Attempts to smear the entire Muslim population on the basis of dubious opinion polls should be seen for what they are – unfounded and dangerous bigotry

The recent ComRes Muslim opinion poll for the BBC Today programme has elicited the usual stream of anti-Muslim commentary from right-wing columnists.

The results of the poll of 1,000 Muslims showed that 85 percent disagree with the statement that ‘Organisations which publish images of the Prophet Mohammad deserve to be attacked’, 94 per cent agreed that ‘If someone I knew from the Muslim community was planning an act of violence I would report them to the police’, 95 percent said that they felt a loyalty to Britain, and 85 percent said the statement ‘I would rather socialise with Muslims than non-Muslims’ didn’t apply to them.

However newspaper headlines focused on the poll finding that 27 percent agreed with the statement ‘I have some sympathy for the motives behind the Charlie Hebdo attacks in Paris’.

Stephen Glover of the Daily Mail declared the findings ‘profoundly chilling‘, concluding that ‘We have to face the fact… that a significant minority do not share British values of tolerance, and that for some of these people violence seems to be a legitimate response to perceived slights to their religion.’

Dan Hodges of the Telegraph described the result as ‘a shocking figure. And an utterly shaming one for Britain’s Muslim community. If this poll is accurate, over a quarter of British Muslims overtly sympathise with the motives of those responsible for the cold blooded murder of 16 journalists, police officers and Jews.’

Hodges concluded, ‘It’s not good for people inside and outside the Muslim community to continue to turn a blind eye to the extremism that continues to fester in the heart of the Muslim community. It’s not good enough for Muslims to keep delivering vacuous homilies about “the religion of peace” when surveys show 27 per cent of Muslims have sympathy with the Charlie Hebdo murderers.’

The poll question, which used the words ‘sympathy for the motives‘ behind the attack is twisted by Hodges to become ‘sympathy with the Charlie Hebdo murderers’ in order to support his view that the Muslim population is offering tacit support for a violent extremism ‘festering’ within their community.

‘Muslim polls’

The ComRes poll is just the latest in a series of opinion polls that purport to reveal the true opinions of Britain’s Muslim population.

However these polls don’t represent unmediated, unbiased access to the opinions of Muslims:

  • Muslim-only polls are inherently unreliable. They do not contain a ‘control’ sample – the same questions are not asked of non-Muslim respondents to establish a baseline figure of support for terrorism. Consequently the polls and the resulting commentary all (incorrectly) assume a zero-level of support for terrorism within the non-muslim British population.
  • The results are influenced by question wording and response options, with certain types of questions likely to generate inflated figures for support for terrorism.
  • Variations in levels of recorded support for terrorism between religious groups can be explained by age and socio-economic factors rather than religion.

An academic study by Maria Sobolewskai of Manchester University examined fourteen Muslim polls asking about support for terrorism and British Islamic identity between July 2005 and December 2006.

Sobolewska found that these polls were often of poor quality.

‘They tend to be small samples which are not drawn probabilistically, and hence over-representing specific types of people: those easier to find and more likely to respond to the pollster…Most often the samples are quota samples or internet-based samples that are largely unrepresentative.’

Pollsters often ‘weight the responses to achieve a sample that represents the demographics of a certain group (and rarely report it) and so, if they do not have enough young people in the sample, they will multiply their responses to achieve ‘more’ young people. This makes the proportions quoted by these polls doubtful at best.’

In the ComRes study quoted above, the 18-34 age group contained 463 respondents and so was was weighted to become 530. This wasn’t mentioned on the BBC page reporting the results. Nor was it mentioned by the articles in Telegraph, the Mail, or The Independent

There’s also the issue of question wording:

‘An examination of how questions are asked by these polls gives the impression that they are ‘fishing’ for the desired answer. The polls conducted soon after Muslim terrorism became a public concern in the wake of the 7/7 bombings simply asked whether terrorist attacks were justified and violence was acceptable. These were simple questions and were normally answered simply and consistently ‘no’.

However, results like those make no headlines and sell no newspapers, and as we move away from the time of the attacks the questioning grew more complex, adding various qualifications and circumstances under which attacks may take place and asking about various other attitudes towards them such as sympathy and understanding.

These later polls then showed much higher levels of apparent Muslim support for terrorism, clearly reflecting the public preoccupation with the potential for tacit and weak forms of support.’ii

The use of words such as ‘sympathy’ and ‘understanding’ in poll questions is clearly problematic. There is compelling evidence that question wording determines the level of support for terrorism recorded and that similar levels of support for violent extremism exists within the non-Muslim British population.

