Why random drug testing doesn't reduce student drug use.

The first study, published in early 2003, looked at 76,000 students in eighth, 10th, and 12th grades in hundreds of schools, between the years 1998 and 2001. It was conducted by Ryoko Yamaguchi, Lloyd Johnston, and Patrick O'Malley out of the University of Michigan, which also produces Monitoring the Future, the university's highly regarded annual survey of student drug use, which is funded by the National Institute on Drug Abuse and whose numbers the White House regularly cites.

The early 2003 Michigan study compared the rates of drug use, as measured by Monitoring the Future, in schools that did some type of drug testing to schools that did not. The researchers controlled for various demographic differences and found across the board that drug testing was ineffective; there was no statistically significant difference in the number of users at a school that tested for drugs and a similar school that didn't.

The White House criticized the Michigan study for failing to look at the efficacy of random testing. So, Yamaguchi, Johnston, and O'Malley added the random element and ran their study again, this time adding data for the year 2002. The follow-up study, published later in 2003, tracked 94,000 middle- and high-school students. It reached the same results as its precursor. Even if drug testing is done randomly and without suspicion, it's not associated with a change in the number of students who use drugs in any category. The Michigan follow-up found one exception: In schools that randomly tested students, 12th-graders were more likely to smoke marijuana.

Results like these would mean budget cuts or death for some government programs. The White House has devised its own rating system, known as the Program Assessment Rating Tool, to help it cull failed initiatives. (These generally turn out to be the type of programs you wouldn't expect a Republican administration to like, but that's another story.) In 2002, PART deemed "ineffective" the Safe and Drug Free Schools State Grants program, the umbrella for school drug testing. The Office of Management and Budget, which runs the PART evaluations, writes on its Web site, "The program has failed to demonstrate effectiveness in reducing youth drug use, violence, and crime." The PART evaluation did not single out drug testing, which is a small part of the overall state grants program. Still, combined with the Michigan studies, what we have here is a bureaucratic pounding. That hasn't stopped President Bush from sounding an upbeat note. In his 2004 State of the Union, he said, "I proposed new funding to continue our aggressive, community-based strategy to reduce demand for illegal drugs. Drug testing in our schools has proven to be an effective part of this effort."