What recent trends in graduate numbers could mean for the future of teacher supply

This blog was funded by The Nuffield Foundation.

New NFER analysis has been looking at trends in the number of graduates by degree subject, and the likely impact of these on the teaching workforce. It suggests that while there are fewer young people entering the labour market overall, higher HE participation rates mean there is a growing number of graduates that could become teachers. Many teachers, including in STEM shortage subjects, come into teaching with a range of degree subject backgrounds that differ from their ITT subject. Indeed, some degree subjects may contain large, untapped pools of potential teachers.

Given the on-going current teacher recruitment crisis, how could policy be focussed on supporting and enabling more graduates from a wide range of degree subjects to enter teaching?

Has the pool of potential teachers been falling or growing over time?

In a recent blog post, my colleague Dawson McLean struck a note of optimism about the trends in career aspirations of today’s young graduates (or ‘Generation Z’). Our analysis found that Gen Z is more likely to desire careers that are pro-social (i.e. about helping others or contributing to society) than previous generations. Given the current teacher recruitment challenges, it is therefore somewhat reassuring that pro-social career intentions have historically been associated with being more likely to become a teacher.

There may be a wide pool of potential teachers in the current generation of new graduates who would be attracted into a teaching career if it was more competitive with other jobs in terms of, for example, pay and flexibility.

However, it is not just the types of careers that cohorts of young people want that matters for teacher supply, but also the sheer number of young people entering the labour market. It has been pointed out by some commentators that, because of demographic trends, the size of the current cohorts of young people are historically small. This could be creating a problem for teacher supply, because a small pool means fewer potential teachers and fiercer competition.

But the data only supports this explanation to a point. Data from the Department for Education confirms that the number of 19-year-olds is currently at a low-point in recent history. However, as teaching is a graduate profession, it is the number of young people coming out of higher education (HE) that really counts. Because of a steadily rising HE participation rate over the last decade – rising from 35.1 per cent in 2010/11 to 46.9 per cent in 2022/23 – the number of young people per cohort in HE has actually been rising while the total number of young people per cohort has been falling.

This suggests that analysis of cohort size for understanding the pool of potential teachers needs to go deeper than simply high-level population trends. Secondary teacher training in particular – which is also where the most acute teacher shortages are – draws on graduates from a variety of different subject backgrounds. The underlying trends in graduate numbers by subject therefore also reveal much about the changing pool of potential teachers.

Trends in graduate numbers by subject reveal a nuanced picture for the teacher supply pipeline

We combined data on UK graduate numbers by subject from the Higher Education Statistics Agency (HESA) and undergraduate applications by subject from the Universities and Colleges Admissions Service (UCAS) to measure these trends. We aggregated the granular subject data to a set of bespoke subject groupings that best matched groups of subjects in teaching. We also used data from the linked ITT Performance Profiles and School Workforce Census (SWC) to establish which degree subjects (using our bespoke subject groupings) are the most prominent feeder subjects for each initial teacher training (ITT) subject.

A limitation of combining this data was the different national coverages: HESA and UCAS data covers the UK while the ITT/SWC data only covers England. This meant caution in combining the data from the two sources. However, the large size of England within the UK and the similarity with other UK nations is likely to mean the data is indicative of the trends within England, even if not perfectly aligned. Further, a new system for coding HE subjects was implemented in 2019/20, which disrupted the consistent time series of data and therefore introduced complexity to our method. The appendix explains more details.

Below I summarise three of the key takeaways from our analysis.

1. Some subjects are more likely to produce teachers than others

We compared the number of UK graduates in each subject group with the number of first-year teachers that had degrees in the same subject groups, to estimate the relative likelihood of a graduate from a particular subject group becoming a teacher. Because of the differences in coverage, the analysis does not yield a straightforward ‘rate’ of entry into teaching. Instead, we express the data as an index, comparing the ratio of the number of UK graduates to the number of teachers in England within the subject group, and comparing that to the overall ratio. Therefore, a subject with a value greater than one is more likely to produce teachers than the average subject, while a subject with a value of between zero and one is less likely to.

The data shows a wide variation between subjects. As might be expected, degree subjects that have a natural vocational link to professions other than teaching are less likely than other subjects to produce teachers: good examples are medicine, architecture and engineering. The Department for Education (DfE) recently launched a pilot aimed at attracting more engineers into the shortage subject of physics, on the basis that the subject background is similar and not many engineering graduates tend to become teachers. Our analysis finds that 18 per cent of physics teachers have an engineering degree, suggesting that it is already a good source of potential teachers.

