The data set contains variables of mean hourly wage "lhwage" and the length of the education measured in years, "educ".
economics
Description
Translation:
1) In this assignment we work with a data set based on research of men in the age group 28-38 in the us in 1980.
The data set contains variables of mean hourly wage "lhwage" and the length of the education measured in years, "educ". A simple regression between the two variables gives the following:
1a) Interpret the estimated coefficients and review how accurate they estimate the coefficients, e.g. by illustrating with a confidence interval.
1b) Explain thorougly why we add "robus" in the regression command.
1c) We expand the model with another variable that measured the amount of years of working experience "exper" and the same variable squared, "exper2". Interpret the coefficient "educ" and explain why and how the coefficient changes compared to the corresponding coefficient in 1a)
1d) Why are the estimates of the effect of work experience so inaccurate? Explain how to test if work experience has an effekt on the hourly wage in the model from 1c)
1e) What is the effect of increasing work experience from 4 to 5 years, and from 9 to 10 years?
1f) We are interested in interpreting the effect of education on hourlyy wage causally, and we are worried that "educ" is not exogenous. We try to use mother's and fathers educational length ("meduc" and feduc") as instrumental variables for educational length. Explain your approach in the method you choose to use, as well as what kind of model we estimate and give an interpretation of the result.
1g) What role does the variables "exper" and "exper2" play in the analysis in 1f)? Should we be worried that neither of these are exogenous?
