NORM REFERENCED TEST INTERPRETATION IN THE DIAGNOSTIC PROCESS

This study examines the extent to which norm-referenced tests can assi st in addressing two independent clinical questions within the diagnos tic process: ''Is there a language impairment?'' and ''What are the sp ecific areas of deficit?'' Children's performance on two tests, the Te st for Examining Expressive Morphology and the Patterned Elicitation S yntax Test, was examined from the perspective of each question. For th e first question, a discriminant analysis using 40 preschool children (20 with specific language impairment [SLI], and 20 with normally deve loping language) revealed 90% sensitivity and 95% specificity for each test. For the second question, an item analysis revealed inconsistent pass/fail rates and low point-to-point agreement for the performance of children with SLI on items targeting the same morphosyntactic struc ture across tests. Given their high discriminant capacity, but inconsi stent item-level performance, the results demonstrate that norm-refere nced tests can be appropriate diagnostic tools for one diagnostic purp ose but inappropriate for addressing another.