ITT becomes more cosmopolitan

Over the past few years, the percentage of the total number of graduates training to be a teacher coming from the United Kingdom has fallen, year on year. On the other hand, the percentage of trainees on these courses from both EEA and ‘other’ countries has increased.  

YEARUKEEAOTHERKNOWN% OTHER% EEAEEA + OTHER
16/17236581295506254592%5%7%
17/18242231294532260492%5%7%
18/19265501422634286062%5%7%
19/20265621470806288383%5%8%
20/21314181747919340843%5%8%
21/22276281210823296613%4%7%
22/23200191201722219423%5%9%
23/24193638801053212965%4%9%
24/252058613811351233186%6%12%
25/262249215432082261178%6%14%

The table has been abstracted from the DfE data catalogue associated with the annual ITT census.https://explore-education-statistics.service.gov.uk/find-statistics/initial-teacher-training-census/2025-26

This year, trainees from countries in the ‘other’ group accounted for a record eight per cent of postgraduate trainees. Taken together with the percentage from EEA countries, some 14% of the current cohort of postgraduate trainees were from these two groups. The home student total was still 86%, but that is the national figure.

This influx of overseas trainees has helped the government meet more of its targets for secondary subjects than had it just had to rely upon home-based students to fill the places. These cosmopolitan students bring fresh perspectives that will help widen the experience of the home students they are studying alongside.

However, as my previous post suggested, these ‘overseas’ trainees are not likely to be spready evenly across courses, or across the country. A significant number will be on courses in London, while few will be on courses located a significant distance from the capital.

What matters more, is what happens to these trainees at the end of their courses. Will they be able to enter the labour market for teachers, and be provided work visas.as important, after training in England, will they want to teacher in this country or will they look to the rapidly expanding international school market for employment opportunities.

Interestingly, of the nearly 4,500 vacancies currently listed on the DfE job site, only 18 appear to say that ‘visas can be sponsored’.  No doubt, when faced with a great trainee and a vacancy that might prove a challenge to fill, attitudes might alter. However, none of the current physics posts sponsor visa students.

Why am I interested in this data? Mostly because the DfE seems to think its job is done with the publication of the ITT census, and the provision of a vacancy website.

Ever since I founded TeachVac in 2013, I have been of the firm belief that as local authorities recede into the background with regard to schooling, so central government needs to know more about the workings of the labour market for teachers. If all 3,500 non-UK trainees didn’t teach in state schools in England, and a number of UK citizens decided to teach overseas, what would be the implications for schools across England? And what would it do to the agenda of lifting young people out of poverty?

Ethnic minority trainee teachers: still huge regional differences in trainee numbers

1n the autumn of 1997, Baroness Estelle Morris, at that time a junior minister in the DfE, in the new Labour government of Tony Blair, opened a conference about recruiting more ethnic minority students to become a teacher. The conference was organised by the then Teacher Training Agency. That conference was held in East London, and was followed by two more in Leeds and Birmingham.

Fast forward to the ITT census produced by the DfE today, and ask the question: how successful has the campaign to recruit certain ethnic groups into teaching been since that first conference nearly 30 years ago? Initial teacher training: trainee number census 2025 to 2026 – GOV.UK

Looking at the group that has found most difficultly in becoming a teacher over the years – Black African/Black Caribbean – there still seem to be big challenges looking at today’s data. Whether these are because students from this ethnic grouping aren’t attracted to parts of the country where there are few of their compatriots or whether there are other reasons cannot be determined just from the numbers.

However, over 500 courses have no candidates recorded from this group in the data published in Table 12 today. Just over 900 courses have between one and four candidates from the ethnic group. A further 83 courses have the number suppressed as being too low, as it might allow an individual to be identified.

A quick review of courses with the highest percentage (over 50% of each course code) shows that 24 are courses run by providers in London; just three are from outside London, and for three the name does not provide a clue to the location.

Looking at the courses with more than 100 candidates from the Black ethnic group: four are located in London – two each from UCL and Teach First – and the fifth is a national SCITT.  

As might be expected, the University of East London, and several other London post 1992 universities, feature in the list of providers with between 25% and 50% of course numbers from the Black group, each with several courses in this percentage range. Most other pre-1992 universities and other post-1992 universities and the SCITTs in London have many of their courses in the 15%-25% group of providers. Few, if any, London providers feature in the list with zero percentage from the black group.

