Tuesday, 30 December 2014

Rejecting Narratives: Data, Islam and Terrorism

The Year of the Monkey

(As always, footnotes and bibliography are at the bottom of the post). Many of my friends in my ‘ideological camp’ do not seem too pleased with the rise of Vox, 538 and The Monkey Cage. I applaud it as an open break from some of the worst journalism we see. Op-ed writers will write streams about housing[1], immigration[2], foreign policy; and not utilise a pool of peer reviewed, robust empirical literature. Vox and 538 are reversing this trend. And I agree that they may sometimes present the literature through their own ideological lens – but once people accept a form of the scientific method, it’s very easy to have a conversation (and, of course, make sounder judgements). 

This divide between journalism and the academic literature is not new. Back in 2010, The Monkey Cage was still a blog and, as always, they were writing a constant stream of posts which debunked journalists unsubstantiated ‘narratives’ by using the empirical literature. A year later, they came out with a constructive paper, the purpose of which was to help journalists. Often, journalists will place great emphasis on a speech, a presidential debate, gaffes etc – when the literature is fairly clear that this stuff doesn’t really matter. In a somewhat comical response to Sides, Francis Wilkinson wrote

...the media's capacity for creating self-serving, fanciful political narratives is more constrained today than ever. An army of spoilsports -- many with Ph.Ds in political science -- has established camp on the banks of the Web... Take John Sides, a political scientist at George Washington who runs the annoyingly excellent Monkey Cage blog. The guy is a total downer.

Every time some reporter starts to have a little narrative fun, Sides gets all political science-y on them... Look, I'm basically on the side of the "narrative" guys. I enjoy making up half-baked theories and then sending them downstream and seeing what happens.

Data driven journalism is clearly a welcome response to this problem. On my side of the Atlantic, we haven’t really seen a comparable change in our press. I’m not about to make any claims about how many articles aren’t based on data (because I don’t have any data) but what I can say is that the problem still exists. One example is Iona Craig who wrote a rather inaccurate, unsubstantiated article about Yemen. She made claims about poverty, the views of Yemeni population and terrorism without citing a single study from a pool of research not only on terrorism, but specifically about Yemen. And rather than accept that the literature might have something to say on the matter (it does, and it says the complete opposite of her anecdotes), she was brazen enough to say that her personal experience trumped the literature, with a dash of anti-intellectualism:  

No further comment is necessary for such an ignorant and arrogant methodology. This long introduction is merely to emphasise what I have been trying to do in my posts on this blog: draw attention to the actual data, studies and research. My efforts have thus far been targeted toward the “anti-imperialist” and anti-war corners. I have tried to show that there is no link, association and causal relationship between civilian casualties, “occupation”, the war on terrorism, poverty and terrorism. The structuralist idea that it is external conditions that cause grievances that lead to terrorism is simply not supported by the data. In fact, the use of military force is actually associated with declines in terrorism (see hereherehereherehere and the last study here).

