Press Box

Driving Mr. Whitebread

The screwy data behind the Wall Street Journal’s “Me & My Car” feature.

As a car guy, I immediately took to the Wall Street Journal’s “Me & My Car” feature when it debuted 35 weeks ago in the “Personal Journal” section. Each week, “Me & My Car” takes a new car, charts the personal interests, age, sex, and marital status of the average driver, and stuffs it into a USA Today-like info box. It’s cool to know that Cameron Diaz drives a Toyota Prius, Lemony Snicket a Volkswagen Passat, and Katie Couric a Ford Thunderbird, or that the Dodge Viper sells well in New York but horribly in Boston.

But even though “Me & My Car” surveys an incredibly wide range of cars—everything from Bentleys to Bugs—I noticed after a few months that the average driver profiles were puddling into the meaningless demographic center. Take, for example, the “Me & My Car” box for last week’s featured car, the Mitsubishi Diamante, excerpted below.

Mitsubishi Diamante
Price: $25,687-$28,447 Average Driver Profile
Sex: Male
Age: 47
Married: Yes
Driver Interests: electronics, gourmet foods, golf, tennis, wine

As it turns out, the average Diamante driver is also the average “Me & My Car” driver. Most of them enjoy golf, tennis, or wine. Of the 35 cars featured so far in “Me & My Car,” 31 of their average drivers are men. Likewise, 32 of the average drivers are married. And most of the average drivers scrunched up in the 44-to-47 age bracket. (The oldest average driver is 52 and the youngest is 43.) I’m sure that if it wasn’t socially unacceptable to put it into print, the Journal could report that the average driver of almost every car is white.

How to explain? Let’s put it this way: If the Journal applied a similar methodology to a weekly feature called “Me & My Crackers,” they’d end up telling us that the average consumer of Premium Saltines is a married person who is 45. And when they canvassed purchasers of Air Crisps, Triscuit, and Chicken in a Biskit, they’d probably come up with the same data point. In other words, the average consumer of almost everything is the average, middle-aged married person. That’s news?

“Personal Journal” editor Eben Shapiro says via e-mail that the Journal gets its “Me & My Car” data for free from R.L. Polk & Co., which calls itself “a premier provider of automotive information solution services.” He confesses that “Me & My Car” ‘s average driver skews so heavily to married guys in their mid-40s because Polk draws its data from state motor vehicle registration records, and car registration seems to be a traditionally male task. “Regardless of who the primary driver is, new cars are typically registered under the male’s name,” he notes. But just because they’re the average registrant doesn’t mean the Journal (or Polk) should call them the average driver. We’re on much safer ground when Polk says the average Subaru Forester driver is female, but shouldn’t the Journal let its hair down with a sexual orientation category so it can come completely honest on the Forester’s demographic? Shapiro also explains that the domination of drivers from the 44-to-47 age group is a function of the demographic bulge. Another explanation for the bulge could be that dads register their teens’ cars, which means “Me & My Car” seriously underreports teen drivers.

Polk didn’t give Shapiro a good explanation for why it alleges most average drivers are married—other than to say most people are married and lots of cars require a two-income family to purchase. (Presumably, marital status isn’t part of car registration.) If Polk can only guess that most drivers are married, on what shaky grounds does it report—and the Journal accept—that the average driver of a Volkswagen Beetle, Audi TT, and Toyota Prius is single? Does the premier provider of automotive information solution services just make things up?

The last info-bits dispensed in “Me & My Car” are no less spurious. The column’s 35 “average drivers” display a remarkable convergence of interests: 17 of them name foreign travel as a hobby or interest; 16 name tennis; 14 name wine; 10 name skiing; and nine name golf. How does Polk conjure these enthusiasms into existence? The organization told Shapiro the data is “self-reported” from production registration forms filled out by buyers of all sorts of products, including toasters and stereos. “Polk says they have a data base of 1/3 of the U.S. population,” Shapiro writes. “Polk obtains the information and cross-indexes it with the names of new car buyers.”

Self-reported data is always bad, but self-reported data from product registration forms!? I hope the Journal relies on better data when it builds its stock listings.

Drive me crazy with e-mail: