This problem is not doable. you have over 10,000 parameters under the
model and only 282 observations. And its a harder problem with so many
missing values. I think you probably need a different strategy: collect
better data, drop some of the variables, or get more observations.
Best of luck with your research,
Garh King
On Sun, 14 Aug 2005, Maria wrote:
Dear all,
I have a problem with imputing missing values in Amelia. I have a
dataset of 144 variables and 282 cases (I know it's no good :))); most
of the variables are count data which are NOT normally distributed nor
could they be transformed into normal. Percent of missingness is rather
high (there are very few fully complete cases, average percent - 18
and it varies from 0% to 90%). When I try to use Amelia just
to look what I might get, it simply vanishes from the screen without
any error messages after I press 'Run' button. I tried to specify most
simple options (a couple of indicator variables, a couple of fully
complete variables) as well as more complicated (e.g., a ridge prior
of 1; I would be grateful if somebody explained me what it exactly
means, I haven't found it anywhere). The result was always the same.
Thanks to everyone who might give some idea as to what to do with
this!
With kind regards,
Maria Pavlova (a psychologist)
Moscow State University
Russia
mailto:haul@rambler.ru
-
Amelia mailing list served by Harvard-MIT Data Center
[Un]Subscribe/View Archive:
http://lists.hmdc.harvard.edu/?info=amelia
-
Amelia mailing list served by Harvard-MIT Data Center
[Un]Subscribe/View Archive:
http://lists.hmdc.harvard.edu/?info=amelia