KNUU AM 970 Didn't Improve Rating Despite New Show

[jondavidvox]

Re: Arbitron reality.

Thanks for the information, David. I truly appreciate it.

However....

The "....opinion of an estimate" quote came from a lecture on ratings by Walt Sabo in 1990. I looked it up in a book at 94Q in Atlanta, and sure enough there it was. (I was in elementary school in the 60's.)

The "legalese" statement is the only real thing in any ratings book. They tell you that "The reliability of audience estimates cannot be determined to any precise degree...." So we take the whole thing on faith, right?

Don't get me wrong....Arbitron has been really good to me. But I took the wins with the same grain of salt as I took the (fortunately rare) losses. But honestly, do you really believe these numbers to be as accurate as you say?

Diaries, and people meters give consistant results....in Philly. But doesn't it follow that most methodologies would work better in a large, established metropolis? Better sampling rates, and response....A more stable populace....Stronger civic traditions....Much better radio?

How about markets from #30 to #50? Here in Las Vegas, the only thing that's consistant in Arbitron numbers is the volatility.

By the way, nice to talk to a real broadcaster, and not a vaporous alias...

Jon-David Wells
Fearless Broadcaster

 

[davideduardo]

Re: Arbitron reality.

> Thanks for the information, David. I truly appreciate it.
>
> However....
>
> The "....opinion of an estimate" quote came from a lecture
> on ratings by Walt Sabo in 1990. I looked it up in a book
> at 94Q in Atlanta, and sure enough there it was. (I was in
> elementary school in the 60's.)

I have a number of books representing the 1970's, 80's and 90's. Nowhere do I find the workd "opinion" in the disclaimers.
>
> The "legalese" statement is the only real thing in any
> ratings book. They tell you that "The reliability of
> audience estimates cannot be determined to any precise
> degree...." So we take the whole thing on faith, right?

Statistics is the only science where error is not a dirty word. We do not take statistics on faith, we take them on the basis that the data is, in nearly all situations, predictable based on sample and cells. However, since each cell has a different sample size... in fact, each station has a different diary return, there is no way to apply one single figure as margin of error in a book. A station with a 10 share might have a margin of error on total week of +/- 5%, and one with a 1 share a margin of error of +/- 25%. And each will have higher errors in specific dayparts and demos.

Given this potential for error, advertisers accept it as being relatively correct within all the other variables of doing an ad campaign.
>
> Don't get me wrong....Arbitron has been really good to me.
> But I took the wins with the same grain of salt as I took
> the (fortunately rare) losses. But honestly, do you really
> believe these numbers to be as accurate as you say?

Yep. I have even replicated them in the US using a different methodology. I might clarify that I actually wrote the software for part of the ratings service in Puerto Rico in the 80's, so am very familiar with the whole process... and I currently design research for a living.
>
> Diaries, and people meters give consistant results....in
> Philly. But doesn't it follow that most methodologies would
> work better in a large, established metropolis? Better
> sampling rates, and response....A more stable
> populace....Stronger civic traditions....Much better radio?

No. Market size, ethnicity, etc. has no effect that I can determine as long as a properly stratified and proportional sample is maintained. Remeber, the PPM sample is less than 20% of the diary sample for a full book in all markets. This is because it is a panel, not a rolling sample.

The good thing about a panel is that true proportionality can be achieved week after week by recruiting the panel proportionally. This is nearly impossible form week to week with the diary, as you do not know which diaries will come back.
>
> How about markets from #30 to #50? Here in Las Vegas, the
> only thing that's consistant in Arbitron numbers is the
> volatility.

This is because the market has a small sample and the margin of error per the back of the book is higher. A larger sample could be used, but none of the users will pay for it and the existing sample is good enough for sales.
>
> By the way, nice to talk to a real broadcaster, and not a
> vaporous alias...

Thanks. Same here.