Today's front page headline in the Sun covers almost the entire page. Accompanied by a chilling photo of a knife-wielding "Jihadi John" in black balaclava, it proclaims "1 IN 5 BRIT MUSLIMS' SYMPATHY FOR JIHADIS".
It is a lie. Even worse, it is a shameful distortion of its own polling data, consciously designed to fuel terror and distrust of Muslims.
The front page juxtaposition of "Brit Muslims", "Jihadis" and frightening photo of an Isis murderer is no accident. It reflects the warmongering editorial tone of the newspaper, which is apparently intent on exploiting any means at its disposal to persuade its readers that British muslims are - at best - ambivalent about terrorism. Its dishonest use of opinion polling is the Sun's latest weapon of choice in this undeclared propaganda war.
The single question on which the Sun's misrepresentation is based can be found here. Its polling company, Survation, asked a sample of just over 1,000 Muslims "Which of the following statements is closest to your view" with four possible responses:
•I have a lot of sympathy with young Muslims who leave the UK to join fighters in Syria
•I have some sympathy with young Muslims who leave the UK to join fighters in Syria
•I have no sympathy with young Muslims who leave the UK to join fighters in Syria
•Don't know.
It's "1 in 5" figure is derived from the combined totals of the first two responses, which come to 19.8%.
Now imagine a poll in 1937, in the midst of the Spanish civil war, which asked British voters whether they had sympathy for those who went to join fighters in Spain. How would you interpret a figure of 20% who said yes? Were they sympathising with the pro-Franco nationalist forces, or the anti-Franco republican forces? We would not have known.
Exactly the same applies now. There are various factions fighting in Syria, including forces dedicated to resisting Isis. They are not "jihadis", and yet the Sun quite deliberately uses its front page to lump together responses which are entirely opposed to each other. The truth is, we have absolutely no idea how many of those one in five have "sympathy for jihadis".
I spoke to the pollster who was responsible for the survey, Patrick Brione of Survation, whose own thoughtful and balanced commentary on his own survey - in stark contrast to the Sun - can be found here. He agreed that it was "certainly possible" that respondents could have been referring to fighting on different sides of the Syrian conflict. In fact, he said, the wording was chosen in order to replicate a question first asked in March this year (in rather less tense circumstances).
As it happens, Survation's March poll asked identical question of both muslim AND non-muslim samples. The resulting data, says Brione, suggest that "attitudes held by the Muslim and non-Muslim populations are not that different." I wonder why that fascinating comparison was entirely missing from the Sun's "analysis"?
It gets worse. On its front page, the Sun's second paragraph declared ominously that the proportion of jihadi sympathisers amongst young muslims aged 18-34 was "even higher at one in four". But when Survation ran the same questions past muslims and non-muslims in March, figures for this age group were almost identical: 30% of young non-muslims were sympathetic towards those going to fight in Syria compared to 32% of young muslims. Do we really think that nearly one in three British young people are jihadi sympathisers?
For obvious reasons, Brione was non-committal about the Sun's use of these data: Survation may advise clients as to how figures might be interpreted but "whatever journalistic interpretation they put on these figures is up to them".
Polling companies mostly subscribe to the British Polling Council whose rules have little to say about how clients publicly interpret their figures, as long as all the questions and tables are openly available. Such transparency is commendable but is frankly of little use compared to a screaming and manifestly false front page tabloid headline.
Some are also members of the Market Research Society whose Code of Conduct rule 56 obliges members to "take reasonable steps to check and where necessary amend any client-prepared materials prior to publication to ensure that the published results will not be incorrectly or misleadingly reported". I would be surprised if that was routinely done for newspaper articles.
It might now be necessary, in the interests of accuracy as well as social harmony, for all reputable pollsters to insist on precisely that kind of copy clearance. Today's Sun headline demonstrates that some journalists and editors have granted themselves licence to spread poisoned untruths under the cloak of "objective" opinion research. Pollsters should not allow themselves to be willing accomplices in this propaganda fraud.
Steven Barnett is Professor of Communications at the University of Westminster.