Not surprisingly, this came as news to many British adults who couldn’t recall ever being asked. And, not surprisingly Twitter was rather sceptical of the veracity of this claim.
Twitter user Alan Travis showed that the figure of 54% was at odds with the data from ComRes
which appears to show very clearly that in answer to what was a quite leading question the response
was, in fact, only 44%.
Another user Sarah Mackie figured out that the 54% figure was achieved by excluding the don’t
know’s. Of course, it is common practice in survey work to exclude don’t knows to present a picture of those who hold an opinion, rather than those unwilling to express one. However, it is unusual for a
respectable polling company to allow figures to be manipulated in order to turn a minority into a
majority.
Indeed, this misuse of the data has led Mark Pack, an ex-election strategist with the Lib Dem’s to report the Telegraph to press regulator IPSO. He cites the Editors code which states:”The Press
must take care not to publish inaccurate, misleading or distorted information “
Many people on the left are understandably sceptical about opinion polls, though often on the
grounds that they were not personally asked. But, this is to entirely misunderstand the nature of a survey. Statistics tell us that a sample of around 2000 can be as accurate as a census of the entire
population. Most opinion polls are based on samples of 2000 people, so does this mean they can be trusted?
That figure of 2000 is based on a random sample of the population being studied. A random sample should be representative of the population and the results of a survey are then +/- 3% of the entire population.
The problem for polling companies is that drawing properly representative samples is expensive and time consuming. Margaret Atwood’s novel ‘The Edible Woman’ is based around a pollling company interviewer whose job was to knock on doors to fill a quota of specific segments of the population.
As far as I know no mainstream polling company now employs an army of skilled interviewers
preferring instead to invite people to join a panel.
Polling companies no longer employ professional interviewers and rely on online surveys instead |
overcome the shortcomings pollsters use a variety of techniques known as weighting. Each company
has its own weighting methodology, but for all of them they are attempting to overcome the
limitations of their own panel in order to appear representative.
The 54% figure is grossly misleading leading to the Express quoting it as meaning:
End of line for MPs: Public BACKS bid to SCRAP PARLIAMENT so Boris can get on with Brexit
They lead the story by repeating the blatantly false claim:
BORIS JOHNSON has received support from more than half of the UK public to deliver Brexit by any means possible, including suspending Parliament.
newspapers that support Brexit, the Conservative Party and, of course, Johnson, who once made up stories about the EU for the Telegraph. So, the fact that they have lied and manipulated figures to suit
themselves is no great surprise.
ComRes, on the other hand, is a respected polling company. It is also a member of the British Polling Council and is expected to abide by their rules. Their objectives, include the following:
- To advance the understanding, among politicians, the media and general public, of how polls are conducted and how to interpret poll results.
Ben Page of rival polling company Ipsos-MORI also took the Telegraph to task in this tweet.
Strangely enough neither Ben Page nor Full Fact choose to highlight the intrinsic problem of online polling using self-selected panels which has become commonplace in opinion polling.
The conclusions for those of us on the left are rather obvious: