This paper compares two methods of analyzing aggregate data that is classified by period and age. Because there is a linear relationship among age, period, and cohort, it is not possible to distinguish the separate effects without employing an identifying assumption. The first method, which is...