A trial program is carried out to lower the recidivism rate of convicts (that is, the rate at which released prisoners commit additional crimes). The program selects men in county jails serving two years or less according to the first letter of their last name: prisoners with last names beginning with the letters A-M receive intense drug rehabilitation, psychotherapy (if needed) and help finding a job after their release. Prisoners with last names beginning with letters N-Z receive no additional aid.
1) How would you estimate the effectiveness of the treatment? Suppose you had information on age Could you use this information to improve your estimate? Please explain. Suppose you had information on the number of prior convictions. Could you use this information to improve your estimate? Please explain.
2) Would experimental effects (such as Hawthorne effects) be a problem in this program? If so, how? Alternatively, what threats to internal validity might exist in this program?