But once again, you are wrong. Every person on any PPM or diary panel is worth a certain number of people. And that number literally changes everyday depending on how many people are holding their meters. Any Arbitron Rep will confirm this fact. That's how weighting works.
In a perfect panel, there is no weighting. That is because a panel is pre-recruited to perfectly match the composition of the group under study, and built in perfect synchronization with the market.
The only time, in a panel, where weighting has to be applied is if the panel does not perfectly represent the universe. But, for all practical purposes, the weighting on a panel is very very small... typically less than +/- 5%, which is less than the overall margin of error of a panel survey.
If the amount of weighting is significant, then there is an issue. But if it is in the low range of single digit numbers, it's insignificant and does not materially affect the outcome of the survey.
In Las Vegas, using the Spring 2011 data, there were 766 in tab meters for 1,571,000 12+ persons, or an approximate per meter value of 2,000 persons. For 490,000 Hispanics 6+ there were 247 meters, or approximately 2,000 persons per meter. For 776,000 women 12+, there were 386 meters and for 775,000 men 12+ there were 380 meters. If you go through all the cells, the per-meter value stays very close to 2,000.
There is none of the 10%... 15%... even 20% weighting up or down we saw with the diary. The weighting is so small that it is far less a factor than margin of error for the given sample size and is a factor that is almost irrelevant in a discussion of the PPM methodology... because, again, we are dealing with a predesigned panel, not a random probability sample such as the diary, phone samples, intercepts, etc.
And now, Spanish dominate PPM holders are worth more than English dominate. Once again, the proof is in the numbers as every Spanish language station is up from book to book.
The most desirable situation for a station is to have its target audience oversampled so that each panelist is "worth less " and weighted down. That means there are more meters in the sample for their listeners, so the margin of error... and wobble... will be less.
The value of a Spanish dominant Hispanic in a panel is the same as the value of an English dominant one... about 2,000 in Las Vegas. While there may be some difference in the months preceding implementation of language proportionality in that the panel, as it turns out, had too many English dominant Hispanics and some had to be dropped while more Spanish dominant ones had to be added. Overall, we are talking about a change of about 12 more Spanish dominants and 12 less English dominant Hispanics to restore proportionality to the survey and to make any weighting minimal or totally unnecessary.
Spanish language stations are at a disadvantage if the Spanish dominant panelists have to be weighted up ("worth more") because it means the reliability of their ratings is less. That's not a desirable situation.
Perhaps in a couple months, Arbitron will add more Spanish dominate(sic) panelist so they won’t have to weight them up so much. But that probably won’t change the outcome. Spanish language stations will still be much more dominate(sic), compared to where they ranked recently.
The amount of weighting is minimal, if you look at the sample report. And within a month or two, it will be negligable... in other words, less than "minimal."
