Quote: JB Down Under "There isn't 20 million households (4k of 20mill is 0.02%!) hence why it is closer to 0.05%.'"
If there's an average of 4 people per household (ie. 16k people covered by the 4k recorders, bearing in mind multiple TVs etc), then it can come out as a little over 0.05%, but not by much.
Quote: JB Down Under "As a methodology and sample group size it is extremely poor and would be laughed at if such a piece of research was put forward in a peer reviewed journal.'"
The methodology isn't disastrous, but it's not great. The problem is, like you say, the small sample size - but even then, if they could actually get a representative sample (difficult) then the figures are not necessarily useless, they're just an approximation. Probably with rather large error bars!