Impact of Maternal Smoking and Global Financial Crisis on Birth Weight

Investigation of Sample Representativeness

a) Any empirical study starts by reporting summary statistics of all covariates used inthe study in a table of descriptive statistics. This is what you did in assignment
1. We will now very simply investigate the impact of non-response on the sam?ple representativeness. Build a new table which reports the sample mean of dbwt,mager, boy, married, black, white, primary, secondary, tertiary and prenat for mothers who did not report a smoking habit (smoker==.) and those who have reported smoking habits (smoker==0 or smoker==1). Test whether statistical dif?ferences in mean covariates exist between of each group of mothers. Based on your results, do you think that not reporting smoking behaviour is purely random?
(b) Estimate a model to explore the association between maternal smoking during pregnancy and infant’s birthweight, controlling for mother’s age, educational attainment and the number of prenatal visits. Interpret the coefficient estimate on smoker.Note that several model specifications are possible. Chose your preferred one and motivate your choice.
(c) Estimate a model to explore the differential impact of maternal smoking on infant’sbirthweight between the three trimesters of pregnancy, controlling for mother’s age, educational attainment and the number of prenatal visits. Formally test whether the association between smoking and infant’s birthweight varies between pregnancytrimester. During which trimester of pregnancy does maternal smoking appear to
be most detrimental to an infant’s birthweight?
(d) Estimate a model to explore whether there is any differential impact of maternal smoking on birthweight between Blacks and Non-Hispanic whites. Formally test this hypothesis.
2. Birth-Weight and the Global Financial Crisis. For this question report all regression results in Table 2.
(a) Estimate by OLS a baseline model (Model 1) that measures the birthweight differences of Black babies compared to Whites. What is the birthweight gap? Is this gap statistically significant?
(b) Extend and estimate Model 1 controlling for the age of the mother with a quadratic form, her educational attainment, and whether the mother smoked during pregnancy(Model 2). How is the gap affected? Is the racial birthweight gap still statisticallysignificant? Briefly comment on your results.
(c) Create three dummy variables pre==1 if the child is born before the Global Financial crisis (0 otherwise), post==1 if the child is born after the financial crisis, and during==1 if the child is born during the financial crisis.
(d) Extend Model 2 to estimate the mean birthweight before and after the financial crisis, conditional on all covariates in Model 2. Test the null hypothesis, H0, that there is no difference in average birthweight between infants born before and after the Global Financial crisis (Model 3).
(e) Modify Model 3 to estimate the effect of the Global financial crisis on the average birthweight of black infants relative to whites.
(f) Test whether the Global financial crisis had a statistically significant impact on the black/white birthweight gap. Briefly comment on your results.
(g) Repeat (e) and (f) separately for i) mothers who only have completed “primary” education ii) mothers who have completed tertiary level of education. Briefly discuss your results.