To be confident in data, we need big numbers. Having group sizes large enough from which to draw firm conclusions allows us to plan support and celebrate success.

Unfortunately, in schools, the data often isn’t “big” enough to achieve this on its own - not even in large secondary schools. The rich tapestry of learning is too nuanced, the school population is too varied and the performance measures are too unstable for us to really get a handle on what’s going on by group, class and cohort performance simply through looking at the data.

This isn’t to say that data cannot be used intelligently, but we need to take care that we are not creating false data “pictures” that are biased by how our schools are set up.

‘Small’ data

For example, it is common to have leaders with responsibility for different groups of students, whether that be pupils from disadvantaged backgrounds, pupils with special educational needs and disabilities, underachievers, poor attenders and so on. However, it is very hard to truly gauge the effect that the school is having on these groups in isolation, for numerous reasons - not least the fact that these are not unique groups. Often the individuals that make up one group of students also fall into another, or several other categories.

To ensure that data is being used to its greatest effect, schools need to engage with it more at its smallest possible denominator: at student level. We also need to allow for more time to work with data - to make plans around it, to have conversations about it and to ensure that individual student requirements are considered.

This might seem like a lot more work, but what should happen is that all this small-data talk will replace the broad-brush approaches and schools will begin to see a positive effect on learning, as data is used more effectively.

‘Not enough focus on the individual’

Schools need to bring their data systems and processes much closer to the action of the classroom. In the same way that a teacher should be able to talk knowledgeably about any individual in their class, as well as giving a broad summary of where the whole class is at, whole-school data should provide both the big and small picture of what is going on across the school. Too often, our data systems are giving a misleading view of the latter because there is not enough of a focus on the individual.

Data should help us to formulate support strategies for individuals, to identify and cover gaps in knowledge, to ensure that each individual is able to enter the classroom in the best frame of mind to learn, and that each individual is appropriately challenged.

Marrying the impersonal world of data with a student-centred approach can sometimes be difficult. The aim should be to always have data to hand, but not let it get out of hand. Data should clarify, not cloud the raft of transitions that every student and teacher needs to navigate. In this way, schools can think big, using small data to help shape, but not govern, their thinking.

Peter Atherton is chief data officer at Minsthorpe Community College, Wakefield. He has more than 15 years’ experience in analysing education data. He blogs at minsthorpedata.com