Comparing top mathematics departments in universities in the UK 2008


This is now outof date. See 2009 version


Ranking, league tables and other data for university maths departments and courses in England, Scotland, Wales and Northern Ireland

At Shanghai Jiao Tong University they started their "league table" the Academic Ranking of World Universities [WP] to measure the gap between Chinese Universities and "world class" (mostly US) universities I was interested to get some idea of the difference between our School of Mathematics and other top mathematics departments in the United Kingdom. Since our formation from the merger of UMIST and the Victoria University on Manchester (VUM) in 2004 the School of Mathematics at Manchester has been on a generally upward trend, and while crude measures should be treated with extreme caution it helps us to get some idea of how we compare (for example on size and income). I was also interested in the reasons some of the newspaper league tables for university maths departments produce some results that are quite surprising relative to how we perceive research reputation of departments.

I was interested to note a survey by the Sunday Times of the opinions of headteachers about the top five universities for several subjects. The headteachers had this order for mathematics: Cambridge, Oxford, Imperial, Warwick, Bath. Lets see how this compares to some rankings using published data.

One starting point is the 2001 Research Assessment Exercise. Of course we did not exist in 2001 and at the time of writing the process of assessing the 2008 RAE has just begun. For the sake of comparison we can combine the UMIST and VUM entries, although I expect we have changed (upwards!) much more than most UK maths departments since 2001 so in most respects this gives a lower bound.

RAE 2001

The RAE 2001 grades were 5*, 5, 4, 3a, 3b, 2 and 1 in decreasing order. In terms of the RAE 2001 grades there are five Universities that gained at least two 5* and on 5 grade from the three Units of Assessment Pure Maths, Applied Maths and Statistics+OR. Cambridge gained top marks with three 5*s, while Oxford, Warwick, Imperial and Bristol all got two 5*s and a 5. For details see a table of all mathematics departments submitting to the 2001 RAE . One could perhaps make a distinction that some departments chose to leave people out of the RAE declaring them as not research active, this is indicated by a B or C (but it is not clear how to interpret this). I have sorted each university by a weighted average by the number of research active staff in each UoA (again of debatable merit!).

Size of departments

Does size matter? Certainly there are some benefits to a large department, for example the coverage of more specialisms which might be expected to give students a wider choice of optional courses. Economies of scale might be expected as well. It is hard to count the number of academic staff accurately from for example the department web sites. it is n't clear who might be emeritus, or on a fractional contract for example, or shared with another department. The 2001 RAE return gives the Full Time Equivalent (FTE) of research active staff. Some Universities split Pure and Applied Mathematics and Statistics in to two or three different departments. For the sake of comparison I have taken the union over the three units of assessment.

Number of research active staff in UK mathematics and statistics departments and weighted RAE score
University Total Research Active Staff (FTE) Weighted average grade by FTE in each unit. 5*=6, 3a=3.5,3b=3 Average grade over three UOA (taken as zero if no return)
University of Cambridge 114.56 6
University of Oxford 81.6 5.59 5.67
The University of Manchester (UMIST+VUM) 75.5 4.55 4.5
University of Leeds 58.25 5 5
Imperial College of Science, Technology and Medicine 54 5.84 5.67
University of Nottingham 50.9 5 5
University of Edinburgh 48 5.15 5
University of Bristol 44 5.75 5.67
University of Warwick 43.75 5.42 5.67
University of Durham 42 5.22 5
One difference between Manchester and most of the other universities with large mathematical communities is that we are all one big School, rather than separate Pure, Applied and Statistics departments. We seem to be the biggest "Maths and Stats" department all under one roof, as of 2008 we have about 80 academic staff.

The numbers of undergraduate maths students is harder to find in published data. However I have found the following:

As far as we can tell from this data Manchester has the biggest population of undergraduate mathematical sciences students (judged by the 2007 intake) by a very small margin over Warwick. Going by the 2001RAE figures for staff it seems unlikely that any other departments would make it in to the top six by size, but Leeds (about 500 total - see brochure), Nottingham (total approx 500 ) and Edinburgh have fairly large undergraduate populations.

