A trove of confidential examiner’s reports leaked to Cherwell show that Oxford’s female undergraduates are performing worse than their male classmates in several STEM subjects.

Chemistry and biochemistry

The 2017 chemistry examiner chairman wrote: “I was particularly troubled by the fact that there was an almost 10% difference between male and female attainment in the easier questions!” He stated that he had “no answers” to explain the difference but thought that “there is something clearly systematic.”

In a letter to the chemistry external examiner in 2017, the Head of Policy for taught courses noted: “There is a significant gender gap in favour of men in the proportions of candidates gaining A* in Chemistry A level, and that a number of other Russell group universities seem to have a similar (although less marked) gap.”

Two years later, she noted that the problem had persisted, suggesting that: “a gender deficit in ‘academic self-concept’” may cause an “imbalance” that is “correlated with examination performance.”

In another letter, she added: “Testing of our students at the start of the course shows a significant gender gap, which does not increase through the course, and there is a similar gap in the problem solving mark of the TSA test, which is taken at the application stage.

“It seems likely that this is a pre-existing problem rather than something created by the tutorial system.”

The department is also considering whether other confounding factors are falsely creating the appearance of a gender gap. Singaporean and eastern European students — who are usually male — tend to outperform British chemistry students. The greater proportion of men among international students accounts for half of the gender difference.

“There is no current plan to reduce the rigour of the examination process,” Gurm wrote, “the department considers it to be of [sic] the highest importance to preserve this.”

Cherwell understands that professors are collaborating with the head of Experimental Psychology on an unspecified project to improve academic self-concept.

In biochemistry, men have almost always had a greater chance of getting distinctions than women since 2010 (although women have been less likely to fail). Only 2015 did not see any difference between genders, which the examiner observed occurredw was in “the cohort with the highest proportion of males.”

Women comprised 35% of chemistry candidates and 52% of biochemistry candidates in 2017.

Computer science

Computer science’s external examiner wrote: “Of the three programs I examined there was not a single student was female [sic].” She urged the department and University to take “serious action to improve the situation.” Last week’s annual access report showed that only 9.8% of UK students admitted to computer science were women.


Men outnumbered women 4 to 1 in physics’ 2018 graduating class. The examiner noted that women were far more likely to switch out of the integrated master’s course to the BA course. Four in ten women transferred, compared to 14% of men.

“The Physics BA course is largely used as an exit route from the MPhys course,” the examiner said. “It appears that female students continue to perform noticeably worse on average than males at most levels.” Half of men received a first compared with 30% of women.


Maths had a similar gender balance to physics — 22% of candidates were women – but the results were more lopsided. The department’s examiners were “concerned to discover” that twice the proportion of male candidates received firsts compared to women, whilst women were two and a half times as likely to get a 2:2 or below.

This “very significant gender discrepancy” comes in the wake of the department’s 2017 decision to extend exam papers’ time limit from 90 minutes to 105 minutes in an attempt to help some female candidates who are “adversely affected by time pressure.”

The gap is about the same size in maths prelims. During 2016-2018, 37% of men earned distinction in contrast with 15 percent of women. Women were 35% more likely to receive a pass or lower. Graduate students in the MSc Statistical Science programme also experienced similar discrepancies. Between the same years, 54% of males earned distinctions versus 28% of females despite the gender balance being 58/42% respectively.

The examiner cautioned readers against making conclusions because “variables in addition to gender can play an important role.”

Other subjects

But a few MPLS subjects are achieving equal outcomes. Material Science’s “Confidential” report concluded: “the performance of male and female candidates was not significantly different.” Women tended to get the highest marks and men the lowest, which reflected individuals’ rankings overall.

Concern for gender disparities is not limited to MPLS either. “The better performance of male candidates is clear,” the History and Modern Languages examiner wrote. Over 2014-2018, 57% of males earned firsts compared with 36% of females. Men averaged fewer than a third of all candidates across those years. “The small number of males taking this joint school could perhaps suggest a greater degree of self-selection,” the examiner said.

There were also female attainment gaps in Music, Jurisprudence, Engineering Science, Classics, Modern Languages, Oriental Studies, Theology & Religion, and English Language and Literature. However, their examiner’s reports did not discuss their equality statistics or Cherwell was unable to acquire them.

The University and multiple women’s groups have been contacted for comment.

For Cherwell, maintaining editorial independence is vital. We are run entirely by and for students. To ensure independence, we receive no funding from the University and are reliant on obtaining other income, such as advertisements. Due to the current global situation, such sources are being limited significantly and we anticipate a tough time ahead – for us and fellow student journalists across the country.

So, if you can, please consider donating. We really appreciate any support you’re able to provide; it’ll all go towards helping with our running costs. Even if you can't support us monetarily, please consider sharing articles with friends, families, colleagues - it all helps!

Thank you!