The academic efficiency and social justice of entry procedures at Oxfo
rd and Cambridge Universities are examined over the past quarter of a
century. For each major subject the mean A-level scores of males and f
emales entering from state and independent schools are compared with m
ean final examination scores in the major subjects. In any comparison
of state and independent cohorts of the same gender, within the bounds
of normal statistical fluctuation, the difference in A-level scope is
a good predictor of the difference in finals score. For example, when
between state men and independent men the difference in A-level score
is zero, the difference between mean finals score is zero also. The o
rigin of female under-achievement is examined In most subjects there i
s pronounced gender inequality due to the following chain of circumsta
nces: (1) to break-even in finals women require at entry better grades
at Advanced Level than men; (2) women used to have much the better A-
levels and so, in finals a quarter of a century ago, they matched and
even-in some subjects-surpassed the men; (3) the A-levels of women ent
ering Oxford and Cambridge Universities fell off during the 1970s; (4)
today female A-level scores are slightly worse than male A-level scor
es, and so female finals scores are much worse, in most subjects, than
male finals scares. The concept of an ideal subject is defined; this
is a subject in which zero difference in A-level score between male an
d female yields zero difference in finals score. Law at Cambridge and
chemistry at Oxford are ideal subjects. Ideal subjects are rare at Oxb
ridge: most subjects exhibit a significant male lead in finals when ma
le and female have equal A-level scores. The most nan-ideal subject at
Oxford is mathematics, in which zero difference in A-level score betw
een males and females yields a male lead in finals score of 13%: at Ox
ford the other nan-ideal subjects are physics (male lead at equal A-le
vels 11%), philosophy, politics and economics (9%), history (8%), mode
rn languages (8%) and English (5%). An ideal subject is a paradigm whi
ch requires even-handedness between male and female cohorts in the fol
lowing parameters: (1) efficiency of course selection from school; (2)
efficiency of teaching; (3) efficiency of finals assessment; (4) late
nt ability. A pronounced relative decline in the A-level scores of gir
ls educated in state maintained schools entering English and Welsh uni
versities occurred in the 1970s; it is attributed to the reform of the
state school system, particularly the growth in mixed-sex comprehensi
ve schools and the decline in the number of female single-sex grammar
schools. A peculiar aspect of the admissions filters at both Oxford an
d Cambridge ensures that state-school educated men gaining entry do so
with A-level scores markedly superior to those of the other three coh
orts.