A friend just told me that Arbitron is doing something new to weight the numbers - something about primary languages, race, etc. Haven't had the time to look on All Access for any news yet. Can someone please explain to me what is going to happen and how it might affect various markets? Also, if you find any articles about this I'd appreciate a link so I can read more.
-
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
New Arbitron weighting?
- Thread starter radiovictim
- Start date
> A friend just told me that Arbitron is doing something new
> to weight the numbers - something about primary languages,
> race, etc. Haven't had the time to look on All Access for
> any news yet. Can someone please explain to me what is
> going to happen and how it might affect various markets?
> Also, if you find any articles about this I'd appreciate a
> link so I can read more.
>
Arbitron is going to seek proportionality in both types of Hispanics, English dominant and Spanish dominant. That means that if a market´s Hispanics are 60% Spanish dominant, there will be 60% of the total Hispcanic diaries place among Spanish dominant Hispanics.
This affects the 21 largest Hispanic markets, and will seek to stabilize wobble in the numbers based on uneven sampling of each segment of the Hispanic population. there is a press release on the Arbitron website.
If the sample is not 100% proportional, weighting will be used to achieve proportionality, just as it is used for all the other stratification variables.
> to weight the numbers - something about primary languages,
> race, etc. Haven't had the time to look on All Access for
> any news yet. Can someone please explain to me what is
> going to happen and how it might affect various markets?
> Also, if you find any articles about this I'd appreciate a
> link so I can read more.
>
Arbitron is going to seek proportionality in both types of Hispanics, English dominant and Spanish dominant. That means that if a market´s Hispanics are 60% Spanish dominant, there will be 60% of the total Hispcanic diaries place among Spanish dominant Hispanics.
This affects the 21 largest Hispanic markets, and will seek to stabilize wobble in the numbers based on uneven sampling of each segment of the Hispanic population. there is a press release on the Arbitron website.
If the sample is not 100% proportional, weighting will be used to achieve proportionality, just as it is used for all the other stratification variables.
M
Marc0n1
Guest
"weighting"
The word "weighting" sounds so much more professional than "fudging," but that's what it is.
When not enough diaries are returned (and minorities' rate of diary return is lower than average), they feed steroids to the diaries they have to make the results look evenly spread. In other words, they fudge.
This is just one of many examples of how the ratings that make or break our careers are shamefully unreliable. If the radio ratings business ever gets real competition, things may improve.
The word "weighting" sounds so much more professional than "fudging," but that's what it is.
When not enough diaries are returned (and minorities' rate of diary return is lower than average), they feed steroids to the diaries they have to make the results look evenly spread. In other words, they fudge.
This is just one of many examples of how the ratings that make or break our careers are shamefully unreliable. If the radio ratings business ever gets real competition, things may improve.
Re: "weighting"
> The word "weighting" sounds so much more professional than
> "fudging," but that's what it is.
>
> When not enough diaries are returned (and minorities' rate
> of diary return is lower than average), they feed steroids
> to the diaries they have to make the results look evenly
> spread. In other words, they fudge.
Weighting is a standard statistical treatment. There is no way to get a perfectly proportional sample unless you do a census.
In fact, there are as many cells weighted down as there are weighted up. If there are, lets say, 1000 diaries in the ¨perfect sample, and some cell comes back undersampled, there is an oversample somewhere else.
In general, the weighting is very small in the arbitron survey... it is unusuall to se weighting of more than 2% up or down in any cell... so even if there were no weighting, the variance introduced is very small.
>
> This is just one of many examples of how the ratings that
> make or break our careers are shamefully unreliable.
Statistics is the only science where error is not a dirty word. Any time you use a poll and not a census, there is a margin of error. That error percentage is related to the sample size. If stations in a market wanted less wobble (error manifested as non-stable results in a stable environment) they could buy a bigger sample. The fact is, stations are satisfied with the current sample size as it is a good cost vs. value point.
The errors are, in fact, minimal most of the time. Like any poll, there are occasional larger errors (outside one standard error) but that will happen in any survey of comparable size no matter what the methodology is.
The Portable People Meeter will have less than a quarter of the participants per 12 week period of the diary method, so the 12 week sample goes DOWN!
> If the
> radio ratings business ever gets real competition, things
> may improve.
>
There have been many competitors. Since 1970, we have seen Pulse, Hooper, Birch, Audits and Surveys, and Burke come and go. The fact is that the users of Arbitron, the advertising agencies and big advertisers, are the ones who were not convinced by any of the other companies and chose arbitron to measure radio.
> The word "weighting" sounds so much more professional than
> "fudging," but that's what it is.
>
> When not enough diaries are returned (and minorities' rate
> of diary return is lower than average), they feed steroids
> to the diaries they have to make the results look evenly
> spread. In other words, they fudge.
Weighting is a standard statistical treatment. There is no way to get a perfectly proportional sample unless you do a census.
In fact, there are as many cells weighted down as there are weighted up. If there are, lets say, 1000 diaries in the ¨perfect sample, and some cell comes back undersampled, there is an oversample somewhere else.
In general, the weighting is very small in the arbitron survey... it is unusuall to se weighting of more than 2% up or down in any cell... so even if there were no weighting, the variance introduced is very small.
>
> This is just one of many examples of how the ratings that
> make or break our careers are shamefully unreliable.
Statistics is the only science where error is not a dirty word. Any time you use a poll and not a census, there is a margin of error. That error percentage is related to the sample size. If stations in a market wanted less wobble (error manifested as non-stable results in a stable environment) they could buy a bigger sample. The fact is, stations are satisfied with the current sample size as it is a good cost vs. value point.
