our Chief Learning Officer, Dr Sue Sentance. She shared the work she and her collaborators have done to develop a research-based approach to teaching programming called PRIMM, and to work with teachers to investigate the effects of PRIMM on students.
As well as providing a structure for programming lessons, Sue’s research on PRIMM helps us think about ways in which learners can investigate programs, start to understand how they work, and then gradually develop the language to talk about them themselves.
Productive talk for education
Sue began by taking us through the rich history of educational research into language and dialogue. This work has been heavily developed in science and mathematics education, as well as language and literacy.
In particular the work of Neil Mercer and colleagues has shown that students need guidance to develop and practice using language to reason, and that developing high-quality language improves understanding. The role of the teacher in this language development is vital.
Sue’s work draws on these insights to consider how language can be used to develop understanding in programming.
Why is programming challenging for beginners?
Sue identified shortcomings of some teaching approaches that are common in the computing classroom but may not be suitable for all beginners.
‘Copy code’ activities for learners take a long time, lead to dreaded syntax errors, and don’t necessarily build more understanding.
When teachers model the process of writing a program, this can be very helpful, but for beginners there may still be a huge jump from being able to follow the modeling to being able to write a program from scratch themselves.
PRIMM was designed by Sue and her collaborators as a language-first approach where students begin not by writing code, but by reading it.
What is PRIMM?
PRIMM stands for ‘Predict, Run, Investigate, Modify, Make’. In this approach, rather than copying code or writing programs from scratch, beginners instead start by focussing on reading working code.
In the Predict stage, the teacher provides learners with example code to read, discuss, and make output predictions about. Next, they run the code to see how the output compares to what they predicted. In the Investigate stage, the teacher sets activities for the learners to trace, annotate, explain, and talk about the code line by line, in order to help them understand what it does in detail.
In the seminar, Sue took us through a mini example of the stages of PRIMM where we predicted the output of Python Turtle code. You can follow along on the recording of the seminar to get the experience of what it feels like to work through this approach.
The impact of PRIMM on learning
The PRIMM approach is informed by research, and it is also the subject of research by Sue and her collaborators. They’ve conducted two studies to measure the effectiveness of PRIMM: an initial pilot, and a larger mixed-methods study with 13 teachers and 493 students with a control group.
The larger study used a pre and post test, and found that the group who experienced a PRIMM approach performed better on the tests than the control group. The researchers also collected a wealth of qualitative feedback from teachers. The feedback suggested that the approach can help students to develop a language to express their understanding of programming, and that there was much more productive peer conversation in the PRIMM lessons (sometimes this meant less talk, but at a more advanced level).
The PRIMM structure also gave some teachers a greater capacity to talk about the process of teaching programming. It facilitated the discussion of teaching ideas and learning approaches for the teachers, as well as developing language approaches that students used to learn programming concepts.
The research results suggest that learners taught using PRIMM appear to be developing the language skills to talk coherently about their programming. The effectiveness of PRIMM is also evidenced by the number of teachers who have taken up the approach, building in their own activities and in some cases remixing the PRIMM terminology to develop their own take on a language-first approach to teaching programming.
Future research will investigate in detail how PRIMM encourages productive talk in the classroom, and will link the approach to other work on semantic waves. (For more on semantic waves in computing education, see this seminar by Jane Waite and this symposium talk by Paul Curzon.)
Resources for educators who want to try PRIMM
If you would like to try out PRIMM with your learners, use our free support materials:
If you missed the seminar, you can find the presentation slides alongside the recording of Sue’s talk on our seminars page.
In our next seminar on Tuesday 1 December at 17:00–18:30 GMT / 12:00–13:30 EsT / 9:00–10:30 PT / 18:00–19:30 CEST. Dr David Weintrop from the University of Maryland will be presenting on the role of block-based programming in computer science education. To join, simply sign up with your name and email address.