Until a few months ago, Jason never thought marijuana was addictive. Even though he had been smoking weed every day for the past 10 years, he didn't think it was a compulsive habit. But then the 43-year-old Seattle resident decided to quit for health-related reasons. "For the first couple days, I didn't feel much different," he recalled. "The third day, I was, oddly enough, very hungry. I became very, very hungry, like I couldn't get enough food in." And that wasn't all. "I started having very extreme dreams, very detailed, increasingly intense, bizarre stories." He said he felt as if, without marijuana, "an essential nutrient was missing."

He endured these symptoms for three weeks until he finally decided to start smoking again. "I was actually surprised, the level of dependence," said Jason, who asked that his last name be withheld to protect his privacy. "I was kind of shocked and surprised that [the withdrawal symptoms] were so pronounced—more like a psych med than stopping coffee and being curmudgeonly and tired."

Is marijuana addictive? The answer varies widely depending on whom you ask. If you ask the federal government or drug-war proponents, the answer is a resounding "yes." If you ask legalization advocates or medical marijuana patients, the answer will likely be "no" or maybe "not really." How can there be such a discrepancy? What does science say about the matter?

The most commonly cited statistic regarding cannabis's addictiveness is 9 percent—that is, about 9 percent of people who try pot will at some point in their lives become addicted to it. But that's a highly disputed number.

The 9 percent statistic comes from a two-year study conducted from 1990 to 1992 of 8,098 participants ages 15 to 54 in the "noninstitutionalized civilian population" in the lower 48 states, the results of which were published in a 1994 report called "Comparative Epidemiology of Dependence on Tobacco, Alcohol, Controlled Substances, and Inhalants: Basic Findings from the National Comorbidity Survey." According to Roger Roffman, a professor emeritus of the University of Washington who has done extensive research on cannabis dependency, the 1994 study represents "the best epidemiological research that we have today." Which is why people continue to fight over it.

Researchers in that study defined "dependence" based on criteria laid out in the revised third edition of the Diagnostic and Statistical Manual of Mental Disorders (commonly known as DSM)—i.e., the bible of mental-health professionals. This particular edition of the DSM was published in 1987 and included seven criteria for the diagnosis of dependence on cannabis. If at least three criteria are experienced in a one-month period, the patient is believed to have cannabis dependence:

1. The substance was taken in larger amounts or over a longer period than the person intended

2. Persistent desire or one or more unsuccessful efforts to cut down or control substance use

3. A great deal of time getting the substance, taking the substance, or recovering from it

4. Frequent intoxication or withdrawal symptoms when expected to fulfill major role obligations at work, school, or home

5. Important social, occupational, or recreational activities given up or reduced because of substance use

6. Continued substance use despite knowledge of having a persistent or recurrent social, psychological, or physical problem that is caused or exacerbated by the use of the substance

7. Need for markedly increased amounts of the substance in order to achieve intoxication or desired effect.

Many people have criticized these criteria. For one thing, they appeared in a diagnostic manual for mental disorders. To some, this is an inherent bias in how marijuana use is viewed—as a mental disorder. And keep in mind this list was made when marijuana was illegal in all 50 states. As Dr. Sunil Kumar Aggarwal, a graduate of UW School of Medicine, has pointed out, both 3 and 5 on that list might not be an issue in a place where marijuana is legal.

Also, the criteria don't take into account pot's medicinal uses, which could explain symptoms such as taking large amounts. In other words, it's possible researchers might have unfairly or inaccurately labeled someone who's not addicted to pot as a pot addict.

And the study was partially funded by the National Institute on Drug Abuse, an organization that some argue has a financial stake in continuing the war on drugs.

Plus, the criteria used to determine addiction are now somewhat out of date. The DSM is now in its fifth edition and has shifted its terminology from "dependence" to "cannabis use disorder." It's defined as "a problematic pattern of cannabis use leading to clinically significant impairment or distress, as manifested by at least two of [11 criteria], occurring within a 12-month period."

Those 11 criteria include:

1. Using more or for a longer period than intended

2. Persistent desire to cut back

3. Lots of time spent using

4. Craving, strong urges to use

5. Use contributes to failures at work, school, home

6. Continued use despite recurrent interpersonal or social problems related to use

7. Use leads to giving up or reducing important social, occupational, or recreational activities

8. Recurrent use in hazardous situations

9. Continued use despite recurrent physical or psychological problems related to use

10. Tolerance

11. Withdrawal.

According to Roffman, "cannabis use disorder" is an umbrella concept that covers both addictive dependent use and nonaddictive abusive use of marijuana. In other words, many of the same criteria that were once considered symptoms of dependence are now part of a concept that also includes abusive—but not necessarily addictive—use. Plus, the DSM V acknowledges a broad range of severity levels: mild (two or three symptoms), moderate (four or five symptoms), and severe (six or more symptoms). It's a much more nuanced view of addiction than previously defined.

