The question is as follow

3. Consider estimating the following multiple regression models with the same sample: 9 ='o+fl~"1+2-'392+m (la) y = 130 + E31131 + 13232 + 13313 + H (1b) 1 (a) If ,83 is positive, which model gives you a bigger R2? What if {33 is negative? Explain. (b) Make TWO recommendations to the data collection team if you want to im prove the precision of your OLS estimator of the coefcient on 31? (c) Suppose 1:3 = iog{a:2), does model (1b) violate any Gauss-Markov assump tions? If so1 which one? If not, why not