Non-Muslim support for terrorism

The underlying assumption behind polls that investigate opinions on terrorism and focus exclusively on the Muslim population is that support for terrorism or ‘violent extremism’ within the general population is zero. This is not, in fact the case.

A poll conducted by YouGov in September 2011 as part of an academic study by Sobolewskaiiiasked a sample of 2,583 non-Muslim respondents a series of questions about their attitudes to terrorism. The sample size for this poll is over two and a half times greater than the recent ComRes poll and over four times the size of the majority of the polls conducted after 7/7.

The study measured the levels of support for terrorism among the general population, and looked at the impact of question wording and the notion of ‘tacit support’ for terrorism. On the basis of the results Sobolewska concluded that:

‘terrorism, political violence, and even suicide bombing receive far more support among the general British population than is usually implicitly assumed, that it is governed by similar hierarchy of causes and targets seen among Muslims and that tacit support is an artefact of measurement and not a distinct analytical concept.’

The first question asked respondents whether they thought terrorism was justifiable for any of eight named causes: animal rights, environment, Islamic extremism, protecting one’s culture, protecting one’s faith, fighting for independence for one’s country, fighting an oppressive regime and opposing a foreign invasion.iv

The resultsv demonstrate that a significant proportion of the non-Muslim British public believe that terrorism can be justified (Figure 1)

Seven percent of British non-Muslims agree that terrorism can be justified for the cause of ‘Islamic extremism’, 8 percent for environmental causes and 9 percent for ‘protecting your religious faith’. In some circumstances over half of British non-Muslims are prepared to offer explicit support for terrorist attacks – even when civilians are listed amongst the targets.

Figure 1

To test whether response options have any impact on results, a second group of respondents was offered a version of the same question with an additional response option: ‘Yes, can sometimes be justified, but only in extreme circumstances’.

Making this response option available led to higher percentages saying terrorism could be justified – rising to 10 percent for ‘Islamic extremism’, 14 percent for ‘environmental causes’, and 15 percent for ‘Protecting your religious faith’.

Support for suicide attacks amongst British non-Muslims was also demonstrated to be non-zero and influenced by the location and type of targets (Figure 2).

Four percent agreed that suicide attacks could sometimes be justified in the case of ‘British citizens opposed to the British government’s wars in Afghanistan and Iraq’, rising to 8 percent in the case of ‘Palestinians fighting Israel in Gaza and the West Bank’, or against military targets.

Figure 2

A differently worded question asked whether respondents could understand the motives for suicide attacks. Around 30 percent of British non-Muslims agreed ‘I do understand why some people might behave in that way’ (Figure 3).

Question wording clearly has an impact – in the case of ‘suicide attacks’ asking whether respondents could ‘understand why’ generated a much higher positive response compared to asking whether such attacks can be ‘justified’.

Figure 3

Further support for the importance of question wording comes from a comparison between results obtained by the Northern Ireland Life and Times Survey 2007 and the European Values Study 2008. Each survey used differently worded questions to measure support for terrorism. The pattern described above is repeated.

The Northern Ireland Life and Times Survey asked 1,179 respondents if they had ‘sympathy’ for the violent attacks conducted by Loyalists and

The percentages reporting some sympathy for violent Loyalism or Republicanism varied according to age and religion, but ranged from 20 percent to over 40 percent (Figure 4)

The European Values Study the following year asked if there were any circumstances where terrorism could be justified.vii

Eight percent of people in Northern Ireland believed that terrorism can be justified in certain circumstances (Figure 5).viii

Figure 4

Figure 5

Far more people are prepared to say that they have some ‘sympathy’ for the reasons for Loyalist or Republican violence compared with those who agree that terrorism can be ‘justified in certain circumstances’.

The evidence shows that question wording influences the size of the constituency expressing some support for terrorism and suggests that ‘tacit support’ is indeed ‘an artefact of measurement and not a distinct analytical concept’.

Socio-economic factors

The Northern Ireland Life and Times Survey results presented above showed that age is a significant factor influencing support for terrorism.

Population groups with a higher proportion of young people will therefore show slightly higher levels of support for terrorism.

An analysis of the 2010 British Citizenship Survey demonstrates that age, socio-economic group and income are linked with support for the use of violent extremism.

