I collected experimental data of subjects performing movements in different
conditions (a Kinesiology study). All subjects performed in all
conditions, but technical problems in acquisition and processing left a
small number of cells blank (6 of 784, affecting 2 of 14 subjects). Can I
handle this with Amelia?
I tried NORM to impute some missing data but have some confusion about
setting the number of iterations (at several points in the process),
setting the number of imputations, and reducing to one data-set.
The goal is to apply repeated measures ANOVAs to test the effects of three
independent variables on 18 or 20 dependent variables. Subjects did
perform multiple responses in each condition, and the number of responses
per subject/condition are not balanced due to several technical issues, but
I hope to work around this (perhaps just average responses/subject/cell and
ignore the imbalance).
I would greatly appreciate any help to prepare my data for testing. I get
output from NORM, but it is not clear to me how to reduce to one
representative data set. "Stata" is not available, so I am not sure if I
can get to the one data set via Amelia, but perhaps I could use the
formulae presented, or enter results from Amelia back into NORM.
Deric Wisleder, Ph.D.
Show replies by date