Hide table of contents


I have been working on a research project into the scale, tractability and neglectedness of child marriage. After 80 hours of research, I thought that there was a relatively strong case that effective altruist funding organisations that fund projects addressing international poverty should consider funding child marriage interventions. I then came across a study that destabilised the premise that child marriage leads to decreased wellbeing across a number of health metrics. I describe in more detail my experience and findings below, and share some tips for those undertaking self-directed research projects to avoid making the mistakes I made (skip to ‘What I will do next time’ for these).


 I had no direct experience researching child marriage, but I was interested to learn about effective interventions and whether it had potential as a possible cause area. I studied Political Science and International Relations at University, as well as some subjects on development, gender and economics, I have also worked as a government evaluator. My goal was to do some preliminary research and determine if child marriage was large scale problem, tractable and neglected. If so, I would share this research with effective altruist funders.

My model:

In October last year, I started a self-directed research project into the scale, tractability and neglectedness of child marriage. I read and collected dozens of sources, analyzed data, contacted a top researcher, compared effective interventions, built a mental model of what the charitable space looked like and identified potential interventions for EA support.

I came to the following findings, based on around 80 hours of research:

  1. Scale/importance
    1. Child marriage is a widespread practice that affects around 12 millions girls per year (UNICEF, 2022)
    2. Child marriage is a harmful practice that increases the risk of negative maternal and sexual health outcomes, domestic and sexual violence, and reduces the likelihood that a girl will complete school. This is the consensus position held by global development institutions (see meeting report from leading global institutions on child marriage UNFPA, 2019). 
  2. Tractability 
    1. There are cost effective interventions that work to prevent child marriage, e.g. the ‘cost per marriage averted’ ranged between US$159 and US$732 in this study (Erulkar, Medhin and Weissman 2017).
    2. The effect of child marriage on quality adjusted or disability adjusted life years has not been quantified so it difficult to compare cost effectiveness with other interventions (EA Forum explainer on these metrics).
  3. Neglectedness
    1. Population Council is a research body focussed on running quasi-experimental programs and creating scalable interventions (Population Council, date unknown). 
    2. The lead investigator into child marriage at Population Council informed me that it is not currently running programs to prevent child marriage because of lack of funding.
  4. Conclusion
    1. Effective altruist funding organisations focused on international health and poverty should consider funding effective interventions to prevent child marriage at a large scale.

What broke my model

Earlier this week, I decided it would be useful to try and quantify the harm of child marriage, or at least some of the harms, into commonly used metrics like quality adjusted or disability adjusted life years (QALYs or DALYs). I anticipated that this would be a key piece of information for EA funders, and it had not been done so far (finding 2b). In doing so, I came across a study that fundamentally challenged finding 1b: child marriage is an underlying cause of many harmful outcomes. Without strong evidence that child marriage causes harm, the other findings on its tractability and neglectedness are significantly less consequential.

Fan and Koski, 2022 examined the data from 58 studies on the effects of child marriage on a range of health outcomes (see my summary Table 1). They found that while data clearly shows that child marriage increases the risk of physical violence, results from other metrics, including impacts on contraceptive use, maternal health, nutrition and mental wellbeing are mixed (Fan and Koski, 2022). Some studies even showed that child marriage was associated with more positive outcomes, such as higher contraceptive use (Fan and Koski, 2022). It is unclear why results are so mixed, but worth noting that these studies take place over many different countries and cultural contexts across Sub-Saharan Africa and South Asia.

Additionally, Woden et al., 2017 reviewed a series of studies into decision-making power and found that it is unclear whether child marriage leads to lessened decision-making power.

This is not to say that child marriage is not harmful, it is linked to:

Girls married before the age of 15 are a less studied cohort than girls under 18. The few studies that look at girls married before 15 as a separate group indicate negative impacts. While limited, studies that show that the following health risks are greater for girls married before the age of 15 for: 


Table 1 below summarizes the information I have collected on the harms of child marriage so far.

Table 1

HarmGirls married before the age of 15Girls married before the age of 18Sources
Physical violence Increased riskIncreased riskFan and Koski, 2022 consistent results across eight studies. One study showed no effect but this study only looked at a period of three months (Erulkar, 2013).
Sexual violenceUnclearUnclearFan and Koski, 2022, citing seven studies. Some showed increased risk and some showed no effect.
Unwanted pregnancyUnclearUnclear

Fan and Koski, 2022, citing seven studies: some found increased risk, some found decreased risk, some found no effect and some found variable effects in different countries. 

Girls married before 15 studied separately by only one study. This study found there was increased risk for this group (Kamal, 2013). 

