Common Orphan Statistics in Context – Part 2

In part 1, I looked into the sources of some commonly used statistics on orphaned children, specifically:

  1. “140 million children are orphans”
  2. “80% of children in orphanages have a living parent”
  3. “2.7 million children are in orphanages”

Here’s three more for part 2. This time I went so deep, I had to request two actual books from a physical library!

4. “There are 8 million children in orphanages worldwide”

Last time I looked into the claim that there are 2.7 million children in orphanages, but a big thanks to Rob Oliver and Leigh Mathews for pointing out that 8 million is more frequently cited as the total.

This Lumos video highlights the 8 million figure

Leigh said that the figure comes from Pinheiro’s 2006 World Report on Violence Against Children for UNICEF, where it says, “By some accounts, as many as eight million boys and girls around the world live in institutional care.” You’ll find it on p. 183 (p. 14 of the PDF) here in chapter 5.

This “account” is Save the Children’s 2003 report A Last Resort: The Growing Concern about Children in Residential Care.

In turn, this report cites the “8 million” figure as coming the 1995 book Roofs and Roots: The Care of Separated Children in the Developing World by David Tolfree.

And then THIS book cites ANOTHER source in a footnote: Defence for Children International, 1985, “Children in Institutions”, DCI, Geneva.

I haven’t gotten my hands on a copy of this…report? book? vinyl record?…um, product, yet. (I emailed DCI but didn’t hear back.) But wow, what a rabbit hole! It’s certainly not what I was expecting: it turns out that this data is at least 33 years old. Definitely not relevant to today, as the children counted back then are now adults, and there’s no reason to assume institutionalization rates stay constant over time.

While 2.7 million could easily be an underestimate, since it relies on government-reported data that probably overlook unregistered orphanages, this 8 million figure seems not to be based on any recent, reliable, or even accessible information at all. I would prefer to stick to 2.7.

5. “For orphans who age out, 10-15% commit suicide before age 18, 60% of girls become prostitutes, and 70% of boys become criminals”

These numbers are ubiquitous. I had an extremely hard time finding their original source, because orphan care NGO after orphan care NGO seemed to copy and paste the same numbers on their websites without any references! As I searched, there were a lot of broken links and dead ends along the way, so I began to realize that the original data must be quite old…

A UNICEF report (p. 89) eventually led me to a 1996 book by Judith Harwin (a physical book!). It says on p. 147:

The Russian Procuracy General found that one in three care leavers becomes homeless, one in five ends up with a criminal record, while as many as one in ten commits suicide.

Her citation is a 1994 article in the Rossiyskaya Gazeta–which is apparently too old to be cataloged in online databases, and though a hard copy seems to exist in my city, I’m not sure it will be possible to use Google Translate on a newspaper microfilm. I can’t find the Procuracy General report, either, unfortunately–but it clearly predates the mid-1990s.

So, again, these data are just too old to apply to the current day. When I asked care leaver and children’s advocate Vadim, he agreed that far too much has changed in the former Soviet Union since twenty years ago for survey data like these to be relevant to today’s care leavers.

But perhaps other research exists that NGOs can use to help garner support for their programs. I would follow the following guidelines when selecting studies to use:

  • Only look at studies specific to your context. If you want to say something about orphanage alumni in China, do not look to research on care leavers from Eastern Europe.
  • Consider how old the studies are. Have significant political or societal changes happened since the research was conducted? I wouldn’t use data more than 15 years old (but the newer, the better).
  • Review the sampling. Look for phrases like “nationally representative” or “generalizable.” In order for a study to be generalizable (i.e., the findings apply to the whole population of care leavers from that country, not just the sample of kids who participated in the survey), the participants should be selected randomly (or as close to randomly as possible) and the sample should be somewhat large.

6. “For every 3 months a child spends in an institution, they lose 1 month of development”

Dr. Delia Pop makes the claim in this video

I hear this one sometimes from adoption, deinstitutionalization, and orphanage improvement advocates. It’s a very effective way to communicate the harm of severe deprivation. But where does it come from, and to what can it be generalized?

I did a search for this phrase in articles and reports and dug into their citations. At first I mis-remembered the quote as children losing one month of brain development, but it turns out this is not the case.

