Single motherhood in Ghana: analysis of trends and predictors using demographic and health survey data

The study investigates how socio-economic and demographic transformations have influenced the rise of single motherhood in Ghana between 1993 and 2014. Motivated by evidence linking single motherhood to poverty and adverse health outcomes for mothers and children, and noting Ghana’s limited social protection for vulnerable groups, the authors aim to quantify trends in single motherhood and identify key predictors using nationally representative GDHS data. Theoretical framing draws on Becker’s economic theory and the economic independence hypothesis, which posits that improvements in women’s economic status may reduce gains to marriage and increase non-marriage or marital dissolution. The authors explicitly test whether women with stronger economic positions are more likely to experience single motherhood than poorer women. They also situate the inquiry within Ghana’s socio-cultural context, including kinship systems and changing family norms, and note demographic shifts such as rising age at first marriage, premarital childbearing, and marital dissolution that may elevate single motherhood risks.

Background literature highlights adverse consequences of single motherhood, including poverty and worse child outcomes (malnutrition, under-five mortality), with single-mother families often faring worse than single-father families. In Ghana, indicators suggest single motherhood is common: median age at first marriage has increased, divorce and separation have risen, and premarital childbirth remains prevalent. Prior Ghanaian studies have focused on marital dissolution with limited attention to single motherhood per se. The economic independence hypothesis suggests higher female economic status may increase non-marriage/divorce, potentially raising single motherhood. Empirical evidence is mixed: some contexts show higher economic status increases separation, while others find single motherhood concentrated among poorer or less educated women. Ghana’s kinship systems (matrilineal vs patrilineal), bridewealth practices, and family organization may shape marital stability and child living arrangements, potentially affecting single motherhood prevalence. The authors note limited data on socio-cultural mechanisms and call for qualitative work.

Study design: Secondary analysis of pooled data from five GDHS waves (1993, 1998, 2003, 2008, 2014), each nationally representative of women aged 15–49 across Ghana’s ten regions. Sampling: Two-stage cluster design based on national census frames (1984 for 1993/1998; 2000 and 2010 for 2003/2008/2014), with systematic sampling of households in selected clusters. Weights: DHS sample weights applied to adjust for selection probability and non-response; robust standard errors via Huber-White to account for clustering. Software: Stata 13. Sample: Restricted to mothers living with at least one of their own children; excluded non-mothers and mothers without any co-resident child. Analytic sample sizes: 1993 (n=3,204), 1998 (n=3,000), 2003 (n=3,380), 2008 (n=2,837), 2014 (n=5,644); total N=17,831. Outcome: Single motherhood (1=yes, 0=no), defined as women not married or in union who had at least one of their own children living with them (premarital birth, divorced, separated, widowed). Predictors: Age (7 groups), education (no, primary, secondary+), occupation (not working; professional; sales/service; agricultural; skilled/unskilled), age at first sex (<15, 15–19, 20–24, 25+), age at first birth (<15, 15–19, 20–24, 25+), current age of child (<5, 5–9, 10–14, 15+), number of living children (1, 2, 3, 4+), contraceptive use (no/yes), wealth quintile (poorest to richest), religion (no religion; Orthodox/Protestant; other Christians; Islam; other), residence (urban/rural), region (10 regions). Analysis: Univariate descriptive statistics; bivariate Pearson chi-square tests for associations; bivariate and multivariate binary logistic regression. Models: Sequential models assessing individual and contextual factors; survey year controlled. Multicollinearity check via VIF indicated no concern (mean VIF=7.98; threshold >10). 1988 GDHS was excluded due to absence of wealth index.

