D Peter Maus wrote:

David Eduardo wrote:
"D Peter Maus" wrote in message
...

A sample.

A sample, well designed.

Represents a stations entire listenership.


My statement stands. A sample influences your decisions. How the sample
is designed and how it represents are decisions based on assumptions
accepted by statistical science.

Statistics _is_ a science, and it accepts the fact that there is always a
margin of error in polling, The degree of error that is acceptable depends
on the use that will be made of the information. The data obtained is,
itself, accurate withing the margin of error. Radio ad sales is tolerant of
a degree of error, as there are more important variables involved in
advertising than just the the margin of error of a survey sample.
Assumptions that cannot be proven, nor demonstrated to be true in any
given instance.

Actually, if you look at margin of error, a properly designed poll...
whether for music to play or audience size, can be pretty much proven by
replication procedures.

The differentiation between statistical analysis and census. One is a
scientific extrapolation. The other is a headcount. One CAN result in the
same outcome as the other, within defined limits of acceptability. But
they are not the same. And can in significantly divergant results.

If there is divergence, it is due to not doing the poll correctly. In this
case, it is quality control. It's just like making a car... faults per 1000
vehicles, etc.

So, reiterating, music is not programmed by your listeners, it's
selection is influenced by a sample. But the decisions are made by
consultants and PD's.

Since replication can verify using a sample to determine the acceptability
of songs, then the issue is implementation... a separate matter. Neither PDs
nor consultants change test results. It is almost plug and play once you
have the results.
Selling the process to your listeners: "Music is programmed by the
listners."

I did a little experiment... in Argentina, we did a 100 person music test.
We also did the test on the air, and ran the test form in a large newspaper
(circulation 1.1 million) We got 40,000 forms back. The test matched the
newspaper results. Then we pulled 100 test forms at random from the 40,000.
The results were also the same.
Critical analysis asserts that cannot be the case.

Any effort I have seen (some done on purpose) to disprove a music test
results when the test itself follows standard techniques has failed.


That's because the axiomatic assumptions are the same in each case.
The statistical science is the same. Of course the results are going to
be the same

David, we're not arguing the test. Nor the science. But the
PD/Marketing claim that the music is programmed by the listeners.

It's not. It's programmed by the PD's and Consultants based on the
results of a sample of listeners.

The two statements are not the same. Your statement implies a sample
of 100%. Which is not the case.

As I said, the results may be, give or take, about the same. But they
are not the same. Anymore than a statistical extrapolation is the same
as a census.

One of the reasons you take as much **** here as you do, is because
of statements that sound more like marketing than discussion.

Which is precisely why he's worn out his welcome everywhere he goes.

He knows that, I know that, but his defence mechanism will never, ever, allow him
to admit it.

It's a mental illness.