Fake CVs reveal discrimination against Muslims in French job market

By Ed Yong | November 22, 2010 3:00 pm

MosqueMeet Khadija Diouf. She is 24 years old, she’s single, she lives in France and she has spent the last three years working in secretarial and accounting jobs. Her surname tells us that she’s descended from Senegalese immigrants, and her first name strongly suggests that she’s Muslim. Hundreds of employers across France will have seen Khadija’s name and none of them would have known the most important thing about her: she doesn’t exist.

Khadija is one of three fake women invented by Claire Adida from the University of California, San Diego. They are all part of a clever experiment that reveals how the French job market is rife with discrimination against Muslims. Adida found that in at least two sectors, a Muslim candidate is around 2.5 times less likely to get a job interview than a Christian one, with all else being equal. These results were backed up by a large survey, which showed that among second-generation Senegalese immigrants, Muslim households earn far less than Christian equivalents.

To some, this won’t come as a surprise, for modern Europe doesn’t exactly seem like a welcome place to be a Muslim. In France, the Senate voted almost unanimously to ban women from wearing Islamic face-veils in public, a move that drew widespread support from other countries. Swiss voters approved a constitutional amendment to ban the construction of new minarets on mosques. In the UK, ridiculous headlines in so-called newspapers regularly portray Muslims as an identically-minded group, out to distort ‘traditional’ values, kill people and generally cause trouble. Negative opinions are growing everywhere.

These examples of religious discrimination are obvious and blatant, but others – such as prejudice in the workplace – are harder to uncover. Adida did it by focusing on France’s Senegalese community, which includes a mix of both Muslims and Christians. To see how they would compare on the job market, Adida created three imaginary CVs. All were single, 24-year-old women, with two years of higher education and three years of experience in secretarial or accounting jobs. Only their names, and small details about past employers, differed.

Khadija Diouf had a well-known Muslim first name and an obvious Senegalese surname and had worked with Secours Islamique, a humanitarian organisation. Marie Diouf had worked for its counterpart Secours Catholique and had an obvious Christian first name. And Aurélie Ménard had a typical French name with no religious connotations and had only worked for secular firms.

In the spring of 2009, Adida collected ads for secretarial and accounting jobs from the French national employment agency and grouped them into pairs, matched for area, sector, company size and position. For each pair, both received Aurélie’s CV while one received Khadija’s and one received Marie’s.

The results were striking. Marie Diouf got a positive response on 21% of her applications; she was clearly an employable (if fictional) young woman. But Khadija Diouf – her exact equal in virtually every respect – got callbacks from just 8% of her applications. For every 100 interviews that Marie was called for, Khadija was summoned for just 38. Even after Adida included a photo on the applications (the same one, showing a woman who was clearly not North African), she found the same bias.


It was impossible to send both Marie and Khadija’s CVs (which were virtually identical) to the same recruiter, given that they were identical except for a few names. Alarm bells would have rung. Nonetheless, it’s clear that Adida matched pairs of recruiters as well as possible, given that Aurélie’s odds of getting interviews were the same no matter which employer read her CV.

Adida essentially did the closest possible thing to a randomised trial (the gold standard in medicine), by assigning identical ‘people’ to be either Christian or Muslim and seeing how they fare.  In her own words, “This experiment thus provides a clear indication that in at least one sector of the French labor market… there is significant religious discrimination.”

The strength of Adida’s experiment lies in isolating the effects of religion from the package of cultural traits that accompany it. That’s far from easy. In European countries, Muslim immigrants tend to come from the same place. In the UK, they largely hail from South Asia. In France, they mostly come from northern Africa.

If Muslims in these countries suffer socially and economically, religious discrimination is just one possible explanation. Others include racial or geographical discrimination (north Africa, for example, has a history of conflict against French imperialism), or differences in education, language or culture in one’s home country. An experiment that tries to look at religious discrimination needs to somehow hold all of those other factors equal. That’s exactly what Adida managed to do.

She also found that the discrimination that her fake applicants faced can directly affect the lives of real Muslims. She relied on a survey done in 2009 by David Laitin, looking at issues of integration among 511 second-generation Senegalese Muslims. Notably, the survey only looked at two ethnic groups – the Joolas and Serers – who have Christian and Muslim members in equal measure. Both groups arrived in France in the 1970s, so neither enjoyed an economic headstart, although the Christians were slightly better educated.

The survey’s data revealed that the Muslim households were significantly poorer than their Christian counterparts, even after adjusting for their initial educational advantage. They’re more likely to fall into poorer income groups and they make around 400 Euros less per month, around 15% of the average monthly salary in France.


If anything, this just scratches the surface of anti-Muslim discrimination in France. There is a common view in France that Senegalese Muslims aren’t “real” Muslims because they have know little Arabic and because they socialise with Africans of all religions. The level of discrimination faced by groups that are more clearly linked with Islam might be even greater. The big question now is why, and it’s something that Adida plans to explore in future studies, using interviews, psychological games and more.