Note added September 2008. I asked one of our librarians and he found the data (published by HESA) on the number of combined undergraduate and postgraduate student for the academic year 2006/2007. This is the latest data they publish. The Open University comes first, but it is not clear how to compare as their undergraduates are part time. The next few are Warwick Oxford Cambridge and Manchester, which were quite close, then Bath and Imperial. Complete table is here

Value of active research grants

I wrote a very crude bash [WP] script for our own purposes that extracts data on the currently held research grants by mathematics (and statistics) departments. It runs every month on a cron job so should update automatically (and will break if EPSRC change their web site). The latest version of the table here. Scroll down to see the results with the various departments in each university combined. As the script is to some extent written by hand I had to type in a couple of lines for each university. If one department not listed suddenly gets loads of grants it will not show up unless I add it! You can check the results individually on the EPSRC Grants on the Web site. Go to institution and the select maths department.

On January 7th 2008 the table is as follows, MORE RECENT VERSION

EPSRC grants held by mathematical sciences departments in the UK 2008
NUniversity Total pounds
1 Bristol £19,366,708
2 Warwick £9,604,880
3 Oxford £9,515,222
4 Cambridge £8,573,353
5 Imperial £6,604,829
6 Manchester £6,568,348
7 Sheffield £4,612,261
8 Edinburgh £4,494,493
9 Nottingham £4,411,390
10 Bath £3,119,602
11 Glasgow £2,993,405
12 UCL £2,987,122
13 Southampton £2,502,640
14 Liverpool £2,127,533
15 Queen_Mary £1,626,088
16 Cardiff £1,467,353
17 York £1,444,890
18 Royal_Holloway £1,396,629
19 KCL £1,360,514
20 Strathclyde £1,275,419
21 Leeds £1,251,566
22 Birmingham £934,290
23 Aberystwyth £886,386
24 Durham £845,066

These results should be taken with a very large hand full of salt. Of course they indicate some merit as at least getting a grant means that referees and panel members approved the proposal. Here are some caveats

Doctoral Training Account

The EPSRC Doctoral Training Account is the main mechanisim by which home (as oppose to overseas) PhD students in mathematics get funded. In mathematics this allocation is decided by a pannel of mathematicians, whereas in other subjects a formula based on grants held is used. This reflects the fact that plenty of good mathematics is done without grants. The data for the 2007 allocation is on the EPSRC website. Some funds go to non Mathematical Sciences departments that hold mathematical sciences grants. Obviously this is interesting to potential PhD students as it shows the departments that have funding, and an idea of the number of students. But of course it does not reflect the numbers of overseas students who have funding from their own country. Some departments will fund postraduate bursaries from other sources as well. The table should also be of interest to academics thinking of moving to a UK mathematics department, and of course prospective undergraduates looking for a department with a lively mathematical culture.

