Daughter circumcision and maternal life satisfaction: a cultural moderation effect revealed across two multicountry studies

Abstract

Introduction

The present paper assessed the relationship between maternal life satisfaction (MLS) and the intergenerational transmission of female genital cutting (FGC, female circumcision). It was hypothesised that the association would be more strongly positive in countries in which FGC is more prevalent (ie, culturally normative), suggesting a practice that is socially reinforcing within sociocultural contexts in which it is common.

Methods

Across two studies with more than 85 000 participants in 15 African and Asian countries, mothers completed surveys reporting on their own FGC experiences and those of their daughters’ and on their educational history and socioeconomic status.

Results

The association between MLS and daughter circumcision was weak but positive for the full sample. Contrary to predictions, in countries in which FGC is uncommon, it was more positively associated with MLS, and in countries in which it is common, it was weakly or negatively associated with MLS.

Conclusion

Results are contrary to the notion that the intergenerational transmission of FGC is a function of happiness deriving from its cultural normativity. They suggest, instead, a diversity of social motives depending on cultural context. Customised messaging to reduce the intergenerational transmission of FGC is discussed.

WHAT IS ALREADY KNOWN ABOUT THIS TOPIC

• Female genital cutting (FGC) has negative implications for health and gender equality and is practised to different degrees in countries in parts of Africa, Asia, the Middle East and their diasporas.

WHAT THIS STUDY ADDS

• This study assessed the intergenerational transmission of FGC—in particular, how daughter FGC relates to maternal life satisfaction (MLS). Contrary to expectations, life satisfaction ratings were higher for mothers of circumcised daughters, although this relationship was moderated by country-level FGC frequency. In countries in which it is more common, daughter FGC is less strongly or is negatively associated with MLS.

HOW MIGHT THIS STUDY AFFECT RESEARCH, PRACTICE OR POLICY

• Study findings may inform anti-FGC messaging. In countries in which FGC is common, messaging should highlight its association with relative maternal dissatisfaction. In countries in which it is uncommon, messaging should highlight alternatives to FGC as an expression of cultural identity.

Introduction

Female genital cutting (FGC), also known as female genital mutilation or female circumcision, involves the laceration of genital anatomy or the stitching of the vaginal orifice.¹ The practice exists in parts of Africa, the Middle East and Southeast Asia, usually undertaken in the context of religious or coming-of-age rituals with the aim of signalling or facilitating chastity, purity and emasculated femininity.^{2 3} It is most common in the rural and traditional areas of countries within which it is practised.⁴

In 2012, the United Nations (UN) General Assembly adopted a resolution to ban this practice.^{5 6} The rationale is as follows: (1) it is unsound from a reproductive health perspective, and (2) it is a form of discrimination against women and girls contributing to gender inequality and constricted social and economic opportunities for them.² Despite efforts to eradicate it, the practice persists to the present day, although with somewhat decreasing frequency over time (there is concern that the UN resolution against FGC constitutes cultural imperialism⁷).⁸

Social and religious norms are thought to be the basis of FGC.2 3 9 10 Consistent with this, the analysis of data collected by the UN International Child’s Emergency Fund (UNICEF) across countries in which FGC is practised revealed that the likelihood of a girl being cut is higher if her mother was cut. In addition, the likelihood of a woman or girl being cut increases with the proportion of cut women in her community.⁴ These findings suggest that the practice reflects community-level normative behaviour that is intergenerationally transmitted—perhaps because the practice is socially reinforced and therefore rewarding within the sociocultural contexts in which it is common.