- Prevalence and trends: Single motherhood increased from 14.1% (1993) to 19.5% (2014); slight decline in 2003 (13.0%) before rising in 2008 (16.3%) and 2014. Overall, 84% in-union and 16% single among mothers (N=17,831). - Pathways: Shift from divorce as dominant pathway (47.5% of single motherhood in 1993) to premarital birth/never-married as dominant in 2008 (34.7%) and 2014 (39.1%). Separation remained substantial; widowhood contribution relatively low (14.5% in 1993 to ~17% in 2014). - Bivariate patterns (weighted): Significant associations with age, education, occupation, age at first sex, age at first birth, current child age, number of living children, contraceptive use, wealth, religion, residence, region, and survey year (all p<0.001). - Multivariate logistic regression (Model II; AIC=12,864.3; Pseudo R²=0.1363; p<0.001): • Age: Compared to 15–19, lower odds at 20–24 (OR=0.58; 95% CI: 0.47–0.71), 25–29 (0.52; 0.41–0.66), 30–34 (0.67; 0.51–0.88); attenuation at older ages. • Education: Higher odds vs no education—primary (1.27; 1.11–1.46), secondary+ (1.31; 1.15–1.51). • Occupation: Lower odds vs not working—professional (0.72; 0.55–0.94), agricultural (0.68; 0.58–0.79). • Age at first sex: 25+ vs <15 (0.58; 0.48–0.70). • Age at first birth: 25+ vs <15 (0.43; 0.32–0.59); 20–24 (0.66; 0.51–0.85). • Contraceptive use: Users vs non-users (0.64; 0.57–0.71). • Child’s current age: vs <5 years—5–9 (2.42; 2.12–2.76), 10–14 (2.71; 2.24–3.27), 15+ (2.63; 1.97–3.51). • Number of living children: vs 1 child—2 (0.43; 0.38–0.50), 3 (0.27; 0.22–0.32), 4+ (0.22; 0.18–0.26). • Wealth: Richer (0.76; 0.59–0.96), Richest (0.57; 0.31–0.56) vs Poorest; middle/poorer not significantly different in final model. • Residence: Urban vs rural (1.21; 1.08–1.36). • Religion: Islam vs no religion (0.58; 0.46–0.74); Orthodox/Protestant and other Christians not significant; other religion not significant. • Region: Lower odds vs Western—Volta (0.60; 0.49–0.74), Eastern (0.81; 0.67–0.97), Brong-Ahafo (0.74; 0.61–0.90), Northern (0.39; 0.30–0.51), Upper East (0.56; 0.44–0.70), Upper West (0.53; 0.42–0.68). • Year effects (ref=1993): 2003 (1.29; 1.08–1.54), 2008 (1.81; 1.54–2.13) elevated odds; 1998 and 2014 not significantly different. - Interpretation: Single motherhood is more prevalent among economically poorer women; delayed sexual debut and childbearing, contraceptive use, employment (professional/agricultural), and Muslim affiliation are protective; urban residence and older child age increase odds.

Findings address the research questions by documenting a sustained rise in single motherhood and identifying key individual and contextual predictors. The shift from divorce to premarital childbearing as the dominant pathway aligns with rising age at first marriage and evolving social norms around premarital birth. The inverse associations with higher ages at first sex and first birth, and with contraceptive use, suggest that preventing early sexual debut and unintended pregnancies could reduce single motherhood. Employment in professional and agricultural sectors appears protective relative to non-employment, possibly reflecting economic stability and selection. Contrary to the economic independence hypothesis, higher household wealth is associated with lower odds of single motherhood, indicating over-representation among poorer women and highlighting poverty’s role in nonmarital fertility and union instability. Urban residence increases risk, potentially due to later marriage and higher premarital fertility in urban settings. Religious affiliation, particularly Islam, is linked to lower single motherhood, perhaps via norms discouraging premarital sex and divorce. Regional differences likely reflect variations in age at marriage and cultural practices. Collectively, these results underscore the importance of addressing premarital fertility and economic vulnerability to mitigate single motherhood and its adverse effects.

Single motherhood in Ghana increased from 1993 to 2014, with premarital childbearing overtaking divorce as the leading pathway. Both individual-level (age, education, occupation, age at first sex, age at first birth, number of children, child’s age, contraceptive use) and contextual factors (religion, region, residence) predict single motherhood. Contrary to the economic independence hypothesis, lower wealth is associated with higher odds of single motherhood, indicating concentration among poorer women. Policy implications include empowering single mothers economically, expanding access to family planning and sexual education to delay sexual debut and first births, and addressing urban-specific risk factors. Future research should use qualitative methods to explore socio-cultural mechanisms (e.g., normalization of single motherhood, men’s attitudes toward economically independent women) and better understand regional and religious differences.

The study uses cross-sectional DHS data, precluding causal inference. Important determinants such as normalization of single motherhood, comprehensive sex education, and men’s attitudes toward partnering with economically independent women are not captured in DHS, limiting explanatory power. Cultural mechanisms (e.g., kinship systems, bridewealth) could not be directly measured. Despite weighting and robust methods, residual confounding and reporting bias may remain.