I have become involved in research on Female Genital Cutting (FGC) in Nigeria, as part of my work, and in preparation for this work I had to read a report on the prevalence and distribution of the practice in Nigeria. This report makes for interesting reading, and in particular it seems to run counter to a lot of the stereotypes and propaganda being put out by various individuals, organizations and movements on this much-discussed topic. Some of the findings in the report I read really contradict what I think is the prevailing wisdom, and so I thought I’d give a little summary and discussion of the contents of the report here. None of the work presented here is my or my department’s research, and I don’t intend to say anything about our research because it’s not published, and I don’t want to present anything in the public domain in any way shape or form until it’s been published officially. This post then is my review of someone else’s research report, with a few opinions and some musing that are entirely my own. Also, I’ll use the term “Female Genital Cutting” (FGC) here rather than “Female Genital Mutilation” (FGM) because that’s the term the report uses, and (for reasons I think we’ll see) it’s probably a more accurate term.

The report I’m talking about here is Chapter 18 of the 2008 Demography and Health Survey of Nigeria (more information on that below), which can be downloaded here.

Female Genital Cutting

FGC is the practice of cutting into or excising parts of the female genitals, or (in its most extreme form) closing up part of the female genitals, with or without the removal of some flesh. The most severe forms of FGC, infibulation, can have potentially life-threatening and/or (usually “and” I think) fertility-impeding results. It’s not clear that all forms of FGC are harmful, though I am told there is a lot of debate about the health consequences of the most minor examples of the practice. The term FGC covers a much wider range of practices than male circumcision (which is also, obviously a form of cutting); from small ritual scars, removal of flesh (e.g. the clitoral hood); excision of the clitoris; and infibulation. I don’t think there’s any debate about the health effects of the more extreme forms of FGC, which are generally accepted to cause loss of sexual arousal and function, physical harm, trauma, sometimes difficulty with basic functions, and extreme difficulty in childbirth, with increased risk of maternal and infant mortality.

In the popular press over the past 10 years, FGC has been associated with Islam, and is presented as a common, even ubiquitous practice amongst muslim communities through north Africa and the middle east. I think you can probably find people claiming it is prevalent all the way across to Indonesia, though I doubt the people you’ll find saying such things are credible, and they probably also have published blog posts on how the unmarked helicopters will soon come for your carbon. Many muslim activists and feminists have pointed out that FGC is prevalent in non-muslim communities in Africa, and that it comes in different forms and isn’t necessarily popular with men or women in the countries where it is practiced. It is also a highly controversial topic in western political debate, often presented as a pressing example of why the west needs to “act” to instill our “civilized” values in these countries; discussion of FGC is entangled with the complex of post-colonial and post-9/11 thought that delivered us the “Arab Exception” (Arab countries can’t support democracy), “liberal intervention” (bombing people can be a viable way to introduce democracy) and, to my mind most odious of all, claims to be defenders of women’s rights from the neo-cons who orchestrated the slaughter of a million people in Iraq.

Of course, discourse that is tangled up with these issues inevitably tends to avoid inconvenient things like “facts,” or tedious tasks like “assaying the available evidence.” I don’t intend to stray from this time-honoured tradition by giving a complete review of the literature on FGC. But I do want to talk about the facts as they have been gathered in this DHS Report. Consider it my small contribution to 0% of fuck all on the topic.


Nigeria is Africa’s most populous country, and is characterized by a wide diversity of geographical regions and a large number of different tribal and ethnic groupings. It was ruled under a dictatorship until about 1999, and is a member of the Commonwealth. Like a lot of countries it is home to a lot of different religions, but primarily Christian and Muslim in a roughly 50-50 split. Nigeria’s first census was in 1866 and its current population 140 million. It has significant oil reserves and the exploitation of these reserves by Shell (and other companies) has been a matter of some controversy, with accusations of both reckless behaviour and a low level conflict between residents of the oil-drilling areas and the central government. I think there are areas in the North of Nigeria where some sort of tribal and/or Sharia law applies. Nigeria is an ex-British colony, which presumably means it has a cricket and a rugby team. The wikipedia run-down on Nigeria is probably all you need to get an overview of a diverse and interesting country still struggling with its post-colonial problems and the (sadly typical) legacy of a post-colonial military dictatorship. Like most developing nations in Africa, it is rapidly reforming and improving its health sector to try and meet the Millenium Development Goals, focussing on those connected to population planning (i.e. maternal and infant mortality). Nigeria’s HIV prevalence is low compared to some African countries – about 3.6% – but still punishingly high by international standards, and primarily transmitted heterosexually.

