Our research
The research behind every Talamo assessment
The research behind every Talamo assessment
The research behind every Talamo assessment
Each one is built with specialists, piloted with pupils in UK schools and standardised by age, from 7 to 16. Here’s how we do it, the evidence behind each result, and what screening can’t tell you.
Each one is built with specialists, piloted with pupils in UK schools and standardised by age, from 7 to 16. Here’s how we do it, the evidence behind each result, and what screening can’t tell you.
Research in numbers
The numbers behind our results
The numbers behind our results
The numbers behind our results
4,580
4,580
4,580
pupils from 85 UK schools in our Cognitive & Dyslexia norms
20+
20+
20+
research schools helping us pilot new tasks and build our norms
95%
95%
95%
agreement between our dyslexia risk model and specialist reviewers
Our process
How we build assessments
How we build assessments
How we build assessments
Nine steps in three stages, from deciding what to measure to the limits of screening. Where a check applies to one assessment only, we say so.
Nine steps in three stages, from deciding what to measure to the limits of screening. Where a check applies to one assessment only, we say so.
Build
Steps 1–3 · What we measure, how the tasks are built and how they’re piloted
Steps 1–3 · What we measure, how the tasks are built and how they’re piloted
1
Step 1 of 9
Defining what to measure
Defining what to measure
Each assessment begins with a clear framework of the skills and processes we need to understand. We work with specialist researchers, assessment publishers, educational experts and practitioners, and deliberately start with more material than the final assessment needs.
Each assessment begins with a clear framework of the skills and processes we need to understand. We work with specialist researchers, assessment publishers, educational experts and practitioners, and deliberately start with more material than the final assessment needs.
The Cognitive & Dyslexia content was defined in 2023 with a specialist dyslexia assessment trainer. Eleven established assessments were reviewed as part of that work, and the content is aligned with the Delphi consensus definition of dyslexia (February 2025).
The Cognitive & Dyslexia content was defined in 2023 with a specialist dyslexia assessment trainer. Eleven established assessments were reviewed as part of that work, and the content is aligned with the Delphi consensus definition of dyslexia (February 2025).
Maths & Dyscalculia focuses on the markers research links most consistently to dyscalculia, such as placing numbers on a number line and comparing their size, alongside maths vocabulary and word problems.
Maths & Dyscalculia focuses on the markers research links most consistently to dyscalculia, such as placing numbers on a number line and comparing their size, alongside maths vocabulary and word problems.
Executive Function measures the cognitive processes behind goal-directed behaviour directly, through performance-based tasks, so schools aren’t relying on classroom observation alone.
Executive Function measures the cognitive processes behind goal-directed behaviour directly, through performance-based tasks, so schools aren’t relying on classroom observation alone.
Cognitive & Dyslexia: three tiers
Cognitive profile
Sample report
Baseline ability
Visual reasoning
Visual reasoning
117
Verbal reasoning
Verbal reasoning
91
Key indicators
Phonological awareness
Phonological awareness
79
Processing speed
Processing speed
82
Working memory
Working memory
84
Impact on reading and spelling
Reading speed
Reading speed
74
Reading comprehension
Reading comprehension
90
Spelling
Spelling
73
Scores from Emily’s sample report. 100 is average for her age.
2
Step 2 of 9
Building for digital
Building for digital
Every task runs online in the browser. Each one opens with spoken instructions and practice items with feedback, and pupils are supervised by a teacher or teaching assistant.
Every task runs online in the browser. Each one opens with spoken instructions and practice items with feedback, and pupils are supervised by a teacher or teaching assistant.
Some of the assessments in the Cognitive & Dyslexia tasks are adapted for digital delivery from the Woodcock-Johnson IV, developed in partnership with its publisher, Riverside Insights, with Talamo building the remainder.
Some of the assessments in the Cognitive & Dyslexia tasks are adapted for digital delivery from the Woodcock-Johnson IV, developed in partnership with its publisher, Riverside Insights, with Talamo building the remainder.
The maths-specific tasks were built with Janet Goring, a field expert in maths difficulties and dyscalculia.
The maths-specific tasks were built with Janet Goring, a field expert in maths difficulties and dyscalculia.
Executive Function uses Talamo’s digital versions of established tasks, including the Sustained Attention to Response Task, a flanker task, rule switching and tower planning.
Executive Function uses Talamo’s digital versions of established tasks, including the Sustained Attention to Response Task, a flanker task, rule switching and tower planning.
3
Step 3 of 9
Piloting with pupils
Piloting with pupils
Before large-scale data collection begins, we test every task directly with pupils: how clear the instructions are, how accessible the interactions are and whether the difficulty is right. Tasks are then refined, replaced or removed.
