27 April, 2009

What is evidence-based parenting?

In essence I have made this title up, analogous to "evidence-based practice", "evidence-based medicine" and the like. The principle is that we should be able to learn what parenting practices have the most effective outcome for children by looking at research.

"But my child is special!" I hear you cry. "What suits him isn't going to suit other children!". Well, perhaps true. However, let me tell you a little bit about how we can study child development scientifically. These are general principles of scientific study but I'm nicking them shamelessly from the lovely Professor Regan.

First we are looking for studies that have some kind of control. So, we all know children grow and that they develop. It's no good saying "We gave all these children spinach, and look, they all grew, and now, five years later, they can all write! Must have been the spinach!" As you can probably guess they probably would have grown, and learned to write, for different reasons. We want studies that have some children treated in one way but some in another.

In some cases it's perfectly fine to do this (spinach versus cabbage). In others it would be very wrong (spinach versus locking in a cupboard - well, as a spinach disliker, perhaps not so very wrong - let's say vegetables versus locking in a cupboard). In this circumstance we must look for sad cases where locking in a cupboard (or no vegetables ever) happens by accident, or is overlooked by accident. We need to be extra cautious in such cases that the children who had no vegetables were not different in other ways to the children who had vegetables. We don't want children who have vegetables to also have had lots of foreign holidays and a French nanny, otherwise we will wrongly conclude that having no vegetables makes you rubbish at learning French.

Second we want the studies to come out with statistically significant results - and preferably for the size of the results to be meaningful. If the vegetable-eating group is a small group and has very variable French skills, and the no-vegetable group is also a small group with very variable French skills, then even if the vegetable group is better at French, this could happen because children's French skills vary naturally. After all, in each group the French skills vary naturally. So, we need the difference to be big enough to be reliable. This is usually reported when a study is written up for a scientific journal. It is also helpful for the journal to report how meaningful the results are - you wouldn't give all children 9 servings of vegetables every day if it helped them to learn one extra French word in 3 years. This aspect of results (called effect size) should be reported in journals these days, but isn't always, and often wasn't in the past.

To go back to your own very special child, if you read a study that said that eating vegetables made no difference to French learning ability, but you were still convinced it did and wanted your child to eat 9 servings of vegetables a day simply for that purpose, consider this: If for children in general the vegetables do not affect French learning ability, this could mean that for some children it did and some it didn't. However, to get a zero average, it has to mean not that some children are positively affected and some are not affected, but that some are positively affected and some are negatively affected. Do you want your child in the latter group? Thought not.

Where the outcome doesn't matter that much - does it really matter that a child walks a week or two early? probably not - and the intervention is minor - a nursery-rhyme a day isn't going to hurt anyone - then if it's fun and doesn't drive you mad, go for it. Where the outcome does matter - particularly with children with disabilities, often the stakes are high - or where the intervention is costly for parent or child - again, some of the disability interventions are time-consuming and/or expensive and/or wearing for parent and child - then it's helpful to look at evidence before committing yourself and your child.

Third we want to see the studies in a peer-reviewed journal. As a scientist, I love to hate the peer review process. On Facebook I am a member of a tongue-in-cheek group called "Down with Reviewer 2!". Peer reviewers are anonymous (in many cases they don't know the author's name either) and can occasionally verge on vindictive. However, good peer reviews spot the flaws in your study's design, suggest really interesting reanalyses, and point out articles you haven't read. In an ideal world, you thank them profusely and include their comments, and your article comes out in the journal whereupon all your friends ring you up and congratulate you on a job well done. In practice, your article goes to 3 or 4 journals and looks nothing like it did at the start when it finally comes out. However, this does mean that you know when you read an article in a peer-reviewed journal, some of your colleagues have judged its quality. Scientists will still read these articles and criticize the science and the writing, but they know they are not the first to do this.

Finally, a bit about myself. I'm a lecturer at a British university in developmental psychology, and specialise in not making up my mind about what I want to study. Some of the things I'm particularly interested in are learning disability, language and literacy, and children in unusual circumstances (carried round in a cradleboard for the first year? Ooh shiny!). This post which was meant to be an introduction to my blog, which was also meant to be a regular slot, has sat in the cupboard (along with the spinach-haters) for quite some time due to an unfortunate accident suffered by my arm over the summer. Better late than never, eh?