Saturday, March 8, 2014

Sagacity on Saturday : Dragging the Westminster parliament kicking and screaming into the 19th Century

"The Honourable Gentleman keeps referring to 19th Century practices in this place. Does he not think that that is a bit unfair? I cannot remember anything quite as modern as the 19th Century here."

SNP leader Alex Salmond, speaking with a slight touch of sarcasm in a House of Commons debate in 1998.  To be fair, since then there have been major modernising reforms, such, the Speaker no longer wears court robes.  At this rate of progress, I confidently predict that hereditary peers will have been abolished by the 53rd Century.

Friday, March 7, 2014

Let's play rounders!

I've just noticed something rather peculiar.  Of the seven BPC pollsters that have published referendum polls so far, there are five that provide actual numbers (as opposed to just percentages) in their datasets.  All five are currently showing a slightly lower No lead on their actual numbers than in their published figures.  There's nothing sinister about that - it's just an unlucky coincidence caused by the effect of rounding.  But it does mean that the Poll of Polls has been overstating the No lead by a smidgeon.


Published figures :

Yes 37%
No 49%

NO LEAD - 12%

Actual figures before rounding :

Yes 37.4%
No 48.7%

NO LEAD - 11.3%

*  *  *

Panelbase -

Published figures :

Yes 37%
No 47%

NO LEAD - 10%

Actual figures before rounding :

Yes 37.0%
No 46.7%

NO LEAD - 9.7%

*  *  *

Survation -

Published figures :

Yes 38%
No 47%

NO LEAD - 9%

Actual figures before rounding :

Yes 37.7%
No 46.6%

NO LEAD - 8.9%

*  *  *

Ipsos-Mori -

Published figures :

Yes 32%
No 57%

NO LEAD - 25%

Actual figures before rounding :

Yes 32.1%
No 56.6%

NO LEAD - 24.5%

*  *  *


Published figures :

Yes 29%
No 42%

NO LEAD - 13%

Actual figures before rounding :

Yes 29.1%
No 41.8%

NO LEAD - 12.7%

*  *  *

Scot Goes Pop Poll of Polls - 

Average based on published figures :

Yes 34.6%
No 48.9%

NO LEAD - 14.3%

Average based on actual unrounded numbers (where available) :

Yes 34.6%
No 48.6%

NO LEAD - 14.0%

You might be wondering why I don't simply use the actual unrounded numbers for the Poll of Polls in the first place.  It's simply a question of practicality - the headline percentages are often published several days before the datasets turn up.  I had assumed that it wouldn't make any difference, because the rounding effects in different polls would always cancel each other out, but clearly that isn't necessarily always going to be the case.

*  *  *

You may already have seen this, but Women for Independence are running an Indiegogo fundraiser with a target of £20,140.  I know it's difficult to decide which of the many pro-independence causes are most worth helping, but the way I'm looking at it is that Yes have three key challenges that need to be met - 1) reducing the gender gap in referendum voting intentions (which is where the Women for Independence fundraiser comes in), 2) nurturing an alternative media to at least partly balance out the severe anti-independence bias of the traditional media (which is where the Wings over Scotland fundraiser comes in), and 3) convincing traditional Labour voters that a Yes vote is a vote in favour of their own values (which is where the Labour for Independence fundraiser comes in).

Wednesday, March 5, 2014

Ipsos-Mori poll : SNP open up big Scottish Parliament lead

So now we turn to the other side of the coin when it comes to the Ipsos-Mori enigma - they may be the most extreme No-friendly outlier in respect of referendum voting intentions, but they've nevertheless reverted to being very much part of the mainstream in showing a solid SNP lead over Labour.  These figures will ease any concerns that had been building up (fuelled in particular by the Cowdenbeath result) that the tide was gradually starting to turn in Labour's favour.

Scottish Parliament constituency vote :

SNP 38% (+2)
Labour 29% (-5)
Conservatives 17% (+2)
Liberal Democrats 9% (+2)

(Note : Ipsos-Mori don't ask for voting intention on the regional list ballot.)

