ARBITRON LOSES ACCREDITATION IN 5 MARKETS / ATLANTA ON THE BUBBLE

[gregg75]

Looks like those PPMs aren't as acurate as they were "promised" to be.
Could there be the possibility of manipulation of data (somewhere)?

http://arbitron.mediaroom.com/index.php?s=43&item=797

[RoddyFreeman]

Looks like those PPMs aren't as acurate as they were "promised" to be.
Could there be the possibility of manipulation of data (somewhere)?

http://arbitron.mediaroom.com/index.php?s=43&item=797

I doubt it's manipulation of data.  From what I've heard, it's probably more related to the sample, proper ethnic weighting, etc.

[upstate29651]

Looks like those PPMs aren't as acurate as they were "promised" to be.
Could there be the possibility of manipulation of data (somewhere)?

http://arbitron.mediaroom.com/index.php?s=43&item=797

I doubt it's manipulation of data.  From what I've heard, it's probably more related to the sample, proper ethnic weighting, etc.

Explain this "proper etyhnic weighting".


G

[agentUrge]

'Splains why they were minimizing the numbers showing Pandora's growth in listener-ship

[jabba17]

Looks like those PPMs aren't as acurate as they were "promised" to be.
Could there be the possibility of manipulation of data (somewhere)?

http://arbitron.mediaroom.com/index.php?s=43&item=797

I doubt it's manipulation of data.  From what I've heard, it's probably more related to the sample, proper ethnic weighting, etc.

Explain this "proper etyhnic weighting".


G

If a market is (say) 20% black, then ideally your panel will be 20% black as well--assuming your demo data to begin with is correct; ATL got burned on overestimated Census figures, especially with Latinos, last year.  One controversial shortcut is to apply a factor to come up with the proper weighting; for example, if your panel is only 10% black, then the ratings of those panelists will be weighted double, and everyone else will be weighted less to compensate.  

Of course, there are limits to how much of this can be done; if sample size gets too small, then reporting may be incorrectly magnified.  For example, what would be the effect on the final numbers if you have only 2 black panelists (out of 100, making for a 2% black panel) and they both listen to WGST all day?  Now you have 2% of the market making up 20% of your ratings (and they looove WGST).

Now blow this out among all demos tracked in the market, and all the possible associated factoring.  

[upstate29651]

Looks like those PPMs aren't as acurate as they were "promised" to be.
Could there be the possibility of manipulation of data (somewhere)?

http://arbitron.mediaroom.com/index.php?s=43&item=797

I doubt it's manipulation of data.  From what I've heard, it's probably more related to the sample, proper ethnic weighting, etc.

Explain this "proper etyhnic weighting".


G

If a market is (say) 20% black, then ideally your panel will be 20% black as well--assuming your demo data to begin with is correct; ATL got burned on overestimated Census figures, especially with Latinos, last year.  One controversial shortcut is to apply a factor to come up with the proper weighting; for example, if your panel is only 10% black, then the ratings of those panelists will be weighted double, and everyone else will be weighted less to compensate.  

Of course, there are limits to how much of this can be done; if sample size gets too small, then reporting may be incorrectly magnified.  For example, what would be the effect on the final numbers if you have only 2 black panelists (out of 100, making for a 2% black panel) and they both listen to WGST all day?  Now you have 2% of the market making up 20% of your ratings (and they looove WGST).

Now blow this out among all demos tracked in the market, and all the possible associated factoring.  

Ok, I see the logic, but that's also assuming ALL blacks listen to an "urban" format, for example.  Can't say I agree with this metric.


G

[wpb1999]

From NYC board:

Can someone explain to me how 2 Califorinia stations, KWIZ and XHTY, show up on New Jersey radio ratings and they are not shown as streaming?

[gregg75]

As for Atlanta my guess would be that the black population is OVERSTATED. Arbitron lists
the black population percentage as 30.4%.........yet 6 of the top 13 stations on the latest
ratings page are mostly black  oriented (around 48%). Take away WWVA and 5 of the top 8
are mostly black oriented (roughly 65%). That could be the Arbiton problem in Atlanta.

Now you'll say some other races also listen to black radio, but blacks also listen to "white
radio" so that probably cancels each other out IMO.

[RoddyFreeman]

As for Atlanta my guess would be that the black population is OVERSTATED. Arbitron lists
the black population percentage as 30.4%.........yet 6 of the top 13 stations on the latest
ratings page are mostly black  oriented (around 48%). Take away WWVA and 5 of the top 8
are mostly black oriented (roughly 65%). That could be the Arbiton problem in Atlanta.

Now you'll say some other races also listen to black radio, but blacks also listen to "white
radio" so that probably cancels each other out IMO.

Arbitron gets its population estimates from the U.S. Census.  So if Arbitron is wrong, the U.S. Census is wrong.

[kilamanjero]

As for Atlanta my guess would be that the black population is OVERSTATED. Arbitron lists
the black population percentage as 30.4%.........yet 6 of the top 13 stations on the latest
ratings page are mostly black  oriented (around 48%). Take away WWVA and 5 of the top 8
are mostly black oriented (roughly 65%). That could be the Arbiton problem in Atlanta.

Now you'll say some other races also listen to black radio, but blacks also listen to "white
radio" so that probably cancels each other out IMO.

You are reaching for sure with this statement about the number of blacks in Metro Atlanta.

According to the US Census, the mean percentage of the core 6 counties (Fulton (44.1%), DeKalb (54.3%), Gwinnett (23.6%), Clayton (66.1%), Cobb (25%), and Douglas (39.5%)) in Metro Atlanta of black inhabitants is nearly 42.1%, now if you add the additional non-core counties the mean percentage of blacks drops to 30.4%.  This is because the majority of the blacks in Metro Atlanta reside in the core counties not the outer counties of the Atlanta radio market.  These are percentages from the 2010 US Census and Arbitron uses these as their numbers.  The only thing with Arbitron is their usage of Census estimates, which aren't accurate at all for a number of reasons, but the decennial count is the only legitimate count.