Subjects with a high relative likelihood of graduates entering teaching include languages, English, theology and religious studies and geography. This is consistent with evidence from surveys of undergraduates about their intentions to enter teaching. These subjects have less obvious links to other professions and are also mostly subjects that tend to have lower average earning potential (the exceptions being languages and geography). This is likely to make the teaching salary – which does not vary by subject – relatively attractive, thereby driving a greater relative likelihood of entry.

2. Maths teachers come from a wider variety of degree backgrounds than in other subjects

Perhaps surprisingly given that it is a shortage subject in schools, mathematics graduates have a slightly higher relative likelihood of becoming teachers than the average graduate. However, maths is a core subject in the school curriculum and a popular A level subject, meaning that schools need many more teachers than mathematics degree programmes alone can provide.

The quantitative skills developed on many degree courses – such as economics, psychology and some business-related – courses also mean there is a suitable pool of graduates from wider sources. Maths teachers display the widest variety of degree backgrounds out of any ITT subject. Our analysis finds that just 38 per cent of new maths teachers have degrees in maths, with 11 per cent having degrees from the ‘business and administration’ group, seven per cent in engineering, six per cent in economics, six per cent in education and four per cent in psychology.

Other shortage subjects also gain a number of their new teachers through a variety of related degree routes that are not the main degree subject linked to that subject. For example, 18 per cent of physics teachers have an engineering degree and ten per cent of computing teachers have a degree in business and administration.

Many chemistry teachers have a biology degree, most likely biochemistry that combines the two disciplines, and seven per cent have a medical degree.

3. The recent trends in STEM graduate numbers have generally been positive for shortage subjects in teaching

We examined the overall trend in the numbers of graduates for each ITT subject, weighting the trends from each of the underlying degree subject according to the extent to which the subject is a feeder degree subject for that ITT subject. The figure below summarises the recent trends for five shortage subjects. The forecast is a prediction based on applications data rather than graduate numbers (i.e. the trend in applications received is used as a proxy for what might happen to graduate numbers once those cohorts get to graduation). It is important to note that these predictions forecasted future graduate numbers with a low degree of precision when tested on known data, so these predictions should be treated with caution.

Graduate numbers in the feeder subjects for four STEM shortage subjects – physics, maths, chemistry and computing – have grown modestly since 2014/15. Graduate numbers in the feeder subjects for physics, maths and chemistry teaching rose by between five and eight per cent between 2014/15 and 2022/23, while the number of graduates in feeder subjects for computing teaching rose by 27 per cent. This suggests that the pool of potential graduates has not shrunk for these subjects, implying that this is not an explanation for the recruitment difficulties in these subjects.

However, in contrast the number of graduates in feeder subjects for modern foreign language (MFL) teaching – mainly languages degrees – has fallen steadily over time, by 24 per cent between 2014/15 and 2022/23. This is likely to have contributed to MFL recruitment being increasingly challenging in recent years despite a high bursary. A further decline in future is likely to make recruitment of MFL teachers even more challenging.

A focus for policy attention should be subjects that are large and growing, with a low relative likelihood of entering teaching

This data invites the policy question of whether there are any degree subjects from which more graduates could be attracted into teaching than currently. High-value candidate subjects to consider would be degree subjects that:

• currently provide few teachers
• have a clear link to a shortage subject in need of more teachers
• ideally, are subjects that have a large and growing number of graduates.

The figure below plots our degree subject groups according to the growth in graduate numbers between 2014/15 and 2022/23 (horizontal axis) and the propensity of its graduates to become a teacher (vertical axis – log scale). The bottom right quadrant features the degree subjects with fast-growing graduate numbers and that currently provide relatively few teachers. Some of these have a large number of graduates (indicated by the size of the bubble) and are linked to key shortage subjects.

Some key candidate degree subjects emerge from this analysis:

Business and administration degrees is the largest subject group in our bespoke subject categorisation. The number of graduates has been growing fast in recent years and it tends to provide relatively few teachers. However, those that become teachers are not just business studies teachers: twenty per cent are maths teachers. Attracting more business graduates where their subject knowledge is good into teaching – i.e. particularly from quantitative disciplines such as accounting and finance – could contribute to filling maths teacher shortages.

Engineering is also a large degree subject, has grown in recent years and is a key feeder subject for physics. As mentioned above, there is already a Government initiative aimed at attracting more engineering graduates into physics teaching. However, engineering is also a considerable feeder subject for maths teaching: 45 per cent of the engineers who become teachers teach maths. Driving harder to attract engineering graduates into teaching could therefore benefit maths as well as physics.

Computing and ICT is a reasonably large degree subject and has grown rapidly in recent years, although is a subject that provides fewer teachers than the average. It is an obvious feeder subject for the shortage subject of computing, but also 12 per cent of computing graduates who become teachers teach maths.