Stillbirth and miscarriageUnclear



Fan and Koski, 2022, citing 2 studies. One found increased risk, one found no increased risk for 15-17 year olds but increased risk for girls married before age of 15 (Paul, 2018).
Use of contraceptionUnclearUnclearFan and Koski, 2022, citing 15 studies. Some found increased use, some found decreased use, some found no effect. One study showed girls married under 15 less likely to use contraception (Habyarimana F, Ramroop, 2018).
Use of maternal health careUnclearUnclearFan and Koski, 2022, citing nine studies. Some showed no effect and some showed decreased likelihood of using maternal health care. Consistent across studies was the decreased likelihood of giving birth in a healthcare facility. However, this was not adjusted for the potentially confounding variable that most child marriage occurs in rural areas where health care facilities are often far away.
Nutritional statusUnclear



Fan and Koski, 2022, citing six studies. Some found increased likelihood of malnutrition, others found no effect. One study found decreased likelihood of malnutrition. One study found increased likelihood of malnutrition for girls married before 15 but no effect for those married between 15 and 18 (Yimer, 2016).
Completing less school school than peers Increased riskIncreased riskConsistent across three studies (Delprato et al., 2015Lloyd and Mensch, 2008Field and Ambrus, 2008.)


Updated positions and next steps

My updated views on child marriage after doing this research are that: 

  1. The causal model that child marriage leads to harmful outcomes, except for the metrics of increased risk of physical violence and decreased school completion, is not stable (see Table 1).
  2. I should direct time into measuring:
    1. How much child marriage leads to physical violence and decreased school years completed, and attempt to quantify this harm through the most appropriate metric (e.g. QALYs or DALYs).
    2. The harms to girls married before the age of 15, and attempt to quantify this harm into the most appropriate metric (e.g. QALYs or DALYs).
  3. If the findings of 2 or 3 are significant, I should research the tractability and neglectedness these problems.
  4. It may be better to look outside of the causal model of child marriage at more specific harms and attempt to address these directly. For example, this study found that no more than 20% of girls who dropped out of school in francophone Africa did so because of marriage or pregnancy (Lloyd and Mensch, 2008).

What I know so far about scale and tractability: marriage before 15

UNICEF estimates that around 5% of women aged 20-24 alive today were married before the age of 15 (UNICEF, 2021). The Population Council demonstrated that providing educational materials to girls at a cost of $20 per girl reduced the risk of girls being married. The adjusted risk ratio was 0.09 (91% decreased chance) with a 95% confidence interval of (0.01, 0.71) or 99% to 29% decreased chance and (significance: p<0.05). However, the results of this intervention were not consistent in Burkina Faso and Tanzania (Erulkar et al., 2017). I plan to review further studies that measure interventions to delay marriage targeted at girls under 15.

What I will do next time

I am glad I undertook this project. I learnt lots of things about child marriage, taught myself some statistical concepts, and things that I will do differently next time to save time and more quickly develop my model.


Before committing time to a research project:

  1. List my underlying assumptions (I likely would have listed 1b)
  2. Search for sources that disagree with these assumptions (I would likely have found Fan and Koski, 2022 within the first 5 hours of doing this project, and disrupted 1b)
  3. Consider whether those assumptions still hold. 

Consensus and data

It is important not to automatically trust a position made by large institutions: 

  1. As soon as I see a claim being made, no matter which authority is making this claim, trace the claim back to its underlying data (I would have traced UNICEF and UNFPA’s claims about 1b back to their source material).
  2. Thoroughly examine the methods and results sections of studies and come to my own conclusions based on the data before reading any discussion or conclusions (I would have found that 1b was weaker than I anticipated, and worked to refine my research questions). 

Find someone to report on my projects to 

I think the act of reporting my process and findings to someone else, and having them challenge my assumptions, would probably have led me to challenge my assumptions faster, and saved me lots of time. I will be reaching out to fellow EAs to fulfill a supervisor/challenger role for future self-directed projects. Also, if you have a project, I am happy to play this role for you. My email address is catherinefist@gmail.com.


If you found this post interesting, you may also like:

How science works and how best to read scientific papers

I really enjoyed listening to Spencer Greenberg and Christie Aschwanden speak about how slow and hard determining even the most simple things using scientific methods is in this podcast. They also spoke about how they read scientific papers, which I found useful.

EA Forum post on violence against women and girls 

If you are interested in the broader topic of reducing gender based violence, I recommend this recent post from Akhil: What you can do to help stop violence against women and girls - EA Forum

A promising maternal health charity start-up 

Sarah Hough and Ben Williamson launched the Maternal Health Initiative (MHI) last year through Charity Entrepreneurship. MHI which aims to increase access to family planning in Sub-Saharan Africa. They have a robust theory of change and I am excited to see the impact they make.


Thanks :)

Thanks to RP and TF who reviewed this post for me and gave me wonderful feedback.






More posts like this

Sorted by Click to highlight new comments since: Today at 1:51 PM

I really appreciate this post, thanks for sharing it (and welcome to the Forum)! 