The survey, which has a sample size of just under 7,000, asked respondents ‘whether it is right or wrong for people to use violent extremism in Britain to protest against things they think are very unfair or unjust’.ix

Violent extremism was defined as ‘taking actions to cause injury or death to people in order to make a political protest’

The results show that ‘In April-December 2009, 85 per cent of people in England and Wales said that it was ‘always wrong’ for people to use violent extremism in Britain to protest against things they think are very unfair or unjust. Eight per cent felt that it was ‘often wrong’ and 7 per cent felt that it was ‘sometimes right, sometimes wrong’. One per cent felt that it was ‘often right’ while less than half a per cent felt that it was ‘always right’.’

Multivariate analysis showed that ‘Age, socio-economic group and income were the strongest predictors of whether someone rejected violent extremism’ (my emphasis).

Young people aged 16 to 19 years and 20 to 24 years were more likely to support violent extremism (Figure 6).

Figure 6

A respondent’s socio-economic group also showed a strong relationship with their attitudes towards violent extremism – those in managerial and professional occupations showed the least support for violent extremism with only 3% selecting ‘always right’, ‘often right’ or ‘sometimes right, sometimes wrong’. On the other hand students showed the highest levels of support for violent extremism with 17% selecting those responses.

Income was also an important factor in explaining attitudes towards violent extremism (Figure 7).

People earning under £5,000 per annum were least likely to reject violent extremism, with a total of 14 percent saying that it was ‘sometimes right, sometimes wrong’, ‘often right’ or ‘always right’ to use violent extremism.

Figure 7

The study found that there were variations in the level of support for violent extremism by religious affiliation, but only ‘If religious groups are compared without taking account of the substantial variations between populations (for example age and socio-economic differences)’ (emphasis in original).

In fact ‘There were no statistically significant differences between Hindus, Muslims, Sikhs and people with no religion.’ (Figure 8)

Figure 8

As far as religious affiliation is concerned the analysis concludes:

‘while Muslims and Hindus are, as a group, less likely than Christians to reject violent extremism, the differences may be explained by their younger age profile and/or socio-economic profile. The Christian population has an older age profile and would therefore contain a larger proportion of people who rejected violent extremism. It may therefore be age, rather than faith, which explains differences between Muslims, Hindus and Christians.’

The Hindu, Muslim, Sikh and ‘no religion’ groups all have a much younger age profile than the Christian population or the general population.

35 percent of the Muslim population are between the ages of 16 and 34. The figures for the Christian and general population are 20 percent and 25 percent respectively.x

The Muslim population is also significantly less wealthy than the Christian population, wage levels for Muslim men are 17% and women 22% less than for Christian males.xi

In 2008 average household wealth for Christians was estimated at £223,000 compared to just £42,000 for Muslims.xii  According to the 2011 Census, the unemployment rate for Christians is 6% compared to 17% for Muslims.

Evidence gathered by a large study of young people between the ages of 12 and 18 in the UK, France and Spain also tends to favour the conclusion that relatively small variations in support for violent extremism are due to socio-economic factors, it concludes:

‘..young people who had experienced social marginalisation and discrimination were highly likely to support the use of violence and more especially, to engage in emotional and physical violence themselves…however, there was no indication that Muslim youths…were more likely than non-Muslims to be emotionally or physically violent towards others, once other aspects of discrimination and social marginalisation had been taken into accountxiii(my emphasis)

The evidence shows that the Muslim population of Britain is not more supportive of violent extremism than the non-Muslim population.

However some have used Muslim poll results to infer that within the Muslim community there is a circle of ‘sympathisers who supply, however unwittingly and passively, the crucial oxygen for those committed to and involved in violent confrontation’.xiv According to Michael Portillo Muslim terrorists are surrounded by ‘a penumbra of disaffected Muslims who may not condemn their crimes or denounce their murderous plots’xv.

Taking the poll evidence at face value these commentators then generate estimates of the number of Muslims who supposedly offer ‘tacit support’ or even form a ‘pool from which new bombers can be recruited’.

The evidential basis for these claims is flimsy; there is some support for violent extremism in all communities and small variations in question wording and response options can produce inflated figures.

If we accept the results of the 2010 Citizenship Survey, over 3.5 million British people think that violent extremism is ‘always right’ ‘often right’ or ‘sometimes right, sometimes wrong’.

