1. 

4. Suppose an investigator assumes that the true model has no intercept (31 = 0). He is considering two estimators of the slope coefficient: the OLS estimator 62 = Zay/1 / Zy and another estimator 62 * = y /C. How different are the estimates from that found in (1)? In your opinion, which of the two estimators 62 * and 62 do you think is better, assuming the investigator is correct in her assumption that 1 = 0? Explain clearly