scooty430 said:
- Arbitron somehow did not notice that this household, this address, was the home of a Top 10 radio personality.
Arbitron (and Nielsen) do not keep track of the addresses of on-air talent or employees of radio stations. We're talking about maybe 75,000 or so people for all the stations in all the rated markets. Heck, most stations I have worked with can't even keep an updated extension list of their own staff. Addresses, by the way, can be considered confidential.
- It took a number of months before Arbitron noticed anything odd.
No, it was spotted fairly quickly. The Spring book was not relased for that market until the end of July. As I said, it only changed the station a few tenths in one daypart, but it did get caught. Remember, the diaries may or may not have been in the talen's own home. The issue was spotted and corrected within days. the issuance was July 24 and the new, retabulated, book was issued on the 11th of August, about a dozen work days after the book was issued.
- It took outsiders to notice the connection between the host's hometown and the diaries. Aribtron did not make that connection.
Arbitron can not and does not track the residences of station staff. They ask a media affiliation question only. What was noticed was the considerable number of quarter hours in the household for one station, and Arbitron investigated.
- No red flags were raised when the host's wife told Arbitron there were six people living in a suburban household between 25 and 34. SIX people in that age range? Not suspicious?
No, the term is "dwelling units" and shared housing often has a number of people in the same age range. It could have been 6 interns at a hospital, 6 grad students, etc. It's quite common.
- No red flags were raised when it showed that they listened to almost every minute of WPRO's morning show, and yet did no other radio listening.
There are so many diaries that have heavylistening in only one daypart that this would not be at all unusual. I've been looking at diaries since 1970 in Beltsville, and you are questioning things those of us who have done it for decades know are quite normal. Not everyone lives in a "Leave it to Beaver" stereotypical home with a mom, a dad, two kids and a fluffy dog. I have gotten two books and one trend reissued myself, so I know there are errors on occasion. But you are looking at stuff that is quite normal in a random probability sample.
- A number of commenters in the Providence thread admit they would do (or have done) the same thing if given the opportunity, only they'd be smarter about it - less obvious.
And the fact is obvious from the outcome that they would not get away with it.
- A few posters also mention that one single diary, in the right demo, can "move the needle." That is not good statistics.
If radio needed a larger sample, it would pay for it. As it is, radio is paying as much as it can afford, so occasionaly there will be sample errors, minor inaccuracies and attempts to influence the book. And yet the system has worked for 43 years with seldom a real complaint
The PPM opens up even more problems, namely people being less willing to wear the things,
The participants are willing to wear them, and that is all that matters.
even smaller sample sizes,
If you knew what a panel was, you would know that we actually have about 4 times as large a sample for the PPM.
people not wearing them consistently,
They do, or they are removed from the panel.
and certain demos not wanting to wear monitoring devices as they distrust authority.
Untrue.