An age difference between sexual partners is a social norm in most societies; typically, the man is older than the woman. The acceptability of unions in which the husband is considerably older than his wife is more variable, although relatively large age differences are found in patriarchal societies. Large age gaps are often accompanied by differences in maturity, life experiences, social position and financial resources, which may make relationships inherently unequal and a source of risk for women’s health.^1–5 For instance, asymmetries in age, education and economic circumstances compromise women’s ability to negotiate safer sexual behavior.^1–3,6–8

The studies that have examined associations between partners’ age differences and contraceptive behavior have concentrated on adolescent females’ nonmarital relationships and on whether protection was used against STIs (especially HIV).^1–4,7–17 The emphasis on adolescents and on nonmarital relationships may divert attention from more common age asymmetries and from the risks that may exist in such relationships. For instance, age differences in marital relationships are usually of little public interest, unlike situations involving unmarried adolescents. Furthermore, there is the need to examine the role that contraceptive use may play in preventing unintended pregnancy in age-asymmetric relationships, given that contraception has been found to be an effective means of fertility control for various health, economic and social reasons.¹⁸

Examination of age differences between marital partners is of particular importance in Sub-Saharan Africa, where, as a result of gender imbalances and cultural influences, males are frequently in charge of what happens (or does not happen) within relationships, including sexual interactions and decisions regarding family size and contraceptive use. The limited decision-making power of women in this setting often impedes their ability to achieve their fertility goals.^19–26 Moreover, in a patriarchal context, pronatal views tend to be stronger among men than among women, and husbands’ real or perceived opposition to family planning may prevent wives who want to delay or stop childbearing from using contraceptives.^3,16,27–32 The situation is no different in Nigeria, a patriarchal society where decision-making powers are traditionally vested in males. Nigeria is the most populous country in Africa, and women have an average of about six children, a level that has remained stable over the past two decades.^33–36 Only about 15% of married women in Nigeria use any method of contraception, and about 20% of married women have an unmet need for contraception.³⁷

Although some studies have shown that women in Nigeria have a degree of autonomy regarding contraceptive use,^26,38 the evidence is inconclusive and women’s degree of autonomy varies by region. Further, whether age differences between marital partners are related to contraceptive use in the country is unknown. Given that age differences might influence relationship dynamics in ways that have adverse effects on women’s health, and that male partners have enormous influence on fertility decisions and contraceptive use in Nigeria, the current study examines the association of age differences between married couples with women’s contraceptive behavior. Specifically, the study seeks to determine whether a large age difference between marital partners is negatively associated with the woman’s ability to use contraceptives.

METHODS

This study used matched couples data from the 2008 Nigerian Demographic and Health Survey (NDHS). The couples data set includes responses from 8,731 women aged 15–49 and their partners. We excluded 1,315 women who were pregnant, 203 who were infecund, 636 who had not had sex for at least six months and 25 whose partner was infecund. The resulting analytic sample consisted of 6,552 sexually active women—6,518 fecund women who had fecund partners and 34 women who had undergone sterilization as a family planning method. The latter group was left in the analytic sample because their use of sterilization as a family planning method may be related to the age difference between them and their partners.

The dependent variable was current contraceptive use. Women were asked if they were doing something or using something to delay or avoid becoming pregnant. Current contraceptive use was coded as 0 for nonuse, 1 for use of any couple method (condom, withdrawal or periodic abstinence) and 2 for use of any non–couple method (pill, IUD, injectable, diaphragm, female sterilization, implant or lactational amenorrhea). Although the methods grouped as "couple methods" include both modern and traditional methods, they all require the involvement and cooperation of the male partner. We examined these methods separately from other methods to explore the possible influence of male partners on women’s contraceptive use and whether age differences between marital partners are associated with use of such methods.

