Use the data in MEAP00_01 to answer this question.
(i) Estimate the model
mathA = (0 + (2 lexppp + (2 lenroIl + (3 lunch + u
by OLS, and report the results in the usual form. Is each explanatory variable statistically significant at the 5% level?
(ii) Obtain the fitted values from the regression in part (i). What is the range of fitted values? How does it compare with the range of the actual data on mathA?
(iii) Obtain the residuals from the regression in part (i). What is the building code of the school that has the largest (positive) residual? Provide an interpretation of this residual.
(iv) Add quadratics of all explanatory variables to the equation, and test them for joint significance. Would you leave them in the model?
(v) Returning to the model in part (i), divide the dependent variable and each explanatory variable by its sample standard deviation, and rerun the regression. (Include an intercept unless you also first subtract the mean from each variable.) In terms of standard deviation units, which explanatory variable has the largest effect on the math pass rate?