Donald Trump’s administration recently ended US government funding to the Demographic and Health Survey (DHS), despite its value for policy-makers and many scientists. But the DHS will probably continue in some form, and now is a good time to consider improvements.
By Frank Götmark
Started by the USAID (United States Agency for International Development) in 1984, data from the DHS survey have been used by thousands of scientists over the years. It was developed from its predecessor, the World Fertility Survey (where data for some countries, such as Bangladesh, go back to 1975). The impressive program covers at least 90 developing countries and comprises more than 350 surveys. They are repeated periodically, allowing study of trends in for example age of marriage, fertility, family size, child and maternal mortality, health and diseases (e.g. malaria).
Some UN organisations, the World Bank, and some countries have helped finance the DHS. But for over 40 years USAID was the primary funder and hosted the database. Its dissolution earlier this year was a hard blow, raising protests among researchers and population experts. The Gates foundation has stepped in to secure availability of data for a short period of time, and to complete some on-going surveys and analyses.
For family planning (FP) in developing countries, the fertility level and women’s and men’s desire for children are of special interest. Is desire for childbearing falling, and to what extent? Many studies have used the DHS to seek answers. Perhaps surprisingly, for many African women the so called ideal family size is still high. In a study from 2020, Ahinkorah and co-authors analysed DHS data from 32 countries in Sub-Saharan Africa (SSA), including 233 000 married and cohabiting women. They found that 20 % of them preferred 0 – 3 children, 37 % preferred 4 – 5 children, and 44 % preferred as many as 6 or more children. Women in rural areas, and women who did not take decisions on family size alone, desired more children.
In 2025, Balya and co-authors included 167 500 married women from 23 countries in SSA, also using DHS data. Of these women, 62.5% desired more children. Higher gender equality in household decision-making, and reproductive health autonomy, correlated with reduced desire for additional children among women. However, the authors state that ”these effects were notably small in magnitude”, and that policy-makers need to address ”deeply rooted socio-cultural factors”.
Factors influencing fertility
The total fertility rate (TFR) in Africa has declined since about 1980, but the mean value is still 4.1, and 4.3 for Sub-Saharan Africa, compared to 1.9 and 1.8 for Asia and Latin America and the Caribbean, respectively (UN Population Division 2024). In SSA, TFR is influenced by many factors which vary by circumstances and region, according to Church and colleagues (2023). Historically, family planning (FP) programs have been successful in reducing TFR, especially in Asia and Latin America, but also in some African countries. Bongaarts (2020) concluded that the reduction of TFR in countries in SSA between 1989 and 2019 was associated with strength of FP program and degree of female education, with the FP program effect slightly larger than the education effect. Yet the mechanisms underlying such results are often unclear and hard to study. One probably important influence is cultural norms for childbearing.
Norms in family and communities
The DHS does not always interview men but when they do, they find that men report on average higher ideal family size than women. According to Church et al. (page 10), men in SSA are more pronatalist and prefer larger families than women, while women, usually the primary caregivers, are more aware of the amount of work it takes to raise children. In 2022, Sarnak and Becker found that in 21 of 32 countries, it was more common for the husband to want more children, while in 29 countries, over two-thirds of spouses were concordant in their fertility preferences. The authors suggest that males should continue to be included in the DHS, and also suggested other improvements to the survey.
Women in SSA report high ideal family size in the DHS. For instance, Odusina and co-authors reported that the ideal family sizes for men and women in Nigeria were 7.2 and 6.1, respectively (DHS data, 2018). In the Democratic Republic of Congo, 43% of the women preferred 6 or more children (Mosuse & Gadeyne, DHS, 2014).
Why do women in SSA prefer large families? It could be that they independently want many children and large families. Alternatively, their answers to the question of ideal family size in the DHS may mainly reflect external socio-cultural norms regarding childbearing. According to the Nigerian study above, ”women are expected to cooperate with their husbands and recognise their views as a mark of respect and obedience. They are not expected to assert their views on family matters but that of their husbands.” This is a generalization of course, for a large population living under varying conditions. Church and colleagues also found evidence that having many children brings social status and prestige, as well as influence and respect, for men and women.
In a paper from 2017, Aisha Dasgupta and Partha Dasgupta refer to such norms as ”socially embedded preferences” that help sustain high fertility rates. They argue that the basis on which women’s expressed desire for children is elicited in the DHS survey misestimates their desire and underestimates their true need for family planning, because their expressed preferences are influenced by the reproductive norms around them. The authors suggest how the survey may be improved to show family size norms that are presently hidden. (Their reasoning is in a long paper that lacks an abstract and conclusion, perhaps a reason why their suggestion has not been discussed much.)
In the DHS, women are asked the question “If you could go back to the time when you did not have any children and could choose exactly the number of children to have in your whole life, how many would that be?” This is an interesting question, but the survey does not ask what a woman’s desire would be if the prevailing fertility practices of others in the community were different back in time. The question is not conditioned on the behaviour of others, and does not clarify individuals’ desires in different situations. The Dasguptas therefore proposed a set of questions where n would stand for family size of, say, 5, 4, or 2, as follows: “If you could go back to the time when you did not have any children and could choose exactly the number of children to have in your whole life, how many would that be, assuming everyone else in your community had n children over their whole life?”
The answers to such a set of questions would be interesting and relevant for the DHS. The questions may seem unwieldy, and probably require extra thought on the part of the respondents in the survey. Therefore, to begin with, a test survey should be conducted separately from the DHS, to obtain preliminary results and knowledge about how to best pose questions in the DHS. Google Schoolar lists 2200 articles that deal with ideal family size and DHS.
Cultural norms and fertility should be studied broadly, including also questions about the role of religious norms for childbearing (see a new review of this subject by Turner and Götmark). The African continent (especially SSA) ranks highest in TFR and degree of religiosity among the global regions (Götmark & Andersson), and religious leaders are influential. Ideally, questions about social norms for childbearing in test surveys should separate three influences: men (spouse), religious leaders, and the community at large.
The DHS is a form of monitoring, common in some fields such as environmental work where it is mainly used for ’early warning’. The ecologists Nichols & Williams emphasized that monitoring should be combined with experiments wherever possible to reveal cause-and-effect. There is usually strong separation in funding for monitoring and science (funding comes from different sources). But given extra funding for research, the DHS could be used experimentally to also analyse the role of family planning (FP) where some areas are subjected to more FP information and services, others remain as conventional controls, and areas are compared before and after the intervention. Bongaarts (2020) summarize such research, but more is needed, especially in Africa.
Hopefully, the DHS will be fully funded again in the future, and also expanded and revised in line with the suggestions above.
I thank Malte Andersson, Stan Becker, Aisha Dasgupta, Partha Dasgupta, Rhoda Mundi, and TOP colleagues for useful suggestions and comments on this blog.
