The Garden of the Blind

I just read an article that discusses yet another study showing that cell phone use while driving leads to accidents. This one actually gives some breakdown between accidents that occur while dialing and ones that occur while talking (although it doesn’t give numbers for both), and it provides numbers for accidents from other distractions.

The study has a number of problems, although it looks like it is better than most. This study videotaped 100 cars and their drivers for a year, and the results are compiled from examining the video. As such, they can analyze “events” that would normally not make it into police reports or other sources of data collection.

There are a number of issues to be had with the study. They found that “wireless devices” (presumably cell phones) were involved in 644 events, including 6 crashes. During most of these, including all 6 crashes, the driver was talking on the phone, rather than dialing, indicating that talking on the phone is more dangerous than dialing. Except that it doesn’t indicate that. We would need to consider how many events occurred relative to the amount of time a phone was in use. It is likely that less than 5% of the time a person spends on the phone is spent dialing (especially with speed dial), so if only 6 crashes occurred and they happened randomly and unrelated to cell phone use, it is quite reasonable that none of the 6 crashes occurred while dialing.

Similarly, what percentages of the crashes or other events occurred while using a cell phone, and how does that compare with the percentage of time people are on the phone? Merely having accidents occur while someone is on the phone does not mean that the accidents are phone related.

The study found that passenger distractions (noisy kids, conversations, etc.) were the second most common factor in events. This too must be considered relative to how much time passengers are interacting with drivers. I talk on the phone about 10 minutes a day while driving on average, but I only spend about 30-60 minutes a week talking to another person in the car with me, so it is reasonable to expect half again more events to occur while I was on the phone than while I was talking to a passenger.

My point is that people will look at this raw data and come to the incorrect conclusion that it is far more dangerous to talk on the phone than to dial a number or talk to a passenger in person, when the study might not give real evidence to that effect at all. If all events and accidents had occurred randomly (which is likely the case, since they typically involve interactions with other drivers who are not influenced by what is happening in your car), then the results mentioned in the article could very likely have been obtained in spite of no actual correlation (much less causation) between cell phone use and “events” or accidents. In any case, it is unlikely that talking on the phone while driving is any more dangerous than having a restless kid in the car with you.