Some aspects I want to highlight: 

  1. The project — trying to translate the known (or assumed) harms from child marriage into the metrics used by related projects that might work on the issue — seems really valuable
  2. Noticing that a key assumption falls through and sharing this is great. I'd love to see more of this
  3. The post also outlines some learnings from the experience
    1. Write out key assumptions and test them / look for things that disprove them
    2. Avoid trusting consensus
    3. Get accountability / find someone to report to
  4. I also like that there isn't the sense that this is the last word on whether working on child marriage is a promising cause area or not — this is an in-progress write-up (see "updated positions and next steps") and doesn't shy away from the fact
  5. And there's an "if you find this interesting, you may also like" section! I'm curious if you've seen: 
    1. Giving What We Can's report from 2014 on this issue? (And the associated page, which also seems pretty outdated.)
    2. Introducing Lafiya Nigeria and the Women's health and welfare and Family planning topic pages. 

Quick notes on the model — I'd be interested in your answers to some questions in the comments (Jeff's, this one that asks in part about the relationship between economic growth (and growth-supporting work) and this issue, etc.). 

  • I skimmed this report on some programs, and in case anyone is interested, it seems: 
    • "In each study country, we tested four approaches: 1) community sensitization to address social norms, 2) provision of school supplies to encourage retention in school, 3) a conditional asset transfer to girls and their families, and 4) one study area that included all the approaches."
  • I'm immediately a bit worried that estimating the impact of these programs is more messy if e.g. one of the harms that stem from child marriage that you track is a loss in education (or loss in nutrition or something) — as presumably e.g. the school supplies program also just directly supports education (so there's potentially some double-counting).
    • (I'm also wondering if, assuming that education delays marriage, more effective education-support programs, like iron supplementation, are just the way to go here.)
      • In general, it seems like there might be a bit of circularity (or, alternatively, loss of information)q if we do something like: "ok, these interventions, which we evaluate on a given factor — how much they delay (child) marriage — are effective to [this degree] at achieving the particular thing we're measuring, which we think is important for [a number of factors

I made a sketch to try to explain my worry about the models (and some alternative approaches I've seen) — it's a very rough sketch, but I'd be curious for takes. 

1-4. Thanks for the positive feedback Lizka 🙂

5. Thanks for sharing those links.
 a) No I had not seen that! I did not know that any EA orgs had done a profile on this issue. 
b) Great to know about those threads, I am exciting to read and learn more in this space.

Thanks for visualising this, I found it useful. I think that I need to better understand what the harms are and how they are caused to evaluate the best model.

Right now, the third model makes most sense to me. This is because the negative outcomes associated with child marriage occur outside of child marriage as well and it would be better to address the outcome as a whole, not just where it is happening in the context of child marriage. 

Your process for working through this is really clear, and I admire your willingness to change your mind after sinking tremendous time into the cause exploration. But I have to admit that I am puzzled that the cited evidence moved your beliefs that much.

  • Physical violence seems a priori like the most important factor in that list of harms, and is clearly negatively associated with child marriage.
  • Mixed/noisy evidence on other outcomes is certainly something to update on, but not with the strength that you have. Noisy evidence says more about the limitations of studies (sample size, respondents being sensitive because this is a frought issue) than about the absence of an effect.
  • You interpret the increased use of contraception as a positive outcome, but that looks like a negative outcome to me, because it seems like a proxy for having more (likely coercive) sex at a young age. It also seems like a second-order benefit compared to the first-order harms of coercion and violence.

More generally, we have to combine evidence with our prior beliefs, and we update less from evidence when our prior beliefs are stronger. In general I think we should have a strong prior belief that child marriage is bad - we are not totally clueless. So this doesn't change my beliefs about child marriage very much.

Hi Karthik, thanks for engaging with my post. I think my post expresses two things: 

a) the shock at finding evidence late into the project that had the potential to undermine other findings 

b) my updated my position in light of new evidence. 

I think I may have accidentally over-expressed a) to demonstrate the importance of looking for evidence that disagrees with positions early when researching topics. To clarify, my position now is definitely not that child marriage is not harmful. 

My updated position is that the range of metrics across which child marriage is definitely and reliably   harmful is narrowed. Unfortunately, with mixed and noisy results it is hard to know what is poor research design and what is lack of effect size. I still think that at the end of this project I will find that there  is a good  case that child marriage should be prevented, based on metrics I have not looked at yet,  physical violence and the likelihood that girls will complete less school. 

If I was trying to determine whether child marriage was harmful or not harmful, I would probably come back with a clear yes. But I have focused on exactly working out what exactly what  harms are taking place because it has implications for the interventions that would be most effective, e.g. whether we focus on delaying the age of marriage to later teenage years because younger marriages are a lot worse or ending all marriages under 18 because all child marriages are equally harmful.  

I take your point on the use of contraception, I do not know the rates of sexual activity in marriages or outside for these populations, but I would assume girls in marriages are having more sex, earlier.

