(a) They consider a range of birth outcomes for children: birth weight (in grams
ID: 3376426 • Letter: #
Question
(a) They consider a range of birth outcomes for children: birth weight (in grams), indicators variables for being less than 1000 g (extremely low birth weight) at birth, less than 1500 (very low birth weight), less than 2000 and less than 2500 g (low birth weight) at birth.1 Table A1 below gives descriptive statistics of these outcomes. The first and second column give the mean and standard deviation of the corresponding outcome for the treatment group before the reform. The third column shows the meen difference between treatment and control groups before the reform and the fourth column afer the reform b0 Descriptive statisdics for the eaentp and the control group before and aher the Birth weigh -04--63 P[-2500) What are your DiD estimates of the smoking ban effect on each of these 5 outcomes? (mas. 3 sentences) (b) To estimate their main results, the authors employ a regression framework. They estimate the following regression model where the W variable are additional cont tains a full list). iables (the foot i. What do Yt, Gi, and De correspond to in the present example? (max. 3 sentences) ii. Why do the authors estimate the effect of the policy with a regression instead of the simple calculations that you did before? (max. 2 sentences) ii. Which coefficient or combination of coefficients gives the causal effect? (c) Table 1 presents the estimates from estimating the regression model. Interpret the effect of Table 1500 s 2000g Birth weight 5452 4614 018 013 (04)Explanation / Answer
i. Here Gi is used for the treatment group. If the person is in the treatment group, it takes the value 1; otherwise 0.
Dt is used here for the difference in the period, before the reform and after the reform. Dt = 1 for the period after the reform; otherwise 0.
ii) We are using here a regression method because we could get all the possible combinations here in one equation. It would not have been possible by doing simple calculations at one time.
iii) In general, the combination of all the coefficients (Beta coefficient) of the control variables tell us the causal effect of these variables on the dependent variable y.
Related Questions
Navigate
Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.