Britain’s former equalities chief Trevor Phillips created a swathe of headlines last week by declaring the nation is “in danger of sacrificing a generation of young British people to values that are antithetical to the beliefs of most of us”. But, argue Bridget Byrne and James Nazroo, this sweeping statement is based on misrepresentation of some seriously flawed research.

In his opening sentence to his Daily Mail article ‘What do British Muslims Really Think?’, Phillips neatly achieves that trick of implanting a thought in your mind at the same time as seeming to disavow it: ‘As a doctrine of religious belief, Islam never held any terrors for me’.

The association of terror and Islam is set up from the outset. In the articles (in both the Daily Mail and The Sunday Times) Philips wrote over the weekend of 9 and 10 April, and in the broadcast on Channel 4 last week (13 April), Phillips warns ‘us’ (assumed to be largely secular, and certainly not Muslim, ‘liberals’) that we must not be complacent against the ‘nation within a nation’ of British Muslims.

He tells us that we ought to be scared – and that we should not be distracted by the likes of Nadiya Hussain of ‘Bake Off’ fame, or Mo Farah. According to the programme, there is a ‘chasm’ opening up between Muslims and non-Muslims.

Using the ever-elastic notion of ‘some’ Muslims, we are told that Muslims are not interested in associating with non-Muslims and hold conservative views about sexuality and gender which are suggested to stand in stark contrast with the rest of Britain.

The programme is compelling, drawing on survey data and delivered with urgency and authority by Trevor Phillips, one of the few public commentators guaranteed to have his voice heard on issues of race, ethnicity and multiculturalism in Britain. Channel 4, Trevor Phillips, and ICM Unlimited, the company commissioned by Channel 4 to conduct the survey, make the emphatic, but wholly inaccurate, claim that this is the largest survey of Muslim people in Britain. And they use this claim to draw unfounded conclusions about the views of Muslim people in Britain.

There have been a number of surveys that have included as large, or larger, samples of Muslim people in Britain over the past decade, which have been well designed and provided good evidence on people’s experiences and attitudes. These include Understanding Society, the Ethnic Minority British Electoral Survey and the Citizenship Survey. In terms of size, the survey they use is by no means unique. However, it is unique in having three fundamental flaws in its design, any of which would lead to serious errors. The conclusions drawn from this evidence are misleading, irresponsible and dangerous.

The first problem is that the survey was designed to only be conducted in areas where the local population is made up of at least 20% Muslim people. Yet only half of all Muslims in the UK live in areas with a 20% or more concentration of Muslims. This gives an unrepresentative sample that completely fails to include those Muslims who live in areas with lower numbers of Muslims.

In the sample, 12% of the interviewees came from areas of extremely high (more than 70%) levels of Muslim residence and 28% from areas with more than 50% levels of concentration. It is perhaps telling that, in the tables provided by ICM, this is referred to as ‘penetration of Muslims’, language suggestive of an alien, enemy invasion. But areas with such high levels of concentration of Muslim residents are unusual, less than 20% of Muslims live in areas with a more than 50% concentration, and these are also some of the most deprived areas of the UK.

Trevor Phillips recognises this criticism in his presentation of the programme, but in doing so makes the ludicrous claim that Muslims living in areas not covered by the survey design are the same as those who live in the areas with a higher concentration of Muslim people.

This is wrong; our economic circumstances, jobs, educational outcomes and, importantly, beliefs and attitudes, are all related to the type of area in which we live. And the type of area in which we live shapes our opportunities and consequently our attitudes.

The second problem is that the survey design meant that once an area was randomly selected from those with a high proportion of Muslim residents, Muslim people within the area were selected to be interviewed in a way that made them not representative of the other Muslim people living in the area.

To make them representative the survey company should have followed a known probability sample design, where everybody in the area had a known chance of being included. Instead, they followed a quota design, which simply meant that the interviewers only had to ensure that they interviewed eight people per area, just so long as they fit some crude pre-selected demographic characteristic. Who they actually interview does not matter. When this approach is taken, the interviewees end up being people who are convenient to interview and who are, as a consequence, not representative.

These two problems mean that the sample of people included in the survey are not representative of the Muslim population of Britain. And then this is compounded by the third problem with the work; that the ‘control’ sample of British people (that is, the survey that is used to draw comparisons with the survey of Muslim people), is not of people living in the same areas as the Muslim people included in the survey. They live in different areas, and area is important, as described above, which means that they are not the right people to draw comparisons with.

Given these problems, it was highly misleading for conclusions to be drawn about Muslim people in Britain, or about differences between Muslim people and non-Muslim people in Britain. And these are well known problems that those involved in the programme must have known about.

These concerns about the quality of the evidence are aggravated when we consider how some ‘facts’ were presented to us, where inflammatory claims are made without empirical justification. The conclusions are amplified by conflating responses of ‘strongly agree’ and ‘tend to agree’. This conflation produces headline-grabbing results but ignores the hesitation and nuance that lies behind the ‘tend to agree’ response. This nuance is absent from a programme which claims that the problem of Muslims in Britain is their religion which tells them what to do and believe ‘every minute of every day’.

The programmes also extrapolates from the raw numbers of the survey responses (which are themselves skewed to, but unrepresentative of, those living in particular situations, as explained above) to the whole population.

For example, we are told that four percent of the Muslims interviewed had sympathy for suicide bombing – and then told that this represents ‘more than 100,000 Muslims’, a dangerous and unsustainable claim.

In fact, according to the survey, 1% of non-Muslims interviewed also are sympathetic to suicide bombing – which, if dealt with in the same manner would ‘represent’ more than 600,000 people in the UK population. If the team responsible for this believe their data, why do they not see this figure as a major matter of concern and report it? Further discussion on the problems with the use of surveys such as this to estimate sympathy for extremism can be found in an earlier blog post on this site.

The programme seeks to ‘explain’ sympathy for violence and extremism among Muslims by associating this with a lack of integration or self-segregation. The idea of self-segregation is something that Trevor Phillips has been long exercised about.

In 2005, he warned that Britain was ‘sleep-walking into segregation’ and he continues to argue for this despite solid evidence to the contrary that he is fully aware of – for example, Nissa Finney and Ludi Simpson’s ‘Sleepwalking to segregation’? Challenging myths about race and migration’ (Policy Press, Bristol, 2009).

Despite Phillips’ alarmism, analysis of the most recent census shows the continuation of the trend for Muslims – like almost all religious groups (and like most ethnic groups) – to be increasingly spread out in terms of where they live.

This irresponsible broadcast follows familiar islamophobic paths; it sets up a ‘them and us’ relationship between Muslim and non-Muslim Britons, with Muslims characterised through a flawed survey as homophobic, extreme, and patriarchal (and non-Muslim Britons as none of these).

It sees integration purely in terms of geographical segregation, where the central problem always becomes one of ‘too many Muslims’. Finally, throughout the programme, Muslims are associated with violence and terror. And the evidential basis for each of these claims is seriously flawed.

Channel 4, Trevor Phillips and ICM should acknowledge these flaws, admit that this broadcast has seriously misled the viewing public, and apologise.

• The views expressed in this article are those of the authors, and not necessarily those of the Centre on Dynamics of Ethnicity (CoDE)