The only slight fly in the ointment is that Johann Lamont continues to have a modestly better net personal satisfaction rating than Alex Salmond, and it was certainly the case in 2011 that leadership ratings were a better early predictor of the election result than headline voting intention figures.  However, Lamont's lead is not all that it appears - in absolute terms more people are satisfied with Salmond, with Lamont benefiting from the fact that fewer people have an opinion about her one way or the other.  I know that a number of us have a reasonable amount of confidence about which way things will go once voters become better acquainted with the charms of the Scottish Labour "leader".

At this stage, of course, we're most interested in what the Holyrood figures say about the referendum race.  There is more than one possible interpretation - you could argue that if a pollster is showing a good position for the SNP, there's no reason to doubt its No-friendly referendum findings.  But the counterargument is that Ipsos-Mori have essentially just corroborated the Holyrood figures produced by Panelbase and Survation, meaning that there is also no longer any particular reason to fear that those pollsters have overstated the strength of Yes due to having a "Nat-heavy" sample.  So you pays your money and you takes your choice.

The other intriguing question is this : how can pollsters that show such similar Holyrood figures be coming up with wildly diverging pictures of the state of play in the referendum?  It seems to me that the explanation must be one or both of the following -

1) As a telephone pollster, Ipsos-Mori are falling victim to 'shy Yes' syndrome, something which the internet pollsters Panelbase and Survation (and indeed ICM) are immune to.

2) Although Ipsos-Mori are broadly getting the balance between the parties right, there may still be a fundamental difference in the types of SNP and Labour voter that they are interviewing, which would explain why Scottish national identity is suspiciously weak in their figures.  In particular, other pollsters may be better at reaching the type of patriotic Labour voter who is moving to Yes.

*  *  *

I have a confession to make - I actually quite like the UK's Eurovision entry this year.  Yes, it's a do-it-by-numbers effort - although it would be unfair to call it derivative of last year's Danish winner, it's not hard to see how someone could use Only Teardrops as a very rough template and end up with Children of the Universe.  But it's a strong song all the same, with a good live performer, and if there's any justice it should easily secure the UK's best showing since Jade Ewen finished fifth with a Lloyd-Webber-penned ballad in 2009.  It might even have an outside chance of winning if the staging is phenomenally good (on past form I doubt that it will be).

I still don't feel that the song represents me, though.  The old public vote system was a kind of ritual that gave us all a stake in the entry, whereas nowadays the whole thing just feels like a private enterprise by the BBC.  This year I've voted in the French and Irish national selections, and for a second year in a row I voted for the winner in Ireland (Heartbeat by Can-Linn).  I haven't had a chance to vote for or against the UK song, and for obvious reasons I don't particularly identify with the UK as a country anyway, so guess which entry I'll feel that I have a stake in?

Incidentally, as I traditionally point out at this time of year, Molly Smitten-Downes' selection means that there has still been no Scottish involvement in any UK Eurovision entry (either as a performer or as a writer) since Scott Fitzgerald performed in Dublin way back in 1988.  Bizarrely, France and Cyprus have both been represented by Scots more recently than the UK have.  That's the typically copper-bottomed type of 'union dividend' that is being brought to the table in referendum year.

In case you've forgotten it, here's a clip of Capercaillie's Karen Matheson singing for France (yes, that's France) at Eurovision in 1996 alongside Welsh singer Elaine Morgan, with the Breton language song Diwanit Bugale.

Tuesday, March 4, 2014

A bit more on polling methodology

I thought you might be interested in this comment left by a user on Reddit. (I stumbled across it because someone was kind enough to link to Scot Goes Pop on the same thread!)

"Met someone last week who until the end of last year worked for Mori. Asked him about the polling situation and he said there's a general understanding among pollsters that most of the weightings used have been badly chosen or chosen specifically in some cases, or not at all in others where there should have been some form of weighting. On the reason why, well he was a little bit coy about this so no insight gleaned here.

I wasn't personally aware of such a thing as poll weighting and for example that some polls last year showing a large No lead were weighted towards Westminster voting intentions, even though the focus of the poll was clearly the independence referendum.

I asked about the likely accuracy of SNP door to door canvassing data collected among the local by-elections and in general. He said it was understood to be pretty good, I asked if 40-50% support for independence was reasonable in this case, he said yep. I asked how much the polling from likes of his recent employers and other large pollsters in the UK could be out based on his observations about the weighting. 10-20% he said."

