Compare two different groups that differ in size?

I think this issue comes up often when we try to compare certain variable between two groups... Thanks Dr MEL!

Thanks!

Usually, the size difference is not an issue. The concern is about the use of a nonparametric statistical test, such as the Kruskal-Wallis test to test the unemployment rate, which unlikely displays a normal distribution. Nevertheless, one reasoning logic of any is that the sample parental distribution resembles the population parental distribution. That is, when the unequal numbers of parental groups are realistic in the population, it is logical for the test.