While it is good that courses in London do seem to be attracting applications from the Black ethnic group, there are still many courses in large parts of the country where that seems not to be the case. Does this matter? Would a ‘token’ representative on a single course in an institution be anything more than a token. Should we encourage such students to be trailblazers r should we accept that outside of the conurbations and a few university towns, graduates from the black ethnic group are still relatively rare.

I went to school in the 1960s with one of the few Black pupils in the school. He went on to become a teacher when Black teachers were even thinner on the ground than now, even in London.

So, there has been some progress, but not enough.

Teaching a global profession? What do the physics ITT numbers tell us?

My previous post contained the good news for the government in the headline data about their annual census of those on teacher preparation courses. Digging down into the details of the census, there is at least one worrying trend. https://www.gov.uk/government/statistics/initial-teacher-training-trainee-number-census-2025-to-2026

The percentage of accepted ITT candidates within each nationality group for selected subjects for 2024/25 and 2025/26

Percentage of accepted candidates
UK and Irish nationalEEA nationalOther nationality
2024/252025/262024/252025/262024/252025/26
Total88%86%5%5%8%9%
Primary94%94%2%3%4%3%
Secondary84%82%6%6%10%13%
STEM Subjects76%74%5%5%19%22%
Physics43%32%3%2%54%66%
English93%93%3%2%4%5%
Mathematics81%81%5%5%13%14%
  1. High Potential ITT (HPITT) route and undergraduate routes are not included in this data.
  2. Subject-level candidate totals will not sum to the total candidate number due to duplication caused by candidates applying for multiple subjects.

The footnote about undergraduate routes should not be of concern as there are relatively few such courses for secondary subjects, and the numbers on primary undergraduate courses have been declining over the longer-term.

Of much more concern is the decline in percentage of accepted candidates for physics from the UK and Ireland, down from 43% last year to 32% this year. This has been balanced by and increase from 54% to 66% for candidates from outside the UK and EEA areas.

As there has bene a dramatic increase in the numbers of trainees in physics, does this matter?

On these percentages, the increase in UK and Irish trainees has been from only around 185 last year to 220 this year. That seems like a very small number and worth investigating to see if I am correct?

If I am correct, then the key issue is, where will the trainees from the rest of the world be able to teach? Will the present government’s stricter policies on immigration mean that they won’t be able to teach in England, or as graduates earning a good salary will they be given visas?

Of course, they may choose to teach in the new British state sponsored selective school being established in both India and the UAE that was recently approved by the Labour government.

British Education is a global export, regardless of the PISA scores of home students, and the destination of trainees, both within the state and private systems, as well as overseas, is an important piece of information Minister should pay more attention to than they do at present.

The number of Uk trainees is likely to be boosted in physics by those training through the High Potential route (Formerly known as Teach First), However, the data for those candidates is not included in the census this year.

No doubt there is room for some interesting parliamentary questions about trainee teachers and where they come from and where they go on to teach, especially for those that receive bursaries and other financial support from the State.

Too many teachers?

Earlier today the DfE published their Annual Census of ITT trainees. Published each December, the census identifies the numbers on the various teacher perpetration routes and some background information about their gender, ethnicity, degree class and routes into teaching. Initial teacher training: trainee number census 2025 to 2026 – GOV.UK

The census provides a helpful indication to schools about the labour market for the following September recruitment. In this case, September 2026.

In recent years, apart for during 2020 and the response to the pandemic, trainees number in many secondary subjects have been lees than the DfE predicted numbers needed to fill vacancies. In the primary sector, falling rolls and erratic recruitment numbers have meant there has been less of a coherent pattern about the balance between supply and likely demand for teachers. Of course, much depends upon assumptions about the turnover in the labour market, and the behaviour of possible ‘returners’ to teaching when reviewing recruitment patterns.

So, what of the current 2025/26 cohort?

subject2024/252025/26
Percentage of Target at census date%%
Physical Education213202
Biology116151
Art & Design64128
Primary88126
History116125
Chemistry62118
Mathematics72113
Geography91111
English99106
Modern Languages4493
All Secondary6188
Computing3780
Physics3077
Classics24573
Design & Technology4070
Music4065
Religious Education7962
Drama4741
Business Studies1530
Other1514

The government can be pleased with some of the best recruitment levels to their targets in almost a generation – covid years excepted – but challenges still remain. Nine secondary subjects didn’t meet their target number, with business studies still recruiting poorly to teaching, along with drama and religious studies where the target was missed by a larger percentage than last year.

