Use the data in MEAP00 to answer this question.

(i) Estimate the model math4 5 b0 1 b1lunch 1 b2log1enroll2 1 b3log1exppp2 1 u by OLS and obtain the usual standard errors and the fully robust standard errors. How do they generally compare?
(ii) Apply the special case of the White test for heteroskedasticity. What is the value of the F test?
What do you conclude?
(iii) Obtain g^i as the fitted values from the regression log1u^
2 i 2 on math4i , math42 i , where math4i are the OLS fitted values and the u^i are the OLS residuals. Let h^
i 5 exp1g^i 2. Use the h^
i to obtain WLS estimates. Are there big differences with the OLS coefficients?
(iv) Obtain the standard errors for WLS that allow misspecification of the variance function. Do these differ much from the usual WLS standard errors?
(v) For estimating the effect of spending on math4, does OLS or WLS appear to be more precise?