Research
The record, from 2010 to now.
Studies, field notes, and briefings from 2011 to 2026. Each one records a change in the standards, the programmes, or the assessment.
2010–2013: Foundations in Computing Literacy
The real-world cost of being left offline
Our first research question was not abstract. It asked what a household loses when the form, the shift, and the appointment have already moved onto a screen.
A certificate an employer can carry across a desk
Foundation, Intermediate, and Practitioner were named in the first years, and they were named so that a pass would mean the same thing in a second workplace.
Which teaching approaches closed the gap fastest
By 2013 we had enough cohorts to compare methods. The fastest gains were not in the rooms with the newest software.
2014–2017: Data Sense arrives
The question moved past “can you use a computer?”
Data Sense was formalised because the computer had started producing figures that ordinary jobs were expected to believe.
How citizens read the charts in front of them
We studied ordinary readers, not analysts, and we asked what a news graphic is allowed to make them believe.
School mathematics and the reasoning a workplace asks for
The gap was not a shortage of arithmetic. It was a shortage of judgement about messy, incomplete, slightly political numbers.
A route into data work that does not start at a university gate
Charity programmes in these years turned toward young people in underserved areas, with a structured path and a certificate a hiring manager could read.
2018–2021: AI Fluency becomes urgent
AI Fluency, and what we refused to pretend
The public arrival of capable machine-learning systems forced a third domain. The aim was never a nation of machine-learning engineers.
Certification when the building is closed
Lockdown would have halted a hall-based exam. We accelerated remote delivery so the pathway did not stop with the bus routes.
A new inequality: fluent access, and the lack of it
Adoption of generative systems was fast, and it was not evenly shared. We measured the split before we wrote the next syllabus.
A certificate is useless if the connection drops
Charity work expanded into devices and connectivity, because remote learning and remote work were only real for the people who could stay on the call.
2022–2026: Integration, scale, and public trust
Three domains, one continuum
A learner may walk from Computing Literacy into Data Sense and on into AI Fluency, or go deep in a single domain. The certificates stay stackable.
Skill trajectories, across age, not just across cohorts
Longitudinal work asks a ruder question than a pass rate: what is still usable three years after the certificate was printed?
Which certification models still hold
We evaluate our own assessment, in public, because an examining body that will not sit still for a question has forgotten who it is for.
What a portion of the fee is already doing
Every certification fee contributes to free community programmes and to open educational resources. That split is a rule, not a campaign.
We have not solved digital inequality
Sixteen years in, the claim we will defend is narrower than the claim we are sometimes offered.