There’s been a fair bit of crowing from “Better Together” about some recent poll results. Which is fair enough – almost everybody likes to shout when they get some good news (though this site has consistently urged caution over polling findings months before a vote, whether favourable or not).
It is, however, always wise to look at the small print.
The Sunday Mail has a surprisingly low-key piece today about a new opinion poll commissioned by the paper through the little-known pollster Progressive Partnership (who aren’t a member of the British Polling Council) and conducted by YouGov.
Oddly, it seems to have produced two very different sets of results.
A phenomenon we’ve reported on numerous times on this site is the strange way that the media will regard the same opinion-poll statistics in radically different ways depending on how the figures relate to their political agenda.
So if 65% of Scots say they think Alex Salmond is a swell and trustworthy guy, the headline will be “MORE THAN A THIRD OF SCOTS DON’T TRUST SLIPPERY SALMOND”. Conversely, if those numbers are reversed on a referendum poll, the banner lead will be “ONLY A THIRD OF SCOTS BACK SEPARATION”.
But there are other ways of misrepresenting numbers, too.
Our attention was drawn this weekend to a survey conducted by the Scottish Chambers of Commerce, which polled 759 Scottish businesses of various sizes about a number of issues relating to independence.
It doesn’t seem to have had a great deal of coverage, perhaps because most of the answers were in the “bleeding obvious” category – business frets about change, and the more change there might be the less they like it.
One set of figures did catch our eye, though.
As people who commission opinion polls occasionally, a thing that puzzles us is why other people who do it ask questions and then don’t talk about the results.
Some polls are done with the intention of being for private consumption only (this is particularly true when they’re commissioned by one side or the other in a debate, rather than by a notionally-impartial newspaper or the pollster themselves), and at other times results will be kept private because the results are unfavourable to the people who commissioned them.
(For the avoidance of doubt, we’ve never withheld any results for that reason.)
But at other times, results will be published but never discussed. Which is why, whenever a poll’s just come out these days, we get ourselves straight over to the polling company’s website and see what’s been left out.
Earlier this week we did a little poking and prodding of the Scotsman’s last ICM poll, and now the full data tables are in for the latest one, so to while away an hour before tea we figured we may as well do a bit of comparing and see how things had changed.
Wait! Come back! There’ll be Miley Cyrus at the end!
We’re still waiting for the full data tables for the ICM/Scotland on Sunday poll that got everyone a little excited at the weekend, and whose findings closely mirrored the Panelbase/Wings Over Scotland one two weeks ago that the same publisher crudely smeared and cast aspersions (which it later retracted) on the credibility of.
In the meantime, even though we’re still technically on holiday, we had a bit of a rummage through the company’s preceding one for the Scotsman papers this morning and picked out some random interesting snippets. We’ll be watching keenly to see if the latest poll has corresponding stats to compare.
Three opinion polls this week have all suggested that Labour’s opinion-poll lead over the Conservatives is continuing to shrink. ICM put Ed Miliband’s party just three points in front, as do Ipsos MORI, while Populus have a mere 1% between the two parties.
For perspective, the same distance out from the 2010 general election, the Tories were 16 points in front. By seven months away from the vote, in October 2009, their lead was an incredible NINETEEN points, and they still couldn’t win a majority.
Who fancies Labour’s chances?
We’ve talked a bit today about a fascinating poll by Survation for the Daily Mail last month, whose findings got very little coverage in the media, perhaps because they revealed the rather off-message fact of how ineffectual two of the No campaign’s three great pillars of propaganda have been proving.
But there was another interesting snippet in the results, which could confound the flurry of recent polls all showing the SNP with a commanding lead for the 2016 Holyrood election, regardless of whether it’s for an independent Scottish Parliament or a devolved one.
Last month there was a mild flurry of activity in the press about the so-called “missing million” – Scots entitled to vote, but who choose for one reason or another not to. Catchy as it is, the phrase seems a significant understatement. Around four million people in Scotland meet voting criteria, but fewer than half of those turned out for the 2011 Holyrood election, and under 2.5m at the 2010 Westminster one.
Obviously that’s a bad thing in principle in its own right. But it could also be seriously distorting polling for the independence referendum, because – perhaps for the only time in their lives – an awful lot of those missing millions ARE going to go out this September and put a cross in a box. And nobody knows which one.
As commissioners of opinion polls ourselves, we know all too well that one of the more unwelcome aspects of the pursuit is that sometimes you don’t like the answers you get – we were noticeably dismayed, for example, by one or two of the things revealed about social attitudes in our last poll.
Today’s media extensively covers a Survation poll for the Daily Record which finds the highest Yes vote in some time (despite an “if the referendum was today” preamble, which generally works against Yes), equating to 45% Yes 55% No, with almost a quarter of Labour voters now planning to vote for independence.
The full tables aren’t yet available for analysis as we write this, so to while away the time we decided to have a proper delve in the last one from the same company, run on behalf of the Daily Mail last month. And a couple of things leapt right out at us.
We’ve commented before on the odd way that newspapers can reveal their bias in the way they phrase their reporting, rather than in the actual content of it, which can be entirely factually accurate. As we noted, a particular giveaway is the angle from which they view statistics, and especially opinion polls.
A poll showing 35% of people backing independence will almost always be reported as “ONLY a third back Yes”, whereas one with the exact same numbers for a different question might be presented as “OVER a third distrust Alex Salmond”. The proportion “one third” is in such a manner portrayed as being both a small and a large one, to suit whatever position the publication wishes to promote.
It’s in such a context that we invite readers to ponder today’s Mail On Sunday.
Wings Over Scotland is a thing that exists.
