Talk:Opinion polling for the 2015 United Kingdom general election/Archives/2015/February

TNS-BMRB phone poll 23-26th Jan (BBC R4 Woman's Hour)
Still appears on TNS official website. Says "TNS conducted a telephone poll between 23rd - 26th January 2015 to better understand UK voter intention.". best, Sunil060902 (talk) 01:13, 4 February 2015 (UTC)


 * I don't edit Wikipedia, but I'm very surprised that the Woman's hour poll is included in the main body of the table. I don't know if the poll is primarily focussed on women. It may be further focussed on women who listen to R4 Women's Hour, which must be a tiny and inevitably unrepresentative sample of the population.
 * If Scotland and Wales belong in separate tables, then surely so does this narrow poll. Even its results show it statistically way out of line with trend. It just doesn't belong in the main table.
 * If this result is included in the 2 week moving averages, it will in itself misleadingly boost the Labour average lead. — Preceding unsigned comment added by 88.104.254.247 (talk) 19:09, 4 February 2015 (UTC)
 * It appears to be a genuine poll carried out in the normal way, but not politically weighted - which is unusual for a political poll. There are tables which show how people recalled voting in 2010, but then no evidence those weightings have been used later to balance the poll. So whether it counts as a genuine voting intention poll is a moot point. I'd say not, but there is no guidance from TNS as to whether it should be considered a VI poll or not. Note that the more recent TNS (Global) poll has a 6% Labour lead, this is not as out of step with that poll as it appears to be with others. Given that we have so many polls, assuming it is a one-off it will only have a limited effect - by the beginning of next week the 20-poll average used for the graph on this page will have stopped including it. Saxmund (talk) 20:09, 4 February 2015 (UTC)
 * I would have to say it is not a real poll, since there have been no other plying by this group before so I believe it should not be included. What is to stop someone adding a poll that they conducted in their university which many editors would view as not relevant to be included. This poll should not be included. SleepCovo (talk) 10:46, 6 February 2015 (UTC)
 * Have reverted your change as whatever else the poll may have wrong with it, TNS-BMRB are a full member of the British Polling Council. See list of company members here: http://www.britishpollingcouncil.org/officers-members/ Perhaps you were confusing the client (BBC Radio 4 Woman's Hour) with the polling company who carried it out (TNS-BMRB)? Andrewdpcotton (talk) 11:36, 6 February 2015 (UTC)
 * In any case, we include non-BPC members. Ashcroft polls are included, and in the past BPIX have been. I note that some other pollsters do not past-vote weight their polls, including Opinium and Ipsos-MORI. Saxmund (talk) 21:58, 7 February 2015 (UTC)

Include it. I think it's hard to justify excluding it in the absence of information that it was from a partial audience. Excluding it doing so would be our bias, which we can't do. Note that TNS's polls generally seem to give Labour better figures than other polling outfits. It could be that they're getting something wrong. Or they could be getting something right. It's not up to us to say. --Dweller (talk) 10:53, 6 February 2015 (UTC)
 * Agreed: this appears to be a genuine poll of a representative sample and so we should include it. Bondegezou (talk) 11:10, 6 February 2015 (UTC)