At the weekend the Sunday Times reported a new ICM poll of British Muslims, conducted for a Channel Four documentary later this week. I’ve written about polls of ethnic and religious minorities in Britain here before – they tend be both controversial and extremely difficult to do. On top of that the topics that tend to get asked carry with them risks of stoking racial tension, so it is crucial that they are done in the most responsible and robust way possible.

So how do you conduct a poll of British Muslims? With difficulty – there is no ideal way, no route that does not include compromises and result in skews and biases. British Muslims are around about 5% of the population, distributed unevenly around the country. Some groups within the community will have come here only recently, and perhaps have poor English. That’s difficult to poll in an affordable way – let’s go through the possibilities. The first would be to go to a very large database of people of known demographics (such as the existing panels of YouGov or another internet polling company) and recruit Muslims from there – this is the easiest route, the company will already know the religion of their panellists and at 5% incidence you’ll probably be able to get enough. The problem is that the British Muslims who join an internet panel are probably skewed towards the well integrated, people who have been here for years or generations and speak fluent English.

What about telephone polling? Well at 5% incidence just randomly ringing numbers and asking if they are Muslim isn’t a feasible route. Things that have been tried in the past include re-contacting Muslims who have taken part in past general polls and indicated a willingness to take part in future polls, snowballing (that is, asking Muslim interviewees for contact numbers for other Muslims who would be willing to take part) or even just ringing up people with “Muslim names” in telephone number databases. These all have their own potential biases.

The final approach is face-to-face sampling, knocking on doors and asking to interview people. This has the same problems as telephone of Muslims only being 5% of the population, but it can be tackled by knocking on doors in areas with a high proportion of Muslims. Here comes the compromise: if you are knocking on doors in Tower Hamlets one in three households will be Muslim, if you are polling in Cornwall only one in five hundred households will be Muslim. Face-to-face polls of British Muslims therefore ignore those areas with a very low percentage of British Muslims, where it is not financially feasible to knock on hundreds of doors for every interview. This inevitably produces a skew towards those British Muslims who live in Muslims areas, but it is a matter of degree how mild or serious it is.

So if we go back to the Ethnic Minority British Election Study back in 2010, they only did interviews in areas that were least 2% BME in the census, which covered 88% of the BME population in Britain. They had a budget that was the best part of a million quid though, and I doubt Channel 4 were willing to go that far for a poll. By necessity ICM’s poll was more limited. It covered areas (Local Super Output Areas to be specific – it’s an ONS defined area of about 1000-1500 people) that are at least 20% Muslim. This covers about 51% of the British Muslim population, meaning the 49% of British Muslims that live in areas with a lower concentration of Muslims were not included in the poll.

For obvious reasons it is likely that a British Muslim who lives in an area where all their friends and colleagues are also Muslim may have different attitudes to a British Muslim who lives in an area where there are few other Muslims and their friends and colleagues are mostly non-Muslim. It also means the poll was probably skewed towards areas of relative social deprivation, and perhaps towards Muslims of particular ethnic backgrounds. The poll would not have been perfect… but then, no other poll of British Muslims would be either. It’s probably the best attempt to poll British Muslims properly that we’ve seen for several years and, given no one is waiting around the corner with a cheque for a million quid to do a more elaborately sampled poll than ICM’s, I think we should probably take this one seriously, but having due regard for the limitations of the sample. This is a poll of those British Muslims living in areas with a comparatively high Muslim population, which may well mean they are less integrated and have more conservative views than British Muslims living in areas that are overwhelmingly non-Muslim. With that caveat aside, what does it actually say about those British Muslims?

Let’s start by confirming the finding of previous surveys – the overwhelming majority of British Muslims identify as British. In fact, more so than the British population in general – 86% of British Muslims identify as British, 83% of the GB population in general. British Muslims are more likely to feel they can influence decisions in their area than most people in GB, feel better represented by their MP and local councillors. In terms of belonging and confidence they are part of the polity, British Muslims seem very well integrated.

