Each week focusses on one topic, and students are taught the content, work through tutorial problems, and complete a practical experiment. One of the reasons for this approach has been to help students prepare for university study. Previous research at Barker has found that students have valued this approach saying it improved their capacity for planning, understanding concepts, and reducing stress (Johnston & Hill, 2018; Hill et al., 2019).

A particularly important part of the model was the weekly tutorials. In a booklet provided at the start of the term, students received sets of problems related to each week’s content. Their homework was simply to produce a full set of worked solutions to that week’s problems, due at the start of the following week. They could complete these problems at any time, but had a dedicated tutorial lesson mid-week where they worked collaboratively on the problems for one hour with teacher support. They also had the final answer to each question (but not the full worked solution) to ensure they were correct, and access to after-hour teacher assistance.

An important feature of this system was that if a student did not complete all problems, saying they did not know how to complete the final few, the homework was not considered complete.

Students were therefore expected to use the available support to work through these challenges rather than avoid them.

In the first week of term, most students considered this approach harsh. However, it proved effective because it encouraged the behaviours we wanted to develop. Requiring the students to complete the questions by the beginning of the following week ensured they practised solving problems, persisted when they got stuck and sought help when they needed it. Quickly the students agreed that this is a very helpful learning strategy in Physics.

More recently, the increase of generative AI has made us reconsider this system, and it has also helped clarify the pedagogical question of what we actually want the students to be able to do. Previously, a student would spend considerable time wrestling with the question, discuss it with friends, or come see the teacher for further guidance – all behaviours we want to encourage as they support a deeper understanding of the question. However, as most of the physics problems students are asked to complete can now be solved very easily by the major AI platforms, a student who is stuck on a problem can ask AI for a solution, read through it, decide that it makes sense, and write it down. As a result, a completed tutorial no longer provides clear evidence of that understanding.

This led to us to question whether completion of all the tutorial questions should remain the measure of learning. Instead of checking only for completion, we are moving to give students one question from the tutorial in class at the beginning of the following week to be completed independently, under closed book conditions.

If they are able to solve the problem, the homework, or rather the learning, is considered complete. If they cannot, the homework is considered incomplete, indicating that they need more time and support learning how to solve these types of problems. At Barker, this can include attending a study hall at lunchtime, giving students another opportunity to work through the material before demonstrating that they can do it.

There are some practical issues to work through, including whether the tutorial should contain fewer questions, or whether the weekly check should be drawn from a designated subset of questions. We also need to be clear with students about the purpose of the process. The goal is not to memorise the answers, rather to develop the ability to solve a range of physics problems on the topic from that week That, after all, was the purpose of the homework in the first place.

In relation to AI, the question for the students then becomes less “can AI help answer this question” and more “Can AI help me learn how to answer questions like this for myself”. Therefore, students are incentivised to use AI platforms that will support the process rather than provide the answer. If they use AI they will use it as a tutor, rather than a solution-generator.

On reflection, this change represents good pedagogy irrespective of AI. The goal of student understanding has not changed. However, AI has made the difference between completing work and genuine learning easier to see.

References

Hill, M., Johnston, D., Taylor, N., & Woolley, D. (2019). Collaborative teaching to prepare Year 11 Physics students for university—Lessons from the second iteration. Learning in Practice, 3(1).

Johnston, D., & Hill, M. (2018). New syllabus, new approach: Preparing Physics students for university studies with a structured approach to the Year 11 course. Learning in Practice, 2(1), 22–27.