The CompareDistributions Class in Quorum can check if data has a normal distribution or does not have a normal distribution. Data without a normal distribution may not have symmetry or have really big values. Some samples may not be normally distributed which can affect what tests in the CompareGroups class can be used on the data set.
|Formal Test||Action in CompareDistributions Class|
Check Normal Distribution
Technical Test Name: Shapiro-Wilk Test
We can use a Shapiro-Wilk Test to check if a sample is distributed normally. To implement this in Quorum, refer to the example below.
If the data is normally distributed, then it would be in a bell-shaped curve. If it is not normally distributed, then more data might be on one side of the curve than the other.
Example of Shapiro-Wilk Test
In the next tutorial, we will discuss Compare Variances, which describes how to compare the variances between groups.