Determinants of children ever born among women in Somalia: a national survey study Authors: Mohamed Mustaf Ahmed, Mohamed Abdirahim Omar, Yahye Sheikh Abdulle Hassan, Abdirasak Sharif Ali Published: 2026 ABSTRACT AbstractBackgroundFertility rates in Somalia are among the highest globally, posing significant challenges to maternal and child health and sustainable development. However, there is limited evidence regarding the specific drivers of high fertility. This study aimed to identify the factors associated with the number of children ever born among Somali women of reproductive age to inform reproductive health policies and programs.MethodsA cross-sectional analysis of a nationally representative household survey was performed. We analyzed the weighted data of 27,352 women aged 15–49 years who participated in the 2020 Somali Health and Demographic Survey. The number of children born was defined as the total number of live births and was modeled as a count outcome. Owing to overdispersion, survey-adjusted negative binomial regression was used to estimate unadjusted and adjusted incidence rate ratios with 95% confidence intervals for demographic, socioeconomic, reproductive, and geographic factors.ResultsThe mean number of children born was 6.22. Higher education level showed the strongest inverse association (higher vs. none: incidence rate ratio = 0.154; 95% confidence interval: 0.049–0.487). A later age at first birth (> 20 vs. ≤20 years) and literacy were associated with fewer children ever born (incidence rate ratio = 0.913; 95% confidence interval: 0.902–0.923, and incidence rate ratio = 0.965; 95% confidence interval: 0.939–0.991, respectively). Compared with rural residents, urban women had more children ever born (incidence rate ratio = 1.019; 95% confidence interval: 1.003–1.035), whereas nomadic women had fewer (incidence rate ratio = 0.979; 95% confidence interval: 0.965–0.993). Women in the middle wealth quintile had more children ever born than the poorest women (incidence rate ratio = 1.033; 95% confidence interval: 1.021–1.046). Compared with current modern method users, non-users intending future use and non-users with no intention to use had more children ever born.ConclusionsFertility rates in Somalia remain high and are shaped by structural and behavioral factors. Expanding girls’ education, delaying childbearing, improving literacy, and ensuring equitable access to contraception may help reduce fertility and improve population health.Keywords: Somalia, Fertility, Children ever born, Contraception, Education, Age at first birth Text box 1. Contributions to the literature• Evidence on fertility patterns in Somalia is limited; this study provides national data on the number of children born to women of reproductive age in Somalia.• It shows how education, literacy, timing of the first birth, and contraceptive intentions are linked to family size in a fragile, understudied setting.• The findings add to the population health literature by highlighting that fertility is shaped by both social conditions and access to reproductive health services.• This evidence can help guide policies that support girls’ education, delayed childbearing, and equitable access to family planning. Introduction The number of children ever born (CEB) is a key measure of cumulative fertility, providing a picture of a woman’s reproductive history and serving as a vital indicator for assessing population dynamics [1]. Although global fertility rates have declined over the past few decades, this trend is not uniform. Sub-Saharan Africa continues to have the highest fertility rates globally, presenting challenges for health systems and sustainable development [2]. According to recent nationwide data from the 2020 Somali Health and Demographic Survey (SHDS), the total fertility rate (TFR) in Somalia was 6.9 children per woman. This figure places Somalia among the countries with the highest fertility rates globally and stands in stark contrast to the regional average for sub-Saharan Africa (4.6 children per woman) and the global average (2.3 children per woman) during the same period [2]. Such high fertility rates have profound implications, contributing to population growth that strains resources and poses risks to maternal and child health [3, 4]. Therefore, understanding the factors driving the number of children born to Somali women is essential for developing effective public health policies and interventions aimed at improving reproductive health and empowering women. Research in high-fertility settings has identified the key factors associated with CEB. Women’s education level is among the strongest predictors. Studies have shown a clear inverse relationship between female education and fertility, as education leads to delayed marriage, increased knowledge of family planning, and greater female autonomy in reproductive decisions [5, 6]. Low contraceptive prevalence and high unmet need for family planning directly