"The colleges of Britain's prestigious Oxford University will lose their
800-year-old right to choose their undergraduates amid government pressure to
admit more students from state schools and lower social classes....
The university admitted that as a result, colleges will lose autonomy and
individuality, according to the newspaper."
Read the whole thing here. Bye bye mortar board and gown. Hello sports jersey and backward baseball cap.
13 comments:
WB,
I am sure you will be interested to know that in point of fact, the move to "regulate" Oxford's college’s admissions is not solely designed to ensure that applicants of lower social status get their fare shake. Because of the way the collegiate system is set up in Oxford, as opposed to other collegiate universities like Yale, it would be possible for the best applicant in a subject (regardless of caste) to be over looked completely.
Here is how it works: prospective students can only apply to one college (at one of Oxford or Cambridge). As it happens, each college has only a certain number of spots allocated each year to the various subjects, math, classics, theology, etc. They also have the 'right,' however, to arbitrarily close those spots on any given occasion and designate them to another discipline. So, suppose you are the best candidate applying to the University to read theology. You apply to Corpus Christi, but unbeknownst to you, they are not accepting students in theology this year. What now? Your name goes into the pot, and perhaps there is another college that, after interviewing all of their own candidates, could give you a shot. More likely, you are forgotten, your hopes dashed, your future destroyed, and you settle for the ivy leagues. Unthinkable!
How to ensure spots are offered to those with equal ability but unequal privilege follows from the above. It is entirely plausible that someone has greater ability than many others applying, but comes from Cowley Technical and Vocational School, and not Eaton. How do you measure these things are fairly judged? How do you ensure 'the best' get in? Should the main criteria be who you rowed for, or whose debating squad you were on? Because Oxford is publicly funded, and not private (like the Ivy's), it is, rightly, responsible to those who keep it afloat - tax payers. Thus, it is open to public scrutiny in a way that Harvard and Yale cannot be.
You and I both know that the long held notion that rich=clever and poor=slow is in rapid decline. Is it possible that the meek too might inherit the earth? Well, in the UK they are certainly working very hard to ensure that they have a good shot at it. More power to'em I say.
BFC
P.S. - nice photo from the ChCh meadow! Did you take it yourself; perhaps whilst picnicking on strawberries and champagne with Garland as he water painted?
darned low class people...don't deserve an education, even if they are smarter and more talented...tut tut
[please note sarcasm]
BFC,
I understand and appreciate your critique of the current mode of admissions to Oxford, but your tacit assumption that the proposed changes will "fix" these problems seems to me to be incorrect.
The proposal basically involves removing the responsibility for admissions from the faculty and giving it to a group of admissions officers. This is how American Universities are run, and I think there are many reasons to question the success of this approach. The central question is, Who are these admissions officers? What sort of qualifications do they have? Upon what criteria do they base their admissions decisions? And most importantly, what makes them better able to select students than professors and tutors, the people who are qualified in their fields and who will be resposible for teaching these students once they are admitted?
This last question seems to me to be the key to this debate. Currently at Oxford, admissions decisions are made by the people who ought to know best the qualities that make a successful applicant, and thus they are made by the group most likely to give applicants a fair hearing based on their individual merits. Would we suggest that admission to PhD programs in America be removed from the control of the individual departments and put in the hands of that nebulous group known as admissions officers? I would say certainly not, and Oxford's commitment to this type of admissions process on every level is thus all the more admirable.
In this way I feel that the answer to your question is simple:
"How to ensure spots are offered to those with equal ability but unequal privilege follows from the above. It is entirely plausible that someone has greater ability than many others applying, but comes from Cowley Technical and Vocational School, and not Eaton. How do you measure these things are fairly judged? How do you ensure 'the best' get in?"
You put the people most aware of the skills needed by a successful applicant in charge of the process.
However, what this system won't give you is affirmative action. Large scale demographic factors do not affect selection in the Oxford system as they do in America because the selections are made on a more individual level - 3 or 4 students are chosen in Theology, and the school wants the best 3 or 4 applicants they can find and are not as concerned about making choices that maintain a particular demographic balance in the university at large. Here, admissions officers are very interested in such a balance; they decide that they should accept 15% Asian students, 10% African American students, etc., and these numbers are in place before the first applicant is considered - factors that certainly have nothing to do with academic promise.
It seems to me that if a school wants to admit the best applicants available, they should use the Oxford model in which each department of each college selects the handful of students whom they feel are the most qualified applicants; and if a school prefers a demographically balanced student body with preference given to historically underrepresented groups - that type of large scale selection is the specialty of admissions officers. I think that the push for changes in the Oxford model is most likely due to this second approach.
ACR
ACR,
Thanks for your response. I will be brief:
1) I harbor no tacit assumption that the proposed changes will fix anything. To read what I have said in that way would be wrong.
2) I was suggesting that because tax payers support and pay for the University, they have a right to observe and critique the Oxford system in a way that is not possible in the U.S. But, I think you are right that there is certain strength to the Fellows making the decisions.
Still, I think the 'Good Will Hunting' question is worth asking. Perhaps I suffer from a hopelessly romantic notion that one's worth is not determined by the grammar school one went to, or the contents of one's bank accounts. In my observation, you might not always be able to tell at dear ol'Oxon.
BFC
Thanks for the response BFC. I also agree with you that Oxford Dons have their own biases when it comes to selecting students, but I just have to return to the question, Who else? The 24 year-old admissions officers in American schools seem to me a poor substitute. I personally harbor the romantic hope that my non-Ivy (and non-Oxon) undergraduate degree will scale the heights and get me into a quality PhD program. Sigh.
Best,
ACR
ACR,
Yes, PhD programs are very hard to get into. My non-ivy, non-oxon, undergrad degree worked just fine, however. Take heart. All will be well!
Merrily,
BFC
Its not that plebes cannot be qualified; but rather just that being a patrician ought to count for SOMETHING. That is, Oxford ought not to give into the notion that education is or ought to be totally comodified and professionalized. College life is better if it is, to some degree, about grace and graciousness.
Ah WB,
I wonder what being aristocracy ought to count for? From my vantage point, it counts for quite a bit, even before University aps are in.
You know my story Will. It has certainly been about grace for this plebe. My hope is that there can be grace for others like me. But grace is not earned, is it? So, if the story is about grace, the point goes to me.
Yet, for patrician like you, I also pray grace upon grace.
Sadly, the race is not to the swift.
Ever,
- BFC
BFC -
Don't let my lisp and pinky ring fool you. Or at least, if I am a patrician, I wish it came with more favors and money. Or some.
I just think that society benefits from cultivating people like Sebastian Flyte. Only, probably less alcohol-soaked.
WB,
Yes, yes. Too true: a good novel is hard to come by these days, what. Admit the Flyte's of the world. We could all benefit from another romantic anecdote. I buy all my clothes in London...
But then it occurs to me, perhaps institutions like Oxford (and the like) should be thought of as more than finishing schools for the "magically beautiful." Perhaps, the world of Sebastian Flyte is now irretrievably gone. Perhaps there is more to life than membership at the Travelers Club, and afternoon tea in the Randolph Hotel. Yes, I am quite certain I heard that somewhere.
I am reminded of the remarkably sophisticated gents who have suffered through non-Oxbridge educations - almost all the Royals for example (save the rotten one). Perhaps they would object to the notion that Oxbridge equals cultivated.
Waugh created a mythical Oxford. It is not the foggy city of dreams where boys can lapse into hopeless spells of drink and depression. It is a place of work, and applicants should be evaluated on their ability to do that work. That is my point.
Toodle pip!
BFC
(p.s. - I too have a pinky ring. No lisp though. Perhaps, therefore, I will never be a patrician. p.p.s. - I don't really think you are a patrician, though the Barbour jacket might suggest differently.)
To return to the main point, I agree that Oxford should be more than a finishing school for the magically beautiful. But its being more than that is compatible with its being THAT. And my complaint is that when they turn over admissions decisions to some subcommittee of the NHS, it will suddenly be at best a meritocracy. At worst, it will become like that dreadful public high school where I went for one year: people staggering around with vacant, purposeless expressions, except whilst purposefully pursuing some crime or other.
I think Oxford had found more or less the right balance. As it stands currently (though not for much longer), there are plenty of people from a working class (and every other) background at Oxford. But there remains a place for legacies and the Lord of Fyssleswaite's second son. And I think that's as it should be. Its what makes these places what they are. Yale was made into Yale by the landed but unintellectual rich folks who left it, as well as, to a lesser extent, the brainy but diverse population of egalitarian fantasies.
There certainly IS more to life than membership at the Travelers, and tea at the Randolph. There's membership at the Carlton and tea at the Lizzie!
Ahem.
Oxbridge is neither necessary nor sufficient for cultivation. All I am saying is that cultivation ought to be a part of its undertaking -- at least as far as undergraduates are concerned. And it always helps to have people around who have had a head start.
I found Oxford to be a foggy city of dreams, where I could very easily lapse into spells of drink and depression... and occasionally visit the library. But I wasn't working on a degree. That was an unfair advantage. But I think you will find that it can indeed be a foggy city of dreams and drink and depression.
I lied. Mea maxima culpa. I have neither a lisp nor a pinky ring. I do have that cope from the previous post, though. And, indeed, a barbour jacket. Its all torn up and it stinks. Surely that makes me the heir to some throne.....
WB %
Agreed - more or less.
Warmest Christmas Wishes,
BFC
To Rread the Working Party Report (from which the proposals are coming) continue below...
WORKING PARTY ON SELECTION AND ADMISSIONS
UNDERGRADUATE ADMISSIONS:
POLICY AND PROCEDURES
November 2005
CONTENTS
FOREWORD pp. i-ii
1. INTRODUCTION p. 1
1.1 The international and national contexts p. 2
1.2 Recent changes in the Oxford admissions system p. 5
2. ASSESSMENT p. 6
3. ORGANISATION p. 10
4. ADMISSIONS MODELS p. 13
4.1 Model A p. 14
4.2 Model B p. 16
4.3 Administrative and IT implications p. 17
4.4 Timetable p. 18
5. INTERNATIONAL APPLICANTS p. 18
6. OXFORD COLLEGES ADMISSIONS OFFICE p. 20
7. SUMMARY AND CONCLUSIONS p. 20
ANNEXE A: Glossary of terminology p. 22
ANNEXE B: Terms of reference and membership p. 23
ANNEXE C: Further discussion of Models A & B p. 24
ANNEXE D: Timetable for proposals p. 28
ANNEXE E: Matrix of subjects’ admissions procedures p. 30
i
Foreword by the Chairman of the Working Party
This report presents the findings of the Working Party established in late 2004 to review
Oxford’s undergraduate admissions policy and procedures.
A previous Working Party chaired by the former Principal of St Anne’s, Dame Ruth Deech,
reported in 2002. Most of its recommendations have been implemented. As a result of these
and other changes, there has been considerable evolution of assessment methods and
procedures over the past few years – with some subjects moving a good deal further than
others.
