Via Lost Remote, MediaDailyNews has posted a news article claiming that Nielsen's minute-by-minute data and "commercial ratings" (i.e. ratings during commercial breaks) indicate that almost no one watches ads when a show has been recorded on a DVR.
While this conclusion falls nicely in line with my preconceptions (I wrote a white paper on product placement as a way of sidestepping DVR ad-skipping last semester), the evidence as laid out in the article doesn't provide much support for the argument it lays out:
"American Idol" posted a 12.1 in "live" numbers during commercials and a 12.2 in "live plus seven day," meaning that less than 1 percent of DVR viewers stopped for the ads. The same tiny increase goes for "Desperate Housewives," which saw a 10.2 during commercial breaks in "live" viewing and a 10.3 in DVR viewing.
Note that, as I understand ratings points, this is a deceptive claim. A 0.1 ratings increase indicates that only 1 in 1000 TV-watching households in the US watched the "American Idol" ads on DVRs... but that has no connection to the percentage of DVR viewers who skipped the ads when watching "American Idol" on their DVR. To get that number, we would need to divide the # of "American Idol" viewers who watched the ads on DVR by the total # of viewers who watched the show on DVR (and multiply by a hundred). Perhaps MediaDailyNews' source material actually did this math. I don't know.
The problem, however, is that you can't tell whether the "less than one percent" statistic quoted is legitimately derived from the information in the article. I'd prefer it if I could back up my position with hard numbers on a sensitive issue like this, instead of being fed (potentially) bogus interpretation by a news outlet. Until we get some better sourcing on a statistic like this, I'd strongly suggest that readers challenge anyone who claims that less than 1% of DVR users watch ads.