables on the democratization
process of almost every country in the world. At the same time I have 30 years for every
country.
1) Is there a possibility to do a fixed effects regression with my imputed datasets with
Clarify (xtreg) or do I have to use Zelig?
2) If I have to use Zelig - what do I have to do or where can I find Information?
3) After this regression I want to use the synthetic control method. I donŽt think that I
can use the Imputed data sets for this. Is there a way to put the imputed datasets
together in one dataset?
Can anyone help me figure out a way to get around this problems?
Thanks in advance for any help you can give.
Sincerely,
Laeticia Klein
_______________________________________________________________________
DSL zum Nulltarif + 20 Euro Extraprämie bei Online-Bestellung über die
DSL Freundschaftswerbung! http://dsl.web.de/?ac=OM.AD.AD008K15279B7069a
-
Amelia mailing list served by Harvard-MIT Data Center
[Un]Subscribe/View Archive: http://lists.gking.harvard.edu/?info=amelia
More info about Amelia: http://gking.harvard.edu/amelia
10:53 a.m.
Laeticia,
I did fixed/between/random effects regression for my own project, and
I had to hand-write the code in R, as Stata's commands for MI data
didn't do xtreg very well. As I wrote my quant chapter, nine months or
so ago, Zelig only included GEE methods for panel data, which also
didn't do what I wanted. So I had to write my own code in R.
The code's not too hard (conceptually) to cook up, if you follow the
algorithm in Gary's original article on MI. Also, I can dig out my
code and share it if you decide you need it.
Best,
-Nathan
----------
Nathan A. Paxton
Ph.D. Candidate
Dept. of Government, Harvard University
Resident Tutor
John Winthrop House, Harvard University
napaxton AT fas DOT harvard DOT edu
http://www.fas.harvard.edu/~napaxton
========================================================
If every professor who backed a lunatic politician were to be sacked,
half the interesting minds in academia would be lost.
- The Economist, 5 Jan 2002
A morning without coffee is like something without something else.
========================================================
On 30 Mar 2009, at 10:37 AM, Laeticia Klein wrote:
Hello,
I want to analyse the effect of several independent variables on the
democratization process of almost every country in the world. At the
same time I have 30 years for every country.
1) Is there a possibility to do a fixed effects regression with my
imputed datasets with Clarify (xtreg) or do I have to use Zelig?
2) If I have to use Zelig - what do I have to do or where can I find
Information?
3) After this regression I want to use the synthetic control method.
I donŽt think that I can use the Imputed data sets for this. Is
there a way to put the imputed datasets together in one dataset?
Can anyone help me figure out a way to get around this problems?
Thanks in advance for any help you can give.
Sincerely,
Laeticia Klein
_______________________________________________________________________
DSL zum Nulltarif + 20 Euro Extraprämie bei Online-Bestellung über die
DSL Freundschaftswerbung! http://dsl.web.de/?
ac=OM.AD.AD008K15279B7069a
-
Amelia mailing list served by Harvard-MIT Data Center
[Un]Subscribe/View Archive: http://lists.gking.harvard.edu/?
info=amelia
More info about Amelia: http://gking.harvard.edu/amelia
5:08 p.m.
Hello everyone,
I was working with imputed datasests once and I created dummies for
countries. I supposed that was the easiest way to get FE estimation with
multiple imputed datasets. Then I used "miest" and included the country
dummies in the regression...From reading your responses to Laeticia, I guess
what I did was wrong...is it?
Another thing that occurred to me, but it's incredibly time-consuming...And
I'm not sure it's right (please let me know if it's not): run the xtreg
regression on each of the imputed datasets. Then averaged the coefficients
plus calculate the SE of each following the formula in Gary King et al.'s
paper...
If what I have done makes no sense at all....please let me know so that I
don't make the same mistakes again/.
thanks
Helen A Brown
On Mon, Mar 30, 2009 at 7:37 AM, Laeticia Klein <Laeticia.Klein(a)web.de>wrote;wrote:
Hello,
I want to analyse the effect of several independent variables on the
democratization process of almost every country in the world. At the same
time I have 30 years for every country.
1) Is there a possibility to do a fixed effects regression with my imputed
datasets with Clarify (xtreg) or do I have to use Zelig?
2) If I have to use Zelig - what do I have to do or where can I find
Information?
3) After this regression I want to use the synthetic control method. I
donŽt think that I can use the Imputed data sets for this. Is there a way to
put the imputed datasets together in one dataset?
Can anyone help me figure out a way to get around this problems?
Thanks in advance for any help you can give.
Sincerely,
Laeticia Klein
_______________________________________________________________________
DSL zum Nulltarif + 20 Euro Extraprämie bei Online-Bestellung über die
DSL Freundschaftswerbung! http://dsl.web.de/?ac=OM.AD.AD008K15279B7069a
-
Amelia mailing list served by Harvard-MIT Data Center
[Un]Subscribe/View Archive: http://lists.gking.harvard.edu/?info=amelia
More info about Amelia: http://gking.harvard.edu/amelia
5513
days inactive
5513
days old
amelia@lists.gking.harvard.edu
3 comments
4 participants
participants (4)
-
Gary King -
Helen Brown -
Laeticia Klein -
Nathan Paxton