We have concluded that to a large degree the admissions system is working well. This is
made possible by the serious effort and substantial resources that are put into it: the time of
college tutors and admissions secretaries, department/faculty organising secretaries and the
staff of the Oxford Colleges Admissions Office (OCAO), and associated expenditures. Above
all, the admission system’s prized asset is the academic judgement of tutors that underpins it
at all stages.
But we have also concluded that there could be further improvements to the system. Some of
these involve relatively modest changes. Others are more far-reaching. The aim of the
changes we are recommending is to provide further assurance that – with more and more
good candidates relative to the available places – the colleges and subject departments and
faculties are doing all they reasonably can, together, to select the very best.
Our most important proposal is that the University should move to a single admissions model
for all subjects, in place of the wide variety of arrangements that currently exists. We have
developed two possible models for consideration by colleges, divisions, departments and
faculties. These are outlined in section 4 of the report, with their respective pros and cons.
Both models build on the experience of the past few years and on best practice as it currently
exists in several subjects.
A common element in both models is that all candidates for each subject would be considered
alongside each other at short-listing and final offer stage by the collectivity of tutors (or a
sub-set of them) in that subject. In the first model, the collectivity of tutors would have the
final say; in the second, colleges would be able to argue the case for an ‘override’ in
borderline cases. By ensuring that all candidates are considered alongside each other, this
would have the benefit of removing the influence that a candidate’s choice of college –
particularly in subjects where there is at present limited coordination – may still have on
his/her chances of obtaining a place. Eliminating the perception that college choice can make
a difference would also help to encourage more applications from good candidates at schools
and sixth form colleges where there is limited knowledge and experience of Oxford.
Annexe D sets out a possible timetable for decisions and implementation. Some of our
proposals can be decided and implemented more quickly than others. As regards the proposal
for a single admissions model, there would need to be at least a two year lead-time between
decision and full implementation, to allow for detailed planning and organisational and IT
system changes. For either of the two models to come into effect in time for the 2008
admissions round, a decision would be needed by June 2006.
There is one proposal – that there should be a review of the structure and position of the
OCAO (see section 6) – on which action is already in hand. The trustees of the OCAO have
ii
decided that this review should proceed forthwith. Dr Richard Repp, former Master of St
Cross, is chairing the review panel which will aim to deliver a preliminary report to the
trustees in December.
The report was commissioned by the Admissions Executive (ADEX), which is composed of
College and University representatives. It will be for ADEX to put formal recommendations
in due course to the Conference of Colleges and to the University’s Educational Policy and
Standards Committee (EPSC), which reports to Council. In drawing up its recommendations,
ADEX will want to take into account the views of the Committee of Tutors for Admissions
(which includes all college admissions tutors), the Conference of Colleges’ Academic Sub-
Committee, college governing bodies, divisions, departments and faculties.
Tim Lankester
President of Corpus Christi College and Chairman of the Working Party
November 2005
1
1. Introduction
1. In their approach to undergraduate admissions, the University and colleges are
committed to three high-level objectives:
(a) to attract applications from the most able individuals, irrespective of socioeconomic,
ethnic or national origin;
(b) to ensure that applicants are selected for admission on the basis that they are well
qualified and have the most potential to excel in their chosen course of study; and
(c) to ensure that the prospects of admission are not affected by college choice.
2. With regard to objective (a), the University and colleges have adopted various
approaches, including attractive and informative prospectuses and websites, bursaries,
open days and a range of access activities. In contacts with prospective students, an
important aim has been to try to ensure that our selection policies and procedures are
seen to be fair and objective and based on academic merit alone, taking account of
potential as well as achievement to date.
3. With regard to objective (b), Oxford uses a range of assessment methods. It relies on
academic staff to undertake the assessments and to make the admissions decisions.
With so many highly qualified candidates, the leading role of academics in the
admissions process is regarded as crucial if the best are to be selected.
4. With regard to objective (c), several measures have been developed so that
oversubscribed colleges are able to redistribute some of their better applicants to other
colleges at various stages in the admissions process. This helps very strong candidates
who are unable to obtain a place at their college of preference to be admitted to another
college.
5. The University and colleges need to do all they reasonably can to meet these objectives
if they are to achieve the highest possible standards in undergraduate learning. All three
objectives, and especially (a), are also implicit in the Access Agreement the University
has signed with the Office for Fair Access (OFFA).1 More generally, our admissions
policies and procedures need to command the highest confidence externally – in
schools and colleges, amongst students, teachers and parents, and in the media and in
government.
6. The intellectual quality of our undergraduate intake is extremely high. Results in Final
Honours Schools have been on a rising trend. In 2005, almost 90% of candidates
achieved a First or Upper Second. Several factors may have contributed to this;
improvements in Oxford’s admissions procedures have almost certainly helped.
7. However, improvements in the way candidates are assessed and in the way that
decisions are taken have not been uniform across all subjects, and better assessment
methods and better organisation in some subjects have highlighted weaknesses
elsewhere. Though those admitted in all subjects are normally extremely able, it is the
view of many – both inside and outside Oxford – that we still fall short in terms of
having systems in place that can ensure that the very best who apply to Oxford are
admitted, irrespective of college choice.
1 Oxford’s access agreement is available at http://www.admin.ox.ac.uk/po/news/2004-05/mar/accessagreement.pdf
2
8. There is a widespread perception in the media and in schools that ‘choosing the right
college’ can significantly improve a candidate’s chances of success. This is
notwithstanding the measures that have been introduced to redistribute2 candidates
amongst colleges. It is supported by anecdotal evidence from schools – when they
sometimes tell us that a candidate for a particular subject whom they regard as their
most able fails to get a place at one college, whilst a candidate whom they regard as
less able gets a place at another college. The fact that some colleges achieve much
better results in public exams than others probably in part reflects differences in the
quality of intake.
9. To the extent that college choice does still matter, this will tend to disadvantage those
candidates with less knowledge of Oxford. Though there is no systematic evidence that
the college admissions system actually deters candidates from applying, feedback from
pupils at schools with limited Oxford connections – most often in the non-selective
maintained sector - suggests that they find our admissions arrangements confusing and
opaque, particularly when making a choice of college.
10. The Working Party (whose terms of reference are at Annexe B) has examined these and
other issues across the various subjects. The Working Party has concluded that, despite
the tremendous amount of time and effort that is put into the admissions process by
academics and administrators alike, there is scope for further improvements, especially
in the way assessment and decision-making are organised. In putting forward its
proposals, the Working Party recognises that no system, however sophisticated or well
intentioned, will ever achieve perfection in ensuring that the best candidates are
selected; but we do think it is possible to move somewhat closer to that ideal position.
11. Under the University Statutes and Regulations, the Educational Policy and Standards
Committee (EPSC) is responsible for policy and standards in respect of access and
admissions. The Admissions Executive, which commissioned this report, makes policy
recommendations to the EPSC, in consultation with the Admissions Committee (the
Committee of Tutors for Admissions). The purpose of this report is to set out policy
recommendations for general discussion to enable the Admissions Executive to make
formal proposals that would command broad support in the University and in the
colleges.
1.1 The international and national contexts
12. To remain internationally competitive, Oxford needs to attract the ablest candidates not
just in this country but also from abroad. In regard to potential students from outside
the EU, a major issue concerns lack of bursaries, which is outside the scope of this
report. For all candidates from outside the UK, our admissions arrangements may be a
deterrent. The average success rate for applicants from other EU countries and
elsewhere has been far below that for UK candidates. This may be partly because
international applicants have tended to apply for the more popular courses, and partly
because some may be less aware of the high standard required for Oxford entry.
Nonetheless, the complexity of our admissions systems, tutors’ limited knowledge of
some overseas qualifications, and difficulties in arranging interviews, have almost
certainly contributed.
2 See Glossary of Terms at Annexe A.
3
13. From a national standpoint, our admissions policies and procedures need to be viewed
against the background of the Government’s 2003 Higher Education White Paper, the
Schwartz Report on admissions to higher education which followed in 2004,3 and the
Department for Education and Skills’ consultation document, Improving the Higher
Education Applications Process, published in September 2005.4
14. The White Paper focused principally on funding issues, but it also – inter alia – had a
chapter on “Fair Access”. In regard to Oxford and Cambridge, it commented as
follows: “In Oxford and Cambridge, the difficulties inherent in running a collegiate
admissions system in a sufficiently robust, rigorous and professional way to ensure that
it is fair have been recognized. We welcome reforms being made by Oxford and
Cambridge to coordinate and centralise admissions, as part of ongoing efforts to widen
access and we would support their rapid extension.”
15. The Schwartz Report set out key principles to which universities should adhere in their
admissions policies. It emphasised the need for fairness and transparency,
professionalism and appropriate institutional structures, and reliable assessment
methods based solely on merit; and it suggested that, in assessing an individual
candidate’s potential, account should be taken of his/her educational and socioeconomic
background. Whilst not questioning the autonomy of universities in regard to
their admissions policies and processes, a key thrust of the Report was that universities,
especially top ones such as Oxford, should achieve a more equitable distribution of
students in terms of socio-economic, ethnic and school background.
16. In the wake of Schwartz, and following the passage of the Higher Education Act 2004,
all universities that wished to charge higher tuition fees from 2006 were required to
conclude an Access Agreement with the Office for Fair Access (OFFA). Oxford’s
Access Agreement, submitted to OFFA in January 2005, formalises the Collegiate
University’s outreach and student support activities. In addition to bursary provision
and other forms of student support, the Agreement underscores the University’s
commitment to continue its extensive programme of outreach work, alongside close
monitoring of applications and admissions data. The University has committed itself to
developing an appropriate benchmark for maintained school applications, based on the
proportion of pupils from the maintained and independent sectors who achieve the
standard Oxford offer of three As at A level, adjusted for the appropriate combinations
of A level subjects for the degree courses we offer. The benchmark will relate to
applications, not to actual admissions. Progress towards the applications benchmark
might in part be helped if our admissions arrangements were simpler and more
transparent.
17. The consultation document, Improving the Higher Education Applications Process,
contains:
a) proposals, uncontentious as far as Oxford is concerned, for improving the
information available to potential and actual applicants;
3 The full report, Fair admissions to higher education: recommendations for good practice, can be found at
http://www.admissions-review.org.uk/downloads/finalreport.pdf.
4 The consultation, Improving the Higher Education Applications Process, can be found at
http://www.dfes.gov.uk/consultations/conDetails.cfm?consultationId=1346.
