SNP hold 36% lead in talebearing TNS poll

The monthly TNS poll is now out, and I'm beginning to think Kezia Dugdale should just go the whole hog and change her party's name to "Labour 21%", because this is the fifth poll in a row from a range of different firms to give Labour exactly the same constituency vote share.  However, that's very slightly misleading, because TNS have suddenly switched to using headline figures that are filtered by certainty to vote.  Using a like-for-like comparison with the filtered figures from last month, Labour have actually 'recovered' from 20% - which of course is a statistically insignificant change.  At this stage in proceedings they might just be grateful for any indication that they are not slipping even further (as the recent Panelbase poll suggested they were), but the snag is that their fate will almost certainly be wholly decided by the regional list ballot, where their previous double-digit advantage over the Tories has been abruptly wiped out.

Constituency ballot :

SNP 57% (-2)
Labour 21% (+1)
Conservatives 17% (+5)
Liberal Democrats 3% (-1)

Regional list ballot :

SNP 52% (-1)
Labour 19% (-2)
Conservatives 17% (+6)
Greens 6% (-3)
Liberal Democrats 6% (+1)

As we used to constantly discuss during the independence referendum, figures that are filtered by certainty to vote are more prone to volatility, because the sample size is smaller, and because there may be a different demographic mix from month to month among those who say they are certain to vote.  So it's important to stress that the advances made by the Tories in this poll would be significant, but not nearly so dramatic, if TNS hadn't switched to using the filtered figures for their headline results.  On both ballots Ruth Davidson's party have made 3% gains on the unfiltered numbers.  That merely takes them back to 15% on the constituency ballot (exactly where they were in the first TNS poll after the UK general election), and still leaves them a relatively substantial 5% behind Labour on the regional list.

It's therefore still open to debate whether TNS have yet fully come into line with other pollsters in showing a mini-renaissance for the Tories.  However, the evidence that the party have at least made some progress recently is now so extensive that it's impossible to discount.  And of course TNS have switched to the filtered numbers for a very good reason, ie. they think that's a better indication of how the election will actually turn out (or would have turned out if it had been held at the 'snapshot' moment that the poll covered).  So although the finding that Labour and the Tories are now in a statistical tie for second place on the list vote should be treated with some caution, it also can't be easily dismissed, especially as it's entirely consistent with the results from other firms.

TNS had always been Labour's warmest comfort blanket - as long as one of the two 'real world' pollsters were consistently showing them with a big advantage over the Tories, it was possible to explain away the last Ipsos-Mori telephone poll as a freak result, and to imagine that other firms were only showing an alarming narrowing of the gap because volunteer online polling panels contain too many politically committed people.  But that comfort blanket has now largely gone, and for that reason I and others will for the first time have to start taking seriously the idea that the Tories may genuinely be in with a shout of replacing Labour as the main opposition party.

I still think it's unlikely, though.  To date, not a single poll from any firm has put the Tories in second place.  The other crucial point that has to be borne in mind is that polls tend to exaggerate the final difference between the constituency vote and the list vote (perhaps because a minority of respondents wrongly gain the impression that they're being asked for a second preference vote).  So although the list ballot is the only thing that really matters for Labour, in a perverse way they are still entitled to take some heart from their slightly bigger 4% cushion on the constituency ballot.

This poll may appear to be a significant blow for the Greens, but in fact they're experiencing a straightforward reversion to the mean.  6% is entirely typical of the vote share TNS has been giving them since the general election, meaning that the spike in their support last month now looks very much like a freakish finding caused by sampling variation.  We know that polls tend to overstate Green support on the list, so on current trends they seem to be heading for another disappointing result.  The good news for them, though, is that in both 2007 and 2011 they did better in the eve-of-election polls than they were doing a few months out.  Their average vote share in the winter of 2010/11 was a little lower than it is now.  So on past form it's not inconceivable they will gain a few percentage points between now and May, see those gains wiped out on election day itself, but still end up with a slightly better result than in 2011.  (And because of the quirks of d'Hondt, 'slightly better' could make the difference between, say, two seats and five.)

TNS still don't seem have caught up with existence of RISE.  However, just one person in the unfiltered sample of 574 said they were planning to vote SSP on the list, so that's probably a very good indication of where RISE find themselves with only three months to go.  Similarly, just one person said they were planning to vote for Solidarity, although we know that Tommy Sheridan has a concentrated core of support in the electoral region of Glasgow that polls may find difficult to detect.  But unless something dramatic happens, it looks like RISE will win zero seats, and that Solidarity will win either zero or one seats (with zero being the more probable outcome).

Incidentally, Aldo's oracular pronouncement last night about UKIP being on course for a Holyrood breakthrough is looking slightly dubious in the cold light of day - just one person in the sample said they were planning to vote UKIP on the list!

*  *  *

SCOT GOES POP POLL OF POLLS

The trend in previous updates of the Poll of Polls was distorted because of firms dropping out of the sample or returning to it.  This time there's a distortion because of TNS switching on their turnout filter, but it hasn't made a huge difference.  We might question the extent of the Tory gains, but there's not much doubt they've been creeping up.  From a psychological point of view, the most important change is that the Greens have slipped back into fifth place on the list.

Constituency ballot :

SNP 52.3% (-0.2)
Labour 20.8% (n/c)
Conservatives 17.0% (+1.2)
Liberal Democrats 5.8% (-0.2)

Regional list ballot :

SNP 47.0% (-0.5)
Labour 19.3% (-0.2)
Conservatives 16.5% (+1.2)
Liberal Democrats 7.3% (+0.5)
Greens 6.8% (-0.7)

(The Poll of Polls is based on a rolling average of the most recent poll from each of the firms that have reported Scottish Parliament voting intention numbers over the previous three months, and that adhere to British Polling Council rules. At present, there are four - Panelbase, Survation, TNS and Ipsos-Mori. Whenever a new poll is published, it replaces the last poll from the same company in the sample.)

*  *  *

It seems that the only supplementary question TNS asked in this poll was on the EU referendum - and irritatingly, they didn't use the actual referendum question, which we know from Ipsos-Mori's findings can make a telling difference.

If there was a referendum on the United Kingdom’s membership of the European Union, how would you vote?

Remain 44% (-3)
Leave 21% (+3)

The percentage changes are from the last time the question was asked in September.  A 6% narrowing of the gap may look significant, but it's difficult to know how to interpret that - clearly there still doesn't seem to be much danger of Scotland voting to Leave, even taking account of the fact that the electorate often proves to be more volatile during referendum campaings.  The Leave campaign may drawn some comfort from the result, although there's no reason to assume that trends in Scotland will run in parallel to trends in the UK as a whole, and indeed there is strong evidence in the Britain-wide polls that neither side has really gained much traction of late.

One thing that Scotland does have in common with the rest of Britain, though, is that referendum vote is strongly correlated with age and affluence.  The older you are, the more likely you are to vote Leave, and the wealthier you are, the more likely you are to vote Remain.  That's a double-edged sword for the Leave campaign, because older and more affluent voters are both more likely to turn out to vote.  The age effect on turnout looks somewhat more significant, though.  For the Holyrood election, TNS found that 84% of over-55s were certain or very likely to vote, compared to 72% of under-55s.  The difference in the equivalent figures for the "affluent" and "non-affluent" halves of the sample was just 6%.  So it could be that differential turnout in the referendum will work in Leave's favour.