The second sex

Gerard McCrum argues that women from state schools are less academically formidable than they were 25 years ago.

Why do men dominate the finals lists at both Oxford and Cambridge, scoring proportionately more than their fair share of firsts ? It is known that women did not perform worse than men in the early 1970s, before the "mixing" of the formerly single-sex colleges. One possible explanation therefore is that when formerly male colleges started to admit women in the 1970s, the result was a teaching atmosphere and institutional culture that adversely affected the academic progress of women. An alternative explanation, more amenable to an experimental test, is that women entering Oxford and Cambridge today are, compared with the men, less academically formidable than they were a quarter of a century ago.

I have tested this latter hypothesis by correlating the A-level scores of male and female students entering Oxbridge with their results in the final-year degree exams, examining the evidence subject by subject.

One aim was to ascertain whether the "value added" academically by an Oxford University undergraduate course, over and above A levels, was the same for men and women. The problem was to determine whether the decline in women's academic performance in finals could be understood in terms of a fall in female added-value during the course (possibly indicating a culture operationally biased towards men) or in terms of a fall in female A-level scores at entry - or as a mixture of the two.

The method used involved working out the difference in average A-level scores between men and women in several specific years and then comparing this with the difference in average finals scores for the same cohorts in the same years. This article outlines results for five subjects at Oxford: English; history; politics, philosophy and economics; modern languages; and chemistry, together with results for four subjects at Cambridge: maths; history; law; and social and political science.

The Universities Statistical Record supplied the A level and finals scores for all Oxford graduates for the years 1972-76, 1977-80, 1981-84 and 1989-93. The finals data was put into numerical form using a Norrington scoring system (scoring 5 for Class 1, 2.5 for undivided Class 2, 3 for Class 2.1, 2 for Class 2.2 and 1 for Class 3). For example, for the years 1972-76 the mean scores in the final honour School of English were 2.654 for men and 2.709 for women: for the years 1989-93 the mean scores were 3.180 for men, 2.972 for women. The percentage differences between these (men's score minus women's score) are -2.1 per cent for 1972/6 and +6.7 per cent for 1989/93: a positive difference signifies a male lead and a negative difference a female lead.

The scoring system for A levels was A=5, B=4, C=3 and so on. For the candidates graduating in English in 1972-76 mean scores were 11.645 for men and 12.166 for women: for the years 1989-93 the scores were 13.290 for men, and 13.182 for women. The percentage differences between the men's and women's A-level scores are -4.4 per cent for 1972-76 and +0.8 per cent for 1989-93.

If we consider the performance of men and women in English finals at Oxford, when the A levels of men and women are equal, the Norrington score of men is predicted to be 5.3 per cent greater than that of women. For equal finals scores the prediction is that women must enter Oxford with a superior A-level score.

According to this evidence the poor performance of women today in the final honour school of English can be divided into two parts.

First is what we can call a "structural factor", possibly a gender differential in Oxford's teaching, institutional culture and assessment. For example, in the years 1981-84 and 1989-93 the male lead in A-level score, although small, produced a far larger male lead in the English finals score.

Second, a falling-off in the standard of women's finals results in parallel with the diminution of their lead at A levels, from complete female dominance in finals between 1972 and 1976 to a small male lead today (1989-93).

I then applied the same analysis to the subjects of chemistry/biochemistry, history, PPE and modern languages at Oxford. The last three are comparable subjects in that men perform far better at finals than women in all three. (I combined chemistry and biochemistry to get a good female sample size -separately the two subjects have similar characteristics.) Chemistry is the least male-dominated subject at Oxford. When A-level scores are equal between men and women, the Norrington score of women in chemistry finals is predicted to be 2.1 per cent higher than that of men.

At Cambridge I examined a pair of Cambridge "macho" subjects, history and mathematics. When A-level scores are equal between men and women, the Norrington score of men in finals for these two subjects is predicted to be 11 per cent higher than that of women.

For law, however, the balance is redressed - at both universities. When A-level scores are equal between men and women entering Cambridge to read law the sexes do as well as each other at finals. Likewise at Oxford. It is known that women do well in this subject in universities across the UK.

Finally I looked at social and political science in Cambridge, which is a female-dominated subject. The data show considerable scatter but it is clear that the "value added" in SPS at Cambridge, over and above A levels, is considerably greater for women than for men.

Put simply, any departure from the "ideal" situation in which, if A-level scores are equal between men and women, then Norrington scores at finals are also equal, can be taken to be prima facie evidence of a gender differential in teaching, institutional culture or assessment. A more complex explanation involves the possibility that the scaling of academic quality at entry to Oxford by an A-level examination is different for men and women. Success in finals depends on the extent to which the student identifies with, is motivated by and relishes the chosen course, and this parameter may not be predicted by the A-level examination for men and women with equal precision.

What of the other hypotheses? Could an innate difference between the intelligence of male and female - due to different brain size, for example - cause the positive weightings towards men's finals performance in history, modern languages and PPE at Oxford and in history and mathematics at Cambridge? Or is there a discrepancy between the priorities of men and women in these subjects, men on average being more diligent and focused? If so why should chemistry at Oxford and law and SPS at Cambridge be so different?

It is far more likely that our attempts at ensuring equality of opportunity between the sexes have not been successful. Perhaps Oxford's menu of courses appeal more to the innate interest of men than women? The success of women in SPS at Cambridge is surely worth looking at.

There has been a decline in the relative A-level performance of female entrants to Oxford. One possible cause is the decline nationally in the academic quality of female university entrants from state schools as measured by A-level performance. In the early 1970s the strongest cohort nationally was women from state schools: the strongest cohort today is men from independent schools and the weakest women from state schools. It is this fact, coupled with the inherent structural problem - most honour schools at Oxford produce finals results in which men score more highly than do women who achieved equal scores with them at A level - that has brought about the current gender deficit.

The decline of the state girl can be seen in individual subjects. For example, in 1974 Oxford history finalists included 36 state women with an average A-level score of 13.0: they were the strongest cohort - the weakest, independent men, scored 11.5. In 1991 there were 38 state women finalists in Oxford history with an average A-level score of 13.2. The scores of the other cohorts in that time increased by 18 per cent (independent men) and 8 per cent (state men and independent women).

My findings are that the unexpected appearance of a gender deficit in the past 20 years is due to a weakening of the originally high academic quality of women from state schools, together with an operational bias towards men within the Oxford honour schools, other than chemistry, classics and law.

Gerard McCrum is emeritus fellow of Hertford College Oxford.