Let us consider the linear probability model y = #1a1 + X202 +e, where y is a binary variable. The matrix X = [ai : X2) is full column rank and E(e|X) = 0. We obtain a random sample {(yi, Til, 2), i =1,., n}. %3D Prove that the estimator derived for at remains consistent in the presence of heteroskedasticity.