Well, Claritas is just plain wrong about Midland & Odessa. The Hispanic population is a lot higher than a third, especially in Odessa.
That's entirely possible. Claritas could be wrong.
However, let me give you one example from the 2000 Census. The 1999 estimates from the Census bureau were of by nearly 360,000 from the actual Census count in 2000. The estimates are done by statistical techniques, not by polling or such. On the other hand, the Claritas estimates, using the Census data and other sources like total homes in the tax base, vehicle registrations, electric accounts, etc., was off by under 100,000.
The fact is that the Claritas estimates for 1999 were very close, in every market, to the actual Census count; the Census in most high growth markets was waaaaaaaaaaaay off.
In any case, error or not, the basis for Arbitron numbers is the 12+ population of the Metropolitan Survey Area, defined by Arbitron, and updated annually by Claritas in the Fall book period.
So far, in a bit over 37 years of dealing with Arbitron, I have not heard any criticism of the population counts for the MSA's. There is much criticism of the definition of the MSA's themselves, mostly from stations that do not cover the entire MSA.... but of the counts themselves, I hear nothing. Maybe this is because Claritas is the premiere demographer of the US, and companies from P&G to Microsoft use them for marketing purposes.
It's entirely possible for Claritas to be off by more in one or two markets than the rest of the country, but the liklihood, based on experience (MSI was the company Claritas bought to bring in high-powered demography) with MSI and Claritas, it has not been seen before.
And, as I said, radio ratings are based on the Claritas data, so what you see in ratings is based on those populations, to which proportionality is obtained by diary and weighting. Were there more Hispanics in that market, it would not change the ratings as each diary keeper has weighted equal value in the sample.
