We did a thing
So it’s very nearly time to start on the poll results. We’ve made you wait long enough. But before the big reveals, we’d better get some quick background out of the way.
We got quotes from all the grown-up polling companies, most of which were in pretty much the same ballpark (with one dramatically expensive exception), but we chose Panelbase because they’ve got an excellent record for accuracy, they were the best at responding to our numerous stupid questions in a friendly and patient way, and if they’re good enough for the Sunday Times they’re good enough for us.
They sampled 1015 Scottish adults, weighted to accurately represent the Scottish population. The breakdown of respondents therefore came out like this:
GENDER
Male: 47%
Female: 53%
(A wee bonus for the No camp there, as women are the more indy-sceptic sex.)
AGE
18-24: 10%
25-34: 19%
35-44: 17%
45-54: 21%
55-64: 21%
65+: 13%
(GROUPED)
18-34: 29%
35-54: 38%
55+: 34%
(Again, if anything a population skewed a touch towards No, ie older.)
VOTES IN 2011 HOLYROOD ELECTION (constituency/region averaged, rounded)
SNP: 37%
Labour: 24%
Conservative: 11%
Lib Dem: 5%
Green: 4%
Other: 1%
Didn’t vote: 16%
Can’t remember: 4%
(4% of Scots can’t remember who they voted for two years ago? Wow. But those numbers are in pretty close proportion to the vote shares in 2011. Nice and fair.)
RESIDENCE
Owned/mortgaged: 66%
Social rented: 25%
Private rented: 6%
Other: 3%
(We’re very surprised at how low the private renting figure is. Is that typical?)
EMPLOYMENT
Full-time employed: 38%
Retired: 17%
Part-time employed: 14%
Houseperson: 8%
Student: 7%
Unemployed: 6%
Incapacitated: 5%
Self-employed: 5%
(Similarly the self-employed figure.)
SOCIO-ECONOMIC GRADE
A: 4%
B: 20%
C1: 22%
C2: 26%
D: 19%
E: 9%
(A pretty even distribution.)
Panelbase will be releasing all the full data tables for study soon. There’s so much information to take in we’re going to break it down into a few chunks to give us all some time to analyse and digest the info, and doubtless our pals at “Better Together” will be looking for something they can spin too. Good luck to them.
Looking forward to this 🙂
Just in time for lunch?
The No campaign’s only spin will be “ha, what a joke, space monsters!!” And as you know, that’s already begun. They probably won’t touch on the results or debate that can come of it, at all.
Looking forward to it. Not sure how the media release of this works though – do Panelbase offer the results to the media? Does someone draft a press release and send it out? Do you call the papers and annoy the hell outta them about it? 🙂
I’m quivering with antici…………pation.
I find it incredible that the percentage of those living in privately-rented accomodation is apparently lower than the percentage of students. In my experience pretty much all students are staying in privately rented accomodation. In fact most people I know rent privately. Maybe it’s an Aberdeen thing though.
“I find it incredible that the percentage of those living in privately-rented accomodation is apparently lower than the percentage of students. In my experience pretty much all students are staying in privately rented accomodation. In fact most people I know rent privately. Maybe it’s an Aberdeen thing though.”
This is what Panelbase said about the figures:
“On the issue of housing tenure, private rental is genuinely surprisingly low. We actually had 17% of our raw sample in that category, but downweighted them as that would be an overrepresentation – here are some government stats:
link to scotland.gov.uk“
This is the most exciting thing ever.
A wee bonus for the No camp there, as women are the more indy-sceptic sex.
Couple of % bias towards females, but not worth worrying about.
I was one of the respondents to the poll. Looking forward to the results.
Cant wait…but i will. Cue drum roll.
The age breakdown shows that there were no panellists under 18. Remembering that some 14 year-olds will be voters by Referendum Day, there is room for someone to complain that the poll is not representative.
I am sure that all the polling companies have this problem, of how to include young voters in their survey – but does anyone know if steps are being taken to address the missing electors?
there is room for someone to complain that the poll is not representative.
Not really, and some of those polled at the higher age end may not be around next year…
Both groups are very small, so much so that any influence would be within standard error for the sample size.
Rev
Should I go to the bookies for a wee punt before your release these and the price tumbles?
The MSM will probably just totally ignore the poll, as it was commissioned (and spun) by a separatist blog. They wouldn’t, after all, like to risk legitimising the hateful output from Rev’s laptop
I would be very interested in an area breakdown. I suspect there are widely varying views in different parts of Scotland a poll taking heavily from West Central Scotland may be very different from what comes out of other areas.
