Client Requests


Your custom request has been sent. We’ll contact you if we have any additional questions.

* Indicates a required field.

  • MM slash DD slash YYYY
  • This field is for validation purposes and should be left unchanged.

Hot Topics

Arbitron Market Information – Does This Stuff Matter?

November 29, 2012

We’ve come a long way from the days when Arthur Carlson on “WKRP in Cincinnati” sat with a brown envelope in his trembling hands before he opened it up to learn his station’s latest Arbitron ratings. Back in the ‘70s, Arbitron’s clients didn’t know when the books would arrive. They waited for the phone to ring to get a few numbers read to them as “Advanced Ratings.” In the early ‘80s, Arbitron began publishing a schedule, so at least you knew when Arbitrends data would be released and when the box of books would be delivered. In 2006 Arbitron switched to the eBook format, and we no longer needed shelves to hold the infamous red-, silver-, blue-, and brown-striped reports.

These days most of us don’t seem to have a reason to open an Arbitron eBook. We access the ratings using a wide variety of software programs such as Tapscan®, Arbitrends®, PPM Analysis Tool, or XTrends. These tools make it easy for us to get the information we are searching for.

So why should you open the Arbitron eBook? One of the reasons is the first section of the report called “Market Information.” This section includes a map of your market, but it also contains important information about the sample that was used to generate the ratings.

Why should you care about that? The ratings are impacted by things other than what you and your competitors do on-air. They are affected by how well the sample matches the population it is meant to represent. If certain parts of the population are under- or over-represented in the final sample, it could benefit or hurt different stations in different ways.

Before I go on, please understand one important fact. There is no correlation between lower intab in counties, demo cells, or ethnic groups that are the core listeners to a given station, and the ratings that come from that smaller sample. The fact is, Male 18-24 intab could be low but the AOR station that depends on that cell could have great numbers. How can that be? If the M18-24 diaries or meters that are in the sample happen to be heavy listeners to that AOR station, that sample will be weighted up, giving the station the famous ratings “kiss.” On the other hand, if those diaries or meters happen to be lighter listeners to that AOR station, those will be heavily weighted and the station’s estimates will plummet.

Starting on page MI-2, you can find the number of diaries or meters that were intab by county, demo cell, race/ethnicity, and other subcategories, such as the respondent’s language preference or the amount of sample that came from cell-phone-only households.

So what should you do with this sample performance data? You’ll want to analyze it to determine if all of the counties, demo cells, and Black or Hispanic (in ethnically weighted markets) portions of the population are well-represented in the sample. Just as important, you’ll want to know if the proportionality of those groups changed significantly from the prior book. These changes in sample proportionality can produce big ratings surprises.

The proportionality index provides an easy reference for determining if a portion of the population was under- or over-represented in the sample. Here’s how you calculate the proportionality index for a given segment of the sample. Let’s calculate the index for Men 18-24:

  1. Calculate the percent of the total unweighted sample from Men 18-24. So if there were 112 meters intab from Men 18-24, and a total of 2,300 meters intab from Persons 6+, Men 18-24 represents 4.9% of the sample (112 / 2,300 = 4.9%). If you are working with diaries, you would use the number of diaries from P12+ as the denominator in this calculation.
  2. Calculate the percent of the total population from Men 18-24. So if there were 450,500 Men 18-24, and a total 6+ population of 8,655,000, Men 18-24 represents 5.2% of the population (450,500 / 8,655,000 = 5.2%).
  3. Divide the result in #1 by the result in #2. In this case 4.9 / 5.2 = .94 or an index of 94. An index below 100 indicates that the demo cell is under-represented, and will need to be weighted up to correct for the lower return. An index above 100 indicates that the cell is over-represented, and will need to be weighted down to correct for the higher return. An index of 100 indicates a perfectly proportionate sample, so coming in at or close to 100 is the goal, because it results in less weighting.

This same process can be used to evaluate sample performance by county or race/ethnicity. Once you’ve calculated the proportionality index for each demo cell, county, and race/ethnicity, that information can be trended. If the indices fluctuated significantly from the last book, that can be a major factor in ratings changes.  If the indices are pretty close to 100 and are consistent, it is less likely that sample performance is causing the changes in your ratings.

So go ahead. Open your eBook and learn more about the sample performance in your market. You’ll understand more about why your ratings are what they are.

-Marc Greenspan, Partner