Jan 2012 ratings

[radiopunk99]

Here are the Jan 2012 top 10 rankers for P25-54 and W25-54. I still don’t get P18-34…so PLEASE, if someone has it - post it!

P25-54:
KOMP (1)
KISF (2)
KWID (3)
Sunny (4)
KKLZ/Mix (5t)
Coyote (7)
KLUC (8)
KXTE/Point (9t)


W25-54:
Coyote (1)
KOMP (2)
KISF (3)
Mix (4)
KKLZ (5)
Sunny (6)
KLUC (7)
KWID/KRGT/KWNR (8t)


Notes from me:

-As expected and previously discussed in prior posts, Spanish language stations are thru the roof due to Arbitron’s newly implemented “language preference” strategy in response to the new census data. Both KISF and KWID were up over 20% from Holiday to January.

-Also, as I mentioned in a previous post, Coyote is up once again for the 4th book in a row. They are now #1 with women and almost doubling KWNR’s share.

-Sunny’s share is literally cut in half now that Christmas music is gone.

-A huge book for KOMP, #1 P25-54 and (even more impressive) #2 W25-54.

-KXTE had a pretty decent book. It’s about time since their direct competition went away months ago.

-Not a good book for Point – just barely in the top 10. I have a feeling BOB-FM might be taking a little and it’ll probably get worse for Point with all of the BOB-FM TV spots running. Bob’s music is clearly aiming at Point.

 

[MC]

Doesn't look like the Republican caucuses did anything for the talk stations. I wonder if advertisers are going to weigh the "new methodology" in Spanish station ratings?

 

[Dr. Daliah]

es una posibilidad, si.....

 

[davideduardo]

Doesn't look like the Republican caucuses did anything for the talk stations. I wonder if advertisers are going to weigh the "new methodology" in Spanish station ratings?

There is no "new" methodology.

Arbitron finally, after 12 years, applied language preference as a "quota" or stratification variable for Las Vegas. This had not been done before as the source for the enumeration... that is, the percentage of Hispanics who are Spanish domianant and those who are English dominant... did not provide data for Las Vegas. Now they do. Now Arbitron can make sure the sample has representation of both Spanish dominants and English dominants within the Hispanic portion of the sample.

Language preference as a stratification variable has been used in markets like Phoenix and LA and Riverside and even Bakersfield for over a decade.

 

[davideduardo]

-As expected and previously discussed in prior posts, Spanish language stations are thru the roof due to Arbitron’s newly implemented “language preference” strategy in response to the new census data. Both KISF and KWID were up over 20% from Holiday to January.

"Language Preference" is not a strategy. It's just part of doing one of the many things the MRC requires to certify the methodology and implementation in Las Vegas.

A very finite and very slowly changing percentage of Hispanics are either English dominant or Spanish dominant. If 30% of the market is Hispanic (just using a round number) and 60% of Hispanics are Spanish dominant, the 18% of the sample should be Spanish dominant Hispanics.

This is no different than the LA market where 9.9 million live in LA County and 3.0 million in Orange County. The sample should be about 22% in Orange and the rest in LA county... if it were not, there would be a failure on the geographic stratification variable and the sample would be wrong.

The Hispanic language based sample has been wrong for decades. Finally, Arbitron can correct this.

The incorporation of language preference as a stratification variable in Las Vegas has nothing to do with the Census. Until the beginning of the January survey, Arbitron had no way of establishing a sample quota for Spanish dominant or English dominant Hispanics, because the provider of the data in all other markets did not do an enumeration in Las Vegas. Now they do, Arbitron buys the data from them (Nielsen, to be specific) and now the survey is going to be much more consistent in including the proportional number of each language preference group, book to book, by properly and proportionally balancing the panel.

 

[MC]

Not a "new methodology" or "strategy" but a "quota." OK, that certainly makes a big difference to advertisers I bet.

 

[MC]

I wonder about Reno and SLC, and their Spanish Language stations rankings now. I had a statistics professor make a big deal about "treatment" of numbers and statistics. It wasn't about "right" or "wrong" or "accurate" or "inaccurate" but how you wanted to look at the data.

Audio and video equipment specs, car horsepower and mileage data, and of course unemployment and inflation statistics have lost a lot of credibility over the years.