I tweet new studies frequently which add to what I call the emerging consensus against Robert Pape’s thesis. I would urge everyone to beware of “The Man with One Study” and carry out your own investigations but what follows are recent studies that I think reinforce that consensus and, in one or another, erodes the Greenwaldian thesis[3]:
  •  Berger, ‘What shapes Muslim public opinion on political violence against the United States?’, Journal of Peace Research,  Volume 51, Issue No. 6 (2014) – “...perceptions of controversial US policies toward Israel, Middle Eastern oil, or the perceived attempt to weaken and divide the Muslim world are not related to support for attacks on civilians in the United States... Approval of attacks on US civilians is shaped, instead, by negative views of US freedom of expression, culture, and people, disapproval of the domestic political status quo and the notion of general US hostility toward democracy in the Middle East” (see this table)
  • Ahmad, ‘The Role of Social Networks in the Recruitment of Youth in an Islamist Organization in Pakistan’, Sociological Spectrum, Volume 34, Issue 6 (2014) – “Findings reveal that young people who joined this organization did not necessarily do so because of their ideological affinity, political or social grievances or because of macro-level events occurring in the national or global arena, such as the U.S.-led war on terror” (see also this quote).
  •  Hultman et al, ‘United Nations Peacekeeping and Civilian Protection in Civil War’, American Journal of Political Science, Volume 57, Issue 4 (2014) – “Using unique monthly data on the number and type of UN personnel contributed to peacekeeping operations, along with monthly data on civilian deaths from 1991 to 2008 in armed conflicts in Africa, we find that as the UN commits more military and police forces to a peacekeeping mission, fewer civilians are targeted with violence. The effect is substantial—the analyses show that, on average, deploying several thousand troops and several hundred police dramatically reduces civilian killings” (see this table)
  •  D'Alessia, Stolzenberg and Dariano, ‘Does Targeted Capture Reduce Terrorism?’, Studies in Conflict & Terrorism Volume 37, Issue 10 (2014) – “Using quarterly data and an interrupted times-series Auto Regressive Integrative Moving Average (ARIMA) study design, we investigated the effect of Abimael Guzmán's capture on the ability of the Shining Path to wage its war against the Peruvian government. Maximum-likelihood results revealed that the frequency of terrorist acts committed by the Shining Path dropped by 143 incidents per quarter a short time after Guzmán was captured” (see this table).

These studies are all from late 2014 – all have come out since my last post. The literature on the Greenwaldian thesis is so clear that it really is like beating a horse that was killed years ago.  I could go on but I wanted to direct my focus to another argument.  At its most extreme, it comes from the right but in lesser forms, you’ll find it amongst many on the left: it’s the idea that terrorism has something to do with religion, or more particularly that it has something to do with Islam. In many ways, this narrative is as empirically weak as the anti-American blowback narrative.


I recently came across a “white paper” put out by Katharine C. Gorka (the kind of person who likes to talk about the persecution of Christians in the Middle East to the exclusion of Muslims). In it, she rebukes every single U.S President since Reagan because they have ‘tried to address the reality of terrorism against the United States by Muslims who claim jihad as their justification, while at the same time trying to avoid any direct condemnation of Islam in the context of that terrorism.’ This is misguided, Gorka claims, because it is Islam that is what motivates these people.

The empirical literature does not support the role that many, like Gorka, give to religion and Islam.[4] I will be looking at the literature in three areas: (a) on the incidence of terrorism (which will show there is nothing unique about Islam relating to terrorism); (b) on the characteristics of terrorists (which will show there is no link between religiosity and terrorism). In the next section, I will discuss three conceptual issues to maintaining a link between Islam and terrorism.

(a) On the incidence of terrorism

Conrad and Milton (2013) give a good starting point for looking at the literature. In their study, they note that if there is a connection between Islam and terrorism, ‘then countries with more adherents to Islam might experience and/or produce more terrorism than countries with fewer followers of Islam.’ After controlling for a wide range of variables, ‘Muslim states do not systematically produce more terrorism than non-Muslim.’ The variables they control for, however, are less convincing in terms of the more general research (this criticism is far too broad to detail here).

Nonetheless, it aligns with much of the data on terror threats within the U.S and Europe.  Loonwatch is rightly derided for having content that is usually un-nuanced and hyperbolic – but they do, in two posts, make particular good use of some official statistics. The first relates to the FBI’s statistics which report that from 1980 to 2005, “Islamic extremists” accounted for 6% of terrorist attacks. More recently, the New America Foundation has produced some data on who is behind terrorist attacks in the U.S. They find that Islamist terrorists have carried out four attacks killing 17 since 9/11 and right-wing terrorists have carried out eight attacks killing 9 people.