Of course we have to be cautious about this, because the catastrophic error of weighting by recalled Westminster vote has now been largely resolved (touch wood), so that should no longer be responsible for distorting the numbers. That said, YouGov are clearly still in denial about the sheer scale of the methodological change that they introduced back in September. A day or two ago, Peter Kellner wrote the same bloody article he's been periodically writing for the last three years about how Scottish public opinion on independence is supposedly set in stone - well, how the hell does he square that claim with the massive 10% drop in the No lead that his own firm reported in September? Presumably we're meant to infer that the shift in opinion was not real, but an illusion caused by the methodological change - but has anyone actually heard Kellner say that directly? I certainly haven't, and no wonder - it would be tantamount to an acknowledgement that YouGov had previously been pumping out grossly distorted figures for months, if not years.

To return to the original point, though, it's interesting to hear some anecdotal evidence that pollsters themselves are privately anxious that they may have been understating the Yes vote, and perhaps still are. As far as canvass returns are concerned, they do come attached to a health warning, because there will always be some people who tell the canvasser what they think he/she wants to hear. But experienced people in the Yes campaign will be aware of that problem, and will be able to make a common sense adjustment to the numbers to take account of it. If the pattern still bears no resemblance to the polls even after that, it could be an indication that at least some of the pollsters are indeed getting their methodology wrong. But I can only speculate on what the canvass figures are actually showing - the only proper insight we've had was the well-publicised Cowdenbeath example.

I also had this little exchange with Duncan Hotdogstall (no relation) on the previous thread about the new Ipsos-Mori poll -

Duncan Hotdogstall : Hi James,

A glance through the data tables is setting off some alarm bells. I'm starting to wonder if I am looking at the wrong tables? Can you advise?

On page 2 the table says "Party support Scottish Parliament Holyrood First Vote Constituency" Does this mean it is weighted to Holyrood? If so then the weighted SNP vote appears to add up to 36% (302/845). We all know the actual, real life vote was 45.4%.

The sample also seems to be less than representative of the general population when it comes to national identity. The chart of page 4 of the tables says that a total of 48% of the sample feel either Scottish only or more Scottish than British. Yet the most recent census data tells us that 62% of the population feel Scottish only. Conversely, 48% of the sample feel more British or British only, whereas the census data gives that figure as around 34%.

Finally, the sample and census data depart again when it comes to place of birth: On page 5 of the chart it says that 78% of the sample were born in Scotland. The census says the figure is 83%.

Therefore, it appears that the sample in this poll is far more "British" than you would find in the usual population and way less representative in terms of SNP voters.

What is your analysis, James? Scottish Skier seemed to think that the result, when adjusted, should be more like Y 37, N 43 and DK 20, but the explanation wasn't very detailed.

Me : The Holyrood figures are current voting intention, not recalled 2011 vote.

It's not really possible to make a direct comparison with the national identity figures in the census, because the census asked the question in a different way. But even comparing to the most recent SSAS, Ipsos-Mori's Scottish figures are a bit too low.

As Oldnat pointed out the other week, the problem with making a comparison with the census figures on place of birth is that the census included children. I've tried to find the figures for 16+ only, but without any luck so far.

I think what Scottish Skier does is take an average of the SSAS national identity figures over the last fifteen years, and adjust according to that. I'm not sure that's wise, because there's fairly clear evidence that Scottish identity has fallen back on the SSAS in recent years. So he's probably making too radical an adjustment.

Duncan Hotdogstall : Thanks James. Although I still can't believe that poll, no one in the real world is experiencing what it claims. I personally know 5 people who have moved to yes in the last month (including a tory, oh yes, he made a right song and dance about it on facebook) but have yet to meet a single person in the world ever who has moved from yes to no.

Sorry for the double post earlier, don't know what happened.

Me : OK, I've done a rough calculation to reweight the Ipsos-Mori figures in line with last year's SSAS national identity data, and it comes out as Yes 32%, No 51%. That isn't directly comparable to the published headline figures because it isn't filtered by certainty to vote, but essentially it has reduced the No lead by about 7%. Not quite as dramatic as what Scottish Skier came up with, but certainly an indication that Ipsos-Mori should be taken with a pinch of salt.