4
b) proposals to do with the information that universities should receive about
applicants. These include the piloting of a single national aptitude test which we
discuss in paragraph 35 below, and a proposal that schools and colleges should
no longer supply predicted A level grades on the grounds that this is unfair to
those who over-perform and to those who under-perform. The Working Party has
not formed a firm view on this latter issue, but it seems unlikely that, given that
predicted grades are an important source of information about applicants and that
they are treated with care by admissions tutors, we would want to give them up;
c) proposals for reforming the current applications process, for implementation by
2008/9. Many of these would be of limited relevance to Oxford, such as the
proposed reform of clearing, a process in which Oxford at present does not
participate. More relevant is the proposal that students who achieve higher grades
than required by their conditional offers should be able to make a new application
and have their original first firm conditional offer protected whilst they do so.
Whether such a proposal could be accommodated within the Oxford admissions
system will require careful consideration, in the first instance by the Admissions
Executive. It would be possible, for example, for Oxford to consider such
candidates in a single pool alongside those who had not met their offer
conditions;
d) proposals for a Post-Qualifications Applications (PQA) system, for
implementation in the longer term. The consultation document offers two
models. In the first, all candidates apply after they receive their A level results
(although they ‘register’ with their preferred universities earlier), and all offers
are made post-A level. In the second model, the majority of places would be
offered ahead of A levels but universities would hold back a small proportion of
places for candidates who might wish to change their existing applications or
might wish to apply for the first time after they had received their A level results.
18. The first PQA model would, in principle, have great advantages for Oxford. Having
candidates’ A level results before the offers were made would simplify our processes
and improve our ability to select the strongest candidates, and especially so if we had
access to scores within the A level A grade. It would also probably improve the
application rate from candidates from the maintained sector. It probably would not
obviate the need for interviews; but if interviews could be done after A level results
become available, the need for pre-interview testing for the purpose of short-listing
could be minimised.
19. The problem with this model is one of timing. Without the A level examinations being
brought forward and/or the start of the academic year being delayed, there would not be
enough time for the admissions process to be completed.
20. The second model suffers less from this timing problem insofar as only a small
proportion of offers would be made post-A level results. But it would suffer from one
very large drawback: it would result in two gathered fields.5 We think it essential that
all candidates face exactly the same level of competition for a place, giving candidates
5 It is important to distinguish between separate gathered fields, with a set proportion of places assigned in
advance to each gathered field, as proposed in the second PQA model, and the proposal to allow students who
achieved better than expected results to submit a new application, put forward by the DfES (see paragraph 17 (c)
above) as part of the proposed reforms to the current applications process. Under the latter proposal, places would
not be set aside for a second round of applications.
5
of equal ability an equal chance, as well as making it easier to apply selection criteria
uniformly to all candidates. It is also hard to see how in this model the many excellent
candidates who would be rejected in the first round could reasonably be excluded from
applying in the second round.
21. Clearly, the introduction of some form of PQA system would have far-reaching
ramifications for our admissions procedures. Several of the changes suggested in this
report, including the detailed proposals for moving to a single admissions model, would
have to be adjusted. However, even though the implementation of our proposals would
take several years, the introduction of PQA – and its timing – seem sufficiently
uncertain that it would be unwise to put them on hold until the position on PQA
becomes clearer.
1.2 Recent changes in the Oxford admissions system
22. The colleges and departments devote huge amounts of time and effort to the admissions
process. In contrast to many universities in the UK, academics are heavily involved
throughout. Their leading role is, in our view, essential if the very difficult choices
amongst the many excellent candidates are to be made, and if confidence in our
systems is to be maintained. The willingness of academics to devote so much of their
time to this activity demonstrates their strong commitment to achieving the very
highest standards of learning.
23. In the past few years there have been significant changes and improvements in the
handling of admissions. These have been partly the result of initiatives taken by
individual subjects, and partly the result of initiatives taken by the colleges collectively.
A shift toward greater co-ordination of admissions within particular subjects and
amongst colleges has been a marked trend.
24. The most important changes of the past few years – many recommended by an earlier
Working Party in 2002 - have been the following:
(a) most candidates are now interviewed twice, often by two colleges;
(b) several subjects use ‘smart’ second choice college allocation;
(c) prior to interviews, oversubscribed colleges increasingly redistribute some of
their candidates to less popular colleges;
(d) all interviewers receive guidance on how to interview and on equal opportunities
legislation;
(e) all subjects have a template interview report form and published selection
criteria;
(f) increasingly, organising secretaries, working in close liaison with the colleges,
co-ordinate much of the admissions process on behalf of their subjects;
(g) there is extensive sharing of information about candidates, pre-, mid- and postinterview,
through meetings, and via web-based systems;
(h) a growing number of subjects centrally rank applicants to determine, or to help
colleges decide, who to invite for interview;
(i) many subjects also centrally rank post-interview;
6
(j) an increasing number of subjects have open offer schemes;
(k) all colleges and departments have signed up to a Code of Practice and a
Complaints procedure;
(l) three subjects – History, Law and Medicine – have introduced a pre-interview
aptitude test as an aid to short-listing decisions;
(m) Medicine now has: central ranking at the short-listing stage, based on GCSE and
pre-interview aptitude test scores (this has enabled the ratio of interviewees to
places to be brought down to a more manageable 2.5:1); two independent
interviews for all short-listed candidates; and ‘blind’ interviews.
25. One recommendation in the 2002 Report that was not accepted by colleges was the
following:6
"Candidates should apply to Oxford (rather than an individual college) and
interviews should be conducted collectively. Candidates may name a college of
preference at the point of application, and this preference will be accommodated if
possible, in the event of the candidate being successful."
26. In view of developments in certain subjects towards ‘collectivisation’ since 2002 and
the beneficial effects this appears to have had, the observed weaknesses in admissions
processes in some subjects, as well as some of the external considerations mentioned
earlier, we have felt it right to return to this issue – and we have done so in
considerably greater depth than was the case in the 2002 report.
27. In examining whether the selection procedures now in place adequately meet the
objectives set out in paragraph 1 of this report, the Working Party has found it useful to
consider the issues under two broad headings:
(a) how applications are assessed: assessment of UCAS and Oxford application
forms, assessment of submitted work, use of contextual information (including
school performance data), pre-interview testing, use of AS module marks, use of
interviews, and so on;
(b) how the assessment and decision-making processes are organised: for example,
who decides on the short-listing of candidates for interview and on where
candidates are to be interviewed, who makes the final offer decisions, and what is
the balance between the colleges and the subjects in making these decisions?
2. Assessment
28. Whilst assessment methods are tailored to the requirements of particular subjects, for
the most part the information that applicants are required to provide is similar across
the university.
29. The key difference is between those subjects that have introduced pre-interview tests
and those that have not. As noted in paragraph 24, History, Law and Medicine now
6 It should be noted that when tutors for admissions voted on this motion individually the idea was passed but was
rejected when they voted as college representatives.
7
have such tests,7 designed to assess aptitude rather than acquired knowledge. Their
main purpose is to support short-listing decisions, allowing these subjects to assess on a
more informed basis the many candidates with roughly similar achievements and
predicted qualifications. As such, the three subjects have found them useful.
30. The reaction of schools to these tests has been mixed – in part because of some
organisational problems – and other subjects considering introducing such tests will
need to take great care to ensure that schools remain supportive.
31. If PQA with additional grades and differentiation at the top end of A levels were
introduced, the need for testing would diminish considerably. However, in the absence
of such changes, and as the number of applicants to Oxford increases and GCSE and A
level results improve further, the extension of aptitude tests to most subjects seems
probable. If that turns out to be the case, the Working Party would wish to see
particular attention given to the organisational aspects of pre-interview tests, so as to
minimise burdens on schools, colleges, and students.
32. It is unlikely that many subjects will be able to have multi-institution tests, as is the
case in Law and Medicine, largely owing to the differences in courses amongst
universities and the potential lack of agreement on the nature and content of tests. If
new tests are to be ‘Oxford only’, special efforts will need to be made – as they have
been in the case of History - to ensure that they are not seen as another barrier to entry
to those from less advantaged backgrounds. This will require special attention to test
design and content so that as far as possible the tests examine aptitude, and so that
results cannot be improved through coaching.
33. To this end, the Working Party recommends the adoption of the following key
principles for the establishment and running of tests:8
(a) proposals for the introduction of new tests should be thought out in detail before
they are presented to the University and the colleges for approval;
(b) tests should be instituted only where there is a real and demonstrable need to
short-list fewer applicants or to obtain information in addition to that provided by
A Level results;
(c) subjects should publish well in advance: a sample test paper; selection criteria to
be used when marking the test; detailed information on practical arrangements for
sitting the test; and information on how the test results will be used;
(d) the organisation of Oxford-specific tests must be professional, with a well
informed single point of contact available for schools, colleges, and prospective
students;
(e) subjects should carefully consider all the access implications of their tests, and
how any impact on access will be minimised – this includes ensuring that tests do
not constitute an unacceptable financial burden;
7 Medicine and Law use tests – the BMAT (BioMedical Admissions Test) and LNAT (Law National Admissions
Test), respectively – developed for several UK universities, whilst History uses the HAT (History Aptitude Test),
which is Oxford specific.
8 The principles have already been endorsed by ADEX and noted by EPSC.
8
(f) subjects should ensure that tests are designed to assess aptitude, address the
published selection criteria and do not contain inadvertent bias, and that any
benefits to be gained from coaching are limited;
(g) as far as possible, pre-interview tests should be piloted in Oxford during the
interview period, before they are used for short-listing. The percentage of
candidates not invited for interview should increase only gradually, as confidence
in the test grows and as statistics are collected that demonstrate that the test is
valid and reliable;
(h) tests should be monitored annually for bias and for validity (i.e. ‘fitness for
purpose’). The success rate of applicants on the whole test and on its constituent
parts should be broken down by various factors, including gender and school
type.
34. Proposals for the introduction of new tests should contain: comprehensive arguments
setting out the need for such tests; consideration of whether multi-institution or Oxfordonly
tests would be more appropriate, and summaries of any discussions held with
other universities and with other departments in Oxford; selection criteria to be applied
to the test; details of how the test will be used for selection, and how this is likely to
develop in future; and a sample test paper.
35. We comment now on the idea of a national aptitude test. Initiatives include the recent
launching of a five-year study into the validity of the use of a Scholastic Aptitude Test
(SAT) in university entrance, sponsored by the Sutton Trust and backed by the DfES,
and the 'uniTEST', an aptitude test designed to assess generic reasoning and thinking
skills across the two broad domains of mathematics/science and humanities/social
sciences. Developed jointly by Cambridge Assessment (formerly the University of
Cambridge Local Examinations Syndicate) and the Australian Council for Educational
Research (ACER), the 'uniTEST' is undergoing its first pilots in schools. We welcome
these initiatives and believe that a national aptitude test may be useful in providing
additional information on candidates’ academic potential. But we doubt whether it
would be able to substitute for our other assessment methods. It is unlikely that a single
generic test could adequately examine students across the spectrum of arts, science and
social science subjects and enable tutors to rank the large number of excellent
candidates without interviews and specific subject testing. Moreover, American
experience shows - most US universities require undergraduate applicants to take the
SAT - that contrary to the original intention, it is possible to coach effectively for the
more generic end of the aptitude test spectrum.