The key independent variable was partners’ age difference, which was derived by subtracting the woman’s age from her partner’s. Women were grouped according to whether they were older than their partner (<0) or 0–4, 5–9, or 10 or more years younger. Since Nigeria is regionally, culturally, socially and economically diverse, and marital partners’ fertility intentions may interact in complex ways, the analysis controlled for a number of related variables. Individual characteristics included in the analysis were the women’s age (categorized as 15–24, 25–34, or 35 or older), religion (Muslim, Catholic, other Christian or other) and type of employment (professional, sales and service, agriculture and other, or none). Other variables were joint measures related to women and their partners, including household wealth (low, medium or high),* place of residence (urban or rural), region (South West, North Central, North East, North West, South East or South South)† and type of marriage (monogamous or polygynous). Three final variables compared women with their spouses. A measure of educational attainment classified couples according to whether both had no education, both had a primary education, both had at least a secondary education, the woman was more educated than the man or the man was more educated than the woman. Another variable assessed whether the two partners had the same number of children, the woman had more children than her husband did (e.g., because she had children from a prior marriage) or the man had more children than his wife did. The last measure examined congruence in fertility intentions: whether both partners wanted a child soon, both wanted to space or limit births, the woman wanted a child soon but the man did not, the man wanted a child soon but the woman did not, or other mismatches in intentions. Individuals were classified as wanting a child soon if they wanted one within two years, were unsure of their desired timing or undecided about their fertility intentions; as wanting to space births if they wanted a child more than two years in the future; and as wanting to limit births if they wanted no future births.

Descriptive, bivariate and multivariate analyses were carried out using Stata version 12. The bivariate and multivariate analyses used multinomial logistic regression modeling to assess the unadjusted and adjusted relative risk ratios of the association of age difference between partners with current contraceptive use (couple method vs. nonuse, and non–couple method vs. nonuse). As the proportion of women who were older than their partner was very small (2%), such women were combined with women who were 0–4 years younger than their partner in the bivariate and multivariate analyses.

It is important to note the possibility of bias in the analytic sample, as men older than 59 were not interviewed in the NDHS. Thus, some women may have been interviewed but not included in the couples sample because their partner was older than 59. Tabulation reveals that the distribution of couples’ age differences varied significantly by women’s age. The percentage of women who were 10 or more years younger than their partner was smallest among women aged 35 or older (Table 1). This is likely due to the exclusion from the couples sample of some older women (especially those aged 45–49) whose partner was 10 or more years older than they were.

Analysis of the NDHS individual data set showed that 11% of women reported that their partner was older than 59 (and therefore ineligible for the couples data set). However, a comparison between women’s reports of their partner’s age and men’s reports of their own age indicated that only 75% of women in the couples data set correctly reported their partner’s age. It is thus difficult to reliably estimate the proportion of women who were excluded from the couples data set (or from the analytic sample of this study) because of the NDHS age restriction. In addition, information on duration of marriage with the current partner was not available in the data set and could not be controlled for in this analysis.

RESULTS

Descriptive

Women in the analytic sample had a mean age of 30.5. Ninety-eight percent were younger than their partners; 43% were younger by 10 or more years (Table 2, page 41). Nearly 60% were Muslim, and 69% were employed. In 32% of couples, neither partner had any education; in 10%, both had only a primary education. In 40% of couples, educational attainment differed between partners; in three-quarters of these cases, the man had more education than the woman. The sample was distributed fairly equally among the three household wealth categories. About 63% of the couples were from a northern region, 69% lived in rural areas and 31% were in a polygynous union. In 53% of the couples, the two partners reported the same number of living children, while in 41% the man reported more children than his partner did. Fertility intentions were aligned in two-fifths of couples: In 18%, both partners wanted a child within two years, while in 23% both wanted a child after two years or did not want any (or any more) children. Only 18% of couples were using a contraceptive method; 10% (or 54% of couples practicing contraception) were using a method that required cooperation of the male partner (i.e., a couple method).

Bivariate

The key independent variable, partners’ age difference, was negatively associated with contraceptive use (Table 3, page 41), indicating that the smaller the age difference between marital partners, the more likely the woman was to use a method. Specifically, couples with an age difference of less than five years or of 5–9 years were more likely than those with a difference of 10 or more years to be practicing contraception, whether they were using a couple method (relative risk ratios, 1.4–1.9) or a non–couple method (1.7–2.2). All of the other independent variables were also associated with current contraceptive use (p<.001 for each; not shown).

Multivariate

The negative bivariate association between partners’ age difference and contraceptive use was no longer significant in multivariate models that controlled for respondents’ individual and joint characteristics with partners (Table 4). However, associations with both contraceptive categories were found for several other measures in the multivariate analyses. The likelihood that a woman and her partner were using a couple method rather than no method was greater if at least one partner had a primary or higher education than if both had no education (relative risk ratios, 1.9–2.2). Most of these associations were also apparent for non–couple methods (2.8–4.0). Household wealth was also positively associated with contraceptive use; couples with a medium level of wealth were more likely than couples with a low level of wealth to be using a couple method rather than no method (1.6), and those in the high wealth group had an elevated likelihood of using a couple method or a non–couple method (2.8 for each).

