Suppose we are interested in estimating the model: y = 0 + 1x1 + u. Suppose E(u|x1) = 0 and V ar(ux) = 2. We collected a random sample following the above model with size N. (a) Suppose we use only the first half observations to estimate the model using OLS. Assume that {xi: i = 1,....N/2} are not all the same. The estimators are noted as 0, 1. Are these estimators unbiased? Are they consistent? Briefly explain your reason. (b) Now suppose we use the full sample to estimate the model using OLS. Assume that {xi: i= 1.....N} are not all the same. The estimators are noted as "0, 1. Compared to 0,1, which of these estimators do you prefer? Explain why.