The second Loonwatch post relates to Europol data which, for an extremely limited data set finds that Islamist terrorists are responsible for less than a percentage of terrorist attacks. These two posts need to be qualified. Firstly, the data is intended to show that there is nothing unique about Muslim perpetrators and terrorism. I am well aware that the risk from Islamist terrorism has grown on U.S soil since 2001 – but the FBI statistic is still a significant data point. Second, for reasons that I have outlined before regarding Al Qaeda being a form of market state terrorism, Islamist terrorism deserves more of our attention – particularly in the UK where we do have separatist groups (which explains an overwhelming bulk of the Europol results) and Islamist terrorism is the biggest terror threat.

Thirdly, the data and research under this head does not apply to the majority of terrorism around the world. Given that I have said that Conrad and Milton study has significant drawbacks, it is significant that in ‘2013, 66% of all fatalities from claimed terrorist attacks were caused by four terrorist groups: the Taliban, Boko Haram , ISIL and al-Qa’ida.’ This is partly related to the previous caveat: Islamist terrorism is a form of ‘market state terrorism’ and causes vast amounts of casualties. But these three qualifications do not negate the point: there is clearly nothing unique to being Muslim and carrying out terrorist attacks (i.e, there are so many other groups that make up a significant amount of terrorism). 

The most significant data point on the incidence of terrorism is not the 66% figure, but the following from Kurzman’s The Missing Martyrs: fewer than 1 in 100,000 Muslims since 9/11 have been recruited by Islamist terrorists (p.11). Those who wish to explain a causal role need to explain why that figure is so low as to be insignificant. It’s a question I ask to blowback-advocates: if foreign policy is the cause, why do we find rejection of Al Qaeda to be the norm? Why, as the subtitle of Kurzman’s book asks, ‘are there so few Muslim terrorists’? In the Ahmad (2014) study quoted above, he says the following:

There are grievances to be found everywhere in the world, many of which never culminate in the membership of a radical party or the formation of a social movements. If there was a direct link between them, society would be swarming with countless organisation and movements struggling for [resolution].

I would ask the reader to replace ‘grievances’ with ‘religion.’ Of course, someone could respond to all of the above by saying ‘all that data shows is that terrorism is not unique to Muslims, that doesn’t mean Islam or being Muslim does not play a causal role in carrying out terrorism.’ That is fair (this subsection is really directed at more EDL types than Eustonite types). Hence, we move on to the literature on the characteristics of terrorists and their supporters.

(b) Characteristics of terrorists

The weakest form of evidence that could be used is to ask what the terrorists themselves think about the role of religion or foreign policy. Glenn Greenwald makes a habit of quoting what terrorists say, followed by “See?!”. Of course, many who make the link between Islam and terrorism will do the same thing. Of course, Greenwald will never cite the literature which shows the opposite. For example, in a recent study published in Studies in Conflict & Terrorism, Botha had a small sample of al-Shabab terrorists who asked why the joined the terrorist group. The results are summarised in the table below: 

 Given recent events, I have seen this quote from Zarqawi being used frequently:

I swear by God that even if the Americans had not invaded out lands together with the Jews, the Muslims would still be required not to refrain from jihad but to go forth and seek the enemy until only God Almighty’s shariah prevailed everywhere in the world (accurate translation from Hashim, Middle East Policy, Vol.1 Issue, No. 4)

In any event, just as we should not use quotes about why people become terrorists, we shouldn’t use the Botha study either. It’s steeped in social desirability biases and in any event, the literature is mixed so doesn’t provide any answers. The literature on the religiosity of terrorists, is, however consistent. Sageman (2008) in his sample of 500 terrorists found that a ‘lack of religious literacy and education appears to be a common feature among those that are drawn to [terrorist] groups.’ MI5’s Behavioural Science Unit (2008) undertook a study ‘based on hundreds of case studies by the security service’ to find the characteristics of terrorists, they found

...far from being religious zealots, a large number of those involved in terrorism do not practise their faith regularly. Many lack religious literacy and could actually be regarded as religious novices. Very few have been brought up in strongly religious households...

