Waiting for Arbitron

Interesting article from NYT: something is wrong about portable people meter (PPM) of Arbitron (US radio audience research company); after a test in October in New York, Arbitron will delay the broader introduction of PPM by nine months.
Why did NYT write that now?
The original news from Arbitron came on November 26 2007.

"Arbitron inc. announced today that it will delay that commercialization of its Portable People Meter (PPM) radio ratings service in nine markets. New York, Nassau-Suffolk and Middlesex-Somerset-Union will be delayed by nine months; Los Angeles, Riverside and Chicago by six months; and San Francisco, San Jose and Dallas by three months."

Radio's challenge: counting the young

By Brian Stelter
Published: December 31, 2007

As radio moves away from an outdated system of audience measurement, it is facing the same problem that confronts other media: figuring out how many young people are listening or watching is maddeningly difficult.

A long-awaited electronic measurement tool, the “portable people meter,” has produced sudden swings in ratings in tests, raising red flags for radio executives. After a test in October in New York, the radio measurement company Arbitron delayed the broader introduction of people meters by nine months.

Arbitron officials say that tallying the number of young listeners — in particular the 18- to 34-year-old demographic group prized by advertisers — has been a perpetual concern. The stakes are high for media companies and advertisers, and for the measurement companies themselves; last month, Arbitron warned its stockholders that earnings would be adversely affected by the people meter delay.

Ceril Shagrin, an executive vice president for corporate research at Univision Communications, a major Spanish-language broadcaster, found that the ratings for Radio La Kalle 105.9 FM in New York fell sharply when the people meters were tested in October. She expressed concerns about the reliability of the Arbitron sample.

“The Hispanic universe is younger, and our audience is therefore made up of more young people. When the sample doesn’t reflect them, then our radio ratings aren’t correct,” Ms. Shagrin said.

Lower ratings can translate into less advertising revenue. The television ratings company Nielsen Media Research has experienced the same sampling challenges.

“I think any survey research operation, no matter what they’re doing, would say that young households are more difficult to recruit,” said Paul Donato, chief research officer for Nielsen.

Mr. Donato said young demographics were well represented by Nielsen’s samples, and Arbitron said its data was improving. November data from the people meter, released on Dec. 7, showed some gains among younger people and minority groups.

Still, Arbitron has work to do; on an average day in New York last month, 575 participants aged 18 to 34 provided data, far short of the company’s goal of 919.

Owen Charlebois, president for technology, research and development, said young people were less likely to participate automatically in a research study.

“We have to work harder to explain why it’s important for them to take part,” he said. “In some ways, there are parallels with voting rates in the U.S.”

Because it costs more for Arbitron to deploy electronic measurements, the sample size becomes smaller, but each member of the sample contributes more data, making the recruitment process that much more important.

“You really have to get people who you know are going to participate,” said Brad Adgate, director for research at the media-buying agency Horizon Media.

Researchers often try to persuade hard-to-reach age groups to participate by paying them more. While Arbitron does not release specific information on incentives, it said it offers 18- to 24-year-olds two to three times as much as the average respondent. The company will increase the age for incentives to 34 in 2008.

Even once recruited, the younger people pose challenges for sampling studies.

“On an average day, half of 18- to 34-year-olds don’t provide usable data,” Ms. Shagrin said. Young people may leave the people meter, which is roughly the size of a cellphone, at home because carrying it is inconvenient in certain situations, as when on a date or while playing sports.

Like other research companies, Arbitron weights the sample to achieve equal representation for each demographic group.

“Those few people who do decide to participate count more than those in other age groups,” potentially skewing the results further, Mr. Adgate said.

Bob Patchen, chief research officer for Arbitron, said young people have less routine in their lives, making them more likely to forget or lose the meter. The company sometimes visits young respondents to encourage participation. It also plans to release “meter skins” that will allow participants to customize their meters, in much the way some young people change the colors and textures of their cellphones.

The larger issues raised by people meters involve the fragmentation of media and the preference of advertisers to reach narrow swaths of consumers. On television, as in radio, programs that appeal to 18- to 34-year-old viewers can become cash machines because advertisers will pay a premium to reach young people. But as media companies use Arbitron and Nielsen to carve ever-thinner slices of demographic data, concerns about the reliability of that data increase.

“Sometimes the ratings, particularly in cable, are really being generated by just a couple of people in the sample,” said Alan Wurtzel, president for research and media development at NBC.

But over time, the data tends to become increasingly reliable, researchers say. Arbitron will spend the next nine months trying to make its customers just as confident.