January 2004, Vol. 10, No. 1

Female Genital Cutting
Potential Data Biases in Studies Exposed

Female genital cutting (FGC), also known as female circumcision and female genital mutilation, has been the topic of much research and attention in the past ten years. Investigators seek to learn the prevalence of the procedure in a certain area, to ascertain the determinants of the practice, and to verify the effects of interventions aimed at reducing the incidence of the custom. Findings from a recent study conducted by the Population Council, however, suggest that results of such investigations should be interpreted with caution. Women interviewed about FGC may falsely deny or falsely confirm having undergone the procedure.

Navrongo Health Research Centre
The study is part of an experimental investigation in health services delivery launched by the Navrongo Health Research Centre in the Kassena-Nankana District of northeastern Ghana. The Centre’s core scientific resource is the Navrongo Demographic Surveillance System. Using this system, developed in part by Population Council demographer James F. Phillips, investigators have recorded demographic events among all residents of the area since 1993.

Yearly surveys collect more detailed information from a subset of women aged 15–49. In 1995 and 2000, for example, among the questions the survey asked women was, “Are you circumcised?” Almost 2,400 of the 5,275 women surveyed in 1995 answered the question again in 2000. Because of high rates of migration, fewer than half the women interviewed in 1995 were re-interviewed in 2000.

The responses from women who participated in both surveys were analyzed and the respondents were placed into four categories. In the first two categories were women whose survey answers in the two years agreed. In both surveys they reported either that they had or had not undergone circumcision. In the other two categories were women whose two survey answers disagreed. Some of these women said in the first survey that they had not been circumcised, but in the second survey said they had been. The researchers assumed that the women underwent circumcision in the time period between the two surveys. The rest of the women reported in the first survey that they had been circumcised, but subsequently said they had not been—an impossible sequence of events.

A relatively high rate of response reversal was not completely unexpected. In the past, nearly all women in the Kassena-Nankana area were circumcised, and women who avoided the procedure were stigmatized. In 1994, however, the government of Ghana passed a law banning female genital cutting, and in 1996 a circumciser in the Kassena-Nankana area was jailed. In the past several years there has been a notable decrease in the stigma attached to women who are not circumcised.

There may be more than one explanation for response reversal. Women may be falsely denying circumcision. Or they may have incorrectly reported circumcision in the first survey because they felt pressured to give what they thought was a socially appropriate response. Because being uncircumcised had become less stigmatized by 2000, these women may have felt more comfortable reporting their true status.

The gold standard for settling this issue would be to conduct physical examinations of survey respondents. This method has drawbacks, however. Medical examinations require specially trained personnel and thus are costly. Moreover, many women may not agree to being examined, producing a biased sample.

Data analysis
The researchers analyzed the survey data to compare the marital status, education, and other characteristics of women who fell into each category. They found that women who in both surveys reported being uncircumcised had the highest education rates of the women surveyed. They were also more likely than other women to have used family planning and less likely to practice a traditional religion or be in a polygamous marriage. Women who reported being circumcised in the first survey but later denied it appear to occupy a middle ground. They are more likely to be educated and less likely to practice traditional religion than are women who reported circumcision in both surveys. And they are less likely to be educated and more likely to practice traditional religion than are women who never reported being circumcised.

Besides biasing data about the incidence of circumcision, the false reporting of circumcision status can influence conclusions drawn about the determinants of the practice. “If we had access only to the survey answers from 2000, we would have underestimated the influence of education, traditional religion, and marital status on circumcision,” says lead researcher Elizabeth F. Jackson, a Population Council staff associate at the time of the study. The researchers conclude that investigators should use caution when interpreting data on female genital cutting that are collected at a single point in time. In such studies, determining the rate of false reporting of circumcision status is more difficult than in other studies.

Source
Jackson, Elizabeth F., Patricia Akweongo, Evelyn Sakeah, Abraham Hodgson, Rofina Asuru, and James F. Phillips. 2003. “Inconsistent reporting of female genital cutting status in northern Ghana: Explanatory factors and analytical consequences,” Studies in Family Planning 34(3): 200–210. (PDF) This study also appeared as "Women's denial of having experienced female genital cutting in northern Ghana: Explanatory factors and consequences for analysis of survey data," Policy Research Division Working Paper no. 178. New York: Population Council. (PDF)

Outside funding
United States Agency for International Development

(Return to issue contents)

See Also

• Female Genital Cutting