Background: The outcomes of a research investigation are presented as statistically significant or not statistically significant, or testing alone, including the reporting of p-values. This information conveys the statistical outcomes, or statistical findings, in response to the hypotheses of interest. A companion issue is the results of the estimation plus testing, clinical or practical significance, or the findings for the hypotheses of interest. Estimation plus testing, clinical or practical significance, findings provide information as to the strength of the finding, the differences that were detected between treatment groups, or other such conclusions. Objectives: Effect sizes are recommended as a measure of estimation plus testing, clinical significance, since they are generalizable and invariant. A case is made for the reporting of estimation plus testing for outcomes in biomedical journals. Design: A review of recent publications reporting effect sizes as well as a review of publication polices for biomedical journals are discussed. Results: Of the 113 articles in 38 medical journals that mentioned effect size, 35% were meta-analyses or systematic reviews. Of the original research reported, 12% reporting effect size were randomized control trials, 54% were descriptive or observational studies. Six of the 16 Public Health/Epidemiology journals contained effect size statistics in 24 articles. Studies reporting meta-analyses accounted for 17% of the total number of Public Health/Epidemiology articles reviewed. Of the 38 medical journals and 16 Public Health/Epidemiology journals reviewed, the "Instructions for Authors" were typically stylistic in nature. Conclusions: When a criterion for testing alone, namely statistical significance, is met and a criterion for estimation plus testing, namely clinical significance is met, then a conclusion of effectiveness may be reached. For a complete interpretation of research results, the authors strongly encourage the reporting of estimation plus testing.
