Beyond the significance test in administrative research and policy decisions.
PURPOSE: To describe confidence interval (CI) analysis and show how it can be used in administrative decisions. ORGANIZING CONSTRUCT: Statistical significance testing should be supplemented, if not replaced, by effect size (ES) estimation and confidence interval analysis. Hypothesis testing based on the statistical significance test is the dominant paradigm in statistics; however, this approach has inherent problems which can ultimately diminish the usefulness of research for operational decisions. After identifying major difficulties with significance testing, the authors use hypothetical examples to demonstrate how ES and CI analysis provide more informative answers to nursing administrative research questions. CONCLUSIONS: CI analysis provides the basis for both improving the interpretation of findings from individual studies and for facilitating the analysis of cumulative evidence. By clarifying the meaning of results, CI analysis can increase the relevance and usefulness of research for health care executives and practitioners.