Organisation

% of Total Maths EPSRC Doctoral Training Account

University of Oxford

10.994

University of Warwick

8.191

University of Cambridge

7.780

Imperial College London

7.548

University of Bristol

6.066

University of Nottingham

4.378

The University of Manchester

3.916

University of Bath

3.598

University of Edinburgh

3.478

University of Durham

3.071

University of Southampton

2.992

University of Leeds

2.803

Lancaster University

2.568

University of Glasgow

2.250

Queen Mary, University of London

2.101

Kings College London

1.828

Heriot-Watt University

1.808

University of Sheffield

1.703

University of Strathclyde

1.647

University of York

1.560

Newcastle University

1.477

University College London

1.467

University of Birmingham

1.433

University of Liverpool

1.420

University of East Anglia

1.401

University of St Andrews

1.357

University of Leicester

1.197

Loughborough University

1.116

Cardiff University

1.051

University of Aberdeen

1.001

University of Surrey

0.951

Royal Holloway, Univ of London

0.892

University of Exeter

0.875

University of Kent

0.797

LSE

0.775

University of Sussex

0.690

University of Salford

0.627

Brunel University

0.624

University of Wales Swansea

0.268

Aston University

0.065

Kingston University

0.055

Durham University

0.050

Coventry University

0.050

University of Wales, Aberystwyth

0.035

University of Reading

0.029

University of Essex

0.011

Goldsmiths College

0.005

University of Bradford

0.002

Royal Society, Wolfson Merit, Leverhulme

The Royal Society administers a number of prestigious research fellowships. These are the Royal Society Research Professorship, Wolfson Research Merit Award, Leverhulme Senior Research Fellowship, as well as the Royal Society University Research Fellowship, Industrial Fellowship and Dorothy Hodgkin Fellowship. The RS web site lists all current fellows on a single web site. On 6th January 2008 this was the whole list. I extracted this to a text file and then used grep to count those held in Mathematics and Statistics Departments. The results as follows:

RS grants and Fellowships
University Royal Soc.
Research
Professorship
Wolfson
Research
Merit
Award
Leverhulme
Trust Senior
Research
Fellowship
University
Research
Fellowship
Dorothy
Hodgkin
Fellowship
Industry
Fellowship
Cambridge 0 2 0 6 0 0
Oxford 0 2 0 4 0 0
Bristol 0 3 0 0 0 0
Warwick 0 0 2 0 0 0
Manchester 0 2 0 1 0 0
Imperial 1 3 0 4 0 0
Sheffield 0 0 0 1 0 0
UCL 0 0 0 1 0 0
Queens Belfast 0 1 0 1 0 0
St Andrews 0 0 0 1 1 0
Bath 0 0 0 0 0 0
Nottingham 0 1 0 0 0 0
Glasgow 0 0 0 1 0 0

Mathematical "stars": FRSs and Highly Cited

How many Fellows of the Royal Society currently work in each maths departments? Its pretty hard to get this data, but probably having a few is some kind of sign of a healthy research active department! Here is the data I have (not counting emeritus and retired). There is no systematic list so I may have missed some people. Also at Cambridge DAMTP arguably includes some people, Hawking for example, better known as physicists than mathematicians. However even with a tight definition of mathematician Cambridge still comes top if the FRS table.

Fellows of the Royal Society in UK maths departments (incomplete)
UniversityFRSsNotes
Cambridge Barlow, Turok, Coates, Gowers, Kelly,
Hinch, Huppert, Shepherd-Barron, Spiegelhalter, Pedley,Wills +etc
Gowers is a Fields Medallist
Oxford Ball, J. Ockendon, Silverman, Segal, Birch, Kirwan, James, Lyons
Heath-Brown, Hitchin, Donnelly, Trefethen
Warwick MacKay, Reid, Stewart, Preiss
Imperial Donaldson, Hayman, Atkinson, Donaldson is a Fields Medallist.
Manchester Higham, Taylor, Wilkie Paris is a Fellow of the British Academy
Buchstabber and Shiryaev are Corr. Memb. Russian Acad. Sci.
Bristol Green, Wooley,
Liverpool Rees Mazya is Member of the Royal Swedish Academy of Sciences

Fields Medalist[WP] are very rare so not really useful for league tables. Probably more important taking a long view is that a department has a culture that will produce one. For example Donaldson and Atiya got their Fields Medals while at Oxford. That said the research culture of a department is likely to be improved by having one there.

It is good for ones research papers to get cited and ISI Highly Cited researchers are ones that have got cited a lot, an indication that their work is valued by other researchers. Some topics in mathematics, for example numerical analysis, seem to generate more citations than others. But maybe that means these areas are more useful anyway?

