A screening assessment is only as good as the evidence behind it. To know whether a task measures what it should, whether the instructions make sense to a seven-year-old, and what “typical” looks like at each age, you need data from a broad range of pupils, taking the assessment in the conditions schools actually use.

We couldn’t do that without schools. More than 20 research schools across the UK, a mix of trust and maintained schools, have helped us pilot new tasks and standardise each assessment. This article explains what that work involves and what it has helped us build.

How a new assessment is built

Each Talamo assessment goes through the same stages: a pilot, refinement, then standardisation.

• Pilot. We start with more tasks than we need and test every one with pupils before any large-scale data collection. Each task is checked at item level for reliability, validity and how well it tracks age.

• Refine. Feedback from schools and the pilot data tell us which tasks to keep, which to change and which to drop.

• Standardise. The final battery is taken by a large group of pupils of every age from 7 to 16. Their results become the norms, the benchmark each pupil’s scores are compared with.

A task keeps its place only while the data supports it. That principle is written into all three of our technical manuals, and it has led to real changes.

What research schools have helped us build

Executive Function. Nine tasks were piloted with 238 pupils in 14 schools between January and March 2026. Two were replaced on the pilot evidence. A selective-attention task added nothing beyond an existing measure of inhibition, so a Flanker task (pupils say which way the middle fish is pointing) took its place. A two-choice visual memory task was limited by its 50% chance of guessing correctly, so a multiple-choice picture memory task replaced it, and that task is still used for research only until its reliability supports scored use. The norms were then built from 560 pupils in 13 schools across England in a spring 2026 study.

Maths & Dyscalculia. Fourteen maths tasks were piloted with 333 pupils in 12 schools between November 2025 and April 2026, and twelve made it into the final assessment. An estimation task was dropped after the pilot showed a design problem that gave artificially low scores to around 8% of secondary-age pupils, and a division task was dropped in favour of the three arithmetic scales we kept. Before launch, Janet Goring, our maths and dyscalculia specialist, reworked one task to focus on applying known facts rather than following patterns. The norms come from 954 pupils in 15 schools across England.

Cognitive & Dyslexia. The content was defined in 2023 with a specialist dyslexia assessment trainer, after reviewing eleven established assessments. The first pilot, in autumn 2023, took place at a specialist dyslexia school, where around 60 pupils with confirmed diagnoses and educational psychologist reports helped tune the earliest scoring models. Since launch, the norms have been rebuilt on a much larger group: 4,580 pupils in 85 UK schools using Talamo, in summer 2026.

What it means for the results schools see

When a school screens a pupil, their scores are compared with pupils of the same age in whole years. That comparison is only meaningful because so many pupils, in so many schools, took the assessments first.

The same care runs through the rest of our evidence. Scales shared between assessments are standardised once, in one cohort, so a score means the same thing whichever assessment it came from. Every report records the version of the norms that produced it. And our technical manuals are open about their limits, for example where samples are smaller at the youngest and oldest ages, because the way those gaps close is with more data from more schools.

Read how we build and test our assessments →

Taking part

In our larger studies, our team visited schools to oversee data collection. Research-only tasks don’t produce reports on individual pupils: the data is used to develop and check the assessments.

Interested in taking part in a future study? We’d love to hear from you. Ask us a question and tell us a little about your school.

Ask a question →