It's chilling...perhaps frightening...to find so many defenders of Arbitron. It's clear that you have all had a healthy dose of their Kool-Aid...which is readily available in written form at any NAB. You won't have any trouble finding their booth, because it is the size of an airplane hangar. I disagree with so much of what has been said that I don't know where to start. To witness the comparatively meteoric rise of The Point when the PPM came out and suggest that diary results have been replicated by this new methodology is kind of a shocker.
The difference is that PPM results are so compacted that there is not much difference between #5 and #15. And the results are not replicated... they are similar in share, very different in TSL and very different in cume. But on P1 and P2, they are even closer than in the share tables.
To suggest that most local direct advertisers have multimedia campaigns is downright ignorant (unless you are talking about car dealers...most of which use agencies of some sort or another).
In larger markets, single location advertisers generally use targeted direct mail or neighborhood cable or suburban print. They can not afford mass market full coverage radio.
And to point out that Arbitron has to be affordable to radio stations is to completely miss the point.
It DOES have to be affordable, or stations would not sponsor it, and Arbitron would leave the market. It's a balance of a reliable sample vs. price for each subscriber.
The biggest abusers of the Arbitron system are the ad agencies and they pay a fraction of what radio stations pay to use the service.
Ad agencys pay cost for Arbitron. They do not share in the expense of doing the survey, they just pay the cost for delivering the results. The only purpose of doing ratings is to have a sales tool to help set pricing, and agencies use ratings to negotiate price. There is no abuse there.
And I have taken a statistics course by the way but it doesn't provide an explanation for why a vast percentage of a rock station's cume would suddenly abandon Led Zep for Garth Brooks in a 13 week period...which happened in Houston in the 90's along with dozens of other horror stories that resulted in PD's renting U-Hauls.
Many stations are cyclical, based on format appeal. And there is the one time in many that the statistical error is more than one standard error. In general, the diary results were stable, and reasonable.
To paint Arbitron as some benevolent, impartial third party interested in truth and fair play for all is to give credit where credit isn't due. Any system that gambles with people's livelihoods by depending on the masses' willingness to provide factual information in exchange for A DOLLAR is not living in the real world.
Arbitron is a business and wants to make a profit. So is radio. When the goals are in congruence, all are satisfied. And the fact is, not eveyone got a dollar (DST) and the PPM pays like a frequent flier program, not by the meter.
The system is not perfect... only a census can approach perfection, and a quarterly census in Houston for radio listening would cost more than the total billings of all radio stations. So we make do with a salmple, and understand that the results are estimates and will wobble a bit, but that is not important within the buying function.
And any system that can be manipulated by a 300 point per week TV campaign, The Birthday Game, or a multiple car giveaway was definitely in need of an overhaul. Are we saying that the PPM is it? Not according to NABOB and dozens of others.
The fact is that contests can change listening, just like advertising heavily can affect sales. That is why both are done... if a station can bring up ratings 52 weeks a year, then the advertiser does not care why... as long as it sustains the listening.