Data glitch? Houston PPM carry rates drop one-third in one week


As RBR first reported yesterday, Houston, we may have a big problem when it comes to PPM carry rates. There is a sudden sharp drop in Houston August Week 1 weekly PPM intab.  The daily intab is in range but the weekly has suddenly dropped by around one-third in most demos. Is this possible? Philadelphia has been on a slow, long-term drop over the last few months as well.

RBR sources privy to the latest information are frankly pretty concerned: "Philadelphia has a slight improvement.  I’m concerned about Houston, however.  It’s so far off that I think there may be something wrong with the sample data we’ve been supplied.  While the Daily sample number is in-line with the last one (actually it’s up slightly), the weekly in-tab number is down dramatically, falling by one-third.  The carry rates have fallen from the mid-70s to the 50 percentage range in nearly all age cells.  Men 18-24 is in the 30’s. Inconceivable this could happen from one week to the next. I’m checking further to make sure the sample data we were given is correct."

There were 2,078 installs in Houston and only 683 intabs for P6+ in Week 1. For P12+ there were 620 intab. In July Week 4, there were 910 intab for P6+, and 811 intab for P12+.

Are those Houston numbers right? If they’re correct it’s not good for the future of PPM. The explanation could be as simple as they downloaded the wrong data (i.e. the weekly sample numbers aren’t the weekly sample numbers; they’re something else and are mislabeled).

"Anything other than that is too scary," said one radio operator.
Here’s what Arbitron Chief Research Officer Bob Patchen is telling one concerned source: "It certainly seems possible that summer schedules and vacations could impact the number of people who qualified for six out of seven days in August week 1, but we need to verify that’s the cause."
We heard Patchen is telling some that the data (the weekly sample) is wrong and that he’s waiting for a fully checked and rechecked version of the sample count before its reissued. The weekly cumes and TSL’s might change because of this "data processing error." However, Arbitron spokesperson Thom Mocarsky tells RBR a different view-see story, below.

RBR observation: This should make good fodder for the Arbitron PPM Fly-in this week. Three more markets go live in 40 days-this may get interesting. In the PPM call-in last Wednesday, Arbitron said radio was willing to dump some respondents that were not cooperative to get more cooperative respondents on the panel. Arbitron claimed they had been going through this process, churning them out. The problem is if that is what had really happened then the daily number would have dropped as well-because if the people aren’t in the sample, they couldn’t be in the daily or weekly sample.
Care to respond to this? We’ll run it. Send to: [email protected]