How to use data transformation to address issues with non-normally distributed data
Some quantitative analyzes of parametric statistics require the assumption of normally distributed data. For example, when researchers want to test differences in paired samples using the t-test, researchers need to test the normality of the data.