Consider the same dataset as in section 5.5. Repeat the analysis of section 5.5 using all observations rather than restricting the sample to ages between 21 and 65 years.
First, declare the survey design.
Second, compare the unweighted mean of hgb and its standard error, ignoring survey design, with the weighted mean and standard error allowing for all features of the survey design.
Third, do a similar comparison for least-squares regression of hgb on age and female.
Fourth, estimate this same regression using regress with pveights and cluster-robust standard errors, and compare with the survey results.