On the good news side, mathematic exceeded its target for the first time in a long while, and the increase to 77% of target in physics teachers is very welcome news.

There will be too many primary school teachers looking for jobs come September, and although course providers will be happy to have recruited 202% of the target for physical education trainees, this over-recruitment does beg the question as to whether recruitment controls should be once again considered as a deterrent to such significant over-recruitment?

Taken with the news, highlighted in my previous post, about attitudes to pay by serving teachers, the government can probably stop worrying abut teacher recruitment for the first time since 2012.

However, all is not good news, if the Curriculum Review is to be implemented in full, attention to recruitment in some subjects will be needed. In that respect, as already suggested by this blog in a previous post, removing the bursary from music seems like a daft idea. Yes, there was a 25% increase in outcome against target, but that still left a third of places unfilled. Music departments in schools are often small and cannot be easily covered by non-specialists, such as the spare PE teachers. Time to think again on the basis of these figures.

Gatsby Survey confirms importance of pay and working conditions for would-be teachers

A Gatsby funded study by a team at London’s UCL has researched assumptions about why people do -or do not- choose to become a teacher in the UK and the US. The findings were that extrinsic rewards drive career choices. The report found that in both countries, extrinsic factors such as salary, working hours and paid leave were the most powerful drivers of career choice. Altruistic motives did play a role – participants were willing to accept lower pay for roles with higher social impact – but these were consistently smaller than the influence of pay and workload.’ New research reveals what really attracts graduates to teaching  | Gatsby Education

These factors were even seen among those undergraduates who already said they were already considering becoming a teacher.

Replies to the UCL study suggested that increasing working hours beyond 40 per week to that of a typical teacher reduced attractiveness of teaching by 15%, and that teachers holiday entitlement increased attractiveness by 11%. Increased salary raised job attractiveness by 9%.

How do these findings compare with previous research? In May 1997, almost 20 years ago, and during another period of challenges in recruiting graduates into teaching, The School Teachers’ Review Body (STRB) commissioned the agency BMRB to investigate what factors influenced the attitudes towards teaching shortage subjects/ This was a small-scale study involving only 82 graduates compared with the 2,000 undergraduates, in both the UK and the US, surveyed by the UCL team.

BMRB students said that

  • Teaching should be a vocation
  • Those sampled felt not all were suited to be a teacher
  • There were serious concerns about both working conditions and stress levels
  • Pay was acknowledged to be a significant factor – although not a deterrent to those determined to teach, a better pay structure would make teaching more attractive to those considering other options.

The views BMRB found ‘were not specific to those studying the shortage subjects … but were common across the different subject areas. ‘

So, the common message from both studies, nearly twenty years apart, and of different participant sizes and survey methods, is that teaching must be competitive in regard to pay and working conditions to attract graduates in a competitive labour market.

Another study, in 2000, for the Office of Manpower Economic (OME), by Whitmuir Research, reported similar finding to the BMRB list, but added, disruptive pupils; lack of parental support and the cost of tuition feed to the list.

A large-scale study of 1,880 final year undergradues across 26 HEIs for the TTA in 1999, distributed through careers services, found more interest amongst women than men in teaching as a career, and amongst those in post-1992 higher education institutions.

A review of where applicants to teaching come from, based on DfE data through the common application process would be a sensible annual outcome in order to see if there are changes in the key undergraduate market with regard to teaching as a career.

It seems likely that the STRB knows the issue around recruiting into the teaching profession. The question every year is – will the STRB stand up to government on behalf of the children of this country and ensure that teaching is an attractive career for graduates across all subjects?

More men looking to teach

Today, the DfE published their first round of statistics about applications to train as a teacher on courses starting in the autumn of 2026. Generally, one has to be cautious about data from ‘applications’ and ‘offer’ statistics published in November, as this is very early in the application round.

However, with more than 20 years of data underpinning my remarks, I think it possible to say something.