I will further consider your view that perhaps I have not weighted my prior beliefs highly enough.  I think I am hesitant about these views because I know that beliefs about when someone should get married and have children are very culturally determined, e.g. my grandmother got married at 19, which is shockingingly young to me, but was normal and what she wanted at the time.

This goes back to my take about how the problem here is actually LMIC governance, but it seems trivially true that child marriage is objectively very bad and  also relatively bad compared to the opportunities that these girls should have, but it might also be true that it's not so bad  compared to the very limited opportunities they actually do have (albeit probably still negative). But the pathway to getting rid of child marriage seems clear: just improve governance & economic growth rates, and the problem will take care of itself as the  economic returns to delaying marriage grow and grow. That seems much more tractable than some sort of mass cultural transformation. 

Hi Sabs, thanks for engaging with the post. I would be interested to see the interventions available to improve governance and economic growth rates, and their cost effectiveness. The interventions to prevent child marriage are relatively cheap (like providing school supplies or conditional asset transfers) and have the benefit of immediately targeting the issue. They also help with economically uplifting families.

I notice you have a table collecting and assessing possible harms from the practice but no similar table collecting and assessing possible benefits. In deciding whether to fight against some practice shouldn't we want to figure out the net effect - benefits minus costs - rather than just costs?

Given how widespread the social phenomenon is, surely there must be some benefits?

( Something something Chesterton's fence...)

Near as I can tell, the people who think it's terrible are in large part motivated by largely-false quasi-Mathusian claims related to "overpopulation". If we set those aside, younger brides tend to have more kids; all else being equal we should assume those kids have lots of extra QALYs (that wouldn't otherwise exist) and also presumably make their parents happy. Are those married as children happier adults on average than those not? How do we balance a claimed higher risk of physical abuse against, say, a lower risk of ending up childless or alone or financially insecure?

Hi Glenra, thanks for engaging with the post. This is something I had not considered at all, and something I will consider for my QALY calculation. More kids may also worsen the economic situation of a family already living in a poor, rural place, and may not increase the happiness of parents. There is not available data on whether child marriage reduces the risk of never getting married, but there is data on the number of births (lots of it referenced here if you are interested Fan and Koski, 2022). Thanks for giving me challenging food for thought.

Thank you for this writeup. I enjoyed reading it.

As someone who is pretty convinced by the capability approach , one thing that I feel is somewhat missing from this (well done) exercise is a consideration of the foreclosing of options that a child marriage entails for the girl. Even if the girl grows up happy and doesn't have the problems that you sought out measures for (schooling, experiencing violence), her life options may have been quite massively curtailed by being  entered into this hard-to-break legal and social arrangement before she was an adult. I think that alone makes it bad.

I'm not saying I have a way to cost this out in terms of capabilities, but I think this consideration merits attention. My guess is when a lot of us think about what is wrong with child marriage we start with intuitions around "losing options for life" but then our training or norms guides us to things that are easier to measure like "school completion." That's not necessarily bad, but I think it would be a mistake to see no effects on the latter and conclude that the former didn't happen.

Again, thanks for posting this and thanks for the work that went into it.

Thanks for reading and engaging with my post Ryan! 

Thanks for raising the capability approach here, I had not heard of it before. Thinking about how child marriage curtails life prospects was one of my personal motivations for looking at this topic in the first place. I was reading The Precipice at the time and thinking about the parallel with the future of humanity being curtailed by existential crises. 

I agree that just because losing options for life  is somewhat unmeasurable does not mean it is not important. Take this example of Stella, interviewed in a BBC podcast. Stella is a 22 year old woman who as a girl who was helped by a grassroots program in Kenya. She experienced a forced marriage and had a son when she was 11.  She is now at university and has a drastically different life, and according to her much better life, than if the program had not intervened.

I do not have a way to include this in my model but I will keep thinking on it. Some things I am  considering related to this: 

1. Divorce rates are really high (similar to US) in some contexts where there is child marriage, e.g. in Ethiopia, so child marriage can be escaped. Obviously if you have children things are very complicated. Accessibility of family planning can help keep options open for girls to divorce.

2. Many things limit the possibility of people's options, like being in poverty or growing up in a rural place. These things would have to be considered as confounding variables.

3. I generally have the view that being married would curtail choice, but others may believe that marrying early allows more choice, e.g. better choice of partner within a community, better access to resources and freedoms that come from being an adult in a household rather than a child in your parents' household. This is a fraught line of reasoning and I am very unsure here. This speaks to some of the criticisms that you note in your post on the capability approach about it being culturally biased towards liberalism.

Thanks for looking into this! This is super interesting.

