Effect Size Calculation in Power Estimation for the Chi-square Test of Preliminary Data in Different Studies

Anna Laszlo ； Agnes Feher ； Anna Juhasz ； Tibor Nyari ； Krisztina Boda ； Jozsef Csicsman ； Ferenc Bari

《Statistics Research Letters》 2卷2期 (2013/05) Pp. 31-43

並列摘要

The validity of statistical analyses applied to identify different factors in many fields depends upon the use of appropriate sample sizes, the lack of which reduces the power of the findings. However, the number of cases collected for data analysis in medical studies is generally limited, for medical, financial and other experimental reasons, and statistical tests are often carried out without power and sample size estimation. Power analysis involves several parameters, the most important of which, the effect size, reflects the degree of the effect expected to be found in the study. An easy-to-use MS Excel calculator has been constructed to determine the effect size for chi-square tests based on 2×2, 2×3 and 2×4 contingency tables, and compared the results obtained with this calculator with those given by GPower, R and, for a 2×2 table, SAS software to demonstrate the practical use of this calculation tool in three studies involving various data.

並列關鍵字

Frequency Data ； Contingency Table ； Preliminary Study ； Effect Size ； Statistical Power Estimation ； GPOWER ； R ； SAS