The research seems to come from Dana Johnson, a well-known international adoption pediatrician and researcher. In one book chapter he wrote:

The adverse effects of the orphanage environment on linear growth are remarkably consistent worldwide. Whether a child is institutionalized in Romania (Ames, 1997; Johnson et al., 1992), the Former Soviet Union (Albers et al., 1997), or China (Johnson et al., 1996; Johnson & Traister, 1999), analysis of growth data from a variety of orphanage systems indicates that children lose one month of linear growth for approximately every 3 months spent in institutional care.

(If you’re interested, here’s a figure.)

First, by linear growth, he means height, not brain development or any other measure of development. When you take this research and phrase it as simply “children lose one month of development…” it sounds much more compelling than “children lose one month of height”!

Here is proof from an NGO newsletter that lots of people think the statistic is about brain development! (From April 2018)

Second, all five studies are of children after they have been adopted to North America. Children chosen for international adoption might be categorically different than the larger population of children in orphanages–maybe they’re healthier, maybe they’re smaller, who knows…

Third, though, I appreciate that the data comes from multiple countries (I hate when data from one country is used to generalize to the whole Global South), it’s still just Eastern Europe plus China–no data from Africa or Latin America.

And finally, and most importantly, it’s just so old…many of these kids’ heights were measured before I was born. There have been very significant changes in Eastern European and Chinese child welfare systems in the past 25 years and I don’t think we can assume this trend still holds true.

The way Hope and Homes for Children used this statistic in the above video, it’s like they are saying, “We can assume today’s baby homes in Uganda are the same as the Chinese and Eastern European institutions of 25 years ago.” That’s a valid hypothesis, sure, but there’s just little reason to assume it’s true without evidence.

So what should NGOs say instead? Well, I would be satisfied if they just prefaced this factoid with, “Studies of Eastern Europe and China in the 1990s found that…” Then the listener is at least a little more equipped to decide if the statistic is relevant to the issue at hand.

Edit [April 16, 2018]: By happy coincidence, I met Dr. Johnson at the Rudd Adoption conference a few days ago. I asked him about this statistic and he said that yes, it refers to linear growth…But his team did do another study about developmental milestones, which was never published. They found that children were delayed about 1 month in milestones like rolling over and walking for every 3 months they were institutionalized. However, he said, these children experience catch-up growth in middle childhood and so the negative effects are not sustained. So I still stand by my original conclusion.

In Conclusion

A lot of (non-peer-reviewed) reports just cite figures that were included in a prior report, even though report had gotten its figure from an even older report, and so on, without anybody realizing the original data was collected decades ago!

This happens for one major reason: Governments collect far too little data on children living outside parental care. As Lumos put it, “All children count, but not all children are counted.”

If there’s one thing that everyone working in the alternative care space can agree on, it’s that our field desperately needs newer and better data collection and research studies. So we have to press donors and governments to budget for it!

2 thoughts on “Common Orphan Statistics in Context – Part 2”

  1. I can only share experience from Thailand where we have been hunting and documenting unregistered children’s homes for 3 years now. A breakdown between provincial registration of homes and central collation of these registrations has also meant that Thailand’s quoted “137 registered private children’s homes” (UNICEF 2015) is now known to be very wrong. Chiang Mai province alone has now registered 52 private children’s homes and has 30 further homes pending (Chiang Mai’s share of the 137 figure was just 17 homes) . There are 76 provinces in Thailand plus the Bangkok special administrative area.
    Through online searching we have documented 242 unregistered homes, although anecdotally we know here are many small homes with no web foot print (our team of 15 web searchers only found 5 of the 18 homes local to us, non are registered).
    So for Thailand at least I would expect the number of children in institutional care to be at least three times any number currently shared by the government (although must add that right now all parties are working to dig into the real picture, including government).

    1. Wow, that’s really something. The country-level mapping of unregistered institutions that you guys did is SO important. Maybe a next step for the Patrowski research (the article resulted in “2.7 million”) would be to compile mapping data like yours and estimate how many children live in unregistered institutions around the world… Then we could have a much more accurate number – and more recent than 1985. 🙂

Leave a Reply

Your email address will not be published. Required fields are marked *