In which I serve up a big fattie for my self-perceived critics

There is an interesting paper that I just ran across which will possibly please a certain segment of my audience. You see, it provides a bit of a test of the hypothesis frequently bandied by my commenters that anti-drug messages backfire. That if you tell adolescents all sorts of bad things are going to happen to them if they try an illicit drug once, and it doesn’t happen, somehow you are actually encouraging them to try the drug again. This general area is an occasional interest of mine and you can read a few thoughts here, here, here, here and here. The paper itself is this one.
Skenderian JJ, Siegel JT, Crano WD, Alvaro EE, Lac A. Expectancy change and adolescents’ intentions to use marijuana. Psychol Addict Behav. 2008;22(4):563-569. [Free PubMed Central version]
This paper describes a secondary analysis of data collected under the National Survey of Parents and Youth which focuses on the efficacy of an anti-drug media campaign. This means that it is, necessarily, correlational in nature, not a prospective experiment*. The purpose of this secondary study was laid out as:

There are many possible reasons for [poor effect of anti-drug messages] including the possibility that the typical campaign often is designed to develop expectancies regarding marijuana use outcomes that may not be experienced by the initiate. Changes in expectancies regarding marijuana, and the effects of such changes on initiates’ intentions to continue use, are the focus of this investigation.

In short, if we deliver lies-to-children to adolescents, do we end up encouraging cannabis use?

The sample:

…consisted of 1,344 adolescents who ranged in age from 12 to 18 years in the 1st year of the research (M = 14.15, SD = 1.52). … All demographic characteristics were weighted to be nationally representative.

For the purposes of this paper, the sample from the parent survey was limited to those who had never tried cannabis at the time of the first interview. The experimental data were derived from interviews with the subjects, conducted at a one-year interval. Questions of interest included Intentions:

“How likely is it that you will use marijuana, even once or twice, over the next 12 months?”
“How likely is it that you will use marijuana nearly every month for the next 12 months?”

as well as questions about Expectancy:

“How likely is it that the following would happen to you if you used marijuana nearly every month over the next 12 months:” ( followed by the expectation outcome “Damage my brain.” …”Lose my friends’ respect” … “Have a good time with my friends.” ..”Be more creative and imaginative” ..”Be acting against my moral beliefs.” .. “Mess up my life,” “Do worse in school,” ..”Lose my ambition.”)

The meat of the statistical analysis focused on the changes in both Intention and Expectancy from the first to the second interview. Some 14.5% of the sample had tried cannabis at that point, a group of additional focus. The bottom line result was that changes in intention were correlated with changes in expectancy and that moreover this correlation was stronger in the subgroup which had tried cannabis. Table 3 has the best summary but note that they’ve done a weird transformation so that all correlations are positive. The shorthand is that perceiving cannabis as less detrimental was correlated with greater intention to use, no matter which of he eight expectancy questions you look at.
Now, before you get all het up over the obvious, go read the Discussion section. This is a correlational study, not an experiment. Nevertheless, the focus here is on change in attitudes between those that actually changed their behavior (started smoking cannabis) and those that did not. Is it because the initiators changed their perception of risk and therefore chose to smoke or because they smoked and found their prior expectation of harm to require adjustment? Doesn’t really matter in some sense, does it? It seems to me that getting that knowledge about risks as accurate as possible conveyed to the adolescents would minimize any change in expectancy, since there would be less room for experience to modify these expectancies.
I do think the authors go a bit off the rails with their “implications”. They acknowledge the essence in their Discussion

Attention should be focused on whether users attained user status prior to a change in expectancies or rather if a change in expectancies resulted in changes to user status.

but then go right on to end on the strong claim that:

Our results do indicate that the best chance for reducing future use might be to persuade adolescents that if they do initiate use, the harms they will experience at some point may prove worse than anticipated, or the expected benefits may not prove as positive. At a minimum, the results warn against overstating marijuana harms in prevention.

yeah, not really sure we know that yet folks.
I’ll finish up with the observation that I was alerted to this by a NIDA NewsScan dated September 11, 2009 (which strangely just showed up in my email box). Oh yes and we should review the Acknowledgments statement from the study:

This research was supported by National Institute on Drug Abuse Grant 5R01DA020879-02.

See, my legalize-eet friends? Is this a crack in the great right-wing prohibitionist conspiracy? Or is this bug, you know, actually a feature….
__
*he says, fruitlessly hoping to head off the usual idiotic correlation != causation triumphalism
Skenderian, J., Siegel, J., Crano, W., Alvaro, E., & Lac, A. (2008). Expectancy change and adolescents’ intentions to use marijuana. Psychology of Addictive Behaviors, 22 (4), 563-569 DOI: 10.1037/a0013020