What does it mean when a country reports that 33% of women aged 20–24 were first married or in a union before turning 18? Does that number describe everyone’s experience?
In Nigeria, UNICEF data reports national child marriage prevalence of 33%. Yet the figure is 67% among the poorest households and 6% among the richest. That is a gap of 61 percentage points within the same country. A single national figure does not show this difference.
This story uses UNICEF data to explore three questions: How does child marriage prevalence differ by household wealth and place of residence? Do those inequalities tend to appear together? And has prevalence changed in the same direction everywhere?
Explore the interactive charts by hovering over points and comparing countries. The data come from different reference years, so cross-country comparisons should be read as descriptive rather than same-year rankings.
A national prevalence figure helps show the scale of child marriage, but it does not tell us how evenly that prevalence is distributed. The first chart compares national prevalence with the difference between the poorest and richest household groups across countries.
Source: UNICEF Child Marriage Data Portal, prevalence workbook (downloaded 2026). Country estimates have different reference years.
Nigeria is a useful example. Its national prevalence is 33%, while prevalence among the poorest and richest groups differs by 61 percentage points. Cameroon has a similar national figure of 30%, but its wealth gap is 46 percentage points. These estimates refer to different years (Nigeria 2024; Cameroon 2018), so the comparison illustrates differing patterns rather than a same-year difference.
What this reveals: Countries with similar national figures can have different levels of inequality within their populations. To see how widespread this difference is, the next chart compares the two household wealth groups directly.
The second chart places the poorest and richest household groups side by side. Each country’s gap shows how much child marriage prevalence differs between those groups.
Source: UNICEF Child Marriage Data Portal, prevalence workbook (downloaded 2026). Country estimates have different reference years.
Among the 111 countries with available wealth-group data, prevalence is higher among the poorest group in 104 countries. The median difference between poorest and richest groups is 22 percentage points. Nigeria has the largest observed gap in this comparison, at 61 percentage points.
The pattern is widespread, but it is not universal: four countries have higher prevalence among the richest group, and three have equal values. These exceptions matter because they remind us that a general pattern does not describe every country.
Household wealth is one way to examine inequality. Another is where people live. Does the same kind of difference appear between rural and urban areas?
The third chart compares child marriage prevalence in rural and urban areas. This adds a geographic dimension to the story, rather than looking only at household wealth.
Source: UNICEF Child Marriage Data Portal, prevalence workbook (downloaded 2026). Country estimates have different reference years.
Of the 120 countries with available rural–urban data, 105 have higher prevalence in rural areas. The median rural–urban difference is 10 percentage points. Seven countries show higher urban prevalence, while eight have equal values.
The median geographic gap is smaller than the 22-percentage-point median wealth gap reported in the previous section. However, the two calculations include different sets of countries, so this is a descriptive comparison rather than a direct measure of which factor matters more.
We have now seen two common patterns: higher prevalence among poorer households and higher prevalence in rural areas. The next question is whether countries with larger wealth gaps also tend to have larger rural–urban gaps.
This chart brings the two inequality measures together. Each point represents a country; the horizontal and vertical positions show its wealth and rural–urban gaps. The size of each bubble represents national child marriage prevalence, so larger bubbles indicate higher national prevalence.
Source: UNICEF Child Marriage Data Portal, prevalence workbook (downloaded 2026). Country estimates have different reference years.
Across 110 countries with complete data for this comparison, the Pearson correlation between the two gaps is 0.791. This is a strong positive linear association: countries with larger poorest–richest differences also tend to have larger rural–urban differences.
The median wealth gap in this matched group is 22 percentage points, and the median rural–urban gap is 9.5 percentage points.
An important limitation: Correlation does not establish causation. Wealth and residence may describe overlapping groups, and other social or economic factors may contribute to both patterns. The chart shows that the two inequalities are associated across the available country estimates; it does not explain why.
So far, the story has focused on differences across countries and population groups. But inequalities also exist across time. Has child marriage prevalence consistently fallen everywhere?
The final chart follows UNICEF historical prevalence estimates for five selected countries. These examples show that changes over time have not all followed the same direction.
Source: UNICEF Child Marriage Data Portal, historical trends workbook (downloaded 2026). Trend years and coverage differ by country.
The selected estimates show substantial declines in Maldives (51% in 1992 to 2% in 2017), India (53% in 1996 to 23% in 2021), and Ethiopia (69% in 1991 to 40% in 2016). In contrast, the selected series increase in Guyana (25% in 1995 to 32% in 2020) and South Sudan (32% in 1985 to 52% in 2010).
These are changes of −49, −30, −29, +7, and +20 percentage points, respectively. The countries have different starting and ending years, and the dates follow the year labels in UNICEF’s trend workbook. These lines should not be treated as identical survey periods or as evidence of what caused the changes.
What this adds to the story: A national figure can hide inequality at one point in time, while an overall impression of progress can hide very different historical trajectories.
The story began with one number: Nigeria’s national child marriage prevalence of 33%. Looking closer revealed a much larger difference between its poorest and richest households. Across the available UNICEF data, similar wealth and rural–urban patterns appear in many countries, and the two gaps are strongly associated.
The historical comparisons add a final reminder: progress is not uniform. Some selected countries show major declines, while others show increases across their available estimates.
The main takeaway is that national averages are a starting point, not the whole story. Understanding child marriage prevalence requires looking at differences within countries, the ways those differences overlap, and how patterns change over time.
Indicator. Child marriage prevalence refers to the percentage of women aged 20–24 who were first married or in a union before age 18, following UNICEF’s standard indicator definition.
Data source. The analysis uses two UNICEF Child Marriage Data Portal Excel workbooks: a prevalence dataset and a historical trends dataset. The prevalence workbook supplies national, wealth-group, and rural–urban estimates; the trends workbook supplies historical estimates for the selected countries.
Preparation. The R script imports the Excel sheets, reshapes the historical year columns, converts relevant fields to numeric values, and excludes missing or invalid observations for each comparison. The wealth gap is calculated as poorest-group prevalence minus richest-group prevalence; the rural–urban gap is calculated as rural prevalence minus urban prevalence. Gaps and changes are reported in percentage points (pp), not percent change.
Coverage. The wealth comparison includes 111 countries, the rural–urban comparison 120, and the joint inequality analysis 110. These counts differ because data availability varies by variable. The historical chart displays five selected countries to contrast different trajectories; these examples are not intended to represent every country.
Analysis and interactivity. Charts were created in R using ggplot2 and Plotly. Interactive hover information helps readers inspect country-level values. The relationship between the two inequality gaps was summarised using Pearson’s correlation coefficient (r = 0.791; 110 countries).
Limitations. The data are observational and include estimates from different reference years. Country coverage is incomplete and differs across charts. Correlation does not show causation, and differences between wealth or residence groups should not be interpreted as the effects of those characteristics alone. Historical trend year labels are used as supplied in the UNICEF workbook; the underlying estimation method should be consulted before making claims about specific survey or cohort dates.
UNICEF. (n.d.). Child marriage data portal: Data centre. https://childmarriagedata.org/data-centre/
OpenAI. (2026). ChatGPT [Large language model]. https://chatgpt.com/
I acknowledge the use of ChatGPT (OpenAI, 2026) during this assignment. I used ChatGPT for technical assistance with RStudio, R Markdown and Plotly, and for reviewing grammar, clarity and the interpretation of my results. I checked the suggestions against the UNICEF datasets and my R outputs before finalising my work.