22 Aug
2010
22 Aug
'10
4:19 p.m.
Hi Donald,
I think there are a few things going on. First, you have some very
highly correlated variables in your dataset. It's very possible that
one bootstrapped dataset could have two of the "policy" variables
identical, which would lead to the kind of problems you are seeing.
Second, you have roughly 100 variables, which leads to roughly 9000
parameters for Amelia to estimate. With only 1500 observations, this
could be problematic. If there are repetitive variables in the data,
you may want to try and omit them (especially if you are not going to
use them in any analysis model).
Cheers,
matt.
On Wed, Aug 18, 2010 at 4:44 PM, Donald Braman <dbraman(a)law.gwu.edu> wrote:
I get the following error when attempting to impute
missing data in the
attached dataset:
-
Amelia mailing list served by Harvard-MIT Data Center
[Un]Subscribe/View Archive: http://lists.gking.harvard.edu/?info=amelia
More info about Amelia: http://gking.harvard.edu/amelia
1 Error in eigen(thetanew[2:nrow(thetanew), 2:ncol(thetanew)],
only.values = TRUE, :
infinite or missing values in 'x'
A summary of the data set reveals no column with entirely missing or
constant values, and if I reduce the tolerance greatly, the problem goes
away.