I noticed the phenomenon while flipping back and forth between Fox and CNN, trying to get war news. Fox's rah-rah news coverage looks and sound like a Saturday morning GI Joe commercial, and CNN's coverage seemed to focus almost exclusively on Aaron Brown's ability to emote on various topics, which forced me to switch back and forth between the two so as to avoid toxic levels of either. When something newsworthy would happen, though, I would stay riveted to one channel until the commercials came. I would then switch immediately to the other news channel. After a couple days of intermittently keeping up with the news - whenever the Internet revealed that something interesting was happening - I slowly realized that I hadn't seen any TV commercials. I had flipped over to alternative news channels whenever they started up.
Having just recently written a column on why advertisers shouldn't have qualms about keeping their media deals in place during the war, I began to get concerned that this wasn't perhaps the best advice. So I called up several friends in different parts of the country to ask them about their own viewing habits, both online and offline. The key issue was whether or not normal news watchers (as opposed to people like me opportunistically rushing in and out of news watching) follow a similar pattern during dramatic live events.
What I found was that most of them had a favorite news channel that they watched without switching much, until such time as something dramatic and live was happening. Then they all tended to switch around with abandon. This applied both online and offline. On the Internet, as the ads aren't generally interruptions in the content, there didn't seem to be much of an observed effect on their attitudes toward the creative (they all still claim to never notice banners, just like they claimed before). On TV, though, they appeared to never even see ads on news channels during the more exciting times, as the switch before they can run.
No one felt that non-news television programming was affected one way or the other.
I don't know precisely how a media buyer could avoid this sort of news event trap, and pulling ads during wartime seems like the wrong tool for the job. And certainly my unscientific "study," (with a sample number of 6) isn't exactly the ironclad research one would want to have to justify any changes in policy.
Perhaps a clause could be inserted in contracts, disallowing ads from running after live foreign coverage, as opposed to recorded content. The other alternative would be to switch media buys to non-news programming. Better still, buy news programming in a medium where users don't twitch back and forth, like print or the 'net.