FILE More and more parole boards are using risk assessment tools w hen trying to determ ine an individual's likelihood of returning to crim e (The Boston Globe, February 2 0 , 2011). M ost o f th e se m odels are b ased on a range of character traits an d biographical facts a b o u t an individual. M any studies have found th a t older p eo p le are less likely to re-offend th a n y o u n g er ones.

In addition, on ce released on parole, w o m en are n o t likely to re-offend. A sociologist collects d ata on 20 individuals w h o w ere released o n parole tw o years ago. She no tes if h e /sh e com m itted a n o th e r crim e over th e last tw o years (crim e equals 1 if crim e com m itted, 0 otherw ise), th e individual's ag e at th e tim e of release, an d th e g e n d e r of th e individual (g en d er equals 1 if male, 0 otherw ise). The accom panying ta b le show s a portion of data;

th e full d ata se t can b e found on th e tex t w ebsite, labeled Parole.

Crime A ge G ender 1 25 1 0 42 1

⋮ ⋮ ⋮

0 30 1

a. Estimate th e linear probability m odel w here crime d ep en d s on age and gender.

b. Are th e results consistent w ith th e claims of other studies with respect to ag e and gender?

c. Predict th e probability of a 25-year-old m ale parolee com m itting an o th er crime; repeat th e prediction for a 25-year-old fem ale parolee.