̽����Ƶ

Your review of the impending reforms to A levels in England (“Remaking the grades”, 15 August) refers to their potentially negative consequences on many universities’ efforts to widen the participation of under-represented groups, citing the significantly reduced take-up of AS levels, used by some admissions tutors as a valuable marker in this cause. However, there is another equally significant problem for those, such as my university, that use another widening participation strategy.

The use of contextual data is supported as good practice for “fair admissions” by many in the sector. The University of Bristol’s approach to this is based on identifying those markers that generate evidence of students’ educational disadvantage, derived from their schooling experiences, which dissipates while at university. So these groups, although entering Bristol with lower average A-level grades, perform similarly by their final degree assessments. This pinpoints, first, which (if any) of the several possible widening participation definitions on offer are valid bases for such a “contextual” approach, and second, how to calibrate the entry grade differentials to represent fair treatment of such applicants. But if these entry qualifications change entirely, such that the grades in the two exam systems are not directly comparable, then the contextual data calculations need to be started from scratch, leading inevitably to some “dead time” before the first cohorts with the new qualifications graduate.

How long this will take depends on the specifics of the contextual approach adopted. Bristol’s model uses entry-exit grade differentials for the three most recently graduating cohorts. (Depending on the vagaries of just one cohort’s experience is a risky alternative, especially so while a new entry qualification beds in.) Let’s take the “minimum disruption” case with students on just three-year degree programmes (assuming the university’s ability to convert results from this trio of cohorts almost immediately into a reformulated and coherent fair admissions policy). We must note that the new A levels will be phased in over two school years (with some new syllabuses taught from 2015, the remainder from 2016) and hence two Ucas cycles. As a result, we would not be able to use the results to inform the contextual data methodology for some seven Ucas cycles.

This would take us well into the next-but-one Parliament. Who knows what the national widening participation policy landscape will look like by then?

Tony Hoare
Director of widening participation research
University of Bristol