Before large-scale data collection begins, we test every task directly with pupils: how clear the instructions are, how accessible the interactions are and whether the difficulty is right. Tasks are then refined, replaced or removed.
Cognitive & Dyslexia was first piloted in autumn 2023 at a specialist dyslexia school, with around 60 pupils with confirmed diagnoses and educational psychologist reports.
Cognitive & Dyslexia was first piloted in autumn 2023 at a specialist dyslexia school, with around 60 pupils with confirmed diagnoses and educational psychologist reports.
Executive Function: nine tasks were piloted with 238 pupils in 14 schools, and a Stroop task was replaced by a flanker task.
Executive Function: nine tasks were piloted with 238 pupils in 14 schools, and a Stroop task was replaced by a flanker task.
Maths & Dyscalculia: fourteen tasks were piloted with 333 pupils in 12 schools, and twelve went forward into the final assessment.
Maths & Dyscalculia: fourteen tasks were piloted with 333 pupils in 12 schools, and twelve went forward into the final assessment.
Measure
Steps 4–6 · Norms, reliability and fairness
Steps 4–6 · Norms, reliability and fairness
4
Step 4 of 9
Standardisation
Standardisation
We test the final assessment across the full age range, 7 to 16, so each result is a standard score compared with pupils of the same age. Every report records the norm version that produced it.
We test the final assessment across the full age range, 7 to 16, so each result is a standard score compared with pupils of the same age. Every report records the norm version that produced it.
Cognitive & Dyslexia norms were rebuilt in summer 2026 on a far larger cohort, weighted to match national school types and dyslexia prevalence, so they reflect whole cohorts.
Cognitive & Dyslexia norms were rebuilt in summer 2026 on a far larger cohort, weighted to match national school types and dyslexia prevalence, so they reflect whole cohorts.
Executive Function norms come from 560 pupils in 13 schools, and Maths & Dyscalculia norms from 954 pupils in 15 schools, all across England.
Executive Function norms come from 560 pupils in 13 schools, and Maths & Dyscalculia norms from 954 pupils in 15 schools, all across England.
Executive Function and Maths & Dyscalculia use first-generation norms, and restandardisation on a larger cohort is planned.
Executive Function and Maths & Dyscalculia use first-generation norms, and restandardisation on a larger cohort is planned.
Cognitive & Dyslexia norms, summer 2026
Cognitive & Dyslexia norms, summer 2026
4,580
4,580
pupils
pupils
85
85
UK schools
UK schools
7–16
7–16
age range
age range
5
Step 5 of 9
Reliability
Reliability
Reliability shows how consistently a task measures the same thing. All nine Cognitive & Dyslexia tasks show good to excellent reliability, where 1.0 would be perfectly consistent.
Reliability shows how consistently a task measures the same thing. All nine Cognitive & Dyslexia tasks show good to excellent reliability, where 1.0 would be perfectly consistent.
Executive Function and Maths & Dyscalculia figures come from their pilot studies, so treat them as early estimates.
Executive Function and Maths & Dyscalculia figures come from their pilot studies, so treat them as early estimates.
Two Executive Function scores don’t have their own estimate yet.
Two Executive Function scores don’t have their own estimate yet.
Want more technical detail?
Want more technical detail?
Ask us and we’ll share it
Ask us and we’ll share it
Reliability at a glance
Cognitive & Dyslexia
0.80–0.97
Executive Function (pilot)
0.84–0.98
Maths & Dyscalculia (pilot)
0.78–0.96
6
Step 6 of 9
Fairness
Fairness
We check whether anything other than the skill being measured could change a pupil’s result, and correct for it where we can.
We check whether anything other than the skill being measured could change a pupil’s result, and correct for it where we can.
Answering timed tasks by touch can make a pupil look faster. Our data shows touch raises scores on Cognitive & Dyslexia’s two timed tasks, so we correct for it.
Answering timed tasks by touch can make a pupil look faster. Our data shows touch raises scores on Cognitive & Dyslexia’s two timed tasks, so we correct for it.
Before Maths & Dyscalculia reports signs of dyscalculia, it checks for language-led difficulty (weak maths vocabulary and word problems with number and arithmetic skills intact), which matters for EAL learners.
Before Maths & Dyscalculia reports signs of dyscalculia, it checks for language-led difficulty (weak maths vocabulary and word problems with number and arithmetic skills intact), which matters for EAL learners.
EAL status never raises a pupil’s dyslexia risk level, and reported conditions never change the Executive Function outcome.