For the moment, her current study stands as a rare treat among research into discrimination. Some research goes on in the lab and benefits from careful experiments, but it’s unclear if their results apply in the real world, with all its complexities and vagaries. On the other hand, real-world experiments often struggle to isolate a single factor like religion, from the many others than entangle it. Adida’s work – inspired by a classic experiment on racial discrimination in America – combines both approaches: a careful, real-world experiment that does its best to unveil the effect of religion, with all else being equal.

Reference: PNAS http://dx.doi.org/10.1073/pnas.1015550107

More on discrimination:

If the citation link isn’t working, read why here

Twitter.jpg Facebook.jpg Feed.jpg Book.jpg


Comments (22)

  1. interesting. one thing i’d like to add is that the pressures may be feedback loops which are from the outside (discrimination) as well as inside (values differences). the “inside” might be a function of what french muslims of senegalese origin define “muslim” to be in their new environment. a better example might be contrasting sikh punjabis and mirpuri pakistani muslims in the UK. both groups to my knowledge arrived without the high human capital levels of east african indians, but they’ve gone in somewhat different trajectories. both are roughly similar ethnically and from neighboring regions of south asia (also, there’s a contrast between the achievements of pakistani british men vs. women).

  2. steve

    It would be very interesting to see how they fared when submitting CV’s to companies in Muslim countries.

    Admittedly, I have my presuppositions.

    I suspect you could find similar results here in America with ‘white’ and ‘black’ sounding names, whether it was sent to predominantly white employers or not.

    A nice article, as usual, Mr. Yong.

  3. anonymoose

    Steve: There have been very similar studies done in the United States using “black” and “white” sounding names, and they also show that the “white” name is much more (nearly 50%) likely to get a job interview, and that the “white” candidate gets more interviews even when making the “black” candidate more qualified for the position. Roland Fryer and Sendhil Mullainathan have both done work on this topic.

  4. Indeed, if you look at the very last paragraph and click on the magic linky words that say “classic experiment on racial discrimination in America”… 😉

  5. HP

    Quibble: Senegal is geographically, ethnically, and culturally in West Africa, not North Africa.

    I only mention it because of this: “Adida included a photo on the applications (the same one, showing a woman who was clearly not North African).” I got all confused because someone who is typically Senegalese is also “clearly not North African.”

  6. Mustafa


    Why would it be interesting to see how this experiment would go in Muslim countries? I think most would agree that muslim countries in general don’t have very good human rights records,…but this illusion that the Muslims of Europe are getting a fair shake must stop. Studies like this help dispel that. Banning the face veil, banning the minarets, and chucking Muslim resumes in the trash are just different manifestations of the same thing.

  7. PL

    Permettez-moi de relever une imprécision dans votre article.

    Si le niqab (voile intégral) a été interdit en France dans les espaces publics (et non le voile) ce n’est pas au nom d’une hostilité à l’islam mais pour défendre la condition de la femme ainsi que l’islam “modéré” ou progressiste contre des formes extrêmes.

    Ce que l’étude que vous citez montre à l’égard du marché du travail est sans doute assez proche de la réalité et de la pratique de certains employeurs malheureusement.

    Il n’empêche que l’on puisse demeurer dans la complexité et ne pas tout “simplifier” dans le sac d’une posture anti-musulmane. La situation est bien plus complexe en France où des imams modérés sont menacés de mort et des femmes contraintes.

    Comme moi bon nombre de citoyens souhaitent l’intégration à la communauté française de tous les islams compatibles avec les règles républicaines et laïques.

    Bien cordialement

  8. DocFranglais

    In one lab I worked in France CVs were quite openly sorted according to postcode. Anyone who happened to live in the countryside or one of the less chic ‘banlieue’ away from one of the main French cities had no chance of being shortlisted. In another lab, after applying for a permanent contract for the temporary job I was already doing, my application was rejected. The jury’s verdict, I was ‘too English and thus had an unfair advantage’. My French GP on diagnosing my mild depression after a year on the dole gave a strange prescription – go to the local Prefecture and apply for French citizenship. Having French nationality would change the way employers perceived me, being an EU citizen holds no sway. The doctor spoke from experience – his wife is Korean. Time for a revolution, I’d say.

  9. “I was ‘too English and thus had an unfair advantage’.”

    Heh. Well we do have an innate awesomeness… 😉

  10. Sophie

    Take the same survey and place the nationality Ukrainian, Bulgarian, etc … and you will have similar results not only for France, but for any western European country. Of course, I am not claiming eastern Europe is even less racist, but at least they dont call themselves developed countries like others …
    There is no need to prove that Europe is racist. What Europe needs is more lessons into how to stop being so racist and to actually acknowledge they have a huge problem in this area.Unlike USA that actually works towards fighting racism, Europe pretty much likes it the way it is. After I made an application for a job within a company that claimed to be “its all about people”, they showed a lot of interest in my cv, the 2nd email I got from them was to ask what was my nationality, and the 3rd email from them was that the offer was no longer available after I told them my nationality. This is Europe.

