|
|
Dear Alex, thank you very much for your help. Paulo On Fri, Apr 6, 2012 at 10:13 AM, Alex Reutter [via SPSSX Discussion] <[hidden email]> wrote:
Hi Paulo,
Here's how I would characterize it:
There is no such thing as a "pooled
dataset" in multiple imputation. When you run a procedure on
a multiply imputed dataset, that procedure is run on each of the imputed
datasets, and then the *results* are pooled in order to give you pooled
output. It's the pooled output, along with each of the multiply imputed
datasets, that you want.
If, at the end of your imputations,
you have a single dataset that you perform analyses on, then you are doing
single imputation, regardless of how you arrived at those imputed values.
Alex
From:
pfreire <[hidden email]>
To:
[hidden email]
Date:
04/05/2012 10:57 PM
Subject:
SPSS pooled
dataset
Sent by:
"SPSSX(r)
Discussion" <[hidden email]>
Dear all,
although some peolpe have posted the same question I couldn´t find an answer
that could fulfil my needs.
The literature says that when calculating values, via multiple imputation,
relplacing missing values in a dataset, one should use a pooled dataset.
I am using SPSS V.18 and when performing a multiple imputation, I just
can´t
find an output that provides me with a pooled complete dataset that I can
use in further statisical analises. all I get are the original dataset
and
the datasets for the various imputations.
I would be gratefuf if anyone could help me out in learling how to get
multiple imputation´s pooled dataset.
Thanks,
Paulo
|
Locked
