Question 5 (0.2 points) Suppose that the causal population regression model (the true model) is given by. where y is the wage, x/ is the individual's education (years of education), x2 is the individual's on-the-job training, and x3 is the individual's ability. Suppose that we know the true population parameters are given as: 50 = 15, 81= 0.2, 8,= 0.1, 8:= 1.0 You do not observe ability and run the following regression model: j=botbut boxte ( 2) where bo, by, and by are parameters and e is the error term. As we already discussed, the error term e in Equation (2) includes Bexg + a Suppose you got three observations (for illustration purpose, assume that this is the population data; i.e. you are only interested in these three individuals.). Years of Education (X1) Wage (Y) On-the-job Training (X2) Error Term (e) 15 15 30 10.5 10 20 10 -9 20 10 40 20 What is the sign of the bias in the estimate of , if you run regression without x3 ? [Blank 1] What is the sign of the bias in the estimate of 6, if you run regression without x3 ? [Blank 2] Note: Give an answer either positive or negative. Blank # 1 Blank # 2