Ipsos-Mori consolidate their status as the extreme outliers

One thing that we can certainly deduce from yesterday's Ipsos-Mori poll is that the Yes campaign are highly unlikely to go into polling day in September in the lead with that particular pollster. But the million dollar question is whether they actually need to do so. The gap between the No lead being shown by Ipsos-Mori and the average No lead being shown by the four online pollsters that have reported so far this year is a stonking 12.25%. In other words, even if the online pollsters end up pointing to a narrow Yes victory by September, the chances are that Ipsos-Mori will still be showing a double-digit No lead. They really have turned into the extreme outliers of this campaign.

That's not to say they are necessarily getting it wrong, of course. They are unique in being the only referendum pollster to be carrying out their fieldwork by telephone, and because of that factor it's hard to shake off the nagging doubt that they may be getting a more representative sample than the online pollsters (at the very least, there's no danger that they're interviewing the same people over and over again as YouGov and Panelbase do). Luckily, however, we do have one other non-online pollster to make a comparison with. TNS-BMRB conduct their fieldwork face-to-face, and in recent polls have been showing a markedly more favourable position for Yes than Ipsos-Mori. So the non-online status of Ipsos-Mori can't in itself explain why they produce such different results from everyone else. A more likely explanation is their ongoing refusal to weight their results by past Holyrood vote, which is now standard practice for most other BPC pollsters. It's also known that they use a 'stealth preamble' when asking the referendum question, but we can only guess how leading or biased that may be.

And whatever the virtues of telephone polling (or indeed face-to-face polling), we shouldn't forget that there's one very good reason for thinking that online polls may be producing more accurate results, namely that 'shy Yes' voters probably find it harder to admit their preference to a live interviewer than to a computer screen.

All the same, this huge degree of uncertainty as to the true state of play is incredibly frustrating and unsettling. If ICM (the UK's "gold standard" pollster) and their online polling cousins of Panelbase and Survation are correct, the Yes campaign are within relatively easy striking distance of victory and an independent Scotland in 2016 is a real prospect. But if Ipsos-Mori are correct, the challenge faced by Yes is bordering on the insurmountable. I know most of us firmly believe that the former scenario is far more likely, but it's scary to think that we simply won't know for sure until the real votes start coming in.

As ever, when pollsters are so far apart the only way of making sense of any individual poll is to look at the trend, and Ipsos-Mori have at least confirmed that the No lead remains three points lower than in the last two polls prior to the publication of the White Paper.

* * *


Since the No vote is unchanged in the Ipsos-Mori poll, it's only the Yes vote that is affected slightly in the updated headline figures of this blog's Poll of Polls.

MEAN AVERAGE (not excluding Don't Knows) :

Yes 34.6% (-0.3)
No 48.9% (n/c)

MEAN AVERAGE (excluding Don't Knows) :

Yes 41.4% (-0.2)
No 58.6% (+0.2)

MEDIAN AVERAGE (excluding Don't Knows) :

Yes 42.0% (n/c)
No 58.0% (n/c)

(The Poll of Polls is based on a rolling average of the most recent poll from each of the pollsters that have been active in the referendum campaign, and that adhere to British Polling Council rules. At present, there are seven - Angus Reid, YouGov, TNS-BMRB, Survation, Panelbase, Ipsos-Mori and ICM. Whenever a new poll is published, it replaces the last poll from the same company in the sample. Changes in the Poll of Polls are generally glacial in nature due to the fact that only a small portion of the sample is updated each time.

For clarity, the Poll of Polls takes no account of polls conducted by bridalwear companies.)

The fact that the median average has become more favourable to Yes than the mean average is testament to how the Yes-friendly end of the polling spectrum is now more heavily populated than the No-friendly end, a complete reversal of the position that held good until recently.

Sunday, March 2, 2014

YouGov poll shows support for independence rising to its highest level of the campaign so far

Apologies for being a bit later than usual with this poll - I've been having the week from hell.  Among many, many other calamities, I - a) fell awkwardly and twisted my knee, and because of circumstances haven't been able to rest it as much as I should, b) have had my faith in one particular type of public servant rocked to the core, and c) have just five hours ago lost my favourite woolly hat, which I've been inseparable from for months!  Ah well, in the words of the theme song to a 1980s sitcom, life goes on, right or wrong...

So down to business. There was a time, not so long ago, when this kind of late night tweet from Blair McDougall...

"New #indyref poll tomorrow. More first thing."

