Measuring quality in pre-kindergarten classrooms: assessing the early childhood environment rating scale

Abstract

Measures that assess the quality of early childcare environments are not only useful to researchers investigating the link between quality and child outcomes but are also useful to policy makers that use such measures to determine funding systems for childcare programs. This project involved examining the most widely-used instrument designed to evaluate the quality of early learning environments, the Early Childhood Environment Rating Scale-Revised Edition (ECERS-R). The ECERS-R has often been used in studies that focus on the impact of early educational quality on children’s development. Before an instrument can be used in such research, it is important that the measure be psychometrically sound. In addition to the ECERS-R use in research, several states use the ECERS-R as part of their quality rating system to determine which programs are to receive government funding. The policy arena, as well, should be concerned that measures put to such use not only have good psychometric properties but also include items that are thought to be important for quality. Using secondary data sets and surveys of field experts, this study sought to answer four main questions. First, the study compared the different scoring methods possible for the ECERS-R and applied them to the same data to determine their influence on the final quality ratings a classroom receives. Second, the study examined how the ECERS-R reflects the views in the field currently about what aspects of a classroom contribute to quality. Third, the psychometric properties of the ECERS-R and an alternate version of the instrument based on expert opinion were examined. Fourth, this study sought to examine how altering the content of the ECERS-R based on expert opinion might influence a classroom’s scores and be important for policy decisions. Results indicated that the ECERS-R scores are generally unaffected by the scoring method used, and the instrument has fairly sound psychometric properties. However, serious concerns were explored about the content of the instrument; recommendations for a different way of thinking about measuring quality were proffered.