Intepreting Empirical Evidence

Josh Wright —  1 February 2010

So, there is some new evidence that state laws banning cell phone usage does not reduce accidents (HT: Orin Kerr).  The Insurance Institute for Highway Safety study has gotten some attention in the media.  See, for example, this NYT piece discussing the researchers “surprise” that they did not find the result they were looking for.   The NYT piece also suggests that there are so many alternative forms of driver distraction that it is not likely such a ban will, on the margin, reduce accidents.  Even worse, a ban might provide additional incentives to engage in substitute forms of communication that could be more distracting.  Of course, the empirical evidence will be interpreted in two ways.  While critics of the bans will point to the study as evidence that the restrictions provide no benefit, proponents of the cell phone bans will point out that the study demonstrates that, in order to be effective, states must also ban other reasonable substitutes for cell phone distractions.

10 responses to Intepreting Empirical Evidence


    I’m doing research for a persuasive campaign about this topic. There definitely seems to be a problem; the NSC “estimates that 28% of crashes per year” have to do w/ talking or texting while driving. Definitely the physical distraction is an issue, but most studies highlight the cognitive distraction so banning hand-held devices isn’t going to solve the problem. A study done by the Harvard Center for Risk Analysis concluded that “use of cell phones by drivers may result in approximately 2,600 deaths” and 570,000 injuries in America/year. So I don’t think we could deny there is a danger. I’m wondering though, can we come up with an effective solution other than banning? I don’t like the idea of giving government more control, but we can’t ignore a problem that’s costing people their lives.


    Joey — Is the “communicate with . . . voices on the radio” to allow shouting at Rush Limbaugh?


    Um, what’s exogenous here? Instituting a ban on hand held phones might decrease crashes but also areas where such crashes are rising are more likely to institute these bans. This is not unlike the positive raw correlation between the number of cops and the amount of crime.


    Drafting clear statutory language is never such a simple task. I mean take the Sixth Amendment, seems pretty clear, but it’s been open for so much interpretation:


    Why is this so difficult? ‘No person operating a motor vehicle shall verbally communicate with any parties while driving except other passengers in the vehicle or voices on the radio.’

    25 words. Lawmakers can even throw in another provision that hand gestures to drivers or passengers of other vehicles is permissible, since that’s going to happen no matter what.


    Here’s what worries me: There have been several recent cases of people convicted of drunk driving for merely being inside a vehicle while intoxicated, even if the vehicle was parked. In one case the engine was off, the keys were on the console, and the car was in a parking space in front of the guy’s apartment. The court’s argument was that the driver could have operated the vehicle, in the future, and that was enough for conviction. A future crime, as in Minority Report. By this logic, I wonder if merely having a cell phone or texting device in one’s car will be grounds for arrest?


    Ah, unintended consequences. My favorite example of this (I’m riffing of the possibility that drivers might substitute more distracting forms of communication) comes from a few years back when there was an effort to require children traveling on planes to sit in car seats-which meant that for a lot of people the cost of traveling by air was going to go up substantially as they would have to buy a seat for each child, even children that otherwise would have traveled on a parent’s lap. And–here’s the fun part–at some point it was pointed out that the predictable consequence of this would be that more people would drive instead of fly, and whatever safety gains would be had from kids being in car seats on planes (and there were very few) would be swamped by the substantially increased risk from putting kids in cars instead of planes. Eventually the plan was scrapped.


    . . . or perhaps more effectively enforce the laws they have. In a jurisdiction with a cell-phone ban, the police are notoriously poor at enforcing the law.


    I stand corrected but unmoved. Feel free to insert the word “some” in front of cell phones in the first sentence if it makes you more comfortable. Or just use the word restriction if you like. The point of the post remains the same. I take it at least one point of the, lets call them “restrictions,” was to improve safety through reduced accidents. I take no stance on the accuracy or quality of the findings — and was merely pointing out that I do not think the will move the debate much because of very differing views on how to interpret such a finding, i.e. a finding of no effect will be seen simultaneously as demonstrating that less should be done and more should be done by different groups. Of course, both could be wrong. And both could be wrong to rely on the study at all! But notice that the post is about how the evidence will be interpreted not whether it is sufficient to justify those interpretations.



    Check your data, doctor.

    None of these laws banned the use of cell phones. Just the use of hand-held cell phones. Using cellphones with headphones etc. is allowed under all these laws. So these “natural experiments” tell us nothing about whether having phone conversations while driving increases the accident rate.