1. Suppose there is a first order MLR model with only two continuous predictors, predictor A and B and no interaction term. Someone wants to compare the impact of the predictor A on Y to the impact of predictor B on Y. Comment on the following approaches on whether it always works, or under what circumstances it might not work. Then propose your own plan of solution.

Approach 1 is to compare the linear coefficients in the MLR model to decide which one is greater.

Approach 2 is to compare the simultaneous confidence interval of the coefficients in the MLR model to decide whether one is greater.

Approach 3 (Your approach), propose your plan and briefly explain the purpose of each step. Please write Approach 3.