Arbitron tells Philly clients all is well

For the second straight week, with March Week 2 now reported, in-tab numbers for PPM in Philadelphia were down. This time, though, Arbitron was quick to tell clients not to panic and that this was the expected result of its intentional effort to improve PPM panel representation of the 25-34 demo.

Arbitron VP Thom Mocarsky emphasized repeatedly to RBR that the in-tab target for Philadelphia is 1,530 and that the company continues to exceed that, despite dropping from the February average daily in-tab of 1,763, which was 15% ahead of target. In keeping its pledge to broadcasters to increase the 25-34 sample, Arbitron said in the first two weeks of March it intentionally reduced the Persons 6+ oversample in Philadelpha to "open up" sampling points for households with 25-34-year-olds. It is using methods like those previously used to boost 18-24-year-old representation. Broadcasters had been pressuring Arbitron to focus on getting in-tabs up to snuff for 18-34-year-olds, where most radio ad buys are made.

"As we use our newly enhanced sampling control system to bring into the panel new households with 18-24 year olds, with 25-34 year olds and with both 18-24 and 25-34 year olds, we expect to see continued improvement in the 18-34 sample. We also expect there will be a better balance between 18-24 year olds and 25-34 year olds," Arbitron said in an email explaining the latest Philadelphia results.

———————————————————————————————-

The Arbitron email:

Subject: March 2008 Week 2 PPM Weeklies Report and our 25-34 sample initiatives

We want to bring you up to date on two important matters:

* the status of the 25-34 sample initiative, which is an important addition to our ongoing program to improve the quality of our PPM panels; and
* the impact that the program is having on the PPM panel in Philadelphia.

The Philadelphia panel has been running substantially over the target of 1,530 panelists among persons 6+.  In February, we had an average daily in-tab sample of 1,763 panelists – 15% greater than the target. 
Given the desire to increase the 25-34 sample as quickly as possible, and given the large cushion in P6+ sample that we had in Philadelphia, we decided to implement an accelerated matching and replacement process as soon as the enhanced sample control software came on line.

In the first two weeks of March, we intentionally reduced the Person 6+ oversample in Philadelphia in order to open up sampling points needed for households with 25-34 year olds.

The reduction of the oversample at the P6+ level is having an impact on the average daily sample sizes (DDI) for many demographics including (in the short term) Persons 18-24. 

Here’s what you will see with today’s release of March 2008 Week 2:

* The Persons 6+ DDI is at 102, down from 110 the week before. (March Week 1 vs. Week 2)
* 18-24 DDIs have dropped in both the Male and Female demographic cells. Men 18-24 went from a 93 to an 80 DDI.  Women 18-24 went from an 81 to a 77 DDI. (March Week 1 vs. Week 2).
* 25-34 has improved markedly March week one vs. week two: Men 25-34 – 73 to 83 DDI; Women 25-34 – 69 to 80 DDI.
* 18-34 DDIs are flat, but now they are much more balanced between the 18-24 and 25-34 demographics.

As we use our newly enhanced sampling control system to bring into the panel new households with 18-24 year olds, with 25-34 year olds and with both 18-24 and 25-34 year olds, we expect to see continued improvement in the 18-34 sample. We also expect there will be a better balance between 18-24 year olds and 25-34 year olds.

Since the 25-34 initiatives are being implemented panel wide, you will see similar patterns among the Black sample.

* Week 2 Black DDI is at 99. We intend to also manage that segment closer to goal as part of our focus on 25-34 improvements.
* The DDI for Black 18-54 Black is at a 93. We expect to see similar improvements the balance between the 18-34 and 35-54 segments as the 25-34 initiatives continue.

We’ve also heard consistently from our customers that in-tab sample levels for the sales demos are more important than the overall persons 6+ in-tab. The overrepresentation  at the 55+ level has already been reduced; we expect that a greater percentage of the Person 6+ sample in Philadelphia will be comprised of Persons 18-54.

Over the next two months, we plan to exceed our guarantee of an 80 DDI for the 18-54 sample, and we anticipate doing so with a better balance between 35-54 year olds and 18-34 year olds. We also plan to be above our benchmark of a 90 DDI in terms of the average daily in-tab target of 1,530 persons age 6 and older.

I’ve attached a worksheet that covers the highlights of the changes to the sample to date. It’s important to note that despite the changes in sample, there have been no significant changes in total listening levels for key demographics week over week, again demonstrating that changes in sample composition are not directly correlated with changes in ratings.

We also realize that we need to improve the timeliness of our communication about our intentions, our actions and the expected outcomes in terms of the PPM sample quality initiatives in your market.  We intend to be more diligent about letting you know in advance when a significant sample enhancement program comes on line and what you can expect to see with the release of the next ratings report.

If you have any questions, please don’t hesitate to call me.

In the first two weeks of March, we intentionally reduced the Person 6+ oversample in Philadelphia in order to open up sampling points needed for households with 25-34 year olds. In February, we had an average daily in-tab sample of 1,763 panelists – 15% greater than the Persons 6+ target of 1,530.

Change in Average Daily In-Tab Panelists
February 2008 PPM Survey Month versus March 2008 Week 1

February 2008 PPM Survey Month versus March 2008 Week 2

Most of the changes in terms of the absolute number for average daily in-tab occurred outside the Persons 18-54 demographic. Person 55+ and Persons 6-17 declined by 123 panelists; Persons 18-54 by a net of 78.

In the first week of March, there was a noticeable decline of approximately 215 persons in Average Weekly In-Tab for Philadelphia. It was more pronounced than the 56-person decline in Average Daily In-Tab. This was due to the temporary increase in panel turnover as we intentionally took sample households offline to make room for 25-34 households. In any given week with higher turnover rates, fewer persons will qualify for the Average Weekly in-tab sample size, which requires at least 6 out of 7 days in-tab.  Both the persons leaving the panel and those entering the panel during the week will be less likely to meet the 6-day in-tab minimum for that week. (Please note, all our sample and DDI targets are based on Average Daily in-tab sample.)

Despite the changes in sample, there have been no significant changes in total listening levels for key demographics week over week, again demonstrating that changes in sample composition are not directly correlated with changes in ratings.

Data (c) 2006-2008 Arbitron Inc.  All Rights Reserved. For use only pursuant to a license from Arbitron Inc. Subject to the limitations and qualifications disclosed in the data and reports.
Arbitron PPM Weeklies(tm) Software has not been accredited by the MRC and Arbitron has not requested accreditation for this version of the software.
Exported reports from the Arbitron PPM Weeklies are not accredited by the MRC.

RBR/TVBR observation: Both last week and this week we heard from both Arbitron officials and constant critic Randy Kabrich about the Philly in-tabs – either they are just as expected, or this is evidence of a serious problem. Kabrich doesn’t think Arbitron can focus on boosting 25-34 numbers at the same time it is focusing on 18-24 participation and keep numbers where they should be for other demos all at the same time, but you can read his comments elsewhere in the RBR/TVBR Blog Log. One comment from Kabrich to RBR/TVBR that Arbitron folks will agree with is "The proof is in the pudding." We and everyone else will be watching to see how the figures come out for the entire month in Philadelphia, as well as for the non-currency panels in New York, LA and elsewhere.