The Wilcoxon-Mann-Whitney test is known for its nonparametric analog, where the t test (the independent samples) can be leveraged if there is no assumption of a dependent variable being a normally distributed variable and where the variable is just an ordinary variable. This is one of the reasons why the Stata code for the Wilcoxon-Mann-Whitney test is similar to that of an independent sample t test. Here is an example:
ranksum fridge_sales, by(model)
Here are the results of this query:
In this case, here is the mathematical formula:
Here, U is referring to normally distributed variables in the given dataset.
n1 and n2 are the number of observations in the related datasets.