Yes, that helps. Thank you.
-
Get involved.
We want your input!
Apply for Membership and join the conversations about everything related to broadcasting.
After we receive your registration, a moderator will review it. After your registration is approved, you will be permitted to post.
If you use a disposable or false email address, your registration will be rejected.
After your membership is approved, please take a minute to tell us a little bit about yourself.
https://www.radiodiscussions.com/forums/introduce-yourself.1088/
Thanks in advance and have fun!
RadioDiscussions Administrators
July '21 ratings
- Thread starter ckg279
- Start date
Take it from someone who subscribes to the book monthly for the past 4 years.
I think Rusty is talking about the July quarter hour listening, which I actually posted on this board, but it was taken down. WECK was 7th 12 plus in raw quarter hour listening, tied with HTT, and beating stations like Star and Kiss.
There are 2 things happening here. Rusty is talking about raw quarter hour listening (popular vote) while Brian is talking about the 12 plus ratings you see on all access( the electoral college).
I am finding that their are many nuances in ratings. It’s just not what the eye sees. WECK shows a 3.2 12 plus, but again, that is not the true number. The true quarter hour number share in July was a 4.6, tied with HTT. Obviously, our 50 to 60 year old numbers crush it.
in July , WECK beat many stations 25 -54 in quarter hour listening. Big time .Who knows why ? It’s not our demo. I don’t want anything under 50. But it has happened many times. Then it gets weighted down when the ratings come out and you see a 3.2
Here’s the process. About 2 days before the actual ratings come out, I receive the actual raw quarter hours and ranks for all stations. In this case in July, WECK tied WHTT 12 plus with a 4.6
2 days later at noon, I get the actual ratings voodoo, which punishes stations like mine because more older people send back diaries, and Nielsen thinks that is unfair, and weights down stations like mine.
So in reality, WECK beat MANY stations in July for quarter hours. 2 days later, after the voodoo WECK finishes at a 3.2.
Do not let All Access and 12 plus ratings be your guide. The real quarter hour shares come 2 days before and WECK does very well. Especially when you consider it’s a 1KW AM , with a few metro translators, versus 50 and 100k full power FM’S
Rustys comments are correct, and I’ve been saying it all along. Stop in WECK sometime and I will show you what I see and you don’t. You would be amazed. It’s always the weighting thing that drives me nuts.
Do you actually believe that BLK is the number 1 station in the market for persons and quarter hours? It’s not. The weighting makes it that way
In any random probability sample, there will not be a perfect representation of every subset of the audience. That is where mathematical weighting comes in.
It's not unfair.
A simple example is a sample of 100 persons. The universe is 50% male and 50% female per the US Census. But we get back 60 women and 40 men, which does not represent that 50/50 balance that exists in the sampled population. So each female return is weighted down proportionally and each male one is weighted up so that the total sample is represented as half and half and reflects the universe.
Of course, the undersampled (and overweighted) portion has fewer respondents so the margin of error within that group will be greater. But the purpose of radio ratings is to get a reasonable approximation, not 100% precision. To be more accurate, the cost increases dramatically.
The weighting simply compensates for lower than standard returns of diaries by African Americans. If the market is 12% Black (as is the case here) but only 10% of the returned diaries is from that group, each one will be weighted up so that Blacks represent the correct percentage in the market.
And as people here have mentioned before, one should not assume that all listeners of WBLK are black.
So while weighting may be a factor, it's not the only reason the station is #1.
Good point... the market has significant (over 5%) Hispanic residents and a percentage of non-Hispanic whites will listen. But in the case Buddy mentions, the weighting is most likely to involve African Americans since that group is more often undersampled due to lower diary return rates. Further, Black and Hispanic communities tend to be younger and younger demos generally have greater issues in diary returns.
Great points David. I wish I was more a student of the ratings methodology. I'm not at all, so I react to what I see. You make some valid points that I did not know.
But I am right now looking at the actual QH shares 12plus in July, and here is how it goes.
WBLK #1 9.2
YRK #2 7.7
BEN #3 7.6
GRF #4 6.9
GR #5 5.2
HTT #6 4.6
WECK #7 4.6
KISS #8 3.9
Breeze #9 3.4
STAR #10 2.9
BFO #11 2.7
EDG #12 2.5
NED #13 2.0
How does our 4.6 at WECK turn to a 3.2 2 days later after I read this? This happens every month like this, and sometimes WECK is in the 5 shares 12 plus, and WAY more in our demo. I just don't understand why the change after the vodoo math 2 days after I get this report.
What is the methodology for ratings? I care much more about sales and client focused ideas than ratings, but I would like to know the yeoman's version of this ratings equation.
My rep has tried to explain to me, and I still don't get it.
But I am right now looking at the actual QH shares 12plus in July, and here is how it goes.
WBLK #1 9.2
YRK #2 7.7
BEN #3 7.6
GRF #4 6.9
GR #5 5.2
HTT #6 4.6
WECK #7 4.6
KISS #8 3.9
Breeze #9 3.4
STAR #10 2.9
BFO #11 2.7
EDG #12 2.5
NED #13 2.0
How does our 4.6 at WECK turn to a 3.2 2 days later after I read this? This happens every month like this, and sometimes WECK is in the 5 shares 12 plus, and WAY more in our demo. I just don't understand why the change after the vodoo math 2 days after I get this report.
