-----Original Message-----
From: Gary King [mailto:king@harvard.edu]
Sent: Thursday, November 11, 2004 3:31 PM
To: Dagmar Radin
Subject: Re: amelia
I'm out of town presently; can you send email to the amelia list and one of
my colleagues or I (after I get back) will figure it out? sorry for th
troubles. thanks, Gary
On Thu, 11 Nov 2004, Dagmar Radin wrote:
Dr. King,
I am a graduate student working on a paper and trying to use Amelia.
Emailing you was my last resort, as I have not been able to move
forward past some problems I am having. I read through the pdf file
several times, and the help menus, but things are not working out. I
have cross secitonal time series data with a number of missing
observations for different variables.
When I try to run the imputations, I get a message stating there are
insufficient observations for estimation, and that I need to use prior
or increase prior if already used. I tried to increase the prior
(AMempri) to more than zero, but the program states that the AMfully
is not properly defined and that I have more defined variables than
there really are, despite the fact that I specified only 2 or 3 out of
26 variables that have full values. For some reason the program
changes the AMfully definitions on its own. When I try to specify the
AMfully again, it tells me I need to define a different value for
AMempri. It is almost a circular problem.
I am sorry to bother you, but I have run out of resources on my own
and nobody in my department seems to have used Amelia or if they did,
they never had such problem. Where should I look next?
Thank you. Regards,
Dasha