Question

does a relationship exist between race (White, black, Asian/pacific islander, other) and satisfaction with the primary care provider (on a scale 1-10, Where one = completely unsatisfied and 10 = extremely satisfied)? ( n=52, satisfaction is not normally distributed). What statical test would you use?

Answer 1

We know that when we determine the relationship between two continuous variable which are distributed normally, then we use pearson correlation.

In this case, data is not normally distributed, i.e. this is a case of non-paramtric statistics because the satisfaction is not distributed normally. Satisfaction is an ordinal variable because we can satisfaction from 1 to 10 on

So, for non parametric statistics where at least one variable is ordinal, then we use spearman correlation for determining the relationship between the variables.

So, spearman correlation is statistical test for this case.