The errors are, in fact, minimal most of the time. Like any poll, there are occasional larger errors (outside one standard error) but that will happen in any survey of comparable size no matter what the methodology is.
The Portable People Meeter will have less than a quarter of the participants per 12 week period of the diary method, so the 12 week sample goes DOWN!
> If the
> radio ratings business ever gets real competition, things
> may improve.
>
There have been many competitors. Since 1970, we have seen Pulse, Hooper, Birch, Audits and Surveys, and Burke come and go. The fact is that the users of Arbitron, the advertising agencies and big advertisers, are the ones who were not convinced by any of the other companies and chose arbitron to measure radio.
T
Tom_Desmond
Guest
Re: "weighting"
> The word "weighting" sounds so much more professional than
> "fudging," but that's what it is.
Weighting, when properly done, is not fudging. Check out a marketing research textbook and you'll find the matter of weighted sampling discussed as a method to *improve* the validity of research.
> The word "weighting" sounds so much more professional than
> "fudging," but that's what it is.
Weighting, when properly done, is not fudging. Check out a marketing research textbook and you'll find the matter of weighted sampling discussed as a method to *improve* the validity of research.
A
abz7001
Guest
Re: "weighting"
> The word "weighting" sounds so much more professional than
> "fudging," but that's what it is.
>
> When not enough diaries are returned (and minorities' rate
> of diary return is lower than average), they feed steroids
> to the diaries they have to make the results look evenly
> spread. In other words, they fudge.
>
> This is just one of many examples of how the ratings that
> make or break our careers are shamefully unreliable. If the
> radio ratings business ever gets real competition, things
> may improve.
>
I couldn't agree more. A major problem with this is that once the targeted group does start returning diaries at a normal rate, the 'weighting' will still be there. That will make the stations targeting this weighted group more likely to be the top stations, even though they might not be. A good example of this is African-American targeted stations in markets like DC and Baltimore. Weighting was used when the diaries weren't being sent back - but now they are, and African-Americans count as more than one caucasian person. As a result, the stations targeting that group are inflated even though they really don't have the listeners. The same will happen to any group that gets 'weighted' if Arbitron doen't make this a temporary, rather than permanent thing.
> The word "weighting" sounds so much more professional than
> "fudging," but that's what it is.
>
> When not enough diaries are returned (and minorities' rate
> of diary return is lower than average), they feed steroids
> to the diaries they have to make the results look evenly
> spread. In other words, they fudge.
>
> This is just one of many examples of how the ratings that
> make or break our careers are shamefully unreliable. If the
> radio ratings business ever gets real competition, things
> may improve.
>
I couldn't agree more. A major problem with this is that once the targeted group does start returning diaries at a normal rate, the 'weighting' will still be there. That will make the stations targeting this weighted group more likely to be the top stations, even though they might not be. A good example of this is African-American targeted stations in markets like DC and Baltimore. Weighting was used when the diaries weren't being sent back - but now they are, and African-Americans count as more than one caucasian person. As a result, the stations targeting that group are inflated even though they really don't have the listeners. The same will happen to any group that gets 'weighted' if Arbitron doen't make this a temporary, rather than permanent thing.
Re: "weighting"
>
> I couldn't agree more. A major problem with this is that
> once the targeted group does start returning diaries at a
> normal rate, the 'weighting' will still be there.
Weighting is used to make minor adjustments to make the value of every age, sex and ethnic cell proportional. Since diaries come in over 12 weeks, Arbitron can adjust the recruit right to the end to get as close to a real proporitonal sample and not have to weigh any cell excessively.
> That will
> make the stations targeting this weighted group more likely
> to be the top stations, even though they might not be.
Untrue. Weighting makes each person in the market count at the same value.
> A
> good example of this is African-American targeted stations
> in markets like DC and Baltimore. Weighting was used when
> the diaries weren't being sent back - but now they are, and
> African-Americans count as more than one caucasian person.
Wrong. The returns for Blacks and Hispanics are fairly similar to Other. Weighting makes all identically proportional. Arbitron adjusts every week during a survey period.
> As a result, the stations targeting that group are inflated
> even though they really don't have the listeners. The same
> will happen to any group that gets 'weighted' if Arbitron
> doen't make this a temporary, rather than permanent thing.
>
Weighting is dynamic, to adjust to proportonality and cell that gets a few to few diaries or a few too many. Weighting increases the accuracy.
>
> I couldn't agree more. A major problem with this is that
> once the targeted group does start returning diaries at a
> normal rate, the 'weighting' will still be there.
Weighting is used to make minor adjustments to make the value of every age, sex and ethnic cell proportional. Since diaries come in over 12 weeks, Arbitron can adjust the recruit right to the end to get as close to a real proporitonal sample and not have to weigh any cell excessively.
> That will
> make the stations targeting this weighted group more likely
> to be the top stations, even though they might not be.
Untrue. Weighting makes each person in the market count at the same value.
> A
> good example of this is African-American targeted stations
> in markets like DC and Baltimore. Weighting was used when
> the diaries weren't being sent back - but now they are, and
> African-Americans count as more than one caucasian person.
Wrong. The returns for Blacks and Hispanics are fairly similar to Other. Weighting makes all identically proportional. Arbitron adjusts every week during a survey period.
> As a result, the stations targeting that group are inflated
> even though they really don't have the listeners. The same
> will happen to any group that gets 'weighted' if Arbitron
> doen't make this a temporary, rather than permanent thing.
>
Weighting is dynamic, to adjust to proportonality and cell that gets a few to few diaries or a few too many. Weighting increases the accuracy.
- Status