Inspirational Ipsos-Mori poll finds majority of Scottish voters still support independence

After our little sneak peek at the unfiltered Ipsos-Mori numbers earlier, here are the official headline results filtered by certainty to vote - and they're even better than we expected for the SNP.

Constituency ballot :

SNP 53% (+3)
Labour 20% (n/c)
Conservatives 16% (-2)
Liberal Democrats 6% (-1)

Regional list ballot :

SNP 49% (+3)
Labour 19% (n/c)
Conservatives 15% (-1)
Liberal Democrats 8% (n/c)
Greens 6% (-1)

Of even greater significance, though, is that Ipsos-Mori have asked the independence question.  This is only the third 'real world' poll (telephone or face-to-face) to test support for independence since the referendum - and, extraordinarily, all three have shown a Yes lead.

Should Scotland be an independent country?

Yes 52.1% (-2.8)
No 47.9% (+2.8)

This is obviously further confirmation (if any were needed) of the divide that has opened up on the topic of independence between 'real world' pollsters and firms that rely on volunteer online polling panels. Bizarrely, it's the complete opposite of the divide we saw during the bulk of the referendum campaign, when online firms (with YouGov sticking out like a sore thumb as the sole exception) tended to be much more Yes-friendly than their 'real world' counterparts.

Doubtless the likes of our occasional commenter "Roger Mexico" will continue to point out that there is no absolute proof that the divergence is caused by data collection method, because it just so happens that the two 'real world' firms (Ipsos-Mori and TNS) are also the only two firms that don't weight by recalled referendum vote - ie. it's possible that there are too many Yes voters from 2014 in the Ipsos-Mori sample, perhaps because they are more eager to take part in polls.  However, I'm extremely sceptical as to whether that's the whole explanation for the divergence, not least because the TNS poll showing a decent Yes lead was politically weighted, albeit by recall of party vote, rather than referendum vote.

Even when different firms seem to have an in-built 'house skew' towards one side or the other, it is of course still possible for all firms to broadly agree with each other on the direction of travel.  As this is the third (arguably the fourth) consecutive poll from an assortment of firms to show some movement towards No, an interesting question is whether we now have convincing evidence of a real swing, or whether it could just be an illusion caused by normal sampling variation.  The reality is that we don't have enough information to go on yet.  The swing showed by Panelbase looked like an innocuous reversion to the mean, because the previous poll had been unusually good for Yes.  The YouGov findings may have been more significant, because the Yes vote was below its recent 'normal range'.  But with Ipsos-Mori, it's impossible to reach any firm conclusions, because this is only their second post-referendum poll on independence, so there's no way of knowing whether a 52% Yes vote is 'unusually low' for the firm or not.

Just as Panelbase did recently, Ipsos-Mori have sought to test whether Britain's exit from the European Union would lead to any change in public opinion on independence.  They've found the same relatively modest increase in support for Yes that Panelbase did - but of course Ipsos-Mori's basic Yes vote is higher than Panelbase's, leading to a much more commanding Yes lead when Brexit is assumed...

Imagine that the UK as a whole votes to leave the European Union in the referendum when voters in Scotland vote to remain in the European Union.  If this led to another Scottish independence referendum being held, how would you vote in response to the question ''Should Scotland be an independent country''?

Yes 57.7%
No 42.3%

The exact figures shouldn't be taken too seriously, because people tend to be bad at working out how they would feel about a subject in hypothetical circumstances.  There's also a danger that some respondents may have felt that Brexit was "expected" to change their view on independence.  So probably the most that can be said is that it's encouraging that Brexit seems to nudge more people towards Yes than towards No.

On the EU referendum itself, Ipsos-Mori have found that Scottish voters currently plan to back the "Remain" position by an overwhelming margin of 62% to 26%.  STV's Stephen Daisley (who has refreshingly reverted to "non-zoomer analyst" mode, just for the day) has made a direct comparison between those figures and the most recent Britain-wide Ipsos-Mori poll showing Remain ahead by 55% to 36%.  However, that could be quite misleading, because the Britain-wide poll took place entirely before David Cameron's failure to secure a credible deal with his EU partners.  So the divergence between Scottish and UK public opinion could well be even more extreme (as indeed other pollsters have suggested it is).

To return to the Scottish Parliament numbers, the recovery in Labour's lead over the Tories for second place isn't as pronounced in the headline results as in the unfiltered numbers, and it's entirely caused by Tory slippage rather than by Labour progress, but it will still be a relief for Kezia Dugdale to see an up-to-date telephone poll giving her a 4% cushion.  It's debatable whether the Tories really have slipped, though - the jump in their support in the November poll looked implausibly huge.  Today's poll doesn't wipe all of those gains out, so we could simply be looking at a more realsitic estimate of the modest advances they have made over the last six months or so.

There's more bad news for RISE in this poll with time rapidly running out - just one person in the whole sample of 749 plans to vote for them.  That's the same as the BNP, and fewer than Solidarity.  The only polling firm that has offered RISE any limited cause for optimism is YouGov - but that may simply be because YouGov (as a volunteer online panel pollster) have a few core SSP supporters on their books.

*  *  *

SCOT GOES POP POLL OF POLLS

Today's update of the Holyrood Poll of Polls is arguably the most useful so far, because all of the polls that make up the sample were entirely conducted in either January or February.  And there's no distortion in the trend this time, because no firms have dropped out of the sample or returned to it.

Constituency ballot : 

SNP 52.4% (+0.6)
Labour 20.4% (n/c)
Conservatives 17.2% (-0.4)
Liberal Democrats 5.6% (-0.2)

Regional list ballot : 

SNP 46.6% (+0.6)
Labour 19.4% (n/c)
Conservatives 17.0% (-0.2)
Liberal Democrats 6.8% (n/c)
Greens 6.4% (-0.2)

(The Scottish Parliament 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 five - Panelbase, Survation, YouGov, TNS and Ipsos-Mori. Whenever a new poll is published, it replaces the last poll from the same company in the sample.)

And here is the independence Poll of Polls.  Bear in mind that this one uses a slightly different methodology, so some of the fieldwork taken into account is much older.

MEAN AVERAGE (excluding Don't Knows) :

Yes 48.7% (-0.5)
No 51.3% (+0.5)

MEAN AVERAGE (not excluding Don't Knows) :

Yes 44.8% (-0.7)
No 47.2% (+0.2)

MEDIAN AVERAGE (excluding Don't Knows) :

Yes 47.9% (n/c)
No 52.1% (n/c)

(The independence Poll of Polls is based on a rolling average of the most recent poll from each of the firms that have polled on independence since the referendum, and that adhere to British Polling Council rules. At present, there are six - YouGov, TNS, Survation, Panelbase, Ipsos-Mori and ICM. Whenever a new poll is published, it replaces the last poll from the same company in the sample.)