There was a mild amount of interest yesterday in the Sutton Trust report on the impact of infant attachment on later development. The press release trumpeted how this showed that 40% of children “miss out on the parenting needed to succeed in life”. This is strong stuff, suggesting that there is some key ingredient missing from this (vast) group of children. The researchers feel that they have find this ingredient. It is secure attachment.
Attachment needs some explanation. The simplest way to explain it is broadly that during the first year, the baby learns to trust their caregiver to reliably respond to their needs. This process can go wrong in a few ways.
A tiny number of children seem to be unable to form attachments at all. A slightly larger tiny number form attachments indiscriminately. These problems appear to be quite largely genetic or neurobiological, and we can quickly move on for our purposes.
Some children (about 15%) seem not to be able to work out whether to trust the parent. This seems to be linked to parents being either frightened, or frightening, and therefore unpredictable. This group, described as having disorganised attachment, have undoubtedly poor outcomes, although we have no idea whether the attachment pattern is a cause of the later problems, or an alarm bell for other underlying problems.
This leaves us with children with organised attachment, who seem to have worked out a strategy for responding to parents. These have been split by attachment theorists on the following basis:
Basically, in lab conditions, the child is separated from their caregiver. They either respond too much, not enough, or ‘just right’, or as statisticians would call it, in the average range. Children at both ends of the range are classified as insecure, and this insecurity is claimed to be due to either parental under- or over-responsiveness. These children have, however, never been shown to have any clinically significantly worse outcomes that those in the ‘secure’ middle. This report makes some claims that these ‘insecure’ children are in some ways worse at communication, but as it’s not a systematic review, it’s hard to evaluate these claims.
As you might have guessed, I have a few problems with this report:
Firstly, they conflate disorganised and insecure attachment to increase the proportion of children with an indicator of poor outcome from 15% to 40%, making their conclusion misleadingly dramatic.
Secondly, they make the assumption that the only predictor of the child’s attachment pattern is the parent’s behaviour. This kind of ‘blank slate’ thinking is simply not consistent with the growing evidence if genetic and temperamental factors in attachment. Among those with organised attachment, any slight difference in outcomes could be due to these factors, rather than parental responsiveness. We just don’t know.
But let’s accept that there is a sizeable number of children in the UK with disorganised attachment, which should act as an alarm bell for potential poor outcomes. The authors present evidence that attachment style can be altered with intensive parenting work, but they offer not a single piece of evidence that changing the attachment style improves outcomes. To use my alarm bell analogy, it may well be that changing attachment style is the equivalent of switching off the alarm, rather than putting out the fire.
Why am I going on about this? Surely there’s no harm in at-risk parents getting help? Well, firstly, of 40% of parents who could now be told that their parenting is harmful, most are doing nothing wrong. This seems terribly unfair.
Secondly there is an opportunity cost. Resources for early intervention are limited and shrinking (for instance 2/3rds of CAMHS services cut since 2010) and diverting resources to attachment-based intervention with no evidence of impact on outcome, when we could instead address parental mental health, poverty, substance misuse, domestic violence and all the factors that set off this alarm, would be a tragedy.
Now I need to stop ignoring my children. They are becoming disorganised. And silly.