Examination of Witnesses (Questions 960
- 979)
TUESDAY 18 JULY 2000
PROFESSOR A H HALSEY,
DR N G MCCRUM
AND PROFESSOR
DYLAN WILIAM
Mr Marsden
960. We seem to have been round the paddock
on all of these various measures and none of them, from what has
been said, offers a magic formula of something that is not culturally
biased or determined. Can anybody give a formula for how we produce
something as a measure of potential for admission that is going
to have the minimum of cultural and social bias which we are all
attempting to deal with?
(Professor Wiliam) The problem is that what you will
find generally is that reducing the bias reduces the accuracy
of the selection so you can actually have what you want by injecting
a greater element of lottery. What would be completely fair would
be for everybody to apply to Oxford and just dispense places on
a lottery. Unfortunately, you would not have any guarantee about
their talent or merit but it would be fair. The more you introduce
this idea of having some notion of merit then the more likely
you are to make a measure of merit in a specific, culturally determined
way that disadvantages others. So I think what is interesting
is the trade-off between these two things. I would say we could
do with moving more towards the lottery end in that currently
we are too focused on measures of merit that are too narrow. What
I was suggesting earlier was a fairly high hurdle but anybody
reaching that should be entered into a lottery.
961. If you had 200 places in Oxford and 400
candidates who got three necessary grade As, what you are saying
is you put them all into a lottery and the first 200 out get places?
(Professor Wiliam) Provided they reach the minimum
threshold of being able to benefit from whatever it is on offer,
yes.
Chairman
962. Does anybody else want to come in on that
one?
(Professor Halsey) I do not know whether this is—
Chairman
963. It is not compulsory, Professor Halsey!
(Professor Halsey) I do not know whether this is worth
pursuing but it sounds to me as if we once had a system of lottery
at birth, leaving women and non-Christians out, and Oxford used
to run more or less the system that you are proposing we return
to. I do not think that is a sensible way of doing it. I do think
there is something called "talent" and everyone here
knows who is the most talented among their brothers and sisters.
964. Professor Wiliam said the lottery would
only include those who had got the three As.
(Professor Halsey) He is giving away his own case
on that view because if he believes that the distribution of talent
is random in the population then he cannot possibly entertain
the idea of going through a biased selective process to produce
some new higher level from which you have a lottery.
965. Professor Wiliam, we will give you a quick
rebuttal if you want.
(Professor Wiliam) We clearly have to work in the
interim. There is no point in sending the people who have not
got the right schooling to Oxford and Cambridge because they would
not be able to benefit from it. If there were a completely culture
and schooling free measure of IQ, I would not recommend we use
it even if we could find it (which I do not think we can), so
I do not think my argument is refuted.
Mr Foster
966. A couple of points which were mentioned
earlier. You said that one of the reasons people did not apply
to Oxbridge typically was students did not think they could reach
the required admissions level. Another point that was made was
it did not differentiate between those that got straight As at
A-level and one way in which you could look at getting the most
talented in is to look at the actual marks achieved at A-level
grade, but both of those, in my view, would be solved with a post-qualification
admissions system. What are your views on that?
(Professor Halsey) I think if it is practical, and
I think it probably is, that it is better to take the admissions
after the A-level results are known, rather than second-guessing
in the admissions colleges the wisdom of the examiners which will
be known later.
(Dr McCrum) It would be ideal if A-level results were
out when admissions were decided. That would be the ideal.
(Professor Wiliam) We have a multi-million pound industry
that would fall apart if universities were forced to make the
decisions before the A-level grades were out because then they
would realise they do not need the A-level grades at all.
Chairman
967. Could you expand on that?
(Professor Wiliam) For example, what happens in Sweden
is 60 per cent of university places are reserved on the basis
of grades given to students by their school teachers. These are
not predicted grades, these are schooling grades. If that is the
most advantageous way for a student to get into the university,
that is the grade the students gets. There is a standardisation
and all students take a test and if the student's grade from the
teacher differs from the grade that is indicated by the test,
the teacher is invited to reconsider but not forced to change
his grade. Realising the weaknesses of this system, because it
is based on how hard they work in school, the Swedes also have
an aptitude test in which 40 per cent of places at Swedish universities
are reserved, and students are able to choose whichever system
gives them the best position in the competitive race. As soon
as we move away from the A-level and go to predicted grades you
might as well have a system of moderating the reliability of teachers'
grades and do without A-levels completely.
Mr Foster
968. I understand from the work that you submitted
that you believe the actual use of predicted A-level grades to
gain a place at Oxbridge or other universities does in fact add
to the bias. Could you elaborate on that?
(Dr McCrum) I do not think the problem is there, I
think the problem is with the interview, but it would definitely
be better to have the decision taken after the A-level grades
were out. There is no question about that. It is very difficult
to decide beforehand and that is part of the problem as it exists
at the moment.
969. What happens to students who miss out on
their target grades?
(Dr McCrum) They usually are rejected but then sometimes
they are given a place in any case, if the miss is not too large.
970. And what about the colleges making sure
that they have got sufficient places filled up with students?
(Dr McCrum) That is a big problem because Oxford has
a system by which so many thousand people apply to Oxford, then
they are split into 35 little colleges and each looks at its position
separately. So if they have 100 places they maybe make 110 offers
and the attrition then comes in—people do not make their
target grades—and they end up having only 80 people coming
up so then they have to very quickly make more offers. It is not
ideal at the moment.
