It is a disappointing fact that there remains a huge gulf in the university participation of pupils depending on their background or the school they went to. Yet despite what many think, the issue here is not aspiration –research suggests that around half of students in the poorest quintile of the socio-economic status distribution aspire to go to university at age 14, even though only around 13% go on to do so.

A recent report from the government’s social mobility taskforce suggests that one way to increase the number of students from poor backgrounds in Russell Group universities is to make greater use of contextual data in university admissions –

in other words the background of students should be taken into account by university admissions officers when considering their applications, giving those from less advantaged backgrounds a better chance of admission.

On the surface this seems like a good plan. There is good evidence to suggest that once they get into university, pupils from state schools perform better than those from independent schools . So why not give them an equal chance of getting in? As the report argues, there is certainly evidence that students from poor backgrounds who have the same A-level scores as their private school counterparts are less likely to get into a Russell Group university, even conditional on applying.

However, once we add in the subjects that these students study at A-level as well as the grades they achieve, the picture becomes rather more murky; social class disparities in the likelihood of an offer from a Russell Group university largely disappear – though school type disparities still hold – perhaps evidence that privately educated kids are coached into making better applications.

Meanwhile, there is plentiful evidence that students from poor backgrounds simply don’t study subjects that enable them to get into university. While the majority of pupils in selective schools study towards the Ebacc group of highly academic GCSE subjects, less than half do so in comprehensive schools, and even fewer in schools with high proportions of free-school meals pupils. Without a grounding in these important subjects, these pupils are unlikely to ever get near to attending an elite university.

It is hard to see how using contextual information would solve this fundamental problem. Even if they were accepted to a Russell Group institution based on contextual factors, students would still have to have achieved the right combination of subjects and grades to allow them to do well on the course – pupils without decent maths and physics A-levels would struggle to pass a physics degree course, for example.

So the students most likely to benefit from this are those who have gained the right A-levels in the right subjects for their chosen degree, but were rejected in favour of a candidate from a better background. For the small number of students falling into this category, contextual data would be a good way to ensure they are treated fairly.

But if we are to dramatically improve the participation chances of our poorer students, do we need to first hold to account the schools that fail to put them on the right paths?