Firstly, applications – and candidates may submit more than one – are up from 13,159 last November to 15,572 this year. Applications from men are up from 5,072 to 6,580, while those from women are up from 7,978 to 9,031. That equates to 1,052 more women applying, or an increase of 13%, but 1,508 more men; an increase of 30%. I cannot recall a time when the rate of increase in applications from men last outpaced those from women.

Part of this increase is probably down to the large increases in applications for mathematics, up from 1,657 last year to 1,929 this year. In computing, the applications are up from 509 to 841, and in physics from 1,694 to a staggering 3,277. All these are subjects that tend to attract more male than female candidates.

Aword of warning, before one becomes too carried away; applications from the Rest of the World are up from 3,540 last November to 5,120 this November. Might this account for part of the increase in male applicant in these subjects? Sadly, that cannot be determined from the published data.

Final year undergraduates are not yet swarming into teaching. No obvious concerns about loss of graduate jobs to AI from the 21 and under age group, where applications are actually down by 34 from 1,276 to 1,242. Presumably, studies still take precedence over job hunting.

However, there is a big increase in the 22-24 age group applying for teaching: up from 3,349 to 3,658 with nearly 200 of this increase from 22 year olds. Maybe summer 2025 graduates that are still job hunting are turning to teaching? There is little difference in interest in teaching from those over 45 years old. However, there has been a big jump (210) in interest from the 40-44 age group.

SCITTs is the only route to have seen fewer applications than in November 2025. This may reflect the fact that the SCITT route maty be less well-known to overseas applicants. Both teacher degree apprenticeships and PG teaching apprenticeships have seen significant increases in applications. It would be interesting to see this table by phase and subject.

On ‘offers’, it much depends upon how providers handle early applications. However, there is a trend with mathematics, computing, chemistry and physics all recording the highest ‘offer’ levels since 2013/14, whereas music has the lowest offer level since 2020/21. Most other subjects are close to where they would be expected to be, although biology, PE and geography are below where they might expect to be. PE probably over-recruited to current courses, and I would expect more caution there this year.

So, overall, a good start that should presage a good recruitment round unless something unforeseen happens.

6500 new teachers: wasn’t that a manifesto pledge?

The DfE has today published the annual Education and Training Statistics for the whole of the United Kingdom.  Education and training statistics for the UK, Reporting year 2025 – Explore education statistics – GOV.UKAs ever, there is a wealth of material and the devil is in the details.

The full-time equitant number of teachers in England increased between 2023/24 and 2024/25. The data for schools in England isn’t new, as it was first reported in June 2025 School workforce in England, Reporting year 2024 – Explore education statistics – GOV.UK

However, it is worth discussing the data again, as it will provide the basis against which any claims about increased teacher numbers will be judged.

Phase201920202021202122202223202324202425
Non-maintained mainstreamTotal769757731071695732557524074100
NurseryTotal119511601100106510751060
PrimaryTotal219960221365221230220265217460214575
SecondaryTotal204715209835213630216075217565219000
SpecialTotal242802502526005271402824029150
TotalTotal530800538415537285541690543930542355
Total maintainedTotal453820461105465590468435468690468260

The changes over one year and the whole of the time period are shown in this table

Totalchange on 202324change on 201920
Non-maintained mainstreamTotal-1140-2875
NurseryTotal-15-135
PrimaryTotal-2885-5385
SecondaryTotal143514285
SpecialTotal9104870
TotalTotal-157511555
Total maintainedTotal-43014440

It is worth noting that as the school population increased, so did the number of teachers in state-maintained schools. Thus, between 2019/2020 and 2024/2025 teacher numbers increased by 14,440, although the decline in teachers in the nursery and primary schools had already started.

However, by 2024/25 the total teacher workforce was some 14,440 FTEs larger than it had been in 2019/2020 as a result of the increase in the number of secondary and special school teachers.

Between 2023/24 and 2024/25, teacher numbers in England continued to increase across both the secondary and special school sectors, but the decline in the primary sector teacher numbers continued. The nursey sector showed little change, but employs few teachers in state nursery schools as opposed to nursery classes in primary schools.

There is a message here for anyone considering a career as a primary school teacher. Before accepting a place on a teacher peroration course; do some homework on job possibilities in the area of the country where you would like to teach, especially if it is not where you are training.

I doubt that we have yet seen the end of the decline in teacher numbers in the primary sector, and it will be a buyer’s market, even in 2027 when those applying for courses starting next September will enter the labour market in large numbers.

