Forecasting the Teaching Workforce in England, lessons learned

Lithuania, which developed the pilot Teacher Supply and Demand model, can learn from the take-aways of other countries that have already introduced such  model. Thomas Lockhart, a statistician in the Department for Education in England, agreed to share the experience on developing The Teacher Supply Model (TSM) in England, which is now used for approx. 10 years. As an expert Thomas Lockhart has also contributed to the development of teaching workforce forecast system adapted for Lithuania.

The Teacher Supply Model (TSM) is now used for approx. 10 years in England. What have been the main challenges towards this achievement?

The main challenge over the years has been ensuring that the right people are brought into the modelling process at the right time to ensure that the best possible decisions are made when adding new assumptions. For example, if a future policy is likely to impact on teacher supply, any assumption that feeds into the model should be evidence-based, and agreed & signed-off by the people involved with designing that policy.

Another key challenge we have faced is ensuring that people interpret it correctly, both internally and externally. For example, failing to meet targets does not necessarily indicate a fault in the modelling. Additionally, the model is produced for each year in isolation, if we miss a target one year we should not simply add the residual to the following year’s target, as many things could have changed within the system during that year which would affect what the next target should be.

By ensuring people understand how to interpret the model and understand it’s limitations, through considerable engagement with stakeholders internally and externally, we believe that we have built a model that is well respected and whose outputs can be trusted to be built from reliable and transparent assumptions.

Does the model outcome effect the student career choices? Has the information about the model been spread widely?

The model outcome may have an impact on student career choices, particularly in those subjects for which teachers are most in need. The model is used to inform analysis on where teacher bursaries should be allocated, so those subjects which may be the toughest to meet targets in, or where the skills are most in demand, are allocated more generous bursaries to help attract graduates into teaching.

The model is very widely known within the sector, largely due to it’s longevity and because the model and it’s results are published each year. The department ensures that there is considerable external engagement with the modelling process by holding quarterly meetings with a wide range of stakeholders. This includes attendees from teaching unions, training institutions, school leaders, and academics.

What measures have been taken to attract people to the teaching profession? Could you please comment more about the school-based teacher pathway?

One key measure to attract people into the teaching profession is financial incentives to train. Tax-free bursaries, worth up to £26,000 (~€29,000) for trainees in priority subjects, and several other financial incentives encourage recruitment of high quality teachers into the profession.

Also, people are given the option to train to become a teacher through various routes, depending on which best suits their situation. This includes training full or part-time or training within a university environment or within a school.

School-based teacher training can be via a School Direct route (‘salaried’ or ‘tuition fee’) or ‘provider-led’ when delivered by a School Centred Initial Teacher Training (SCITT) provider.

School Direct (salaried) is an employment-based route that provides trainees with a salary while they train, and they won’t need to pay any tuition fees. This route is available for both primary and secondary training courses, and is delivered by a partnership of schools.

School Direct (tuition fee) and provider-led courses run by SCITTs are non-salaried routes, funded by charging trainees a tuition fee and delivered by a school partnership. Eligible candidates can access a student loan to cover the tuition fee for non-salaried courses.  These courses are school-based, but programmes may also include training from other partners, such as a university, college, schools or other organisations, with the final recommendation for QTS made by the ITT provider.

How is the effective monitoring and development of the model organized?

The model goes through a very clear annual update process, with meetings scheduled from the beginning of the year up until July (when the results are produced) to discuss the assumptions that will feed into the model and to analyse the latest data on the school workforce that becomes available in April/May. During the second half of the year, further analysis is done on many of the assumptions that underpin the model to see whether they are still relevant and whether they should be changed.

All assumptions are taken to a group of people from across the department to discuss and agree. A Senior Responsible Officer, who has the ultimate responsibility for what goes into the model, chairs this.

In terms of modelling the accuracy of the model, this is very hard to do (if not impossible!), there are a number of other factors within the system which could either mask or exacerbate any issues with recruitment that the TSM could show. So, whilst we regularly attempt to assess the accuracy of the TSM, such analysis looks at the accuracy/effectiveness of the whole teacher modelling and recruitment system.

Are there any plans to adopt this model to other policy areas?

All UK government departments maintain models to estimate the impact of policy areas that are within their remit. The format of this model will be seen in many other government models as there is a range of principles that government analysts are expected to meet when building models.

The size and scope of these models will vary depending on how often they will be used and how critical the analysis is. Smaller pieces of analysis that feed into one off policies will be comparatively small whereas a model like this, which is a business critical model which needs to be updated each year, will be much more detailed and undergo considerable scrutiny.

Your experience working with Lithuanian representatives, any take-aways and recommendations?

I really enjoyed my time working with Lithuanian colleagues, it was a pleasure to work with people who were clearly dedicated to try and improve their system and try to learn as much as possible from other countries.

From an analysts perspective, it was clear straight away that the technical skills were there to produce complex analysis and to build a comprehensive and detailed model. As a consequence, my advisory role very much focused on how to get the systems in place to ensure that the model could be clear and transparent as possible and that all assumptions had been thought about.

It was also interesting to see the challenges faced from another education system, many of which are ones shared across many countries (particularly recruiting to STEM subjects).

My main recommendation taking this forward is that a system is put in place to ensure that this model is regularly updated and built upon. If it isn’t I can envisage that a similar project to build a new model will take place in the future, which would be unnecessary given the hard work that has already been put into this project.