When the Baseline Is Broken: Data, Power and the Voices We Keep Missing

This blog accompanies Episode 10 of Data Talks: the Podcast from Data in Schools – "A Conversation about Data for Good, with Frances Akinde and Dr Nicole Ponsford." Hosted by Matthew Savage and Paul Swanson, Data Talks is available on Apple Podcasts, Spotify, and YouTube.

A blog by Matthew Savage and Dr Nicole Ponsford, Founding CEO of the GEC

What happens when the instruments we use to see children are themselves the source of their invisibility?

That question sits at the very centre of everything we claim to care about in education. And yet much of the data infrastructure of contemporary schooling was built in the 1960s. We are, in many ways, still measuring children with tools designed for a world – and a worldview – that most of us would now reject.

The data was never neutral to begin with. Frances Akinde's work at the intersection of race and SEND makes this vivid: when a school leader can say, in all seriousness, "that child isn't double-disadvantaged so he won't be our focus right now," something has gone badly wrong with the system of thought that produced the category in the first place. Categories become ceilings. Measurements become excuses.

For Dr Nicole Ponsford, this is also personal. As a child who should have been counted as eligible for free school meals but whose family refused the label for reasons of shame and dignity, she knows firsthand that data which is never collected cannot inform the decisions that shape a life. And when you layer in the incentive to underidentify SEND to reduce spend, or the baseline figures submitted to the DfE without a school even realising they are wrong, unreliability starts to look less like oversight and more like architecture.

This is where the work of the GEC (Global Equality Collective) becomes so important. Nic's doctoral research and Kaleidoscopic Data Inclusion Roadmap – created through participatory action research and built with schools - now used across 400 schools in over 30 countries – was born out of a refusal to accept that the only demographics worth measuring are the ones that are convenient to measure. By collecting intelligence across 60 different demographic identifiers and crucially allowing people to self-identify, the GEC Platform surfaces the intersectional lived experience of the smallest and most marginalised groups – precisely those whom standard data dashboards or MIS systems tend to render invisible.

What strikes us both about Kaleidoscopic Data is that it is, at root, an epistemological act. It says: we will not let the powerful decide what counts as knowable. It refuses the tidiness that reductive data systems offer, and insists that human complexity is a signal to be honoured rather than a problem to be smoothed away.

The data we have inherited reflects what was easy to capture decades ago, and those who built their authority around it have a vested interest in treating it as the only data that matters. Good intentions and flawed implementation can produce outcomes nobody wanted. The problem of data in schools is never purely technical: it is always also political.

These are threads Matthew has been pulling at for some time – the idea that there are many ways of knowing a child, and that the numerical and quantifiable represent only one of them. Warm data, kind data, slow data: ways of seeing that are rooted in observation and relationship, that ask who decided what counts as evidence, and whose experience gets left outside the frame. What Frances calls "data through an equitable lens" belongs to the same impulse – a demand for a more honest engagement with what a child actually is than any aggregated score can offer.

This is one of those rare conversations that keeps expanding after it ends. As well as schools trusts and regions now using the GEC Platform as their missing intelligence layer for inclusion, Nic's Inclusion Index, launched each term from live data and visible to schools around the world, is proof that this is not merely theoretical. Frances's work – insisting that leaders disaggregate their SEND data by ethnicity before they claim to understand it – is proof that the conversation can happen at the chalkface. And Matthew's thinking on ‘a plurality of knowing’ a child is a reminder that the alternative to reductive measurement is not vagueness: it is a richer, more honest account of what schools are actually for.

Rep. John Lewis called it good trouble. In this particular episode of the Data Talks podcast, Frances and Nic are definitely making it – and so are we. When you listen to the episode yourself, we hope it finds you ready to ask different questions of the numbers in front of you.

Listen to Episode 10 of Data Talks: the Podcast on Apple Podcasts, Spotify, or YouTube.

Explore the work of the GEC at thegec.education.

Learn more about Matthew's work on data, assessment and belonging at The Mona Lisa Effect.