Suppose that we fit the full model y = 50 + ,61X1 + mm + a to n statistically independent observations and obtain the least-squares estimator of ,61 . X1 and X2 are fixed (that is, they are not random variables). Then we fit the subset model y = u + lxl + .9 to the same n observations and obtain another least-squares estimator of ,31 . Discuss which least-squares estimator of ,61 is more biased for the parameter ,61 and which has smaller variance. State assumptions underlying your discussion