Use the data in FERTIL2 to answer this question.

(i) Estimate the model children 5 b0 1 b1age 1 b2age2 1 b3educ 1 b4electric 1 b5urban 1 u and report the usual and heteroskedasticity-robust standard errors. Are the robust standard errors always bigger than the nonrobust ones?

(ii) Add the three religious dummy variables and test whether they are jointly significant. What are the p-values for the nonrobust and robust tests?

(iii) From the regression in part (ii), obtain the fitted values y^ and the residuals, u^. Regress u^

2 on y^, y^ 2 and test the joint significance of the two regressors. Conclude that heteroskedasticity is present in the equation for children.

(iv) Would you say the heteroskedasticity you found in part (iii) is practically important?