#231 Productive failure with Manu Kapur

Bibi Groot, Eedi's Chief Impact Officer, returns to unpack the second Eedi and Google DeepMind AI tutor trial: 1,179 students, four arms, ten weeks. She and Craig discuss where AI tutoring matched expert human tutors, the null result on Star Maths, an unexpected gender effect, tutor wellbeing, safety data and what comes next. View the show notes here: podcast.mrbartonmaths.com/235-building-an-ai-tutor-with-google-deepmind-the-sequel-with-bibi-groot-eedis-chief-impact-officer/

Sign up for your free trial at HoD Space at hodspace.com.Access the show notes here: podcast.mrbartonmaths.com/234-hod-space-ep5-ai-resource-creation-exam-marking-supporting-science-teachersIn this episode of the podcast, Jo Morgan and Craig Barton discuss:1. AI resource creation Craig leads on a topic he expects them to disagree about, with Jo describing herself as an AI sceptic who is nonetheless noticing that most of the resources in her latest Gems post show signs of AI involvement. Craig sets out how he works: putting the ethical questions about training data to one side, investing heavily in the first conversation, challenging the output, and asking the AI to summarise its design principles for next time. Jo reports a quick test of three tools on area of a triangle with exact trig values, and they discuss whether AI could ever match Don Steward, whether it will make teachers better resource designers, and the time-saving uses HoD Space members shared in a recent thread.2. AI and exam marking Jo brings this one, prompted by what she describes as the heaviest remarks and appeals season she has known, with parents now requesting scripts and errors still surfacing only at appeal. She asks whether exam boards should move to AI marking. Craig makes the case for examining as professional development, then draws on his conversation with Daisy Christodoulou about early automated essay marking, gaming the system and comparative judgement. They consider where AI might outperform human markers, why a mark lost to a machine may feel less forgivable than one lost to a person, and why Jo has declined an AI marking trial for her own department.3. Maths in science Jo picks up a HoD Space post on numeracy across the curriculum, a phrase she would rather replace with maths in other subjects. Drawing on her experience as both a parent and a head of maths, she describes science lessons using calculators, rearranging formulae and standard form well before the maths curriculum reaches them, and makes the case for each department knowing what the other teaches and when. Craig recalls a disagreement with Adam Boxer about lining up equals signs, and the pair discuss how to share approaches, such as avoiding "move it over the equals sign", without criticising colleagues' teaching.

Sign up for a no-commitment free trial at hodspace.com. Access the show notes here: podcast.mrbartonmaths.com/233-hod-space-ep4-teacher-talk-finding-good-assessments-and-the-most-stressful-times-of-the-year/In this episode of the podcast, Jo Morgan and Craig Barton discuss:1. Teacher talk Jo brings a member's post about developing teaching and learning across a department, where a passing line about getting teachers to talk less drew some strong replies. Craig argues that the quantity of talk matters less than its quality, describing how he records teachers' explanations and plays them back in coaching, the research on redundancy and mazes in speech, and Kris Boulton's near-silent examples in atomisation. Jo reflects on over-exampling as a trainee, the choice of first example when expanding double brackets, why she still values longer periods of practice with some classes, and the shortage of videos of real maths lessons.2. Assessments students have not already seen Jo raises a problem members are describing at both GCSE and A-level: students arriving at mock exams having already worked through the exam board's secure papers. She explains how this came to light at her school, why swapping in a different past paper no longer solves it, and how her department now writes its own papers by rebuilding old ones question by question, along with the workload, grade boundary and further maths difficulties that brings. Craig draws on his Inside Exams series to explain why a good paper is so hard to write, then offers to have AI produce one, a challenge Jo accepts.3. The most stressful time of the year Craig puts a question to Jo that will also go out to HoD Space members as a survey: is there a point in the academic year that heads of maths find more stressful than the rest? Jo notes that staff absence is stressful whenever it lands, then makes the case for four contenders in turn, September, the run-up to Christmas, the spring term and the first half of the summer term, before choosing her winner. Craig compares her answer with his own memories of exam season as a classroom teacher without a head of department's responsibilities.