36. The Working Party considered several other assessment issues:
(a) submitted written work, whether requested or not, should generally not be used as
a basis for short-listing decisions because of uncertainties as to authorship, but it
could be used in interviews where it would become clear whether or not it was
the candidate’s own work;
(b) more systematic use should be made of contextual information, especially that
relating to school performance. The contextual information would only be to
inform judgements about academic ability, not for the purposes of positive
discrimination. The Oxford Colleges Admissions Office (OCAO) should provide
colleges with school performance data so that tutors can take into account school
9
performance in borderline cases;9
(c) there is a wide division of views in Oxford on the usefulness of AS scores in the
assessment of candidates. Those in Oxford who favour asking for AS scores
argue that since some applicants provide AS scores already, it is only fair to ask
all applicants for this information; and that it is worth having this information as
an additional tool for distinguishing amongst candidates. On the other hand, it is
argued that AS scores are not a useful predictor of potential for success at Oxford
- both because of the nature of the examination and because the approach of
schools to the taking of AS's differs widely, particularly in the re-sitting of
modules. From 2007, UCAS will ask for AS scores in all applications. Until then,
the Working Party was of the view that Oxford should not request this
information. Where candidates already provide this information, and from 2007
when provided by UCAS, it should be viewed with considerable discretion,
taking into account, where possible, the school context.
37. There is the broader question of the efficacy of interviews as part of the admissions
process. It was evident to the Working Party that most Oxford academics consider
interviews to be an essential element in the overall admissions process, believing they
allow for differentiation amongst candidates with almost identical paper qualifications
and enable the spotting of potential that might otherwise have been missed.
38. Taking into account best practice already in place at Oxford, the Working Party
believes there are various ways in which interviews can be made a more effective
selection tool.
39. First, there should be agreed standard formats for each subject, and there should be an
agreed interview mark scale. All candidates should be assessed against carefully
designed selection criteria (as opposed to assessing candidates against each other).
40. Second, the fairness and effectiveness of the interview process would be enhanced if all
candidates received a minimum of two independent interviews. This would give all
candidates greater opportunities for demonstrating their ability and would provide a
useful cross-check on tutors’ assessments. To avoid the last minute rush that often
takes place in arranging second interviews (where they do take place), these should be
scheduled in advance – with candidates informed in a timely manner where, when and
with whom they will meet. The number of interviews should be pre-set for each subject
in advance, and no additional interviews should be carried out.
41. Third, there should also be a pre-set number of interviews in all Joint Schools. Ideally,
the number of pre-set interviews would be more than two (and where there are three
subjects, more than three) so that in at least one subject in the Joint School there can be
two independent interviews. Alternatively, each subject in the Joint School could carry
out one interview, with representatives from both (or all) subjects combining to carry
out an additional independent interview. It is unlikely that subjects within a Joint
School would combine for all their interviews, as the length of interview slots would
not allow each individual subject to acquire enough information.
9 Cambridge has developed a simple model for adjusting GCSE scores for school performance, outlined at
http://www.admin.cam.ac.uk/offices/admissions/athandbook/ . Whilst it would be inappropriate to use contextual
information mechanistically, subjects may wish to investigate whether such models bring added value to the
admissions process.
10
42. Fourth, interviews should be of sufficient length to allow tutors adequately to explore
candidates’ abilities and potential. In addition, second interviews should be conducted
along the same lines as first interviews, and tutors conducting second interviews should
have received the same supporting material provided to the tutors who undertook the
first interview.
43. Fifth, interviews should be scheduled in such a way that no candidate need stay in
Oxford for more than one night, other than for reasons of distance and travel.
Exceptions should only be made if subjects choose to conduct aptitude tests during the
interview period.
44. Sixth, and as a consequence of several of the preceding points, more radical shortlisting
would be required. The ratio of interviewees to places should normally be
between 2.5 and 3.5 to 1. It is likely that the upper limit will only be appropriate for
subjects with relatively small numbers of applicants or large numbers of tutors. A
maximum ratio of around 3 to 1 would be suitable for most subjects. It is important that
a lower limit be set in order not to undermine the benefits of access work, and because
it is very difficult to differentiate amongst applicants who are not at the very top or very
bottom of the cohort. This would ensure that a sufficiently wide ‘middle band’ of
candidates is short-listed.
45. The Working Party also believes that the Collegiate University as a whole would
benefit from the implementation of open offer schemes in all subjects. This would
reduce the number of candidates accepted to the University who had not met their offer
conditions, and would obviate the need for individual colleges to make more offers
than they had places.
46. The Working Party concluded that, with existing assessment methods and procedures,
colleges can already be fairly confident that they are selecting the best candidates from
amongst those who apply or are presented to them. With the changes suggested in this
section of our report, they could do even better.
3. Organisation
47. Where Oxford continues to be rather less successful, despite the efforts that have been
made in recent years, is in ensuring that college choice – at least in some subjects -
does not affect the chances of admission. We now turn to this question.
48. The organisation of the admissions process in different subjects varies enormously. To
an outside observer the variety may seem bewildering and hard to comprehend. The
differences mainly concern the extent to which:
(a) there is co-ordination amongst colleges and across an individual subject in the
decisions to invite some candidates for interview and not to invite others;
(b) the subject or the colleges organise the interviews;
(c) subject-based ranking is used and the degree to which this ranking determines the
final decision to offer a place.
49. There are currently wide variations in the degree of co-ordination across the different
subjects, and in the way that decisions are taken in regard to both short-listing and offer
of admission. The matrix at Annexe E illustrates what is a veritable mosaic. In recent
11
years, most subjects have moved along the spectrum towards greater co-ordination.
Some have moved further than others. However, even within the relatively co-ordinated
subjects, procedures may vary from college to college. For example, in a number of
subjects that have implemented open offer schemes, not all colleges offering that
subject participate in the scheme.
50. Examples of subjects at different points on the spectrum are set out below:
(a) English, PPE: colleges are largely autonomous in relation to short-listing,
interviewing and final selection; colleges also divide into groups that operate
largely independently of each other;
(b) History, Law, Modern Languages: there is a fair degree of co-ordination and
some central ranking of candidates, but the colleges still mostly make
autonomous decisions;
(c) Medicine: central ranking largely determines who is interviewed. The department
organises the interviewing of candidates, with second interviews/colleges
allocated to candidates at random. Interviewing colleges make the final selection
decisions and there is no post-interview central ranking;
(d) Biochemistry, Engineering, Materials: central ranking plays a large but not
determining role in deciding who is interviewed and who is admitted. Each shortlisted
candidate receives interviews at two colleges and the second college is
allocated on a ‘smart’ basis;
(e) History of Art: the whole process is centralised with final decisions being made at
the departmental, rather than the college, level.
51. Why this diversity along the ‘co-ordination spectrum’ exists is not wholly clear. The
extent of co-ordination has tended to be greater in small subjects, and in the sciences
generally. Where there are relatively few candidates and places, ranking the whole
body of candidates may be easier. Likewise, if subject tutors in different colleges know
each other well, they may be more likely to accept the judgement of their colleagues.
52. The experience of subjects where there is a relatively high degree of co-ordination
illustrates what can be achieved when admissions reforms are pursued co-operatively.
In subjects that have not moved as far along the co-ordination spectrum, two arguments
are often brought to bear against such co-ordination, and particularly against central
ranking. The first maintains that available quantitative methods – a prerequisite to
establishing a central ranking - inadequately capture qualitative assessments. The
second contends that inconsistent application of assessment criteria by some tutors can
skew central rankings.
53. The experience of several subjects which use central ranking in the short-listing and
final decision stages suggests that neither argument poses a substantive obstacle to
effective central ranking. If quantitative methods are applied to a sufficiently broad
range of qualitative inputs, including two independent interviews at minimum, the
central ranking that emerges accurately reflects the qualitative judgements.
Engineering, Materials and Biochemistry interview short-listed candidates at two
colleges and so the effects of outlier marking – already limited through the use of
common marking schemes and the possibility of cross-correlating interview scores - are
diluted in the final rankings. Once this is considered, continuing doubts about the
12
comparability of marks and the ‘ordering’ of candidates, essentially comes down to the
issue of whether tutors feel they can trust the judgement of their colleagues in other
colleges, either in the choice of criteria they use to assess candidates or in the way they
apply them.
54. Furthermore, tutors in all subjects do rank candidates whom they have seen and wish to
accept or reject, which rather weakens the argument that the ranking is unable fully to
encompass qualitative judgements. As regards the issue of mutual trust, all tutors
accept cross-comparability of marking in University Finals Examinations, and are
content to trust their colleagues to ‘rank’ their students and to place them in particular
‘categories’ or degree classes.
55. It is obvious that, other things being equal, the more co-ordination there is across a
subject, the better the chances that the best candidates who apply to Oxford will be
identified and selected. This is not to decry the work and skills of those who conduct
the assessments in individual colleges. As indicated earlier, from amongst those whom
they assess, they generally do a good job of picking the best. But working on their own
or without close liaison with colleagues across the colleges, they have no way of
knowing whether their best candidates are better than other candidates who may have
been rejected by another college. It is precisely at the margin of each college’s field,
and in the absence of centralised ranking and co-ordination, that it is most likely that
college choice will affect a candidate’s prospects of a place. This is because, without
central ranking and high levels of co-ordination, colleges are more likely to fill their
places from their own cohort of first choice applicants than to look outside that cohort
for candidates of higher quality.
56. In subjects that have moved to a more co-ordinated admissions process, the patterns of
offers made to candidates have changed significantly. This includes: (i) Medicine,
where the introduction of central short-listing, two independent interviews per shortlisted
candidate, and blind interviews, have dramatically increased the number of
candidates taken at a college other than the one they chose or were allocated; (ii)
Biochemistry, where the introduction of central ranking has resulted in a significant
number of candidates being made offers at colleges which had not interviewed them;
and (iii) Mathematics, Medicine and Physics, where the implementation of ‘open offer’
systems has resulted in under two percent of the final intake being composed of
candidates who did not meet their offer conditions in full.
57. The evidence that changes in process have resulted in changes in outcome in these
cases suggests that the previous processes did not fully meet the University’s principle
that candidates’ college preferences should not affect their prospect of admission. It
should be noted, however, that one side-effect of the move to collective decisionmaking
in various subjects has been that the responsibility for decision-making has
become blurred. Whilst final decisions on admissions remain with colleges, there are
intermediate decisions and processes that departments and faculties are undertaking
over which colleges have little or no control; yet it is the colleges that are ultimately
accountable for all admissions decisions at every stage of the process and it is the
colleges that have to deal with complaints from candidates. This blurring of the link
between responsibility and accountability can cause difficulties for colleges; can be
difficult for schools and candidates to understand; and makes the handling of
complaints more complex both for the Colleges and the candidates. Moreover, it
undermines the efficacy of the existing complaints mechanism, as neither departments
13
nor colleges can properly hold the other to account for instances of poor process.
