Effect Size Estimation and Confidence Intervals

Abstract

We describe a six-step estimation framework for research that starts with the formulation of research goals in terms of “How much?” questions. Such questions are best answered by effect size (ES) estimates and confidence intervals (CIs) calculated from data, where the ESs estimates and CIs are point and interval estimates of population parameters. These estimates usually provide the best basis for the interpretation of research findings. Such an estimation approach includes use of precision-measured by CI width-for the planning of research, and extends naturally to meta-analysis. We explain why an estimation approach is usually highly informative, and much better than traditional null hypothesis significance testing (NHST). We describe a range of ES measures and explain how to calculate CIs for these measures. We emphasize choice of ESs that most closely correspond to the research questions; construction of figures that support good understanding of ESs and CIs; and the importance of meta-analytic thinking-the consideration always of the current study in the context of past and possible future studies that address similar questions. Psychology can improve its research greatly by adopting estimation much more widely. Keywords: estimation; effect sizes; confidence intervals; meta-analysis; precision