ISI Highly Cited researchers in UK Maths Depts
UniversityHighly cited researchers
Oxford Ball, Cox, Quillen, Silverman,Trefethen,
Cambridge Clayton, Spiegelhalter, Fokas, Lickorish,
Imperial Donaldson, Liebeck
Bristol Goldstein, Green
Manchester Higham, Hammarling (hon post), Dongarra (fractional post)
Warwick Roberts, Stewart, MacKay
LSE Atkinson
King's College London Davies
UCL Dawid
Open University Jones
Queen Mary Smith
Glasgow Titterington

Entry grades

The Times Good University Guide has listings by subject. (What follows was accessed Jan 2008) Here is their table for mathematics. They have used RAE 2001 scores, entry qualification and something called "graduate prospects". You can reorder them to rank by one of the criteria. For example according to them the top 20 by ranked by entry qualifications (UCAS points). There is some explanation of the methodology is given indicating that it is based on some very old UCAS figures! No wonder this and other newspaper tables give very strange results for mathematics. Some also use the results of the teaching quality assessment the government carried some times ten years ago.

More up to date information on entry requirements can be found on the UCAS web site, e.g. by a search for course code G100 (Maths). Here are some examples as of Jan 2008 (some are slightly more complicated than stated, follow links for more details). Most of the top universities do not use UCAS points tariff as their entry criterion, in England and Wales the grades on three A-levels is the usual standard. The top maths departments find that this is not sufficient to select on mathematical ability and use things such as STEP and AEA results, module grades, for example on the core mathematics modules C3 and C4, and Further maths A or AS grades. UCAS tariffs for students who are accepted may include things like A-level general studies that are often excluded from entry requirements, as well as stand-alone AS scores.

A-level entry grades for G100 Maths degree courses in the UK
Cambridge AAA + STEP
Oxford AAA (varies by college)
Warwick A(M),A(FM),B,2-STEP
A(M),A(FM),A,A
Imperial A(M),A(FM),A
A(M),A,A, STEP
A(M),A,A,B
Leeds A(M),A,A
A(M) B 3-STEP
A(M)A a(FM)
Bristol A(M)AB
Bath A(M),A,A or A(M),A,B
Manchester A(M with AB in C3-4)AB
Nottingham A(M)AB (A in FM if taken)
UCL A(M),A,B(FM)
Durham A(M) AB (FM at AS if not A2)

So the whole there is still basically "COWI" Cambridge Oxford Warwick Imperial asking for three As and some, and then the next bunch who will let you get a B as long as you do well enough in maths. In contrast to the Times GUG based on 2001 entry requirement the current entry requirements are broadly in line with the other rankings..

The hilarious Guardian University Guide

The Guardian compiles a league table noted mainly for its comic value. The biggest gaffs where the listing of maths departments at Bangor and Hull with quite high league table positions after they had closed down. The table continues to produce some quite unexplected ranklings. The 2008 edition is out already. The current top 10 are Cambridge, Oxford, St Andrews Portsmouth Warwick Imperial Aberdeen Leicester Lancaster Kingston. Well Portsmouth will be pleased! The methodology is given. The PDF seems to indicate that they are also using out of date information. The student data is taken from 2004/2005, and the leavers destination from 2004/2005. The entrance qualification are the 2005/2006 figures. It makes one wonder why they bother making a new league table each year! One other interesting feature of the Guardian table is the inclusion of spend per student. So any university that wants to go up the league table just needs to burn money, not spend it on anything useful. You would think that Manchester would do quite well on that having recently invested £ 43m on the new Alan Turing Building,[WP] which is at least half for maths, however I suspect again old figures are used and they may not attribute capital project to subject "cost centres". However such investments clearly have an impact on maths students' experience.

Good University Guide

The Good University Guide produces table by subject including Mathematics(choose maths on the menu). As other tables it is based (again I am looking in 2008) on old data including the 2001 RAE and the 2004-2005 HESA entry qualification data and destination data. The current version lists the top ten as Cambridge Oxford Bristol Bath Durham Warwick= Imperial LSE UCL and St Andrews. As I understand it a university that did not submit RAE returns in all three of Pure Applied and Stats was not penalised (eg LSE only submitted in Stat which seems a bit strange.

The GUG is compiled by a consortium called Mayfield University Consultants, largely current or former university administrators, and the University of Sheffield. It seems this group used to be responsible for the Times Good University Guide, but have split from them. The Times GUG are now produced by Exeter Enterprises, owned by Exeter University.

