Tuesday, June 27, 2017

Tell them to buy an ad

Q: What's the best way to get some free publicity for your business?*
A: Dangle a "study" in front of a local journalist!

DETROIT - Detroit is the best driving city in America, according to a new study.

QuoteWizard published a new study ranking the best and worst driving cities in America, ranking 75 cities.

The final rankings are sum of weighted means calculated from these parameters

  • Accidents
  • Speeding tickets
  • DUIs
  • Citations (running a red light, using a cellphone while driving, etc.) 
If you're wondering what that all means, you might be tempted to click through to the study itself:

After QuoteWizard compared the best and worst drivers by state, we wanted to know more about how drivers in America's 75 most populous city metro areas stack up against each other. Here's how we did it:

How do you statistically determine bad driving? We sampled incident stats from users of our website with over two million data points from 2016. To quantify over driver standards for comparison, we weighted incident counts for each city with its occurrence percentage. The final rankings are sum of weighted means** calculated from these parameters:

You don't know a lot more than you did before, but you do know one important thing: "Two million" is just a number with some zeroes until you know what it's two million of and how the two million were chosen. In this case, it's "data points" drawn in some undescribed fashion from "users of our website," though there's no indication of how those are related in turn to the "incident stats" that were apparently "sampled" in some undescribed way.

Should you draw any conclusions about whether Detroit is a good or bad "driving city" from the study? Given that the link from the anchor text complaining about Detroit's "oft-uninsured drivers" goes to a story about a fatal crash involving two unlicensed drivers at an intersection*** 50-plus miles from our snug little office in midtown, you make the call.

That's sort of the point. You don't need a course in statistics to ask what a crash in the wilds of Livingston County has to do with driver habits in Detroit. You just need to ask. And you don't need a course in statistics to ask what a writer means by "incident count," "city" and "occurrence percentage," not to mention why and how the means are weighted, or even why users of an insurance comparison website would be a good representation of a city where a huge proportion of drivers are uninsured. That's why we tell people in quant classes to write for their smart friends in the English department: Being able to explain your data is a good indication that you understand your data. Using numberish-sounding words that you don't explain is often strongly correlated with, well, bullshitting.

This isn't "fake news" in the 2016 sense; it's the old-school kind that has always gotten past enough gatekeepers to do its work. The traditional response is "tell them to buy an ad."

* Whose purpose seems to be summed up in this note: By clicking "Find Discounts & Check Rates" I provide my signature, expressly authorizing telemarketing calls from this website, our marketing and re-marketing partners, and up to eight insurance companies or their agents or partner companies at the phone number, including wireless numbers, and address provided, in order to deliver insurance quotes or to obtain additional information for such purpose, via live, pre-recorded or auto-dialed calls, text messages, or email for up to 180 days.** Yes, this would suggest that the TV station didn't read before copy-pasting.
*** One of the drivers was from Oakland County, which at least is in the three-county metro area, though it's not the county that contains Detroit.


Post a Comment

<< Home