Reading the final report of the Munro Review of child protection, my attention was caught by what turned out to be a minor typographical error. The last sentence of paragraph 5.8 appears to refer to reference 95, the Royal Society paper Brain Waves Module 2: Neuroscience: implications for education and lifelong learning, but in fact cites reference 94. Reference 94 is an independent report commissioned by the current UK coalition government, published in January 2011, written by Graham Allen, Labour MP for Nottingham North and entitled Early Intervention: The Next Steps.
Early intervention is summed up as follows by the ‘Early Intervention Review Team’;
“Early Intervention is an approach which offers our country a real opportunity to make lasting improvements in the lives of our children, to forestall many persistent social problems and end their transmission from one generation to the next, and to make long-term savings in public spending….” (p.vi)
In 2008, Graham Allen had written another paper on early intervention, this one co-authored with Iain Duncan Smith and entitled Early Intervention: Good Parents, Great Kids, Better Citizens published jointly by the Centre for Social Justice (CSJ) and the Smith Institute. Iain Duncan Smith is a former leader of the Conservative Party and currently Secretary of State for Work and Pensions. In 2004, he founded the Centre for Social Justice, a centre-right think tank, just after his period of party leadership ended. The Smith Institute is a left-leaning think tank set up in 1996 in memory of the former Labour Party leader, John Smith. Ed Balls, later to become the Labour government’s Secretary of State for Children, Schools and Families, worked for the Institute between 2004 and 2005.
In previous posts I’ve complained (at some length) that the model of child development being used by children’s services pays little attention to recent biological research. It would be unfair to suggest that biology is entirely absent however. There are many references to physical development in this literature, Aldgate et al’s book contains a chapter on genetic and biological influences and the Munro report cites the National Research Council’s From Neurons to Neighbourhoods: The Science of Early Childhood Development – chapter 8 contains a comprehensive summary of brain development. The second chapter of each of the two documents written by Graham Allen is also dedicated to brain development. It’s these two chapters I want to concentrate on in this post.
I’m not entirely clear why it was deemed necessary to refer to brain development in papers about early interventions intended to forestall social problems. After all, few people would want to see evidence from brain scans before they could be persuaded that sanitation, a balanced diet or education have good outcomes for individuals and for the population as a whole. And given the technical problems with brain scanning and the interpretation of the resulting images, there are other more reliable ways of measuring the effectiveness of interventions. Allen and Duncan Smith’s Early Intervention: Good Parents, Great Kids, Better Citizens justifies the inclusion of material on brain development as follows;
“We make no apology for presenting, as laymen, a considerable body of medical evidence in this chapter. When economic resources are under intense pressure, and facing strong claims from well-established programmes and special interests, we believe that this medical evidence points overwhelmingly in favour of a shift to Early Intervention.” (Good Parents p.45)
What’s the evidence?
The first few pages of chapter 2 of Good Parents focus on results from three large-group, longitudinal studies purported to show that early adverse childhood experiences result in later health risks such as smoking, alcoholism, illicit drug use, obesity and high level promiscuity (Good Parents p.54). At first glance, the conclusions presented are persuasive, but when you look a little more carefully, the picture isn’t quite so clear-cut.
I’ve mentioned several other documents that in some cases refer to each other. To clarify how they are linked, I’ve mapped out the connections here:
Two of the studies, Farrington and West’s Cambridge Study in Delinquent Development, a Prospective Study of South London Males From Ages 8–32 and the Dunedin Multidisciplinary Health and Development Study are prospective – that is, they started with children and have periodically sampled their health, development and behaviour over many years.
According to Allen and Duncan Smith the Cambridge study showed that adult offending could be predicted in childhood (Good Parents p. 51). That’s not quite what the study records. The 2006 report, which tracked the participants up to the age of 48, found that there were predictive factors in childhood for adult offending. In other words, some factors were predictive of behaviour for a particular group, not for particular individuals. The highest correlation between childhood factors and persistent offending was for children having a convicted parent or sibling. What this means is that children with previous offenders in their families are more likely to offend, not that offending can reliably be predicted in individual children. A significant number of children from families with an offender didn’t commit crimes, whereas some children from non-offending families did.
