So-called "distracted driving" has become a big public health issue in recent years. The majority of U.S. states now ban texting behind the wheel, while a handful prohibit drivers from using handheld cellphones at all (though many more ban "novice" drivers from doing so).
But studies have reached different conclusions about how much of an added crash risk there is with cellphone use.
In the new report, Richard A. Young of Wayne State University School of Medicine in Detroit finds that two influential studies on the subject might have overestimated the risk.
The problem has to do with the studies' methods, according to Young. Both studies a 1997 study from Canada, and one done in Australia in 2005 were "case-crossover" studies.
The researchers recruited people who had been in a crash, and then used their billing records to compare their cellphone use around the time of the crash with their cell use during the same time period the week before (called a "control window").
But the issue with that, Young writes in the journal Epidemiology, is that people may not have been driving during that entire control window.
Such "part-time" driving, he says, would necessarily cut the odds of having a crash (and possibly reduce people's cell use) during the control window and make it seem like cellphone use is a bigger crash risk than it is.
The two studies in question asked people whether they had been driving during the control windows, but they did not account for part-time driving, Young says.
So for his study, Young used GPS data to track day-to-day driving consistency for 439 drivers over 100 days.
He grouped the days into pairs: day one was akin to the "control" days used in the earlier studies, and day two was akin to the "crash" day.
Overall, Young found, there was little consistency between the two days when it came to driving time.
When he looked at all control windows where a person did some driving, the total amount of time on the road was about one-fourth of what it was during the person's "crash" day.
If that information were applied to the two earlier studies, Young estimates, the crash risk tied to cellphone use would have been statistically insignificant.
That's far lower than the studies' original conclusions: that cellphone use while driving raises the risk of crashing four-fold.
And, Young says, the results might help explain why some other studies have not linked cell use to an increased crash risk.
A researcher not involved in the work said that the two earlier studies may well have overstated the crash risk from using a cellphone.
But that doesn't mean you should feel free to chat and text away at the wheel, according to Fernando Wilson, an assistant professor at the University of North Texas Health Science Center in Fort Worth.
A number of other studies, using designs other than case-crossover, have suggested that cellphone use, and particularly texting, is hazardous on the road, Wilson told Reuters Health.
"In wider policy, I don't think this study is going to change the conversation about distracted driving," Wilson said. "Most of the conventional thinking is that we need to do something to reduce it."
In his own study published last year, Wilson looked at information from a government database that tracks deaths on U.S. public roads. He found that after declining between 1999 and 2005, deaths blamed on distracted driving rose 28 percent between 2005 and 2008.
And the increase seemed to be related to a sharp rise in texting. ("Distracted driving" refers to anything that takes the driver's attention off the road, from fiddling with the radio to talking to other people in the car.)
Other studies, Wilson noted, have used mounted cameras to show that drivers' behavior becomes more risky when they are using cellphones.
All of those studies have limitations, and cannot pinpoint just how big a risk driving-while-texting (or talking) might be. Wilson said the current study highlights a limitation in case-crossover studies.
But the new study, itself, has shortcomings. Applying the GPS findings from this study to the two earlier ones, done with different drivers, in different countries, is tricky, both Young and Wilson point out.
"It's possible that the (earlier) study findings were overstated," Wilson said, "but it's difficult to know by how much."
According to the National Highway Traffic Safety Administration, about 450,000 Americans were injured in crashes linked to distracted driving in 2009. Another 5,500 were killed.