Dynamic Fertility Preferences in Nigeria: A Bayesian Spatio-Temporal Analysis of DHS Data, 2003–2024

Background: Fertility preferences are an important indicator of reproductive intentions and demographic change, but national averages can obscure substantial geographic and life-course differences. This study examined how women's preferences for additional children have changed across Nigeria and whether these changes vary by age, individual characteristics, and geographic location between 2003 and 2024. Methods: Data were obtained from five nationally representative Nigeria Demographic and Health Survey rounds conducted in 2003, 2008, 2013, 2018, and 2024. The analysis included married or cohabiting women aged 15--49 years. Fertility preference was defined as a binary outcome indicating whether a woman desired another child. A Bayesian hierarchical binomial model was fitted using Integrated Nested Laplace Approximation. The model incorporated nonlinear age effects, survey-year effects, individual and household characteristics, and spatial variation across states using the Besag--York--Mollié 2 model. Posterior estimates and 95\% credible intervals were used to assess associations and changes over time. Results: Fertility preference declined substantially at the national level by 2024, although considerable variation remained across states. The posterior odds of desiring another child were lower in 2024 than in 2003, while differences between some intervening survey years were less pronounced. Women with more living children had lower odds of desiring another child. Higher educational attainment and greater participation in household decision-making were also associated with lower fertility preference, whereas Muslim women had higher odds of desiring another child. Fertility preference varied systematically across the reproductive life course, with a nonlinear age pattern. Spatial estimates showed persistent geographic clustering but also substantial changes in the relative level of fertility preference across states between survey rounds, indicating that the fertility transition has not occurred uniformly across Nigeria. Sensitivity analyses showed that the main substantive and spatial findings were broadly robust to alternative prior specifications. Conclusions: Nigeria has experienced a substantial decline in fertility preference over the study period, but the transition remains geographically uneven and varies across the reproductive life course. These findings suggest that national fertility trends can mask important subnational differences and support geographically targeted reproductive health planning, continued investment in female education and women's empowerment, and improved access to family-planning services.