Hard facts show that contextual admissions are right and fitting

Geoff Lucas, the general secretary of the Headmasters' and Headmistresses' Conference, wants to put an end to the "discrimination" against independent-school pupils in university admissions ("Uneven playing fields", 18 March). He is right to argue that statements about admissions must be based on evidence. However, he is wrong about what the evidence shows. He is right that there should be no discrimination against applicants on the basis of the schools they attend. But he is wrong to say that this is what I believe. He wants an evidence-based account: I hope he is ready for the results.

Admissions decisions have to be based on students' achievements and potential. Contextual admissions do not mean that individuals are not treated on a case-by-case basis. A-level grades and school backgrounds are taken into account when assessing potential, among a host of other factors: that is what universities do and have always done.

Lucas makes five claims about the inadequacies of the Higher Education Funding Council for England's "Schooling Effects" papers (2003 and 2005). Interestingly, they refer to discussions covered in the original paper, rather than being something that the University of Buckingham's Alan Smithers has "demolished". The claims have been made before and have been refuted by Hefce.

Lucas claims that:

- Schooling effects are small: this is correct but irrelevant, since no one has ever suggested that a student with 2Es has the same chance of success as one with 3As. Hefce simply argues that there is a schooling effect.

- Other factors have a far greater effect: yes, other factors have effects, but on different issues (eg, dropout), but this has nothing to do with the existence of a schooling effect.

- "A-level grades could not be disaggregated from the A-level points scored": the first Hefce paper explicitly describes the limitations of the data, including that actual grades were not available. There's no reason to believe that having them would have made a difference to the existence of the schooling effect.

However, a study of outcomes at the University of Oxford (Zimdars, 2007) found that students from independent schools did less well than state-school pupils, all other things being equal. It concludes that "state-school applicants performed better than private-school applicants" after controlling for prior achievement. It shows that students from state schools were favoured at admissions, but given the differences in HE achievement, this "was not only justified but didn't go far enough".

- "What does 'degree success' mean?": this was explored in depth in the Hefce work. It is vital to note that Smithers did not criticise the methods used to allow for differing difficulties in gaining a good degree - he simply ignored them. Other critiques of the Hefce reports (such as McCrum et al, 2006) did so, too.

- "If...undergraduates from state-school backgrounds did slightly better at university overall, this would hardly be surprising": here Lucas paraphrases Smithers, who goes much further than the Hefce study. It merely summarises different explanations for the schooling effect. Smithers seems to say that there is less a university education can "pull out" of independent-school students. Were this true, it would provide a rationale for taking school types into account in admissions.

Lucas then cites the Schwartz report on fair admissions: it is a pity he did not read Appendix 4 of the report, "Statement on the Schooling Effects of Higher Education Achievement", which evaluated the Hefce research. It concludes it was "the most rigorous analysis available".

Lucas does not look at primary sources but relies on assertions that have been refuted. Table 7 of the 2003 report shows that state-school entrants in selective institutions with AAB got more firsts and 2:1s than those from independent schools. Hefce comments: "We did find a consistent difference between students from different school types. Before reporting this finding, we made repeated attempts to 'explain it away', all of which failed" (2006).

Then there is the other evidence. The Sutton Trust's recent report on postgraduates found that "comparing like-for-like students...those educated at independent schools were 4 per cent less likely to achieve a first- or upper second-class degree than otherwise similar students educated in state schools".

The most extensive work in this area is being undertaken by Tony Hoare at the University of Bristol. Using data from three years of entry, he has found that students from lower-performing schools do better than those from high-performing ones in final-year results by a margin that would justify admitting them with between one to two grades lower (for typical AAA offers) and three grades lower (for ABB offers). This reinforces the Zimdars study's conclusion that current contextual offers don't go far enough.

Far from the evidence not supporting my claims, it is Lucas' assertions that are unsubstantiated. But some good has come out of his suggestion that we look at the evidence. It shows that schooling effects should lead to more differentiated offers when considering students' potential to benefit from a university education.