4. Admissions models
58. The Working Party believes it is time that Oxford moved to a single admissions process
for all subjects which draws on best practice and comes as close as possible to
achieving the University’s admission and selection objectives. This would not preclude
variations in the details of the assessment process for different subjects (for example,
whether or not subjects choose to run pre-interview tests); but it would mean the
adoption of some common key practices.
59. Quite apart from whether the different models currently in place across the subjects are
effective, the existence of such a variety of models has operational disadvantages. The
administration of admissions in colleges has become increasingly complex and
burdensome in recent years, partly because of the different arrangements for different
subjects. Current IT systems for undergraduate admissions (in particular, ADMIT) are
struggling to cope with these different arrangements. There will need to be up-grading
or, more likely, replacement of these systems in any case in the next few years whether
or not a single admissions model is adopted; but with a single admissions model for all
subjects, it might be reasonable to expect that development costs for a new upgraded
system would be less. A single undergraduate admissions model would also be more
understandable for candidates and their schools and colleges.
60. The Working Party has developed two models, either of which, if adopted, would
remove the influence that choice of college may still have over the chances of gaining a
place. In doing so, they would move the University and colleges a step further towards
the ideal goal of trying to ensure that the very best candidates in every subject who
apply to Oxford are admitted. As explained in the previous section of this report, there
are some subjects that have moved close to this goal already. Our aim is to establish a
single model which will move all subjects to this position.
61. The two models have one key element in common: all candidates for a particular
subject would be considered alongside each other at both the short-listing and final
offer stages by the collectivity of tutors (or a sub-group of it). They differ in the
following key aspects:
(a) in Model A, applicants would not state a college preference when they applied –
they would do so only after they had been offered a place. In Model B, they
would continue to have the option of stating a college preference at the time of
application;
(b) in Model A, the collectivity of tutors would have the final say as to who would be
interviewed and offered a place. In Model B, colleges would retain the final
responsibility for short-listing and making offers: they would be guided by, but
would not be bound by, the collectivity of tutors.
62. The use of the word ‘subject’ in the following sections denotes the collectivity of
college tutors (or a sub-set of them) in a particular subject.
14
4.1 Model A
63. The main features of Model A are as follows.10
(a) all applicants would apply to the University in the first instance;
(b) colleges would inform organising secretaries how many places they had for their
subjects;
(c) applications would be dealt with by subjects, not colleges;
(d) candidates would be ranked11 and, using this central ranking, subjects would
decide whom to short-list for interview;
(e) subjects would pre-arrange interviews, setting the number and the timetable –
organising secretaries of Joint Schools subjects12 would liaise with their
colleagues in the relevant single honours subjects when determining the number
and time of Joint Schools candidates’ interviews;
(f) subjects would conduct interviews – at least two independent interviews per
short-listed candidate;
(g) interviews would normally take place in colleges but might also take place in
departments and faculties, and many international candidates would be
interviewed overseas;
(h) candidates would be ranked, and offers would be made by the subject - the
numbers equating to the places the colleges had notified plus a number of open
offers to allow for candidates turning down their places or failing to meet the
conditions of their offers;
(i) those offered a place would be guaranteed a place at a college;
(j) after being offered a place, candidates would be asked if they wished to state a
college preference - candidates could name up to two colleges, by order of
preference;
(k) each college would choose initially from the pool of candidates that had
expressed a preference for it, but the college would be under no obligation to
accept those who gave it as their first or second preference;
(l) if colleges were unable to fill their places from those who had named them as
first or second preference, the collectivity of tutors would seek to agree college
placings by mutual consent (in the last resort, if agreement could not be reached,
there could be a ‘smart’ allocation of any remaining candidates);
(m) candidates who did not name a college of preference would be placed by mutual
consent amongst the collectivity of tutors, with ‘smart’ allocation used, if
necessary, as a last resort.
64. The main benefits of this model would be as follows:
10 Annexe C provides further details, as well as possible variations, on Model A.
11 Ranking should be based on a total score of 100, which provides sufficient gradation in the rankings to allow
candidates to be distinguished from one another. The score could include several weighted elements.
12 For a more detailed discussion of Joint Schools arrangements under Model A, see Annexe C, paragraph 5.
15
(a) Oxford could be as confident as possible that, given the range of assessment
methods used, in all subjects the best applicants would be offered places;
(b) tutors could be sure of teaching those students who were from the best of those
applying to Oxford;
(c) colleges that hitherto received relatively few, or relatively weak, applications in
particular subjects could be sure of getting very good students;
(d) in subjects where co-ordination is currently rather limited, the possible impact of
college choice would be overcome;
(e) in all subjects, there would be no longer the perception that admission might be
affected by college choice;
(f) applicants with limited knowledge of Oxford would find it easier to apply, and
the University might receive a larger number of well qualified applicants from
the maintained sector;
(g) it would be simpler, more transparent and would command greater public
confidence;
(h) in subjects where there is already a significant degree of collectivisation, there
would no longer be any blurring between responsibility and accountability. The
University, through the departments and faculties, would be exclusively
responsible and accountable for all decisions.
65. The arguments against this model can be summarised as follows:
(a) tutors want to see and make their own judgments on students they are going to
teach;
(b) tutors would not put the same effort into the selection process if those admitted
were not necessarily going to be tutored by them;
(c) it would involve central ranking of candidates (but see paragraphs 53-54);
(d) tutors might feel less responsible for ensuring their students succeed
academically, if they did not take the decision to admit them;
(e) colleges would have less incentive to undertake access work, and to do all in their
power to attract and select the most able students;
(f) colleges would lose an element of autonomy, and individuality, insofar as they
would no longer play the principal role in admissions decisions.
66. These counter-arguments are held strongly by some and therefore cannot be dismissed
lightly, especially as any changes that reduced the commitment on the part of tutors to
the admissions process and to the teaching of those admitted would be seriously
detrimental to Oxford. The Working Party, nonetheless, feels that to some degree they
are more a matter of perception than of substance. In particular:
(a) the desire on the part of tutors to make their own judgements on candidates
whom they may end up teaching is understandable as an ideal, but it suggests a
lack of confidence in the judgement of their fellow tutors that rarely in our view
reflects reality;
(b) with increasing specialisation both in degree courses and amongst academics,
16
college tutors see less of their ‘own’ students than in the past, as more teaching is
done outside the college. Undergraduates now have academic relationships with a
variety of tutors across the colleges. Although the college tutor has overall
responsibility for the individual, the college tutor/undergraduate relationship is
only one amongst several tutor/undergraduate relationships;
(c) it is already not uncommon, on account of sabbaticals and other absences, the use
of more than one team of interviewers or the allocation of students through the
open offer system, for a college tutor not to have interviewed undergraduates for
whom he or she ultimately will be responsible. In addition, the lack of a college
preference at the time of application would eliminate any possibility of
inadvertent bias that might result from tutors feeling that they had a particular
responsibility towards their first choice applicants;
(d) individual colleges would still have an incentive to try to attract the best students
and to undertake access work insofar as successful applicants would still be able
to state a college preference after they had been offered a University place. Many
of those tutors involved in access work in any case regard what they do as
intended to benefit the University as a whole rather than just their own colleges;
(e) In subjects such as Biochemistry, where there is currently a high degree of coordination,
decision-making has already shifted to a considerable degree from
individual colleges to the collectivity of tutors. It is debatable how significant the
accompanying loss of college autonomy has been in these subjects, and whether
it has really affected colleges’ individuality. Any reduction in college autonomy
and perhaps individuality have to be weighed against the better undergraduate
intake at the margin at the less popular colleges, as well as the enhanced equality
of opportunity for all candidates that should result from this model.
4.2 Model B
67. The main features of Model B are as follows:13
(a) applicants would apply to the University;
(b) applicants would state a single college preference, should they wish, or submit an
open application;
(c) applicants’ UCAS forms and pre-interview tests (if used), would be graded14 by
the subject;
(d) based on this grading, applicants would be ranked. The subject would take a view
on how many candidates should be short-listed and, based on the rankings, the
subject would produce a short-list of candidates for interview. Those short-listed
by the subject would all be interviewed, but colleges could also choose to
interview a small number of candidates who were not on the short-list;
(e) candidates would be interviewed by two separate colleges: the preferred college
(or assigned college in the case of open applications), and a college allocated
according to the ‘smart’ allocation process;
13 See Annexe C for a discussion of blind interviewing, which some subjects use at present; the Working Party did
not include it in Model B, deciding that other aspects of Model B would achieve the same end.
14 For example, in Biochemistry three tutors independently grade all candidates’ UCAS forms.
17
(f) applicants would be centrally ranked again and the collectivity of tutors (or an
elected sub-group of them) would meet to decide to whom to give an offer;
(g) if candidates were ranked highly enough to merit an offer, but could not be taken
by either of the colleges that had interviewed them, the expectation would be that
tutors would co-operate to ensure that they would be taken by a college that still
had places to offer, without further interviews being conducted;
(h) colleges could not be obliged to take candidates according to the central rankings,
but if tutors wished to make offer decisions that conflicted with the rankings, they
would have to make a case to the meeting of tutors referred to in (f);
(i) tutors would collectively determine, according to the central rankings, the
candidates to whom open offers should be made.
68. The benefits of this model in several respects parallel those of Model A:
(a) it would ensure that the very best candidates are offered places (provided any
‘over-ride’ of central rankings by colleges – see paragraph 67 (h) – is very much
the exception and the case for over-ride is made on academic grounds alone15);
(b) by the same token, and with the same proviso, it would no longer be possible to
claim that college choice influenced the chances of being admitted, and all
colleges would be sure of admitting very good candidates.
69. In addition, it would avoid or mitigate the various disadvantages listed in paragraph 65
which would – in the minds of some – accompany Model A. But equally, there are
some advantages of Model A which it would not have, namely:
(a) there might still be the perception that college choice does matter and does
influence the chances of getting a place;
(b) this perception, along with the perceived need to choose a college at the time of
applying, might continue to deter some candidates;
(c) whilst simpler and more transparent than the current plethora of arrangements for
different subjects, it would lack the simplicity and transparency of Model A;
(d) the blurring of responsibility and accountability between colleges and
departments would continue, as would the situation where two or more colleges
can be accountable for different parts of the admissions process for a single
candidate, both leading to concomitant effects on the complexity of the
complaints process and on public perception.
4.3 Administrative and IT implications
70. The adoption of either model would have significant administrative and IT system
implications. The administrative work-load for departments and faculties would be
increased – especially where there is at present limited coordination at
department/faculty level. The administrative work-load for colleges, especially under
15 It is expected that cases such as these will be very rare, but an example of such a case might be as follows: A
student scores fairly poorly in the Mathematics admissions test. The teacher’s reference states that this student has
missed large portions of the syllabus through illness. Tutors agree that the student has greater potential than her
test score seems to indicate, and they believe that her poor result was due to missed schoolwork. Had this not been
the case, tutors believe that the candidate would have obtained a better mark and appeared higher in the rankings.
