Date of Award
Education and Human Services
Education -- Research, Education -- Research -- Methodology, Meta-analysis, Effect sizes (Statistics), Meta-analysis research methodology bias
Effect sizes are the statistic generated by meta-analyses, a commonly used statistic in education research. Meta-analyses are widely used by education practitioners, administrators, and policy makers as a means to decide best classroom or school practices. It has been suggested by authors in other fields, most notably Jennions and Moller (2001), that effect sizes have declined over time due to various sources of bias. This paper examines the question of whether shifting effect sizes can be observed in educational research and attempts to explain possible causes of this observation. It uses the methodological framework used by Jennions and Moller (2001) and applies it to educational meta-analyses conducted from 1970 to the present. It finds that, contrary to the findings of Jennions and Moller (2001), that effect sizes in educational research have increased over time. Likely explanations regard systemic bias in the conduct and publication of educational research. The paper concludes with recommendations for future research to examine causal factors contributing to this phenomenon
Stephens, Joshua A., "Longitudinal Stability of Effect Sizes in Educational Research" (2013). ETD Archive. 281.