GfK figures highlight the flaws in Rajar methods
GfK Media’s radio audience figures have caused a stir with Rajar. But instead of criticising, Rajar should be getting its own house in order, says Mark Jarvis
At first sight, the results of GfK Media’s National Broadcast Media Survey, which covers TV and radio audiences, confirm what many had suspected, that research respondents’ memories are fallible and paper diaries are not a reliable method of measuring media consumption.
However, the audience levels recorded by the electronic wrist monitor device used by GfK are claimed to reflect all listening – including in cars, offices and out of home. This brings a new dimension to the numbers and should enable measurement of radio, television and even cinema consumption to be conducted using the same standards. At the moment the disparity between the Rajar figure for Radio 4, which reported 10 million listeners, and the GfK research figure of 17.9 million, demonstrates how much the two methods vary.
But flawed research is not a novelty in the broadcast industry, as anyone who has had to work with BARB data over the past 18 months will testify. The fact that Rajar is researching electronic audience measurement systems, including the wristwatch system, supports the belief that paper diaries are outdated and that the time has come to move to a more robust methodology.
Rajar has argued that GfK’s offering is inaccurate, while accepting that its current system is flawed. And GfK’s parallel test, in which people used both diaries and a wristwatch over a week, proved the advantages of the latter. When recording data in a diary, people forget about short bouts of station surfing, and say they listened for longer uninterrupted blocks than was actually the case. But the wristwatch isn’t perfect either. It relies on the user to remember to put it on and then to use it correctly. It is also going to be more susceptible to malfunctions. But it is less likely to report inaccuracies than a diary method.
Rajar’s definition of reach when using the diary method is classed as five minutes or more in any 15-minute period. The GfK technology allows reach to be calculated minute by minute and more people in the country are going to have listened for one minute-plus than five minutes-plus, so it is valid to claim that GfK has got closer to what the industry actually wants. Rajar has pointed out that its sample is nine times bigger than GfK’s, but GfK is only aiming to report national listening, so its sample size is large enough. Rajar uses a much larger sample because it covers local stations.
We should applaud the efforts of Kelvin McKenzie’s Wireless Group, which commissioned the survey, for drawing attention to the flaws in the current system. However, a measurement system that has the potential to measure across media must be introduced with industry consensus, following rigorous testing, if the results are to be honoured as a viable trading currency.
Mark Jarvis is broadcast director at Carat
