Please give a more detailed explanation of what you are trying to do. I have seen an idea about the format of items not forcing dependence when writing items, e.g., Who was Jack's friend? [Mary, Jill] What is her chore? [Fetch water, shepherding]. So that responding Mary and shepherding results in two wrong answer even though Mary had a little lamb.

Are you trying to get a correlation matrix among dichotomous items after partialling out the total items correct? like the following syntax?

compute readscor = sum(item01 to item10). Partial Corr variables = item01 to item10 by readscor.

Hope this helps.

Art Art@DrKendall.org Social Research Consultants University Park, MD USA (301) 864-5570

Jude Cosgrove wrote: > Hi > > I would like if possible to carry out tests of local independence > between pairs of reading test items (i.e., the probability of getting a > correct response on one item should be independent of a correct response > on another item within a passage). The only method that I have found so > far is cumbersome in that it involves no less than three test item > analysis software packages. There is a test which computes the > correlation between residuals of pairs of items but again after a search > on SPSS home page I cannot find anything that appears to be analagous to > it. > > Has anyone carried out such a test in SPSS? > > Thanks in adavance. > Jude > > Jude Cosgrove > Research Associate > Educational Research Centre > Drumcondra > Dublin 9 > > jude.cosgrove@erc.ie > > Ph +353 1 837 3789 > Fax +353 1 837 8997 > >