The Dunedin study looked at the health and development of 1037 babies born in Dunedin, New Zealand, between 1972-73. Data from the Dunedin study has been used in over 1000 publications but I couldn’t find which one Allen and Duncan Smith were referring to. They claim that nurses’ assessments of which of a group of 3 year-olds were at risk, predicted criminal convictions, violent behaviour and domestic abuse at age 21. They conclude;
“the fact is that children who are likely to have poor outcomes, including adult criminality, can be identified at age three when they are still riding their tricycles.” (Good Parents p.52)
Not exactly. As Allen and Duncan Smith themselves point out, not all of the at-risk children offended, and some of the not-at-risk children did – 18% exhibiting violent behaviour and almost 10 % abusing their partners (Good Parents p.51).
The third study, the Adverse Childhood Experiences (ACE) Study, by contrast, is retrospective; it relies on self-reports about childhood maltreatment, family dysfunction and health status, and therefore on that notoriously unreliable data source, human memory. You can read the questions that were posed to participants in Preventing child maltreatment: a guide to taking action and generating evidence. It’s published jointly by the World Health Organization and ISPCAN, The International Society for the Prevention of Childhood Abuse and Neglect. My curiousity about Preventing child maltreatment: a guide to taking action and generating evidence was initially piqued by the title. I have no problem with taking action against child maltreatment, but do have concerns about ‘generating evidence’. Evidence is usually ‘gathered’ or ‘found’ – implying that it’s already out there, researchers just have to go and look for it. ‘Generating evidence’ suggests that, like Bettelheim, your case might not actually have strong evidence behind it so you need to create some.
I was also concerned by a reference in the Foreword to the idea that
“the traditional “privacy barrier” between the domestic and public spheres has inhibited the evolution of policies and legal instruments to prevent violence within the family and provide services for those affected by it.” (p.vi)
I’d predict that the prohibition of violence is as likely to be effective as the prohibition of alcohol consumption, but that violence might be lessened if its causes were to be addressed. Furthermore, the ‘traditional “privacy barrier”’ isn’t about “privacy” – a relatively recent development in human history – but about protecting the individual from the abuse of power by the state. I’m sure the author, who’s had extensive experience with the UN, is aware of that. But I digress.
I also had concerns about Box 1.1 (p.8). It’s entitled Child maltreatment and damage to the developing brain and is adapted from a pamphlet published in 2001 by the National Clearinghouse on Child Abuse and Neglect Information and the National Adoption Information Clearinghouse called In Focus: Understanding the Effects of Maltreatment on Early Brain Development. The pamphlet lists 31 references, 11 by Bruce Perry and one by Allen Schore. In 13 pages, Schore’s work is cited 17 times and Perry’s 40 times. These names crop up again in the papers by Allen and Duncan Smith.
My concerns about Box 1.1 and the In Focus pamphlet weren’t so much about what they said, as about their emphasis. Firstly, maltreatment and neglect of children is, by definition, harmful – that’s why they are called maltreatment and neglect. We already know that certain practices cause harm to children, at the time they happen, immediately afterwards, and, in some cases, throughout life. We don’t need evidence from brain scans to tell us that. But maltreatment and neglect are being presented as if pre-existing evidence of harm isn’t sufficient to persuade legislators that more stringent legislative measures are required to prevent maltreatment and neglect, so neurobiological findings are being recruited for this purpose. Secondly, although there is certainly evidence to suggest that maltreatment and neglect have a negative impact on brain development, they are only two of the factors that do so. In other words, you could predict with some confidence that maltreatment and neglect would result in ‘abnormal’ brain development, but you can’t assume that because someone’s brain has developed abnormally, that they were maltreated or neglected as a child. Thirdly, there’s an implicit assumption in the way the evidence is presented that maltreatment and neglect are the primary cause of ‘social problems’, when social scientists have been aware, for decades, that those causes are many, varied and have complex interactions.
In short, the evidence doesn’t appear to support the idea that the predominant cause of social problems is child maltreatment or neglect. Allen and Duncan Smith call for a study along the lines of the Dunedin study to be carried out in the UK “in order to provide definitive evidence on the benefits of Early Intervention” (Good Parents p.52). I can’t see why another study is necessary – the Cambridge study makes clear that the causes of antisocial behaviour are complex and that patterns of behaviour change significantly over the lifespan.
Update 13/12/16: Thanks to @PaulWhiteleyPhD on Twitter for drawing attention to this article in Nature analysing the Dunedin study findings.
In the next post, I want to look at what Graham Allen and Iain Duncan Smith have to say about brain development.
Photographs of Graham Allen and Iain Duncan Smith from Early Intervention Early Intervention: Good Parents, Great Kids, Better Citizens.