Research School Network: Reflections on ​“Modelling Evidence-based practice in initial teacher training: causal effects on teachers skills, knowledge and self efficacy” A review of the main findings from Ambition Institutes recently published research on modelling in teacher training

Alongside questioning and explanation, effective modelling is a fundamental part of an effective teacher’s toolbox. Nearly, if not all, classroom teachers understand the role modelling plays in bridging the ​“knowing-doing gap” in their classrooms, yet it is a practice the is starkly lacking in the majority of teacher training/​development programs. This void goes against the majority of empirical research into the value of modelling in teacher training, which argues that models play an important role in CPD and allow novice teachers the opportunity to attend to and notice important features of teaching practice.

To be fair there is a general shortage of, as Hill et al (2021) put it, effectiveness research into teacher education, primarily due to the challenges of isolating the effect of variations between programs, therefore the recently published research from Ambition Institute entitled ​“Modelling evidence-based practice in initial teacher training: causal effects on teachers’ skills, knowledge and self-efficacy” has fostered great interest.


The project is founded in the school of thinking that skills are ​“improvable abilities” and as such models can support the developing of teaching skills by helping novice teachers develop a mental image of the focal teaching practice. The research focused on participants skills in questioning for retrieval with approximately 90 participants (all enrolled in primary initial teaching training courses for the 22/23 academic year) were all given an evidence summary to read regarding effective questioning for retrieval, including suggestions such as cold calling, wait time and the use of partial hints when students say they ​“don’t know”. The participants then completed a classroom simulation (using avatars and scripts to ensure consistency) whereby their questioning for retrieval skills were judged by trained observers. Following this the participants were then split into 3 equal groups;


- Group 1 were asked to restudy the evidence summary before completing another classroom simulation
- Group 2 were shown a video model of evidence-based questioning before completing another classroom simulation
- Group 3 were shown a video model which was integrated with text guidance highlighting the teacher’s decision-making process (relating to the evidence) before completing another classroom simulation.

The authors wanted to test 4 hypotheses:


1) Exposure to a video model of some evidence-based teaching practice will improve pre-service teachers skills vs re-reading the evidence base
2) Exposure to a video model in which important aspects of the practice are highlighted and underlying knowledge is stated will improve pre-service teachers skills vs exposure to similar videos without the highlighting aspect
3) Exposure to a video model of some evidence-base detaching practice integrated with underpinning knowledge will enhance pre-service teacher’s knowledge vs just re-reading the underpinning knowledge.
4) Exposure to a video model of some evidence-based teaching practice will increase pre-service teachers’ self-efficacy in their use of the practice vs simply re-reading the theory.

The findings of the study supported some but not all of the hypotheses above. For example, participants that only restudied the evidence summary (without any form of modelling) saw no significant improvements in their questioning between the two simulations, while participants allocated to either group 2 or 3 (receiving some form of modelling) almost doubled their questioning effectiveness scores. However, there was no discernible difference between the improvements of participants that just received the video model and those that received the model with the underlying theory and rational highlighted. This would seem to fly in the face of the theory that novice teachers may struggle to notice the fine nuances of practice in the complex cauldron of an actual classroom.

Furthermore, there was minimal difference noticed in participant knowledge of the underlying theory no matter what group they had been allocated, while reported self-efficacy scores (based on questionnaires completed after each simulation by each participant) showed similar improvements across all 3 groups.

So, we need to consider what this means for teacher educators and in school continued professional development.

The dominant message coming from the research is that teacher educators and school leaders should ​“seriously consider” incorporating models into professional development. The researchers suggest that programs might consider incorporating video libraries exemplifying good practice. This is something that we had begun doing at Durrington prior to the pandemic and is something that we have had on the agenda to return to. Our current video libraries consist of videos modelling high quality formative assessment and metacognitive instruction – however these are perhaps currently under utilised and also require a wider scope to cover other T&L priorities such as live modelling. The advantage of video models includes that they can be used at any time, be watched by multiple people simultaneously (which is not possible is doing an in-class observation) and, if using staff within your own school, can enable a sense of social norming – i.e. if they can do it so can I which, theoretically, should develop teacher self-efficacy.


Interestingly the common school of thought is that any video model would benefit from integrated explanation of the theory behind the actions modelled, however the results suggest that modelling with or without theory integration is equally successful. However, before schools and teacher educator programs go out and begin producing video after video without the theory integration based on these results, we need to remember that participants in this study had been exposed to an evidence summary of the theory behind questioning for retrieval, and therefore their theoretical understanding may have been higher than the staff schools and training providers will normally be working with. Schools therefore may need to consider how evidence informed their teaching staff are and/​or who they are aiming the models at (i.e. ECTs or experienced staff) prior to deciding whether to integrate the theoretical explanation into the model.

The form of the modelling also needs to be considered – for example longer videos of whole or substantial parts of lessons provide a more ​“authentic” model, however when working with novice teachers it is possible that the quantity of such models may result in them missing the most important parts of the model. Alternatively, schools (and many are) could choose to use models and practice opportunities outside of the classroom, in the mould of instructional coaching for example. This has its benefits of allowing the focus of the model to be at the centre of discussion and practice, without any external distractors or competing priorities, but can of course (if not subsequently implemented into the actual classroom) lack authenticity and impact in the classroom.

If you want to read the full research report, you can do so here.

Ben Crockett

Assistant Director – Durrington Research School