Question

Which of the following would be a violation of our assumption of no perfect multicollinearity?

Group of answer choices

Including age and age² as independent variables in a regression estimating the effects of age on wages.

Including years of education and years of experience as a independent variables in a regression estimating the effects of education on wages.

Including both weight in pounds and weight in kilograms as indepdent variables in a regression of weight on heart disease morbidity.

Answer 1

Answer :

Option c) Including both weight in pounds and weight in kiolograms as independent variables in a regression of weight on heart disease morbidity.

The assumption of no perfect multicollinearity implies that the independent variables in a regression model should not be perfectly correlated. But here when you include both weight in pounds and weight in kilograms, that basically implies the presence of perfect correlation since weight measured in any unit will always move in the exact same way.

That is to say that if weight in pounds is to increase then weight in kilograms will also increase and to the same extent and vice versa. Thus these two variables if added in the regression model, violates the assumption of perfect multicollinearity.