Nonprofit, nonpartisan journalism. Supported by readers.


DWI numbers that mean precisely nothing

Rarely do you see numbers butchered as badly as in the Minnesota Department of Public Safety’s announcement Wednesday that Hennepin County was the deadliest county in the state for alcohol-related deaths and injuries from 2004 to 2006. Ramsey County placed second, Anoka third, and Dakota fourth.

The Star Tribune and Pioneer Press reported the story on their websites Wednesday without even noting that the report did not adjust for each county’s population. This renders the rankings meaningless.

Late Wednesday night, and for the printed paper, the Star Tribune did at least note that the populous counties ranked higher, and it played down the rankings in the story, but the headline continued to say, “Hennepin tops list of worst counties for drunken driving.” The Pioneer Press story totally ignored the link to population.

It’s a good bet that Hennepin County also leads the state in cancer deaths, property taxes collected, number of obese residents and even fishing licenses. But Hennepin County also has about twice the population of the next-biggest county, Ramsey — which, lo and behold, finished as the second-most deadly county in the Public Safety Department report.

With my pocket calculator, I recalculated the safety department’s data to account for population — in other words, on a per-capita basis. I just divided the number of alcohol-related deaths and serious injuries over the three-year period by the population of the county, based on the 2000 Census.

DWI Chart

Look at these two lists. On the left is the state’s ranking of the most deadly counties. On the right, is the deadliness rank of those same 15 counties on a per capita basis, plus — in case you’re curious — the three metro-area counties that didn’t make the state’s top 15.

Nuances in the numbers
That comes close to turning the ranking on its head. Cass County is five times as deadly as Hennepin, once you adjust for population. That’s not strictly the same as saying a Hennepin resident has one-fifth the chance of dying or being seriously injured by a drunk driver, because the fatality and injury victims may not be residents of the county where the accident occurred. But it’s almost certainly true that the risk to a typical Hennepin resident is far lower than to a typical Cass County resident.

Often, an argument can be made that there’s a better factor to adjust for than population. For example, these deaths and serious injuries could be expressed as a percentage of miles driven in each county. This won’t align exactly with population, because in metro areas, people drive fewer miles per capita. But this is a nuance, and the choice can create a lively debate.

Still, comparing numbers of anything in Hennepin County with the numbers in counties that are 20 times smaller, and not adjusting for population or something similar, makes no sense.

With analysis like that, you could conclude that Russians are wealthier than Swedes, or that New Jersey is a more agricultural state than Iowa.

Joel Kramer is MinnPost CEO and editor.

Comments (4)

  1. Submitted by Christina Capecchi on 12/07/2007 - 09:08 am.

    Funny to see how dramatically the list changes when you put the numbers in like-terms. Thanks for doing this, Joel.

  2. Submitted by Derek Reise on 12/06/2007 - 03:34 pm.

    What gets me about this is that I bet it wasn’t hard to get these numbers. Cut and paste the two sources of numbers (DWIs and population) in a spreadsheet, then divide and sort.

    I know that reporters are increasingly spread thin. But it’s not like this calculation entails a major piece of investigative journalism. It’s something an intern or other lackey could be asked to do. (Do journalists have interns? If not, maybe they should; some real research could be done now and then.)

  3. Submitted by Craig McHenry on 12/07/2007 - 12:54 pm.

    Thanks for this story, Joel.

    This thoughtful analysis is exactly what I was hoping for from MinnPost, but what all of our media should be providing as well. Even when considering the most run-of-the-mill news release, journalists and their editors should ask, “What is it that this person or organization is trying to get me to say, is there something they are not telling me, or is there an important angle they are missing?”

    Curiosity and skepticism should be the top prerequisites for journalists… and news consumers as well.

  4. Submitted by Paul Linnee on 12/07/2007 - 01:52 pm.

    A couple of years ago, the MN DPS issued a release fortelling a “record increase in traffic fatalities”, unless Minnesota driver shaped up.

    As a former cop, who was on patrol in the early 1970’s, I remembered that annual death tolls were the 1,300 range in MN back then, and in the year referenced above, I think we might have been headed for something like 750 fatalities, or so.

    So I got suspicious, did a little independent research and proved my point — to myself — that not only were we NOT headed for a “record”, but in terms of fatalities per vehicle miles driven we were at an all-time low!

    I shared my view with the then “Information Officer” for the State DPS (Ms. Swanson) who, if I recall correctly, grudgingly accepted the critisism (and the implicit criticism of the media for running the release without checking any facts at all), but urged me to remember that ‘we are trying to change public behavior here’.

    So it’s OK to bellow, exaggerate and lie if your intent is pure? It is apparently also O.K. to report such behavior with nary a critical eye.

Leave a Reply