18
Model A, would be reduced. Further work would be needed on this, including how the
extra costs to be borne by departments/faculties would be funded.
71. As indicated in paragraph 59, existing IT systems will need to be up-graded or replaced
in the next few years with or without major change in admissions processes.
Replacement could be either by the extension of the new student record system (OSS)
to the whole of undergraduate admissions or through a new bespoke system. If a single
admissions model were to be adopted, up-grading or replacement would clearly need to
be tailored to it in terms of functionality and timing. Further work on alternative IT
options and their costings needs to be undertaken urgently, and appropriate budgetary
provision obtained in due course.
4.4 Timetable 16
72. The Working Party has concluded that, if Model A were adopted, ideally all subjects
would move to this model at the same time. Given the need for extensive preparation,
the Working Party considers that all subjects could move for the 2008 admissions
cycle, for students applying for admission in 2009. Support for implementation in
departments would need to be provided by the central University.
73. The Working Party also considered the possibility of a phased move to Model A. A
phased move would not hold back subjects that wished to adopt Model A earlier than
2008. Moreover, it could be argued that such a substantial change should only be
undertaken gradually, with the more co-ordinated subjects moving earlier than others.
74. However, a phased implementation would present significant disadvantages:
(a) the already difficult situation in regard to the administration of admissions (see
paragraph 59) would become virtually impossible to manage;
(b) it would be very hard to explain simply and coherently to candidates - in person,
in the printed Prospectus, and on the University’s website - the significant
differences amongst the application and selection procedures in different
subjects, and the reasons behind them;
(c) it could potentially create difficulties for UCAS, which from last year has
required that all applicants to Oxford provide in their application a ‘campus code’
that denotes either their choice of college, or their choice of an open application;
(d) IT systems – essential for storing and processing applicant data, making
information available to colleges and departments, and communicating with both
UCAS and applicants - could not support effectively such variations in
admissions models.
75. If it were decided to adopt Model B it would not be necessary for all subjects to move
together, but all subjects should do so by the 2008 admissions cycle (for entry in 2009).
5. International applicants
76. The Working Party also reviewed the assessment of international applicants in order to
ensure that the equity of the admissions process regarding all, not just UK applicants,
16 Annexe D provides a detailed timetable for implementation.
19
was addressed. In addition, the Working Party was mindful of the University’s wish to
recruit the best candidates, irrespective of country of origin.
77. It was noted that the success rate of overseas applicants (including continental
European candidates interviewed in Oxford, and other candidates interviewed abroad)
is much lower than that of UK students. There are a number of possible reasons for
this: poor English language skills; inappropriate qualifications; poor references and
personal statements (owing to a lack of understanding of the kind of information
required); and over-clustering in certain subjects such as Law (although even within
these subjects the success rates for overseas applicants are lower).
78. It was also suggested that success rates for overseas applicants from outside Western
Europe are lower because they are not interviewed in Oxford. Although interview
panels are sent to North America and the Asia-Pacific region, candidates may be
interviewed by a tutor who is not a specialist in an appropriate subject area. However,
this argument is belied by the fact that EU candidates who are interviewed in Oxford
also have low success rates – lower in fact than candidates interviewed in the Asia-
Pacific region.
79. The Working Party concluded that there was little evidence that calling overseas
candidates for interview in Oxford would lead to higher success rates. In addition, in
the face of serious international competition for students, sending interviewers abroad
demonstrates Oxford’s commitment to global access and to international students in
general. However, there is little sense in interviewing all international applicants
abroad, regardless of quality. For the sake of equity (and logistics), overseas candidates
should be considered for short-listing at the same time as all other applicants. Only
short-listed candidates should be interviewed abroad.
80. This could be achieved by the following means:
(a) the short-listing deadlines for all candidates would be brought forward to mid-
November (or even earlier);
(b) overseas Arts interviews would take place in Week 8 of Michaelmas Term;
(c) overseas Science interviews would take place in Week 9 of Michaelmas Term;
(d) separate interview teams would be used for North America and the Asia-Pacific
region (making four teams in total);
(e) overseas interviewers would take part in all relevant tutor meetings in the
subjects of candidates they interviewed, in order to represent the interests of those
applicants.
81. This timetable would be feasible if written work were not used for short-listing but only
for discussion during interview (as would be the case for those interviewed at Oxford).
The later timing of overseas interviews would also allow use of these candidates’
written work during interview, which is not currently possible. The major complication
in this timetable is presented by pre-interview tests, for which results may not be
available until mid-November at the earliest. Candidates in relevant subjects could
20
either be given shorter notice of their interview date, or the possibility of earlier preinterview
tests could be considered.17
82. The Working Party also concluded that Divisions should appoint overseas interviewers
and that tutors in the same subject or in a cognate subject within the same Division
should interview candidates. This would increase tutors’ confidence in the results of
these interviews. In addition, an increase in pre-interview testing would benefit
overseas applicants by providing another mechanism for demonstrating their potential.
6. Oxford Colleges Admissions Office
83. Irrespective of the model of undergraduate admissions adopted, the Working Party
believes that a review of the structure and position of the Oxford Colleges Admissions
Office, including its relationship to the University, should take place. This review
should be undertaken regardless of whether other changes in admissions procedures are
pursued, and should be set up as soon as possible. A restructured Admissions Office
operating as part of the central University, and possibly as part of a single
graduate/undergraduate admissions office, would seem to have clear advantages, whilst
still being able to serve the interests of the colleges. Changes in the operational
affiliation of the Admissions Office would only happen after full consultation with the
staff and unions concerned.
7. Summary and Conclusions
84. The Working Party concluded that, in general, individual colleges – supported by
departments and faculties - do a good job in assessing and selecting applicants for
undergraduate study. Action, however, is recommended in the following areas:
• Unless and until PQA is introduced, pre-interview testing will probably have
to be extended to more subjects. Various principles are suggested for the
setting up of any new tests (paragraphs 33-34).
• Written work, whether requested or unsolicited, should not be used for shortlisting
but used only as a basis for discussion during interview (paragraph 36
(a)).
• Contextual information on candidates needs to be used more systematically
(paragraph 36 (b)).
• AS module marks, when provided by candidates, need to be treated with
caution; and for the time being, Oxford should not request these marks
(paragraph 36 (c)).
• Various measures are suggested to make interviews a more effective and fairer
selection tool (paragraphs 39-44).
• Open offer schemes should be extended to all subjects (paragraph 45).
85. Despite the various measures that have been introduced in recent years to redistribute
strong candidates amongst colleges, the Working Party considers that further action is
needed to ensure that, in all subjects, as far as is reasonably possible the very best are
17 At present, the BMAT, LNAT and HAT are sat on the first Wednesday in November.
21
offered places somewhere in the University. In order to achieve this, and to make the
admissions system simpler and more transparent, the Working Party has concluded that
all subjects should move to a single admissions process. It has developed two possible
models. Both of them build on current best practice in several subjects where tutors,
acting collectively, already play a major role in decision-making on short-listing and
final offers. The adoption of a single model would extend collective consideration of
candidates to all subjects. The respective advantages and disadvantages of the two
models need to be weighed in deciding which one – or a close variant of either – might
be adopted (paragraphs 58–69).
86. International candidates should be subject to the same short-listing process as UK/EU
candidates, and where interviews have to take place overseas, they should be
interviewed by subject specialists (paragraphs 76-82).
87. The structure and position of the Oxford Colleges Admissions Office need to be
reviewed (paragraph 83).
88. Annexe D provides a suggested timetable for our proposals. Some of them can be
introduced relatively quickly. Others would require a longer lead time. For a single
admissions model to be implemented in time for the 2008 admissions round, a decision
would be needed by June 2006.
22
Annexe A
GLOSSARY OF TERMINOLOGY
Blind interviews – interviews conducted without tutors knowing their interviewees’
college preference.
Central ranking (or ‘ranking’) – the process undertaken at subject level to rank (i.e.
to order) applicants. Ranking can take place at any point in the admissions process, but
is particularly important for short-listing and for making final admissions decisions.
Ranking based on a total score of 100 provides sufficient gradation to allow candidates
to be distinguished from one another.
Independent interviews – interviews conducted by two different individuals (or
panels), representing more than one college. In the context of Oxford admissions, two
interviews within a single college are not considered ‘independent’.
International – status of all candidates who are not resident within the UK. However,
it should be noted that some candidates resident in the UK have ‘overseas’ status for
fees purposes.
Open application – application without a college preference.
Open offer (or ‘over offer’) - the offer of a place without a college specification. Open
offers are determined at the subject rather than college level. Subjects that run ‘open
offer’ schemes offer places to some candidates on the basis that their college place will
only be determined once examination results are known in August. Candidates are then
allocated to colleges once it is known which colleges have places available through
having ‘lost’ candidates (either through applicants rejecting their offers or through
failing to meet their offer conditions). Open offers are essentially ‘over offers’ – offers
over and above the number of places that all colleges have collectively for a particular
subject. Subject-based open offer schemes therefore obviate the need for colleges to
over-offer in particular subjects.
Organising Secretary – academic responsible for coordinating admissions for a
particular subject.
Redistribution – the process by which candidates are invited for interview by a college
other than the one they originally chose or were allocated. Redistribution aims to even
out the number and quality of applicants interviewed across the colleges. Candidates
who are redistributed are those who are strong but who are unlikely to be offered a
place by their original first choice college, due to the large number of strong candidates
for that subject applying to that college.
Smart allocation - a procedure that aims to even out both the ratio of applicants per
place or interviews per tutor, and also the quality of applicants at each college. A
‘smart’ second choice allocation ensures that strong applicants who are unlikely to
receive a place at their college of preference are seen by colleges with weaker fields,
which are more likely to be in a position to offer them a place.
Annexe B
Working Party on Selection and Admissions
Terms of Reference
1. The Admissions Executive first established in 2002 the Working Party on Selection and
Admissions. It was reconstituted by the Admissions Executive in Michaelmas Term
2004 with the following terms of reference:
(a) to review selection procedures and arrangements in individual subjects;
(b) to consider the potential impact of recent developments, including the Schwartz
Report and the establishment of the Office for Fair Access;
(c) to review progress made in relation to the changes in 2002 and 2003; and
(d) to make recommendations to the Conference of Colleges and to Council’s
Educational Policy and Standards Committee, through the Admissions Executive.
