I'm using the fwdmsa package to identify deviant cases in a Mokken scale
analysis. I've run into a problem., separate from the one I posted
previously. The problem comes with items that are "easy" by IRT standards. A
good scale should include a range of difficulties; yet when I include "easy"
items in a forward search I continuously run into the problem that these
items demonstrate no variability in the smaller subsamples selected for
forward search. I can resolve the problem by only running the analysis
without the easier items, but this is not ideal for for the analysis.
The data are 150 responses to a 37 item test. For the first 15 items, they
look like this (the variable names are cumbersome and so have been removed)
:
> head(by364.data)
1 1 1 1 0 1 0 1 1 0 1 0 1 1 1
1
2 1 1 1 0 1 1 1 1 1 1 0 0 1 1
1
3 0 1 1 1 0 0 1 0 1 1 1 0 1 1
0
4 1 1 1 0 0 1 1 1 1 1 1 1 0 1
1
5 1 1 1 1 0 1 0 1 1 1 1 0 1 1
0
6 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1
You can see that items 1, 2, 9 and 13 demonstrate no variance in this small
sample and thus muck up the analysis. This happens even when I have items
with plenty of variance, but seeing as they are easy items, they might have
a .8 - .9 probability of endorsement, so even in samples of 25-50, still run
into this variance problem. I have seeded the forward search with specific
cases that create the necessary variance, but these artificial subsamples
tend to loaded with deviant observations and thus are excluded after a
couple of steps, creating the lack of variance problem.
Any suggestions?
Robert
--
View this message in context:
http://r.789695.n4.nabble.com/fwdmsa-package-Error-in-search-normal-X-samp-verbose-FALSE-At-least-one-item-has-no-variance-tp4493531p4493531.html
Sent from the R help mailing list archive at Nabble.com.
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
