The results from the 2011 National Survey on Drug Use and Health (NSDUH), released this week, are not very different from the 2010 results. Reaching for evidence of success in the war on drugs,  the federal government is highlighting a 14 percent drop in the number of 18-to-25-year-olds who reported using prescription drugs for nonmedical purposes in the 30 days prior to the survey. USA Today notes that the change occurred "amid federal and state crackdowns on drug-seeking patients and over-prescribing doctors." Pamela Hyde, who runs the agency (the Substance Abuse and Mental Health Services Administration) that sponsors the survey, says "these findings show that national efforts to address the problem of prescription drug misuse may be beginning to bear fruit and we must continue to apply this pressure to drive down this and other forms of substance use." Here are three reasons to be skeptical of this argument:

1. If nonmedical use of prescription drugs had stayed the same or gone up, Hyde still would be arguing that her numbers show the need for continued "pressure to drive down this and other forms of substance use." That's the great thing about drug use surveys: No matter what the data are, they always reinforce the case for more taxpayer money.

2. The government tends to notice drug trends, such as the increase in methamphetamine use during the 1990s or the increase in nonmedical use of prescription painkillers during the first decade of this century, after they have already peaked. That may seem like a disadvantage from the perspective of someone who counts on the government to stop people from consuming politically incorrect chemicals, but it positions government officials to take credit for declines in drug use that would have happened with or without their belated interventions.

3. To the extent that the "pressure" championed by Hyde works, it does so by discouraging doctors from prescribing opioids, which may seem like a great idea to drug warriors but is apt to be perceived differently by patients suffering from severe chronic pain who rely on these medications to make their lives livable. Since there is no way to objectively verify pain, even the most diligent physician has to put a certain amount of trust in his patient when deciding what to prescribe. Efforts to discourage doctors from believing their patients may well frustrate some malingerers, but only at the cost of condemning others to avoidable agony. It is hard to see how this tradeoff, which sacrifices the welfare of legitimate patients for the sake of protecting fakers from their own recklessness, can be morally justified.

Drug czar Gil Kerlikowske provides another illustration of how any data can be used to support any point a prohibitionist wants to make. The NSDUH report says the share of 12-to-17-year-olds reporting past-month marijuana use in 2011 (7.9 percent) was "similar to the rates in 2009 and 2010" (7.4 percent in both years). Kerlikowske nevertheless latches on to the new number as an excuse to reiterate his argument that legalizing marijuana for medical use and talking about legalizing it for recreational use send "a bad message" to the youth of America, encouraging them to believe cannabis is not all that dangerous (which happens to be true, but never mind) and tempting ambivalent teenagers to try it. "Marijuana is still bad news," he tells USA Today. "I think [teenagers] are getting a bad message on marijuana. I think that the message that it's medicine and should be legalized is a bad message."

Is there any evidence this "bad message" translates into more marijuana consumption by teenagers? Kerlikowske seems to be fixated on a single data point: The NSDUH measured a statistically significant increase in past-month use by 12-to-17-year-olds between 2008 and 2009 (from 6.7 percent to 7.4 percent). The timing suggests Kerlikowske himself may be to blame, since he took over the Office of National Drug Control Policy in early 2009. Not surprisingly, that is not the explanation he favors, but it makes at least as much sense.

NSDUH numbers indicate that the share of teenagers reporting past-month use of marijuana was slightly lower last year than it was in 2002, the first year of the survey, despite all the intervening publicity attracted by the marijuana reform movement. The Monitoring the Future Study, which focuses on students, indicates that the share of seniors reporting past-month marijuana use rose from 21.9 percent in 1996, the year that California became the first state to legalize the medical use of marijuana, to 23.7 percent in 1997. But the rate immediately started falling, and it has never been that high since, even as 16 more states and the District of Columbia followed California's example. Furthermore, studies that compare marijuana consumption in states with and without medical marijuana laws find no impact on use by teenagers. And if public receptiveness to repealing marijuana prohibition encourages teenagers to smoke pot, what are we to make of the fact that past-month use by 12th-graders peaked at 37 percent in 1978, even though the percentage of Americans favoring legalization has almost doubled since then? Kerlikowske's persistent reliance on this argument, despite all the evidence against it, should be condemned for what it is: a crude, fact-free attempt to intimidate reformers by portraying their advocacy as an act of child endangerment.