In the past, under such conditions, schools have preferred to employ experienced teachers leaving new entrants to look for posts in schools for which they may not have been trained. Often the jobs will either be in schools with a higher deprivation index score or small schools with mixed age classes. Neither of these teaching situations may have been encountered during a one-year teacher preparation course, and can be challenging for new teachers if not adequately supported.   

Don’t be afraid to ask about job prospects at interview, especially if you are paying your own tuition fees.

Has Labour gone mad?

Queen Elizabeth’s School, a selective grammar school in North London, is to open an affiliate fee-paying branch school in Dubai – becoming the first state school to open overseas. Queen Elizabeth’s School to open fee-paying school in Dubai | Tes

I am going to state my opposition to this proposal outright. If we had a sufficiency of high-quality teachers for all our schools, then I might, just might, look on this as part of the export drive using resources not currently needed for the home market.

But the blunt truth is that we don’t have enough qualified teachers for our secondary schools. It is bad enough private schools offering UK teachers jobs overseas, but most of them probably weren’t in the state system anyway.

Here we seem to have a state funded school spending leadership time becoming part of a global brand, and at the very least risking taking a couple of hundred teachers out of the UK system to teach middle class children in the UAE and India.

 Even if the investment is funded by Global Education, a company with a strong base working with universities and higher level vocational providers, I am not sure why a Labour government has allowed the DfE to approve this move?

I do think there should be a policy designed to maximise UK revenue from our strong background in education across the board, but a government’s first duty is to its own citizens, and this move by a state school, along with the growth of our private school’s overseas campuses, risks the education of our own citizens by sucking teachers overseas, and away from schools that badly need them, not only in some of our most deprived communities.

The DfE must make clear both why it approved this venture, and what happens if lots more state schools want to go down this road as a means of earning income to support the homebase.

As regular readers know, I am a strong support of democratic accountability for our schooling, and the academy system doesn’t provide that support to our system. Rather it provides fragmentation and encourages this sort of move all the while costing the system millions of pounds in unnecessary CEO’s salaries and other overheads.

This move reminds me of the Attlee government struggling with the aftermath of the Second World War and restricting sales of cars and other items in the home market to boost exports. Here we have a Labour government opening the doors to sending UK teachers to educate children of parent s that can afford their fees, and to directly set up in competition with private schools.

I might have understood a Conservative government sanctioning this move, but not a Labour government.

Please tell me I have missed some important value here.

.

Fine words butter no parsnips

What is one to make of a government that announces an expansion of the place of the creative arts in the National Curriculum review literally weeks after cutting the bursary for trainee teachers of music? Labour’s determination to recruit new teachers doesn’t include music | John Howson 8th October 2026

If I am being kind, it would be that one part of the DfE doesn’t know what the other is doing. Recruiting trainee music teachers has been a challenge over the past few years, and with universities eyeing the future of music degree courses, recruitment probably won’t get any easier.

Did a Minister, when sanctioning the bursary withdrawal, ask what the forthcoming Curriculum Review might have to say about the subject?  If so, why was the bursary withdrawn if the creative arts re to play a larger part in the new curriculum?

Hopefully, someone at Westminster will ask this question over the next few days. Perhaps media arts programmes might also like to interrogate a Minister about this curious state of affairs.

Of course, it is possible that the talk of expanding provision is just that, and the government has no real intention of putting funds behind any expansion in order to make it happen. Blame can then be laid at the door of schools for not switching resources into the creative subjects.

After all the government just said that

A new core enrichment entitlement for every pupil – covering civic engagement; arts and culture; nature, outdoor and adventure; sport and physical activities; and developing wider life skills.’ New curriculum to give young people the skills for life and work – GOV.UK

Not much meat on the bone there. Delving into the detailed response from the government we find that

We recognise the Review’s concerns around access to music and that some schools require support to deliver music well, including from specialist teachers, particularly to help pupils to develop their knowledge and skills in learning to read music and play instruments. We continue to invest in instrument stocks through the music hubs. Our £25 million investment will provide over 130,000 additional instruments, equipment and other music technology by the end of 2026, with around 40,000 already in the hands of teachers and pupils. We will consider how we maximise the impact of this investment to ensure the opportunity of and access to a reformed music curriculum is fully realised.”  Government response to the Curriculum and Assessment Review page 34.

Not much joined up thinking there. Encouraging singing has a much lower capital cost than instruments, and can capture more pupils – see the great scheme at Debry Cathedral that has over 900 possible singers.

The first sentence of the paragraph bears no relation to the rest of the paragraph, so don’t hold out hopes that music will achieve more than lots of instruments sitting on shelves or being played by children whose parents can afford the lessons.  

I am very disappointed in the music section of the government’s response, especially that now I chair the Oxfordshire Music Board and so music is a particular interest of mine.

Overseas teachers in England. More or less?

How far have teachers from outside the United Kingdom helped keep schools in England staffed during the period when there were teacher shortages? Although it takes a great deal of research to know what and where these teachers are working in England, the DfE in its evidence to the STRB (Teachers Pay Body) did provide some interesting data about changes in numbers of these teachers by their country of origin, between the 2015/16 and 2023/24 November teacher census returns. https://www.gov.uk/government/publications/evidence-to-the-strb-2026-pay-award-for-teachers-and-leaders data annex

For the purpose of this blog, countries have been divided into three groups: EEA – effectively all of Europe; countries with 20th century links to the United Kingdon, either as current Commonwealth countries or for other historical reasons, and countries that do not fit into either of the two other groups.

Taking the EEA countries first. It might be expected that post-BREXIT the numbers their had reduced. This is true for some countries, including  France and Germany, and, more interestingly, for the Irish Republic, where there was a loss of more than 900 teachers between 2015/16 census and the 2023/34 census.

EEA2015/162023/24difference
France22102085-125
Germany645605-40
Ireland35202595-925
Netherlands2252250
Sweden9590-5
-1095

Elsewhere in the EEA list of countries, there were more teachers in 2023/24 than in 2015/16

EEA2015/162023/24difference
Austria60600
Belgium951005
Bulgaria100205105
Czech Republic7510025
Denmark65650
Finland60600
Greece260590330
Hungary17527095
Italy485850365
Malta30300
Norway253510
Other EEA153015
Poland11551540385
Portugal255430175
Republic of Croatia406020
Republic of Latvia458035
Republic of Lithuania11016050
Romania350740390
Slovak Republic15018030
Slovenia, Republic406020
Spain12552100845
Switzerland50555
2905

There were nearly 2,000 more EEA teachers in England in 2023/24 according to these numbers. Greece, Italy and Poland between them accounting for nearly half the increase in EEA teacher numbers, and Spain alone, a further 40% of the total.

For countries with historic links to the United Kingdom there has been a marked decline in teachers from Australia, New Zealand and Canada recorded in the DfE census, and increase in teachers from Jamaica, some countries in Africa, and from the Indian sub-continent.

LINKS TO UK2015/162023/24difference
Australia16851290-395
Canada15801330-250
Guyana6045-15
New Zealand745480-265
Sierra Leone8575-10
Trinidad & Tobago10595-10
-945

Jamaica, India and Pakistan and South Africa together account for the bulk of the increase in teachers from this group of countries.

LINKS TO UK2015/162023/24difference
Bangladesh10011515
Cyprus559540
Ghana515665150
India8651615750
Jamaica7451550805
Kenya14516015
Malaysia7510025
Mauritius11513520
Nigeria580860280
Pakistan280560280
South Africa15751815240
Sri Lanka11016555
Uganda709020
Zimbabwe37545075
2770

Teacher numbers from other countries not in the above two groups tend to be small in number.

Israel was the only country with fewer teachers, down from 60 to 55; a loss of just five teachers.

ROW2015/162023/24difference
Algeria559035
Brazil6012565
Cameroon709020
China145315170
Colombia559540
Iran13016030
Morocco558530
Other ROW9551540585
Russia8012040
Turkey10017070
Ukraine359560
USA845985140
1285

China and the USA were the only two countries providing more than 100 teachers during the period between 2015/16 and 2023/24.

As Michael Gove provided QTS to teachers trained in the USA over a decade ago, the number of teachers from the USA seems surprisingly small. However, it may not include those teaching in international schools in England that are part of the private sector.  

While it is clear that a substantially more ‘overseas’ teachers were recorded in the 2023/24 census than in the 2015/26 census, their numbers alone would not have been enough to have solved the teacher supply crisis. Might they have made a difference to the percentage of teachers from some ethnic groups?