Answer the following questions, providing full details and discussion:
You are asked by the Government to investigate determinants of women's labour force participation.
CW2018.gdt file contains cross-sectional data of 50 states.
|
Variable name
|
Description
|
|
PARTICIPATION
|
Participation rate of all women (over age 16) in %
|
|
EDUCATION
|
Female high school graduates (over age 24) in %
|
|
MARRIAGE
|
|
|
Marriage rate of women (at least age 16) in %
|
|
|
You are expected to use the following linear models:
Model 1. PARTICIPATION = f(EDUCATION)
Model 2. PARTICIPATION = f(EDUCATION, MARRIAGE)
Both models should be estimated with a constant.
Present regression output for each of two models above.
Question 1. What is the interpretation of the coefficient for MARRIAGE in the Model 2?
Question 2. Are the signs of the EDUCATION and MARRIAGE estimated coefficients in the Model 2 in line with the expectations based on the theory? Justify your answer with some academic references.
Question 3. Is the Model 2 statistically significant?
This requires you to: state the null and alternative hypotheses, determine a critical value at 1% significance level, calculate the F statistic and present your decision.
Question 4. Can the MARRIAGE variable be deleted from the Model 2?
This requires you to: state the null and alternative hypotheses, determine a critical value at 5% significance level, calculate the F statistic and present your decision.
Question 5. Undertake the Ramsey's RESET test based on squares and cubes for the Model 2.
This requires you to: state the null and alternative hypotheses used, determine a critical value at the 10% significance level; calculate F statistic and present your decision.
Question 6. How could the model be improved? Illustrate your answer with reference to the relevant academic literature. 500 words maximum.