DARE effectively prevents youth drug use.

The program arrived in the school on a Tuesday morning, as it always did, with a uniformed Los Angeles police officer and a box of workbooks. In 1983, Daryl Gates, the LAPD's autocratic chief, and the Los Angeles Unified School District had created Drug Abuse Resistance Education, DARE, as a classroom-based response to the city's crack cocaine epidemic. Officers would teach fifth and sixth graders to recognize drugs, resist peer pressure, and make healthy choices. The curriculum was sixty-five lessons long and cost the department almost nothing to deliver.

By 1987, DARE operated in thousands of school districts across the United States. Congress had earmarked federal law enforcement money specifically for drug education, making DARE the obvious recipient of those funds. President Reagan praised it. Nancy Reagan's "Just Say No" campaign gave it a national slogan. Principals liked having a uniformed officer in the building. Parents felt reassured. Local police departments got positive press. Everyone, in other words, had a reason to want DARE to be working.

No one had actually tested whether it was.

The first serious evaluations began in the late 1980s and early 1990s, as researchers applied the standard tools of program evaluation, control groups, longitudinal tracking, statistical analysis, to a program that had been scaled to serve 26 million children before anyone asked whether it worked. What they found was uncomfortable. A 1991 study by Dennis Rosenbaum at the University of Illinois at Chicago found that DARE graduates, compared to a control group two years later, showed no difference in drug use rates. The students knew more facts about drugs. But knowing facts about drugs and not using drugs, it turned out, were unrelated.

In 1994, Susan Ennett and colleagues at Research Triangle Institute published a meta-analysis in the American Journal of Public Health that pooled the results of eight methodologically rigorous DARE evaluations. The weighted mean effect size for drug use behavior was 0.06, a figure not statistically distinguishable from zero. More pointedly, Ennett found that interactive prevention programs, programs where students practiced refusal skills with peers rather than listening to an adult lecture, showed effect sizes four to seven times larger. DARE was not just small in effect. It was small compared to programs that were themselves modest in scope.

The federal government's response was cautious. The Department of Justice commissioned its own review. The Government Accountability Office, examining six long-term studies in 2003, reached the same conclusion: DARE's impact on illicit drug use was not statistically significant. That year, the U.S. Surgeon General's office placed DARE on its list of programs that scientific evidence had found to be ineffective.

The political reaction was revealing. Several states, including Utah and New Mexico, dropped DARE from their schools after the negative evaluations. Many others did not. DARE America, the nonprofit that ran the program, disputed the research, sponsored its own studies, and continued expanding internationally. The program remained in roughly 75 percent of U.S. school districts through the late 1990s, long after the scientific consensus had formed.

What made DARE so durable was precisely what made evaluating it inconvenient. The program was not only a curriculum; it was a symbol. It represented the belief that education and good intentions could substitute for structural interventions in drug policy. Questioning its effectiveness meant questioning that belief and, by implication, the political consensus built around it.

DARE eventually retooled its curriculum in 2009, shifting toward a model with more interactive components and less officer-led lecture. The new version, developed with input from Pennsylvania State University researchers, had modestly better evaluations. But by then, the underlying research question, does talking to children about drugs prevent them from using drugs, had accumulated decades of disappointing answers, regardless of how the talking was packaged.