Assessing the potential to teach: key insights from our Future Teaching Scholars assessment centre

04 October 2020

As part of our commitment to research and development, we have invested in the review of our Future Teaching Scholars Assessment Centres. Our new Data Insights report examines the validity and reliability of the assessment centres, and correlations between candidates’ performance in different aspects of the initial assessment and later classroom performance. Here, we summarise our findings and their significance for teacher recruitment, and provide seven recommendations for improved teacher selection.

Recruiting and retaining talented teachers is a challenge for education systems around the world. According to UNESCO estimates, 68.8 million teachers will need to be recruited globally to meet Sustainable Development Goal 4. However, simply recruiting teachers will not be enough to meet this challenge:  we need to recruit high quality teachers who provide high quality lessons to improve learner outcomes – and who will remain in the teaching profession. There is also a crisis in teacher retention.  Many teachers leave teaching early in their careers. Furthermore, high staff turnover constitutes a major drain on the resources available to develop a strong education system. In the US, for example, schools lose between $1 billion and $2.2 billion each year in attrition costs as teachers move between schools or leave the profession altogether. In England, where early career teachers often make the decision to leave within the first three months of starting to teach, workload can heavily influence this decision. There is much that can be done to reduce teachers’ workload (to find out more, read our report on teacher workload for the UK Department for Education here), but the personal characteristics of teachers also appear to be important predictors of both turnover and success.

Our Future Teaching Scholars programme seeks to recruit high-quality maths and physics teachers in England, and has already demonstrated success: by the end of their first term in teaching, Scholars were already performing in line with many expectations for the end of a full year of teacher training. The Future Teaching Scholars (FTS) programme was established to help address the challenge of teacher recruitment and retention for maths and physics – an area for which recruiting and retaining talented staff has been particularly problematic. It is a six-year route into teaching, comprised of three-year undergraduate degree, alongside additional training, conferences and online learning, before an initial teacher training year (as an employed unqualified teacher), and two further years of support after attainment of Qualified Teacher Status.  The first cohort of scholars qualified as teachers after their ITT year in 2019/2020. As our research has showed, by the end of this year, they already demonstrated many areas of  performance expected after this full year of training by the end of the first term, placing them approximately one or two terms ahead of their peers.  To recruit the Scholars, the FTS programme uses an assessment centre, a common tool in recruitment in the UK. In addition to a competency-based interview, a reflective activity and a group problem-solving exercise, the assessment centre uses classroom roleplays, assessed by two practitioners – one of whom is a serving subject-specific teacher from an outstanding Teaching School – which seeks to measure a candidate’s innate ‘mental set’ prior to teaching. Specifically, this includes their ability to monitor and quickly identify a potential problem and take action (referred to by some education researchers as ‘with-it-ness’), and their emotional objectivity, amid a broader framework of competencies and values.

Although assessment centres are widely used, there is a lack of studies of education assessment centres and their validity and reliability. If this method of selection is to be utilised effectively, we need to know how best to use them to assess teacher retention and candidate competences – and measure these with validity and reliability. And if such areas can be measured, can they predict teacher effectiveness, teacher retention, or student outcomes? In response to these questions, Education Development Trust undertook a rigorous data analysis on its FTS assessment centre to help ensure its effectiveness and value for money. In this analysis, we measured not only the validity and reliability of the centres (analysing internal consistency and construct validity), but also correlations between performance in different aspects of initial assessment and later classroom performance (predictive validity).

 Our findings suggest that the Future Teaching Scholars assessment centre appears to be able to make valid predictions about classroom performance years in advance of an individual being employed in a school. Our analysis provides some insights as to which candidate competences or attributes appear to predict stronger classroom performance. We found the Future Teaching Scholars assessment centre to be internally consistent and have construct validity in respect of candidates’ behavioural competence and their application of subject knowledge in the classroom, as well as in relation to their opinions about becoming a teacher and working with young people.

Our analysis suggested that, in terms of later classroom ability, teacher training programmes should pay greater attention to the observation of a candidate’s skills (for example, in a classroom simulation), than the extent of their desire to work with young people. Specifically, the candidates’ ability to explain subject-specific concepts – and their problem-solving abilities – at the assessment centre appeared to predict their later teaching performance in areas such as structuring and designing learning, maximising learning time, high expectations, and giving feedback to learners. The candidates’ passion for their subject similarly predicted their ability to use learning time well. We also found that candidates who scored highly for empathy during classroom and group problem-solving simulations were less likely to drop out of the programme prior to being employed in a school in their fourth year. Conversely – and counterintuitively – candidates who expressed the highest levels of passion for working with young people (during the competence-based interview) were more likely to perform less well in the classroom once in teaching.

These findings are potentially far-reaching because teacher training programmes often put great store during recruitment interviews on a candidate’s passion for working with children. The predictive validity analysis suggests that recruiters should place more emphasis on the observation of a candidate’s actual skills (during a classroom simulation and ability to reflect afterward) rather than the strength of their motivation to work with young people.

While the assessment of the performance of the Scholars in the first four years of the programme has informed this analysis to date, they will continue to be assessed until the end of their third year as serving teachers. At that point, further comparisons can be drawn between test centre scores and their ongoing classroom performance, as well as teacher retention rates and their students’ examination outcomes. However, in the meantime, we believe that the data analysed in this study provides us with sufficient evidence to make seven key recommendations for improved teacher selection, which we hope will help schools and other organisations to get the most out of potentially time-consuming and costly selection methods.

Seven recommendation for improved teacher selection:

1. Selection processes should include some form of behavioural simulation, such as a short teaching activity, to enable assessors to judge the candidate’s skills in practice.
2. Classroom simulations should include elements of unexpected challenge to draw out a candidate’s mental set (e.g. their ‘with-it-ness’ and emotional objectivity). For example, the ‘learners’ could be briefed in advance to provide challenges such as low-level disruption or non-understanding to test the candidate’s reactions.
3. Simulations should develop and apply a clear and consistent scoring rubric. This may involve training assessors or observers in the use of the rubric to ensure a consistent method of assessment.
4. Learners should be role-played by those with real classroom experience – for example, serving teachers. School pupils may also be used in the role-plays, provided they are well briefed and are able to be consistent in their behaviour or approach across candidates.
5. It may be helpful to identify candidates who display empathy and cultural awareness in initial teacher recruitment, as evidence suggests that they are more likely to remain in training.
6. In weighting candidates’ scores during a selection process, more emphasis should be placed on how a candidate demonstrates their abilities while engaging with others than on what they may say about themselves in an interview context.
7. The results of a selection process can be fed forward into a trainee teacher’s development planning, to ensure targeted support in areas where they have a greater need for skills development.