STATISTICS HAPPEN

On the heels of a survey (http://www.nida.nih.gov/DrugPages/MTF.html)
conducted by the University of Michigan, the media has sent mixed or
incorrect messages to the public about teen use of cigarettes, alcohol, and
other drugs. While long-term trends in increases or decreases in smoking and
drug use can be measured with this survey, many changes from 2003 to 2004
were not statistically significant.

Despite this, the findings were reported - if not trumpeted - by the media
as if they had great meaning. Worse yet, by emphasizing which drugs are
being used by current drug users - and neglecting the larger context - the
public is misled about drug abuse by teens.

USA Today, for example, headlined its report
(http://www.usatoday.com/news/washington/2004-12-21-drug-study_x.htm) with
"Survey: More teens using Oxycontin."

Sure, the percentages of 12th-graders reporting having used Oxycontin went
from 4 percent in 2002 to 4.5 percent in 2003, and now to 5 percent in 2004.
The 1 percent increase from 2002 to 2004 is statistically significant.
However, highlighting this small-yet-noteworthy increase in Oxycontin use
over this two-year period (an increase that was not noted in 8th or 10th
grades) eclipses the brighter picture of drug use among teens: Since 2002,
the use of illicit drugs in the last year has declined by 2.2 percent among
12th graders.

The increase in Oxycontin use may be more accurately described as a gain in
popularity among drug users. In the same time period, the use of LSD any
time over the past year went down 1.3 percent, and the use of ecstasy went
down 3.4 percent among 12th graders. Perhaps those who have used Oxycontin
this year would have used LSD two years ago; these shifts in drug popularity
are not indicative of more teens becoming addicted to, or even trying,
illicit drugs.

Another source for confusion were the reports on smoking. The Associated
Press (http://www.sfexaminer.com/article/index.cfm/i/122204b_youth) heralded
"progress" in preventing teen smoking by leading with "Federal study finds
modest drop in teen drug use, smoking in 2004." Indeed, smoking has been
reduced significantly since the early 90s, but the changes from 2003 to 2004
are almost entirely insignificant. The survey asked students four relevant
questions about smoking:

- - had they ever smoked;

- - had they smoked in the past thirty days;

- - had they smoked in the past day, and;

had they smoked more than half a pack in the past day.

The change in percentage (from 2003 to 2004) among students answering all
four of these questions was statistically insignificant for 8th and 12th
graders. Among 10th graders, the changes in percentages were statistically
insignificant for two of the questions. Only the decline in ever having
smoked among 10th graders (from 43 percent to 40.7 percent) and the decline
in those 10th graders smoking half a pack or more a day (from 4.1 percent to
3.3 percent) were statistically significant, hardly justifying the AP claim
that "This was the eighth consecutive year that smoking rates among surveyed
teens dropped."