  11. Doug

    I have to agree with PL. Banning the Niqab isn’t an example of anti-muslim sentiment. It’s just standing up for the basic human rights of women. It’s why I support anti-mask laws in the United States. Seperate but equal never worked… why woukd a society that cares about it’s citizens enable that kind of mysoginy?

  12. Linda

    Doug, there are plenty of Muslim women (here in the Netherlands, at least) who choose to wear a niqab or other form of veil of their own free will, as an expression of their devotion and both cultural and religious identity. Banning veils and headscarves in public is not going to liberate the women who are actually suppressed and forced to wear them by their husbands and fathers – they will merely be unable to leave their houses from now on.

  13. “Banning veils and headscarves in public is not going to liberate the women who are actually suppressed…”

    Ding ding ding. You don’t stand up for basic human rights by removing a person’s choices. You do it by working towards an environment where a real choice is possible.

  14. Brian Too

    Random thought.

    I’m struck by the expressions of dismay (only talking North America here) by immigrants over discrimination. Well, sure.

    However this ignores history and humanity. As far as I can tell, immigrants of ALL stripes, with the possible sole exception of English immigrants, were discriminated against. Take a look at just a partial list off the top of my head: Chinese, Irish, Portuguese, Ukrainian, Polish, Japanese, Puerto Rican, Finns, Swedes, Africans (sorry for the amalgamating), Germans, Italians, Czechs, Vietnamese, Philippine, the list goes on and on.

    It’s often easier and more convenient to reject a stranger, and lump them in a category of ‘different and inferior’, than it is to open your heart and mind to them. Especially if there is any plausible category to put them in.

    What I’m saying (asking really) is, is this a passing phase in European history? I suspect I know the answer but we have to be careful not to assume that today’s circumstance is tomorrow’s destiny.

  15. And thus, Mary Anne Evans published as George Eliot, so she COULD get published.

    As Brian Too says, discrimination against the latest wave has always been the case — they’re too different, they’re not one of us, etc.

    The difference between then and now is that in my grandparents’ day coming to the US meant they assimilated by making their kids learn English and tried desperately to learn it themselves, and when someone asked what their home country was, they looked askance at the questioner and said, “America!”

    My father in law grew up speaking German as his first language in Ohio, but flew B-17s bombing Germany during WWII. I asked him once if that bothered him, dropping bombs on Germans when he had grown up in a German family. He looked at me like I was crazy and said, “Hell no, I’m an American!”

    I absolutely do not defend a firm that rejects applicants based on criteria that the firm is under the law to not consider. I also absolutely defend the right of people to NOT assimilate in a new environment. However, I also believe that with every choice we make, we also must accept the consequences and responsibilities. We can yell about not being treated fairly, but we must accept that sometimes, we bring things on ourselves by digging in our heels.

    No easy answers…by the way, does the French Government have the equivalent of the US Equal Opportunity Commission (with enforcement powers via the Justice Dept)?

  16. Swift Loris

    In the fifth paragraph, why is there a comma after each of the names? There’s absolutely no grammatical reason for them to be there. Did you do that, Ed, or was it an editor? Whoever’s doing it, stop!

    [I do it specifically to annoy you – E ;-p]

  17. Daniel J. Andrews

    I did a resume experiment a while back. I submitted the same resume to a large number of government jobs but using a female name for one version, and a male name for the other version. Even I was stunned by the result which I expected. I received one job interview offer for the male resume, and 7 for the female name. That, I think, is the result of our government’s policy on hiring called “Equal Opportunity”, which appears to be anything but.

  18. Hasan

    Gentlemen – Very discouraging! Muslims in France need to show their value based upon their major accomplishments. They need to showcase their expertise and if possible quantify their major accomplishments supported by facts and figures.

  19. Thank you for reporting on our PNAS article – however, I’d like to specify that this was collaborative work with David Laitin (Stanford University) and Marie-Anne Valfort (Université Paris I Panthéon Sorbonne). The original identification strategy was Professor Laitin’s, and the article was co-authored by the three of us. Thanks again.

  20. Awesome research. Was that published in France too?

  21. jacky

    Well that’s why I left France they are too rascist you cannot found houses jobs .

  22. Rien en Islam n’oblige les femmes de se couvrir le visage et les mains complètement . Bien au contraire dès qu’une personne a choisi de se convertir à la religion musulmane ils ne peut intervenir sur ce que Dieu et son messager ont ordonné ou recommandé.


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Not Exactly Rocket Science

Dive into the awe-inspiring, beautiful and quirky world of science news with award-winning writer Ed Yong. No previous experience required.

See More

Collapse bottom bar