This is a report of fieldwork methodology and preliminary findings fro
m a controlled retrospective study of vulnerability to narcotic addict
ion comparing three groups of males closely matched by neighborhood, a
ge, and race: a community-wide sample of narcotic addicts; a nonaddict
ed control sample of age-eleven peer associates of the addicts; and a
nonaddicted community control sample of age-eleven peers who did not a
ssociate with the addicts. Fieldwork experience and preliminary findin
gs suggest a strong selective association between friendship and devia
nce among narcotic addicts and their age-eleven associates - a relatio
nship that is much less common in the two control groups, particularly
community controls. Association with older deviants was also more cha
racteristic of addicts than for the peer or community controls. Experi
ence in the study also attests to the feasibility and merit of capturi
ng historical data via retrospective, reconstructive sampling methods.