Cognitive Daily

There is little doubt that the cognitive demands of conversation can affect our awareness of the world around us. Everyone has a story of a near-miss collision with some clueless airhead driving who was jabbering away on the cell phone. A co-worker once tearfully told me of the time she was in an argument with her boyfriend while parked in his car at the side of the road. Furious, he got out of the car and slammed the door. He never noticed the passing car that hit him and instantly killed him. Was this a freak accident, or does conversation—and not just cell phone conversation—impair our ability to drive and assess the traffic around us?

Although there is a growing consensus that talking on cell phones—even hands-free phones—is a distraction that impairs driving ability (we’ve reported on one study by David Strayer and William Johnson confirming this notion), many researchers have suggested that in-person conversation may not have the same effect, because passengers can see the traffic patterns and slow the conversation when a difficult driving situation arises. A group led by Leo Gugerty designed two experiments to try to determine if car passengers adapted their conversation for tough driving situations.

Gugerty’s team used a simple driving simulator for their task (you can see it here—it’s more sophisticated than what Strayer and Johnson used, but still not exactly a realistic reproduction of real driving). In their first experiment, the team used the same task as Strayer and Johnson: the “passenger” gives the driver a word, then the driver must repeat a new word that begins with the last letter of the original word. But instead of simply navigating a path, drivers had to perform several tasks designed to replicate real driving, like remembering the locations of other vehicles, avoiding crashes, detecting hazards, or remembering when vehicles were in the car’s blind spot. In a second version of the task, designed to approximate talking on a hands-free phone, the conversants were placed in adjacent cubicles where they could not see the driving similator. Drivers were also tested with no conversation. To motivate them to try their best, driver-passenger teams were told that the best two teams would receive a $25 reward.

Gugerty et. al did not find a significant difference between driving performance with on-board passengers and remote conversants—in both cases, driving was significantly worse than with no conversation. What’s more, passengers did not slow the conversation during hazardous driving situations—in fact, the in-car conversation was faster than the remote conversation.

But perhaps the research participants simply felt that the verbal task was more important than the driving task, and so neglected the driving task. In a second experiment, the research team decided to try giving separate rewards for driving and conversing: each team could earn up to $3.75 for good driving, and up to $3 for good performance on the verbal task. They also made the verbal task more difficult for drivers: the passengers simply read words off a computer screen, and only the drivers had to generate new words. Since passengers never had to think of new words, the drivers would have to respond more frequently. As before, researchers found no difference between in-person and remote conversations: passengers did not slow their conversations to help the driver, even in difficult driving situations.

But with this second, more difficult verbal task, drivers performed even worse: here’s a graph comparing the first experiment (easy verbal task) with the second (hard verbal task):


The decrement of the harder verbal task was largest for crash avoidance, cars recalled, and reaction time—hardly insignificant aspects of driving ability.

The authors are careful to point out that their task is by no means an accurate model of real driving, or real conversation. Participants tended to get into a rhythm in the verbal tasks, and this rhythm seemed to guide the pace of talking much more than the driving situation. Perhaps real conversation, especially with real passengers, is more adaptive to driving situations.

On the other hand, as my co-worker’s tragic example demonstrates, sometimes conversations can seem more important than the driving situation. It’s not difficult to imagine other such conversations: what if a lawyer is negotiating a multi-million-dollar contract on a cell phone? Would she give precedence to that conversation, or the more mundane task of merging onto the interstate? Gugerty et al.’s research is a strong reminder that the substantive demands of conversation are a significant drain on cognitive resources. There are some situations where we may not be able to trust ourselves not to talk, even when our own lives are at stake.

Gugerty, L., Rakauskas, M., & Brooks, J. (2004). Effects of remote and in-person verbal interactions on verbalization rates and attention to dynamic spatial scenes. Accident Analysis and Prevention, 36(6), 1029-43.


  1. multitasking

  2. #2 Scott Reynen
    December 21, 2005

    My family has long played games during long trips. For example, we would try to find words on signs starting which each letter of the alphabet in order, or we would do anagrams. This makes me think maybe we shouldn’t allow the driver to play. But it seems like there’s also some risk in losing attention when driving becomes monotonous.

  3. #3 grubstreet
    December 21, 2005

    The design of this study seems a bit strange to me. Why not just tell people to have a conversation rather than play word games? And the passenger ought to be able to say things like “hey, look out for that bicycle!,” because that’s what real passengers do (and a crucial difference between a cell-phone conversation & an in-car conversation)

  4. #4 Dave Munger
    December 21, 2005

    Grubstreet, your objections make a lot of sense. However, a difficulty with that sort of conversation is that it’s impossible to control. Different drivers will converse differently, so how do we measure the effect of conversation? Arguably, though, if that objection can be made for the research, then one can also say that it’s impossible to make any sort of general claims about the relationship between conversation and driving.

    Even if we can’t model “real” conversation, I’d be interested in seeing some kind of research where the penalties for “crashing” the car were more severe. Just a $.75 difference between driving rewards and verbal rewards doesn’t seem like enough. How about $25 for perfect driving, with $5 penalties for every mistake?

  5. #5 Alex Barnett blog
    December 22, 2005

    Don’t talk and drive

    Don’t talk and drive.
    Well, don’t talk about stuff that’s hard to talk about, that’s the message that Dave…

  6. #6 Andy Affleck
    December 22, 2005

    I’ve always felt that conversations have a focus point which is roughly between the two speakers. When you are talking to somebody in the car, your attention is focused within the car. When you are speaking to somebody on the phone, that focus point is now somewhere outside of your immediate location. It is, in a sense, as if you have to mentally force yourself to be back in the car when driving gets tricky. I have never had an attention focus issue when talking to somebody in the car. In part, because they are also looking out, but mainly because my mind’s location is still rooted within the car. When I speak on a phone, I can feel myself not entirely in the here and now but partially elsewhere as if my mind cannot be both here and speak to somebody “out there.” This is very hard to explain and I would bet even harder to devise a study to detect. But it’s why I’ve always avoided speaking on the phone when driving.

  7. #7 Maggie's Farm
    December 28, 2005

    Wednesday Afternoon Links

    From Dan, EOD technician, on the reality of Muslim culture: WillismsThe Real Scandal, and the NYT Toys with Treason: NERepubDem chances in 2006 squandered by Moonbats: Capt’n EdHow does church-going make you wealthier? The EconomistHilary fund-raising…

New comments have been disabled.