The Demography and Health Surveys

The Demography and Health Surveys (DHS) are a standard population survey funded and developed by USAID and implemented in a variety of developing nations on a roughly 5-yearly basis by local statistics offices in conjunction with USAID and sometimes the UN (in Nigeria, the UN Population Fund has been involved). The surveys are conducted nationally and are intended to provide a representative sample of the population. They focus on population-planning and women’s health issues, especially fertility and infant mortality, because these issues are the crucible in which health development happens. Some surveys include sub-samples for examination of specific topics, including fingerprick samples to estimate HIV prevalence. In many instances these surveys are the only way in which developing nations are able to get a snapshot of key health problems like infant mortality, due to poorly-functioning central population registers, or limited funds for aggregation of data.

The DHS are implemented in a door-to-door interview style on a random sampling basis that is analysed in a manner similar to the analysis of  national household surveys in the west – for example Australia runs a National Drug Strategy Household Survey every 5 years along similar lines. The data is then analysed and published in reports on the DHS website, and also made available free of charge upon application to interested researchers (e.g. Yours Truly). They provide a wealth of opportunities for research into development and health issues in the developing world. The 2008 Nigerian DHS sampled about 33,000 women and 16000 men, which is a huge sample by anyone’s standards, and an excellent opportunity to ask all sorts of interesting questions of Nigerian women. Note that the DHS surveys usually only concern themselves with women aged 15-49, and that is what “women” in the remainder of this blog post should be taken to mean.

FGC in Nigeria

So, chapter 18 of the report first gives us the rather surprising result that only 61% of women aged 15-49 have ever heard of FGC in Nigeria. Assuming that those who have never heard of it have never experienced it, they then calculate a stark figure of 30% of all Nigerian women having experienced some form of FGC. This figure doesn’t seem to be afflicted by missing values, but we can see from Table 18.1 that a mighty 45% of all women who have had FGC do not report the type they have received. The DHS divides FGC into three categories, and 45% of women who had experienced FGC classed themselves in the “cut, flesh removed” category. What this means is not clear and inspections were not done, so whether these women were mostly seriously affected or suffered nothing more sinister than a female equivalent of male circumcision is not known from this data. Eighty percent of women had been cut before their first birthday, suggesting that this practice in Nigeria is not a coming of age ceremony or adulthood rite, but something committed largely post childbirth. The majority of cutting was administered by a traditional circumciser or traditional birth attendant, but 2% were performed by doctors.

Some of the demographics of FGC are show in Table 18.1, and here we start to see our western notions of the causes and reasons for FGC running into the brick wall of reality. FGC is much more prevalent in urban rather than rural areas, and its prevalence is proportionate to wealth. That is, the richer you are, the more likely you are to have experienced FGC. My original image of FGC was that it was a phenomenon of poor, rural women but the opposite appears to be the case. We’ll get on to discussing confounders for this conclusion in a moment, but let us first consider one other remarkable fact – the more educated a Nigerian woman is the more likely she is to have experienced FGC. Rather than being a phenomenon of poor, ignorant rural women is it actually a fad of the urban upper class? Perhaps a cohort effect? There does appear to be a linear relationship between age and circumcision and, since circumcision occurred mostly before age 1 this linear relationship does suggest a cohort effect.

The problem we have with drawing strong conclusions about the demographics of FGC from this report is that there is no multiple regression model. We note that FGC is more common in urban areas, but we can also see that it is much more common in the South than the North; and Nigeria’s cities are all in the South. If a tribe is distributed unevenly across the country and has a strong history of FGC, then it will confound conclusions about the demographic associations. It could also be that there is some historical effect, perhaps because a tribe that controlled access to education and wealth during the dictatorship also had a high prevalence of FGC. Nonetheless, the closest you can come to an iron law of development and health is that education improves women’s rights, so it’s interesting that FGC is more prevalent amongst educated women. However, in the absence of a multiple regression model, conclusions about the meaning of these demographics are weak.