While FGC has been investigated with respect to the influence of family practices and cultural norms, psychological factors have less often been investigated. If the practice is rewarding to those who condone or engage in it, one might expect it to be associated with life satisfaction or happiness.¹¹ That is, perhaps the motivation for individual families to engage in the practice with their daughters is a psychological one—it is consistent with increased happiness or life satisfaction of significant others. If maternal happiness derives from alignment or compliance with social or cultural norms, then FGC should be most strongly associated with maternal happiness or life satisfaction in countries in which the practice is more common.¹² This possibility is evaluated in the present paper. We explore whether maternal life satisfaction (MLS) is associated with the intergenerational transmission of FGC (as reflected by daughter circumcision) and whether that relationship is moderated by country-level FGC. The hypothesis underlying the present study was as follows: we expected that the association between MLS and daughter circumcision would be more strongly positive in countries in which FGC was more prevalent. Support for the hypothesis would suggest that the intergenerational transmission of FGC is perhaps a function of maternal happiness deriving from its status as a culturally normative practice. If the hypothesis were not supported, then the impact of country-level normativity on the intergenerational transmission of FGC would seem to be independent of MLS. Study results will inform our understanding of the role of MLS, country-level norms and their interaction on the intergenerational transmission of FGC.

Testing this hypothesis requires controlling for the influence of several possible confounds. Such confounds include individual-level and country-level variables that may themselves be associated with MLS or daughter FGC. These include mother education level, household wealth and country-level wealth. For instance, the proposed moderating influence of country-level FGC may simply reflect differences in country-level wealth. Including country-level wealth as a covariate predictor within relevant statistical models would allow for examining the effect of country-level FGC on the association between daughter FGC and MLS that is unique from that of country-level wealth. At the within-family level, it is necessary to similarly control for mother education and household wealth, both of which are likely associated with a cultural practice such as daughter FGC and with MLS. Finally, mother FGC is also included as a covariate predictor, thereby allowing for an estimate of the association between MLS and the intergenerational transmission of FGC, controlling for mother’s own personal experience of FGC.

Methods

Study design and data source

The present study used data collected by UNICEF as part of its Multiple Indicator Cluster Surveys (MICS) project.¹³ MICS is the largest source of internationally comparable data on children and women worldwide with data collected in 120 countries, across 28 years, with 7 distinct rounds of data collection.¹³ We used data from MICS 5 and MICS 6 collected, respectively, between 2013–2016 and 2017–2022.

MICS data are deidentified, publicly available and organised by country.¹³ Data for the present study existed in two separate files: a file labelled ‘FG’ contained daughter circumcision data, and all other study variables were contained in the ‘WM’ file. Data preparation involved merging variables across these two files for each country and then merging cases across countries.¹⁴ In this way, final datasets were generated for MICS 5 and MICS 6, analysed as Study 1 and Study 2, respectively.

Study participants

Data for the present investigation included responses from adult women who self-identified as mothers, in countries for which daughter FGC data were available. Data were collected by trained interviewers in a private or semiprivate one-to-one interview setting.¹⁴

Participants were female offspring of women interviewed as part of the MICS data collection effort for whom valid responses existed for daughter FGC. Participation was further limited to those offspring for whom valid responses existed for the dependent variable of the present study, MLS. Table 1 presents frequencies for these two variables for each country and the total sample for Study 1 and Study 2. Also presented is the number of cases for which valid responses existed for both variables; those cases comprise the final sample for the present study. Additional columns in table 1 provide information on the percentage of participants who experienced FGC and country level per capita gross domestic product (GDP). Additional demographic characteristics were as follows: average age of mothers, 30–34 years; average number of daughters per mother, Study 1, 2.61 (1.37), and Study 2, 2.55 (2.51); and daughter average age at circumcision, Study 1, 1.58 years (2.45), and Study 2, 2.14 years (3.07).

Table 1.