Psychology is a large and fast-growing degree subject. While it is around average for the propensity of its graduates to become teachers, psychology graduates who become teachers teach a wide range of subjects. Many psychology graduates in teaching are primary teachers, but around 15 per cent are maths and science teachers. However, a substantial maths component within the psychology degree, which isn’t universal on such programmes, would be required to make for suitable maths teachers.

Medical and dentistry degrees are another large and fast-growing degree subject area. While these degrees have a very strong vocational track away from teaching, the science background required for medicine and dentistry may make for suitable teachers. Indeed, almost half of new teachers with medical-related degrees are science or maths teachers.

How can policy action capture these sizeable pools of potential teachers?

There is a pressing challenge of recruiting more teachers, especially in shortage subjects. While it is not straightforward to do so, policymakers should be considering the ways that could tempt and enable more graduates from non-ITT subjects, but whose subject knowledge from their degree would make them a suitable teacher in a shortage subject, into the teaching profession.

First, the Government and ITT providers should look at ITT admissions guidance and processes to ensure that teachers from a wide range of career backgrounds can enter teaching in shortage subjects if they meet the other selection criteria and can demonstrate good subject knowledge. Relevant subject knowledge may not be directly obvious from the degree subject they hold, potentially meaning this is challenging for ITT admissions staff to assess. For example, the quantitative content of a psychology degree is likely to be highly variable, but some applicants may have acquired enough knowledge to be good maths teachers.

Relatedly, enabling teachers to gain further subject knowledge as they train should also be supported with funded subject knowledge enhancement (SKE) courses. Again, assessing which trainees could benefit from SKE and which may not need it may be challenging based on degree subject alone, so their individual needs could be explored by ITT providers. Continuing subject knowledge enhancement courses could also be beneficial for early-career teachers once they enter the classroom.

Second, Government and ITT providers should consider how courses are marketed, ensuring that students on a wide range of degree courses with relevant subject knowledge for shortage ITT subjects are made aware of the attractiveness of a route into teaching. For example, explicit acknowledgement in marketing material that graduates with a wide range of degree backgrounds could be welcome in maths ITT may encourage more to apply. Support with applications for those with degree subjects that have a less obvious link to an ITT subject could also be considered, although who would provide such support or have the available resources to be able to offer this is not clear. Finally, ITT providers may gain some benefit from establishing links with key university departments in non-ITT subjects, which could generate new pathways for interested graduates.

Many thanks to Mateo Medellin and Dawson McLean for the data wrangling and analysis that informed this blog. This project was funded by the Nuffield Foundation, but the views expressed are those of the authors and not necessarily those of the Foundation.

Appendix

We combined data from a number of sources to establish the trends in UK graduate numbers by degree subject. First, we used HESA data from 2014/15 to 2018/19 (Table 22) and from 2019/20 to 2022/23 (Table 52). We converted the subjects into a bespoke grouping, combining some degree subjects into larger groups where they were less relevant to teaching (e.g. medicine, allied medical degrees and dentistry) and splitting out other groups where they were grouped together despite having links to different ITT subjects (e.g. physics and chemistry within ‘physical sciences’).

The two data sources used different coding systems for grouping subjects (the former used JACS and the latter used HECoS). There is no simple lookup relation between the two systems. We therefore harmonised our bespoke groupings as best as possible across the two coding systems, linking each bespoke group to each of its relevant one-and two-digit JACS codes and Tier 3 Common Aggregation Hierarchy (CAH) codes (CAH codes are aggregations of individual, related HECoS subject codes).

However, there were challenges in ensuring that each grouping had a consistent definition across time. Examining the data revealed that while many of the groupings demonstrated expected continuity, others were misaligned due to changes in subject categorisation. We therefore re-indexed the data from 2019/20 onwards so that it linked to the previous data and illustrated the trend, even though there is very likely to be some degree of misalignment in some groups. This means that the time series should be treated with some caution.

The currently available data on graduate numbers only covers up to 2022/23, while the current focus for teacher recruitment is those who graduated in 2024. We therefore set out to also use data on accepted applications from UCAS as a predictor for future graduate numbers. However, when we tested accepted application numbers as a predictor of future graduate numbers using historic data, we discovered that the correlation was positive but low. This implies that the quality of the prediction is also likely to be low. Therefore, while the positive correlation suggests that the prediction contains some information, the forecasted data using applications should be treated with plenty of caution.

Finally, we used data from the ITT Performance Profiles and SWC to be able to link our bespoke groupings of degree subjects to ITT subjects, according to the undergraduate degree subject of first year teachers (recorded in the SWC) by ITT subject (as recorded by the ITT-PP). We aggregated this data across ITT cohorts 2017-2022 to maximise the available sample size. This work was undertaken in the Office for National Statistics Secure Research Service using data from ONS and other owners and does not imply the endorsement of the ONS or other data owners.