A couple of thoughts:

  1. I wouldn’t expect child marriage to be bad primarily because it is bad for girls’ health (although it is hard for me to imagine becoming pregnant before 15 isn’t linked to higher maternal morbidity, as you note). Child marriage seems bad because aspects of being a child bride/mother seem intrinsically hard/undesirable. Here’s an example of what I mean: I assume many married children are having sex that isn’t very consensual, since their partners likely are not chosen by the children themselves and many are quite young. I agree that DALYs can indirectly convey the mental health harms of, e.g., non-consensual sex, but also think non-consensual sex is bad independent of whether it is well captured by DALYs. (I assume “sexual violence” is referring to something different?)

  2. The idea of assessing whether a pregnancy was “wanted” in a kid who grew up in a culture that permits child marriage seems somewhat fraught. Presumably, most of these cultures place a high premium on women having kids (and probably a lot of kids, hence the child marriages). This isn’t to say that we should completely ignore kids’ preferences (adapted preferences are still preferences!), but I don’t think we should take them at face value, either. I’d be eager to see more qualitative research on child brides, since this could help clarify nuances here.

  3. My prior would be that outcomes for kids married before 15 and kids married between 15 and 18 would be different. You also suggest before the table that there is some evidence for this. My reaction to the “unsures” in the table for kids < 15 is thus “we’d probably find an effect if there was more data,” but I’m curious whether you drew different conclusions.

  4. And lastly, one of the cited studies mentions that there is a relationship between child marriage and fertility (i.e., child brides have more kids and start earlier). I didn’t read the whole paper, but if this is true, it seems potentially quite important, because it likely has implications for socioeconomic development. Preventing child marriage strikes me as a relatively non-problematic way of reducing fertility (which I know not everyone thinks is good) and deferring fertility (which definitely does seem good—having a 15 year old mother probably isn’t great for the kid).

Thanks again for this really interesting and thoughtful report!

Hi Lily,

Thanks for your considered and thoughtful response, it touched on a lot of thoughts I have had throughout this project.

1.  My intuition from the outset of this project that child marriage would be intrinsically bad because it would expose children to unwanted sex and pregnancy. I have tried to leave that intuition to one side as much as possible because I know ideas about gender and adulthood/childhood are pretty culturally entangled.

In saying that,  my ideal model of harm would definitely take into account whether someone wanted to get married and psychological wellbeing. Unfortunately, the available data does provide a lot of clarity.I will look into more qualitative data on what girls think about their marriages to address this. 

In the meantime, here are some survey results from Erulkar, 2017, who is a researcher at Population Council on this topic (Table 1). You will see that the results change dramatically between the different groups, some marriages being wanted and some being unwanted, arranged and not arranged, when they found out about the marriage, etc. (Erulkar, 2017). Bear in mind that the Tanzania group had the lowest n, and that the sample sizes in general are not that large. Also note that these are pretty much raw results as percentages, no upper or lower limits provided.

Table 1: What girls surveyed think about their marriages (Erulkar, 2017)

Psychological wellbeing

Fan and Koski, 2022 summarise the data on psychological wellbeing as:

  • one US based study and one small Iranian study showing decreased psychological well-being for girls married before 18
  • a study from Niger and Ethiopia showing decreased wellbeing for girls married before 16 but not 16-17 (Table 2)
  • a study from Ghana that found child marriage protected against measures of stress.

In the study from Niger and Ethiopia, it suggests a pattern of decreased psychological wellbeing for girls married at 12, 13, 14 for both countries, and 15 and 16 for Ethiopia (Table 2)(John et al., 2019). However, please note the standard deviation for these approximations is more than 20 points for each. For me this says:

  • child marriage may be worse the younger you are when you are married
  • the harms of child marriage are likely to be drastically different depending where its happening
  • available data has a lot of uncertainty.

Table 2: Psychological well being

How sexual violence is measured

I agree with you that measures of sexual violence do not address all non-consensual sex and this is a harm that should be included in the overall model of harm of child marriage. I will look for some data on surveys about non consensual sex more broadly as part of this project going forward, unfortunately early signs are that this likely has not been measured. Fan and Koski, 2022 note that all 16 studies they considered in their review on this topic either use the Demographic and Health Surveys (DHS) data or word their questions very similarly to ask about sexual violence:

h)  physically force you to have sexual intercourse with him when you did not want to?

i)  physically force you to perform any other sexual acts you did not want to?

j)  force you with threats or in any other way to perform sexual acts you did not want to?

I think there is a pretty strong argument that if you were a child married to an adult and you did not believe they were allowed to say no, you could experience a lot of non-consensual sex without answering yes to any of these questions.

2. Yes I agree it’s incredibly fraught! I will look for more qualitative data on this and provide it in this thread when I have it.

3. I have the same view that child marriage under 15 would likely show to be much more harmful compared to getting married post 15 if it was broadly studied, but this is only my opinion. Hopefully I should be able to run this data myself using the DHS data.

4. Yes, I left out fertility on purpose because there is disagreement about whether reducing fertility is good or bad, and I do not have a firm view on this. I think increasing access to family planning is an even more non-problematically good thing to do regarding fertility. This is what the Maternal Health Initiative is working on in Sub-Saharan Africa.