UniStat

The on line university statistics site UniStats would seem to be a way to check league table data. However it seems to use the same out of date HESA data. For Manchester it says our average UCAS score is 420 and only 42% of our graduates get a job after six months. Both of which sound totally fictitious to me. The details for the University of Manchester even list the web site as the Victoria University of Manchester's now defunct URL. Looks to me as if this site is pretty much pointless unless you are interested in historical data.

Google Page rank

The Page rank [WP] is a number in the range 0 to 10 used by Google to sort search results and is uses an eigenvector of a graph of web links to decide which pages are most important. So presumably important maths department's web pages have lots of important links to them. The automated checker failed so I replaced these by a snapshot on 4/6/2008.
URL Page rank
Cambridge www.maths.cam.ac.uk 7
Oxford www.maths.ox.ac.uk 7
Warwick www.maths.warwick.ac.uk 6
Imperial www.ma.ic.ac.uk N/A
Bristol www.maths.bris.ac.uk 6
Manchester www.maths.manchester.ac.uk 6
Durham maths.dur.ac.uk 5
Leeds www.maths.leeds.ac.uk 6
Heriot Watt www.ma.hw.ac.uk 6
St Andrews www-maths.mcs.st-andrews.ac.uk 6
Southampton www.maths.soton.ac.uk 6
Birmingham www.mat.bham.ac.uk 6
UCL www.ucl.ac.uk/Mathematics 5
Aberdeen www.maths.abdn.ac.uk 5
Edinburgh www.maths.ed.ac.uk 6

At the time of writing that does not look very interesting. Most of the reasonably sized maths departments at the time of writing (Jan 2006) have a Page rank of about 6, so it is currently not a very discriminating measure. A list of all maths departments in UK Universities is maintained at Heriot-Watt, if you want to check others. There are some quite respectable departments with Page rank zero. Perhaps because they changed their URL recently. In time there will no doubt be more useful tools to asses the impact of departments on line. Already Webometrics ranks Universities by the impact of their on line publications.

Added Sept 2009. The Google Directory has a list of uk maths departments sorted by page rank (but the page rank still does not vary very much between th top 30).

More citations nonsense

There are not many world league tables of mathematics departments. One based on citation was published in 2002 by ScienceWatch. It lists the top 25 in the world by the number of citations and impact.

On number of citations no UK departments even made it in to the table! On the "impact" UK departments in the table were 9th Imperial, 11th Cambridge, 15th Oxford, 18th Warwick. Agreeing with the traditional rating as COWI being the top four (but in a different order). Of course many pinches of salt here needed as usual when interpreting citation data. That said it is certainly a good thing to write papers that other researchers cite.

I have recently noticed a paper by Bollen et al that suggests the use of something more like the Page Rank could be a better indicator of quality rather than popularity in networks of journal citations. No doubt such a thing will be implemented and if it is used, eg to allocate funding, people will start to optimise against that criterion. In mathematics citation based measures of the importance of work suffer from the effect that the importance of some mathematical result only shows a considerable time later. For example Fritz John's 1938 paper on the ultrahyperbolic equation had to wait half a century before it found use in 3D x-ray tomography, which had not been invented at the time he did the work.

