a. Consider the regression using three explanatory variables, FERTILITY, PUBLICEDUCATION, and LNHEALTH that you did in Exercise
Question:
a. Consider the regression using three explanatory variables, FERTILITY, PUBLICEDUCATION, and LNHEALTH that you did in Exercise 3.3.6. Test whether PUBLICEDUCATION and LNHEALTH are jointly statistically significant, using a partial \(F\)-test. State your null and alternative hypotheses, decision-making criterion, and decision-making rules. (Hint: use the coefficient of determination form for calculating the test statistic.) Provide an approximate \(p\)-value for the test.
b. We now introduce the REGION variable, summarized in Table 4.11. A box plot of life expectancies versus REGION is given in Figure of life expectancies versus REGION is given in Figure 4.8.
Describe what we learn from the table and box plot about the effect of REGION on LIFEEXP.
Table 4.11.
Figure 4.8.
c. Fit a regression model using only the factor REGION. Is REGION a statistically significant determinant of LIFEEXP? State your null and alternative hypotheses, decision-making criterion, and decision-making rules.
d. Fit a regression model using three explanatory variables, FERTILITY, PUBLICEDUCATION, and LNHEALTH, as well as the categorical variable REGION.
d(i). You are examining a country that is not in the sample with values FERTILITY \(=2.0\), PUBLICEDUCATION \(=5.0\), and \(\mathrm{LNHEALTH}=1.0\). Produce two predicted life expectancy values by assuming that the country is from (1) an Arab state and (2) sub-Saharan Africa.
d(ii). Provide a \(95 \%\) confidence interval for the difference in life expectancies between an Arab state and a country from subSaharan Africa.
d(iii). Provide the (usual ordinary least squares) point estimate for the difference in life expectancies between a country from subSaharan Africa and a high-income OECD country.
Step by Step Answer:
Regression Modeling With Actuarial And Financial Applications
ISBN: 9780521135962
1st Edition
Authors: Edward W. Frees