...would have been a portent of doom. I did go to bed on Friday night wondering if I would awake to a poll showing a drop in the pro-independence vote, but I needn't have worried - instead it was the second YouGov poll in succession to show the Yes vote rising to its highest level of the campaign so far. It seems that 'bad is the new good' for McDougall. If this is the type of poll he likes to crow about these days, he must be getting a tad jittery.

Should Scotland be an independent country?*

Yes 35% (+1)
No 53% (+1)

*YouGov always add a preamble to the actual referendum question - hopefully when the datasets are published it will turn out to be the more neutrally-worded preamble that was first used in September.

Presumably the reason that the No campaign felt they could at least have a stab at spinning this as a good news poll for themselves is the fact that they've increased their own support by 1%, just as Yes have done. The snag is, though, that if the headline No lead remains the same but the number of undecided voters fall, that means by definition that the true lead is actually declining. Here is the position when undecideds are stripped out -

Yes 40% (+1)
No 60% (-1)

So Yes have broken through the psychological 40% barrier for the first time with a pollster that is traditionally extremely unfavourable to them. And that's the key point that must always be borne in mind when assessing YouGov's headline numbers - Yes don't necessarily need to be in the lead by the end of the campaign. It's quite conceivable that Yes could still be four or five (or even more) points behind with YouGov on the eve of polling day, and yet be level pegging or ahead on an average of all the pollsters. As David Halliday pointed out a few hours ago, YouGov's final poll of the 2011 Holyrood campaign had the SNP ahead by just 3% on the list ballot. The actual election result was an SNP lead of 18%.

When the datasets are published, the first thing I'll be looking out for is the gap in voting intentions between higher and lower income respondents - it's been suspiciously low in recent YouGov polls, leading me to wonder if the firm has a serious problem in its sampling process.

Probably the biggest significance of this poll is that it's the final piece in the jigsaw that proves the coordinated announcements on the currency from the three London parties failed to have the desired impact, and if anything probably led to an increase in the Yes vote. It also demonstrates beyond a scintilla of doubt that the Press & Journal regional "poll" that the No campaign and half the mainstream media embarrassed themselves by getting so excited about a few days ago was not worth the paper it was printed on.

* * *


The pro-independence campaign's progress with YouGov means that they also rise to 34.9% support in this blog's Poll of Polls - the highest Yes figure recorded to date.

MEAN AVERAGE (not excluding Don't Knows) :

Yes 34.9% (+0.2)
No 48.9% (+0.2)

MEAN AVERAGE (excluding Don't Knows) :

Yes 41.6% (n/c)
No 58.4% (n/c)

MEDIAN AVERAGE (excluding Don't Knows) :

Yes 42.0% (n/c)
No 58.0% (n/c)

(The Poll of Polls is based on a rolling average of the most recent poll from each of the pollsters that have been active in the referendum campaign, and that adhere to British Polling Council rules. At present, there are seven - Angus Reid, YouGov, TNS-BMRB, Survation, Panelbase, Ipsos-Mori and ICM. Whenever a new poll is published, it replaces the last poll from the same company in the sample. Changes in the Poll of Polls are generally glacial in nature due to the fact that only a small portion of the sample is updated each time.

For clarity, the Poll of Polls takes no account of polls conducted by bridalwear companies.)

It's only rounding issues that prevent Yes creeping up slightly on the mean average with Don't Knows excluded. The median average is unchanged because YouGov remain firmly on the No-friendly end of the spectrum, with Angus Reid continuing to provide the mid-point.

* * *

UPDATE : I've just posted this as a comment below, but on reflection I thought it was worth adding to the main post -

On a related topic, I've been having an exchange with a unionist Twitter troll who has a bee in his bonnet about the possibility that Yes may only have appeared to break the 40% barrier with YouGov due to their vote being rounded up. Well, we'll never know, because YouGov only publish percentages in their datasets, not raw numbers. However, ICM do provide raw numbers, and having had a look at them I'm delighted to report that their last poll was a touch better for Yes than we originally thought -

Yes 37.4%
No 48.7%

So the true lead was only 11.3%, which is hardly a million miles away from the 7% lead ICM reported in their sensational January poll. With Don't Knows excluded, it's -

Yes 43.4%
No 56.6%

So rounding certainly worked against the Yes side in the way that poll was reported. Hopefully those numbers bode well for future ICM polls, though.