Los Angeles ratings for the month of May???

[davideduardo]

K.M. suggested rating the foreign-language stations separately from the English-language stations. Since AM radio listenership is so much lower than FM radio listenership, I'm wondering if it would be a good idea to also rank AM stations separately from FM stations.

In some places, AM is rated separately. In this hemisphere, IBOPE in Argentina separates the two bands.

One of the disadvantages of doing radio ratings this way is that the total reach of radio can not be measured since a big majority of people use both bands, but there is no "all radio" data to be had.

 

[davideduardo]

What if the listening pattern was real? No one has been able to prove otherwise.

Per the notice that Nielsen released and the angular view at the actions taken by the MRC, the KXOS household was removed because it was not in compliance with Nielsen standards. That, in other words, means the "pattern" was not real.

 

[davideduardo]

I find myself wondering if, in markets such as L.A. with a substantial number of stations programming in a specific language other than English (Spanish, in this case) there shouldn't be a separate set of ratings for those stations. I woudl think that with the PPM technology it wouldn't be difficult to do, and it would end the "weighting" problem David talked about above.

The problem here is that the market has several million persons who listen to both English language and Spanish language stations, so because of that co-mingling of about 20% of the market, you couldn't separate books based on language. Further, the percentage of those users varies somewhat over time, so neither population base for English only, Spanish only and Both changes regulary.

In any case, weighting is used on every subset of the ratings sample... including age cells, gender, income, education, ethnicity, language usage and geographical location. That's because there is no true "perfect" sample or panel, and even when you get one that is as perfect as possible, many of the subsets will be slightly under proportionality and others will be slightly over. So weighting is used almost everywhere, but general in rather small amount, to make the sample, after weighting, perfectly representative of the population in general.

You did remind me, though, of the Arbitron special survey of a few years ago... in diary days... of Asians. The universe was defined as Asians and they only surveyed members of that group, using techniques that included recruiting and follow up in any of the major Asian languages. It was, I believe, only done once and paid for by Art Liu's MultiCultural Broadcasting.