29 Apr
2014
29 Apr
'14
9:22 p.m.
Dear Amelia Team,
I've been having problems with Amelia crashing frequently. I can run it
with a few splines, but if I interact time and cross-section, it crashes.
This is the case whether I run it on my data, the example data, in R, in
AmeliaView, or on someone else's computer. Help?
Also, whenever I try to include similar variables -- alternative measures,
for example -- I get a collinearity warning. I'd like to be able to impute
everything, including all my robustness check measures, together. The
documentation makes it sound like Amelia can handle that.
Thank you,
Calla Hummel
5 May
5 May
3:52 p.m.
Hi Calla, It sounds like the issue is collinearity, which is something you
have to deal with before giving it to Amelia. you can diagnose the issue,
if it isn't obvious, by starting with a simple model and adding variables
gradually until you see the problems. In all likelihood, you're asking it
to do something that turns out to be impossible, such as if all the dummy
variables add up to the constant term.
Gary
--
*Gary King* - Albert J. Weatherhead III University Professor - Director,
IQSS <http://iq.harvard.edu/>- Harvard University
GaryKing.org - King(a)Harvard.edu - @KingGary <https://twitter.com/kinggary> -
617-500-7570 - fax 812-8581 - Assistant <king-assist(a)iq.harvard.edu>du>:
495-9271
On Tue, Apr 29, 2014 at 9:22 PM, Calla M Hummel <calla.hummel(a)utexas.edu>wrote;wrote:
Dear Amelia Team,
I've been having problems with Amelia crashing frequently. I can run it
with a few splines, but if I interact time and cross-section, it crashes.
This is the case whether I run it on my data, the example data, in R, in
AmeliaView, or on someone else's computer. Help?
Also, whenever I try to include similar variables -- alternative measures,
for example -- I get a collinearity warning. I'd like to be able to impute
everything, including all my robustness check measures, together. The
documentation makes it sound like Amelia can handle that.
Thank you,
Calla Hummel
--
Amelia mailing list served by HUIT
[Un]Subscribe/View Archive: http://lists.gking.harvard.edu/?info=amelia
More info about Amelia: http://gking.harvard.edu/amelia
Amelia mailing list
Amelia(a)lists.gking.harvard.edu
To unsubscribe from this list or get other information:
https://lists.gking.harvard.edu/mailman/listinfo/amelia
15 May
15 May
10:54 p.m.
Hi Calia,
One other thing to consider is that adding splines of time along with an
interaction with the cross-section can add *many* addition variables to the
imputation model. For a cubic polynomial, that's 4 parameters for each unit
in the data. That will eat up degrees of freedom very quickly and lead to
the kind of errors you are seeing. Gary's suggestion would help find the
right balance between complexity and parsimony.
Hope that helps!
Cheers,
Matt
~~~~~~~~~~~
Matthew Blackwell
Assistant Professor of Political Science
University of Rochester
url: http://www.mattblackwell.org
On Mon, May 5, 2014 at 3:52 PM, Gary King <king(a)harvard.edu> wrote:
Hi Calla, It sounds like the issue is collinearity,
which is something you
have to deal with before giving it to Amelia. you can diagnose the issue,
if it isn't obvious, by starting with a simple model and adding variables
gradually until you see the problems. In all likelihood, you're asking it
to do something that turns out to be impossible, such as if all the dummy
variables add up to the constant term.
Gary
--
*Gary King* - Albert J. Weatherhead III University Professor - Director,
IQSS
[iq.harvard.edu]<https://urldefense.proofpoint.com/v1/url?u=http://iq.har…
Harvard University
GaryKing.org[GaryKing.org]<https://urldefense.proofpoint.com/v1/url?u=ht…
- King(a)Harvard.edu -
@KingGary[twitter.com]<https://urldefense.proofpoint.com/v1/url?u=https://twitter.com/kinggary&k=p4Ly7qpEBiYPBVenR9G2iQ%3D%3D%0A&r=jLgdG6f%2BQq4pzHWI0S37ROhc5Jfy9q9oKEsPDdQXskc%3D%0A&m=ipSSygXtDr53LbWTQ5YYXrAJfZ5Z3WcM82mVqYjY0tA%3D%0A&s=dc87271086947bc96178686d27e0ec490eb4212064c230ed8a1f1fb8d6d6372d>
-
617-500-7570 - fax 812-8581 - Assistant <king-assist(a)iq.harvard.edu>du>:
495-9271
On Tue, Apr 29, 2014 at 9:22 PM, Calla M Hummel <calla.hummel(a)utexas.edu>wrote;wrote:
Dear Amelia Team,
I've been having problems with Amelia crashing frequently. I can run it
with a few splines, but if I interact time and cross-section, it crashes.
This is the case whether I run it on my data, the example data, in R, in
AmeliaView, or on someone else's computer. Help?
Also, whenever I try to include similar variables -- alternative
measures, for example -- I get a collinearity warning. I'd like to be able
to impute everything, including all my robustness check measures, together.
The documentation makes it sound like Amelia can handle that.
Thank you,
Calla Hummel
--
Amelia mailing list served by HUIT
[Un]Subscribe/View Archive:
http://lists.gking.harvard.edu/?info=amelia[lists.gking.harvard.edu]<htt…
More info about Amelia:
http://gking.harvard.edu/amelia[gking.harvard.edu]<https://urldefense.pr…
Amelia mailing list
Amelia(a)lists.gking.harvard.edu
To unsubscribe from this list or get other information:
https://lists.gking.harvard.edu/mailman/listinfo/amelia[lists.gking.harvard…
--
Amelia mailing list served by HUIT
[Un]Subscribe/View Archive:
https://urldefense.proofpoint.com/v1/url?u=http://lists.gking.harvard.edu/?…
More info about Amelia:
https://urldefense.proofpoint.com/v1/url?u=http://gking.harvard.edu/amelia&…
Amelia mailing list
Amelia(a)lists.gking.harvard.edu
To unsubscribe from this list or get other information:
https://urldefense.proofpoint.com/v1/url?u=https://lists.gking.harvard.edu/…
3640
days inactive
3656
days old
amelia@lists.gking.harvard.edu
2 comments
3 participants
participants (3)
-
Calla M Hummel -
Gary King -
Matt Blackwell