An R web based application that will calculate sample size for the Mann-Whitney U Test
Abstract
This project intends to develop an R web-based application that can easily be usable by researchers who are not familiar with coding to calculate sample size for the Mann-Whitney U test with the primary focus on the effect size as an input among other parameters such as the significance level and the power of the test. The application was developed using the R environment integrated with Rstudio with the help of packages Shiny and ShinyWidgets. The dataset used was ―Students’ CGPA‖ which contains 200 observations of the CGPA of first year and third year students of Makerere University. The application returns the required number of sample size as 56 per group, effect size as 1.56, median 1 as 3.25 and median 2 as 3.7. The application is able to calculate mean, standard deviation, Z-score, effect size (r), median 1, median 2 and which are all used at different steps to reach at the final step of calculating the required sample size. It is recommended that the study is extended to support other nonparametric tests.