What is the methodology for ratings? I care much more about sales and client focused ideas than ratings, but I would like to know the yeoman's version of this ratings equation.
My rep has tried to explain to me, and I still don't get it.
Based on unweighted sample returns, your listener group obviously are "better" at filling out and sending back diaries. So in your demos, there was oversampling, and in others there was undersampling. The oversampled are weighted down and the undersampled group weighted up. That makes the various parts (age groups, gender, ethnicity, county, income, etc) be in proportion to the actual market, even if the sample was disproportional.
Weighting is the entire answer. That means that an oversampled group member may be worth 0.8 and an undersampled group member might be worth 1.2. When those weight-adjusting multipliers are applied, then you get a true proportional sample.
That's not going to satisfy Buddy. Nielsen is rigging the game to cheat WECK. WECK has more listeners than any station in the history of Radio. What happens if someone "expires" the day they send the diary back? They must get "Weighted" underground...
Of course, stations with those demographics lose a potential listener each time a Lombardo hearse passes by...
Thank you for the simple explanation. I understand more than I did.
Your so jealous. You know you are...just say it...in the meantime, put an extra cream in my coffee when I order.
Exactly how long have you been in Radio?
If you don't understand Nielsen by now, it's because you don't want to. Whining when the numbers don't go your way is like blaming the teacher when you get a C on a math test.
Nielsen is not a perfect system, but your station is not a victim of ratings discrimination...
Last edited:
I may unknowingly be looking for a simple answer to a complex process*, and I suppose I will be further displaying a somewhat embarrassing ignorance. But comments about over/under samples has me interested. How is an initial (presumably, random?) sample drawn? From what raw sources are the individual humans identified/taken? And how granular is the process... Zip codes? Legal municipalities? Fire districts? Census tracts? Census blocks? Terrain features?
* - I do subscribe to the thinking that if one cannot explain something so that a seven year old will understand, one doesn't really understand it.
* - I do subscribe to the thinking that if one cannot explain something so that a seven year old will understand, one doesn't really understand it.
I made my first visit to Arbitron in Beltsville in 1970. I've been to Laurel and to Columbia, always looking at diaries and talking to the staff members. I've been on Arbitron committees, industry panels and committees and negotiated contracts.
At no point have I felt I fully understood Arbitron or Nielsen because there is constantly a moving target situation. Changes in methodology, changes in how terms are defined, changes in ascription rules, changes in how people are recruited and dozens more.
And just when I think I am caught up, they change something else.
As I have mentioned before, in a conversation with Owen Charlebois who was first head of the BBM in Canada and then Arbitron, he said that there are three kinds of subscribers:
- Those who went up, who are brilliant programmers and managers.
- Those who staid the same, who are notoriously quiet.
- Those who went down, who know it was due to an Arbitron sampling error.
Last edited:
Any random probability sample tries to get a smallish number of people to participate which represents a miniature set of the whole group they want to understand.
Depending on the research company, they use some kind of commercial listing of residents or households or dwelling units. They then contact them and enlist participation. They try to recruit in proportion to the population they are studying.
Generally from commercial data providers who keep updated lists of the population. They use everything from the Census and the Census Bureau's annual survey to things like auto registrations, home ownership, and whatever other public data they can acquire and consolidate.
That all depends on the scope of the research. An election may want only residents of a city or county. Nielsen works on counties, not cities or smaller units (although they try, by ZIP codes to represent all areas of a county proportionally).
The issue here is that Nielsen does two different kinds of survey. The Diary markets are a random probability sample. The PPM is a semi-permanent panel. Each is recruited differently as the diary has a new panel each week, while the PPM has, perhaps, 4% to 8% turnover each month.
The point was that Nielsen is not rigging the methodology in favour of some stations. If Nielsen said to station owners "Pay us more than your competition and we'll boost your ratings"--I expect that would be fraud. That would be a scandal that would put the company out of business.
Most Radio veterans have a working knowledge of Nielsen. I agree that it can be convoluted and maze like trying to interpret all of it...
In a way, any random probability sample is rigged against the lower ranking players. The margin of error for a 2 share station is considerably higher than that of a 4 share station, which in turn is higher than that of a 6 share station.
A 7 share station is within the margin of error with a 6.1 or even an 8 share. But a 1 share station may be a 1.8 or a 0.5.
That 7 share station may wobble by +/- 15% but the 1 share station may wobble by as much as 50% to 60%... or even more.
What if the "Margin Of Error" is in error?
A 1 share station could actually have an 8 share. Or they could have a Zero share.
Up is down - Down is up.
Using the word "rigged" implies fraud. Stations are free to cancel Nielsen anytime and tell the ad agencies to screw off. "We don't need no badges. I don't have to show you any stinking badges..."
However, that station's competition reserves the right to take those same Nielsen reports to those same advertisers and use them to make a case for moving advertising to them. A station that cancels Nielsen has to be confident in their relationship with their local clients, and that their relationship can withstand the salesmanship at the competition.
- Status