Posted by
Joost van Ginkel on
Apr 06, 2010; 7:33am
URL: http://spssx-discussion.165.s1.nabble.com/Multiple-imputation-tp1075737p1075738.html
Dear Sungrok,
To answer your other two questions:
2) You should always report the pooled results, especially when the
results of the original data and the pooled results differ. If there are
differences between the results of the original data and the pooled
results, this means that the missingness is probably not missing
completely at random. If this assumption is violated, listwise deletion
may give biased results.
3) SPSS does not report pooled R-squared's and standardized beta's,
probably because no rules have been defined for pooling these
statistics. Maybe you can compute R-squared by simply computing the
average R of the imputed datasets, and by squaring the result. But maybe
other people have some better suggestions?
There are, however, rules for combining the F-values (Schafer, 1997, pp.
113-114), only they are not implemented in SPSS (yet). If you want to
combine the results of the F-values, this may be done using an SPSS
macro named MI-mul.sps which can be downloaded from my personal page:
http://www.socialsciences.leiden.edu/educationandchildstudies/childandfamilystudies/organisation/staffcfs/van-ginkel.html
I just happen to work at an improved version of this macro, namely
MI-mul2.sps, which is nearly finished now. If you wait one or two more
days, you can download it from the same link. If you can't wait that
long you can just use the old version, which is a lot less user
friendly.
Good luck!
Joost van Ginkel
Joost R. Van Ginkel, PhD
Leiden University
Faculty of Social and Behavioural Sciences
Data Theory Group
PO Box 9555
2300 RB Leiden
The Netherlands
Tel: +31-(0)71-527 3620
Fax: +31-(0)71-527 1721
-----Original Message-----
From: SPSSX(r) Discussion [mailto:
[hidden email]] On Behalf Of
Sungrok
Sent: 05 April 2010 20:10
To:
[hidden email]
Subject: Multiple imputation
Hi, all.
I have some questions about multiple imputation (I run MI using SPSS
17).
1. Should I have to report results based on the original dataset or
imputed datasets?
2. If there was a difference between original and imputed datasets,
what do I have to use? I mean there ara rules or criteria for decision?
3. In spss 17, after MI, I run hierarchical regression as below.
----------------------------------------------------
SORT CASES BY Imputation_.
SPLIT FILE LAYERED BY Imputation_.
REGRESSION
/DESCRIPTIVES MEAN STDDEV CORR SIG N
/MISSING LISTWISE
/STATISTICS COEFF OUTS R ANOVA CHANGE
/CRITERIA=PIN(.05) POUT(.10)
/NOORIGIN
/DEPENDENT T_PTSD
/METHOD=ENTER FEScoh FESconf entage D_enrank
/METHOD=ENTER CE TO_mimpo Cohesion Desire Undesire
/METHOD=ENTER ELDER SLE SS
/RESIDUALS HIST(ZRESID) .
-----------------------------------------------------
The problems of this result were I couldn't find R, R square,
adjusted R square, and F value of pooled imputation. I mean I could find
these results for each imputed data, but pooled imputation data. How can
I report R, R square, adjusted R square, and F value? In addition, I
couldn't find beta (standardized coefficients) of pooled imputation
(there were only unstandardized coefficients: B and Std. Error).
Please give me some direction. Thank you so much in advance.
Sincerely,
Sungrok
--
View this message in context:
http://old.nabble.com/Multiple-imputation-tp28142986p28142986.htmlSent from the SPSSX Discussion mailing list archive at Nabble.com.
=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command SIGNOFF SPSSX-L For a list
of commands to manage subscriptions, send the command INFO REFCARD
**********************************************************************
This email and any files transmitted with it are confidential and
intended solely for the use of the individual or entity to whom they
are addressed. If you have received this email in error please notify
the system manager.
**********************************************************************
=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD