Citation

McCartney K, Rosenthal R. Child Dev. 2000; 71(1): 173-180.

Copyright

(Copyright © 2000, John Wiley and Sons)

DOI

unavailable

PMID

unavailable

Abstract

Real decisions for real children are influenced by the papers developmentalists write, regardless of whether we ever intended our papers to be used in the policy arena. Yet most social scientists seldom analyze data in ways that are most useful to policymakers. The primary purpose of this paper is to share three ideas concerning how to evaluate the practical importance of a finding or set of findings. First, for research to be most useful not only in the policy arena but also more generally, significance tests need to be accompanied by effect size estimates. The practical importance of an effect size depends on the scientific context (i. e., measurement, design, and method) as well as the empirical literature context. Second, researchers need to use all existing data when weighing in on a policy debate; here, meta-analyses are particularly useful. Finally, researchers need to be careful about embracing null or small findings, because effects may well be small due to measurement problems alone, particularly early in the history of a research domain.