Per country participant information

Participating country	FG cases*	MLS cases†	Final sample‡	Daughter FGC frequency§	GDP per capita
Study 1: MICS 5
Benin	9722	1073	1059	0.3%	$4056
Cote d’ Ivorie	19 840	2676	2676	12.9%	$6538
Guinea	8938	10 397	1257	45.4%	$3187
Guinea-Bassau	8742	1342	1342	28.8%	$2190
Kenya	2245	288	288	<0.1%	$5704
Mali	15 205	2297	2253	77.4%	$2517
Mauritania	12 855	1458	1457	51.2%	$6424
Nigeria	17 078	207	207	23.9%	$5860
Senegal	5660	384	383	2.5%	$9209
Total (average)	100 285	10 987	10 922	(11.5%)	($5,076)
Study 2: MICS 6
African Rep.	15 274	19 074	15 234	1.8%	$967
Gambia	12 417	16 357	12 411	36.7%	$2510
Ghana	7357	10 840	7357	0.4%	$6498
Guinea-Bissau	8536	11 142	8534	19.9%	$2190
Iraq	9090	3046	2086	0.6%	$10 862
Nigeria	17 772	24 825	17 767	8.8%	$5860
Sierra Leon	13 541	18 197	13 532	6.4%	$1931
Togo	4510	6045	4509	0.3%	$2608
Total (average)	88 497	109 526	81 430	(9.4%)	($4178)

*FG cases = cases for which a valid response existed to the ‘daughter circumcised’ question.

†MLS cases = cases for which a valid response existed for the MLS question.

‡Final sample = cases for which a valid response existed for both questions.

§Percentage of final sample cases for which response to ‘daughter circumcised?’ was yes.

FGC, female genital cutting; GDP, gross domestic product; MICS, Multiple Indicator Cluster Surveys; MLS, maternal life satisfaction.

The present investigation was not conducted or conceived of with public or patient input.

Measures

MLS was assessed with a single item: ‘Taking all things together, would you say you are very happy, somewhat happy, neither happy nor unhappy, somewhat unhappy, or very unhappy’? A ‘smiley card’ served as a visual prompt depicting five facial icons ranging from upturned, smiling mouth (very happy = 1) to a downturned mouth (very unhappy = 5). Participants could verbalise or point to the relevant response. For the present paper, responses were reverse coded so that higher scores reflected happiness.

Daughter FGC was assessed by a single dichotomous question asked of mothers for each of their daughters: ‘Has she (daughter) been circumcised’? Likewise, mother FGC was assessed with a single dichotomous question addressed to mothers: ‘Have you been circumcised’? Valid response options for both items were yes = 1 and no = 2. For the present study, responses were reverse coded so that higher scores reflected circumcised.

Mother education was assessed with a single item: ‘What is the highest level and grade or year of school you have attended’? Response options included the following: without formal education = 0, primary education (grade 1 to 5) = 1, secondary education (grade 6 to 10) = 2, and higher secondary education and above (grade 11, 12 and above) = 3.

Household Wealth scores were from the MICS Wealth Quintile Index.¹⁵ Quintiles were assigned to households in the entire MICS dataset for each country. They were generated using principal component analysis based on the following factors: ownership of household goods and amenities; persons per sleeping room; type of floor, roof, wall, cooking fuel and sanitary facility; and source of drinking water.

Country-level wealth was estimated for each country using GDP at purchasing power parity per capita, which improves on nominal GDP by accounting for the relative cost of living (https://www.worldometers.info/gdp/gdp-per-capita/). GDP values for 2022 were used and are reported for each country in table 1.

Country-level daughter FGC frequency was calculated for each country as the percentage of daughters in the final sample for each country for whom their mother identified as having been circumcised. Values for each country are reported in table 1.

Data analysis/statistical analysis

We used moderation analyses to investigate if the association between MLS and daughter FGC differed as a function of country-level daughter FGC rates. The statistical model generated to evaluate the primary hypothesis of the present study involved using MLS as the dependent variable, daughter FGC as the independent variable, country-level FGC as the moderator variable and the following variables as covariates: maternal education, household wealth, maternal FGC and country-level wealth. Analyses were conducted using the PROCESS macro for Statistical Package for the Social Sciences.¹⁶ For all statistical tests, the significance level was pegged at p<0.01. Missing data were treated as missing at random, and listwise deletion procedures were used.