ICM also asked about various measures of social conservatism. British Muslims were consistently more socially conservative than the British population as a whole, strikingly so in questions about attitudes towards homosexuality. 33% of Muslims thought boys and girls should be educated separately, 47% disagreed it was acceptable for a homosexual to teach in a school, 52% disagreed that homosexuality should be legal, 39% said wives should always obey their husbands.

The next section explored the issue of anti-Semitism. In terms of attitudes towards Jewish people themselves, British Muslims were not hostile. ICM asked respondents to express their feelings towards different religious and ethnic groups on a thermometer. On average Muslim respondents rated their feelings towards Jewish people at 57, compared to 64 among the GB control sample. Not a huge gulf, though it was larger than the gap on parallel questions about Catholics, Protestants and so on. The Muslim sample were, however, less likely to say they thought anti-Semitism was a problem in Britain today and were significantly more likely to agree with a range of anti-Semitic tropes than the wider GB population were. Around a third of British Muslims agreed with statements about Jews having too much power and influence in Britain and the world, compared to about one in ten in the control sample.

The final part of the survey dealt with attitudes towards violence and terrorism. This is the often the most controversial part of polls of British Muslims, and the bit that is often rightly criticised. It is important to be careful with wording and it is crucial that there is a control sample of non-Muslims to avoid painting Muslims as unusually supportive of violence or terrorism when non-Muslims would actually answer questions in the same way. The ICM poll does well on both, asking a broad range of different scenarios and issues, all also asked to a GB control sample.

Asked about the use of violence in general, answers of British Muslims and the GB control sample were not that different. Sympathy for violence against government injustice or police injustice were similar. Muslim respondents were more sympathetic for violence in defence of religion, the GB sample were significantly more sympathetic towards violence to protect one’s family.

Asked about support for terrorism, British Muslims were more likely to say they were sympathetic to terrorism than the GB control sample, but the net figures were extremely low in both cases.

Asked about organising radical groups, the GB control sample was the more sympathetic. 11% would sympathise, 74% condemn. Among British Muslims the figures were 6% sympathise, 75% condemn.

Now asked about making threats of terrorism, 6% of British Muslims said they would sympathise, 79% condemn. The figures in the GB control sample were 2% sympathise, 95% condemn.

of terrorism, 6% of British Muslims said they would sympathise, 79% condemn. The figures in the GB control sample were 2% sympathise, 95% condemn. Asked about actually committing terrorist actions, 4% of British Muslims said they would sympathise, 83% condemn. In the GB control sample 1% would sympathise, 95% would condemn.

The survey then asked more specifically about issues around ISIS. 7% of British Muslims said they supported the principle of ISIS’s aims – the creation of a caliphate – 67% were opposed. However support for the principle of an Islamic State does not necessarily imply support for ISIS’s actions, asked if they supported how ISIS was attempting to set up an Islamic State support fell to 3%.

So overall, we have a picture of a British Muslim community that identifies with Britain. It has views that that are much more socially conservative than Britain in general, particularly on homosexuality. The overwhelming majority of British Muslims condemn terrorism and ISIS, but a tiny minority do not. There is nothing here that is a huge surprise, but it has been a long time since we’ve had any hard data to back it up with. In terms of the way the poll was done, remember that the sampling did only cover areas with a comparatively high Muslim population. It’s not as crude as picking by local authority – taking LSOAs means it will include pockets of Muslims people across the country. It doesn’t cover the 49% of Muslims who live in areas that are less than 20% Muslim though, where I think it likely British Muslims are more integrated and have more in similar with their non-Muslim neighbours. Even if that does make a difference though, and views of other British Muslims are less distinct from the rest of British society, this poll should give us a good guide to the 50% of British Muslims who live in areas with a comparatively high Muslim population.

The full data is here.