College recruitment before, during and after the COVID-19 pandemic

16th April 2024 by Timo Hannay [link]

Update 16th April 2024: See also summaries on the Gatsby Foundation website and FE News.

For several years we have been gathering data about vacancies advertised on the websites of schools and colleges in England. The results have appeared in many analyses concerning school recruitment activity, including our joint annual reports with Teacher Tapp and The Gatsby Foundation, as well as occasional forays into roles such as technicians and other support staff.

This post extends the scope again by analysing in detail college recruitment activity in England between September 2018 (ie, about 18 months before the start of the pandemic) and the end of February 2024. It looks separately at further education (FE) colleges, sixth form colleges and university technical colleges (UTCs). Sincere thanks to The Gatsby Foundation for once again supporting this work.

For details of the methods used, see Footnote 1 and for a description of the taxonomies we developed and deployed to detect relevant positions, please refer to Footnote 2. As with our recent study of support and auxiliary staff in schools, but unlike our regular reports on teachers and technicians, this analysis applies our automated advert-identification system to our archive of web pages without checking each item manually, which would be impractical over such a large archive. The data are therefore very likely to include false positives (ie, mentions of positions outside the context of a recruitment advert) as well as false negatives (ie, failure to detect an advert, which is a potential error common to both approaches). We expect these errors to be relatively small (where we have checked for them, false positives are typically in the range 10-15%) and to some extent they should also cancel each other out. But more to the point, as long as any biases are reasonably small and consistent, they still allow for analysis of changing recruitment activity over time even if the absolute values might overstate or understate the true numbers of adverts. Because we can't be certain that all of the items detected here correspond to adverts, they should be considered 'putative adverts', though for brevity we refer to them below simply as 'adverts'.

The results show both that the effects of the pandemic have been substantial and in many cases are still ongoing. They have also varied considerably by college type. To summarise:

• FE colleges showed only a brief and modest downturn in online recruitment activity during the pandemic (notably in spring 2020), but an extremely large increase since then, with numbers of adverts up 50% or more in 2021 and 2022 compared to pre-pandemic norms. Things have since subsided to something closer to normal pre-pandemic levels.
• Unlike schools, FE college recruitment does not appear to be particularly seasonal, with roughly equal levels of activity all year round. However, before the pandemic slightly more recruitment (about 55% of the annual total) tended to happen in the second half of the calendar year (July to December); since the pandemic this trend has reversed, with more activity in the first half of the calendar year (January to June).
• The most common subject groups for FE vacancies were construction, engineering, digital, business and health. Their proportions did not change much during the course of the pandemic.
• Like schools, sixth form colleges showed more seasonality, with the majority of recruitment activity occurring in the first half of the calendar year and low levels during the summer. This persisted during and after the pandemic.
• However, unlike schools and more like FE colleges, sixth form colleges did not show much overall reduction in recruitment activity during the pandemic. They nevertheless did display big increases in the post-pandemic period, with unprecedented peaks in the springs of 2021 and 2022. Activity reduced in 2023, but remains well above pre-pandemic levels.
• UTCs are smaller in number, so provided less statistically robust results, but overall showed rather similar patterns to sixth form colleges. The main differences were that the post-pandemic increase in recruitment activity was even larger.
• These results suggest that colleges – especially FE colleges – show different patterns of recruitment to schools and, furthermore, that the COVID-19 pandemic has had substantial, in some cases ongoing, effects. It is not possible to discern from these results alone whether the recent trends are due to a greater propensity among colleges to advertise vacancies on their website, genuine increases in staff turnover or general tightness in British labour markets, but they seem likely to be caused by a combination of such factors.

 

Further education colleges

This section looks at further education (FE) colleges. These include land-based colleges as well as art, design and performing arts colleges, but we have omitted specialist designated colleges. The Association of Colleges (AoC) lists a current total of 171 of these in England. For the analysis below, we found 195 separate identifiers (Unique Reference Numbers, or URNs) in the Department for Education (DfE) database. The latter number is higher for various reasons, including the fact that the number of colleges fluctuates over time, that the URNs of individual institutions can change, and that some institutions have multiple URNs (eg, corresponding to different sites). We have included all URNs for all state FE colleges that were active for any time during the period analysed.

