Question: a. Consider the lifespan data discussed. Perform the regression analysis using all of the data points, and then delete those data points that have a standardized residual outside the interval (-3, 3), in addition to deleting those points that have a leverage value in excess of 2p/n. Continue doing this until there are no large standardized residuals and no large values.
What does a large leverage value mean in terms of what is being measured with this data set?
b. Consider Figure relative to the number of points that you eventually deleted because they had large leverage values. What does this suggest about using 2p/n as the cutoff for leverage values?
c. Consider the "data acquisition" for these data. Would you have any concerns about the quality of these data, as Shapiro and Kiichner had for the data analyzed?
