Using 11th grade subjects, two screening instruments designed to ident
ify youth with substance use problems were compared. The main results
were: 1) that the instruments gave different diagnoses for about 1 out
of 5 youth who are using alcohol or drugs; 2) that the probability of
being identified by these tests increased only among youth who were b
ecoming drunk multiple times each week; and 3) that these instruments
do not predict adolescent problems or substance use over time substant
ially better than a simpler measure of the frequency of substance use.