 

[davideduardo]

Not a "new methodology" or "strategy" but a "quota." OK, that certainly makes a big difference to advertisers I bet.

Having proper "quotas" is essential to advertisers and to MRC accreditation.

In it's simples form, if a market is 51% female and 49% male, then the sample should be 49% male and 51% female.

As you drill deeper, let's say 18-24 is 7% of the population and there the male-female is 48% women and 52% men and 12% of those have household income of $50 k or more. Then the whole sample would be 7% from 18-24, and that 7% should mirror the gender and income (and geography and ethnicity and Hispanic language preference and educational level and so on...) of the market.

So, if you have a sample of 1000, going back to the start, you would have 510 females in the sample and 490 men. That's how quotas work. And that is why coming very close to mirroring the actual population is essential for advertisers to rely on ratings.

 

[davideduardo]

I wonder about Reno and SLC, and their Spanish Language stations rankings now.

Neither market qualifies for language weighting as a variable due to small market size and smaller Hispanic populations. The Hispanic numbers will continue to be erratic in both places.

 

[holyradio]

DavidEduardo,

I like the quote you use in your Radio-Info.com profile..."half of the truth is a whole lie". This new language preference “quota” has really screwed up the general market. Both KISF and KWID are up over 20% since the Holiday book and the general market stations are paying the price. The whole truth is that Arbitron would not have spent the money to implement language preference if they were not boycotted and pressured by Spanish Language companies. If this new system was so important in order to get an accurate sample, Arbitron would be executing it in every market around the country (not just the markets where the Hispanic population is above 10%). And you don’t think Spanish Language companies have any influence over the MRC???

 

[MC]

Reno back in 2000 had 19.2% Latino and 29% Mexican American, although the surrounding market might be different. Must be more by now.? So looks like we will see some ratings changes in the future.

"The racial makeup of the city was 77.5% White, 2.6% African American, 1.3% Native American, 5.3% Asian, 0.6% Pacific Islander, 9.3% from other races, and 3.6% from two or more races. Hispanic or Latino of any race were 19.2% of the population. Mexican Americans made up 29.0% of the city's population." factfinder census.gov

The number of listeners a radio station drives to the Internet, facebook, etc. would be important as well to advertisers.

 

[davideduardo]

DavidEduardo,

I like the quote you use in your Radio-Info.com profile..."half of the truth is a whole lie". This new language preference “quota” has really screwed up the general market.

Language preference proportionality is not new. It has been done for over a decade in all the major Hispanic markets except Las Vegas.

All that achieving proportionality on both Spanish dominant and English dominant Hispanics does is insure that the total sample has the right percentage for each; the total Hispanic sample does not change at all.

Both KISF and KWID are up over 20% since the Holiday book and the general market stations are paying the price.

Nobody is "paying the price" since reality is the object. The truth is that Arbitron was including too many English dominant Hispanics in the PPM panel and not enough Spanish dominants. So, once the correct balance was struck, we have a better view of the behaviour of the entire market, including its subsets.

The whole truth is that Arbitron would not have spent the money to implement language preference if they were not boycotted and pressured by Spanish Language companies.

The "pressure" to use language preference as a stratification variable began back in the mid to late 90's and it came from this simple fact: with growing Hispanic populations, it became apparent that Spanish language stations would jump and drop by enormous percentages from book to book... meaning that the sample was always out of whack.

The MRC decided to intervene, and leading Hispanic targeted broadcasters such as Heftel, Tichenor, SBS, Radio Unica, Excel and others started documenting the lack of programming based logic in the immense wobbles.

Arbitron accepted the need to get proportionality on language preference, particularly since English dominants tended to have a higher response and return rate and thus got more than their share of representation. But they could not implement for several years since the old Control Data mainframe software which had been ported to the current platforms had run out of fields to use for new stratification variables and had to be totally rewritten, something that took nearly 3 years to do.

If this new system was so important in order to get an accurate sample, Arbitron would be executing it in every market around the country (not just the markets where the Hispanic population is above 10%).

Arbitron buys the data from Nielsen. Nielsen only does language enumeration in the markets we know about... they just started doing Las Vegas last year. Enumeration is very costly and is not available from any other source. So many markets don't have it because the data does not exist.

