Parametrizing age, period, and cohort effects: an application to United States delinquency rates, 1964-1973