There is also support from Nielson (2014) (a study that will be discussed further below). These aren’t cherry picked studies, as Patel (2011) notes ‘overall, the available research does not support the view that Islam drives terrorism or that observing the Muslim faith—even a particularly stringent or conservative variety of that faith—is a step on the path to violence.’

Note that we find the same results when we widen the pool to those who support terrorism. Berger (2014) uses a vast amount of survey data from across the Muslim world to assess Muslims’ views on terrorism, U.S culture and norms and their religiosity (it was quoted above in the list of studies against the Greenwaldian thesis). Berger’s (fourth) hypothesis was ‘among Muslim publics, support for political violence against US civilians is associated with greater religiosity.’ What Berger found was no clear relationship between Islam and support for terrorism:   

...earlier findings about the pacifying role of religiosity in the Muslim world find confirmation in the case of Egypt where those who claim to adhere to the Islamic precept of five daily prayers (3.7%) are substantially less likely to support terrorism against US civilians than those who did not (10.8%)... The relationship is the reverse in Pakistan [where] the probability of a respondent endorsing attacks on civilians in the United States increases from 4.3% if he does not pray five times a day to 13.5% if he does. [There is a] lack of a pattern in Indonesia.

If there was a link between Islam, we would expect terrorists and their supporters to be the most observant, pious and have the highest measures of religiosity. But we don’t. This explains why you shouldn’t be surprised when you read that those heading off for Syria purchase ‘Islam for Dummies’, the 9/11 highjackers went to a strip club, Zarqawi was a pimp and Bin Laden’s computer had a ‘considerable quantity of pornographic videos.’

Religion and Causation

There will be those who are not satisfied by the above. They may even engage in the following sophistry: being a Muslim is not related to terrorism, but Islam is. Islam does have an association/causal role in terror, but only amongst Muslims who take their faith seriously. Before moving on to the conceptual issues with this argument, it’s worth noting how weak this argument is through the Nielson (2014) study. Nielson looked at ‘27,124 fatwas, articles, and books by 101 contemporary clerics’ and tried to measure the characteristics that made them more likely to support ‘Jihadi ideology.’ In his study is hidden this nifty graph:

It’s very strange that the more educated a cleric is about Islam, the less likely it is that he will support ‘Jihadi ideology.’ Nielson had previously found that ‘clerics with the best academic connections had a 2-3% chance of becoming jihadist. This rose to 50% for the badly networked.’ In fact we know that Quran readers are more likely to support democracy (see this table from Hoffman and Jamal (2014)).

Leaving aside all of the this literature, there is a conceptual issues about causation, agency and responsibility. Many pride themselves on giving Middle Eastern Muslims agency, a concept that seems to have been robbed of them by the prevalence of structuralist views. But the argument about the connection between Islam and terrorism, seems to me to rob people of agency as well. Here is Sam Harris talking about the connection between Islam and certain behaviours (he is not necessarily talking about terrorism here):

I am never blaming Islam for all the bad things Muslims have done in history. I am only blaming Islam for the things that Muslims have done on the basis of the doctrine of Islam... All I am asking for is honest conversation about the logical connection between ideas and behaviour (38:16).

It should be apparent that one thing that I have not sought to do in this post is to define what ‘Islam’ says about terrorism. My own view is that Islam supports all manner of barbarity (in terms of the way it views women, gays, legitimate punishments) but that a reasonable interpretation does not endorse terrorism. A mainstream interpretation does not support terrorism. But this really isn’t an important point: what Harris is doing is shifting blame from the individual to a set of beliefs. This is a, at times, subtle but definitely important change in emphasis for a number of reasons. No longer are we blaming the individual who acted in a free, informed and deliberate manner but we are focusing on a text.  When you take away from someone blame and responsibility for their decision, you rob them of agency.