mediate fertility rates. Where access to modern contraception is limited by logistical, social, or cultural barriers, women may have more children than they desire [7]. Sociocultural norms, religious beliefs, and marital patterns also influence fertility rates. In many traditional societies, large families are culturally valued, and status may be linked to childbearing [8, 9]. Early marriage extends the reproductive period and is correlated with a higher number of children ever born [10, 11]. The focus on Somalia is particularly timely and essential as the country emerges from three decades of state fragility and a prolonged data vacuum. The 2020 SHDS represents the first comprehensive opportunity in a generation to analyze fertility trends among urban, rural, and historically under-studied nomadic populations. This study is necessitated by recent developments, specifically the massive scale of internal displacement and a total fertility rate of 6.9, which far outpaces current economic reconstruction efforts. Understanding the drivers of high parity in this fragile setting is a critical priority for the Somali Federal Ministry of Health. By identifying the specific socioeconomic determinants of CEB, this study provides the empirical evidence required to transition from generalized regional health strategies to targeted, decentralized interventions. Ultimately, these findings have direct implications for implementing the National Health Sector Strategic Plan, offering a blueprint for reproductive health programs that can effectively reach marginalized nomadic groups and internally displaced communities. While these determinants are well established in many high-fertility settings, their specific manifestations and relative importance in Somalia remain significantly under-researched, presenting several critical gaps in the literature that this study aims to address. First, recent nationally representative data are scarce because of the country’s long history of instability, meaning that much of the existing knowledge is based on outdated information or localized samples. This study fills this data gap by utilizing the comprehensive 2020 SHDS. Second, most fertility research originates from stable, low-income countries, leaving a contextual gap in our understanding of fertility dynamics within fragile and post-conflict states such as Somalia, where factors such as displacement, disrupted health systems, and the presence of nomadic populations may uniquely shape reproductive behaviors. Finally, this study addresses a methodological gap by applying a comprehensive multivariate regression model to concurrently analyze the independent effects of a wide range of demographic, socioeconomic, and reproductive factors, thereby providing robust and policy-relevant evidence. By filling these gaps, this study offers crucial insights for policymakers and public health stakeholders working to improve reproductive health outcomes in Somalia and other fragile settings. Methods Study design and objective This study employed a cross-sectional analytical design to investigate the individual- and community-level determinants of the number of children born to women of reproductive age in Somalia. The primary objective was to identify the sociodemographic, reproductive, and contextual factors associated with higher or lower CEB in this population sample. Data source Data were drawn from the 2020 SHDS, a nationally representative household survey that used a multistage stratified cluster sampling approach to capture urban, rural, and nomadic populations across all regions of Somalia. Enumeration areas were selected with a probability proportional to size, and within each selected cluster, households were sampled systematically to yield a dataset weighted to reflect the national distribution of women aged 15–49 years [12]. Sample size and sampling procedure The analytical sample comprised women aged 15–49 years with complete information on CEB and key explanatory variables. After excluding records with missing data, the final weighted sample comprised 27,352 participants. The sampling weights provided by the survey were applied to account for the complex design and to ensure the generalizability of the findings. Measurement of variables The outcome variable, CEB, was defined as the total live births reported by respondents at the time of the interview and treated as a non-negative count. The explanatory variables included administrative region (Awdal as reference), place of residence (rural [reference], urban, nomadic), women’s education (none [reference], primary, secondary, higher), literacy (cannot read [reference] vs. able to read whole sentence), ever use of contraception (yes [reference] vs. no), marital status (married [reference], divorced, widowed), sex of the most recent live birth (male [reference] vs. female), household wealth index (poor [reference], middle, rich), media exposure (weekly or more [reference], less