Arbitron's sampling methodology is FARmoreimportantthanMrorMrsDePetro's misdeeds

[Don A]

> And people take the numbers seriously!


Part of why people in the industry take it so seriously is that the numbers 
are relatively stable month after month.

What was key in DePetro's situation was the apparent "out of the blue" spike 
in listeners for no apparent reason.

Can one month be out of whack?  Probably.

Can 12 months be out of whack?  Not as likely.

Like in sports when averaging statistics they have been known to throw out 
the lowest and highest numbers before averaging.....thinking they are 
abberations.


For instance, from what I see, WKLB had last year between a 3.2 and a 3.7.

The relative stability of that number gives some creedence to the accuracy, 
reliability and methodology.

If there suddly was a 2.5 or a 5.0 then we could more readily assume that 
there is some abberation in the methodology.


Was it possible that someone could cheat and get a 1 month spike?  Yeah, 
probably.

Is it possible they could keep on cheating every month for 12 months?  Not 
likely.


This is why DePetro's story came to light.