EAL status never raises a pupil’s dyslexia risk level, and reported conditions never change the Executive Function outcome.
How each answer was given
How each answer was given
Recorded for every answer: touch, pen, keyboard or mouse
Recorded for every answer: touch, pen, keyboard or mouse
Corrected only when clearly touch
Corrected only when clearly touch
An age-banded offset on Cognitive & Dyslexia’s two timed tasks
An age-banded offset on Cognitive & Dyslexia’s two timed tasks
Norms come from pupils using a keyboard and mouse
Norms come from pupils using a keyboard and mouse
Check
Steps 7–9 · Human review, specialist benchmarks and honest limits
Steps 7–9 · Human review, specialist benchmarks and honest limits
7
Step 7 of 9
AI, with a person checking every report
AI, with a person checking every report
Talamo uses AI to help interpret each pupil’s profile, within rules set by our research team. For Executive Function, a fixed rule sets the outcome and AI writes the explanation. We built the process so that every step can be checked.
Talamo uses AI to help interpret each pupil’s profile, within rules set by our research team. For Executive Function, a fixed rule sets the outcome and AI writes the explanation. We built the process so that every step can be checked.
Pseudonymised inputs only: the model sees scores and context, never a pupil’s name, date of birth, any identifier or the school’s name. Pupil assessment data is stored in the UK.
Pseudonymised inputs only: the model sees scores and context, never a pupil’s name, date of birth, any identifier or the school’s name. Pupil assessment data is stored in the UK.
Every output is stored with its exact input, and every report records the prompt, model and norm versions, so any result can be traced.
Every output is stored with its exact input, and every report records the prompt, model and norm versions, so any result can be traced.
Before it was adopted in June 2026, the dyslexia risk model ran side by side with the statistical model it replaced on more than 5,300 reports.
Before it was adopted in June 2026, the dyslexia risk model ran side by side with the statistical model it replaced on more than 5,300 reports.
Questions about how we use AI?
Questions about how we use AI?
Book a call with our team
Book a call with our team
Pseudonymised scores
Pseudonymised scores
No names or identifying details
No names or identifying details
AI model
AI model
Interprets the profile within rules set by our research team
Interprets the profile within rules set by our research team
Checked by a person
Checked by a person
Every report, before the school sees it
Every report, before the school sees it
Report released
Report released
Normally within 48 hours
Normally within 48 hours
8
Step 8 of 9
Checked against specialist judgement
Checked against specialist judgement
We test our results against the judgement of specialist reviewers, and against the scores of pupils with a reported diagnosis.
We test our results against the judgement of specialist reviewers, and against the scores of pupils with a reported diagnosis.
Against a specialist-reviewed benchmark set that includes diagnosed and suspected dyslexia, our dyslexia risk model agrees with reviewer judgement in 95% of profiles.
Against a specialist-reviewed benchmark set that includes diagnosed and suspected dyslexia, our dyslexia risk model agrees with reviewer judgement in 95% of profiles.
In the Cognitive & Dyslexia norm cohort, pupils with a reported dyslexia diagnosis scored clearly lower on spelling than pupils with none, but no lower on reasoning.
Level of Need matched human reviewers exactly in 92% of 200 benchmark profiles.
Level of Need matched human reviewers exactly in 92% of 200 benchmark profiles.
British Dyslexia Association
Talamo was selected through the BDA’s two-stage tender process, which began in December 2025, and met the BDA’s requirements. The BDA considers Talamo a suitable screening tool for children aged 10–16. Our Cognitive & Dyslexia assessment is standardised for ages 7 to 16.
Report scales, checked by reviewers
Dyslexia risk
95% agreement
Unlikely
Unlikely
Possible
Possible
Likely
Likely
Very likely
Very likely
Against a specialist-reviewed benchmark set
Level of Need
92% exact match
Low
Low
Medium
Medium
High
High
Very high
Very high
Against 200 profiles scored by human reviewers
9
Step 9 of 9
Honest limits
Honest limits
Screening is a snapshot, not a diagnosis. Results are described as “consistent with” dyslexia or dyscalculia, and only a specialist assessment can confirm a diagnosis. Results should always be read alongside classroom information and attainment data.
Screening is a snapshot, not a diagnosis. Results are described as “consistent with” dyslexia or dyscalculia, and only a specialist assessment can confirm a diagnosis. Results should always be read alongside classroom information and attainment data.
As with any screening tool, some pupils with dyslexia or dyscalculia will be missed and some without will be flagged.
As with any screening tool, some pupils with dyslexia or dyscalculia will be missed and some without will be flagged.
