It’s month five for our rolling sequence of Survation polls, conducted as always in partnership with the Daily Record and Dundee University’s 5 Million Questions. The June results are here, and the Record have the indyref results. The big question shows 47% Yes, 53% No again, the same as last month’s result. Having said that, last month Yes’s 47% was 46.6% rounded up, and this month it’s up to 47.1% rounded down, with No correspondingly down from 53.4% to 52.9%. That makes for an unchanged headline figure, but the No lead at one decimal place has fallen from by 1% from 6.8% to 5.8%. Confusing, but that’s rounding for you.

Onto the Holyrood results. Usual background: I’m comparing vote shares to the previous month’s figures: but seat numbers are still shown as the change on the 2011 result. Seat projections continue to be from Scotland Votes, who don’t include UKIP in their methodology. The ‘kippers would be expected to win a small number of regional list seats at this level, although it remains unclear at which party’s expense those gains would come (roughly likely to be in proportion to list seats, i.e. costing Labour most, then Tories, then Greens). With all that in mind, here are this month’s figures.

Parties Constituency Region Total
Vote share (+/-) Seats (+/-) Vote share (+/-) Seats (+/-) Seats (+/-) %
SNP 44.1 (-2.1) 53 (±0) 36.9 (-2.4) 7 (-9) 60 (-9) 46.5
Labour 30.6 (+2.3) 15 (±0) 25.7 (-0.5) 22 (±0) 37 (±0) 28.7
Conservative 13.3 (+0.3) 3 (±0) 12.9 (+2.5) 13 (+1) 16 (+1) 12.4
Liberal Democrats 5.1 (-1.1) 2 (±0) 7.3 (+1.2) 5 (+2) 7 (+2) 5.4
Scottish Greens 1.9 (-0.6) 0 (±0) 8.1 (-1.9) 9 (+7) 9 (+7) 7.0
UKIP 4.1 (+0.9) 0 (±0) 8.1 (+1.1) 0 (±0) 0 (±0) 0
Others 0.7 (+0.2) 0 0.9 (-0.1) 0 (-1) 0 (-1) 0

Screen Shot 2014-07-11 at 01.00.05The first oddity is that this would see every last constituency go the same way as 2011. Although Labour are a bit up on last month, it doesn’t win them any more seats: they are at best treading water on these results. The SNP, however, would be down enough to lose their overall majority, and, as per the May result, would either need to run a very strong minority administration, or look for any other party to form a coalition with them. Despite that minor dip on 2011′s landslide, it’s an extraordinarily strong position for a governing party to retain more than seven years after taking office.

Looking at the smaller parties, it’s been a better month for the Tories and to a lesser extent the Lib Dems: both would be marginally up on their 2011 score, with the Tories now in a clear third place (last month they were just 0.4% ahead of the Greens on the list). As for the Greens, they’re 1.9% down on the list, and would elect two fewer MSPs than June’s poll indicated. I still think Patrick Harvie and Alison Johnstone would be pretty pleased to have seven more colleagues, though. With these numbers the second slots on the Lothian, Glasgow and perhaps even Highlands and Islands lists would be promising places for Green candidates, and selection will be competitive.

Making sure the panel is just right.

Making sure the panel is just right.

Before that, though, the small matter of the indyref. There’s been a pretty rough squabble about how to poll that. Are Survation right, or are YouGov? Well, YouGov were the most wrong about the AV referendum, the most recent similar vote. And in 2011 YouGov underestimated the SNP constituency vote by more than 3% and their regional vote by more than 9%, well outside the margin of error. You can even get odds on which side of the argument will be vindicated in September.

Sure, I’m biased, given Survation are our house pollsters, but their methodology is transparent, unlike YouGov’s. The latter have a weighting system for “red Nats”, but won’t say what it is, nor whether other segments are weighted for. More generally, Kellner’s argument, despite YouGov’s substantial underestimate of SNP votes in 2011, is that Survation have the wrong sort of SNP voter in their panel.

If you’re still not sure who to back in the battle of the pollsters, here’s a wee graph from @bgreysk on Twitter (precedes this month’s Survation result). The trend lines are the best guide, and from that YouGov look like the complete outlier. On this evidence, I think Ladbrokes would be easy to take to the cleaners given they’re offering 7/4 on Survation to be closest, but any bets are of course to be made at your own risk.