It may be informative to consider these results in conjunction with the DfE's Further Education Workforce data collection process, though the resulting public reports only provide snapshots of vacancies at the end of each academic year.

Figure 1 shows a breakdown by role type. Across all years, two-thirds of adverts were for teacher positions (red) and just under a fifth were for support positions (yellow), with the remainder made up mainly of coach (green) and instructor positions (blue). The proportion of teacher positions has been broadly stable over time, representing 64% of adverts in 2018/19, 65% in 2022/23 and 69% so far in 2023/24.

(Use the menu below to select different years. Click on the figure legend to show or hide different role types. Hover over the segments to see see corresponding data.)

Figure 2 shows a breakdown by subject group. Across all years, about half of all adverts were for learning support, or for subjects relating to construction, engineering, digital, business or health. These proportions tended to vary only modestly from year to year: compare, for example, 2018/19 with 2022/23 and with 2023/24 to date.

(Use the menu below to select different years. Click on the figure legend to show or hide different subject groups. Hover over the segments to see see corresponding data.)

Figure 3 shows how the numbers of adverts changed over time. Looking at all role types, in the pre-pandemic period before 2020 there was some seasonality, with sharp troughs in the winter holidays and a peak in the summer. There was also a conspicuous reduction in spring 2020, when the first national lockdown was introduced in England. But the most notable trend is the large increase in numbers of adverts from the late spring and summer of 2021, as the early-2021 national lockdown was eased. Activity remained at very elevated levels – around 150-200% of the pre-pandemic baseline – for well over a year and only since spring 2023 have they begun to return to something more like pre-pandemic levels.

Broadly similar temporal trends can be seen for both teacher and support positions. There is some evidence for it, too, among coach postions, but less so for instructor or leader postions, which, as we have already seen in Figure 1 above, are relatively small in number. Click here to see all roles again.

(Use the menu below to select a role type to view. Hover over the lines to see corresponding data values.)

Figure 4 shows the same data, but broken down by subject group. The patterns for the most common subject groups such as learning support, construction, engineering, digital, business and health were broadly similar, confirming that the overall changes were not just caused by disproportionate increases in a few subject areas. This is consistent with the relatively stable proportions already seen in Figure 2. Click here to see all subjects again.

(Use the menu below to select a subject group to view. Hover over the lines to see corresponding data values.)

Figure 5 shows the seasonality by quarter. For all role types, activity has been very evenly distributed across quarters, especially in years less affected by the 2020-2021 pandemic. The other noticeable trend is that before the pandemic there was a slight bias towards quarters 3 and 4 (green and blue), with about 55% of activity coming in the second half of the year. Since the pandemic this has been reversed, with a slight majority of adverts appearing in the first half of the year (red and yellow). These overall patterns are broadly similar for teacher and support positions. Other positions are too rare for a robust analysis and so are not included here. Click here to see all roles again.

(Use the menu below to select a role type to view. Click on the figure legend to show or hide different quarters. Hover over the graph to see see corresponding data.)

Figure 6 shows the numbers of adverts in each academic year superimposed, which enables easier comparisons. Weekly data for all subjects show broadly similar seasonality from year to year, with clear troughs during the winter holidays.

Viewing the cumulative numbers (once again, for all subjects) allows overall year-on-year changes to be seen more clearly. Compared to 2018/19 (red line), which was the last pre-pandemic year, activity in 2019/20 (yellow) was somewhat lower, followed by a rebound in late 2020/21 (green) and a huge increase in 2021/22 (blue). Advert numbers then fell somewhat during 2022/23 (purple) and are now tracking close to pre-pandemic norms so far in 2023/24 (black) – though these trends vary considerably by subject. Click here to see all years again.

(Use the menus below to switch between weekly and cumulative views, and to select a subject group to view. Click on the figure legend to turn individual years on or off; double-click to see one year on its own. Hover over the lines to see corresponding data values.)

