problem-11.3
problem-11.3
A boxplot will show that the income data, like most income data, is
heavily skewed.
> boxplot(income ~ race, data=female.inc)
An analysis of variance using oneway.test() would be
inappropriate. However, the three population distributions appear
to be roughly the same shape. The Kruskal-Wallis test then is
appropriate.
> kruskal.test(income ~ race, data=female.inc)
Kruskal-Wallis rank sum test
data: income by race
Kruskal-Wallis chi-squared = 11.79, df = 2, p-value = 0.002748
The small p-value is not consistent with the assumption of equal
mean wages for all three races represented in the data.