Cognitive & Dyslexia doesn’t assess rapid automatised naming, reading-aloud accuracy or writing, and doesn’t measure visual stress.
Cognitive & Dyslexia doesn’t assess rapid automatised naming, reading-aloud accuracy or writing, and doesn’t measure visual stress.
Executive Function reports never use diagnostic labels, and its indicator isn’t yet validated against clinical outcomes. The maths indicator reports signs of dyscalculia, not a verdict, because its evidence base is younger.
Executive Function reports never use diagnostic labels, and its indicator isn’t yet validated against clinical outcomes. The maths indicator reports signs of dyscalculia, not a verdict, because its evidence base is younger.
In this report
In this report
Cognitive & Dyslexia report
Results
Insights
How to support
“It’s worth noting though that this is a dyslexia screener and not a formal assessment. Whilst our research has shown that our screener is highly accurate, if you feel that the results do not match your expectations, we would recommend getting a second opinion.”
From the “In this report” page of a Talamo dyslexia report. Read it in a sample report (PDF)
Reading a result
Screening result
Likely to be dyslexia
Read alongside classroom evidence
Classroom work, observations and attainment data
Specialist assessment where needed
A formal assessment is needed to confirm dyslexia
See how this evidence shows up in a pupil report
See how this evidence shows up in a pupil report
Our three assessments
Our three assessments
Our three assessments
Each test page has the full detail: what pupils do, what each task measures and what the report shows.
Each test page has the full detail: what pupils do, what each task measures and what the report shows.
Compare the three tests
Cognitive & Dyslexia
Cognitive & Dyslexia
Cognitive & Dyslexia
Measures the cognitive and literacy skills linked to dyslexia and estimates a pupil’s risk of dyslexia.
Measures the cognitive and literacy skills linked to dyslexia and estimates a pupil’s risk of dyslexia.
9 tasks
4,580-pupil norms
Reliability 0.80–0.97
Explore Cognitive & Dyslexia
Executive Function
Executive Function
Executive Function
Measures sustained attention, inhibitory control, working memory, cognitive flexibility and planning through a performance-based battery.
Measures sustained attention, inhibitory control, working memory, cognitive flexibility and planning through a performance-based battery.
8 scored tasks
560-pupil norms
Reliability 0.84–0.98 (pilot)
Explore Executive Function
Maths & Dyscalculia
Maths & Dyscalculia
Maths & Dyscalculia
Screens the cognitive foundations and applied skills that maths depends on, with signs of dyscalculia as a secondary indicator.
Screens the cognitive foundations and applied skills that maths depends on, with signs of dyscalculia as a secondary indicator.
15 tasks
954-pupil norms
Reliability 0.78–0.96 (pilot)
Explore Maths & Dyscalculia
Research schools
Working with research schools
Working with research schools
Working with research schools
More than 20 UK schools, both trust and maintained, work with us on research. They try new tasks first, and they helped build the norms for Executive Function and Maths & Dyscalculia.
More than 20 UK schools, both trust and maintained, work with us on research. They try new tasks first, and they helped build the norms for Executive Function and Maths & Dyscalculia.
Piloting new tasks: pupils try early versions, and their feedback and results decide what we keep, change or drop.
Piloting new tasks: pupils try early versions, and their feedback and results decide what we keep, change or drop.
Our team on site: in our larger studies, we visited schools to oversee data collection.
Our team on site: in our larger studies, we visited schools to oversee data collection.
Research data only: research-only tasks don’t produce reports on individual pupils.
Research data only: research-only tasks don’t produce reports on individual pupils.
Interested in taking part in a future study? Tell us a little about your school.
Leo Thornton, our co-founder and Chief Product Officer, trying out tasks on a school computer.
Leo Thornton, our co-founder and Chief Product Officer, trying out tasks on a school computer.
From the research blog
All articles
Research
Why early screening matters: what the evidence says in 2026
The 2026 evidence on early identification: a new definition of dyslexia, an income gap in diagnosis, and what it means.
·
Leo Thornton
Research
How we work with schools on research
Behind every Talamo result are pupils in real classrooms. How schools help us pilot tasks and build our norms.
·
Leo Thornton
Research
Why early dyslexia screening builds lifelong learners
Practical steps for primary schools, from choosing when to screen to talking to families about the results.
·
Sophie Dick
From the research blog
All articles
Research
Why early screening matters: what the evidence says in 2026
The 2026 evidence on early identification: a new definition of dyslexia, an income gap in diagnosis, and what it means.
·
Leo Thornton
Research
How we work with schools on research
Behind every Talamo result are pupils in real classrooms. How schools help us pilot tasks and build our norms.