In this conference recording, Erulkar talks about the potential of delaying child marriage as a way to increase the gaps between generations, as governments in Sub-Saharan Africa are (in her view) struggling to keep up with the number of young people and provide enough services. This is something I have not explored yet.

I do, in general, think that preventing child marriage could have positive socioeconomic outcomes. One of them is if it increased the education of girls. A popular and cheap intervention for child marriage is to fund girls’ education through supplies or paying school fees. Erulkar, 2017 costed providing school supplies at $20 per girl per year. Girls and women increasing education has positive effects on countries’ economic prosperity (World Bank, 2018).

Thank you again for responding so thoughtfully to my post! Also thank you for introducing me to the term 'non-problematically'! Very useful in this space where pretty much everything is somewhat problematic.

Thanks so much for this really thoughtful response!

I appreciate how carefully you've reviewed the data, and how open-mindedly you're approaching this issue. I think your assessments largely seem reasonable, but we might diverge here:

My intuition from the outset of this project that child marriage would be intrinsically bad because it would expose children to unwanted sex and pregnancy. I have tried to leave that intuition to one side as much as possible because I know ideas about gender and adulthood/childhood are pretty culturally entangled.

I agree that ideas about gender and adulthood (and marriage, the value of tradition, etc) are informed by our cultures, personal experiences, political views, and so on. And when it's possible to study an issue rigorously and reach meaningful results without relying on priors, we should do so. My worry here is that we've both identified challenges inherent to researching this issue (e.g., that many girls who are having non-consensual sex will not screen positive on existing surveys). Given this, I think it'd be a mistake to draw conclusions solely on the basis of existing data. In my view, we have many independent reasons to think child marriage is bad and not a lot of reasons to think it is good, and the data doesn't move me much in this regard (i.e., I'm not being confronted with a lot of information that makes me think child marriage is good; it seems like most existing data suggests it is bad or unclear).

I'm also struck by the differences between Burkina Faso, Ethiopia, and Tanzania in the table you shared (e.g., 76% of girls in Ethiopia did not want to get married, and 75% of girls in Tanzania did!). This makes me think that child marriage might be a much more important cause in some places than in others, which would be consistent with other GHD initiatives, for which there is often substantial variation in cost-effectiveness for the same kind of program across contexts. Thank you again for all of your work on this!

I think you make a really good point that our priors should lead us to seek out missing data points to either confirm or dispute these priors. This is a really useful meta-conclusion  for this project, thank you for highlighting it.

Yes the differences are striking, also a lot of the data I have looked at so far is in Sub-Saharan Africa, where there is already a lot of diversity in findings, but the legal and cultural landscape in South Asia is also completely different and I contains a lot of diversity also. As you say,  I think this is more than a question of ITN for child marriage but ITN for child marriage in x place, knowing that the levels of harm, how common child marriage is, and how tractable it is are likely to vary greatly.

Thanks again for your engagement, it's really helped my thinking!

Presumably, most of these cultures place a high premium on women having kids (and probably a lot of kids, hence the child marriages). This isn’t to say that we should completely ignore kids’ preferences (adapted preferences are still preferences!), but I don’t think we should take them at face value, either.

I agree there is a lot of instinctive force behind this; it definitely seems like some preferences are in some sense 'illegitimate' e.g. this). However, it also seems like a potentially fairly symmetrical argument. Should we also decline to take people's preferences for fewer children at face value if they grew up in a place that de-emphasizes family, or education-related preferences if someone grew up in an are that emphasized schooling? Without a general theory of what makes a preference legitimate it seems hard to object in specific cases.

Thanks for sharing this Catherine! (and congrats on your first forum post! woop!) 

I really appreciate you sharing your views so openly and where you found you were wrong - and your steps for next time sound really great!  I'm curious if you've come across Charity Entrepreneurship / their resources on intervention prioritization during your research?

Also, I think it would be really cool if there were group intervention-prioritization projects where people could investigate causes they think are neglected / promising. 


(Finally, on small grammar thing: when I read the summary I was confused by this sentence:

I then found a source that undermined a key premise: child marriage is clearly harmful across a number of health metrics. 

I might phrase the last sentence to be more clear that the source gave evidence against child marriage having clearly harmful health effects. At first I read it the opposite and was confused. )

Thanks Vaidehi! 🙂

I am generally familiar with Charity Entrepreneurship but had not seen the page you linked. This will be super useful as I continue this project and other future projects, thank you.

I agree, that group would be awesome!

Thanks for the wording pick up, I have rewritten the sentence so it is hopefully clearer now.

This is a great post Catherine, thanks for sharing your findings!

Thanks Kirsten :) great to be in a community to openly discuss ideas in good faith.

Super interesting. Well done for this research and well done for changing your mind based on the evidence, especially given how much time you dedicated to this!