Religion and Tribal Effects on the Prevalence of FGC

Looking at Table 18.1 again, we note that FGC is much more common in the South than the North – overall prevalence in the North is about 13% vs. about 35-40% in the South. It’s almost non-existent in the North East. These, incidentally, happen to be the areas of Nigeria with the largest Muslim populations, while the South has the highest prevalence of Christians. Unless there is a seriously skewed distribution of FGC between Muslims and non-Muslims, the striking conclusion of that result is that Muslim women appear to be much less likely to have experienced FGC than non-Muslim women. In fact some areas of the North the majority of women don’t even know what FGC is. One noteworthy point though is that Northern circumcision is more likely to be reported as the most severe kind (sewn closed); but also the lowest rates of refusal to report the type of FGC are in the North. Of course, the broad categories of definition and the large proportion of missing data make this information almost meaningless.

The authors wisely chose to eschew reporting on religion and its relationship to FGC, which is a no-win game for all concerned. If you find high levels of FGC amongst Christians, then right-wing cultural warriors will crucify you for attacking religion; if you don’t report them, right-wing cultural warriors will accuse you of covering up the true prevalence amongst Muslims. The best thing to do is to try strenuously to avoid your health research being used to confirm or disconfirm other people’s biases. Our concern as health researchers is behaviour, not categories of people, and categories of people are only of interest in so much as they form useful markers for behaviour, or useful media for changing behaviour; this is why HIV was renamed from GRID (Gay-related Immune Disorder) when the behaviour that caused its transmission was identified – the category was no longer analytically useful. Categories like “gay” or “Muslim” (or indeed, I suppose, “gay Muslim”) may be useful as health promotion tools (you can, for example, disseminate health promotion messages in gay mosques!) but as analytic categories they are always superceded by direct research into behaviour. Especially when, as in this case, “Muslim” as a category is so general that it spans multiple regions, levels of wealth and education, and tribe.

Next we can also notice a large variation in rates between tribes. I am under the impression that most Nigerian tribes include members of several religions, though I could be wrong. I can’t find evidence online for the proportion of the various tribes that are Islamic and I don’t trust Wikipedia on this (the source is given as the “World Christian Database” and some of the “tribes” listed there don’t seem to match those in the DHS), but if we assume that the three biggest figures given there are correct in at least their ranking, we can see that the three tribes with the largest Muslim populations (Yoruba, Hausa, Fulani) have the two lowest rates of FGC (and one has the highest). Again there is huge confounding here, since the Yoruba tribe is mostly in the Southwest (a hotbed of FGC) and also has a heavy mix of religions – Wikipedia puts the Muslim population at 8 million, but the overall Yoruba population in Nigeria at 29 million. Similarly sketchy Wikipedia accounts suggest the Fulani are primarily Muslim, and they have a very low prevalence of FGC (8.5%). So, we can do ecological analysis wikipedia style[1]: a tribe with a minority population of Muslims has a very high FGC rate, while a tribe with a majority population of Muslims, credited by wikipedia with spreading islam to Nigeria, has a very low FGC rate.

Shall we join the dots? There is weak evidence from this DHS report, relying on an ecological-level analysis of population prevalence of FGC, that being Muslim is protective against FGC, and FGC is most prevalent amongst Christians. The only way to know if there is any substance to this conclusion is to do a multiple regression at the individual level, adjusting for wealth, education and location, to see whether there is any association between religion and FGC. My guess is that when one does this one will find that FGC is a cross-religious phenomenon, driven primarily by class- and tribe-specific cultural factors that transcend religion.

