Unintended Consequences of Being Smart
8th March 2010
Shocking news!
The Insurance Institute for Highway Safety released a study of vehicle crashes that compared states with laws banning cell phone usage while driving to states without such laws. Even if you don’t know much about history or time series statistics, if you just look at the figures in the report, you can see the Headline: the Laws Have NO Effect.
Everyone is baffled.
“Absolutely, we were surprised by these results,” says Adrian Lund, president of IIHS and HLDI.
“The study raises as many questions as it answers,” says GHSA executive director Barbara Harsha.
Before we get into the merits of the case, just think about the science and the persuasion behind all of this. We cannot randomly assign drivers to different conditions, most importantly here, cell phone usage, then observe what happens. The absence of control is fatal to the quality of the science and the inferences we can confidently draw. Yet, smart people persistently believe in the Observational Research Tooth Fairy and so we get Laws and Regulations, Nudges and Nags from well intentioned folks who say more than they really know; the sign of FauxItAll.
Worse still is the unscientific orientation of many researchers in Observational Research. Most strain to confirm a hypothesis and design data collection and analysis to find anything that supports the hypothesis and almost never actively pursue disconfirming evidence. There’s nothing wrong with entertaining alternative explanations, unless, of course, you already know the true answer and you’re just trying to convince the yokels.
Now, take uncontrolled Observational Research, a confirmation bias, and then add Small Effects and you’re ready for disappointment. You don’t need an Excel spreadsheet to tell you that traffic accidents are extremely rare events compared to the amount of total driving and further that people drive and talk A LOT and that accidents while driving and talking are also extremely rare events compared again to the total. Thus, people drive A LOT, but rarely have accidents; people drive and talk A LOT, but rarely have accidents. The difference between Driving+Phones versus just Driving is a Small Effect.
Thus, we have the commonplace Perfect Storm for failure. People using science as persuasion in their use of Observation, avoidance of contrary evidence and explanations, and those small effects. Now, put that persuasion to work in State legislatures and you’ve got Laws and Regulations that produce no effect.
There is no doubt that conversation produces cognitive load and hence distraction for drivers. Whether on a cell phone or just yackety-yak with a passenger, mere talk requires mental effort and capacity. It’s just not that important for vehicle crashes. There is a clear break in the scientific chain of effect between the obvious distraction and the actual wreck. People tried to draw a straight line from distraction to crashing when it is clear from these data that there are intervening steps and processes that mitigate the distraction. Nobody looked for those intervening steps.
It’s rather like the silly FDA and Food Police efforts with various kinds of warning labels – whether portion size, calorie count, any other information. They expect that Warning Labels like this will cause people to eat less and lose weight as if a Warning Label functions like a double-wrapped strip of duck tape over a hungry mouth. It doesn’t, it won’t, and it can’t.
But, if you’re doing FauxItAll science, nothing else matters.
Posted in Business, Government, Health, Science | Comments Off