I consider the principles at work in a conversation about free speech to apply equally here. Currently in the U.S there is a political controversy regarding ‘rhetoric’ and the extent to which it is responsible for the death of two NYPD officers. “They have created an atmosphere of severe, strong, anti-police hatred in certain communities” Rudy Giuliani says “For that, they should be ashamed of themselves.” And it’s nonsense. People are not there to be acted upon by elites or texts or Anjem Choudary or Muhammed. They choose to act in that way, and if they choose to act in that way after reading the book and accepting it, the emphasis should be on their acceptance, not the book.

In response to  Giuliani’s statements, Kevin Drum wrote a pretty funny post on Mother Jones listing various deaths that the ‘right’ could be blamed for. He concluded with this: ‘Maybe lots of people support lots of things, and we can't twist that generalized support into blame for maniacs who decide to take up arms for their own demented reasons.’ I would go further, Islam at its most heinous interpretation cannot be blamed for the actions of its adherents. That's what causation is about. It can be criticised on rational grounds in the abstract. This point is a rather moot one given the empirical literature doesn’t allow people to make claims about the link between Islam and terrorism. But even if the literature reversed overnight, this point would still stand:

Muslims may act upon the organized collection of beliefs that comprise their faith... But it’s not because those ideas sprang to life, jumped up out of the pages of the Quran and into the minds of Muslims who were captive to their actions... Muslims make conscious choices to act and when they do, for good or bad, that capacity must not be diminished by fixating on lifeless doctrines.

Update (03/01/2015): I've been really pleased with the overwhelmingly positive response to this post. It's been shared, retweeted and viewed more than anything I've written. What's particularly pleasing is that it seems to have gained traction across the ideological divide. I'm glad the first section of this post summarised my views and gave further studies on the lack of a link between foreign policy and terrorism - it's usually people who rightly deny the link between Islam and terrorism that posit the blowback argument.

On Harry's Place (and below here), there have been some criticisms of this post. I don't consider any of them strong (I would say that wouldn't I?) and only two of them are really worth responding to. The first relates to the data used in the section on the 'incidence of terrorism'. WetWork who calls me up on use of my statistics from the New America Foundation. WetWork is right to argue that New America Foundation under-estimates Islamist terrorism and I should have noted that. But I also should have noted that the data 'may well understate the toll of violence from right-wing extremists' too.

