Huynh, H., Michaels, H., Ferrara, S.(1995). Statistical Procedures To Identify Clusters of Items with Local Dependency. Paper presented at the annual meeting of National Council on Measurement in Education, San Francisco, CA.

Abstract: This paper compares three statistical procedures to identify clusters of items with local item dependency (LID). The first procedure is based on the inter-item correlation matrix pooled across groups of examinees with similar raw scores. The second method relies on the partial correlation matrix, with the total raw score as the partialing variable. Within the framework of the Rasch/masters partial credit model, Yen's Q3 provide the basis for the third procedure. The three methods were implemented for the Mathematics Content subtest data from the 1993 Maryland School Performance Assessment Program (MSPAP). It was found that the three processes yielded essentially the same LID results.