For the Florida data file (shown partly on page 295), giving countywide data in Florida for several variables, a moderate positive correlation (r = 0.47) exists between crime rate and the percentage who are high school graduates. The percentage living in urban areas is also strongly correlated with crime rate (r = 0.68) and with high school graduation rate (r = 0.79).

(a) Explain why the association between crime rate and high school graduation rate could disappear, or even change direction, when we control for the percentage living in urban areas.

(b) Under the control in (a), if the association disappears, which type of relationship is more plausible—a spurious relationship or a chain relationship? Explain.