But there is a broader point to make. This criticism is based on the premise that I am arguing that Islamist terrorism doesn't make up a disproportionate amount of terrorism. I am not. I explicitly acknowledge that 66% of the terrorism in the world is from four Islamist terrorist groups. I explicitly reject the Conrad and Milton study. I made explicit reference to what that subsection was doing and what it was not doing:
I will be looking at the literature in three areas: (a) on the incidence of terrorism (which will show there is nothing unique about Islam relating to terrorism)... Of course, someone could respond to all of the above by saying ‘all that data shows is that terrorism is not unique to Muslims, that doesn’t mean Islam or being Muslim does not play a causal role in carrying out terrorism.’ That is fair (this subsection is really directed at more EDL types than Eustonite types). Hence, we move on to the literature on the characteristics of terrorists and their supporters.
It's supposed to be a response to those who make the claim that 'not all Muslims are terrorists but all terrorists are Muslims.' To that extent, the New America Foundation, FBI statistics, the Conrad and Milton are useful. It's telling that WetWork doesn't criticise the FBI statistics. A commentor below, however, does criticise the FBI stats because it fails to take into account 'qualitative difference in crimes against property versus crimes against persons.' I agree, which is why I say 'for reasons that I have outlined before regarding Al Qaeda being a form of market state terrorism, Islamist terrorism deserves more of our attention' - and in that sentence I linked to an old post of mine which states
The substance of what Bobbitt says is important: they are global, decentralised and they cause a lot more deaths – and they intend to... And as crude as it is to suggest, larger or more frequent death tolls and attacks enhance the response to society in terms of consent. This is precisely why Islamist terrorism is such a threat: because of the intended fearful response it seeks.
Of course, I don't expect people to have read all the links I put in the post. But Anonymous' criticism suffers from the same fault of not realising the aim of that subsection: to show that terrorism is not unique to Muslims. The second criticism worth responding to comes from Lamia on Harry's Place. It's worth quoting the salient points in full:
Mugwump is implictly using a 'no true Muslim' argument of his own - namely that if Muslim terrorists are not highly educated and qualified in Islamic study then they are not true Muslims and it is not fair to in any way blame 'Islam'... I think the theological path is unhelpful and impractical to the point of, well, pointlesness, but Mugwump is essentially treading the same path as his opponents, albeit from the other end of the garden.  
Both are, in my view, erroneously using too narrow a definition of 'Islam' or indeed 'religion'. They both view Islam as a group of texts; they judge people to be proper Muslims in so far as they consider those people to accord with those texts. And here they get into a shouting match as group (a) will insist that Muslim terrorists are the real Muslims, because look at this verse, etcetera, while Mugwump and co will argue that on the contrary they are not real Muslims, because real Muslims who are better educated in their religion understand that the verses in fact mean this (or that; or whatever), etcetera, and only people with imperfect understanding of the religion are terrorists. 
...I would suggest that it is of no value whatsoever who is theologically correct because neither is practically correct. That is because a religion is not merely - or, perhaps, even mainly - its texts. It is both texts and associated humans and human structures (i.e. its congregations, believers, adherents both devout and nominal, preachers, theologians etc of the time, and the organising structures, both dogmatic and actual factual). That is to say, a religion is a mix of what it says and what its followers actually believe and - more importantly - actually do. 
The point is that while in certain respects a religion does not change much or at all - most usually and obviously with regard to its core texts - in other respects it can obviously change, and possibly greatly, especially in relation to the wider state and society... Likewise both Anjem Choudary and Majid Nawaz are constituents of 'Islam' today. All this is subject to change - for better or, of course, for worse.
I disagree. Islam's history is full of different schools of thought, different interpretations (something which a lot of people like Spencer and Gellar are loath to accept). But four things have to be stated: first, a definition of a religion which moves away from a doctrinal analysis is useless. Lamia accuses me of engaging in a 'No True Scotsmas' fallacy; the flaw of Lamia's view is to engage in 'Everyone is a Scotsman' fallacy. Second, the lack of a causal link here should be clear. It's bad enough when we hold a human being responsible for the acts of another human being. This becomes even more problematic when you try to make a causal link between a dead person/lifeless doctrines that are moulded by the a free actor. To put the points together: a Beatlist can listen to Lucy in the Sky, stab everyone and we can say (i) that is a legitimate part of Beatles-ideology and (ii) we can criticise the Beatles for their role in the stabbing. It's absurd.

Third, I am very explicit about why my analysis does not rely on a particular interpretation of Islam being the 'true' one. Agency eradicates the need to do this. (To this end, Jacobin makes a very weak argument in which he misunderstands what I'm trying to say about 'agency' when he says 'Mugwump is impressed and persuaded by the ideas of Adam Smith and JS Mill, so Ayman al Zawahiri was impressed and persuaded by those of al Banna'). Fourth, if we are really going to hold the faith responsible for an act because of its believers, then surely we have our evidence in the fact that the overwhelming majority of Muslims do not approve of Al Qaeda? Surely religiosity of adherents is relevant in assessing the link between doctrine and 'adherents'? It seems to be the distinction that Lumia is making doesn't affect the empirical analysis given above.

Which brings me to my concluding remark: in this post I extol the virtues of data, studies and peer reviewed literature. The most signficant section in this post is 'characteristics of terrorists' - and no one has directly addressed it. Even if you agree with Lumia, that section simply will not allow you to come to a different conclusion. There are other criticisms of the post on Harry's Place (some people ignorantly give weight to quotes from Islamists; others come up with howlers to undermine data like 'academia is no bastion of objectivity' and 'Having data isn't enough'; others still accuse me of Islam apolgetics) but they are weak enough that I don't want to spend time responding to them as people will see through them.


