Communities across California now find ride-sharing apps such as Uber and Lyft as common-place conveniences, and the companies like to tout their effectiveness in combatting drunk driving. However, whether or not these services are making a dent in drunk driving accidents and deaths depends upon which study is considered.

As Fortune explains, a study released in 2016 and published by the American Journal of Epidemiology found no correlation between a drop in traffic deaths related to alcohol in communities where ride-sharing is available. Researchers analyzed data in the 100 most populous metropolitan areas from both before and after the introduction of ride-sharing apps and looked at the number of road fatalities. The researchers, who were from the University of Southern California and Oxford University, found “no association” between a service’s availability and a decrease in road fatalities, whether they specified alcohol-related deaths or any traffic death.

However, as the New York Times explains, there are studies that have found evidence that ride-sharing reduces drunk driving. Uber, which was first introduced in San Francisco in 2010, commissioned a study in 2015 with Mothers Against Drunk Driving that found they gave the most rides at times when drunk driving accidents reach their peaks. The study also noted that there was a 10 percent drop in drunk driving arrests after Uber came to Seattle. Several independent studies seem to draw similar conclusions. One study of New York City found that alcohol-related accident fell by 25 to 35 percent since ride-sharing became available in 2011. The study included all boroughs except Staten Island, and it compared their traffic data to places where Uber and Lyft were not available.