than weekly, not at all), sex of household head (male [reference] vs. female), age at first birth (≤ 20 years [reference] vs. >20 years), and contraceptive intent (using modern method [reference], non-user intends to use later, non-user does not intend to use). To capture the effects of gender dynamics, proxies for women’s empowerment, including maternal employment status and involvement in household decision-making, were incorporated into the analysis. Reference categories were defined to enable the direct interpretation of incidence rate ratios. ‘Cannot read at all’ was used as the reference category for literacy because it represents the dominant group (93.55% of the sample), ensuring statistical stability and allowing IRRs to be interpreted directly as the effect of becoming literate. Statistical analysis Analyses were conducted using Stata 16, incorporating sampling weights, clustering, and stratification to reflect the complex survey design. Descriptive statistics were first used to summarize the sample characteristics, distribution of children ever born, and regional variations. Given that the outcome variable (CEB) is a count and initial diagnostics indicated overdispersion (i.e., variance greater than the mean), we compared several count data regression models, including Poisson, negative binomial (NB), zero-inflated Poisson, and zero-inflated negative binomial. Based on the log-likelihood, Akaike information criterion (AIC), Bayesian information criterion (BIC), and diagnostic tests, the NB model was determined to be the most appropriate for the data. Our inferential analysis to identify factors associated with CEB involved a two-stage modeling process. First, we conducted a bivariate analysis by fitting separate survey-adjusted NB models for each of the explanatory variables. The purpose of this step was to estimate unadjusted incidence rate ratios (IRRs). An unadjusted IRR measures the crude association between a single predictor and the CEB outcome without accounting for the influence of other variables. This provides an initial assessment of the strength of each determinant. Second, we constructed a multivariate survey-adjusted NB regression model to estimate the adjusted incidence rate ratios (aIRRs). Variables that were statistically significant in the bivariate analysis (e.g., p < 0.20) or considered theoretically important were included in the final model. The aIRR measures the independent association between a specific explanatory variable and the CEB outcome after controlling for the effects of all other variables in the model. By “adjusting” for these potential confounders, the aIRR provides a more precise and reliable estimate of each factor’s net effect on fertility. For all models, robust standard errors were used for hypothesis testing, with statistical significance set at P < 0.05. Variance Inflation Factors (VIFs) were calculated for the final multivariate model to assess multicollinearity among the explanatory variables, confirming that it was not a concern in our analysis. Results Characteristics of the study population The descriptive characteristics of the study population are presented in Table 1. The weighted analysis included 27,352 reproductive-age women. The largest proportion of respondents was in the 25–34 age group (41.76%), followed by the 35–44 age group (38.96%). A significant majority of women (87.85%) reported having no formal education, and only 6.45% were literate (able to read a complete sentence). Most women were married (89.03%) and resided in urban areas (62.17%). A very low percentage of women (2.22%) reported having ever used contraception, and only 0.11% (n = 29) were currently using a modern method. Table 1Characteristics of the study populationVariableCategoryFrequency (Weighted)Percentage (%)Respondent’s age group15–243,01211.0125–3411,42341.7635–4410,65738.9645–492,2608.26Age at first birth<=2018,17966.46> 209,17333.54Mean (SD)19.6 (3.7)Sex of household headMale18,85568.94Female8,49631.06Current marital statusMarried24,35089.03Divorced1,6826.15Widowed1,3194.82Highest educational levelNo Education24,02987.85Primary2,75110.06Secondary5652.07Higher70.02Literacy statusIlliterate25,58793.55Literate1,7656.45Wealth quintilePoor12,15544.44Middle11,59842.40Rich3,59913.16Media exposureWeekly or more2,5179.20Less than weekly9463.46Not at all23,88887.34Sex of most recent live birthMale14,46552.89Female12,88747.11Ever used contraceptionYes6082.22No26,74497.78Contraceptive use and intentionUsing modern method290.11Non-user, intends to use later1,6866.17Does not intend to use25,63693.73ResidenceRural6,59224.10Urban17,00462.17Nomadic3,75513.73RegionAwdal2530.92Woqooyi Galbeed1,3384.89Togdheer8203.00Sool6992.55Sanaag1,0954.00Bari1,8106.62Nugaal1,0073.68Mudug2,2778.32Galgaduud2,0217.39Hiraan1,2634.62Middle Shabelle1,9927.28Banadir7,31226.73Bay1,9126.99Bakool7532.75Gedo1,1114.06Lower Juba1,6916.18Weighted frequencies; totals may not sum to 27,352 due to rounding