To: Bobbie Mircheva <bmircheva@american.edu>
From: Gary King <king@harvard.edu>
Date: 09/08/2007 03:39AM
cc: amelia@lists.gking.harvard.edu
Subject: Re: [amelia] negative values

use each data set as if it contained fully observed data, and then take
the 5 estimates and combine them as described in our APSR paper (which you
can find on our web site).  Clarify in Stata will do this automatically if
you use one of its models, or Zelig in R will do it for a much larger and
growing range of models (both are on my web site).  But its easy to do by
hand too. Gary

On Fri, 7 Sep 2007, Bobbie Mircheva wrote:

> Gary,
> Thanks a lot for the quick response. I did what you suggested and it worked great. I
> do have one more question if you do not mind. Since I know have 5 output, which one
> am I supposed to use. Or is there a way to compile them? I am working with STATA and
> tries to I tried to use Ken Schieve's program, but somehow could not manage to make
> it work. I still need to do more manipulations with the data (take first differences,
> etc) before I start running regressions.
> Any suggestions would be greatly appreciated.
> Thanks again!
> Borislava Mircheva
> PhD Candidate
> Department of Economics
> American University
> https://eagle1.american.edu/~bm6265a/
>
> -----Gary King <king@harvard.edu> wrote: -----
>
>       To: Bobbie Mircheva <bmircheva@american.edu>
>       From: Gary King <king@harvard.edu>
>       Date: 09/07/2007 06:20PM
>       cc: amelia@lists.gking.harvard.edu
>       Subject: Re: [amelia] negative values
>
>       one way is to transform the data to an unbounded scale.  e.g., if all
>       observed and unobserved values of a variable are positive, then transform
>       the positive ones by taking the log.  then impute with Amelia (on the
>       unbounded log scale).  then transform the resulting imputed and observed
>       values by taking the exponential.
>
>       you can also use the priors in Amelia II to help as well.
>
>       Gary
>
>       ---
>       Gary King
>       David Florence Professor of Government,
>       Director, Institute for Quantitative Social Science
>       Harvard University, 1737 Cambridge St, Cambridge, MA 02138
>       http://GKing.Harvard.Edu, King@Harvard.Edu
>       Direct 617-495-2027, Assistant 495-9271, eFax 812-8581
>
>
>
>       On Fri, 7 Sep 2007, Bobbie Mircheva wrote:
>
>       > Hi,
>       > I have a big panel dataset on countries in Sun-Saharan African with
>       > information on trade and infrastructure between 1990 and 2005.
>       > Unfortunately some of the infrastructure data is missing and I was
>       using
>       > Amelia to fill in the gaps. However, the program returned a bunch of
>       > negative values. How can I fix this?
>       > Thanks a lot
>       > Bobbie
>       > - Amelia mailing list served by Harvard-MIT Data Center
>       [Un]Subscribe/View
>       > Archive: http://lists.gking.harvard.edu/?info=amelia
>       >
>
>
>
>