The Validity of Direct Observation Measures of Pupil Classroom Behavior

Abstract

An evaluation of the validity of direct observation measures of pupil classroom behaviors is presented. The evaluation is based on a survey of analyses bearing on their construct, criterion-related, and treatment validities. Three types of measures are discussed. These include molar measures, where the focus is on broad dimensions of classroom behavior (e.g., “on-task” vs. “off-task”), molecular measures, where the focus is on more specific dimensions of behavior, and molecular-composite measures, where the data are collected in terms of molecular units but scored in terms of composite categories. Consistent support for the validity of molar and molecular-composite types of measures was revealed, but support for the molecular schedules was less consistent. The paper also contains a set of recommendations for users of these instruments and some suggestions for future research on the measures.