Exercise 11.11 showed a regression analysis for statewide data on y = violent crime rate, x1 = poverty rate, and x2 = percentage living in urban areas. When we add an interaction term, we get ˆy = 158.9 − 14.72x1 −

1.29x2 + 0.76x1x2.

(a) As the percentage living in urban areas increases, does the effect of poverty rate tend to increase or decrease?
Explain.

(b) Show how to interpret the prediction equation, by finding how it simplifies when x2 = 0, 50, and 100.