An Introduction to Confidence Intervals for Both Statistical Estimates and Effect Sizes / Mary Margaret Capraro.

This paper summarizes methods of estimating confidence intervals, including classical intervals and intervals for effect sizes. The recent American Psychological Association (APA) Task Force on Statistical Inference report suggested that confidence intervals should always be reported, and the fifth edition of the APA "Publication Manual" (2001) said confidence intervals were "the best" reporting device. The correct use of confidence intervals uses intervals to compare results across prior studies and of prior studies with current studies. Confidence intervals provide a graphical tool to integrate or synthesize results across studies. They invoke two primary concepts, intervals and confidence levels. Intervals are determined by the standard errors of statistics, and confidence levels are chosen by the researcher and given as percentages. In this way, a range null hypothesis is tested rather than a point null hypothesis. New software has reduced the difficulty of establishing confidence intervals. Combining effect size with confidence intervals is the wave of the future in continuing efforts to make research understandable for the reader. Appended are esci (exploratory software for confidence intervals, La Trobe University, Australia)-created figures. (Contains 5 tables, 9 figures, and 30 references.) (Author/SLD)

Microfiche. [Washington D.C.]: ERIC Clearinghouse microfiches : positive.