This story was produced in collaboration with The Guardian as part of a partnership to report on new gun ownership data.

The share of Americans obtaining a gun without first undergoing a background check is dramatically lower than previous estimates, researchers have determined. The finding reshapes one of the most prominent assumptions of the U.S. gun debate.

Just 22 percent of current gun owners who acquired a firearm within the past two years reported doing so without a background check, according to a new national survey by public health researchers at Harvard and Northeastern universities shared in advance with The Trace and The Guardian.

For years, politicians and researchers have estimated that as many as 40 percent of gun sales are conducted without a background check — a statistic based on an extrapolation from a 1994 survey. The new survey, published in the Annals of Internal Medicine, found that the current proportion of gun transfers conducted without a background check is about half of the oft-cited figure.

One interpretation of the survey results is that efforts to boost screenings at the local level may be succeeding, even in the absence of federal legislation. Gun owners in states that require background checks on all private gun sales were much less likely to report acquiring a firearm without having first been subjected to a screening than those in states with no universal background check law.

The study’s authors hailed the new statistics as good news: “We’ve been moving in the right direction,” says Deborah Azrael, a researcher at the Harvard School of Public Health.

The federal background check system has put a stop to more than 2.4 million gun transactions since its implementation in 1994 — but checks are not required on sales between private parties, like many of those made at gun shows and arranged online. The expansion of background checks has been a key political battleground in the gun control conversation.

Since the 2012 massacre at the Sandy Hook elementary school, President Barack Obama and other Democrats have made the case for new gun laws by arguing that as many as 40 percent of guns in America are sold without a background check — a statistic criticized by gun groups as outdated and inaccurate.

The figure was revived again in the 2016 presidential election. In October 2015, Hillary Clinton earned “three pinocchios” from the Washington Post Fact Checker blog when she said during a campaign rally that “Forty percent of guns are sold at gun shows, online sales.”

Azrael hopes the new research can help inform better debate — and policy — concerning background checks. “It’s crazy that nobody has asked these questions since 1994,” she says. “I mean, should we be citing 20-year-old statistics in support of contemporary policy? Probably not, but the problem is that there has been no effort to maintain any kind of ongoing check on what has been happening.”

Philip Cook, a prominent gun violence researcher at Duke University who conducted the 1994 survey, says the new, smaller estimate does not undermine the argument that the U.S. needs a federal law instituting universal background checks on gun sales. In fact, he says, the finding that a smaller number of guns are acquired without background checks could be an advantage for supporters of stricter gun control laws.

“The headline is that we as a nation are closer to having a hundred percent of gun transactions with a background check than we might have thought,” says Cook. “So, it’s more attainable, and cheaper, to pass a universal requirement than it would be if 40 percent of transactions were still being conducted without these screenings.”

The National Rifle Association, which has called the 40 percent estimate a “lie,” did not immediately respond to requests for comment on the new background check statistic.

Gun control advocates say enforcement of the current background check system is as much a policy goal as expanding background checks.

“It’s not that the laws are ineffective, it’s that they’re so weak,” says Joseph Vince, a former agent for the Bureau of Alcohol, Tobacco and Firearms who currently heads the Criminal Justice programs at Mount St. Mary’s University. “With the legislation currently in place, it’s incredibly difficult for law enforcement to track down guns that emerge from the secondary market.”

Over the past two decades, 19 states have moved to regulate private gun transfers, including, as of January 1, Nevada. While research suggests that background checks can help reduce gun offenses, measuring the full impact is difficult given the myriad factors that can influence crime rates. Some studies also suggest that background checks can disrupt the flow of interstate gun trafficking.

The new survey also found that in states that had passed universal screening laws by July 1, 2013, just 26 percent of gun owners said they had obtained a gun through a private sale without a background check, compared to 57 percent of purchasers who live in states lacking such requirements.

Overall, researchers found that half of guns transferred privately in all states within the past two years were obtained without a background check.

While the share of gun owners who obtain firearms without being screened has shrunk, the updated survey results expose the holes that remain in the background check system. Among owners who purchased their most recent gun from a friend or acquaintance, 77 percent did so without a background check. If the purchase was made online, roughly 45 percent of respondents didn’t face screening.

“Compared to ‘94, there is this shift toward more people getting background checks,” says Matt Miller, a professor of epidemiology at Northeastern and co-director of the Harvard Injury Control Research Center. “But we still can’t lose sight of the fact that there are still millions of people every year who are getting guns, often from friends and acquaintances, without them.”

The new study shows that more Americans than ever are buying their firearms as opposed to inheriting or receiving them as gifts. Eighty percent of gun owners who obtained their most recent firearm inside the last two years purchased it, compared to 60 percent of people who obtained a firearm more than five years ago.

The new background check statistics come from the most rigorous survey of American gun ownership in more than a decade — but, like any self-reported survey data, it has limitations, including potential bias from respondents’ faulty memories. Because the survey was conducted online, the researchers wrote, the results may be less subject to the bias that affects surveys conducted over the phone, where people may be more likely to give live interviewers the answers they think are more socially acceptable.

The researchers asked 1,613 adult gun owners if they had undergone a background check for their most recently acquired firearm. The researchers also asked if the person who sold respondents the gun had asked to see a firearm license or permit before going through with the sale. If the answer to either question was yes, the respondent was listed as having gone through a background check. The team later asked the gun owners when they had acquired their most recent firearm.

The research suggests that the uptick in background checks may be a recent development. When the survey was conducted, in 2015, about 1 in 5 respondents who had acquired a firearm in the past two years said they had undergone a background check. But 57 percent of gun owners who reported acquiring their most recent gun before 2010 said they hadn’t undergone a check.

Because the survey relied on the memories of the participants, the researchers wrote, the more recent gun acquisition data might be more accurate.