Consider the HERS data. Suppose we want to test if SBP varies across race (the nonwhite variable). However, a simple t-test may not reflect the real effect of race. Therefore, we would like to test it after controlling the effect of other variables, such as exercise, smoking, and age. You may categorize age into three categories, [44 55), [55 65), 65 and above. Use these variables to create 3 2 2 = 12 groups, from each group draw random samples of nonwhite and white women. You may use 2 nonwhites and 2 whites from every group, form paired data, and carry out the paired t-test. Compare the paired t-test result with the regular t-test. Finally, on the paired data, consider a multiple linear regression of SBP on the nonwhite variable after adjusting the effect of age, smoking and exercise. Interpret the results and compare them with the previous results.