Apologies for the extensive footnotes that have very little to do with the substantive part of this post. I’m off work and have time so I thought I’d handle a few small things in the footnotes.

[1] In a particularly heinous Guardian video, Owen Jones tells us that the real issue with the housing market is property developers. Not once does he get anywhere near the real problem: supply-side restrictions on housing. The most enlightening research on this subject comes from Hilber and Vermeulen (2010) who find that house prices would be ‘21.5 to 38.1 per cent lower if the planning system were relaxed.’ This as Niemietz notes in Redefining the Poverty Debate is likely understating the issue considerably because they assume that no planning controls existed prior to 1974 and the model assumes further development restrictions (p.79). Again, this literature is fairly consistent. The ramp up in property prices, by the way, explains pretty much the entire rise in capital as a percentage of national income (the main measure Piketty uses in Capital for defining rising inequality) as you can see from the graph below taken from Atkinson (2011):

Note also that capital/income ratio is ‘actually stable or only mildly higher’ when you measure property not in terms of its value but the rents that it gathers (a more accurate measure of the wealth that it produces). This is a finding from Bonnet et al (2014) (see this chain of tweets for a summary and the important graphs).

[2] Given that this is a post about Islam, I wont bore the readers of my footnotes with what I hope they already know about the economic benefits of increased immigration. There is however some literature that addresses how well Muslims integrate into Western culture. There is a lot of polling to suggest, for example, that British Muslims have abhorrent views when it comes to homosexuality and free speech (see here for an aggregate of poll results). As I write later in this post, we shouldn't rob these individual of agency by saying that Islam is the cause. We should hold them accountable for their views (and fortunately, in many areas we see a divide which shows how there is no necessary connection). But it’s also worth somewhat downplaying the ‘creeping Sharia’ or ‘Muslims are going to change the character of our nation’ line of argument (not only because of the dubious demographic surrounding the issue).

This line of argument is not entirely without merit. Bisin et al (2007) find that Muslim immigrants integrate at a slower pace than non-Muslim immigration: 

Bisin et al’s study however, has not been replicated indicating that there was some kind of error (Arai et al (2011)). Bisin et al (2011) however accounted for the bad results and said that their results could be replicated, although somewhat weaker that the graph above. Inglehart and Norris (2012), however, find more optimistic results. Given Bisin et al’s weak results, the lack of replication it is worth focusing on Inglehart and Norris’ more robust findings:

the analysis demonstrates that the basic values of Muslims living in Western societies fall roughly half‐way between the dominant values prevailing within their countries of destination and origin. This suggests that migrant populations living in Rotterdam, Bradford and Berlin are in the process of adapting to Western cultures, while at the same time continuing to reflect the values learnt through primary socialization in their original countries of origin... n the long‐term, the basic cultural values of migrants appear to change in conformity with the predominant culture of each society.

... although Western Muslims are consistently located between Islamic and Western societies, there is no evidence that generational change, by itself, will transform the situation so that the cultural differences between Muslim migrants and Western publics will disappear: younger Westerners are adopting modern values even more swiftly than their Muslim peers.

Thanks to Ben Southwood for directing me toward these studies. I’m sure people can cherry pick Bisin et al or Inglehart and Norris – but at least then people aren’t making empirical claims without any empirical research. I have explained why I think the latter study is more rigorous but am open to saying the literature is not, at this point, conclusive. One indication however is to look again at these graphs and see that religiosity is declining. This is significant because once you account for religiosity, many of the socially conservative views of Muslims can be explained away (Lewis and Kashyap, 2013). See also these results from the U.S where Muslims are far more integrated.

