Item Analysis Help


I am working on a paper for a presentation later this month, and I am stuck in doing an analysis. For simplicity sake, I will only describe as much as is needed.

An inventory has twelve items. Each item consists of a stimulus and a reaction from the participant. For each item, the participant is asked to rate how well (using a likert scale) one responded to the stimulus.

I contend that each item may have been affected by responses of previous items. So item #3 was affected by responses to #1 and #2, # 6 was affected by items 1 though 5, etc.

A correlational analysis shows significance among all of the items, but this does not account for previous responses.

I did partial correlations, but later read that partial correlations should not account than more than a few variables. I was accounting for as many as ten variables.

A path analysis showed some promise (and seemed similar to my faulty partial correlations), but the chi square was way too large to find a good model. I assume that my model (where each item is connected to one another) has to be the true model, as that was the way in which it occurred.

What should I do? :eek: