How Measurement Error Affects the Interpretation and Understanding of Effect Sizes


  • This study was supported by the National Institute for Early Education Research and U.S. Department of Education (R305A060021).

Correspondence concerning this article should be addressed to Margaret R. Burchinal, Department of Education, 2001 Berkeley Place, University of California, Irvine, Irvine, CA 92697-5500; email:


ABSTRACT—It is important to understand the role of measurement error when interpreting effect sizes, especially when studying development in very young children. The accuracy of measurements of children’s developmental skills depends on the age of the child and the type of development being assessed. On average, measurement error is lower in older than in younger children and for assessments of cognitive, language, and achievement skills than of social–emotional skills. Standardized effect sizes tend to be lower when measurement error is larger because the variability in the scores, based in part on error of measurement, is used in computing those effect sizes. Because more reliable measures will lead to stronger effect sizes, it is more difficult to obtain large effect sizes in studies of very young children than of older children or in studies than in less robust assessments of children’s developmental skills.