And you don’t think Spanish Language companies have any influence over the MRC???

Broadcasters and advertisers alike have the ability to make their views known to the MRC. But the MRC, which was created in part to allow advertisers to know that ratings are done properly and in part to keep the Federal Government out of the ratings business, is very impartial and motivated only by good methodology well implemented.

 

[davideduardo]

Reno back in 2000 had 19.2% Latino and 29% Mexican American, although the surrounding market might be different. Must be more by now.? So looks like we will see some ratings changes in the future.

The metro survey area is 21% Hispanic, or about Hispanic 100,000 persons. But this is market 124, it is only a 2-book non-continuous measurement market with only $20 million in radio revenue for 27 commercial stations. Nielsen does not do language enumeration there, so Arbitron won't do language weighting. They do perform procedures so that the total sample is proportional for Hispanics, but not on the language subsets.

 

[holyradio]

David, since you seem to know everything about this issue... I thought you might know the answer to this question. Do you have to be a legal US citizen to be on the Arbriton PPM panel in Las Vegas?

 

[davideduardo]

David, since you seem to know everything about this issue... I thought you might know the answer to this question. Do you have to be a legal US citizen to be on the Arbriton PPM panel in Las Vegas?

Nope.

You don't have to be a citizen to be in the US. You can be a legal resident, too (with a "green card" which is, actually, pink). Or a tourist. Or a student.

You don't have to be a citizen or even a legal resident to be counted in the US Census. You just have to live in the US.

Since Arbitron uses a derivative of the Census (processed by Claritas) to define the "universe" of each market, it's pretty obvious that neither citizenship nor legal residency would be a requirement. Having some kind of permanent residence is, but only so that the docking of the meters can take place and so the meters can be attributed to a particular sampling unit. Otherwise, anyone is eligible.

 

[MC]

A web search seems to indicate a lawsuit was settled out of court.

Quota, dominant, this market, not that market, etc., it all sounds like a bunch of "Greek" to me.

People take ratings with a grain of salt to begin with, especially when it comes to Spanish language stations.

 

[davideduardo]

A web search seems to indicate a lawsuit was settled out of court.

There was no lawsuit.

Quota, dominant, this market, not that market, etc., it all sounds like a bunch of "Greek" to me.

Every poll, every panel, has a quota on every variable such as age groups, gender, ethnicity, income, education, place of residence, etc. The quotas are set up so that the survey is a very precise match, in miniature, of the group being studied.

The reason why small samples are used is the same reason why the Census only gets taken every 10 years: cost. The radio industry needs a survey many times a year, not every 10 years. So instead of consulting each person multiple times a year, a representative sample is consulted... because that is the only affordable way to do it.

Yes, this is Greek in some ways. The keyword is Epsilon.. gooogle epsilon +statistics to see.

People take ratings with a grain of salt to begin with, especially when it comes to Spanish language stations.

People in the advertising industry, for whom ratings are done, don't take ratings with "a grain of salt." They invest many billions of dollars a year based on ratings.

When the proper stratification variables are proportional, the ratings for Spanish language stations are just as reliable as those for any other station of comparable audience size.

 

[MC]

No lawsuit? Guess the NY Times had it wrong:


http://www.nytimes.com/2009/01/08/nyregion/08arbitron.html

 

[davideduardo]

No lawsuit? Guess the NY Times had it wrong:


http://www.nytimes.com/2009/01/08/nyregion/08arbitron.html

The context was "the radio industry" and there was no lawsuit by the industry or any part of it (and, certainly, none in Nevada, the subject of this section). Theonly "lawsuit" was what even those of us involved in the debates about the quality of the initial PPM sample would call "frivolous." It was a grasp for power or publicity by the NY State Attorney General's office... no stranger to such meddling in things they knew little about... and the same office in the state of New Jersey (where Sen. Menéndez, who got dragged into this hails from).

 

[Pinegroves]

You guys are getting too technical for me. I gotta go to my Rogets Thesauras to understand what your talking about. Did I spell that right? That's like your Funk and Wagnalls. Let's start talking about some ES stations that are doing well like KOMP, Coyote and whoever else has made some significant strides in this PPM market.