What I don't know is whether or not Nielsen's screening process prevents outliers from being either PPM carriers (in the top 50 markets) or diarykeepers. Usually, if a household is chosen, all members in that household are participants. (They don't like to tell any of us how they go about that.)
Yes, they do. Recruiting is at the household level. Someone in a "dwelling unit" (can be a family of several people sharing a residence) is the team leader and makes the agreement that everyone in the household will participate. Meters are sent, and everyone must use them. If anyone does not, the leader is warned that the individual in question is not carrying the meter. If that meter does not come back into compliance, the whole unit is dropped. Credits good for "rewards" are based on full compliance, and the better the group participates (no skipped days, etc.) the more they can earn.
And Nielsen has made a big deal out of the PPM "hearing" listener exposure that may have been ignored under the diary system. Wouldn't that have a skewing effect similar to an outlier carrying the device?
No.
First, by "outliers" I meant "people who don't have the listening habits of the larger majority... because that group has gone to new media predominantly. So they are members
outside of the target of certain stations who listen anyway. Those stations generally don't target, research or program to those people, but they get them because the core younger demo members are not using much radio.
Those numbers are shares, and a 3 or 4 share in 18-34 today is, in rating, about 15% of what it was 20 years ago.
The "phantom cume" (the term we used pre-PPM) was made up of secondary listening choices. A listener might write in a diary about their two or three favorite stations, but not write in those occasional "alternative choice" stations as they were not top of mind when filling in the diary. A classic case would have been someone in LA in 1999 who put on KFWB a moment when they got in their car to get the traffic before putting on Stern or Rick Dees or whomever on their "favorite" station; that never got written into the diary.
I also noted somewhere on RD in the past week or so that the migration of the old "auditorium testing" for music to online has made it difficult to detect the outliers. In fact, I believe it was David who posted that.
Outliers are detectable by charting all the respondents; people who are generally outside the normal range are visible. It is harder, but still very doable.
So I think you have brought up a valid point in terms of ratings accuracy, Flip. It may be within the margin of error, but what if it isn't?
In seeing several 6+ numbers this evening, I notice that nobody is taking into account that a station going from 8.0 to 7.2 is well within the normal wobble of a sample of that size. That is why nobody actually in radio looks at just two months. We like to look at 4, 5, 6 months in our core.
A good thing about Lance's ratings summaries is that he covers more months. Oh, and he gets the formats much more accurately than some of the sources.
researchdirectorinc.com
