Considerations for test selection: How do validity and reliability impact diagnostic decisions?

Abstract

Nine preschool and school-age language assessment tools found to have acceptable levels of identification accuracy were evaluated to determine their overall levels of psychometric validity for use in diagnosing the presence/absence of language impairment. Eleven specific criteria based on those initially devised by McCauley and Swisher (1984) were applied to each of the selected tests in order to determine each test’s overall level of psychometric validity. Results indicated that each of the selected assessment tools met at least eight of the 11 criteria used to evaluate each assessment tool. Five tests met 10 out of 11 criteria. Findings are discussed to assist clinicians in applying psychometric criteria to these selected tests, as well as those not reviewed as part of this current review of standardized assessment tools. A decision tree is included within the discussion of this study’s findings to aid clinicians in the selection of standardized assessment tools that are most appropriate for clinical use, based on their psychometric characteristics.