 

Sixth form colleges

This section looks at sixth form colleges. These include only state-funded standalone colleges, not independent colleges or sixth forms that are integrated with secondary schools. We have also omitted 'sixth form centres' and '16-19 free schools', which are often associated with schools. This final list corresponds to 111 separate URN identifiers in the DfE database.

Figure 7 shows a breakdown by role type. Across all years, nearly 70% of adverts were for teacher positions (red) and a quarter were for support positions (yellow), with very small proportions of other kinds of position. In particular, instructor postions are all but absent. The proportions of teacher positions were not only slightly higher than for FE colleges (see Figure 1) but also somewhat more volatile, with a peak of 73% in 2019/20 (the first Covid-affected year), and a trough of 65% in 2021/22. So far, 2023/24 has been somewhere in between, at 70%.

(Use the menu below to select different years. Click on the figure legend to show or hide different role types. Hover over the segments to see see corresponding data.)

Figure 8 shows a breakdown by subject group. Across all years, over half of adverts were for learning support, or for subjects relating to science, digital, design, business and social science. These proportions varied noticeably from year to year: compare 2018/19 with 2022/23 and with 2023/24 to date.

(Use the menu below to select different years. Click on the figure legend to show or hide different subject groups. Hover over the segments to see see corresponding data.)

Figure 9 shows numbers of adverts over time. Looking at all role types, the seasonality here has tended to show troughs in the summer and winter holidays, and peaks in the spring – a pattern that is more similar to schools than to FE colleges. Curiously, the national lockdowns between spring 2020 and spring 2021 did not coincide with an obvious reduction in advert numbers for sixth forms, though there were noticeable post-pandemic increases, with much higher spring peaks in 2021, 2022 and 2023 than was typical in the pre-pandemic period.

Teacher and support positions show somewhat different patterns, though both seem to have increased post-COVID. Other position types are too rare to be meaningfully analysed. Click here to see all roles again.

(Use the menu below to select a role type to view. Hover over the lines to see corresponding data values.)

Figure 10 shows the same data broken down by subject group. Most major subject groups such as learning support, science, digital, design, and business showed post-pandemic increases, indicating that this change was broadly based. The main exception was social science, which shows a more ambiguous pattern. Click here to see all subjects again.

(Use the menu below to select a subject group to view. Hover over the lines to see corresponding data values.)

Figure 11 shows the seasonality. Across all role types, activity is concentrated in quarters 1 and 2 (red and yellow). This is driven in particular by teacher adverts, about 65-70% of which appear in the first half of the year (red and yellow). Support positions are less seasonal, though they have also tended to predominate in the first half of the year, at least since the pandemic. These patterns are unlike FE colleges (see Figure 5) but much more similar to schools. Click here to see all roles again.

(Use the menu below to select a role type to view. Click on the figure legend to show or hide different quarters. Hover over the graph to see see corresponding data.)

Figure 12 superimposes the numbers of adverts in each academic year. Weekly data for all subjects show broadly similar seasonality, with peaks in the spring and troughs in the summer and winter.

Viewing the cumulative numbers for all subjects, about 2,600 adverts appeared in 2018/19 (red line), the last pre-pandemic year. There was a slight increase in 2019/20 (yellow), though note that this had been running even further ahead of the previous year until the first national lockdown in spring 2021, whereupon the gap closed. 2020/21 (green) showed an increase on these levels and 2021/22 (blue) was much higher still. Advert numbers then fell back in 2022/23 (purple), but remained well above pre-pandemic norms and so far in 2023/24 (black) are once again on a similar track. Click here to see all years again.

(Use the menus below to switch between weekly and cumulative views, and to select a subject group to view. Click on the figure legend to turn individual years on or off; double-click to see one year on its own. Hover over the lines to see corresponding data values.)

 

University technical colleges

This section looks at university technical colleges (UTCs). The Baker Dearing Educational Trust, which represents UTCs, lists 44 of these as currently active, but we find 55 separate URNs in the DfE database. As for FE colleges, the discrepancy arises because the number of colleges can change over time, as can the URN associated with any particular college. We have included all URNs for all UTCs that were active at any time during the period analysed.

