Originally Posted by Fudd:
“How many of those in the consolidated viewers were new viewers and how many were cult lovers who adored the episode so watched it over again?”
Originally Posted by Fudd:
“I know my mate (a real obsessive!) watched it again about 7 times; I watched it another three and have watched Day of the Moon twice again so far. Ask across this forum; I'm sure quite a few people have watched at least once again on iplayer if not more.
As we do not know the breakdown of new viewers v repeat viewers the consolidated figures will hold up well, and do the programme a good service. But, as much as I'm loving the show at the moment, I'm disappointed such a dent has occured in a two parter.”
All of them were new viewers. You're confusing the consolidated viewership (that's everyone who watches a specific broadcast either live or recorded, but with any individual viewer only counted once), with the weekly reach (which can include iPlayer and BBC Three/HD showings, and where an individual can, in practice, be recorded as viewing more than once). The consolidated figure of 8.86 million is entirely unique viewers.
All the multiple viewings, via iPlayer or other means, of you and your mate, would make no difference to that ratings figure whether you were part of the BARB panel or not.
Originally Posted by mossy2103:
“I see your point, however, you would then question any official consolidated figure then. In other words, by your logic (logic that I am not saying is wrong at this stage), every consolidated figure (for every programme, not just DW) would be subject to the same "error". Which still means that we end up comparing like for like (between the same programme, between different programmes in the same series or different series, and between different programmes entirely. In fact, even overnights would be subject to the same problem id a PVR recording is watched or re-watched on the same day
Which in my mind makes it a bit of a red herring, or is something that is catered for within margins for error or similar. Maybe someone with ratings knowledge could comment?”
See above.
Quote:
“After all, these are just figures from statistical analysis of 5,100 households, and about 11,300 viewers at the end of the day, not a comprehensive survey of all UK households.”
Actually - that's pretty much what it is. A very large and comprehensive, carefully selected demographically representative panel of viewers with redundancies and checks built in. You could make the panel even bigger, and still not get a noticeably more accurate reading of the public at large. Plus, as was once famously demonstrated in a US presidential election, try to sample above a certain level (even up to the point of sampling everyone) and your accuracy can actually fall, hilariously so.
General elections, with millions of voters, can be predicted with a tiny margin of error via polls of just a 1000 people. When we're dealing with polling of fact (what people watched) rather than opinion or intention, and with a panel ten times that size, then there's little to worry about in terms of accuracy.