Perhaps the most careful effort to measure the effects of compensatory preschool education was the Perry Preschool Project begun in Ypsilanti, Michigan in 1962. Children, mostly three years old, were randomly assigned to treatment (58 children) and control (65 children) groups between 1962 and 1965. Children in the treatment group received two academic years of schooling before they entered the regular school system at about age five, while children in the control group did not. The project collected information on the children through age 19, an exceptionally long follow-up period. Using information generated by the study, analysts estimated that two years of preschool generated social net benefits (1988 dollars) of $13,124 at a discount rate of 5 percent. [For a more complete account, see W. Steven Barnett, “Benefits of Compensatory Preschool Education,” Journal of Human Resources 27 (2) 1992, 279-312.]
a. Before seeing results from the project, what would be your main methodological concern about such a long follow-up period? What data would you look at to see if the problem exists?
b. Benefit categories beyond the age of 19 included crime reduction, earnings increase, and reductions in welfare receipts. If you were designing the study, what data would you collect to help measure these benefits?