(These kinds of posts are super important, and we should think of these like we think of 'negative' / statistically insignificant results in science - the incentives to publish them are not very strong and we should encourage these kind of posts more )

Thanks for the positive feedback :) .  I agree, I think most research conducted finds very little, because proving things is really hard. This Clearer Thinking podcast that I linked in the original post goes into this more deeply for anyone interested in why it's so hard.

Thanks for this, really valuable work! 

I'm curious what the qualitative description of child marriage usually looks like in these cases- I have two very rough mental images: 

  1. A 14 year-old girl learning very little at school/ barely attends school. She's very unlikely to continue studying past the age of 16. Her (very low-income) parents struggle to continue supporting her and would rather she married earlier to reduce their burden and make a bit of bridewealth money (maybe to concentrate resources on another child). She gets married, her husband takes on responsibility for her (he might be more responsible/ caring than her parents), and her life outcomes don't change much from if she were to get married at 17. 
  2. A 14 year-old girl is learning quite a lot at school. She dreams of going to college/ sixth-form/ university and could even afford to if she got a part-time job, but family/ cultural pressure leads her to get married early. She has a child at 15, is forced to stay in her village, and all of her plans go to waste.  

Could it be that people like to imagine something more like the second scenario when the first is more common? 

I think the "family/cultural pressure" to marry is very likely to be downstream of an unexpected pregnancy as a result of rape. I have never seen any estimates for for the percentage of girls in SSA for whom sexual initiation (for want of a better phrase) comes through rape, but anecdotally I wouldn't be surprised if it were over 70% in many countries. Again "child marriage" is not the problem here, the problem is likely a) weak growth = less reason to delay marriage and build human capital and b) massive, endemic sexual violence that is just absolutely everywhere.

Thanks for looking into this and writing up what you found!

I'm confused why "contraceptive use" is an evaluation criterion ? What harms does this cover that "unwanted pregnancy" doesn't?

Thanks for engaging with the post Jeff :) 

My understanding is that contraceptive use is a positive outcome in the prevention of girls/women contracting STIs from their husbands in this context. Child marriage puts girls at risk of contracting STIs in two ways:
- The marriage is polygamous. This survey is a very small sample size, but provides a sense that at least in Burkina Faso and Tanzania, some marriages are polygamous. 
- As the men girls marry are older than them, they are likely more sexually experienced than the girls and have a higher risk of already having STIs and passing them on to their wives.

Erulkar, 2017


contraceptive use is a positive outcome in the prevention of girls/women contracting STIs from their husbands in this context

Wouldn't that only be true for condoms, and condoms aren't that common? Ex:

Injections, implants and pills were found to be the most common contraceptive methods used by women in this study. There were few women who used condoms (either male or female) for contraception, female sterilization, or who reported using traditional methods of contraception. -- Contraceptive use and discontinuation among women in rural North-West Tanzania

Separately, even condom usage seems like it wouldn't be a great proxy for how likely these girls/women are to get STIs, since marriage likely reduces their number of partners?

Ah good point! I was assuming they meant condoms. I will rethink this as a metric.

I agree with Lizka's comment, "(I'm also wondering if, assuming that education delays marriage, more effective education-support programs, like iron supplementation, are just the way to go here.)

  • In general, it seems like there might be a bit of circularity (or, alternatively, loss of information)q if we do something like: "ok, these interventions, which we evaluate on a given factor — how much they delay (child) marriage — are effective to [this degree] at achieving the particular thing we're measuring, which we think is important for [a number of factors

Anything that is proven improve the education of girls may well be a great place to  start on the potential intervention front.

I agree, child marriage would probably be only one outcome we would be aiming to address. The broader picture might be an effort to increase girls' education through a simple and scalable intervention. 

I really enjoyed this post. Small note: The links in the "If you found this post interesting, you might also like" section are not working.

Hi Tsunayoshi, thanks for reading the post. Oh no! I just clicked them all and they worked for me... are they still not working for you? Let me know and I can investigate the problem further.

This was a really neat write up.  As someone trying to do research myself (without much experience) I found it useful and I might take you up on the offer to help one day.

My intuition is that addressing the harms themselves is a more effective approach (i.e. violence and ceased education), but it is  just an intuition.

There is a typo here: "The adjusted risk ratio was 0.09 (81% decreased chance)..." should be 91% - an effect size that would be amazing if true, but does seem unlikely unfortunately.

Hi Scott, great to hear from someone else doing research without much experience. Would be great to bounce ideas off each other and review each other's work in future. My email is catherinefist@gmail.com :) 

Yes that is also my intuition at this stage, as the negative outcomes of child marriage also occur outside of that context and we would want to address the whole problem, not just where it occurs within child marriage.

Thanks for the typo pick up! I have edited the post.

Very interesting seeing your process for all this, thanks for laying it out.

