This excerpt taken from the ARB 10-K filed Feb 28, 2008.
Radio Audience Measurement Services
Collection of Listener Data Through Diary Methodology. We use listener diaries to gather radio listening data from persons aged 12 and over in sample households in the 300 U.S. local markets in which we currently provide diary-based radio ratings. Participants in Arbitron surveys are selected at random by landline telephone number. When participants (whom we refer to as diarykeepers) agree to take part in a survey, we mail them a small, pocket-sized diary and ask them to record their listening in the diary over the course of a seven-day period. We ask diarykeepers to report in their diary the station(s) to which they are listening, when they are listening and where they are listening, such as home, car, work or other place. Although survey periods are 12 weeks long, no participant keeps a diary for more than seven days. Each diarykeeper receives a diary, instructions for filling it out and a small cash incentive. The incentive varies according to markets, and the range is generally $1.00 to $6.00 for each diarykeeper in the household and up to $10.00 in certain incentive programs for returned diaries. In addition to the cash incentives included with the diaries, further cash incentives are used at other points in the survey process along with other communications such as follow-up letters and phone calls to maximize response rates. Diarykeepers mail the diaries to our operations center in Columbia, Maryland, where we conduct a series of quality control checks, enter the information into our database and produce periodic audience measurement estimates. We receive and process more than 1.4 million diaries every year to produce our audience listening estimates. We measure each of our local markets at least twice each year, and major markets four times per year.
Collection of Listener Data Through PPM Methodology. In our PPM service, we gather data about encoded audio material through the use of our PPM meters. We randomly recruit households to participate in the service (all persons six and older in the household). The household members are asked to participate in the panel for a period of up to two years, carrying their meters from rise to retire each day. Panelists earn points based on their compliance with the task of carrying the meter. Longer carry time results in greater points which are the basis for monthly cash incentives. Demographic subgroups that are less likely to comply, such as younger adults, are paid higher premiums based on their compliance. We consider the amount of the cash incentive that we pay to the PPM panelists to be proprietary information.
The meter collects the codes and adds a date/time stamp to each listening occasion. At the end of each day, panelists place their meters in a docking station and the information is downloaded to Arbitrons facilities for editing and tabulation.
We issue a currency report for 13 four-week measurement periods per year. We issue weekly reports to station subscribers for programming information. Users access the currency data through a software system.
Response Rates and Sample Proportionality. It has become increasingly difficult and more costly for us to obtain consent from persons to participate in our surveys. We must achieve response rates sufficient to maintain confidence in our ratings, the support of the industry and accreditation by the MRC. Response rates are one quality measure of survey performance among many and an important factor impacting costs associated with data collection. Overall response rates have declined over the past several years. If response rates continue to decline further or if recruitment costs significantly increase, our radio audience measurement business could be adversely affected.
One of the challenges in measuring radio listening, whether by using diaries or electronically, is to ensure that the composition of survey respondents is representative of the market being measured. We strive to achieve representative samples. A measure often used by clients to assess quality in our ratings is proportionality, which
refers to how well the distribution of the sample for any individual survey matches the distribution of the population in the local market. For example, if eight percent of the population of a given market is comprised of women aged 18 to 34, ideally eight percent of the diarykeepers or panelists (collectively, survey participants), as applicable, in our sample are women aged 18 to 34. Therefore, each survey participants listening should statistically represent not only the survey participants personal listening but also the listening of the demographic segment in the overall market. In striving to achieve representative samples, we provide enhanced incentives to certain demographic segments to encourage participation. Households identified as having at least one member who is Hispanic receive bilingual materials. We also use bilingual (Spanish-English) interviewers for households where Spanish is the preferred language.
In our PPM service, we also use a measure known as Designated Delivery Index (DDI). DDI measures sample proportionality based on how many persons in the sample represent a particular demographic compared to the number of persons expected to be in the sample based on the markets sample target. In recent years, our ability to deliver good proportionality in our surveys, both diary-based and electronically measured, among younger demographic groups has deteriorated, caused in part by the trend among some households to disconnect their landline phones, effectively removing these households from the Arbitron sample frame. As consumers adopt modes of telecommunication other than telephone landlines, such as mobile phones, it is becoming increasingly difficult for us to reach and recruit participants.
We have committed extensive efforts and resources to address the decline of response rates and to maintain sample proportionality. Currently, we manually dial mobile-phone-only households for PPM recruitment only. We have conducted a number of research tests over the past two years to develop more efficient ways of contacting mobile-phone-only households to recruit participants. We have not yet announced a date for including these mobile-phone-only households in the diary service.
In recent years, we have announced a comprehensive set of initiatives to bolster response rates and improve sample proportionality among African-American, Hispanic, and young male respondents in our diary-based markets. These initiatives include providing for increases in cash incentives and other survey treatments. We continue to research and test new measures to address these sample quality challenges. The most significant response rate initiatives in 2007 included the completion of the rollout of the 2006 response rate and proportionality action plan and the opening of a third Arbitron owned and operated participant-interviewing center during the first quarter of 2007. Our experience is that internal interviewing centers outperform the outsourced calling center vendors that they replace. We believe that additional expenditures will be required in the future with respect to response rates and sample proportionality.
As the PPM service is rolled out, we expect to experience similar challenges in the operation of an electronic measurement service as are faced in the diary-based service, including those challenges related to response rate and sample proportionality. We also expect that additional measures to address these challenges will be implemented and require expenditures in addition to those required for our diary-based service. On August 31, 2007, we announced new initiatives and investments to address PPM panel maintenance concerns experienced in both the Houston-Galveston and Philadelphia markets. On February 1, 2008, we announced a series of four sample quality benchmarks that we intend to pursue with our PPM services to enhance confidence in PPM ratings as a currency. Rather than serving as guarantees, these benchmarks establish the starting point for where our PPM samples are currently performing, and identify the levels we intend to work toward in 2008 through a program of continued improvement. Specifically, these benchmarks concern total sample size, sample size for persons aged 18 to 34, average daily percentage of the installed panel that provides useable data (the average daily in-tab), and response rates. Currently, we have at least 30 initiatives in the testing or implementation stage for the PPM service that are designed to improve either response, compliance or both.
In December 2007, we announced a sample size guarantee that would provide a partial rebate to our customers for PPM radio ratings in any local market for a measurement period in which our delivered average daily in-tab among persons aged 18-54 falls below 80 percent of our published average daily in-tab target for that market. In Houston-Galveston and Philadelphia, the sample size guarantee took effect with the release of the December 2007 PPM survey month (November 13 December 10, 2007), which we released on December 31, 2007. In future PPM local markets, the sample size guarantee will take effect with the release of the third currency PPM
survey period. To date, we have exceeded our published targets in all PPM local markets, and, therefore, no amounts are payable to any PPM customer under the sample size guarantee.
Small Market Initiatives. We are currently reviewing our services in U.S. local markets ranked 100 and smaller (approximately 200 markets). In concert with a group of broadcasters known as the Owner Operator Caucus, we have begun work in two phases. First, we plan to improve the qualitative information aspect of our current diary service beginning with the Fall 2008 Survey. Next, we are working on new methods of measuring media behavior in small markets along with more extensive qualitative information to better serve the needs of our clients. We expect the testing of the new methods to take place over the next two years.
Quality Improvement Initiatives. We continually invest in quality improvements for our radio audience measurement services. In addition to the initiatives described above in Radio Audience Measurement Services Response Rates and Sample Proportionality, we expect our 2008 quality improvement initiatives to include the following: