The paths to enrichment

18th June 2026 by Timo Hannay [link]

Especially for pupils in Years 6, 11 and 13, the school year in England ends with a crescendo of formal testing. Yet education should be about much more than cramming and examining, and the government would appear to agree. Last November, to coincide with the publication of the Curriculum and Assessment Review, the Education Secretary announced that there would be "a new enrichment entitlement for all that includes civic engagement, culture, nature and adventure, and sport". Ofsted would also assess these activities during their inspections. The schools white paper that followed in February 2026 went on to say that the DfE would "publish the Enrichment Framework this academic year." It has now been released.

To help inform debate on this important topic, we present here an analysis of current extracurricular and enrichment provision among mainstream state secondary schools in England, as described on their own websites. We would like to express our thanks to the Gatsby Charitable Foundation for supporting this work.

In summary, we find that:

• School clubs and societies vary widely in type and scope, with Duke of Edinburgh (DofE), football, reading, drama and art being the most common.
• Based on our taxonomy of 75 club types, a median of 13 and a mean of just over 15 different kinds of club were offered per school. But this varied greatly by school type, with large, urban schools, and those with low deprivation, high attainment and/or high Ofsted ratings tending to offer more.
• There were also disparities in the types of clubs provided. For example, specialist biology, chemistry or physics clubs (but not general science clubs) were more common in schools with lower deprivation, higher attainment and better Ofsted ratings. Maths, computing and STEM clubs also shared some of these trends.
• The analysis revealed regional preferences too, from cycling clubs in the North East to rugby in the South West.
• Among schools with high deprivation, breakfast clubs were the most over-represented (while also being a rather different kind of 'club') and Duke of Edinburgh awards were the most under-represented.
• Many other trends were apparent. For example, schools with high proportions of ethnic-minority or EAL pupils were more likely to provide debating and university-entrance clubs, but less likely to provide rugby or role-playing games. Rural schools were more likely to offer DofE, but less likely to offer basketball. Catholic schools had many more activities concerned with faith or singing, but very few of them had an LGBTQ+ society.

To find out more, read on.

First a few words about our methods. The information analysed here was obtained from school websites by an automated web crawl conducted in January 2026. This resulted in a cache of 2.3 million web pages, PDF documents and other file types. Plain text was extracted from these and filtered based on its semantic content in order to identify pages that putatively contained information about school clubs or societies¹. This resulted in approximately 78,000 pages, which were then fed to OpenAI's gpt-5.4-mini large language model (LLM) along with a prompt giving instructions to list any specific school clubs or societies unambiguously mentioned in the text. The resulting lists for each school were then keyword-matched against a manually curated taxonomy of 75 club types, providing a matrix showing the presence of absence of each club at each school. This was supplemented with manual spot-checking, especially for schools or clubs with unusually high or low numbers. The LLM's performance was found to be comparable to that of a diligent human curator, with very few blatant errors. Of 3,446 mainstream state secondary schools in England, we obtained substantial website content from 2,945 (85.5%) and at least some club information from 2,591 (75.2%).

Of course, this analysis also relies on school websites being comprehensive, accurate and up to date, which, to say the least, is not always the case. In addition, we have aggregated some clubs (eg, different types of instrumental music, foreign languages or niche sports) into single categories, which will tend to understate the total number of distinct clubs on offer. We also have no measures of participation in each club. For all these reasons, what follows should be considered indicative rather than definitive.

Figure 1 shows the proportions of school websites that mentioned each of the 75 club types in our taxonomy. The most frequently occurring were Duke of Edinburgh (DofE) awards, football, reading and drama, all of which were found in more than half of websites. Science clubs were of middling frequency, but separate biology, chemistry or physics clubs (including topics such as medicine, biochemistry and astronomy) were relatively rare.

If anything, these are likely to be underestimates, especially for the most common types of clubs. True, there is a possibility of false positives (ie, mentions of clubs that were not in fact offered by the school), but it seems more likely that we are under-reporting as a result of failing to find the relevant information on the school website, or it being absent in the first place. As a reference point, we understand that around 230-300 state schools in England provide a Combined Cadet Force, or CCF, club (depending on whether we include 'junior partners' that piggyback on another school's CCF provision). This is contained within our 'Cadets' category (which also includes other similar kinds of clubs). Of these, 198 school websites specifically matched the terms "Combined Cadet Force" or "CCF". This suggests that we found 65%-85% of relevant schools. Bearing in mind that we only had website data from about 85% of schools in the first place, this suggests that we identified most, though probably not all, relevant schools within our sample set.

(Hover over the bars to see corresponding data values.)

How many different types of clubs does each school provide, and how does this vary by school location or type? That information is shown in Figure 2, which provides both mean (blue columns) and median (red) values. Median values tended to be somewhat lower because mean values were influenced by a small minority of schools with very long lists of clubs. That said, both mean and median values displayed similar trends.