Sign up for your free trial at hodspace.comIn this episode of the Mr Barton Maths podcast, Craig and Jo discuss three burning issues from the Hubspace community this week:How do you make homework work in the age of AI?How do you handle students moving sets and communicate this to parents?How do you stop students from being confused by the equal sign? View the show notes here: https://podcast.mrbartonmaths.com/232-hod-space-ep3-ai-and-homework-setting-students-and-the-equals-sign/

Craig speaks to Manu Kapur, Professor of Learning Sciences and Higher Education at ETH Zurich and the researcher behind productive failure. They cover what the framework is and is not, the mechanisms behind it, how long students should struggle, what AI tutors mean for learning before instruction, and the strongest challenges to his work. Access the show notes here: https://podcast.mrbartonmaths.com/231-productive-failure-with-manu-kapur/

Sign up for your free trial at HoD Space at hodspace.com.Access the show notes here: ⁠https://podcast.mrbartonmaths.com/230-hod-space-ep2-notes-in-exercise-books-offsted-and-a-level-class-sizes/In this episode of the podcast, Jo Morgan and Craig Barton discuss:1. What actually goes in the exercise bookPrompted by a HoD Space member's survey request, Craig puts four options to the community — I do, we do and independent work; I do and independent work; independent work only; or no exercise books at all — and reports where members landed. He sets out why he thinks copying down worked examples is a pace killer and an illusion of understanding, Jo explains why she models on the board but never asks students to copy it, and the two of them disagree about whether key words and definitions have earned their place in the book.2. What the new Ofsted framework means for heads of mathsJo asked HoD Space members for recent first-hand experience of inspection under the new framework, and reports back on what came out of it: deep dives have gone, middle leaders may not be asked to talk about their own department, and questions about the maths curriculum are landing with line managers instead. She and Craig work through what that means in practice — the workload around annotated seating plans, the pressure schools are putting on exercise books, the focus on gaps in Year 7 prior knowledge — and land on the same advice from two directions: make sure the person who line-manages you can articulate what happens in maths lessons and why.3. A-level class sizesA thread from a member who had seen their A-level groups grow turned into the episode's biggest surprise. Jo describes teaching 28 in further maths and hearing from a school with 32 in a room that seats 32, before setting out why this bites from three angles — the head of maths trying to recruit and carrying an ever-larger key stage five, the classroom teacher who physically cannot get round the room, and the student who was told A-level classes would be smaller than GCSE ones. Craig brings in the Campaign for Mathematical Sciences research on the scale of the problem, and Dylan Wiliam's warning that cutting class sizes only improves outcomes if teachers do something different with the time — and if the extra teachers you hire are as good as the ones you have.

Craig speaks to Joe Liemandt, principal of Alpha School and founder of Trilogy Software, about the model behind the school's two-hour academic day. They cover selection effects, what does and does not scale, the data the apps generate, the future of the teaching profession, and Alpha's plans for the UK. Access the show notes here: https://podcast.mrbartonmaths.com/229-ai-in-education-with-joe-liemandt/

Sign up for your free trial at HoD Space at hodspace.com. Access the show notes here: ⁠https://podcast.mrbartonmaths.com/228-hod-space-ep1-praise-tick-sheets-year-7-sets-first-departmental-meeting/⁠In this episode of the podcast, Jo Morgan and Craig Barton discuss:1. Praise, and the problem of the "invisible pupils"Prompted by a question from HOD Space member Megan, Jo and Craig dig into Pritesh Raichura's seating plan tick sheet — where a teacher aims for 30 to 50 individual praises a lesson — and ask whether that's realistic in a maths lesson built around long stretches of independent practice, before turning to how a head of department supports a teacher who wants to try something like this, and what it would take to roll it out across a team.2. Setting Year 7Jo explains why her school groups new Year 7s using SATs results rather than a baseline test, how two early end-of-unit tests are used to sense-check those groupings, and why keeping the whole business of setting deliberately low-key matters as much as getting the groups right.3. Planning your September INSET meetingJo makes the case that the first department meeting of the year is the most important one you'll run all year, and talks through what goes into hers — from lesson expectations and homework routines to the admin nudges teachers need and the single department focus that will shape the year ahead.

In an unscheduled episode, Kris Boulton analyses a trainee teacher's volume-of-prisms sequence. He and Craig Barton work through key design principles — simplifying the rule, stripping out the arithmetic, starting with the general case, and building testing and expansion sequences that force genuine mathematical thinking. Access the show notes here: podcast.mrbartonmaths.com/227-breaking-down-an-example-sequence-with-kris-boulton

Returning guest Zach Groshell, author of Just Tell Them, explains how he designs Direct Instruction maths apps for Alpha School. They cover the US school system, why most education apps fail, how atomisation and mastery work on a screen, the limits of data and motivation, and why he chose a different school for his own daughter. Access the show notes here: podcast.mrbartonmaths.com/226-ai-in-education-with-zach-groshell