Hi, Everybody in the forum:

I am hoping someone might have some advice for me regarding analysis of a measurement device. I have a scale on which participants imitate a number of movements performed by the tester. We are currently recording 5
different kinds of errors that they might make (these are coded as the actual number of each kind that they make, not a 'present/absent' score), along with a 0-3 overall 'rating'. Obviously, this overall rating is correlated with some of the errors. I also have 2 groups (patients and controls). Among other things, one of the main questions is how well the different errors types discriminate between groups. Further, a question of interest is whether some of the items are redundant, and could be gotten rid of, or if there are too many items altogether (after a certain point, no additional information is gleaned by having them do more movements). I'm a little unsure where to start, and any advice would be greatly appreciated.

Thanks in advance !

Marnie LaNoue, M.S.