Re: Ratings for Dummies
>
> Question. How do they really figure out the ratings?
> Interviews with real listeners or something?
>
In synthesis...
Arbitron recruits by phone using SSI databases for each market, plus random digit dialing to catch unlisted numbers. They recruit all members of a household that accepts.
A household receives the diaries by mail. Each has an incentive, designed to create appreciation and guilt. A follow up call confirms receipt and clarifies the how to part of it.
A diary keeping household starts on a Tursday and ends on the next Wednesday. They fill in all listening on pages labeld by the day, with places for the time and tick boxes for listening location and whether it is AM or FM.
Arbitron attempts to get a proportional sample each week... that means the number of respondents in each cell corresponds to the proportion of that cell in the market. Cells are based on age, sex, ethnicity and geographic distribution in the market. If 40% of the market is Hispanic, 40% of the diaries should be Hispanic... and each week they shuffle depending on the returns from prior weeks.
A diarykeeper gets a call after the weekend, early int he week and at the end of the survey week to make sure they are filling the diary in.
Diaries are mailed back, and manually processed. First, they are digitally scanned, and the processor uses a screen image to tab in the data in the book. The processer has pop up files of station names, slogans, frequencies, program names, etc. to help.
Each book has 12 7-day cycles.
The number of diaries varies by market population. In general, in each 12 week book, the average diary represents about 1000 persons. So the chances of getting called each year are one in 250. The reason the sample is not larger is cost, as the sample is the largest cost in ratings or research. Stations, in general, do not feel like spending the money to increase the sample, as it takes four times the sample to cut the margin of error in half, and ratings are very expensive already.
(Please... don't take offense at the header... it is meant in jest.)