GfK attacks Rajar over audience measurement tests

The other devices, Arbitron's Portable People Meter and Eurisko's Media Monitor, were approved by Rajar for further testing.

The level of difficulty of the conditions depended on a number of variables, such as radio volume, background noise volume, whether the wearer was moving and where the device was placed – from "normal" (ie where the manufacturer advised wearing it), to under a jacket, in a pocket and in a bag.

GfK said the engineer judged that no listening should be registered in most cases.

Matthias Steinmann, chief executive of Telecontrol, said: "We are astonished by the results. We cannot agree with the results of tests that have been presented in such a way that a casual observer will conclude that the [MediaWatch] registers less than one third of all listening. Such claims are not credible."

Rajar said the audiometers were not expected to record listening in difficult conditions.

Out of 1,389 listening sessions, GfK/Telecontrol's MediaWatch identified 32% of listening sessions correctly, against the Arbitron PPM's 59% and the Eurisko Media Monitor's 70%.

However, Rajar's results show that even if the five most difficult levels were discounted, the equivalent figures would be 35% for GfK/Telecontrol's MediaWatch, against the Arbitron PPM's 64% and the Eurisko Media Monitor's 76%.

A Rajar spokeswoman iterated that the tests were designed independently by the National Centre for Social Research.

She said: "GfK/Telecontrol is out as it stands. But if it improves its product Rajar is happy to review its participation in the tests."

Rajar is now designing a second round of tests, which will focus on how people using the audiometers actually use and react to them in practice. The tests will take place within the next two or three months.

Steinmann announced yesterday that Telecontrol would make the MediaWatch's data capture processes more frequent and was conducting its own new series of tests with an independent scientific engineering institute.

He also queried the "normal" position in which the audiometers were worn during the tests.

"When the MediaWatch was worn normally, the other two devices were carried in a much more exposed fashion (round the neck in the case of PPM), which it is hard to imagine would be the norm for respondents in a live survey," he said.

If you have an opinion on this or any other issue raised on Brand Republic, join the debate in the <a href="http://forum.brandrepublic.com" target="_blank">Forum here