Using the Sobolewska’s 2011 YouGov study results we can infer that over 3 million non-Muslims believe that terrorism in the name of ‘islamic extremism’ can be justified, that over 13 million non-Muslims say they ‘understand’ why some people would conduct suicide attacks, and that millions of non-Muslims support terrorist attacks in a range of contexts and for a variety of causes – even when civilians are explicitly referred to in the question.

If the few hundred thousand Muslims who provide similar responses to the same questions represent a circle of ‘tacit support’ for terrorism then what are we to make of the fact that a group of non-Muslims greater than the size of the entire Muslim population of Britain think that ‘islamic’ terrorism is justified?

The inescapable conclusion is that the government’s anti-terrorism policy, with its exclusive focus on surveillance of the Muslim population is completely misguided, if not counterproductive.

Attempts by newspaper columnists to smear the entire Muslim population on the basis of ‘Muslim’ opinion polls should be seen for what they are – unfounded and dangerous bigotry. They contribute to a climate of hostility towards Britain’s Muslim population that increasingly finds expression in Islamophobic hate crime.xvi

Far from defending ‘liberal British values of tolerance’ these commentators arguably embolden the minority of violent extremists who target innocent Muslims.


i Maria Sobelewska, ‘Religious extremism in Britain and British Muslims: threatened citizenship and the role of religion’ in ‘The New Extremism in 21st Century Britain, Roger Eatwell and Matthew J. Goodwin (Eds.), Routledge 2010, p. 25. Available online:

ii Ibid. p. 28

iii Maria Sobolewska, ‘Measuring support for terrorism: a survey experiment and an attempt at a comparison‘ quoted with kind permission of the author

iv Question: ‘There are many causes of terrorist attacks in the modern world. By terrorist attack we mean any intentional violent action against civilians, military, police or governments. Are there any circumstances under which you think that terrorist attacks can ever be justified for any of the following reasons? Please say yes or no in each case…’

The responses offered for selection were:

  • Yes, can sometimes be justified
  • No, can never by justified
  • Not sure

v Results:

  • ‘animal rights’ – 8 percent
  • ‘environmental causes’ – 8 percent
  • ‘Islamic extremism’ – 7 percent
  • ‘Protecting your religious faith’ – 9 percent
  • ‘Protecting your cultural values’ – 13 percent
  • ‘Fighting for independence for your country’ – 42 percent
  • ‘Fighting against an oppressive regime’ – 52 percent
  • ‘Fighting against foreign occupation of your country’ – 54 percent

vi ‘Question: thinking about the reasons why some [ Loyalist / Republican ] groups have used violence during the troubles, would you say that you have any sympathy with the reasons for the violence – even if you don’t condone the violence itself?

The response options were

  • ‘a lot of sympathy’,
  • ‘a little sympathy’
  • ‘no sympathy at all’
  • ‘don’t know’

vii Question: ‘Terrorism is everyday news. In principle, most people are against it, but there is still room for differences of opinion. Which of these two statements do you tend to agree with? Would you say… “

Available responses:

  • There may be certain circumstances where terrorism is justified
  • Terrorism for whatever motive must always be condemned
  • Neither

viii The same survey was conducted on mainland Britain with 6 percent selecting the same response.

ix The available responses were

  • Always right
  • Often right
  • Sometimes right, sometimes wrong
  • Often wrong
  • Always wrong

x Census 2011 – the proportion of the population between the ages of 16 and 34 for the Hindu, Sikh and No Religion groups are 37%, 34% and 35% respectively

xi Simonetta Longhi and Lucinda Platt, Pay Gaps Across Equalities Areas, Institute for Social and Economic Research University of Essex, Equality and Human Rights Commission 2008, p. ix. Available online:

xii Karen Rowlingson, ‘Wealth inequality: key facts’, Birmingham University, Policy Commission on the Distribution of Wealth, December 2012, p.19

xiii McVie, S & Wiltshire, S 2010, ‘Experience of Discrimination, Social Marginalisation and Violence: A Comparative Study of Muslim and Non-Muslim Youth in Three EU Member States‘. in: Scottish Centre for Crime & Justice Research. European Agency for Fundamental Rights. p. 77. Available online:

xiv Shamit Saggar, Pariah Politics, Oxford University Press, 2009, p. 214

xv Ibid, p.213

xvi Metropolitan Police Crime statistics show a 133 percent increase in Islamophobic crimes for January 2015 compared to January 2014.

Tagged under: