Let Y denote the hourly wage of an individual in the U.S. Let Y A be the wage of an individual without a college degree in the U.S., and Y B the wage of an individual with a college degree. 

a) How are means E(Y A) and E(Y B) more helpful for description than only E(Y )? 

b) How could E(Y A) and E(Y B) be used to make better predictions than only E(Y )? 

c) Why can't we interpret E(Y B) − E(Y A) as the causal effect of a college degree on wage? Hint: what other factors might make E(Y B)−E(Y A) large, even if the effect of a college degree itself is small?