Figure 13 shows a breakdown by role type. Across all years, 75% of adverts were for teacher positions (red) and most of the rest were for support positions (yellow). There were very few coach (green) and instructor postions (blue). Even more than for sixth form colleges (see Figure 7), there was high year-to-year variability in the the proportions of teacher positions, from a peak of 83% in 2019/20 to a low of 70% in 2020/21 (both Covid-affected years). So far, 2023/24 is somewhere in between, at 73%. This higher year-to-year volatility may be due in part to the smaller number of UTCs compared to other college types.

(Use the menu below to select different years. Click on the figure legend to show or hide different role types. Hover over the segments to see see corresponding data.)

Figure 14 shows a breakdown by subject group. Across all years, over half of all adverts were for learning support, or for subjects relating to science, maths or engineering. These proportions have varied considerably from year to year: compare 2018/19 with 2022/23 and with 2023/24 to date. The proportion accounted for by learning support adverts seems particularly volatile, having risen from 6% in 2019/20 to 16% so far this year.

(Use the menu below to select different years. Click on the figure legend to show or hide different subject groups. Hover over the segments to see see corresponding data.)

Figure 15 shows how the numbers of UTC adverts varied over time. In the pre-pandemic period there were usually around 20 or fewer adverts per week, with troughs in the summer and winter holidays and peaks in the spring (as for sixth form colleges and secondary schools). The pandemic period starting in March 2020 did not coincide with obvious reductions in activity compared to previous years, but there were very large increases in 2022 and 2023.

These trends were broadly evident in numbers of adverts for both teacher and support positions. As for sixth form colleges, other position types are too rare to be meaningfully analysed. Click here to see all roles again.

(Use the menu below to select a role type to view. Hover over the lines to see corresponding data values.)

Figure 16 shows the same data, but broken down by subject group. Patterns for the most common subject groups such as science, mathematics, and learning support. Engineering is perhaps the main exception, with an very early increase from January 2021 onwards, just when the second national lockdown was coming into effect – though note that the sample sizes are relatively small. Click here to see all subjects again.

(Use the menu below to select a subject group to view. Hover over the lines to see corresponding data values.)

Figure 17 shows the seasonality. For all role types, activity is relatively concentrated in quarters 1 and 2 (red and yellow). The overall pattern is more similar to sixth form colleges (see Figure 11) than FE colleges (see Figure 5). As for sixth form colleges, teacher positions are more seasonal than support positions. Recruitment of the latter group has also shifted a bit towards the first haf of the year relative to pre-pandemic norms: compare 2022 and 2023 with 2019. Click here to see all roles again

(Use the menu below to select a role type to view. Click on the figure legend to show or hide different quarters. Hover over the graph to see see corresponding data.)

Figure 18 superimposes the numbers of adverts in each academic year. Weekly data for all subjects show roughly similar seasonality, with peaks in the spring and early summer, and troughs in the late summer and winter.

Viewing the cumulative numbers for all subjects, about 500 adverts appeared in 2018/19 (red line), the last pre-pandemic year. There was a very similar number in 2019/20 (yellow), but 2020/21 (green) showed an increase on these levels, while 2021/22 (blue) and 2022/23 (purple) were much higher still. So far, 2023/24 (black) continues to track at these unusually elevated levels. Click here to see all years again.

(Use the menus below to switch between weekly and cumulative views, and to select a subject group to view. Click on the figure legend to turn individual years on or off; double-click to see one year on its own. Hover over the lines to see corresponding data values.)

 

Hire and higher

It is clear that in college recruitment, as in so many other areas of education and wider society, the effects of the 2020-2021 pandemic still reverberate. FE colleges appear to have returned to something approaching normal, but sixth forms and UTCs continue to show unusually high levels of activity, at least judged by vacancies posted to their own websites.