Results

Results are organised into two studies. Study 1 reports on results for the MICS 5 dataset, and Study 2 reports on results for the MICS 6 dataset. Study 2 is a direct replication of Study 1.

Study 1 results

Table 2 provides means and SD for all individual-level variables along with their bivariate correlations. MLS was positively associated with education and, surprisingly, negatively associated with household wealth. MLS was also positively associated with mother and daughter circumcision. Household wealth and education were negatively associated with both mother and daughter circumcision. Mother circumcision was a positive predictor of daughter circumcision.

Table 2.

Correlations and means and SD for individual-level variables for Study 1 and Study 2

						Study 1	Study 2
Variable	1	2	3	4	5	M (SD)	M (SD)
MLS	–	0.06**	−0.04**	0.09**	0.05**	4.31 (8.83)	3.81 (1.10)
Education	0.07**	–	0.43**	−0.21**	−0.16**	1.69 (0.91)	0.90 (1.04)
Wealth	0.06**	0.41**	-–	−0.10**	−0.08**	2.91 (1.39)	2.88 (1.42)
M_FGC	0.03**	−0.24**	−0.12**	–	0.54**	0.58 (.49)	0.48 (.50)
D_FGC	0.07**	−0.16**	−0.08**	0.42**	–	0.31 (.46)	0.14 (.34)

Study 1 correlations are above the diagonal, and Study 2 correlations are below the diagonal.

*p< .01, **p<.001.

D_FGC, daughter female genital cutting (dichotomous); M_FGC, mother female genital cutting (dichotomous); MLS, maternal life satisfaction.

Table 3 provides the results of the regression estimates for all variables entered in the moderation analysis for which MLS is the dependent variable. All entered variables were significant predictors except household wealth. Model summary statistics were as follows: R=0.208, R ²=0.043, Mean Square Error (MSE)=0.664, F (7, 10 772)=69.65, p<0.001. For the interaction term (daughter FGC × country-level FGC), the test of the highest-order unconditional interaction revealed ∆R ²=0.008, F (1, 10 772)=94.99, p<0.001.

Table 3.

Results of regression analyses for variables predicting MLS for Study 1 and Study 2

Predictor	Coefficient	SE	t	99% CI
Study 1
Maternal education	0.045	0.010	4.68**	0.020 to 0.069
Family wealth	−0.020	0.006	−3.22	−0.037 to −0.004
Country GDP	0.000	0.000	12.14**	0.000 to 0.000
Country FGC frequency	0.006	0.000	12.52**	0.005 to 0.007
M_FGC	0.070	0.022	3.23**	0.014 to 0.126
D_FGC	0.425	0.050	8.66**	0.295 to 0.555
D_FGC × country FGC frequency	−0.008	0.001	−9.75**	−0.011 to −0.006
Study 2
Maternal education	0.039	0.004	8.72**	0.027 to 0.050
Family wealth	0.037	0.003	11.55**	0.029 to 0.046
Country GDP	0.000	0.000	49.03**	0.000 to 0.000
Country FGC frequency	0.015	0.000	35.37**	0.014 to 0.016
M_FGC	0.130	0.009	14.17**	0.107 to 0.154
D_FGC	0.071	0.024	2.96*	0.009 to 0.133
D_FGC × country FGC frequency	−0.005	0.001	−5.29**	−0.007 to −0.002

*p<0.01; **p<0.001.

FGC, female genital cutting; FGC, female genital cutting; GDP, gross domestic product; MLS, maternal life satisfaction.