Membership
2. The membership of the Working Party was as follows:
Sir Tim Lankester (President of Corpus Christi College, Chair)
Dr Diana Walford (Admissions Executive, Principal of Mansfield College)
Dr Bill Macmillan (Pro-Vice Chancellor (Academic))
Mr William Swadling (Tutor for Admissions, Brasenose College; CUF Lecturer,
Faculty of Law)
Dr Emma Smith (Tutor for Admissions, Hertford College; CUF Lecturer in English
Language and Literature, Faculty of English Language and Literature)
Dr Mark Wormald (Tutor for Admissions, Corpus Christi College; University
Research Lecturer, Department of Biochemistry)
Dr Richard Cross (Admissions Organising Secretary, Theology; CUF Lecturer,
Faculty of Theology; Tutorial Fellow, Oriel College)
Dr Dave Popplewell (Admissions Organising Secretary, Psychology; Tutorial
Fellow, Brasenose College)
Ms Ann Clayden (Project ISIDORE) (from January 2005)
Ms Kirstie Fieldhouse (Admissions Secretary, Oriel College)
Vice President (Access), OUSU (Until July 2005: Ms Linsey Cole; from July 2005:
Ms Charlynne Pullen)
Miss Jane Minto (Director, Oxford Colleges Admissions Office) (until February
2005)
Ms Louise Horsfall (Acting Director, Oxford Colleges Admissions Office) (March-
July 2005)
Secretaries to the Working Party: Dr Litsa Biggs (Oxford Colleges Admissions
Office) and Ms Lynne Hirsch (Central Administration)
23
24
Annexe C
Models A and B
1. This annexe provides, for ease of reference, the key features of both models and
discusses several points in detail that the body of the report addressed only in brief. The
use of the word ‘subject’ in this annexe, denotes the collectivity of college tutors (or a
sub-set of them) in a particular subject.
Model A
2. The key features of Model A are outlined below:
(a) all applicants would apply to the University in the first instance;
(b) colleges would inform organising secretaries how many places they had for their
subjects;
(c) applications would be dealt with by subjects, not colleges;
(d) candidates would be ranked1 and, using this central ranking, subjects would
decide whom to short-list for interview;
(e) subjects would pre-arrange interviews, setting the number and the timetable –
organising secretaries of Joint Schools subjects would liaise with their colleagues
in the relevant single honours subjects when determining the number and time of
Joint School candidates’ interviews;
(f) subjects would conduct interviews – at least two independent interviews per
short-listed candidate;
(g) interviews would normally take place in colleges but might also take place in
departments and faculties, and many international candidates would be
interviewed overseas;
(h) candidates would be ranked, and offers would be made by the subject - the
numbers equating to the places the colleges had notified plus a number of open
offers to allow for candidates turning down their places or failing to meet the
conditions of their offers;
(i) those offered a place would be guaranteed a place at a college;
(j) after being offered a place, candidates would be asked if they wished to state a
college preference - candidates could name up to two colleges, by order of
preference;
(k) each college would choose initially from the pool of candidates that had
expressed a preference for it, but the college would be under no obligation to
accept those who gave it as their first or second preference;
(l) if colleges were unable to fill their places from those who had named them as
their first or second preference, the collectivity of tutors would seek to agree
1 Ranking should be based on a total score of 100, which provides sufficient gradation in the rankings to allow
candidates to be distinguished from one another. The score could include several weighted elements.
25
college placings by mutual consent (in the last resort, if agreement could not be
reached, there could be a ‘smart’ allocation of any remaining candidates);
(m) candidates who did not name a college of preference would be placed by mutual
consent amongst the collectivity of tutors, with ‘smart’ allocation used, if
necessary, as a last resort.
3. Variants to this model are of course possible. For example, candidates could make a
choice of college at the time of application, and this could be ‘hidden’ from all
selectors until after offers had been decided. However, from the point of view of
transparency there would be disadvantages – applicants might still believe that their
choice could affect their prospects of admission. In addition, the factors of stress and
mystique sometimes associated with choosing a college would remain.
4. Another variable in the above model is the composition of interview panels. Although
interviews would take place mainly in colleges, it would not be necessary for interview
panels to be composed of tutors from the same college. Mixed-college panels might be
preferable as they would give representatives of a larger number of colleges the
opportunity to interview the candidates. For example, if each candidate had two
independent interviews with mixed panels consisting of two tutors from different
colleges, four colleges would have seen the candidate. In addition, mixed college
panels could reinforce for candidates the point that interviews were subject-based.
5. Model A would ensure greater consistency in the way Joint Schools candidates were
assessed and selected. Each Joint School would require an Organising Secretary who
would work closely with the Organising Secretaries in each component subject (with
the exception of subjects such as Economics and Management, where the component
subjects are not single honours courses). There would be a clear quota of places for
each Joint School, and interviews for all candidates would be organised in advance, as
for single honours applicants. As well as being ranked within their own cohorts, Joint
Schools candidates could also be ranked alongside those in the component single
honours schools. Should they be unsuccessful in obtaining a place for their chosen
course, they could then be considered for a single honours course. Such a system would
avoid the problems currently associated with some Joint Schools, which include: the
difficulty of organising second and lower choice interviews for these candidates; the
fact that many colleges do not have separate quotas for these students and may only
take a Joint Schools candidate once every few years; the difficulty in some subjects of
considering these candidates for one of the component single honours subjects, should
they not be suitable for the joint course; and finally the communications difficulties
amongst tutors in different subjects, spread across the colleges, during the busy
admissions period.
6. It should be noted that in this model, colleges would retain the incentive to undertake
access and recruitment work, as applicants would still have the option to name a
college of preference and colleges would be able to choose from amongst those who
named them.
7. The possible implications of Model A for St. Hilda’s (female only), Harris Manchester
(mature only) and Permanent Private Halls (PPHs, institutions with a religious
character) would need to be considered. However, the impact of the model is unlikely
to be very different to that of current procedures. Under the proposed model, as at
26
present, St. Hilda’s, Harris Manchester and PPHs would still be able to fill their places,
and candidates would be able to decline their offers if they so wished. As at present,
under the proposed model candidates could name a PPH or Harris Manchester as their
second choice. (The key difference is that under the proposed model the second choice
would not be restricted to PPHs and Harris Manchester, as under current procedures.)
8. Under this model appropriate procedures would need to be developed for candidates
applying for organ awards, who at present are interviewed in advance of the December
gathered field.2
Model B
9. The key features of Model B are outlined below:
(a) applicants would apply to the University;
(b) applicants would state a single college preference, should they wish, or submit an
open application;
(c) applicants’ UCAS forms and pre-interview tests (if used), would be graded3 by
the subject;
(d) based on this grading, applicants would be ranked. The subject would take a view
on how many candidates should be short-listed, and based on the rankings, the
subject would produce a short-list of candidates for interview. Those short-listed
by the subject must be interviewed, but colleges could also choose to interview a
small number of candidates who were not on the short-list;
(e) candidates would be interviewed by two separate colleges: the preferred college
(or assigned college in the case of open applications), and a college allocated
according to the ‘smart’ allocation process. All tutors could interview either the
same number of candidates, or else the same ratio of candidates to places at their
college;
(f) applicants would be centrally ranked again and the collectivity of tutors (or an
elected sub-group of them) would meet to decide to whom to give an offer;
(g) if candidates were ranked highly enough to merit an offer, but could not be taken
by either of the colleges that had interviewed them, the expectation would be that
tutors would co-operate to ensure that they would be taken by a college that still
had places to offer, without further interviews being conducted;
(h) colleges could not be obliged to take candidates according to the central rankings,
but if tutors wished to make offer decisions that conflicted with the rankings, they
would have to make a case to the meeting of tutors referred to in (f);
(i) tutors would collectively determine, according to the central rankings, the
candidates to whom open offers should be made.
2 Whilst constituting a separate gathered field for the admissions exercise, evidence suggests that the academic
performance of organ scholars is equal to that of other undergraduates. An analysis of available degree results
from 2001 to 2004 indicates that of 22 organ scholars for whom results are known, four achieved Firsts and
thirteen achieved Upper Seconds.
3 For example, in Biochemistry three tutors independently grade all candidates’ UCAS forms.
27
10. The Working Party considered whether Model B should incorporate ‘blind’
interviewing, and decided to recommend against it, for the following reasons (i.e.
colleges would know the candidates’ college of preference):
(a) ‘blind’ interviewing is intended to eliminate or at least vastly reduce the potential
for tutors to give places preferentially to their own first choice candidates. A
central ranking system where tutors collectively select, according to the rankings,
the candidates to whom places will be offered, readily achieves this aim;
(b) ‘blind’ interviewing would be extremely complex to operate for all subjects under
the current system (and under Model B), whereby applicants select a college at
the time of application. There would be an additional administrative load on
Organising Secretaries, and there would be duplication of effort and resources.
11. It could be argued that ‘blind’ interviewing has led to significant improvements in the
pattern of offers in Medicine and that, without blind interviewing, tutors would be able
to adjust the grades of their own first choice candidates in order to ensure that they
appeared higher in the rankings. However:
(a) the introduction of two independent interviews and double marking of all pieces
of assessment information would essentially eliminate the concern of grade
inflation;
(b) the change in the pattern of offers in Medicine - the marked rise in the number of
candidates taken by colleges other than their first-choice college, in itself a
positive development - cannot be attributed to any single factor as several
changes to procedures were introduced at the same time. These included a preinterview
test; extensive short-listing (which narrowed the field of interviewees
and therefore guaranteed that applicants were of comparatively high quality);
interviews at two colleges for all short-listed candidates; and blind interviews.
Any one of these factors - or a combination thereof - could have contributed to
the changed offer pattern;
(c) candidates in Medicine are not taken by colleges that have not interviewed them,
and blind interviews are desirable in order to maximise applicants’ chances of
obtaining a place at one of the two colleges that had interviewed them. Blind
interviews ensure that the Medicine process is as fair as can be in a system where
the final selection decisions remain exclusively in the hands of individual college
tutors and are not guided by central rankings.
28
Annexe D
Timetable for proposals
Short-listing and pre-interview tests
1. The ratio of short-listed candidates to places should not normally exceed 3.5 to 1 – to
be implemented by subjects as soon as is feasible (see below).
2. Introduction of pre-interview tests in subjects that are already in the planning stage –
2006 admissions round.
3. All other subjects to seriously consider introducing pre-interview tests - during 2005-
06. Earliest introduction of such tests probably not before the 2007 admissions round
(latest decision by December 2006).
4. Keep under review emerging proposals on a national aptitude test – ongoing.
Written work
5. Written work to be used at interview and not for short-listing – for implementation in
the 2007 admissions round ( decision by Trinity Term 2006 in time for January
2007 Undergraduate Prospectus).
Contextual information
6. Subjects to consider how better to use contextual information – for review in 2005-06
and implementation in the 2006 admissions round (decision by July 2006).
7. OCAO to provide colleges with school performance data for use in the 2005
admissions round.
AS marks
8. AS marks, where provided, to be viewed with considerable discretion – for
implementation in the 2005 admissions round (decision in November 2005).
Interviews
9. Subjects implement a minimum of two independent interviews for all short-listed
candidates, to be organised in advance – for review in 2005-06 and implementation,
where possible, in the 2006 admissions round (decision by July 2006). It is likely
that the timetable will slip to the 2007 round for subjects where more radical shortlisting
is needed (see pre-interview testing, above).