[3] In the interests of transparency, I wanted to bring people’s attention to Romano et al (2013). I hadn’t come across this study until last week. It finds that ‘non-Anbar SOI rather than the troop surge reduced casualty rates in Iraq.’ This study does very little to counter the wealth of literature available on the Surge and how it was both Sons of Iraq and U.S troops that were required to reduce violence (Biddle et al 2012). It also does nothing to counter Smith’s (2007) findings either (see an elaboration of each in my last post).

[4] Generally, contrary to the widespread belief, it does not appear that religion is particularly violent either. See this article by Scott Atran, in particular: “The Encyclopedia of Wars surveyed 1,763 violent conflicts across history; only 123 (7 percent) were religious. A BBC-sponsored "God and War" audit, which evaluated major conflicts over 3,500 years and rated them on a 0-to-5 scale for religious motivation (Punic Wars = 0, Crusades = 5), found that more than 60 percent had no religious motivation. Less than 7 percent earned a rating greater than 3.”


Ahmad, ‘The Role of Social Networks in the Recruitment of Youth in an Islamist Organization in Pakistan’, Sociological Spectrum, Volume 34, Issue 6 (2014

Arai et al, 'On Fragile Grounds: A Replication of 'Are Muslims Immigrants Different in terms of Cultural Integration?', Journal of the European Economic Association (2011)

Atkinson, ‘Wealth and Inheritance in Britain from 1896 to the Present’, London School of Economic and Centre for Analysis of Social Exclusion (2011) available at < http://sticerd.lse.ac.uk/dps/case/cp/casepaper178.pdf>

Berger, ‘What shapes Muslim public opinion on political violence against the United States?’, Journal of Peace Research,  Volume 51, Issue No. 6 (2014)

Bisin et al, 'Are Muslim Immigrants Different in Terms of Cultural Integration?', IZA Discussion Paper (2007)

Bisin et al, ‘Errta Corrige: Muslim Immigrants Different in Terms of Cultural Integration?', Journal of the European Economic Association (2011)

Bonnet et al, ‘Does housing capital contribute to inequality? A comment on Thomas Piketty’s Capital in the 21st Century’, Sciences Po Economics Discussion Papers (2014) available at <http://spire.sciencespo.fr/hdl:/2441/30nstiku669glbr66l6n7mc2oq/resources/2014-07.pdf>

Botha, ‘Political Socialization and Terrorist Radicalization Among Individuals Who Joined al-Shabaab in Kenya’, Studies in Conflict & Terrorism, Volume 37, Issue 11 (2014)

Conrad and Milton, 'Unpacking the Connection Between Terror and Islam', Studies in Conflict & Terrorism, 36:4, 315-336 (2014)

D'Alessia, Stolzenberg and Dariano, ‘Does Targeted Capture Reduce Terrorism?’, Studies in Conflict & Terrorism Volume 37, Issue 10 (2014)

Hultman et al, ‘United Nations Peacekeeping and Civilian Protection in Civil War’, American Journal of Political Science, Volume 57, Issue 4 (2014)

Inglehart and Norris, 'Muslim integration into Western cultures: Between origins and destinations', Political Studies, Volume 60, Issue No. 2 (2012)

Lewis and Kashyap, ‘Are Muslims a Distinctive Minority? An Empirical Analysis of Religiosity, Social Attitudes, and Islam’, Journal for the Scientific Study of Religion, Volume 52, Issue No. 3 (2013)

Nielson, ‘Networks, Careers, and the Jihadi Radicalization of Muslim Clerics’ (2014) available at <http://www.mit.edu/~rnielsen/jihad.pdf>

Patel, ‘Rethinking Radicalisation’, Brenan Centre for Justice (2011) available at <http://brennan.3cdn.net/f737600b433d98d25e_6pm6beukt.pdf>

Romano et al, ‘Successful and Less Successful Interventions: Stabilizing Iraq and Afghanistan’, International Studies Perspectives (2013)

Sageman, Leaderless Jihad: Terror Networks in the Twenty-First Century (University of Pennsylvania Press, 2008)