Of the four Scottish constituency polls conducted by Ashcroft in Liberal Democrat-held seats, by far the most damaging for them was West Aberdeenshire and Kincardine. At least in the other three they were in second place (albeit a very distant second place in two cases), and so didn't have to worry about being undermined in their campaign message of "It's a straight fight between the SNP and the Liberal Democrats here! Only the Liberal Democrats can stop the SNP and their crazy dream of making Ed Miliband the Prime Minister of an independent Scotland!" In West Aberdeenshire and Kincardine, though, they had slipped into third place behind both the SNP and the Tories, effectively killing that message stone dead. So in an attempt to get back in the game, they've released their own self-conducted poll of the constituency, which - hey presto - shows them slightly ahead. The Press & Journal have dutifully swallowed the Lib Dem propaganda in reporting the poll, so here is a helpful cut-out-and-keep guide to the inaccuracies, distortions and omissions in the article -
1) It's not - as repeatedly claimed - a "Survation poll". Survation carried out the fieldwork and then passed the raw data on to the Lib Dems to do whatever they liked with, but if it had been a "Survation poll" it would be subject to British Polling Council rules. Survation would be responsible for deciding which weightings to apply, and also for publishing the datasets on their website. Instead, this is a Lib Dem poll, which the Lib Dems have selected the weighting procedures for. It is not subject to BPC rules, and therefore we will presumably never see the datasets.
2) There is no reference whatsoever to the fact that all Lib Dem-conducted constituency polling commits the cardinal sin of asking leading questions prior to the main voting intention question. If you ask people questions designed to get them to say what a jolly decent chap their local Lib Dem MP is, and only afterwards ask them how they plan to vote, you make it much harder for them to admit that they are planning to vote against the Lib Dems, because that answer would seem logically inconsistent.
3) There is also no reference to the fact that the Lib Dems apply an even more extreme form of the discredited 2010 recalled vote weighting to their results than Ashcroft does, which almost certainly means that the SNP vote in the constituency has been downweighted on an industrial scale. Given the pre-weighting and post-weighting pattern we saw in Ashcroft, I would be astonished if the SNP weren't ahead in the raw data that the Lib Dems received from Survation.
4) A great deal of fuss is made over the fact that the Lib Dem poll gave respondents the names of the local candidates while the Ashcroft poll didn't, but in fact the approach taken by both polls is flawed. If it's true that naming the candidates produces the most accurate results (and that is by no means clear) then it should be done in a single voting intention question asked before any other questions. Both the Ashcroft and Lib Dem polls asked the headline voting intention question later in the sequence.
5) The P&J makes us feel all nostalgic by noting that this poll will be a "blow for the SNP". Well, let's look at it this way. There were two polls conducted in the constituency more or less simultaneously. The first, for all its shortcomings, was conducted impartially, and showed the SNP miles ahead, with the Lib Dems in a distant third. The second poll was conducted by the Lib Dems themselves, and showed the Lib Dems ahead. The P&J may think there's some kind of dilemma in deciding which of the two polls to put more faith in, but I'm not sure many in the SNP will share that view.
However, even if a few people are foolish enough to take the Lib Dem poll at face value, the result actually suits the SNP quite well, because it muddies the waters of the whole anti-SNP tactical voting issue. Are the Tories or the Lib Dems the real challengers to the SNP in the constituency? Take your pick, chaps, but it'll be a leap of faith.