10. Subjects to agree standard formats and mark scales for their interviews – for
implementation in the 2006 admissions round (or 2007 at the latest) (decision by
October 2006, or October 2007).
29
Open Offers
11. A ll subjects which do not have open offer schemes to have them in place for the 2006
admissions round .
If Model A is adopted:
12. Subjects to move to Model A for the 2008 admissions round (decision by the end of
Trinity Term 2006) .
If Model B is adopted:
13. Subjects to move to Model B as soon as is feasible and in any case in time for the
2008 admissions round (decision by the end of Trinity Term 2006) .
International applicants
14. Short-listing should be incorporated into short-listing of UK/EU candidates and done
on the same basis – for implementation in the 2007 admissions round (decision by the
end of Trinity Term 2006).
15. Interviews should be conducted, whether in Oxford or abroad, on the same basis as for
UK/EU candidates – for implementation in the 2007 admissions round (decision by
the end of Trinity Term 2006) .
Oxford Colleges Admissions Office
16. Review to be conducted in Michaelmas 2005 (decision already made by OCAO
trustees) .
30
Annexe E
-
Matrix of subjects’ admissions procedures
Data sharing Assessment Short-listing Interviews Final decisions
Subject
Excel/Access/
Word
Web-based
Pre-interview
tests
School
performance
data
Central
ranking
Subject
decisions
Redistribution
Smart second
choice
Two
independent
interviews
Interviews on a
single day
Blind
interviews
Central
ranking
Subject
decisions
Open offers
Archaeology &
Anthropology
1
Biochemistry
Biological
Sciences
Chemistry
Classical
Archaeology &
Ancient History
Classics
1
Archaeology &
Anthropology: Some colleges collectively make open offers but not all colleges participate.
31
Data sharing Assessment Short-listing Interviews Final decisions
Subject
Excel/Access/
Word
Web-based
Pre-interview
tests
School
performance
data
Central
ranking
Subject
decisions
Redistribution
Smart second
choice
Two
independent
interviews
Interviews on a
single day
Blind
interviews
Central
ranking
Subject
decisions
Open offers
Classics Joint
Schools
2
3
Earth Sciences
Economics &
Management
4
Engineering
Science &
Joint
Schools
English &
Joint
Schools
2
Classics &
English: Ranking is restricted to the written tests. Ranking does not apply to all applicants as different applicants take different tests, depending on their level of Latin
and Greek. The ranking is not regarded as an ‘
overall score’.
3
Classics &
Modern Languages: An ‘
overall score’ is used, as computed in Modern Languages.
4
Economics &
Management: Will use redistribution from next year.
32
Data sharing Assessment Short-listing Interviews Final decisions
Subject
Excel/Access/
Word
Web-based
Pre-interview
tests
School
performance
data
Central
ranking
Subject
decisions
Redistribution
Smart second
choice
Two
independent
interviews
Interviews on a
single day
Blind
interviews
Central
ranking
Subject
decisions
Open offers
Experimental
Psychology &
PPP
Fine Art
Geography
History &
Joint
Schools
History of Art
5
Human Sciences
Law &
Law/LSE
Materials Science
&
MEM
5
History of Art: Does not centrally rank candidates at the start of the final decision process; uses ranking only at final stages, to decide where to cut off offers.
33
Data sharing Assessment Short-listing Interviews Final decisions
Subject
Excel/Access/
Word
Web-based
Pre-interview
tests
School
performance
data
Central
ranking
Subject
decisions
Redistribution
Smart second
choice
Two
independent
interviews
Interviews on a
single day
Blind
interviews
Central
ranking
Subject
decisions
Open offers
Mathematics &
Joint Schools,
Computer
Science
Medicine
Modern
Languages &
Joint Schools
6
Music
Oriental Studies
&
EMEL
PPE
6
Modern Languages &
Joint Schools: It sometimes uses open offers, but this is a
college, not subject, decision.
34
Data sharing Assessment Short-listing Interviews Final decisions
Subject
Excel/Access/
Word
Web-based
Pre-interview
tests
School
performance
data
Central
ranking
Subject
decisions
Redistribution
Smart second
choice
Two
independent
interviews
Interviews on a
single day
Blind
interviews
Central
ranking
Subject
decisions
Open offers
Physics &
Joint
School
7
Physiological
Sciences
Theology &
Joint
School
NOTES TO THE TABLE
In general, a
blank field indicates that a
subject does not include a
particular feature in its admissions process. A
tick indicates that a
subject does
use that feature. For some fields, more than one tick is used, to denote the fact that there are ‘
degrees’ of adoption of a
particular practice. The
meaning of these ticks (
and their absence) is defined below, in italics.
Data sharing
The table assumes that subject tutors share information through subject meetings, telephone and email correspondence. It therefore seeks to
record whether subjects use additional means of sharing data such as an Excel spreadsheet or Access database maintained by the Organising
Secretary, or a
web-based admissions system that allows access to all tutors involved in the admissions process.
7
Physics: The mark from the mathematics test administered during the interview process is used to provide a central ranking; this ranking is used to assign second interviews and,
together with interview grades, to assign places.
35
(a) Excel/Access/Word: A
software package is used to share information. Microsoft Office is the most common software, but other packages
or bespoke databases may be in use.
A
tick (
)
indicates use of such a
system.
(
b) Web-based: Data is shared through a
secure website (
either intranet or internet).
A
tick (
)
indicates use of such a
system.
Assessment
(
a) Pre-interview tests : these are either multi-institutional, or Oxford-only.
A
tick (
)
indicates use of such a
test.
(
b) School performance data: the matrix refers to information about the average GCSE and A
Level performance of students at a
particular
school.
•
A
blank field indicates that a
subject does not use this information.
•
One tick indicates that tutors use this information to determine potential primarily in the case of borderline or open offer candidates,
or in the case of candidates who are very evenly matched in all other indicators.
•
Two ticks indicate that tutors routinely use this data as part of the overall information available about each candidate.
Short-listing
The matrix assumes a
minimal level of information sharing during short-listing for all subjects. It assumes, therefore, that in the absence of more
extensive co-operative processes, second and third choice colleges (
or all colleges within a
Group or offering a
subject) are able to question or
veto de-selection decisions either by email or in person at a
subject meeting. This is entirely consistent with the ‘
blank field’ definition in the
‘
subject decisions’ section below.
(
a) Central ranking: ranking refers to the process of sorting candidates into order on the basis of agreed criteria. In the context of
undergraduate admissions, this means creating a
list of candidates (
within a
particular subject) ordered by their ‘
score’. This score is used
to assign them a
‘
rank’ that indicates their position in the overall cohort of applicants. The nature of applicants’ ‘
scores’ may vary by
subject. They may consist of a
weighted combination of different assessment elements, or they could be the result of an academic
judgement of a
candidate’s overall performance across all assessment elements. ‘
Central’ ranking means that the rank order is determined
at a
subject level and that all candidates within a
subject are ranked on the basis of the same criteria.
36
Central ranking:
•
A
blank field indicates that candidates are not given an ‘
overall score’ of any kind, and that candidates are not ranked across the
entire cohort.
•
One tick indicates that candidates are given an ‘
overall score’ of some kind, but this is not used to generate a
rank order or to give
every candidate a
ranking number.
•
Two ticks indicate that all candidates within a
subject are ranked on the basis of an ‘
overall score’ of some kind.
(
b) Subject decisions: This category is used to indicate the extent to which decisions are taken collectively within a
subject. It is assumed that
as colleges are the admitting bodies, tutors in every subject are able to take independent decisions. However, in practice, the extent to
which this occurs varies, and the table aims to capture this variation.
Subject decisions:
•
A
blank field indicates that decisions are essentially made by tutors in individual colleges. This includes decisions that are
simply ‘
ratified’ (
rather than made) in a
subject meeting.
•
One tick indicates that decisions are made by tutors in colleges, but under the guidance of the Organising Secretary and
potentially other tutors as well. For instance, this could include collaborative decision-making during a
meeting, or tutors deselecting
or selecting certain bands of candidates in accordance with guidance issued by the Organising Secretary.
•
Two ticks indicate that decisions are essentially made at the faculty/department rather than college level.
Interviews
(
a) Redistribution: The process by which candidates are invited for interview by a
college other than the one they originally applied to or were
allocated. This college becomes in effect the candidate’s new college of preference. Redistribution aims to ensure that strong candidates
are not de-selected before interview simply because the college they applied to has an excellent field, is over-subscribed and cannot
interview all of its applicants. Redistribution is carried out either through the OCAO in accordance with agreed guidelines and deadlines,
or through subjects’ web-based systems or Organising Secretaries.
A
tick (
)
indicates that redistribution is used (
regardless of how it operates in practice).
(
b) ‘
Smart’ second choice college allocation: Any process that allocates candidates to lower choice colleges not purely on a
numerical basis,
but also on the basis of an applicant’s quality and a
college’s ‘
need’ (
which is determined by the number of high quality applicants that it
37
has). Such a
process aims to ensure that strong applicants who may not be offered a
place by their first choice college are seen by colleges
that are most likely to be able to offer them a
place.
A
tick (
)
indicates that some kind of ‘
smart’ process is used to allocate to lower choice colleges.
(
c) Two independent interviews: Interviews by different panels that do not represent the same college. They are not therefore interviews within
the same college, even if they are conducted by different tutors. For example, independent interviews can be interviews at two colleges,
interviews at a
college and at the faculty/department, or two mixed-college interview panels. As at least some candidates are given such
interviews in all subjects, the aim of the table is to identify subjects where this is done for all short-listed applicants.
A
tick (
)
indicates that all short-listed candidates are given such interviews.
(
d) Interviews over the course of a
single day: Interviews that allow at least some candidates the option of not staying the night in Oxford.
That is, they are interviews over the course of a
single morning and afternoon, not just over the course of any 24 hour period. Again, this
applies to all short-listed candidates.
A
tick (
)
indicates that all short-listed candidates have interviews scheduled over a
single morning and afternoon.
(
e) Blind interviews: Interviews conducted without tutors knowing their interviewees college preference.
A
tick (
)
indicates that a
subject uses this process.
Final decisions
(
a) Central ranking: For definition and key see ‘
Short-listing’, above.
(
b) Subject decisions: For definition and key see ‘
Short-listing’, above.
(
c) Open offer: This is an offer of a
place, without a
college specification. These offers are determined at a
subject rather than at a
college
level. A
number of subjects currently run open offer schemes under which a
certain number of candidates are made offers but are told that
their college will only be determined once examination results are known in August. Candidates are then allocated to colleges once it is
known which colleges have places available through having ‘
lost’ candidates (
either through applicants rejecting their offers or through
failing to meet their offer conditions).
A
tick (
)
indicates that a
subject uses an open offer scheme.
Post a Comment