Stata group Assignment 1 PDF

Preview

Summary

First Group STATA assignment. Code and results posted. Analysis of results and summaries provided. 3 man group. 3 assignments in total....

Description

STATA Group Assignment #1

Harsha Ramachandran - 170199110 Raphael Tang - 180859080 Michelle Wu Huang - 190878900 Xiangqi Meng - 173107760 Hongtao Cui - 191898950 Fares Abdallah Awad - 171510140 Jiang Chang - 170893890

1. Summarize the mean, standard deviation, minimum, and maximum value of the following variables: schooling_years, annual_income, family_income, age. Copy and paste the STATA output below.[1 mark]

2. Summarize the mean, standard deviation, minimum, and maximum value of the following variables for individuals above the age of 35: schooling_years, annual_income, family_income, age. Copy and paste the STATA output below. [1 mark]

3. Create the following scatterplots with a linear line of best fit included. Label the axes so that they are easily understood [1 mark for each graph] a) annual_income VS schooling_years

b) annual_income VS family_income

c) schooling_years VS family_income

4. Based on the scatter plots you have constructed, what is a possible omitted variable concern if you run a regression of annual income on schooling years? Is the bias positive or negative? [1 mark] The omitted variable family_income has a positive effect on annual_income as derived from plot (b). The variables schooling-years and family_income are observed to be positively correlated from plot (c). This indicates that the result of omitting family_income from the regression causes positive bias.

5. Estimate a regression with annual income as the dependent variable and schooling years as the explanatory variable.

a) Report the coefficient on schooling years and the standard error of the estimated coefficient below. [1 mark] The coefficient of schooling years is 1845.892, with an estimated standard error of 175.6969. b) Is the estimated coefficient statistically significant? How can you tell? [1 mark] The estimated coefficient is statistically significant as the p-value is less than 0.05, and the t statistic 10.51 is well above 2. c) Consider a one standard deviation decrease in schooling years. What is the associated change in annual income from this regression? [1 mark] A one standard deviation decrease in schooling years or (- 3.312358) multiplied by the estimated coefficient of the regression being (1845.892) leads to a 6127.8483 decrease in annual income.

6. Add indicator variables for the province of residence of the individual to the regression estimated in question 5. A.

Do the

added province indicators explain a lot of the variation in the change in annual income? How do you know? [1 mark] The added indicators do not explain a lot of the variation in the change in annual income. This is because the adjusted R^2 does not change by significant amounts.

B. Does the coefficient on schooling years change much when the province indicators are added? Does this mean that including province fixed effects can mitigate some omitted variable bias concerns? [1 mark] The coefficient does not change by a significant amount upon the addition of the indicator variable. This means that the fixed effects do not mitigate omitted variable bias concerns in this case. 7. Add family income as an additional regressor to the regression estimated in question 5. A.

Report the coefficient on schooling years and the standard error of the coefficient below. [1 mark] Coefficient - 1378.63 Standard Error - 151.6446

B. Does including family income as an additional regressor mitigate some of the omitted variable bias? [1 mark] Yes, adding an additional independent variable like the family income as an additional regressor to the regression will mitigate the omitted variable bias downward. This is shown by the significant change in the coefficient and the increase in the adjusted R-squared.

8. Construct an interaction variable between schooling years and age. Now run a regression with annual income as the dependent variable. The explanatory variables will be schooling years, age and the interaction variable you have constructed. A. Report the coefficient on the interaction variable and the associated standard error. How will you interpret this coefficient? [1 mark]

Coefficient - 14.467 Standard error - 57.605 The coefficient is positive, which means that if schooling years and age increase together at the same time, there will be a positive effect on annual income. This means that schooling will become more and more effective as people get older. 9. Is heteroskedasticity a concern when you are running a regression with annual income as the dependent variable, and schooling years as the explanatory variable? How can you tell? [1 mark]

Yes, heteroskedasticity is a concern when you are running a regression with annual income as the dependent variable, and schooling years as the explanatory variable. For homoskedasticity, the distribution of points on each x is relatively consistent. On the other hand, the distribution of points on different x for heteroskedasticity is relatively inconsistent. When x is large, the points are more dispersed....