The research behind Performance Learning

Why we publish this page

If we make claims about results, you should be able to check the evidence. This page summarises each study in plain English: what was studied, who took part, what it found and its limits.

One thing up front: several of these studies are co-authored by our founder, Dr Tej Samani, and use data from Performance Learning’s own programmes. That is not the same as a fully independent evaluation, so we point out each study’s limits.

What do the studies say?

1. Bridging the gap between high and low performing pupils (2017)

Bridging the Gap Between High and Low Performing Pupils Through Performance Learning Online Analysis and Curricula. Tej Samani, Kaśka Porayska-Pomsta and Rosemary Luckin. Artificial Intelligence in Education (AIED 2017), Springer. Read it on Springer

• What was studied: the approach and tools Performance Learning uses to help teachers and learners develop metacognitive skills, and how the company developed and validated them step by step.
• With whom: the abstract describes Performance Learning’s work with front-line teachers and learners rather than a single trial group.
• Main finding: the paper argues that the approach has value in real classrooms, potential for research on metacognition and is ready for artificial intelligence, especially for modelling learners’ non-cognitive skills.
• Limits: this is a short conference paper that describes an approach rather than reporting a controlled trial. Dr Tej is the first author.

2. Students’ self-reports and exam outcomes (2018)

Leveraging Non-cognitive Student Self-reports to Predict Learning Outcomes. Kaśka Porayska-Pomsta, Manolis Mavrikis, Mutlu Cukurova, Maria Margeti and Tej Samani. Artificial Intelligence in Education (AIED 2018), Springer. Read it on Springer

• What was studied: whether students’ own reports about non-cognitive parts of learning, such as their feelings and lifestyle, could help predict their academic results.
• With whom: the paper analyses data gathered by Performance Learning in two UK secondary schools, involving 48 students aged 16 to 17 preparing for pre-university exams.
• Main finding: the authors describe it as a preliminary analysis. The results point to the predictive potential of these self-reports, to the importance of students understanding themselves as learners, and to the value of including self-reports in models of how students learn.
• Limits: the authors note that the main limitation is the small number of students. The study was led by researchers at UCL Knowledge Lab, and two of the five authors, including Dr Tej, were affiliated with Performance Learning.

3. Retention and progression in further education (2021)

Improving the Retention and Progression of Learners Through Intelligent Systems for Diagnosing Metacognitive Competencies – A Case Study in UK Further Education. Tej Samani, Ana Isabel Canhoto and Esin Yoruk. Book chapter, Springer, 2021. Read it on Springer

• What was studied: whether learners’ self-reports on metacognitive dimensions, collected with Performance Learning’s assessment technology, can help predict whether learners stay on and progress in their courses.
• With whom: a case study of one further education college in the north of England during the 2019 to 2020 academic year, where learners were assessed at the start and middle of the year.
• Main finding: the authors report that the results confirm the technology’s potential to identify learners’ metacognitive needs early, and that it helps learners develop ways to improve their retention and progression. The college reported estimated savings of up to £60,000 from the support it put in place.
• Limits: this is a single case study without a comparison group, and the savings figure is an estimate based on assumed retention. Dr Tej is the first author. Professor Ana Canhoto, a co-author, is a special advisor to Performance Learning.

4. Evidence, policy, education and neuroscience in the UK (2024)

Evidence, policy, education and neuroscience – The state of play in the UK. Michael Thomas, Paul Howard-Jones, Jeremy Dudman-Jones, Lucy R. J. Palmer, Astrid E. J. Bowen and Roisin Perry. PsyArXiv preprint, March 2024. The article was later published in Mind, Brain, and Education. Read the preprint on OSF

• What was studied: the state of educational neuroscience in the UK: how well researchers and teachers talk to each other, how approaches to improving outcomes are evaluated, what teachers think about using neuroscience in the classroom, and the government’s move to bring cognitive science into teacher training.
• With whom: the article draws on a survey of a representative sample of more than 1,000 UK teachers and a case study of one high school teacher.
• Main finding: it is an overview of the field rather than a single experiment. In discussing how interventions are evaluated, it uses Performance Learning as an example and reports the 2.4 grades result explained below.
• Limits: Dr Tej is not an author. Professor Thomas is a special advisor to Performance Learning, and the Performance Learning data came from Dr Tej’s own doctoral research.

Where does the 2.4 grades figure come from?

Students using Performance Learning improved by an average of 2.4 grades against their predicted grades (Thomas et al., 2024, an analysis of 940 students led by Birkbeck’s Centre for Educational Neuroscience).

The figure is reported in the 2024 article above, led by Professor Michael Thomas, Professor of Cognitive Neuroscience at Birkbeck, University of London, and Director of the Centre for Educational Neuroscience since 2010. Professor Thomas is a special advisor to Performance Learning. The data came from Dr Tej Samani’s doctoral research, which he made available to the authors.

Question	What the article reports
Who took part?	940 students aged 15 to 16 with low previous attainment, in schools across the UK
When?	Data collected between 2016 and 2018, before the pandemic
What was compared?	Actual summer exam grades against the grades their schools predicted from December and January mocks, with Performance Learning in between
Which subjects?	Averaged across maths, biology, chemistry, physics, history and geography
Result	An average of 2.4 grades above prediction (effect size 1.58)
Comparison group	314 low-attaining students in the same schools, allocated at random not to take part, improved by 0.1 of a grade on average

What are the limits of the 2.4 grades figure?

• It comes from our founder’s research. The data were collected for Dr Tej’s doctoral research, not by an independent evaluator.
• It is a comparison with predictions. School-predicted grades are themselves estimates, so the main result shows how far students beat their predictions.
• It applies to a specific group. The students were low-attaining 15 to 16-year-olds before the pandemic. Results may differ for other ages and starting points.
• It cannot show which part worked. The authors note that programmes combining many elements are hard to evaluate formally, so it is not clear which elements caused the change.
• It is an average. It describes a group, not what any one child will achieve.

What does wider evidence say about metacognition?

Performance Learning focuses on metacognition: helping young people understand how they learn, plan, check their progress and manage how they feel about learning. This is a well-researched area.

The Education Endowment Foundation’s Teaching and Learning Toolkit rates metacognition and self-regulation approaches as high impact for very low cost, based on extensive evidence, with an average of about eight months’ additional progress. The EEF also notes that this impact can be difficult to achieve in practice, and that teachers need support to teach these strategies well.

That is general evidence about metacognition, not an evaluation of Performance Learning.

How we use the research

How students feel, behave and manage their learning is at the heart of Performance Learning. The Foresight assessment looks at learning behaviours, wellbeing and attainment factors such as recall. See how schools and families have used it on our impact page, and meet our founder and advisors on Dr Tej Samani’s profile and our about us page.