Are Boston's cell-phone addicts "high-risk" drivers?
posted at 8/27/2012 7:38 AM EDT
Hiawatha Bray, an excellent reporter on computer technologies, stumbles at science. Recently he calls attention to a study of cell-phone use while driving, appearing in an online preview this month for the journal of the Association for the Advancement of Automotive Medicine. [ Nan Zhao, Bryan Reimer, Bruce Mehler, Lisa D'Ambrosio and Joseph Coughlin, Self-reported and observed risky driving behaviors among frequent and infrequent cell phone users, Accident Analysis & Prevention, preprint August 8, 2012, abstract at http://www.sciencedirect.com/science/article/pii/S0001457512002734 ]
We suspect that Mr. Bray did not inform readers about the source, cited above, lest they draw their own conclusions--likely to be contrary to his. He claims, clearly falsely, that the research shows people in Greater Boston who use cell phones more frequently while driving tend to be "high-risk" drivers. [ Hiawatha Bray, It's not the cellphone but the driver that's high-risk, Boston Globe, August 27, 2012, at http://www.boston.com/business/technology/2012/08/26/not-the-cellphone-but-the-driver-that-high-risk-not-the-cellphone-but-the-driver-that-high-risk/TrEoka1wNf6TfX8HtroajK/story.html ]
As knowledgable people often say about speculation: "case not proven." What the research shows instead, in words of those who performed it, is: "Individuals who reported frequently using cell phones while driving were found to drive faster, change lanes more frequently, spend more time in the left lane and engage in more instances of hard braking and high acceleration events." However, those behaviors would not necessarily make them "high-risk" drivers. More frequent risk-taking behaviors might be compensated by greater driving skills or less hazardous driving circumstances. [Zhao, et al., op. cit., abstract]
What the researchers would need to do to strengthen their case is to examine historical accident records and see whether measured and reported behaviors are significantly associated with greater accident rates or more serious accidents. They might then try factor analysis to distinguish between strengths of association of accidents and cell-phone use versus accidents and risk-taking behaviors. Their recently reported sample of 108 participants drawn from Greater Boston would be far too small to provide confidence in such statistics and likely would need to be 20 to 100 times greater.
This particular research is caught in the conundrum that confounds "soft science," such as nutrition. Potentially valuable topics may be suggested by small-scale, short-term research projects. However, the outcomes of such projects very rarely have enough strength to inform individual choices, let alone public policy. To develop pilot studies into genuinely significant results involves long-term commitments, which researchers may be unwilling to invest, and substantial costs, which often lack ready funding sources.