On the tractability and neglectedness questions, how do you account for other interventions that impact child marriage rates? I'd assume that programs aimed at general economic growth and education also raise average marriage ages

Thanks for engaging with the post :) 

I think Lizka's comment  visualises this phenomena well. I think the answer would probably be to be targeting the negative outcomes themselves rather than child marriage.

This is an example of a general phenomenon I don't really understand: the lack of attention to how to make poor nations rich.  If we can figure that out, a lot of plausibly-bad phenomena like child marriage will just disappear of their own accord. I would bracket all kinds of developing world health and social interventions into this cluster of "things that are vastly less important than improving developing world governance". I don't remotely understand what the point is of mental health interventions like StrongMinds, in particular. I'd also be quite depressed if my government was as dreadful as most governments are in sub-Saharan Africa, and even the best mental health treatment surely can't last for long given the stark reality. I'm not saying that EA should pivot into figuring institutional mechanisms for Amazon to buy out Zimbabwe, but it does seem a vastly more effective cause area than most other things you could spend your time on in this field.

  • I would bracket all kinds of developing world health and social interventions into this cluster of "things that are vastly less important than improving developing world governance"

This is why we have the ITN model, right? Improving governance is super important, but it seems to fail the tractability test most of the time. Trying to improve institutions with government officials who profit from the existing, more extractive institutions is a massive, intractable, relatively thankless task. Also, outsiders are in a worse position to create institutional change, so you have to spend a lot of your time and resources just trying to get a place at the table. For the neglectedness part, loads of people are dedicated to trying to make poor nations rich- it's been the objective of development economists/ IMF/ World Bank people/ many of the smartest people in developing countries way before RCTs and micro-interventions came into fashion. If you can find neglected sub-areas, that's obviously great, but low-hanging fruit seems rare. 

  • "I don't remotely understand what the point is of mental health interventions like StrongMinds. I'd also be quite depressed if my government was as dreadful as most governments are in sub-Saharan Africa, and even the best mental health treatment surely can't last for long given the stark reality..."  

The fact that people have a valid external reason to be depressed doesn't mean that it's pointless treating their depression. People in poor countries can have decent mental health, as any life satisfaction survey should be able to demonstrate. From solely first-order effects, if you think their data is valid, it's probably a better way of improving lives than most Give Well interventions. There are also some second-order, or 'trickle-up' effects for most micro-interventions. The human capital theory of development argues that culture, education, health etc. lead to greater productivity, which leads to a more educated and mobile populace, which leads to better institutions and leads to growth. This piece explains this part of the debate- should also note that randomistas think that their interventions are robustly good, even if they cause comparatively little growth.   

  • " I'm not saying that EA should pivot into figuring institutional mechanisms for Amazon to buy out Zimbabwe, but it does seem a vastly more effective cause area than most other things you could spend your time on in this field." 

I agree that we should be considering ambitious institutional interventions (SEZs / charter cities, or at least something like growth diagnostics), but this one is surely a non-starter. It's hard enough getting a low-income country to marginally decrease agricultural tariffs, let alone selling your whole country to Amazon.

Responding just to the comment about StrongMinds – I think mental health is an incredibly complicated issue, and mental illness is very multi-factored, so even if some people in sub-Saharan Africa are depressed due to bad governance, others may be depressed due to reasons that mental health services would alleviate. In any event, the fact that depression in sub-Saharan Africa is not even remotely close to 100% means the statement "I'd also be quite depressed if my government was as dreadful as most governments are in sub-Saharan Africa" is basically a non sequitur. 

Agree that improving economic growth in LMICs + international wealth redistribution would be effective in solving lots of social problems in LMICs, but both are highly intractable in my opinion, so would probably not solve a specific social problem more cost-efficiently than a targeted intervention aimed at that social problem. 

(But FWIW, I don't think improving economic growth in LMICs and international wealth redistribution are so intractable that they have no place in the EA movement)

and yet AI alignment is apparently tractable whereas "improve LMIC govenance" isn't? EA confuses me sometimes. We have a hypothetical solution to a hypothetical problem vs concrete solutions to concrete problems -we just need to figure out the implementation!

I think AI beats LMIC governance on scale and neglectedness in the ITN framework, so would deserve greater attention from EA even with equal tractability

we've already done massive historic governance interventions in LMICs and still do through more indirect means today, so the idea that we can't more efficiently intervene and this is some massive intractable problem is so for the birds IMO. This logic speaks more to the inherent dislike and distrust of power within EA than anything else IMO.

Would be interested in historical examples of this, and also on elaboration on what the indirect means today are.

(I think philanthropic funding of economic policy research in India pre 1991 would be one example?)

the historical examples I had in mind are various empires, or "empire moments" as such as MacArthur in Japan.

Today the IMF is a reasonable effective cudgel for institutional reform, but I don't think it would take much to expand its operations and make them more ambitious both on the level of the cash it lends and the degree of involvement in recipient governance that it has.