The question is, if that same survey in the early '90s had been completed today, would the 9 percent rate hold? More recent research suggests: "not quite."

The 2013 National Survey on Drug Use and Health, a study conducted by the federal government, found that an individual is more likely to become dependent on or to abuse cannabis the earlier he or she first tries it—11.5 percent of adults who started using at age 14 or younger become dependent on it, versus only 2.6 percent of those who try it when they're at least 18. Clearly, 2.6 percent is less than 9 percent, and even considering the younger demographic, an average of those two statistics doesn't quite equal 9 percent. However, dependence and abuse also appear to be linked to how much a person smokes: Among daily users, the National Survey on Drug Use and Health found that up to 50 percent can be expected to develop a problem with weed.

But once again, if you start to pick apart the data, you'll find there's more nuance to the story. The survey used criteria in the DSM IV to determine those who were dependent on or abusing pot. The criteria of dependence are the same as in the DSM III, but the survey makes an important distinction between "dependence" and "abuse"—dependence being a more severe problem because it involves "the psychological and physiological effects of tolerance and withdrawal," while abuse is determined by the presence of problems at work, home, and school, with family or friends, with physical danger, and/or trouble with the law. What they're calling "abuse" includes getting in trouble with the law, but getting in trouble with the law isn't evidence of "abusing" drugs necessarily. So the decision to lump together dependence and abuse rates might, therefore, be misleading.

In fact, a 2004 report, "Prevalence of Marijuana Use Disorders in the United States," looked at two large national surveys conducted 10 years apart and found that abuse was much more prevalent than dependence when looking at marijuana use disorder cases: Abuse represented approximately 75 to 80 percent of total marijuana use disorder cases. Using that percentage, one could estimate that about 2 percent, not 2.6 percent, of people who try pot when they're at least age 18 will become dependent on it, or about 8.6 percent of those who start at age 14 or younger. (The report used criteria from the DSM IV.)

(To confuse matters more, there is also a distinction made between physical dependence and addiction—addiction being the compulsive use of a drug despite harmful consequences, and physical dependence involving tolerance and withdrawal.)

Considering all this, let's say the 9 percent addictiveness claim is a bit shaky. There's still ample evidence that, for some people, marijuana can be addictive. It may not be as addictive as alcohol or have withdrawal symptoms as pronounced as heroin, but that doesn't mean it's not addictive at all. The evidence? Thousands of people seek treatment for marijuana addiction every year. According to the National Survey on Drug Use and Health, 845,000 people in the United States aged 12 or older reported receiving treatment for marijuana use in 2013.

Back when Roffman was first attempting to study marijuana dependence as a young researcher, anecdotal evidence showed him that addiction was a misunderstood and underreported problem. Although the common perception is that the federal government solely perpetuated mass hysteria around pot use ("reefer madness"), Roffman says there was actually a split in how marijuana was viewed by scientists and drug-war advocates. In fact, when he was beginning to apply for grants to study marijuana dependence and treatment, he was turned down due to the lack of scientific data that such a problem existed.

The existing data at the time, he said, was from "treatment agencies across the country where people being treated for compulsive marijuana use were almost always concurrently being treated for compulsive alcohol use or compulsive use of speed or depressants." To prove the topic of marijuana dependence was a worthy one, Roffman set up a quick study at the University of Washington, recruiting volunteers to man a phone line and developing a questionnaire. "Then we went into the media and asked for adults who smoked marijuana and had concerns or questions about their use to call for an anonymous interview. And once those calls started to come in, they came in droves."

The results of those 225 interviews were that "many of these individuals were people who believed that they needed to cut back or quit and couldn't on their own, they tried and they weren't succeeding, and very few of them were concurrently abusing alcohol or any other drugs," said Roffman. The results were convincing enough that the National Institute on Drug Abuse gave him funding to do his first counseling study of ways to help adults dependent on marijuana, back in the mid-1980s.

Marijuana addicts also have their own 12-step program, Marijuana Anonymous, which defines a marijuana addict not in medical terms but in self-definition: "Whether or not our addiction is psychological, physical, or both, matters little," states its website. "When it comes to the use of marijuana, we have lost the power of choice. It is strictly up to the individual to decide whether he or she feels addicted to marijuana."

