skills for effectively resisting these pressures alone or in combination with broader-based life-skills training.

Reviews published since 1990 have generally concurred with Botvin’s conclusions regarding the relative effectiveness of social skills approaches as opposed to information-only and affective approaches to classroom-based instructional programs. In addition, however, they have raised questions about the use of different modalities (other than classroom instruction), suggested that different content (other than social skills training) might also be effective in classroom instruction, and have suggested that the delivery mechanisms and methods, duration, and timing may be important moderator variables.

Modalities

Table 7.1 presents estimates of the magnitude of effects of different school-based prevention approaches. In addition to showing effect sizes for the major modalities listed above for which more than one study was available, it also shows a separate breakout for the Drug Abuse Resistance Education (D.A.R.E.) program, which is of special interest to policy makers because it is the most widely used classroom instructional program— in 1998 it was used in 48 percent of the nation’s elementary schools—and because it enjoys substantial federal support (Gottfredson et al., 2000). The table shows that (a) very few studies are available to assess the effects of modalities other than classroom instruction; (b) the range of average effect sizes observed from study to study is broad; and (c) the mean effect size for each category is in the small range, but this masks considerable variability and most categories include studies with negative as well as positive effects.

Certain of the modalities—counseling, social work and therapeutic interventions that do not use cognitive behavioral or behavioral methods; tutoring, mentoring, and other individual-attention strategies; and recreational, enrichment, and leisure activities—appear ineffective in reducing substance use because the average effect across all studies in these categories is negative and no study shows a positive effect. However, only a small number of studies have examined effects on drug use for these categories of activity.

The average effect size obtained from the 12 studies of D.A.R.E. for which effect sizes could be computed was 0.03—too small to be practically meaningful, but almost identical to the non-D.A.R.E. studies included in the category of skill-building classroom instructional programs that do not emphasize cognitive-behavioral methods.

Also noteworthy is the finding that the magnitude of the effects for changes to the school environment are generally larger than the magni-



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement