Hi Elisa,
If you suspect that the imputation model itself varies across the units, it
seems reasonable to use different imputation models for those units. So (a)
appears fine. As for (b), Zelig is probably the most straightforward way to
analyze regression models based imputations. See section 4.9 of our
documentation for more:
http://r.iq.harvard.edu/docs/amelia/amelia.pdf
Hope that helps!
Cheers,
matt.
~~~~~~~~~~~
Matthew Blackwell
Assistant Professor of Political Science
University of Rochester
url:
http://www.mattblackwell.org
On Sun, Dec 30, 2012 at 6:41 AM, Elisa D'Arcangelo <
darcangelo.elisa(a)googlemail.com> wrote:
Good morning,
I am using amelia II on mass spec data, in order to deal with missing
values. I have two questions:
1) my experimental setup is as follows: the intensities of the same
proteins were measured in ctrl, condition 1, condition2, condition3,
condition4. This model implies that certain proteins might not be present
at all in a certain condition, while being present in several other
conditions. hence, I don not want amelia to impute these values. In other
words, I want amelia to only impute WITHIN conditions. At least this is the
way I though of this problem and I sub-devided the data frame into 5,
according to conditions, and fed amelia each condition separately.
Is this approach OK?
2) I perform m=10 imputations, and fit each imputed data set to a linear
model, hence I get 10 slightly different outputs. In R, what is the best
way to pool the results of m data sets?
Thank you very much for your help, I appreciate it.
Have a good New Years!!
Regards,
Elisa
--
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