MathSciNet

MathSciNet (MSN) lists and reviews publications in mathematics. It takes a fairly narrow view of mathematics, certainly including everything in pure mathematics but missing some applied papers. I wrote a script that extracts publications for the year 2007. I expect not all 2007 papers have made it in to MSN this early in the year (12th Jan 2008)**. The 2006 totals are greater. I only checked the institutions listed and only the codes for maths and statistics departments (see MSN Institutional codes) Some people might list their address as some interdisciplinary centre in which case their paper might be missed. Certainly the numbers of papers published in the year is going to be much bigger. For example any of my own papers are not in engineering journals and therefore do not show up. As it counts the number of papers on the MSN database with at least one author from each institution two authors from one institution counts as one paper. However an author from each of pure and applied departments in one university counts as two for that university (I should have used OR), similarly a paper authored by an author from Oxford and Warwick would count as one for each. This measure is one that is easy to compare with mathematics departments in other countries for example some of the top US departments Berkeley 210, MIT 200, Princeton 184, Chicago 144, Cal Tech 99, Stanford 98, Harvard 91. In Russia Stekelov Institute of Mathematics St Petersberg 192, Steklov Institute of Mathematics Moscow 124.
UniversityNo. of Pubs in
MSN in 2007
Manchester 163
Cambridge DPMMS+DAMPT 151
Oxford MI+OCIAM+Stats 151
Imperial 140
Warwick M+MI+S 124
Queen Mary 114
Leeds (several)98
Bath 73
Edinburgh 73
KCL 70
Bristol M+EngM 68
Glasgow 62
St Andrews (several)* 61
Birmingham51
Durham 51
Loughborough 49
LSE M+S+OR 48
Southampton 48
UCL 48
Nottingham 43
Herot-Watt 25
Sheffield 24

* St Andrews has seven institution codes 4-STAN-? for maths and stats departments so results may be inaccurate.

**Updated 23th Sept 2008, including all the departments with at least 30 staff in the 2001 RAE + Birmingham. Numbers gradually increase for 2007 as the indexing catches up.

I have not yet got around to working out papers per individual, but you can do that approximately from the 2001 RAE figures.

Overall rankings of universities

"The Times Higher Education" provides an annual league table of universities but does not go down to individual subjects However some of these measures are only likely ever to be applied to whole universities or large subsets of them and are never the less of interest, for example, to a student choosing where to study maths.

The 2007 Lists of Top 10s includes the top 10 on a survey of employers. Out of the UK universities they list Cambridge 1st, Oxford 2nd, LSE 3rd, Manchester 5th and Imperial 8th. It is interesting that the UK does so well so there may be some bias. On the whole large companies concentrate their recruitment effort on a few (usually large) universities and this might be a useful figure for prospective students to choose a university within the UK. The scores are given for each university in the Top 200 Universities in the World list.

US universities slot in the gaps with Harvard 4th, MIT 5th and Stanford 6th. By contrast the table of top 10 citations per staff member contains no UK universities, 8 US and one from France and one from Switzerland. However this is not likely to be especially relevant score for mathematics.

Conclusions

Before I started this exercise I shared the view of most of the UK mathematical community that "COWI" (but the order debatable) are the top four mathematics departments in the UK, but that the ranking of the next few is very much a matter of opinion. At the time of writing one would certainly have to rank Bristol up with COWI on the basis of research grants and RAE and Manchester close at least on grant income, research activity and sheer size. That said there is excellent research work and teaching going on in smaller departments, most notably Bath, many others excellent in their specialisms. Also before looking at some of the details some of the newspaper league tables and the GUG puzzled me as to their mathematics rankings, especially how Manchester seemed to rank surprisingly low and some relatively unknown maths departments rise up their rankings. This seems to be down mainly to the use of some very old figures from RAE2001 and HESA, and in the case of Manchester probably some confusion in the statistics caused by merger. Also tables such as the Guardian ignore research quality. However for students with a passion for mathematics a research active departments would be essential for a satisfactory undergraduate experience. I expect I will revisit some of these issues in 2009, in the light of RAE2008 and other new data.

There is an increasing tendency in Higher Education to employ "metrics",that is crude numerical values as a substitute for an analysis of quality. I think it is useful for academics and heads of maths departments especially to look at some of the above figures to try to understand if they have created an atmosphere in which excellent mathematics can be done, if they are applying for and being awarded a reasonable number of grants, if the work is getting published and if the academics get to be recognised as leaders in their field. From the point of view of students (undergraduate and postgraduate) and individual academic staff we would like to know if a department has a culture and "climate" in which we can achieve our mathematical potential, and to enjoy ourselves doing so. These are of course very difficult to quantify!

Please feel free to post comments, corrections etc, on this blog entry.

I also wrote some thoughts on advising prospective undergraduates on where to study mathematics in the UK

Bill Lionheart School of Mathematics University of Manchester. January 2008.