Survey Method Questioned as WHUR Surges
Saturday, July 26, 2008; Page C07
According to the latest audience ratings, the most popular radio station in the Washington area for the past three months was WHUR, the R&B-and-soul outlet owned by Howard University. WHUR's share of the audience, in fact, grew by almost 20 percent, according to the Arbitron quarterly survey released yesterday.
All of which raises a question: Really?
How could WHUR (96.3 FM) -- a station that played essentially the same type of music, had the same on-air personalities and didn't undertake any special audience-building promotions during the spring -- surge so far, so fast?
WHUR's ratings increase has all the makings of what radio people like to call a "wobble," a freakish, one-time aberration caused by Arbitron's increasingly creaky data-collection methods.
In brief, the ratings are determined by a representative sample of listeners throughout the Washington area. Arbitron sends out diaries to these listeners and asks them to write down what they listened to for a single week. After collecting diaries from 12 sample groups for 12 weeks, Arbitron crunches the numbers and figures out who's gaining or losing listeners.
But the system has always been fraught with problems, says Richard Harker, a veteran radio researcher in Raleigh, N.C. For example, he says, Arbitron doesn't know how many in its sample group will actually return their diaries, which can create statistical turbulence in the final results. Arbitron typically gets older listeners to return diaries, but tends to come up short with younger audiences and "ethnic" listeners -- radio code for primarily African Americans and Latinos. WHUR's audience is mostly younger African Americans.
In the event that a particular group is underrepresented in the overall sample, Arbitron tries to correct that by giving special credit, or weight, to the diaries it did receive from the group's members.
But, as Harker points out, that might not address a geographic problem. If too many diaries come back from one part of a region and not enough from another, stations that have a strong following in the first area gain will gain at the expense of a station whose listeners are in the under-surveyed part.
"Wobbles have been a part of the process as long as Arbitron has been doing ratings," Harker says. Sophisticated advertisers "kind of roll with it. They tend to look at one ratings period as just a single data point" and use an average of three or four periods in determining where to place their ads.
Wobbles, of course, work both ways -- one station's gain can be another's loss. The biggest loser in the spring ratings period, for instance, appears to have been WLZL (99.1 FM). The Spanish-language pop station's overall share of the audience fell by almost 33 percent.
WHUR's general manager, Jim Watkins, said yesterday that he hadn't reviewed the ratings yet and he declined to comment on them. But he offered a general observation: "The whole rating system is manic-depressive. Before a ratings book comes out, most general managers think, 'Am I going to have a liquid lunch or a solid lunch when I tell my investors and owners how we did?' I've been around long enough to know it can go either way."
This time, he said, "I'm having a solid lunch."