Characteristics of the study population Weighted frequencies; totals may not sum to 27,352 due to rounding Frequency distribution of the dependent variable The dependent variable, total children ever born, had a weighted mean of 6.22 (SD = 2.80) and ranged from 0 to 17. The distribution, detailed in Table 2, reveals a concentration of women at parities 4–8, which accounted for nearly half of the study population (12.29% at parity 4 and 13.29% each at parities 5 and 6). Only 0.14% of women reported zero CEB. Table 2Frequency distribution of total children ever born per womanTotal children ever bornFrequency (Weighted)Percentage (%)0380.1415932.1721,7326.3332,5489.3143,36212.2953,63513.2963,63613.2973,23411.8282,6839.8192,3558.61101,3524.94111,0293.76125792.12132580.94141820.6715730.2716160.0617480.18Mean (SD)6.22 (2.80) Frequency distribution of total children ever born per woman Geographical variation of total children ever born per woman Geographical variations in the weighted mean of children ever born across Somalia are presented in Fig. 1. The mean CEB exhibited considerable regional variation, ranging from a minimum of 5.27 in Woqooyi Galbeed to a maximum of 6.62 in Sool. The regions reporting the highest mean CEB (over 6.40) included Sool (6.62), Bay (6.61), Togdheer (6.49), and Lower Juba (6.49). Conversely, the lowest mean CEB rates (below 6.00) were observed in Woqooyi Galbeed (5.27), Hiraan (5.83), and Middle Shabelle (5.88) regions. Geographical variation of total children ever born in Somalia Model selection The assumption of no serious multicollinearity was supported by the VIF analysis (mean VIF = 1.13). Given the count nature of the dependent variable and the observed overdispersion (variance of 7.85 is much greater than the mean of 6.22), four count data regression models were evaluated: Poisson, NB, zero-inflated Poisson (ZIP), and zero-inflated negative binomial (ZINB). As shown in Table 3, the NB model showed the lowest AIC (132702.77) and BIC (132719.21) values, indicating the best fit. The likelihood ratio test comparing the NB and Poisson models confirmed overdispersion (χ²(1) = 799.35, p < 0.001). Thus, the NB model was selected for multivariate analysis. Table 3Model selectionModelLog-likelihoodAICBICPoisson-66749.064133500.13133508.34Negative Binomial-66349.387132702.77132719.21 Zero-Inflated Poisson-66749.064133502.13133518.56 Zero-Inflated Negative Binomial-66349.387132704.77132729.43Lower AIC and BIC values indicate a superior model fit Model selection Lower AIC and BIC values indicate a superior model fit Determinants of the number of children ever born The factors associated with the number of children ever born were analyzed using bivariate and multivariate negative binomial regression, with the results presented as IRRs in Table 4. An IRR greater than 1 indicates a higher rate of childbearing, whereas an IRR less than 1 indicates a lower rate compared to the reference category. Table 4Factors associated with the number of children ever born*VariableCategoryUnadjusted IRRp-valueAdjusted IRR95% CI for Adj IRRp-valueAge at first birth≤ 2011> 200.911< 0.001***0.9130.902–0.923< 0.001***Sex of household headMale11Female0.9940.2740.9960.984–1.0070.479Marital statusMarried11Divorced0.818< 0.001***0.810.791–0.829< 0.001***Widowed1.0090.4791.0130.989–1.0390.296Women’s educationNo Education11Primary0.961< 0.001***0.9480.929–0.968< 0.001***Secondary0.894< 0.001***0.8840.850–0.920< 0.001***Higher0.160.002**0.1540.049–0.4870.001**Women’s literacyCannot Read at all11Able to read whole sentence0.944< 0.001***0.9650.939–0.9910.010*Wealth quintilePoor11Middle1.031< 0.001***1.0331.021–1.046< 0.001***Rich1.0080.3531.0070.988–1.0260.458Media exposureWeekly or more11Less than weekly1.074< 0.001***1.081.046–1.114< 0.001***Not at all0.970.001**0.9630.944–0.982< 0.001***Sex of most recent live birthMale11Female0.978< 0.001***0.9780.967–0.988< 0.001***Ever contraceptive useYes11No0.909< 0.001***0.9180.883–0.955< 0.001***Contraceptive intentUsing modern method11Non-user, intends to use later1.2310.050*1.3021.057–1.6030.013*Does not intend to use1.1970.0891.2951.052–1.5940.015*ResidenceRural11Urban1.020.005**1.0191.003–1.0350.016*Nomadic0.9940.4370.9790.965–0.9930.004**RegionAwdal11Woqooyi Galbeed0.9490.013*0.9430.904–0.9830.005**Togdheer1.079< 0.001***1.0731.032–1.115< 0.001***Sool1.087< 0.001***1.0841.045–1.125< 0.001***Sanaag1.0350.0591.0290.992–1.0670.125Bari1.0120.551.0060.968–1.0450.766Nugaal1.0090.6311.0070.970–1.0450.732Mudug1.010.6081.010.972–1.0490.614Galgaduud1.0260.1781.0090.971–1.0480.648Hiraan0.9540.016*0.9680.932–1.0060.099Middle Shabelle0.9670.0830.9760.940–1.0140.221Banadir1.0410.023*1.0160.979–1.0540.407Bay1.087< 0.001***1.0370.989–1.0860.13Bakool1.072< 0.001***1.0851.046–1.126< 0.001***Gedo0.9910.6460.9870.949–1.0260.506Lower Juba1.0150.4530.9960.958–1.0340.82IRR Incidence rate ratio, CI Confidence interval. Significance levels: *p < 0.05, **p < 0.01, ***p < 0.001. An IRR > 1 indicates