A new study casts doubt on Uber’s claim that ride-sharing has reduced drunken driving.

Researchers at Oxford University and the University of Southern California who examined county-level data in the United States before and after the arrival of Uber and its competitors in those markets found that ride-sharing had no effect on drinking-related or holiday- and weekend-related fatalities.

One reason could be that, despite the soaring popularity of Uber and other ride-sharing services, there still may not be enough ride-share drivers available yet to make a dent on drunken driving, the authors said. They also suggest that the tipsy riders who now call Uber are the ones who formerly would have called a taxi. For others, the odds of getting a DUI are still so low that many would prefer to gamble rather than lay out money for a ride-sharing service. Drunks, after all, are just not rational.

“The takeaway for me is that there’s still tons of room for improvement when it comes to reducing drunk driving fatalities,” David Kirk, a co-author of the study, said Wednesday in an interview via Skype.

The new  study — which was published July 22 in the American Journal of Epidemiology — adds a new element to the debate over the merits of the popular and disruptive business model.

Uber pointed to other recent research that back up its contention that ride-sharing cuts down on drunken driving. A paper co-authored by researchers at Providence College and Stonehill College found reductions in DUI’s of as much as 62 percent, as well as declines in fatal accidents.

The company also said surveys show at least 80 percent of its riders have said that the service helped them avoid drinking and driving, and that its peak usage coincides with times when people are barhopping and partying.

“We’re glad Uber can provide an alternative to drunk driving and help people make more responsible choices,” Uber spokeswoman Brooke Anderson said in an email. “Our ridership numbers show that trips peak at times when people are more likely to be out drinking and 80% of riders says that Uber has helped them personally avoid drinking and driving.”

Drunk driving is the largest cause of traffic fatalities, claiming more 10,000 lives a year, according to the National Highway Traffic Safety Administration (NHTSA).

Uber, meanwhile, is now operating in 490 cities around the world, and its use as an on-call designated driver is part of its marketing appeal. The San Francisco-based company’s website claims that drunk driving-related crashes fell by 60 per month among drivers under 30 years old  in the California markets where UberX began operating, thereby preventing an estimated 1,800 crashes since July 2012. Uber also has partnered with Mothers Against Drunk Driving to reduce drunk driving. “A city with Uber has . . . fewer drunk drivers on the streets,” Uber has said.

But Kirk, a professorial fellow at Nuffield College and associate professor of sociology at the University of Oxford, and co-author Noli Brazil, a researcher at USC’s Spatial Sciences Institute, said people should be skeptical of such broad claims.

The authors analyzed the county by county traffic fatalities from 2005 to 2014 in the 100 most populated metropolitan areas, comparing those areas where Uber had arrived to those where ride-sharing had not, and then controlling for differences. The authors examined drunk driving-related, weekend- and holiday-specific crashes. They found no association with the number of subsequent traffic fatalities in those categories.

One factor may be the sheer size of the problem and still relatively low number of Uber drivers. Although approximately 450,000 people now drive for Uber, there are 210 million licensed drivers in the United States — and an estimated 4.2 million adults who drive impaired, the study says.

“Sure, there are over 1 million arrests for drunk driving a year in the United States.  . . .[B]ut that’s nothing compared to the number of people and the number of incidents  of drunk driving,” Kirk said. “In the grand scheme of the massive volume of drunk drivers on the road, it’s hard to foresee Uber making a dent, unless the growth continues like it has for several more years.”

Kirk said that the authors were somewhat surprised at the lack of research on Uber’s impact on drunken driving. After scouring the literature, the authors could find only one other academic study of Uber’s impact on drunken driving. That study — which was conducted by Brad Greenwood and Sunil Wattal at Temple University — examined data from California and found that the low-cost UberX service had “yielded a significant reduction” in traffic fatalities while the luxury UberBlack had not.

The growth of Uber and its competitors suggests that more research needs to be done.

“I’m not saying Uber might not have an effect in the future,” Kirk said.  “As of now, I think there are other opportunities to lower drunk driving.”

—This post has been updated.