Methodology and data sources

The analysis presented in these pages is mainly based on household survey data from Demographic and Health Surveys (DHS) and Multiple Indicator Cluster Surveys (MICS). The respondents are women aged 15–49, except in Egypt, northern Sudan and Yemen, where the sample of respondents includes only women who have been married.

These surveys focus on two types of prevalence indicators. The first addresses FGM/C prevalence levels among women and represents the proportion of women aged 15–49 who have undergone FGM/C. The second type of indicator measures the status of daughters and calculates the proportion of women aged 15–49 with at least one daughter who has undergone genital mutilation or cutting.

So far, the FGM/C modules have been implemented in 29 countries:
 
•  Benin (DHS 2001, DHS 2006)
•  Burkina Faso (DHS 1998–1999, DHS 2003, MICS 2006)
•  Cameroon (DHS 2004)
•  Central African Republic (DHS 1994–1995, MICS 2000, MICS 2006)
•  Chad (MICS 2000, DHS 2004)
•  Côte d'Ivoire (DHS 1994, DHS 1998–1999, MICS 2006)
•  Djibouti (MICS 2006)
•  Egypt (DHS 1995, DHS 2000, DHS 2003, DHS 2005, DHS 2008)
•  Eritrea (DHS 1995, DHS 2002)
•  Ethiopia (DHS 2000, DHS 2005)
•  Gambia (MICS 2005-2006)
•  Ghana (DHS 2003, MICS 2006)
•  Guinea (DHS 1999, DHS 2005)
•  Guinea-Bissau (MICS 2006)
•  Kenya (DHS 1998, DHS 2003)
•  Liberia (DHS 2007)
•  Mali (DHS 1995–1996, DHS 2001, DHS 2006)
•  Mauritania (DHS 2000–2001, MICS 2007)
•  Niger (DHS 1998, DHS 2006)
•  Nigeria (DHS 1999, DHS 2003, MICS 2008)
•  Sierra Leone (MICS 2005, DHS 2008)
•  Somalia (MICS 2005-06)
•  Senegal (DHS 2005)
•  Sudan (DHS 1989–1990, MICS 2000, Sudan Household Health Survey 2006)
•  Tanzania, United Republic of (DHS 1996, DHS 2004–2005)
•  Togo (MICS 2006)
•  Uganda (DHS 2006)
•  Yemen (DHS 1997)
•  Zambia (Zambia Sexual Behaviours Survey 1998, 2000, 2003, 2005)

Data on FGM/C inform policymakers of critically important variables in an effort to better understand the practice and develop policies for its abandonment, but they must be analysed in light of the extremely delicate and often sensitive nature of the topic.

Under-reporting is often considered a possibility in practising communities situated in countries that have legislation outlawing it. While numbers of FGM/C-related prosecutions are few, the threat can be a significant deterrent to admitting to the practice. To overcome some of the risks of under-reporting, some modifications in the questionnaires used to collect information on FGM/C have been introduced. In the case of Sierra Leone (MICS 2005) and Liberia (DHS 2007), survey questions were adjusted to eliminate the direct reference to FGM/C, and rather reference entrance into women-only societies, such as the Bondo in Sierra Leone and the Sande in Liberia. Questions were rewritten to ask specifically if the woman has been initiated to the society and who initiated her. Initiation to the society occurs when a women is cut. Although it is no different from the FGM/C experience in most other countries, initiation to the society holds broader implications for the respect and status of women in the community.

In addition, the practice may often occur during early childhood, so memories of age or type of practitioner may be obscured. The issue is further complicated when there are multiple procedures: one incident of the initial infibulation and another when the woman is unsealed after marriage. While survey methodologies and analyses attempt to account for these elements, the reader should approach the data in this report with a reasonable degree of prudence.

Examining trends in the prevalence of FGM/C is now possible given unprecedented data availability over the last decade. This said, challenges arise when examining trends, particularly when establishing a connection between programmatic activities and changes in prevalence levels over time. Evidence of change over time in the practice of FGM/C can be obtained by comparing prevalence among women aged 15–49 by age cohort. This comparison, however, may not capture recent changes, due to the length of the retrospective periods involved.

For instance, in the case of a country where most of the girls are cut below the age of 1, even the youngest respondent (in the age cohort of 15–19) would report on an event that occurred 14 years before. Any change that occurred after this period will therefore not be reflected in the data. Another technique consists in comparing prevalence among mothers and daughters. This analysis can be partially skewed for a number of reasons, namely a mother’s reluctance to discuss her choices related to FGM/C and the age at which girls are typically cut. If, for example, the characteristics of the practice in a certain area results in girls undergoing FGM/C closer to adolescence, this will result in low rates of mothers (particularly young mothers) with at least one daughter cut. These girls, however, are still at risk of being circumcised in the near future.