Substantial regional differences emerged. Compared with women in the South West, women in all other regions except the North Central were significantly less likely to be using a couple method rather than no method (relative risk ratios, 0.3–0.7), and women in all three northern regions were less likely to be using a non–couple method (0.03–0.31). In addition, couples in which at least one partner wanted to space or limit childbearing were more likely than those in which both partners wanted a child soon to be using either type of contraceptive method rather than no method (1.8–5.5). The likelihood of contraceptive use was greater when the woman wanted to space or limit and her male partner wanted to have a child soon than when the reverse was the case.

Other variables were less consistently associated with contraceptive use. Respondents were less likely to be using a couple method than no method if they were aged 15–24 rather than aged 25–34 (relative risk ratio, 0.6), and more likely to be doing so if they were non-Catholic Christian rather than Muslim (1.6) or had a sales or service job rather than no job (1.5). Finally, women were less likely to be using a non–couple method rather than no method if they were in a polygynous rather than monogamous marriage (0.4), or if they reported having more children than their husband did rather than reporting the same number (0.5).

DISCUSSION

This study examined the association of age differences between marital partners with women’s contraceptive behavior in Nigeria. Nearly all of the women in the study were younger than their partner, and two-fifths were younger by 10 or more years. This might be related to the patrilineal kinship organizations found in Nigeria. In patrilineal societies, age differences between partners tend to be relatively large, in part because such unions are arranged more frequently than would occur by chance or by the couples’ own choices alone.^2,4,5

Although the size of the age difference between partners was negatively associated with contraceptive use in the bivariate analysis, the association was no longer statistically significant after adjustment for women’s and couples’ characteristics. The reason for this unexpected finding is not apparent and was not explained by additional statistical tests.‡ However, a possible reason for the difference between the findings of this study and those of other studies^1–4,7–17 could be that negative reproductive health outcomes are more common in nonmarital relationships and among adolescents than within marriage and among older women, respectively. Moreover, women are often able to use contraceptives covertly, and exercise some degree of autonomy on use. Therefore, the influence of their husbands may not be as substantial as one might expect on the basis of age differences.

The exclusion of some women from the couples data set because their husband was older than 59 (and hence not surveyed) may have affected the study findings. For instance, if it is true that 11% of surveyed women had partners older than 59, then the differences across age-groups in the percentage distribution of women by spousal age difference may have been inflated in the analytic sample, especially among younger women. The influence of this bias on our findings is difficult to determine, but may have contributed to the lack of an association in the multivariate analysis between age differences and contraceptive use. This scenario and other possibilities (such as having a larger or smaller percentage of women whose partners were 10 or more years older than they were) remain a matter of conjecture. A survey that does not restrict data collection by male partners’ age would shed further light on this issue.

Couples’ educational attainment, wealth and fertility intentions were associated with the use of contraceptives. Further, there was a notably elevated likelihood of women using contraceptives to favor their own fertility intention when it did not match their partner’s. This finding is a departure from previous studies that documented a reduced likelihood of contraceptive use by women if their fertility desires conflicted with their husband’s real or perceived desires.^{23,24,29,39,40}

Couples living in the South West region were generally more likely than those living elsewhere in the country to be using a method of contraception, a result consistent with the more favorable indicators of socioeconomic status and reproductive health among people in the South West region than among those in the other regions of the country.^38,41–44 Women in polygynous unions were less likely than those in monogamous marriages to be using a non–couple method of contraception rather than using no method. This could be due, for example, to competition among co-wives for their husband’s attention and resources. Having many children is reportedly a powerful strategy that women in polygynous unions use to get more leverage over their co-wives and their husbands, even if they themselves do not wish to have more children.^38,41 Women aged 24 or younger were less likely than older women to be using a couple method of contraception rather than no method, probably because of ambivalence in their fertility intentions.⁴⁵

In conclusion, this study found an association between large spousal age differences and women’s contraceptive behavior in Nigeria in a bivariate analysis, but not in multivariate analyses. Further study, including use of qualitative methodology, is recommended to explore and better understand this finding.

REFERENCES