"Armin Walpen never interfered with research".

In issue 26 of Werbewoche, Guiseppe Scaglione expressed strong doubts and concerns about radio research using Radiocontrol. However, Manuel Dähler, Head of the SRG Research Service, was unable to comment due to time constraints and promised a reply at a later date. Here is the answer - in the form of an interview conducted in writing with Manuel Dähler.Radiocontrol (RC) is regarded as the radio currency in Switzerland, although it does not record radio listening from the Internet. Shouldn't the SRG research service be fair in pointing out in every RC publication that a growing proportion of radio usage is not recorded? Manuel Dähler: It is clear and undisputed that RC does not yet measure radio listening from the Internet. What is much less clear is how often and for how long radio is used via the Internet today, i.e. how "large" the part of radio usage that is not recorded is. Accumulated "clients" or "visits" over a month are unsuitable for a comparison with the daily reach from RC because they are not based on people and also do not differentiate between domestic and foreign access. In addition, many radios restrict use via the Internet by limiting the number of simultaneous accesses to the stream for cost reasons. We will only get more reliable information from NET-Metrix - the new Internet research - which measures the use of radio (and other) streams on a person-by-person basis. In the reach rankings of MA Strategy 2005 and RC 2005, no station is placed as differently in the young target group as Radio 105. Why is that? MA Strategy does not show reach, but the narrower user group (ENK). They therefore compare survey results ("how often do you listen to ...") with the technical measurement, in different samples. In contrast to measurement, the image of a station or awareness due to advertising campaigns can also play a role in surveys. Looking at the population as a whole, the differences in ranking are not that great; other radio stations are also ranked similarly differently in the MA Strategy than in RC. However, it is difficult to assess Radio 105, as no other target group station or cable radio station is included in the MA Strategy. Despite this, RC reported more listeners during this time than during normal broadcasting operations. This calls the whole RC system into question? as far as I know, the transmitter failure lasted about an hour. Unfortunately, we were not informed of this incident until about a year later, which made the investigation practically impossible: After this long period, we were no longer able to reconstruct what was actually heard on Radio 105's frequencies during the outage. Radio 105 has also been unable to provide us with any recordings to date.In weeks 11 and 12 of this year, Radio 105's listening time in the 35 to 44 target group suddenly shot up from 5 minutes to 307 and 278 minutes per day. How can this be explained? RC shows two types of minutes: Total usage in minutes (total minutes listened to a radio divided by all people including non-users) or usage per listener (total minutes listened to a radio divided by all users of that radio). In 2005, these values for Radio 105 were 0.2 minutes of total usage and a good 25 minutes of usage per listener respectively. It is true that between March 23 and 29, 2006 (weeks 12 and 13), the usage per listener for Radio 105 was significantly above the average (total usage rose to 1.2 minutes on one day). These days were therefore specially monitored. The high values can be attributed to a person between 35 and 44 who listens to Radio 105 for long periods every day. The watch is worn for at least 13 hours a day, the use of Radio 105 is not en bloc but spread throughout the day and the person also listens to other stations (Basilisk, Radio Monte Carlo Suisse, SWR3). There is no reason to consider the data invalid or manipulated. Once again we had a real fan in the measurement, as we had for Radio 105: March 02, April 03, October 03, March 04, December 05 or February 06. That is part of the nature of the sample. Exceptional does not mean wrong by any means! The fact that one watch is given so much weight, even for a station with a regional language distribution, shows that the RC samples are too small, especially for special-interest stations, and that the RC panel is therefore not representative.the sample is larger than ever, which means that the influence of the individual person is lower than ever: in German-speaking Switzerland, we conducted 9,800 media study interviews every year. Today, we conduct 571 interviews every week. So after around 17 weeks, we reach the same number of people as we used to in 52 weeks. However, if we look at individual days, things look different. For specialty channels that only look at a section of the sample, we recommend analyses using quarterly and semester values. With regard to "representativeness", the question should rather be asked whether the busy Radio 105 listener belongs in a sample representative of the population: Yes, of course! But since they exhibit "rare" behavior, not all that often. If the figures are analyzed over a quarter, he is one of thousands. What do you say to the theory that RC primarily reflects the radio use of conservative sections of the population, but hardly that of young, progressive social groups, because of the clock as a measuring instrument and the recruitment of panelists via landline telephone? Although socio-demographic characteristics (age, gender, education, income, occupation, etc.) are systematically controlled for, they are poorly suited to depicting progressiveness. Using the EuroSocioStyles, however, we were able to show that the RC panel is significantly more progressive than the Telecontrol panel (TV research). The main difficulty with these comparisons lies in the lack of population-representative data. And in difficult definitions: What criteria characterize a hiphopper, for example? A critical question may also be raised here: Is it actually cooler for a hiphopper to answer MA Strategy questions over the phone? In 2004, something happened that should never have happened: SRG boss Armin Walpen intervened in the research in favor of Radio Grischa. This still arouses suspicion today. Why did you never communicate these events more openly or explain how Grischa's problem was ultimately "solved"? And what was different from those other radios that also showed significantly lower values after the panel was expanded, but were unable to block their data? Armin Walpen never interfered with the research. He demanded that the data be made known to the customers before it was released to the general public. Before the data was released, discussions were therefore held with all radios whose values had undergone major changes. Approval for data release was given by the individual stations. Grischa, which experienced the largest downward corrections due to the error, agreed to release the data in the Reporter much later than the others. However, Grischa's data was always available in the planning tool for radio advertising. Both the User Commission itself and external experts appointed by the User Commission as well as the MWKK were consulted to correct the error and subsequently check the data before it was released again; they all recommended the release. However, the responsibility for the release after the correction ultimately lay with me and I guarantee that the figures from Grischa or other radios were not manipulated. And anyone who analyzes the data will see that nothing was "solved" here. Markus Knöpfli