Conducted after the first Scottish leaders’ debate on STV on Tuesday, but with much of the interviewing completed before the second instalment on BBC on Wednesday, the poll must have come as a bitter disappointment to the Labour party. Far from suggesting that the party was finally making some progress in clawing back the SNP’s large lead, YouGov reported that in fact that lead was larger than had been recorded in any of its previous polls. The SNP were credited with 49% of the vote, up three points on the company’s previous poll conducted at the end of March while Labour were estimated to be on just 25%, down four points.
A movement of that size is rather bigger than would normally be expected to occur as a result of the chance fluctuations to which all polls are subject, though we might still want to suspend our judgement and await further polling evidence before presuming that the SNP’s lead has actually grown. In the meantime today’s poll edges the SNP vote in our poll of polls back up from 45% to 46%, while Labour remain on 28%.
The prospect that the SNP might hoover up the vast bulk of Scotland’s representation on little more than 45% has led to quite a lot of speculation about the possibility of anti-SNP tactical voting. Might the SNP’s haul of seats be significantly diminished if, for example, Conservative and Liberal Democrat voters were to vote Labour in seats where Labour and the SNP were locally the principal combatants, and if voters switched in a similar fashion to the Conservatives and the Liberal Democrats where they had the best chance locally of defeating the SNP?
The Channel 4 portion of the poll tries to assess this possibility. Respondents were asked for whom they would vote if they thought that only Labour and the SNP had a realistic chance of winning in their constituency, and similarly what they would do if the contest locally were between the SNP and the Conservatives, and the SNP and the Liberal Democrats.
At first glance the poll suggests that quite a lot of anti-SNP tactical switching could occur. For example, if convinced the battle locally was between Labour and the SNP, a half of Liberal Democrat (50%) supporters and almost as many Conservatives (44%) said they would back Labour instead. In each case only around one in ten said they would switch to the SNP.
Conservative supporters would apparently be even keener on switching to the Liberal Democrats if they had the best chance of defeating the SNP (58% say they would do so, while only 8% would move to the SNP). The feelings are though only partially reciprocated by Liberal Democrat voters faced with a Conservative/SNP battle, who would switch to the Conservatives rather than the SNP by a ratio of approximately two to one – 39% say they would vote Conservative, 21% for the SNP.
However, the mood amongst Labour voters is very different. If they thought they were living in a constituency in which the Conservatives were challenging the SNP they are almost as likely to say that they would switch to the SNP (30%) than they are to the Conservatives (31%). The two parties might have been on the same side in the referendum, but many a Labour voter evidently still has considerable antipathy for the Conservatives. The picture is much the same when Labour voters are faced with the prospect of a Liberal Democrat/SNP battle; while 34% say they would switch to the Liberal Democrats, 27% say they would back the SNP.
Still, at least the apparent willingness of Conservative and Liberal Democrat voters to vote against the SNP might help ensure that some seats remain beyond the SNP’s grasp. And that indeed does indeed appear to be the case – though perhaps not to the extent that Peter Kellner anticipates in his analysis of the YouGov/Channel 4 poll.
If the changes in vote share since 2010 implied by the YouGov poll were to be replicated everywhere, the SNP would win 54 seats, Labour 4 and the Liberal Democrats one (unlike YouGov and The Times I project the Tories to lose Dumfriesshire under this assumption). That projection also gives an estimated vote for each of the parties in each constituency. We then assume that in every seat where a party is heading for third place or less, the supporters of that party switch to either the SNP or their opponents in the proportions identified by the poll. On that basis seven of the 54 seats that the SNP are projected to win might be prised from its grasp, with Labour picking up four of them, the Conservatives two and the Liberal Democrats one.
However in three of these cases we are assuming that a party that is currently either first or second in a constituency would accept that their party was now lying third and no longer has the best chance of defeating the SNP. That might be thought improbable. And if it does not happen the tally of seats the SNP might fail to win would be no more than four, two of them picked up by Labour and one each by the Conservatives and the Liberal Democrats.
Tactical voting may help deny the SNP a few seats, but in truth it seems incapable of recreating on May 7 the unionist coalition that succeeded in defeating independence last September.