Murder and education Example 13 found the regression line yn = -3.1 + 0.33x for all 51 observations on y = murder rate and x = percent with a college education.

a. Show that the predicted murder rates increase from 1.85 to 10.1 as percent with a college education increases from x = 15% to x = 40%, roughly the range of observed x values.

b. When the regression line is fitted only to the 50 states, yn = 8.0

- 0.14x. Show that the predicted murder rate decreases from 5.9 to 2.4 as percent with a college education increases from 15% to 40%.

c. D.C. has the highest value for x (38.3) and is an extreme outlier on y (41.8). Is it a regression outlier? Why?

d. What causes results to differ numerically according to whether D.C. is in the data set? Which line is more appropriate as a summary of the relationship? Why?