a higher rate of children ever born, whereas an IRR < 1 indicates a lower rate compared to the reference category Factors associated with the number of children ever born* IRR Incidence rate ratio, CI Confidence interval. Significance levels: *p < 0.05, **p < 0.01, ***p < 0.001. An IRR > 1 indicates a higher rate of children ever born, whereas an IRR < 1 indicates a lower rate compared to the reference category Bivariate analysis In the bivariate analysis, most factors were significantly associated with fertility. Regional disparities were also observed. Women in Togdheer (IRR = 1.08, p < 0.001), Sool (IRR = 1.09, p < 0.001), Bay (IRR = 1.09, p < 0.001), and Bakool (IRR = 1.07, p < 0.001) had higher fertility rates than those in Awdal, while women in Woqooyi Galbeed (IRR = 0.95, p = 0.013) and Hiraan (IRR = 0.95, p = 0.016) had lower rates. Sociodemographic characteristics were strong predictors. Urban women had a 2.0% higher fertility rate (IRR = 1.02, p = 0.005) than rural women. An inverse relationship existed between education and fertility; women with primary (IRR = 0.96, p < 0.001), secondary (IRR = 0.89, p < 0.001), and higher education (IRR = 0.16, p = 0.002) had lower fertility rates than those without education. Literate women had a 5.6% lower rate of cervical cancer screening (IRR = 0.94, p < 0.001). Marital status was key, with divorced women showing an 18.2% lower fertility rate (IRR = 0.82, p < 0.001) than married women did. Reproductive and economic factors were significantly associated. Women in the middle-wealth quintile had a 3.1% higher fertility rate (IRR = 1.03, p < 0.001) than those in the poor wealth quintile did. Media exposure showed varying effects: women with less than weekly exposure had a 7.4% higher rate (IRR = 1.07, p < 0.001), while those with no exposure had a 3.0% lower rate (IRR = 0.97, p = 0.001) than those with weekly or more exposure. A later age at first birth (> 20 years) was associated with an 8.9% lower fertility rate (IRR = 0.91, p < 0.001). Women who never used contraceptives had a 9.1% lower fertility rate (IRR, 0.91; p < 0.001). Having a female child was associated with a 2.2% lower rate (IRR = 0.98, p < 0.001) than having a male child. Multivariate analysis After adjusting for covariates in the multivariate model, the associations for most variables persisted, although some were attenuated. Regional disparities remained significant, with higher fertility in Togdheer (IRR = 1.07, p < 0.001), Sool (IRR = 1.08, p < 0.001), and Bakool (IRR = 1.09, p < 0.001) and lower fertility in Woqooyi Galbeed (IRR = 0.94, p = 0.005) than in Awdal. The protective effect of higher education was stronger in the adjusted model. Women with higher education had 84.6% lower fertility than those with no education (IRR = 0.15, P = 0.001). Literacy remained significant, with literate women having a 3.5% lower fertility rate (IRR = 0.97, p = 0.010). The association with residence was nuanced; urban women had a 1.9% higher rate (IRR = 1.02, p = 0.016), and nomadic women had a 2.1% lower rate (IRR = 0.98, p = 0.004) than rural women. Later age at first birth was associated with an 8.7% lower fertility rate (IRR = 0.91, p < 0.001). In the adjusted model, non-users who intended to use contraception later (IRR = 1.30, p = 0.013) and those who did not intend to use it (IRR = 1.30, p = 0.015) had higher fertility rates than current modern method users. The association with never using contraceptives remained inverse, showing an 8.2% lower rate (IRR = 0.92, p < 0.001). The relationship with media exposure persisted, with less-than-weekly exposure linked to an 8.0% higher rate (IRR = 1.08, p < 0.001) and no exposure to a 3.7% lower rate (IRR = 0.96, p < 0.001). Discussion This study analyzed the number of children ever born and its determinants among women of reproductive age in Somalia. The findings show the persistence of high fertility, with a weighted mean CEB of 6.22. The analysis reveals that fertility outcomes are shaped by individual, household, and regional factors, notably women’s education, age at first birth, and type of residence. The inverse relationship between educational attainment and fertility is the most significant determinant of fertility decline. Women with higher education have a CEB rate up to 84.6% lower than those with no schooling do. This aligns with the literature from sub-Saharan Africa and other high-fertility regions. Studies in Ethiopia and Ghana have shown that each additional level of schooling corresponds to a significant reduction in the number of children ever born [1, 13]. The mechanisms underlying this association are complex. Education enhances health literacy, providing women with knowledge of reproductive biology and contraception for informed family planning [5]. Education fosters female autonomy, equipping women with the confidence to participate in household decision-making and negotiate reproductive preferences [6, 14, 15]. Time spent in