27 Oct
2002
27 Oct
'02
11:56 a.m.
The main rule is to include any variables in the imputation model that
Amelia runs that you will use in any of your analysis models. that
includes any (nonlinear) transformations of variables or lags or
differences over time. So yes, you should include them.
Gary
: Gary King, King(a)Harvard.Edu http://GKing.Harvard.Edu :
: Center for Basic Research Direct (617) 495-2027 :
: in the Social Sciences Assistant (617) 495-9271 :
: 34 Kirkland Street, Rm. 2 HU-MIT DC (617) 495-4734 :
: Harvard U, Cambridge, MA 02138 eFax (928) 832-7022 :
On Sun, 27 Oct 2002, Diogo Lemieszek Pinheiro wrote:
Dear Dr. King I am sorry to bother you again, but I
have one more
question on AMELIA: In my model I am going to use both "level" variables
and change over time. SHould I use amelia on a dataset that only has the
level variables ot should I also include the change variables (in this
case, I have the ln of all variables, and change over time is simply
measured as t- t-1). Thanks Diogo L. Pinheiro
Quoting Gary King <king(a)harvard.edu>du>:
On Tue, 22 Oct 2002, Diogo Lemieszek Pinheiro
wrote:
Dear Dr. King,
First of all, let me thank you for creating AMELIA and distributing it
free. It
is an excelent software.
I got a few questions, sorry if they seem trivial:
1- When combining the results from the m datasets generated by Amelia,
how can I obtain the R squared and other "goodness of fit" indicators?
the same way you combine any of the other quantities of interest.
2- Should MI only be used for independent
variables or t doesnt matter?
you should use all your variables, x and y. it matters.
3- Using clarify, do I have to use the estsimp
command for every model I
am runnign from the imputed datasets? Using setx eliminate the need to
do that? If so, everytime I run stata again I have to use the estsimp
and then setx commands or is there a way to save it all?
you can run amelia once and then do as many analysis models using the
same imputed data sets.
Gary King
: Gary King, King(a)Harvard.Edu http://GKing.Harvard.Edu :
: Center for Basic Research Direct (617) 495-2027 :
: in the Social Sciences Assistant (617) 495-9271 :
: 34 Kirkland Street, Rm. 2 HU-MIT DC (617) 495-4734 :
: Harvard U, Cambridge, MA 02138 eFax (928) 832-7022 :
Once again, I am sorry if the answer to those
questions is obvious. I am
just
starting to learn about MI...
Thanks in advance
Diogo L. Pinheiro
Graduate Student
Dept of SOciology
EMory University