Divorce Hotspots Revealed: Tanzanian Study Maps Where Marriages Fall Apart

Marriage in Tanzania is more than a personal bond; it is the backbone of family life, economic cooperation and social identity across a nation of more than sixty million people. Yet a new nationwide analysis has revealed that roughly one in seven ever-married Tanzanian women has experienced divorce or separation, and that the risk of a marriage ending is far from evenly spread across the country. The study, published in Discover Psychology by Deogratius Bintabara of the University of Dodoma’s Department of Community Medicine and Ramadhani Mohamedi of the university’s College of Informatics and Virtual Education, is among the most comprehensive attempts to chart the prevalence, drivers and geography of marital dissolution in a sub-Saharan African setting. Drawing on data from more than ten thousand women, the researchers combined sophisticated statistical modelling with spatial mapping to show that where a woman lives may matter almost as much as who she is.

The evidence base for the study is the 2022 Tanzania Demographic and Health Survey and Malaria Indicator Survey, a nationally representative exercise conducted with a stratified two-stage cluster sampling design. From this survey the researchers extracted responses from 10,837 ever-married women aged 15 to 49 years, a sample large enough to support both fine-grained subgroup analysis and reliable regional estimates. The outcome of interest, marital dissolution, was defined to include both divorce and separation, capturing the full spectrum of unions that had ended short of widowhood. Because the survey records the region of residence of each respondent, the dataset also allowed the team to move beyond national averages and ask a question that few demographic studies in the region have answered with such precision: are marriages more fragile in some corners of Tanzania than in others?

The headline finding is stark. Fifteen percent of the ever-married women surveyed reported that their most recent or only marriage had dissolved. That figure may appear modest against global comparisons, but the regional breakdown tells a more dramatic story. Geospatial mapping revealed pronounced disparities, with the highest concentrations of dissolution clustered in the Southern zone of the country. In Lindi, more than a quarter of ever-married women, 25.56 percent, reported a dissolved union, while neighbouring Mtwara recorded 23.84 percent. These southern coastal regions stand in sharp contrast to parts of the country where dissolution rates remain well below the national average, a gradient that the authors argue should inform the design of locally targeted family support and social policy interventions rather than one-size-fits-all national programmes.

To understand what lies behind these numbers, the researchers employed multilevel mixed-effects logistic regression, a statistical technique designed for data with a nested structure. Women in the survey are clustered within communities, and communities within regions, so ignoring this hierarchy would risk underestimating uncertainty and misattributing effects. The mixed-effects framework simultaneously models individual-level predictors and community-level context, allowing the team to estimate adjusted odds ratios, the strength of each association after accounting for all other variables in the model, together with 95 percent confidence intervals. The approach also yields the median odds ratio, a measure of how much the odds of dissolution vary between communities purely because of unmeasured contextual differences, and the team checked for multicollinearity among predictors using variance inflation factors before settling on the final specification.

The individual-level results are striking in their consistency with demographic theory. Each additional year of a woman’s age was associated with a small but robust increase in the odds of dissolution, with an adjusted odds ratio of 1.05 and a confidence interval of 1.04 to 1.06, reflecting the simple fact that older women have lived through more years in which a marriage could end. Lower educational attainment emerged as a significant correlate, with women who had never attended school or left early facing elevated odds compared with their more educated peers. Education, the authors note, is widely understood to operate through multiple channels: it expands women’s bargaining power within households, improves access to information and services, and reshapes expectations about marriage itself.

Perhaps the most eye-catching individual determinant was employment. Women who worked for cash had nearly three times the odds of marital dissolution compared with those who did not, with an adjusted odds ratio of 2.73 and a 95 percent confidence interval of 2.23 to 3.35. The association demands careful interpretation. Cash employment can give women the economic independence to leave unsatisfactory or abusive unions, so higher dissolution among employed women may partly reflect empowerment rather than distress. Conversely, financial strain and long working hours can themselves destabilise marriages. The cross-sectional design of the survey cannot disentangle cause from effect, and the authors are careful to frame the finding as an association that flags a group deserving of attention rather than proof that earning money breaks marriages.

Crucially, the study demonstrates that the forces shaping marital stability do not stop at the individual. Community-level household wealth and urban residence both influenced the likelihood of dissolution, with urban women nearly twice as likely to experience a dissolved union as their rural counterparts. This urban penalty is consistent with a broader literature linking city life to weaker kinship oversight, greater anonymity, higher living costs and alternative social norms around marriage and separation. The community wealth gradient suggests that the socio-economic environment in which a woman is embedded, not merely her own household’s resources, shapes the durability of her marriage. The persistence of community-level variation, captured by the median odds ratio, reinforces the authors’ argument that interventions aimed solely at individual women will miss part of the picture.

The spatial analysis adds a dimension that conventional regression alone cannot provide. By testing for spatial autocorrelation, the statistical tendency of neighbouring areas to display similar values, the researchers confirmed that marital dissolution is not randomly distributed across the map but clusters in identifiable geographic patterns. The Southern zone hotspot, encompassing Lindi and Mtwara, stands out on the generated maps, and such clustering implies that unmeasured regional factors, whether cultural, economic or historical, are shaping local marriage dynamics. Spatial demography of this kind is increasingly recognised as a powerful public health tool: identifying hotspots allows governments and non-governmental organisations to direct counselling services, legal aid, economic support and family planning resources to the places where family instability is most concentrated, maximising impact per unit of investment.

Why does this matter beyond Tanzania? Marital dissolution carries documented consequences for women’s economic security, children’s educational outcomes, mental health and housing stability, and it is rising across much of sub-Saharan Africa as urbanisation, education and labour market change reshape family life. The Tanzanian findings speak directly to that transition. The strong associations with cash employment and urban residence suggest that as more women enter the paid workforce and more families move to cities, patterns of marital stability will continue to shift, potentially in regionally uneven ways. Policymakers across the region, the authors argue, should therefore integrate family stability considerations into broader social and public health strategies, from girls’ education to urban planning, rather than treating divorce as a purely private matter.

The study’s methodology also sets a benchmark for future demographic research in the region. By pairing multilevel modelling, which respects the hierarchical structure of survey data, with geospatial techniques that expose geographic clustering, Bintabara and Mohamedi have produced an analysis that is both statistically rigorous and directly actionable. The use of de-identified, publicly available DHS data meant that no additional ethical approval was required, and the underlying survey itself obtained informed written consent from all respondents under approvals from Tanzania’s National Institute for Medical Research, the Zanzibar Medical Ethics and Research Committee and ICF International’s Institutional Review Board. The authors declare no competing interests and received no specific funding for the analysis. Their message to Tanzania is ultimately constructive: marital dissolution is a measurable, mappable and modifiable phenomenon, and with the right spatially informed policies, the regions where families are most vulnerable can be identified, understood and supported before instability takes its widest social and health toll.

Subject of Research: Prevalence, determinants and spatial patterns of marital dissolution among ever-married women in Tanzania

Article Title: Prevalence, determinants and spatial patterns of marital dissolution among ever married women in Tanzania using population-based survey data

Article References: Bintabara, D., & Mohamedi, R. (2026). Prevalence, determinants and spatial patterns of marital dissolution among ever married women in Tanzania using population-based survey data. Discover Psychology. https://doi.org/10.1007/s44202-026-00920-3

Image Credits: AI Generated

DOI: 10.1007/s44202-026-00920-3

Keywords: marital dissolution, Tanzania, demographic and health survey, divorce, separation, multilevel modelling, geospatial analysis, spatial autocorrelation, women's employment, urbanisation, sub-Saharan Africa, family stability