Does job training work? To measure the effectiveness of government training programs, it is usual to compare workers’ pay before and after training.

But many workers sign up for training when their pay drops or they are laid off. So the “before” pay is unusually low and the pay gain looks large.

(a) Is this bias or random error in measuring the effect of training on pay?

Why?

(b) How would you measure the success of training programs?

AppendixLO1