Roffman, who himself identifies as a former marijuana addict, says part of the denial of marijuana's addictiveness is the image of addiction that's been perpetuated in the media. "It does not lend itself to the image of a junkie having a horrible, horrible experience," he said. "It is not the delirium tremors of the alcoholic. But there is a syndrome of withdrawal from marijuana. It's documented. It doesn't occur in the same level of intensity with all people who use marijuana heavily and then stop. And that makes it all the more complicated."

And there is a proven biological basis for dependence. The 2006 book Cannabis Dependence, which Roffman coedited, points to the fact that humans have a specific cannabinoid receptor in the brain and a "compound that activates the same receptor sites in the brain" as THC, called anandamide. "With greater understanding of the neurochemical basis of cannabis's reinforcing effects on brain systems, the reasons for the persistence of marijuana use have become more apparent," the book states.

Roffman says that acknowledging marijuana's addictiveness and its potential health effects is sometimes considered the enemy to legalization, but that such ideas do not need to be mutually exclusive. Roffman began his academic career as an activist and chair of the state chapter of NORML (National Organization for the Reform of Marijuana Laws), on the stump for decriminalization. But as his own addiction became more of a problem, he turned his eye toward dependence treatment.

While he was and continues to be a strong supporter of legalization, he believes the government needs to do a better job of educating the public about marijuana's potential health impacts. And he chafes at the knee-jerk reactions to discussions about marijuana addiction. "When you do talk about it, it's very likely that somebody is going to squash you, is going to put you down, is going to say, 'Marijuana is not addictive, you're a wuss. Marijuana is not addictive, you're just not using it right. Marijuana is not addictive, your wife is a (bad word)'... people who are absolutely certain that you can't become addicted to marijuana. I would agree with them if the science backed them up. But right now, all that's backing them is ideology and a lot of like-minded people on the web who are pounding their chests and saying, 'Marijuana is not addictive.' It is addictive."

In terms of his own withdrawal symptoms, Roffman says when he stopped using marijuana, he felt "pretty irritable" and "a lot of craving," which fall in line with the commonly accepted withdrawal symptoms of cannabis addiction: anxiety, irritability, restlessness, decreased appetite or weight loss, and sleep difficulties, including strange dreams. Less common symptoms include physical effects (chills, stomach pain, shakiness, and sweating) and a depressed mood.

There's also this: In the 2004 Marijuana Treatment Project Research Group, a sample of 450 chronic cannabis users seeking treatment as part of a multisite intervention trial, determined that almost all participants (96 percent) had unsuccessful attempts to quit or cut down, 95 percent said they continued to use marijuana despite recurrent psychological or physical problems, and 83 percent reported that large amounts of their time were spent using or recovering from marijuana use. Withdrawal symptoms occurred in 77.6 percent of the sample.

Roffman acknowledges that more studies need to be done to fully understand marijuana's impacts on health, but as long as marijuana continues to, in his words, "be held captive" in Schedule 1 of the federal Controlled Substances Act, scientists' ability to study marijuana will continue to be limited. While a major provision of Initiative 502, which legalized marijuana in Washington State, was the earmarking of revenue to fund more research, prevention, and education efforts on substance abuse, Republicans in the state legislature have threatened to redirect those funds to other services because they don't want to raise taxes.

Meanwhile, a proposal for a marijuana research license (which would allow applicants to grow marijuana for research purposes) could help pave the way for more scientific understanding of the drug. "We are only beginning to scrape the tip of the iceberg when it comes to marijuana research," Senator Jeanne Kohl-Welles wrote in an e-mailed statement. "Up until recently, the federal government's war on drugs has stigmatized marijuana and all but prohibited research into the plant's effect."

While acknowledging marijuana's potential for dependence, Roffman doesn't advocate for blanket abstinence. Rather, he seeks a more measured, informed approach toward cannabis use. "I think that the evidence is pretty abundant that people who use marijuana moderately—unless they're in a certain subgroup—are likely to be fine, and maybe even more than fine in terms of its enriching of their lives," he said, adding that he'd caution against use for those who are pregnant, have cardiovascular disease, or are vulnerable to schizophrenia.

Even regular, heavy users like Jason, the Seattle guy who recently tried to stop using marijuana, agree with him. "I don't think a substance that has the potential to be so mind-altering, behavior-altering, and consciousness-altering can be taken lightly, and I feel like it really is," he said.

Ultimately, however, Jason decided the benefits of marijuana outweighed the adverse impacts of not smoking, which is why he decided to start up again. "If this keeps me afloat, I'm down for it," he said. "There are far worse things to have to do."