This Sunday, millions of Americans will sit down in front of their television or computer, crack open a few beers, and watch the Super Bowl. But if those viewers live in a housing project for the homeless, that booze could get them booted back out to the street. Many homeless housing projects have strict abstinence policies, and require residents to be completely sober. Permitting alcohol, many community organizers reason, would enable addictions and promote a downward spiral into continued drinking and declining health.

Now, a study suggests that requiring alcohol abstinence from residents of homeless projects might be misguided. In at least one project where residents were allowed to drink, alcohol consumption decreased, as did alcohol-related health problems, researchers reported  on January 19 in the American Journal of Public Health.

A total of 95 residents were tracked in the 1811 House—a housing project in Seattle. After two years, the residents' median number of drinks per day decreased from 28 to 17, about a 40 percent drop. And the number of alcohol-related health problems, such as delirium tremens, decreased as well. "We found that these people are human beings, and can moderate their drinking," says Susan Collins, at researcher at the University of Washington, and lead author on the study, "and that starts with getting housing."

The 1811 House is what experts call a "housing first"—it provides housing before asking residents to address their addictions, psychological disorders or other existing problems. The alternative, a "treatment first" model, tries to get homeless people clean and sober before offering them housing. Each day in the United States, there are about 640,000 homeless people, 17 percent of whom are considered “chronically homeless.”

Sam Tsembaris, a psychologist and one of the founders of the housing first movement and CEO of Pathways to Housing, is unsurprised by the results in Seattle. "It's very consistent with what we're finding," in other places such as New York City, Washington, D.C., and Philadelphia, he says. And Daniel Malone, a public health researcher and director of housing at the 1811 House and co-author of the study, says that other housing first projects are starting to see the same thing in Anchorage and Fairbanks, Alaska.

Project residents might drink less under lax alcohol rules for many reasons, says Collins. Street people tend to drink their alcohol faster that the average person for a number of reasons. Practically, they have nowhere to store things they buy. So while you or I might drink one or two beers and keep the rest of the six-pack in the fridge, homeless people might drink them all for lack of storage. And since they are never sure when booze might be stolen or confiscated by friends, enemies or the police, the homeless often finish drinks as quickly as they can.

Homeless people also often use alcohol to help them to fall asleep or to deal with the physical and psychological pain that can accompany homelessness. Housing can provide relief from many of these stresses, says Malone, which could lead to less drinking.

Studies have shown that housing first models also help cities save money, according to the National Alliance to End Homelessness. In New York City, a unit of housing cost about $17,277 a year, and saved nearly all of that—$16,282 in public costs such as jail time and trips to the emergency room. A program in Denver saved $15,773 each year per unit that cost $13,400.

Yet housing first projects can be unpopular among both charities and the public. The Seattle program from the study was dubbed "bunks for drunks" at some town hall meetings. "There is a lot of reluctance to give people who have behavioral problems publicly subsidized housing," says Nan Roman, president and CEO of the National Alliance to End Homelessness. Some of that comes from the belief that housing those with substance abuse problems will encourage them to keep using drugs and alcohol—a hypothesis that the Seattle study obviously calls into question. Others object for moral or religious reasons, Collins says.

And, as the Seattle study shows, residents don't stop drinking all together. Many are still addicted to alcohol, and while decreased consumption shows that providing housing won't send them into a tailspin of drinking, it might not get them sober either. And, as with any addict, relapse is always a risk. But, Tsembaris says, the process of overcoming alcoholism isn't easy. Just like non-homeless alcoholics, abstinence is a constant and difficult struggle. "It takes a while to improve mental health and addiction," Tsembaris says.

And a requirement for sobriety among housing projects might translate into putting the cart before the horse. "One of the things we know about homelessness is that it's very difficult to become sober when you're homeless," says Timothy Hilton, a sociologist at Northern Michigan University. "We know that sobriety takes several attempts for most people," he says, and the longer people are in housing the more likely they are to make it through both treatment and relapses.

In many ways, housing projects and homeless outreach is combination of art and science —and often the science lags behind social services [that constantly try out new ways to increase retention and recovery rates. “Why rely on intuition because we don’t have great data,” says Hilton. But science is still the best way to tell whether a new idea or approach is actually helping people or not, says Roman. "Just because someone tries something new, it may work, it may not work," she says, "I think the research is what ends up sorting that out."