Table 4 provides the results of the moderation analysis evaluating conditional effects of country-level FGC frequency on the association between daughter FGC and MLS. As shown in table 4, daughter FGC is significantly related to MLS, and country-level FGC frequency significantly moderated that relationship. This interaction is illustrated in figure 1. The interaction was probed by testing the conditional effects of daughter FGC on MLS at three levels of country-level FGC frequency, at the 16th percentile, the 50th percentile and the 84th percentile. As shown in table 4, daughter FGC was significantly positively related to MLS at country-level FGC values of 12.90 and 28.80 (the 16th and 50th percentiles, respectively) and significantly negatively related to MLS at a value of 77.40 (the 84th percentile). The Johnson-Neyman technique showed that the relationship between daughter FGC and MLS was significant at values of country-level FGC <43.54 and >56.97.

Table 4.

Conditional effects of daughter FGC on MLS at values of the moderator for Study 1 and Study 2

Moderator value	Coefficient	SE	t	99% CI
Study 1
12.90	0.316	0.041	7.80**	0.212 to 0.421
28.80	0.182	0.030	6.15**	0.106 to 0.259
77.400	−0.227	0.030	−7.62**	−0.304 to −0.150
Study 2
0.400	0.069	0.024	2.92*	0.008 to 0.131
6.400	0.042	0.020	2.10	−0.009 to 0.093
36.70	−0.099	0.019	−5.16**	−0.148 to −0.050

*p<0.01; **p<0.001.

Figure 1.

Maternal life satisfaction-daughter circumcision association by country-level daughter FGC frequency for Study 1. FGC, female genital cutting.

Study 2 results

Means and SD for all individual-level variables, along with their bivariate correlations, are presented in table 2. MLS was positively associated with education, household wealth, mother circumcision and daughter circumcision. Household wealth and education were negatively associated with MLS and daughter FGC. Mother FGC was a positive predictor of daughter FGC.

Table 3 provides the results of the regression estimates for all variables entered in the moderation analysis for which MLS is the dependent variable. All entered variables were statistically significant predictors. Model summary statistics were as follows: R=0.247, R ²=0.061, MSE=1.14, F (7, 69 369)=644.62, p<0.001. For the interaction term (daughter FGC × country-level FGC), the test of the highest-order unconditional interaction revealed ∆R ²=0.000, F (1, 69 369)=28.03, p<0.001.

Table 4 provides the results of the moderation analysis evaluating conditional effects of country-level FGC frequency on the association between daughter FGC and MLS. As shown in table 4 (Study 2 results), daughter FGC is significantly related to MLS, and country-level FGC frequency significantly moderated that relationship. This interaction is illustrated in figure 2. The interaction was probed by testing the conditional effects of daughter FGC on MLS at three levels of country-level FGC frequency, the 16th, 50th and 84th percentiles. As shown in table 4, daughter FGC was significantly positively related to MLS at a country-level FGC value of 0.40 (the 16th percentile) and significantly negatively related to MLS at a country-level FGC value of 36.70 (the 84th percentile). At the intermediate value of 6.40 (50th percentile), the relationship was statistically non-significant. The Johnson-Neyman technique showed that the relationship between daughter FGC and MLS was significant at values of country-level FGC <3.24 and >23.31.

Figure 2.

Maternal life satisfaction-daughter circumcision association by country-level daughter FGC frequency for Study 2. FGC, female genital cutting.

Discussion

Results across the two studies reported here are contrary to predictions. In both studies, the association between daughter FGC and MLS was negatively related to country-level FGC. In both studies, a reversal effect occurred, whereby in countries in which FGC is common, daughter FGC was associated with lower levels of MLS and in countries in which FGC is uncommon, daughter FGC was associated with higher levels of MLS. These results suggest that the intergenerational transmission of FGC cannot be attributed to positive maternal emotions stemming from conformity to cultural norms concerning FGC. Rather, the intergenerational transmission of the practice appears to occur despite its negative association with maternal happiness in countries in which the practice is common.