Opinions on the Continuation of FGC in Nigeria

Men and women were asked in this survey what they thought the reasons for FGC were, and whether they would submit their own daughters to it. Responses to the former question for men and women seem broadly similar: there are no benefits to FGC is the commonest response for both sexes (Tables 18.6.1 and 18.6.2). The second most common response (and it’s a distant second overall) was “to prevent pre-marital sex and preserve viginity” and this was heavily over-represented in the South, where the practice is most common. Seven percent of men indicated it was done to improve sexual pleasure for men, vs. 4 % of women. This set of results hardly supports the view from abroad of populations of benighted savages, ignorant of the true consequences of FGC and persisting in silly notions about its benefits. This foreign view of FGC is further dispelled by women’s reports of whether they aim to subject their own daughters to circumcision: only 6% of circumcised women intend to do this, and the proportion of women intending to do it is lowest amongst the highest-educated. That is, where it is most prevalent it is least popular. It is also least popular with the youngest women, who are most in control of girl children’s fate.

I think there is a well-established counter current of feminist thought in the West – so well established that I can’t even be bothered googling it – that women play a significant role in decisions about whether to circumcise daughters, as part of the generally well-established role that women play in policing the boundaries of femininity. This idea is well presented in the inscription at the front of Alice Walker’s book on FGC, in which the trees in the forest notice that the woodsman’s axe has a wooden haft. I think it’s pretty likely that the traditional birth attendants and circumcisers responsible for cutting the majority of women in this sample are also mostly women. So at the point where women decide it’s not happening to their daughters then it probably doesn’t take much in the way of empowerment of those women’s decisions to see their will enacted.

Interestingly, the DHS includes a lot of questions about the household’s attitudes towards women’s equality, couched in practical terms (“he lets me make purchasing decisions”) and ideological questions (“it’s okay to beat a woman if she burns dinner”). So it is possible that some enterprising social scientist could analyse attitudes to the continuation of FGC and work out exactly what is required to empower women’s decisions to make it go away. However, I suspect the most likely answer will be “nothing.” Nigerian women appear to have decided they don’t want the next generation of girls to have FGC, and my guess is that will be sufficient for the practice to die out by itself within 2 generations.


This report presents weak evidence against two commonly-held stereotypes of FGC:

  • It is a predominantly Muslim practice
  • It is a problem of poor, uneducated rural women

at least in Nigeria. It also gives weak evidence in favour of the possibility that FGC is much more common in Christian communities; or at least that being Muslim is protective against FGC. These conclusions cannot be drawn definitively without individual-level analysis of attitudes and practices in a proper multiple regression, that adjusts effects of religion and tribe for region, personal beliefs, wealth and education. To do such analyses is not the responsibility of health researchers, who should be avoiding feeding the fires of this kind of debate (in my opinion) and focussing on the practical aspects of the issue. Maybe. It’s also possible that there is a cohort effect in the prevalence of FGC in Nigeria, in that it may have been widely practiced 20 or more years ago and/or amongst a particular tribe or group of tribes.

This report also suggests that FGC is going to disappear rapidly in Nigeria through natural attrition, and that the single cheapest, easiest way to make this happen is to empower women to have full control of the perinatal process. When women can choose the timing of childbirth and have access to good medical care, and are able to assert their own authority over the child-rearing process and associated decisions, the results in this report suggest that they will choose overwhelmingly against FGC for their own daughters. The report also suggests that the majority of men don’t have strong views on the importance of FGC, and so it is unlikely to continue even if western feminism does nothing about it. As is usual in development health, the simplest way to reduce the dangers to women’s health and to improve children’s health is to enable women’s self-empowerment. But in this case it is not clear that increasing girls’ access to education will protect them against the phenomenon.

This really is a topic crying out for detailed sociological research and I hope it’s being done!

A side point on the analysis of hysterical political campaigns: as usual, the right wing shock-jocks are wrong, and my standard method for making a first-pass judgement about a political problem worked. This method is: if group A hates group B, and group A is claiming that group B do nasty behaviour x (child-eating, FGC) then probably the opposite is true. In this case group A is right wing shock jocks / neocons (and some feminists); group B is that most nefarious and broad of categories, “Muslims” and my conclusion is, as ever, correct in all of its particulars.

fn1: i.e. get it completely wrong!