Visible Learning: Getting hung up on the hinge point

If you were in the UK in May 2014 then it's possible that you will have heard that John Hattie attended three days of conferences in London, hosted by Osiris Education. The interest in Visible Learning will have dramatically increased as a result of this. Therefore it's a good time to announce my most recent findings regarding just one small aspect of Visible Learning: effect sizes.

Hattie is synonymous with the effect size of 0.4 This is the value of the "hinge point". Hattie says that teachers and leaders should look at the impact that teachers have on student learning and then do something about it. If a teacher looks at their own impact over time (say 12 months) and finds that their teaching has an impact equal to an effect size of less than 0.4, then that teacher needs to do something differently.

There are going to be renewed discussions regarding effect sizes. In advance of this I want to offer the following:

The effect size of 0.4 is derived from the average effect size of the 800 meta-analyses (50,000 research papers) that Hattie has thus far considered in his analysing of the last 20 years research. This research looks at student achievement only. For the impact of schooling on the health and well-being of children we will have to wait for work in Germany to be released.

When you look at other international measures of achievement (PISA, PIRLS, TIMSS) then 0.4 is again the effect size associated with a year's input of schooling.

So at this point you'll be thinking "maybe this is a good international measure but what about the national picture."

This is where my maths has now taken me.

Let's look at the starting point of over half a million pupils in England as they enter secondary school. The measure here is their KS2 SATs results. These pupils then spend the next 5 years in education with a variety of teachers. Let's now look at the end point of the pupils. Here I have used their KS4 results. To calculate an effect size you need 2 pieces of data. You also need to know the standard deviation of both sets of data.

What then is the average annual effect size for those circa 500,000 pupils who started secondary education in 2006?

0.37 or as Hattie would say "that's just not good enough".

Let's look at another cohort starting secondary in 2008. The effect of 5 years of English education equates to an average annual effect size of 0.42. That's more like it!

Is an effect size of 0.4 a good measure of 1 years education in England? It would certainly appear so in secondary education for those 2 cohorts.

Is the hinge-point worth getting hung-up on? I think so. It's a great place to start for teachers who want to be evaluators of their own impact. There are caveats regarding the use of effect sizes, just as there are with any aspects of data analysis. But with effect sizes, schools and teachers can see more readily whether pupils are receiving a year's learning for a year's teaching.

Do you "Know Thy Impact"?