Best Practices in Test Construction for Developmental-Behavioral Measures: Quality Standards for Reviewers and Researchers

Abstract

Developmental-behavioral measurement is fundamental to clinical determinations about children’s and families’ needs. Such measurement consists of three main types (listed in order of complexity): (1) screening, (2) mid-level assessment, and (3) diagnostic tests. The three types of measures depend on the same psychometric precepts although screening tests, despite their inherent brevity, depend on an additional construct: proof of accuracy. In this chapter, we focus on standards in test construction with additional emphasis on the psychometry of screening tests because screens are deployed more frequently and serve as a fundamental decision point for whether more complex measures are needed. For example, screening test results help identify whether children require further vision, hearing, or lead screening referrals for further evaluation by special education services and/or to developmental- behavioral pediatricians or other subspecialists. The powerful role of screening tests in decisions that profoundly affect families’ lives means that screens must be especially well-constructed. This review of methods in psychometry highlights how to research and review developmental-behavioral measures including screening tests. Crucial to test selection is an understanding of principles and policy in standardization, reliability, and validity, including accuracy computations in the case of screening tests, and utility-practical considerations in measurement.