The following is not part of the extended least squares assumptions for regression with a single regressor: A)var(u_i | X_i)= $\sigma _ { u } ^ { 2 }$ B)E(u_i | X_i)= 0. C)the conditional distribution of u_i given X_i is normal. D)var(u_i | X_i)= $\sigma _ { u , i } ^ { 2 }$

Question

Quizplus · Accepted Answer

The correct assumption for the extended least squares is that the variance of the error term given the regressors is constant ($\sigma _ { u } ^ { 2 }$), not varying across observations as implied by $\sigma _ { u , i } ^ { 2 }$. This implies homoscedasticity, not heteroscedasticity.

The Following Is Not Part of the Extended Least Squares σu2\sigma _ { u } ^ { 2 }σu2​

The Following Is Not Part of the Extended Least Squares $\sigma _ { u } ^ { 2 }$