In countries in which daughter circumcision is culturally normative, why are mothers of circumcised daughters less happy than mothers of uncircumcised daughters? One possibility concerns the potential negative impact on psychological well-being of exposure to community-level FGC. For instance, one prominent African sociologist, theologian and expert on female circumcision identified her own opposition to the practice as arising from witnessing the suffering and carnage arising from the practice in her native Kenyan community.¹⁷ Moreover, recent literature suggests that FGC may be associated with adverse mental health outcomes, although this conclusion is tempered by equivocal findings across generally low-quality studies.^18–20 Indeed, for both samples reported here, levels of MLS were higher for mothers who reported having experienced FGC themselves and who reported it for their daughters. Nevertheless, it may be that cultural forces (perhaps coercive) overpower individual-level maternal aversion to the practice in societies in which it is common. This would be consistent with mothers, as younger women, having relatively lower status and power than men or older women in countries in which FGC is practised.⁴ As such, positive emotions associated with the practice of FGC may reside elsewhere within community social hierarchies.

On the other hand, it could be that maternal unhappiness is associated with higher levels of compliance with cultural customs. If so, this effect is in addition to any influence regarding cultural conformity of the individual- and country-level variables controlled for in the present study, including maternal education, household wealth, country-level wealth and mother’s own personal circumcision history. Of course, given that the data are cross-sectional and not longitudinal, the present studies cannot disentangle issues of direction of effect, nor can they rule out unmeasured third variable effects. This constitutes a weakness of the present study.

In countries in which the practice is uncommon, FGC might symbolise a linkage to a threatened identity or heritage. For instance, the term ngaitana, which means ‘I will circumcise myself’’, became a slogan in response to national and international sanctions against the practice in Kenya.¹⁷ As such, the practice of daughter FGC in cultural contexts in which it is uncommon may embody resistance against social norms perceived as threatening to self and culture. In that context, FGC may reflect or inspire a sense of personal agency and life satisfaction. Indeed, in describing those who defend the practice, Wangila notes that ‘the agency of the individual is becoming accepted in Kenyan communities as a way of resisting social norms (that are contrary to FGC)’.¹⁷ In sum, depending on the sociocultural context, the practice of FGC may reflect different and even contradictory personal and social motivations.^21–23

An implication of these findings for efforts to reduce the practice of FGC is that different motivational forces must be recognised and addressed. For instance, in countries in which FGC is common, it should be communicated that the practice does not bring maternal satisfaction. In countries in which FGC is uncommon, efforts should be concerned with identifying and supporting alternative forms of cultural self-expression, perhaps particularly among those individuals or communities who are minoritised or feel excluded from expressing their own cultural values and heritage.

Comparisons across the two studies for the purpose of investigating change across time are difficult given that only two countries were data available for both MICS 5 and MICS 6, Guinea-Bassau and Nigeria. For those two countries, and for the total sample, rates of both mother and daughter FGC were smaller for MICS 6 than MICS 5. This is consistent with evidence that FGC is on the decline worldwide.⁹ Average MLS scores for the present sample were lower from MICS 5 to MICS 6 as well. Future research might use longitudinal designs to investigate how changing rates of FGC relate to MLS changes and vice versa.

The two studies reported in this paper explored how a psychological variable, MLS, relates to the intergenerational transmission of FGC. Future research should further explore the role of this and other psychological factors as they relate to FGC. It is worth considering, for example, how the practice relates to the psychological well-being of non-maternal family members. Indeed, few studies have explored differences in cross-parent or cross-generational attitudes related to FGC.^21–23 Similarly, it would be worth exploring how seemingly contradictory motives may act to maintain a cultural practice across different cultural contexts. It would also be useful to explore how life satisfaction ratings relate to the different forms of FGC (eg, genital nicking vs labia stitching), which may be differentially associated with disability and trauma.¹⁸ Regarding the psychological significance of FGC, it would be useful to explore how it may be in the service of cultural conformity or cultural non-conformity, depending on the cultural context.

Data availability statement

Data are available upon reasonable request.

Ethics statements

Patient consent for publication

Not applicable.

Ethics approval

The study was reviewed and deemed exempt by the Washington State University Institutional Review Board (IRB#20245-001).