Across all schools, a median of 13 and a mean of just over 15 different types of club were offered (see rightmost columns), but this varied a great deal across different groups of schools. In general, large and urban schools, or those with sixth forms, offered wider ranges of clubs. Consistent with this, so did those with higher proportions of ethnic-minority or EAL pupils. The same was true of grammar schools, and those with fewer low-attainment pupils or higher Key Stage 4 attainment. Schools with lower levels of in-school deprivation or local deprivation, as well as those with higher Ofsted ratings, also had broader provision. And so did certain types of faith schools, as well as those in particular regions, notably London and the South East.

(Use the menu below to view different school types. Hover over the columns to see corresponding data values.)

So much for the range of clubs on offer, what about differences in various types of club? This is shown in Figure 3. For example, when looking by in-school deprivation level, there was very little difference in the provision of general science clubs, but huge disparities when it came to more focused biology, chemistry or physics clubs. Broadly similar trends can be seen when looking by local deprivation level, Ofsted rating or grammar school status. Maths clubs varied modestly by in-school and local deprivation, but showed bigger differences by prior or current attainment, and by Ofsted rating. Computing clubs varied little by local deprivation, but more by in-school deprivation and great deal by prior and current attainment, as well as by Ofsted rating. More generic STEM clubs showed less variation all round.

Grammar schools were more likely to offer chess and EPQs, but less likely to offer cookery or dance. Across the regions, cycling predominated in the North East, netball in the North West, citizenship in Yorkshire and the Humber, cadets in the East and West Midlands, social action in the East of England, business in London and the South East, and rugby in the South West. Make of that what you will. Above all, explore these results yourself using the interactive figure below – our brief summary above has barely scratched the surface.

(Use the menus below to view different club and school types. Hover over the columns to see corresponding data.)

Figure 4 makes it easier to get an overview of these trends. For each club type, it shows the difference between the prevalence across all schools and that across the subset of schools selected in the menu. For example, among those with high in-school deprivation, breakfast clubs were the most over-represented (while also being a rather different kind of 'club' offering an alternative form of 'enrichment') and DofE awards were the most under-represented activity. (Note also that most club types show up as under-represented, reflecting the fact that high-deprivation schools have a narrower range of clubs to begin with.) Those with low in-school deprivation show roughly the mirror image of this, and schools with high or low local deprivation show very similar patterns. Grammar schools tended to have better provision of subject-specific science, music, debating and drama, but were unlikely to have breakfast or homework clubs. Schools with strong prior or current attainment showed broadly similar, if less extreme, patterns. So did those with Outstanding Ofsted ratings.

Many other stereotypes were at least partially confirmed: schools with high proportions of ethnic-minority or EAL pupils were more likely to provide debating, basketball and university-entrance clubs, but less likely to offer rugby or role-playing games. Catholic schools had many more activities concerned with faith or singing, but very few had an LGBTQ+ society. Rural schools were more likely to offer DofE, but less likely to offer basketball. In London debating clubs were most over-represented, while in the North East it was breakfast clubs. Small schools were less likely to offer just about everything compared to large schools.

(Use the menu below to view different school types. Hover over the bars to see corresponding data values.)

To provide a further perspective, Figure 5 shows how the presence or absence of clubs varies by in-school deprivation (ie, pupils eligible for free schools means, or FSM) and by Ofsted rating. Clubs are organised into colour-coded groups, and circle sizes indicate the overall prevalence of each club type.

Interestingly, these groups of clubs showed a variety of trends. For example, differences in Art club prevalence varied mainly by Ofsted rating. In contrast, Performance clubs varied more by deprivation level. Life Skills and Support clubs varied by both factors: Duke of Edinburgh and Wellbeing tended to be less common in schools with higher deprivation, while Cooking and Breakfast clubs were more common. Debating and University clubs were more common in schools with high Ofsted ratings, while Homework clubs were less common. STEM varied by both factors, with Biology and Computing clubs skewing the most against high-deprivation, low-Ofsted schools, while general Science clubs were much more evenly distributed.

(Explore further by clicking on the figure legend to turn different groups on or off, or double-clicking to show one group on its own.)

There have been some previous surveys on extracurricular activities (for example, see this 2025 report from the DfE), and the recently published Milburn Report on young people and work mentioned that "Participation [in school enrichment activities] is lower for children and young people eligible for free school meals and children in need." But to our knowledge this is the first large-scale analysis of enrichment activities as reported by schools themselves on their own websites. We therefore hope that this report provides useful insights into an important, often under-appreciated and inadequately researched area of education. Our sincere thanks again to the Gatsby Charitable Foundation for supporting this work. We welcome comments, suggestions and questions to: [email protected].

Footnotes:

1. In short, we obtained the 'embeddings' for each file using OpenAI's text-embedding-3-small model and compared these to the average embedding of a manually curated training set of about 100 web pages containing information on school clubs, societies and other extracurricular activities. Files with a 'dot product', or similarity score, of 0.8 or higher were included in the further analysis.