Does Data Matter?

In recent years there has been an increasing call for banning something called "distracted driving" - which seems to devolve always back to cell phone usage in the car.

I must admit that I am bothered when I see a drivers with  cellphones in their hands.   Indeed, when the talk about producing a ban to all but hands free usage began in California, I told the two vice presidents that worked for me that I would buy them a unit for their car so they could talk hands free.

But the Atlantic Cities magazine had an article yesterday which reproduced a chart originally created in a Center for Disease Control and Prevention report which used World Health Organization data to suggest that the US is awash in distracted drivers.   I put all those links in because from my point of view the writer in this unknown magazine jumps to a conclusion that is not warranted.

Here is a second chart from the CDC report.  This reports data on a phone for text and email messages while driving.   Notice that the US numbers are about the same for frequent tester/emailers as other countries.  We are at the high end for occasional texters/emailers as two other countries.  It is pretty clear to me that talking on a phone, hands free, is no more distracting  than perhaps switching a radio station, asking kids in the back seat to quiet down or even rolling down the window.

But the distracted driving paranoids want to ban all cellphone usage.   Indeed, if you listen to some of them they argue that safe driving requires one to not have passengers nor a radio.

The methodology (Highlighted below in an excerpt from the report) in the report is worth reading, from my perspective it calls into question whether the results between European and American habits are even accurate -

HealthStyles and EuroPNStyles are online surveys designed by Porter Novelli (Washington, DC), a worldwide social marketing and public relations firm, and conducted among persons aged ≥18 years to examine health-related attitudes and behaviors. The HealthStyles data analyzed in this study were collected in the 2011 fall HealthStyles survey, conducted in the United States during September 30–October 5, 2011. The fall HealthStyles survey was sent to a random sample of panelists who had completed the 2011 spring HealthStyles survey. The spring HealthStyles survey was drawn from a panel containing 50,000 persons randomly selected through probability-based sampling to be representative of the noninstitutionalized U.S. civilian population; 14,598 panelists were selected to participate in the spring HealthStyles survey, and 8,110 panelists completed the survey (response rate: 56%). The fall HealthStyles survey was sent to 5,315 of the persons who had completed the spring HealthStyles survey; 3,696 (70%) completed the fall HealthStyles survey. Respondents who completed the survey received reward points (worth approximately $10) and were eligible to win a prize through a monthly sweepstakes (prizes generally were worth less than $500). HealthStyles survey data were weighted to match U.S. Current Population Survey proportions for the following nine characteristics: sex, age, annual household income, race/ethnicity, household size, education, U.S. Census region, metro status (i.e., residence in a metropolitan statistical area [MSA] versus a non-MSA), and prior Internet access.

The EuroPNStyles survey was conducted in July 2011 in Belgium, France, Germany, the Netherlands, Portugal, Spain, and the UK. The sample was randomly drawn from Synovate's Global Opinion Panel, recruited via Synovate partnerships with select websites, portals, and Internet service providers in Belgium, France, Germany, the Netherlands, Spain, and the UK. In Portugal, the sample was randomly drawn from the Global Market Insite's Panel. Panelists were selected to match each country's census proportions for age and sex, and quotas were set to reach 1,700 adults in all countries except for Spain and Portugal, where quotas were set to 850 adults. The survey's response rate in 2011 was 34%, with 10,338 persons completing the survey. Respondents received reward points for completing the survey, and the final data were weighted by age and sex to match each country's census proportions.

I wonder how much this "research" actually cost the CDC to produce.  But even if this is a task for the CDC, the report is useless without some additional thinking about the conclusions.  For example, is there data that "distracted driving" incidents are lower in California (where hands free restrictions are enforced) than they are in states where there is no such ban?