The “science” of drug and alcohol prevention: the case of the randomized trial of the Life Skills Training program

Abstract

Changes requiring greater accountability among federal agencies in the United States, along with specific criticisms of prevention activities funded by agencies such as the Department of Education, have led to an increased emphasis on what are called “science-based” or “research-based” interventions in recent years. Federal agencies such as the Department of Education and National Institute on Drug Abuse (NIDA) have produced documents describing such interventions and advocating their widespread use and dissemination. The most widely advocated of these prevention interventions is the Life Skills Training (LST) program, the effectiveness of which, its supporters argue, has been demonstrated using rigorous research methods. The research study that has attracted most attention is the randomized trial conducted with white middle-class adolescents in New York State, as this purports to demonstrate that the LST program can reduce alcohol and illicit drug use 6 years after initial implementation. In contrast to the advocates for the LST program, I argue that this longitudinal trial does not meet the rigorous methodological standards claimed on its behalf. Indeed, it violates one of the fundamental principles of a randomized trial by restricting key analyses to selective sub-samples of the experimental group. I estimate that about 7.5% of those who initially received the LST intervention in the trial were included in the most recent set of analyses reported. This falls considerably short of the proportion of intervention group participants required at follow-up in a methodologically sound controlled trial. Thus, contrary to what its advocates claim, the study tells us little about the long-tern effectiveness of the LST program in reducing alcohol and illicit drug use among adolescents.