Anticipation mounts
Ever so slightly O/T
I see Blair McDougal has written a diatribe n the Huffington Post, which frankly is the usual
rhetoric, comments are however welcome and not prone to the Bitter Together censorship
link to huffingtonpost.co.uk
James Kay says:
The age breakdown shows that there were no panellists under 18. Remembering that some 14 year-olds will be voters by Referendum Day, there is room for someone to complain that the poll is not representative.
TNS include 16 and 17 year olds in their face to face interview polls, although the group is so small it makes no difference to the overall figures.
I was suprised about the residence thing and done a quick search.
“The owner-occupation rate in the United Kingdom is on a par with the United States and Australia at around 70 per cent.”
link to uk.ask.com
““The owner-occupation rate in the United Kingdom is on a par with the United States and Australia at around 70 per cent.””
I wasn’t surprised about that, have covered it before. What surprised me was how much of the rental sector was social rather than private.
I just peed myself..
Quite heartened at this reply in the”huffingtonpost”though.
“I’m neither pro- or anti- separation. I haven’t bothered to delve into the detail so can’t comment on the viability of Salmond’s politics. But d’you know what? He ain’t no oil painting but he does come across as sincere and with the courage of his convictions. I can take him a lot more seriously than I could the capricious Farage wanting to take over England. Salmond is, to me, a good (probably honest) politician who has the makings of a good statesman.”
Refreshing to say the least from”brightonsbadgirl”.
Just wondering if the private rental is so low because a lot of the private lets are paid for with housing benefit claimants. Could there be a mix up with people saying they are in social housing, even though their benefit is paid to a private landlord?
A quick issue to raise about the reliance of such polls – unless they are demographically and geographically spread even, they are less reliable than you think. For example, if you sample folk in Lenzie in east dunbartonshire, which is an affluent area, you really need to balance thar by sampling, say, easterhouse in Glasgow, a poor area, just 5 miles up the road. the responses you get from these areas will be different not on the grounds of typically recorded demography (age, gender) but on the basis of geographic demography (the socio-economic markers of the area in which a person lives). The reasons for that are more complicated than rich and poor, but are really important when planning campaign tactics based on the results Of any poll and when predicting voting patterns. You’ll need to look at that sort of info for application of the results of your poll on the local campaign groups affected.
“but on the basis of geographic demography (the socio-economic markers of the area in which a person lives)”
As we’ve already noted in the demographics post, socio-economic balance/weighting was applied. This is a proper grown-up polling company we’re dealing with, fella.
First thing to note in my opinion is that this isn’t a representative sample of the population as a whole if the answers are truthful. The number saying they voted is 84% (even if they can’t mind who they voted for!). Turnout was just over 50% in that election.
That doesn’t deligitimise the results of course, but you’re probably getting an oversampling of those who regularly vote because those same people are more likely to also answer pollsters.
In terms of the political affiliations of those who voted (I’m going by the regional vote only as the Greens didn’t stand in the constituency seats) it’s a pretty decent sampling. All 5 parties who got regional seats are slightly oversampled and it looks like people may have forgot they voted for the likes of the Scottish Christians or Respect (the minor parties who make up the “others” here got 7% of the entire vote) and said they voted for another party. Every major party got less in the regional vote than in the constituency one and Labour done worst out of that so could just be that people mind voting for a Labour candidate, although they didn’t vote for them on the regional seat.
In other words this is a very good sample that is generally representative with the caveat that those who may be less likely to vote haven’t been reached in their actual share of the population. That will be a problem with pretty much any polling imo, but it is a wee bit of an issue.
@wee 162;
Good point. Those that said they voted ‘y’ may otherwise be people who if they had remembered to vote, that is what they would have voted for. + – 3% is allowable, the data so far (panel base have yet to forward to WoS) may explain it better. I’m currently skeptical, due to reading too many polls in anticipation and seeing rather strange, contradictory percentiles.
On the additional plus side its keeping you too busy to have a boot in at current Rangers debacle ;o) with any luck Labour will implode the same way as we get closer to the vote and polls start to show that they will be in trouble……The rush of Labourites to the exits to embrace independence would be a joy to see, if somewhat sickening.
24% labour voters when Labour polled 32%? Also Panelbase are not really the best way of doing it as it is a paid group, people are paid to take part so it can be influenced by people saying certain things about themselves to maximise their participation and get the cash.
“24% labour voters when Labour polled 32%?”
Come on, Jim, sharpen up. The demographic includes non-voters, so EVERY party’s number is lower than the share it got at Holyrood, by roughly the same proportion, because Holyrood vote share obviously doesn’t include non-voters.
Very much looking forward to seeing the results. 🙂
The actual average percentage of those who voted SNP in 2011 (average of R and L to nearest percent) was 45%. So with only 37% figure of the panel saying they voted SNP, that must constitute another significant bias in favour of the No side.