971. What mechanism do they use for increasing
these offers?
(Dr McCrum) If someone has not missed their target
by too much they will say, Don't worry about that, you can come
up anyway."
972. Are any other criteria used?
(Dr McCrum) I do not think so. It is a black art that.
Chairman
973. A black art?
(Dr McCrum) The head of college and the admissions
tutor usually do it themselves. It all has to be done very quickly
within hours in the middle of the long vacation so the people
who are there—and it is usually the admissions tutor and
the head of the college—usually decide that.
974. I take it that it is not done by lottery?
(Dr McCrum) No.
Mr Foster
975. Do you offer individuals financial incentives
or bursaries or charitable support?
(Dr McCrum) The bursar is breathing down their necks
of course. If they only have 80 people coming up and he wants
100, obviously there is a bursarial interest there.
(Professor Wiliam) We have not talked about the reliability
of A-levels today. Certainly my own research suggests that between
a half and third of students might miss their true score by one
grade. That is not a criticism of the A-level boards; it is a
fundamental limitation of the technology of assessment. There
was a proposal many years ago to move to a 20-grade scale and
for each A-level to be given a margin of error three points either
way. This would go back to Professor Halsey's idea of giving the
marks but encourage the admissions tutors to see themselves that
there is a margin of error. If the mark is 60, in a class of 30
students, there will be one whose mark is ten marks away from
where it should have been if the test had functioned reliably.
It would not be everybody but there would be one or two in a class
of 30 for whom the marks are way out and we ought to be educating
people in the inherent unreliability of these measurements.
976. Despite the reliability of A-levels as
the only thing that we have as a mechanism to use, it would still
be, in your opinion, better to do it post qualification when you
have your A-level grades for students and then to look at where
they go for higher education? They could argue, "Well, I
have the grade. I know I am capable," particularly for those
people who have not had any experience of higher education who
need the encouragement and confidence, they have the grade and
are therefore eligible to apply?
(Professor Wiliam) In a system in which A-levels are
the only currency it is obviously better to wait until the grades
are around. My view would be that if you move towards a system
where that is not the only currency then we would have other indicators
which show that people are ready to benefit from going to university.
That would be in place and you would know about it half way through
the upper sixth year, maybe even before the end of the lower sixth.
Chairman
977. The Chairman usually gets the privilege
of sweeping up at the end. Some of the asides are as important
as the answers to the main questions and I have to take Professor
Wiliam up on one thing he said as an aside which seemed to be
a devastating critique of mixed ability teaching at one stage
when he was talking about the advantage that privately educated
children have over state school educated children. It did seem
to be a very fundamental critique of mixed ability teaching.
(Professor Wiliam) I assume you meen the compositional
effects of the presence of higher attaining students. I do not
think it is a critique of mixed ability teaching actually, because
if you go for ability grouping of any kind then you are taking
the higher achieving students away from the lower achieving students
so those lower achieving actually do less well. I happen to believe
that this country's use of ability grouping, which is greater
than any other developed country, is the main cause of the fact
that the spread of achievement in this country is greater than
any other developed country. We have a long tail of under-achievement
and we also have a long tail of over-achievement. It is a completely
symmetrical distribution as far as we can determine, but the spread
is massive. Our best is as good as the best anywhere in the world
and our tail is far too long, and I believe it is as a direct
result of our adherence to homogeneous ability grouping.
Dr Harris
978. There is one thing that I would like to
mop up, which is Dr McCrum's view about interviews. I have a fundamental
problem with your idea that because there is this acceptance by
us that the interview should go and we should rely on things like
what sport people have done and their musical instruments I would
argue that there is at least a case that that would discriminate
much more in favour of independent school people if you have ever
seen the sports facilities and participation rates at independent
schools compared to average state schools, ditto musical instruments.
Indeed, what basis do you have to suppose that the acceptance
rates would not be even more disparate if there was not the interview?
I accept medicine is probably the worst example one could choose
if you wanted to comment on the Oxford admissions process. In
medicine they have an interview and an assessment test and if
people do not do well in interview because of reticence and they
have a comprehensive background that is actually taken into account
and that can actually work in their favour, therefore. Have I
missed something?
(Dr McCrum) No, because if you have a batch of people
who look all the same academically, GCSE and A-level predictions,
then you look at other things and you say, "These people
are the same. We could use a lottery on them but instead we will
use an interview." I do not think the interview helps. So
the other way to do it is to look at the different types of people
there, A has the same academic qualification as B but has been
to a state school, and you can take that into account.
Chairman
979. We really are now coming to the last few
knockings of the innings. One thing we have heard in this Committee
time and time again is the necessity in Oxford and Cambridge to
demystify the process. Do you think that is important? That is
a short term and part of the mystification of Oxford and Cambridge
is a collegiate system, is it not? Would there not be an advantage
of having applications just to the university for the university
to determine, and get rid of the college intervention? Certainly
that that has been an opinion expressed to us in other evidence.
(Professor Halsey) There are two views on that in
Oxford. One is that college autonomy is what makes the difference
to the quality of the education which must, at least in part,
rest on the capacity of the dons in that college to choose the
people they will teach. The other view is yes, of course, it is
time it was made into a university-wide affair and it has been
made into a university-wide affair in subjects like mathematics
and it does not really matter about the quality of teaching if
it is universalised rather than collegiate. I find that very difficult
to decide for myself.
| 