Why is this? A potential confounding factor is the possibility that our tracking of college websites and vacancies has improved over time – we certainly make continual tweaks to the system. But the numbers of institutions (technically, their identifiers, or URNs) did not changed markedly from year to year (the variation was less than ±5% about the mean for FE colleges, less than ±3% for sixth forms and less than ±6% for UTCs, and in all cases the numbers tended to be slightly lower in more recent years). Perhaps colleges have become more likely to post vacancies on their websites. We cannot rule this out, and indeed the sheer scale of the post-pandemic increases seen for UTCs in particular suggest some change in behaviour. Of course, this in itself might be an indicator of increased difficulty in hiring. Beyond these factors, it seems overwhelmingly likely that genuine increases in staff turnover and vacancies have at least contributed to these trends, especially given recent tight labour markets. Indeed, the vocational orientation of many colleges may mean that they compete more directly with other sectors and are therefore even less insulated than schools from these broader employment trends.

We hope to be able to continue tracking college recruitment activity, and perhaps to supplement the analyses presented here with others, for example by region. In the meantime, if you have questions or comments then we would be delighted to hear form you: [email protected]. To keep up to date with more analyses like this one, sign up for our free monthly-ish newsletter. SchoolDash Insights subscribers can explore related trends for schools and colleges in the Recruitment section. Non-subscribers are welcome to request a trial account or a demo.

 

Footnotes:

1. The websites of all further education (FE) colleges sixth form colleges and university technical colleges (UTCs) in England were visited each weekday night and the contents compared to the previous day in order to detect new advertised positions. Positions were detected by comparing this new text against taxonomies of expected roles and subjects (see Footnote 2). This process does not capture all relevant vacancies because: (a) we only looked for roles that matched our taxonomy, (b) not all positions are advertised on college websites, (c) even when they are, they are not necessarily presented in a way that can be automatically detected, and (d) websites are sometimes unresponsive or unavailable. Note also that, as for schools, a small proportion of colleges recruit at a group level. These adverts are detected by the system but cannot always be reliably associate with a particular institution, so they are omitted from the analysis. The proportions of such colleges have been gradually rising over time and currently stand at 9.7% of FE colleges, 5.4% of sixth form colleges and 9.1% of UTCs. This will also tend to result in under-counting. Conversely, there may be a degree of over-counting too, because: (a) some positions are cross-posted on multiple sites, (b) a single advertisement may contain multiple descriptions of a role that are easily confused for multiple roles, especially by our automated system, and (c) not all mentions of a role are in the context of a recruitment advert. Our ongoing analysis of secondary schools involves manual checking in order to correct for false positives, which typically results in corrections to about 15% of detected positions and a reduction of about 10% in total advertisement counts. Because this study of colleges extends back over many years, manual curation involved only spot checks, so a degree of over-counting is to be expected, albeit offset against potential under-counting. For all these reasons, the data presented here should be thought of as being based not on a comprehensive list of all vacancies but on a subset – effectively a survey of colleges, though one in which essentially all relevant institutions were contacted (via their websites) and in which the vast majority provided a response (ditto).
2. Positions were detected by looking on each web page for new text that comprised of both a 'title' term and a 'subject' term in close proximity (eg, "Lecturer in Engineering", "Sports Coach" or "Learning Support"). Roles without both a title and a subject were not included because this tends to lead to large numbers of false positives. The taxonomy for titles included 31 different terms grouped into five different types: Coach; Leader; Teacher; Instructor; Support. The results presented here use these types. The taxonomy for subjects used 583 terms organised into a hierarchical structure and ultimately aggregated into 27 groups: Agriculture, environmental and animal care; Business and administration; Catering and hospitality; Construction and the built environment; Creative and design; Design and technology; Digital; Education and childcare; Engineering and manufacturing; Foundation learning; Hair and beauty; Health; Historical, philosophical and religious studies; Languages, literature and culture of British Isles; Learning support; Legal, finance and accounting; Mathematics; Media and communication; Other languages, literature and culture; Performing arts; Protective services; Psychology; Sales, marketing and procurement; Science; Social sciences; Sport; Transport and logistics. Both taxonomies were developed in collaboration with The Gatsby Foundation using vacancies data from the Association of Colleges as well as from the college websites analysed here. We intend to continue developing and using them for future analyses.