Research
All the writing in one place. More research and essays will be added here as they are finished.
- 6 essaysCorporate SeriesLearning at work: whether training worked, how adults really learn, compliance as practice, data that helps people instead of watching them, and why workplace English stalls.
- What the Calculator Can Teach Us About AIAI that hands over answers harms learning. Like the calculator, it needs designing in: the learner designs and judges, the AI does the labour, and assessment checks the thinking without it.
- How People Learn: The Principles Behind the SystemThe ten learning-science principles the system is built on, what the research says about each, and where each one lives in the product, in schools and at work.
- 7 essaysK12 SeriesThe system around learning in schools: what gets measured, who checks the educator's judgement, what makes an educator effective, and where AI and EdTech fit.
These essays take the principles in How People Learn into training at work, and set out what measuring it properly would need. Read How People Learn →
- 1
Why Most Corporate Training Can't Show That It Worked
Key Takeaways- Attendance and enjoyment don't show learning.
- Start from the work.
- Measure the same way, without names.
- 2
One Brain, Two Settings
Key Takeaways- The same methods work.
- Prior knowledge is the real difference.
- Drop the myths.
- 3
Compliance as Practice, Not a Once-a-Year Box-Tick
Key Takeaways- Skills fade fast.
- Practice beats watching.
- Little and often.
- 4
Data at Work: A Check, Not a Verdict
Key Takeaways- Feedback on the task, not the person.
- Rankings and monitoring backfire.
- Design it in.
- 5
Why Workplace English Stalls
Key Takeaways- A weekly class is too little time.
- Unpractised language fades, and progress stalls at intermediate.
- Start from the real work.
- Conclusion
One Team, One Quarter: The Cycle in Practice
Key Takeaways- Start from the work, end in the work.
- Most of the cost is the learners' time.
- Run it with one team first.
It sets out how people learn only in passing, because that is covered in How People Learn. The system is built on both. A Corporate series applies the same thinking to adult and workplace learning. Read How People Learn →
- 1
Who Checks the Teacher?
Key Takeaways- The law asks schools to develop the whole person, but schools only check exam results.
- Teachers' judgements of students are often right, but they can be biased, and nobody checks them.
- When a teacher's unchecked judgement decides a student's future, poorer students lose out most.
- 2
What a Report Card Cannot Say
Key Takeaways- Habits like effort and self-control shape a student's life as much as grades do.
- Schools should record what students do in class and compare each student only with their own past.
- A record of student behaviour must never count towards a grade, or students will start gaming it.
- 3
The Case for the Treasury
Key Takeaways- Governments judge whether education worked by the tax people pay years after leaving school.
- By the time that tax figure arrives, it is too late to help the student.
- Schools already collect early warning signs, like attendance and grades, but governments never ask for them.
- 4
Good Design, Not Learning Styles
Key Takeaways- Learning styles have no evidence behind them, and labelling children "visual" or "hands-on" does harm.
- A good learning tool adapts to what a student already knows about the topic.
- A tool like that has to be built with teachers who know how students go wrong in the subject.
- 5
What Makes a Great Teacher, and Why It's Important
Key Takeaways- The best teachers know where their students will go wrong.
- A teacher's fluency, intelligence or confidence tells you little about how much their students will learn.
- Nobody tests or records a teacher's knowledge of student mistakes, before or after they are hired.
- 6
The Tool Was Never the Hard Part
Key Takeaways- A new classroom tool only helps when someone decides what it is for and who it will help.
- Time saved by AI gets filled with other work unless the school removes another task.
- AI tools work best when built on a list of common student mistakes, which exists for maths but not yet for writing.
- Conclusion
When a Number Starts to Count
Key Takeaways- When a number decides something important, like a grade or a job, people will bend it.
- Measure students often with little riding on each result, and have someone outside check the marks that still count.
- When someone outside checks a teacher's marks, the gap between poorer and richer students narrows.