Public Policy

Does “Just say no” work?

Some of the biggest public health problems involve the use of drugs in alcohol.  Individuals use drugs and alcohol because they receive some psychic benefit.  However, this has a cost to their own health and often the health of others (e.g., drunk driving, increased homicide rates).  Whether or not the government should be involved in convincing people not to drink, smoke or take drugs is one question.  Another is whether it actually effective.

A paper by Hornik et al. (AJPH 2008) examines the effect of the National Youth Anti-Drug Media Campaign between 1999 and 2004.  The paper finds that ad campaigns like “Soccer: My Anti-Drug” generally had no effect on drug or alcohol use.  However, the campaign did seem to slightly increase marijuana use.  

Why does advertising not work?  Most people already know about the health costs of drugs, alcohol and smoking.  Thus, advertisements do not serve to change the public perception.  Any government who wants to start a advertising campaign against trans fat in order to reduce obesity must contend with the fact that most people already know which food are and are not good for them.  Thus, this type of advertisement likely has little impact.

Should public health officials just give up on stopping people from using drugs and alcohol?  Maybe not entirely.

The Economist cites the health benefits of Mikhail Gorbachev’s anti-booze campaign.  “Mr Gorbachev’s anti-booze campaign—although hugely unpopular—raised life expectancy by fully three years between 1985 and 1987. After 1992 the state monopoly on alcohol (and health checks on its quality) collapsed. As anybody who lived in Russia at the time will recall, the effect was spectacular—and catastrophic. Death rates returned to their long-term trend.”

1 Comment

  1. I don’t have a subscription, so can’t see more than just the abstract. Could someone please explain their methodology?

    How did they identify the impact of exposure on drug use? Was exposure randomized somehow or do they exploit some quasi-experimental design? It seems to me that if you are just asking kids (i) if they saw the commercial and (ii) if they use drugs, you probably could get a POSITIVE correlation. (i.e., the commercial would catch the attention of kids that do/might use drugs; however, some “good” kids would simply ignore it since it’s irrelevant to their lives, and then not recall seeing the commercial several months later.) This is only one potential story you could tell; would these estimates show a bias?

Leave a Reply

Your email address will not be published. Required fields are marked *