The same scientific study that medical marijuana activists are using to blast a marijuana-legalization initiative, it turns out, debunks their own leading argument.

Yesterday, Anthony Martinelli, the treasurer of a new campaign to oppose pot legalization on the fall ballot, claimed that Initiative 502 would subject sober pot smokers to DUI charges if they'd smoked pot the day or week before. Over the past few months, that talking point has become the primary line of attack to stop legalization in Washington State—and now their campaign is picking up donors. But at the time of our interview, Martinelli couldn't produce the study to validate his central argument.

Martinelli did send me that study later yesterday afternoon, and I checked it out, but it shows the opposite of what he said it did.

Here's the gist of their argument: Both Martinelli and the No On I-502 website cite Karschner et al to claim that the initiative is too restrictive for pot smokers who drive by establishing a cutoff of 5 nanograms of active THC per milliliter of blood. Any driver who exceeds that limit would be automatically guilty of DUI. The study "proves that users fail these tests six or seven days after using, and [I-502] takes away driving for all cannabis users," says Martinelli.

However, let's look at what Karschner et al actually says. Publishing their work in December 2009, the researchers measured levels of active THC, the psychoactive compound in marijuana, in 25 heavy, long-term marijuana users over a seven-day period. The report finds that only one of the participants had active THC levels above 5 ng/mL on "admission (day 1)," which suggests that the woman was high when she arrived, and the rest had levels lower than 5 ng/mL. The woman had reported smoking four blunts per day. However, despite Martinelli's claims, the day after smoking, her THC levels had dropped to 2.9 ng/mL. Every other participant's THC levels had dropped even lower by day two. By day six, the levels were undetectable in most of the participants—and all of them were far below 5 nanograms. (And for the record, this research concerns active THC, not inactive THC metabolites.)

In short: According to the study they cite, none of the heavy pot users showed THC levels above the cut-off that I-502 would establish for drivers, except for one person on the same day she was admitted. Meanwhile, other studies show drivers with more than 5ng/mL of THC have a higher risk for crashes.

As the anti-legalization campaign repeatedly points out, the science on THC and driving is minimal. That's true, and per se cutoffs are imperfect. I agree with them on these counts. But the science that does exist shows significant potential for impairment among drivers over 5ng/mL. "At concentrations between 5 and 10 ng/mL approximately 75-90% of the observations were indicative of significant impairment in every performance test," reported Ramaekers et al in 2006.

Likewise, Grotenhermen et al found in 2005: "The crash risk apparently begins to exceed that of sober drivers as THC concentrations in whole blood reach 5–10 ng/mL." But THC levels drop rapidly after smoking marijuana. "THC concentrations in serum typically drop to 1–4 ng/mL within 3–4 hours after consumption," the researchers reported (the concentrations in serum are equivalent of about 0.6 to 2.5 ng/mL of THC in blood, the report says, which is obviously even lower).

So here's what we can glean: The claim that THC will persist in high levels a day after people are high, even in chronic users, appears unfounded. Further, active THC levels above 5 ng/mL are associated with higher risk of accidents, but the THC levels tend to drop below that level within a few hours after smoking pot.

If there are other scientific studies that prove otherwise, I'm happy to see them. But this is the evidence they point to. And that evidence doesn't support the claims of No On I-502. That said, I agree with No on I-502 that cut offs that result in automatic DUIs are bad policy (I believe that defendants should retain a defense in court, despite THC levels). But it's a small price to pay to eliminate roughly 10,000 marijuana arrests each year in Washington State. And it's far more lenient than current state law: zero tolerance of any THC in the bloodstream for drivers.

The reason we're talking about this at all is because strict DUI provisions will help get this thing passed (and help stave off attacks about stoned driving that have helped defeat other pot initiatives, such as Prop 19 in California). A poll last May by Quinlan Rosner Research found that the DUI provision alone prompted 62 percent of voters to say they were more likely to support I-502, and only 11 percent said it would make them less likely to support the initiative.

So if we want to stop arresting 10,000 people for marijuana each year, we need to pass I-502 this fall. And we can ignore ginned up claims by medical marijuana activists that I-502 will ensnare drivers who are sober. For now, they seem to be fighting I-502 because it won't let them drive high.