Then shouldn't the data for the meter wearers on short trips to, say, Los Angeles, be counted in the Los Angeles market ratings for the time they're there? This seems like a major design flaw if it goes with the person, regardless of where they are for a week or whatever.
When the "panelist" is within the LA market, and their listening registers LA stations, why wouldn't that add to the survey sample in the LA market, even if it's only for a day? That's different than someone DX-ing a distant signal, like listening to KGO in Pasadena at night, which would make more sense (regardless of what's being programmed on KGO now!). I don't think the FM stations from LA, mentioned earlier in this thread, could ever be heard "over the air" anywhere in the Bay Area to be picked up by a people meter. Even if it was a rare tropospheric signal scattering kind of day. And even if they could once a year or so, I seriously doubt that has anything to do with them showing up in the SF market data.
And if it's online webstream listening, I thought the Arbitron listings indicated that separate from the broadcast signal.
So what gives with Arbitron? Are they collecting market data or personal monitoring history of the chosen few?