schooling creates opportunity costs for early marriage and childbearing, delaying motherhood and shortening the reproductive period [16, 17]. In the Somali context, where most women lack formal education, this factor represents a fundamental barrier to fertility transition. The analysis showed that a delayed age at first birth was strongly associated with lower lifetime fertility. Women who initiated childbearing after age 20 years had an 8.7% lower rate of CEB (Adj. IRR = 0.913), consistent with the demographic patterns in Kenya, Uganda, and Bangladesh [6, 17, 18]. Earlier childbearing extends a woman’s reproductive period, increasing potential lifetime births. In Somalia, this pattern is embedded in cultural norms favoring early marriage for girls, which is seen to secure family alliances and protect honor [19, 20]. Childbearing begins after marriage to demonstrate fertility and solidifies the union. Limited educational and economic options for adolescent girls make early marriage and motherhood the dominant path, perpetuating high fertility [13, 18, 21]. This is reinforced by divorced women having a 19% lower rate of CEB (Adj. IRR = 0.810). This association reflects the reduced pregnancy risk after marital dissolution, which is common in traditional societies [14]. The low prevalence of contraceptive use and intention (only 2.22% ever-users) indicates a critical gap in family planning services. The counterintuitive finding that women who never used contraception had a lower CEB rate (Adj. IRR = 0.918) is likely a statistical artifact of reverse causality. Over 93% of women reported no intention of using contraception, indicating a profound lack of demand. This is driven by a weak health system with limited access [10], widespread misinformation [7], and strong sociocultural and religious opposition that frames large families as a divine blessing [3, 11, 22–24]. Unlike patterns in stable development contexts, where urbanization typically drives fertility decline [1], our study found that urban women in Somalia show slightly higher fertility than their rural counterparts (Adj. IRR = 1.019). This seemingly paradoxical finding likely reflects the unique nature of urbanization in the country, which has been heavily influenced by the mass displacement of rural populations due to conflict and climate shocks. These internally displaced persons (IDPs) often settle in informal camps on the peripheries of urban centers, bringing with them the high-fertility norms and limited access to reproductive health services that are characteristic of their rural origins [25]. This phenomenon can create an ‘urbanization of poverty,’ where the urban aggregate fertility rate is inflated by this large, vulnerable subgroup [26]. Furthermore, our analysis distinguished between settled rural and nomadic populations, finding that nomadic women had significantly lower fertility (Adj. IRR = 0.979). This may be explained by factors such as prolonged breastfeeding, spousal separation due to migratory livelihoods, and nutritional stress, which affect fecundity [27, 28]. Thus, the urban-rural differential observed is not a simple inversion of the norm but a complex outcome of distress-driven urbanization and the distinct lifestyles of Somalia’s non-urban population. Policy and public health implications These findings have urgent implications for public health policies and development in Somalia. A multisectoral, rights-based approach is required [29]. First, prioritizing investments in female education is the most critical long-term strategy. Concurrently, there is an urgent need to rebuild the reproductive health infrastructure to ensure access to modern contraceptive methods, particularly through innovative delivery models for remote and nomadic populations. However, strengthening the supply is insufficient without addressing demand-side barriers. Therefore, culturally resonant health campaigns must engage community elders, religious leaders, and men to foster dialogue on the benefits of birth spacing for mothers and children [30]. Finally, programs that enhance women’s economic empowerment can provide them with the agency to exercise reproductive autonomy. Strengths and limitations The primary strength of this study is its use of a large, representative dataset, allowing generalizable conclusions. However, several limitations must be acknowledged. First, the cross-sectional design precludes the establishment of causality; only associations can be identified. We use language that reflects correlational rather than causal relationships. Second, the data relied on self-reporting, which is subject to recall bias for variables such as contraceptive use and age at first birth. Third, conflict in Somalia introduces data quality challenges: (a) insecurity likely prevented survey coverage in conflict areas, potentially leading to selection bias if populations have different fertility patterns; (b) high displacement rates may affect residence classifications and urban-rural differences; (c) excess mortality from conflict, particularly among men, may have altered household structures and marriage markets, affecting fertility but not captured in our models; and (d) displacement-related trauma and food insecurity may influence fertility behaviors and reporting accuracy. Fourth, the survey fails to capture certain paternal characteristics beyond education, such as the father’s age and occupation, which were omitted because of high non-response rates in the SHDS 2020 dataset. Furthermore, while the study tested proxies for women’s empowerment, such as employment and decision-making, these factors were not statistically significant in the final adjusted model, suggesting that their influence may be captured by more dominant socioeconomic factors, such as maternal education and household wealth. Fifth, our inability to distinguish between urban residents and recent migrants limits the interpretation of the urban fertility paradox. Sixth, the small sample of contraceptive users (2.22%) limits the statistical power and adds uncertainty to the estimates. Future research should prioritize: (1) longitudinal studies to establish temporal sequences and causal pathways; (2) qualitative research exploring male fertility preferences, couple communication, and the cultural meanings of childbearing in Somalia; (3) studies examining fertility differentials among displaced populations and by displacement duration; and (4) intervention research testing family planning and girls’ education programs in conflict-affected settings. Conclusion Fertility in Somalia remains markedly high and is shaped by a confluence of structural and behavioral factors operating at the individual, household, and regional levels. Women’s educational attainment emerged as the most influential determinant, with schooling appearing to delay childbearing, expand reproductive knowledge, and strengthen autonomy in family decision-making. Age at first birth, literacy, and marital status further contributed to lifetime fertility patterns, while the unexpectedly higher fertility observed among urban women relative to rural and nomadic counterparts likely reflects distress-driven urbanization and the concentration of displaced populations on the peripheries of cities. Persistently low contraceptive uptake, coupled with limited intention to use modern methods, signals a profound gap in both the supply and demand for family planning services. These patterns underscore that fertility transition in Somalia cannot be achieved through service delivery alone and requires a multisectoral, rights-based approach. Sustained investment in girls’ education stands out as the most consequential long-term lever, complemented by rebuilding reproductive health infrastructure and innovative delivery models that reach nomadic and displaced communities. Culturally resonant engagement with community elders, religious leaders, and men is essential to shift demand-side barriers, while programs that strengthen women’s economic empowerment may reinforce reproductive autonomy. Future work should prioritize longitudinal and qualitative inquiries into the pathways linking education, displacement, and fertility, alongside intervention research tailored to conflict-affected settings. Coordinated action across education, health, and social protection systems offers a credible pathway to improved maternal and child health outcomes and more equitable demographic progress in Somalia. Authors’ contributions MMA conceptualized the study. The study design was developed collaboratively by ASA, MMA, MAO, and YSAH. Material preparation and data collection were carried out by YSAH and MMA. MAO analyzed and interpreted the data. The first draft of the manuscript was written by ASA, MMA, and YSAH. All authors contributed to subsequent revisions of the manuscript and approved the final version. Funding The authors received no specific funding for this study. Data availability The data used in this study are from the 2020 Somali Demographic and Health Survey (SDHS). These data are available from the DHS Program upon reasonable request through the official DHS data access portal. Declarations Ethics approval and consent to participate This study used secondary data from the 2020 SHDS. The SHDS was conducted in accordance with established ethical guidelines and received ethical approval from the Somalia National Health Research Ethics Committee and the ICF Institutional Review Board. Informed consent was obtained from all participants before data collection. The present study involved secondary analysis of anonymized survey data and therefore did not require additional ethical approval. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement The data used in this study are from the 2020 Somali Demographic and Health Survey (SDHS). These data are available from the DHS Program upon reasonable request through the official DHS data access portal.