How Not to Collect Data

The local media yesterday was reporting on a survey conducted by AAA, 495 Express Lanes and Orange Cones No Phones on cell phone use and accidents in constructions zones. I was greeted with the following from WUSA:

According to a 495 Express Lanes and Orange Cones. No Phones survey of 402 law enforcement officers in Northern Virginia from March 21 to March 30, 2012, texting beats out speeding, aggressive driving and not obeying changing traffic patterns in terms of dangerous behavior.

Other media outlets have picked up the same story with the same theme that cell phone use in construction zone is hazardous.

But I have to say this has to be one of the worst ways to collect such data. Why conduct a survey of law enforcement officers to get this type of data? Shouldn’t the local and state police departments have actual records of accidents and their cause? Their data may not be complete as many accidents that do not involve personal injury never merit the attention of the police. But to rely on the memory of selected officers with no real way of measuring how many of the accidents they know about. In addition they may not even know the results of the final investigation as to the cause of the accident seems to be foolish. Many of them may even be just remembering the same horrific accident when they respond to the survey.

I liken this to asking the public what the weather was a on the second Tuesday last month and reporting the results as proof that it rained.

If you want good data go to a reliable source.