Is there a way to bound variables via relationships with other variables in
the dataset when multiply imputing with Amelia? For example, if we have a
household income variable and a household savings variable (both with some
missingness) in our dataset, is there a way to specify that savings must be
less than income for each observation in the imputed datasets? Priors and
logical bounds currently in Amelia don't seem like they are set up to do
this exactly.
Thanks!
--
Patrick Lam
Department of Government and Institute for Quantitative Social Science,
Harvard University
http://www.people.fas.harvard.edu/~plam
I am trying to plot the the observed and imputed values (with confidence
intervals) in a time series cross sectional data set, using tscsPlot.
Here is my imputation command:
EXT_impute <- amelia(imputeVars, m=20, cs="tcid", ts="age", intercs=TRUE,
polytime=1)
The imputation runs fine, and the dataset includes the cs and ts variables
in the arguments (corresponding to the first and second variables, 'tcid'
and 'age', respectively):
> EXT_impute$arguments$cs
[1] 1
> EXT_impute$arguments$ts
[1] 2
However, when I run the tscsPlot command, I receive the following error
(changing "tcid" to 1 does not resolve the error):
> tscsPlot(EXT_impute, cs="tcid", var="extPOM")
Error in tscsPlot(EXT_impute, cs = "tcid", var = "extPOM") :
the cross-section unit is not in the data
Why is the tscsPlot not reading the cs unit that I specified in the
imputation command? Any help would be greatly appreciated! Thanks in
advance.
Dear Amelia users
We are running impuations. In the model we are currently running we
intentionally avoid to use ordinal-specifications. We have some
specifications for nominal variables. We have successfully done multiple
imputations with the same dataset specifying many ordinal variables and
some more nominal variables.
However, now we are running into the problem that for some reason we get
for many single-imputation runs the error mentioned in the header of
this email. Some runs seem to work fine (up to now). From
https://lists.gking.harvard.edu/pipermail/amelia/2011-February/000547.html
I see that this problem might be related to the specification of the
nominals and collinearity among them in their "decomposed" nature. Note
that we have "task 1 failed" and not "La.svd(x, nu, nv)" (followed by
"error code 1..."). If this is still related to near-singularity
problems, I have difficulties to understand why this happens now in a
model where we have no ordinals and fewer nominals.
Do you have a suggestion?
Best,
Philippe
--
===============================================================
Philippe Sulger
Visiting Scholar
Institute of Criminology
Sidgwick Avenue
University of Cambridge
CB3 9DA Cambridge, UK
===============================================================
Email: philippe.sulger(a)econ.uzh.ch
ps596(a)cam.ac.uk
Web: http://www.econ.uzh.ch/faculty/sulger.htmlhttp://www.crim.cam.ac.uk/people/visitors/philippe_sulger
===============================================================
Dear Listserve,
I'm not sure whether to include features of survey design (clustering,
stratification, sampling weights) and features of survival analysis
(follow-up times and the failure indicator) in the imputation model.
With regard to features of survey design, the answer seems to be yes,
based on the information in Footnote 18 in King et al. (APSR 2001).
However, the footnote only addresses how to handle strata (using
dummies) and I've not seen this mentioned elsewhere.
Also, I have a quadratic transformation of a variable in my analysis
model as well as an interaction between two variables. The variables
underlying the quadratic transformation and interaction need to be
imputed. Is it correct that the quadratic and interaction terms should
be in the imputation model also?
Thanks very much for your help.
Andrew
Upenn
Dear all,
in my dissertation on FDI and human rights, which is situated in IPE and comparative human rights, I'm arguing that many political science fields already commonly apply multiple imputation methods. Comparative human rights literature, and the literature on globalization and human rights are still using 'list wise delete.'
From your experience, in which other fields in Political Science is multiple imputation already well accepted and commonly used?
Is it commonly used in (some fields in) Economics?
If anyone can point me to some literature or review articles on that, or just 'name' some fields, that would be great.
Thank you very much in advance!
Best,
Nicole
Nicole Janz
Doctoral Researcher
University of Cambridge
Politics and International Studies
www.nicolejanz.de | nj248(a)cam.ac.uk | +44 (0) 7905 70 1 69 4
Skype: nicole.janz
Dear Amelia Mailing List, There appears to be a bug in AmeliaView for
Windows. I have Windows 7 professional 64-bit. I have R 2.15.1. I
successfully installed the program to my desktop but I'm not able to
open the software. A DOS prompt appears but then is gone in a flash
and no session opens. Thanks in advance for your advice! Andrew
Dear Amelia-Users
I run imputations with some auxiliary variables that are nominal or
ordinal. Do I have to classify them as ordinal or nominal for the
imputation, or is it sufficient only to classify the variables of
interest to be be nominal or ordinal? In other words: does it affect the
quality of imputation in any sense if I don't classify them?
Thank you for your efforts.
Best,
Philippe
--
===============================================================
Philippe Sulger
Visiting Scholar
Institute of Criminology
Sidgwick Avenue
University of Cambridge
CB3 9DA Cambridge, UK
===============================================================
Email: philippe.sulger(a)econ.uzh.ch
ps596(a)cam.ac.uk
Web: http://www.econ.uzh.ch/faculty/sulger.htmlhttp://www.crim.cam.ac.uk/people/visitors/philippe_sulger
===============================================================