After a major PPM carry rate data error being caught earlier this week (8/22/07 RBR #164), Arbitron reissued PPM Houston August Week 1 (7/26-8/1) Weeklies data yesterday, coinciding with the Arbitron PPM Fly-in conference. The re-release contained the missing data Arbitron said caused the precipitous drop-off or carry rates in that week. http://www.rbr.com/epaper/issue164-07-mla.html
Here is the new data from Arbitron:
"The Average Daily sample size for the weekly report–1,189. The Weekly (unified, 6 out of 7 day) sample size–880 (not 683!)."
Arbitron notes that the real average weekly (unified, 6 day out of 7 day) sample for August Week 1 in Houston was 880, better than the equivalent in-tab numbers for the first two weeks in July (the survey weeks which straddled the July 4 holiday week) and the 3rd week in July.
Arbitron also noted in the PPM Fly-in that the in-tab measure that really matters is Average Daily in-tab-the number that is used to calculate all the AQH ratings.
The average weekly (unified, 6 day out of 7 day) in-tab sample are used to calculate only the weekly cumes. And as John Snyder demonstrated at the PPM Consultant Fly In, the cumes for an in-tab based on the partial sample that was improperly processed was just about the same as the full properly processed sample because of weighting.
We note that weekly in-tab numbers are not the worst week ever, even though we are down another 3.3% from the previous week.
See the new data releases here
Says consultant Randy Kabrich: "Even though Arbitron told the industry they were correcting the sampling issues, if Arbitron doesn’t put the proper quality controls in place to catch a simple problem like this which a 5 year old could catch by comparing in-tab numbers, they will never see the other problems that aren’t so obvious. Add the 3% this week with all the other weeks of decreases and pretty soon all the single digit decreases begin to add up."
He adds, "910 last week in July; 880 this week; another 3% drop from the last week in July. The weekly in-tab is less than half the mythical 1,800 sample that Arbitron loves to quote as being the sample size for Houston."
There are two all time historical lows in the newly released data: Hispanics were down 10% from the last week. Hispanics 6+ and 12+ dropped 10% week to week in the just re-released Houston data–from 250 to 225 In-Tab 12+ and 302 to 270
"It was also an historical record low for Males 18-24, with 57 Daily In-Tab versus 34 Weekly In-Tab, which translates to a ratio of 59.6% in a demo they said they were working on improving," noted Kabrich.
At the Fly-in, Arbitron Chief Research Officer Bob Patchen opened with the Houston data glitch. They then had the panel management people there to talk about how they do their jobs. Yes, they try to get everyone in the household to participate but, in the end, if they can’t get everyone to participate they throw out the household.
Said Patchen: "The data glitch reporting error is not a real change in sample size. For some reason quality control didn’t catch it. Will be re-issued in the next several days. It should have been caught but it wasn’t. We are reviewing quality-control procedures."
Patchen also admitted they’ve been missing their sample targets in Philadelphia and Houston and that daily estimates are weighted up each day.
He described how stable PPM is compared to diaries and that the Philly sample should be up to speed in September; Houston in October. Philadelphia’s
18-34 was ‘well below’ what they want it to be.
Said Adam Gluck, Project Leader, Arbitron, at the fly-in: "A bigger sample size gives you more reliable results-that’s what we’re seeing in PPM."
"That’s pretty much the exact opposite of what their spin doctors said the other day, isn’t it? They said that sample size had no relationship to just about anything?," claimed Kabrich.
RBR observation: PPM is a better method, but they are asking broadcasters to pay more for it than diaries–and fund it. Arbitron didn’t get a credit line to pay for this itself. They have great cash flow with no debt. They didn’t build it first and then sell it to radio. They are asking radio fund it. As well, broadcasters don’t own it. Let’s at least get it right.