Dear colleagues,
I am new to the list (as of a minute ago), and I am subscribing at the
suggestion of Solaris to find out how to combine F statistics across
multiply imputed data sets. Before I subscribed, I checked out the impute
list archives, and I couldn't find any information or references to
information about combining F statistics. So I hope I'm not dredging up the
same old same old.
I have a 3 x 2 x 2 x 2 mixed design ANOVA (the first three factors are
repeated measures) from an experimental procedure focused on teaching
21-month-olds new words. My original sample size is about 50, but due to
one or another missing value here and there, I only have COMPLETE data on 27
kids. The reason the data are missing has nothing to do with the children
themselves, but are due primarily to the experimenter sticking her head in
front of the camera (and other experimenter errors).
In any case, I can easily enough calculate the missing values through
Solaris, but then I get five complete sets of F statistics, and I don't know
how to combine either the F statistics or the p values. Thanks in advance
for any help you can provide.
Regards,
Wally Dixon
----------------------------------------------------------------------------
Wallace E. Dixon, Jr. |
Chair and Associate Professor | Adults are obsolete children.
of Psychology |
Department of Psychology | -Dr. Seuss
East Tennessee State University |
Johnson City, TN 36714 |
(423) 439-6656 |
----------------------------------------------------------------------------
