The evaluation was longitudinal in nature, with data from the first sample of controls (non-DARE participants) collected in 1991, data from the second sample of controls collected in 1996, and data from the experimental group (DARE participants) collected in 1996. Participants in the evaluation included 253 high school students in 1991 and 93 non-DARE participants and 49 DARE participants in 1996. Analysis of self-administered survey responses showed significant differences only between the 1991 sample and the entire 1996 sample for frequencies of both alcohol and drug use. Comparisons between 1996 controls and 1996 experimentals proved nonsignificant, as did all other comparisons among the three groups. As further evidence of the DARE program's limited utility, the DARE program did not result in significantly older ages of first time alcohol and drug use. Results suggest the factor of time between 1991 and 1996 may be the causal element in observed differences. The self-administered survey form is appended. 11 references and 2 tables