GENDER AND SOCIAL-INEQUALITY AT OXBRIDGE - MEASURES AND REMEDIES

The men and women who enter the nation's universities each year are, f or that year, the pool of candidates from which Oxford and Cambridge d raw undergraduates. This pool is sliced into tranches according to A-l evel score. Oxbridge's achievement in drawing from the pool is measure d by its penetration rate (P) of the AAA tranche: this is the number t aken front the AAA tranche divided by the total number in the tranche. The Oxbridge mean penetration of the AAA tranche in 1974 was 0.51, an d in 1993, 0.38. This drop in the mean value is due to growth of the n ational AAA tranche by grade inflation at A-level. The significant par ameter measuring inequality in Oxbridge admissions is the distribution of cohort penetrations around the mean. In 1974, for the AAA tranche, the extremes were men from independent schools (P = 0.84) and women f rom state schools (P = 0.30), an indication of extraordinary social an d gender inequality. In 1993 gender inequality was essentially nil, an d the extremes were men and women from independent schools (P similar to 0.50) and men and women from state schools (P similar to 0.30). The tranche model for identifying an equality distribution of cohort numb ers is based on the principle of equal cohort penetration: this implie s that a score of AAA, for example, from a comprehensive school, is as valid for entry to Oxbridge as a score of AAA from an independent sch ool. The equality target numbers for Oxbridge are calculated for each A-level tranche and for each cohort, by multiplying the national cohor t numbers in the tranche by the mean tranche penetration. These number s are then summed over each tranche to yield the equality target numbe rs for each cohort. The result of this computation is an equality dist ribution of numbers with the same overall total and the same mean A-le vel score as the status quo distribution of numbers. The percent chang e for each cohort from the status quo distribution to achieve equality is then known. In 1974 the only significant changes required were a c ontraction of similar to 20% in the independent male cohort numbers to be transferred to state school females. In 1993 the contractions requ ired were of order 8% in both independent school cohorts to be transfe rred to the state cohorts. The penetration rate is equal to the produc t of the application rate and the acceptance rate as defined by Robbin s. The available data permit a precise calculation of penetration rate but not application rate or acceptance rate. Two steps are proposed t o kick-start a rapid increase in the state cohort penetration: (1) the elimination of the interview; (2) the identification of a target for the state entry based on the tranche model to be achieved within a spe cified period of time.