·
Leo Thornton
From the research blog
All articles
Research
Why early screening matters: what the evidence says in 2026
The 2026 evidence on early identification: a new definition of dyslexia, an income gap in diagnosis, and what it means.
·
Leo Thornton
Research
How we work with schools on research
Behind every Talamo result are pupils in real classrooms. How schools help us pilot tasks and build our norms.
·
Leo Thornton
Research
Why early dyslexia screening builds lifelong learners
Practical steps for primary schools, from choosing when to screen to talking to families about the results.
·
Sophie Dick
Assessment content and review
The British Dyslexia Association and Riverside Insights
The British Dyslexia Association and Riverside Insights
The British Dyslexia Association and Riverside Insights
Riverside Insights’ Woodcock-Johnson IV content underpins seven of the nine Cognitive & Dyslexia tasks. Here’s how each organisation is involved.
Riverside Insights’ Woodcock-Johnson IV content underpins seven of the nine Cognitive & Dyslexia tasks. Here’s how each organisation is involved.
Riverside Insights
Woodcock‑Johnson IV content
Riverside Insights has developed research‑backed educational and clinical assessments for nearly a century.
Talamo has a strategic content agreement with Riverside and has adapted elements of its established Woodcock‑Johnson assessment content for digital delivery, combining trusted assessment design with automated scoring and accessible reporting.
British Dyslexia Association
Signposted by the BDA
The British Dyslexia Association works to create a dyslexia‑friendly society in which dyslexic people of all ages can reach their potential.
Talamo was selected through the BDA’s two‑stage tender process and met the BDA’s requirements. The BDA considers Talamo a suitable screening tool for children aged 10–16, and our Cognitive & Dyslexia assessment is signposted by the BDA.
Riverside Insights
Woodcock‑Johnson IV content
Riverside Insights has developed research‑backed educational and clinical assessments for nearly a century.
Talamo has a strategic content agreement with Riverside and has adapted elements of its established Woodcock‑Johnson assessment content for digital delivery, combining trusted assessment design with automated scoring and accessible reporting.
British Dyslexia Association
Signposted by the BDA
The British Dyslexia Association works to create a dyslexia‑friendly society in which dyslexic people of all ages can reach their potential.
Talamo was selected through the BDA’s two‑stage tender process and met the BDA’s requirements. The BDA considers Talamo a suitable screening tool for children aged 10–16, and our Cognitive & Dyslexia assessment is signposted by the BDA.
Riverside Insights
Woodcock‑Johnson IV content
Riverside Insights has developed research‑backed educational and clinical assessments for nearly a century.
Talamo has a strategic content agreement with Riverside and has adapted elements of its established Woodcock‑Johnson assessment content for digital delivery, combining trusted assessment design with automated scoring and accessible reporting.
British Dyslexia Association
Signposted by the BDA
The British Dyslexia Association works to create a dyslexia‑friendly society in which dyslexic people of all ages can reach their potential.
Talamo was selected through the BDA’s two‑stage tender process and met the BDA’s requirements. The BDA considers Talamo a suitable screening tool for children aged 10–16, and our Cognitive & Dyslexia assessment is signposted by the BDA.
Questions about our research
Questions about our research
Questions about our research
Quick answers about accuracy, norms, data and review.
Quick answers about accuracy, norms, data and review.
How accurate is Talamo?
How were the norms built?
Is pupil data safe?
Who reviews the reports?
Still have a question?
Our research
Have a question about our research?
Have a question about our research?
Have a question about our research?
Ask us anything about how our assessments are built, tested and checked, or book a call to talk it through.
Ask us anything about how our assessments are built, tested and checked, or book a call to talk it through.
Cognitive & Dyslexia norms
Summer 2026
Pupils
4,580
UK schools
85
Reliability
0.80–0.97
Ask a question
We’re happy to explain
Get updates from Talamo
A monthly email with news, research and event updates for SENCos, schools and families.
© Talamo Limited 2026. Registered in England and Wales, company number 13714418. Registered office: 3 Buckland House, William Prance Road, Plymouth PL6 5WR.
Get updates from Talamo
A monthly email with news, research and event updates for SENCos, schools and families.
© Talamo Limited 2026. Registered in England and Wales, company number 13714418. Registered office: 3 Buckland House, William Prance Road, Plymouth PL6 5WR.
Get updates from Talamo
A monthly email with news, research and event updates for SENCos, schools and families.
© Talamo Limited 2026. Registered in England and Wales, company number 13714418. Registered office: 3 Buckland House, William Prance Road, Plymouth PL6 5WR.