Question

We want to measure the impact of this magnificent casual/applied metrics class on your starting salary in Summer 2020 (right now, you are in the market and have some potential last round interviews). There are 30 of you enrolled in this class among the total 100 master students in the department who will graduate at the end of this semester. a) If we know everyone's salary, what are the sources of biases in this analysis? Come up with five potential confounders and explain their impacts Since you have rejected the hypothesis of the unbiased estimator, I forced to propose an instrument. I announced in the class) that whoever wants to participate the Angrist talk (which is prepared for the Econ faculty only), I can manage that for free (it is the end of the fiscal year cycle, and department has some extra cash to share with the enthusiastic students via the causal- master club). In the end, 25 of you showed up at the Talk, who got the average salary of 65K. Five of you didn't bother, and it turned out you had the better alternative since your starting salary was 2K more compared to your peers who wasted their time (based on your opinion). Six non-enrolled master students came as well with 66K salaries on average), since they overheard your discussion about the Talk, and I let them in since I had extra tickets, and didn't want to waste them, Pareto improving). The average salary of the remaining of the master cohort was recorded as 63K b) What is the Wald estimator? Calculate it. c) Could you reason the difference between 65K versus 67K salary? d) Was I proposing the valid IV? Why or why not Explain.

Answer 1

Solution: page no: 0 (a) Bias is disproportionate weight in favor of against an idea téh thing * Here usually in a way that is closed minded, prejudical (8") unfaisa. people may develop biases for cols against an individual , a group on a beliep. - Mainly, In Science and Engineering , abias Systematic Sverige (b) Here the Etimator is called the Wald Estimalor. Apten Wold (1940), the grouping Estimator * The Wald Estimată? Can also be obtained from the formula (9.45). Heu vouables are measured in derus for non-intercept model. (c) The main difference beliween 65k and 67k is the difference in Swrplus and deficit of particular poroduct * It also improves the average Salary of Employees. (d) When it comes to back froung here the value. added Senricus are more ciupõltant : X-