12.96 Multiple choice: Correlation invalid The correlation is appropriate for describing association between two quantitative variables

a. Even when different people measure the variables using different units (e.g., kilograms and pounds).

b. When the relationship is highly nonlinear.

c. When the slope of the regression equation is 0 using nearly all the data, but a couple of outliers are extremely high on y at the high end of the x-scale.

d. When the sample has a much narrower range of x values than does the population.

e. When the response variable and explanatory variable are both categorical.