It’s official! Research using Cherwell’s online archive has revealed huge differences in the number of scandals recently uncovered at different colleges. Our computer generated leaderboard placed Wadham top for notoriety, closely followed by Balliol, Brasenose and Corpus.
A sample of 5860 stories was fed to the Cherwell supercomputer to look for scandal indicating words. Articles had to be about a single college, and have appeared in the news section in the last five years. To our supercomputer, all scandals look the same, so there was equal weighting for “Oriel library shut amid sex rumours” and “Brasenose in pyjama palaver”.
Any of the following words made an article ‘scandalous’: disaster, scandal, cock-up, corrupt, illegal, incompet, lying, cheat, crime, theft, racis, sexis, bias, apolog, bribe, discrim, orgy, sued, deny, deni, steal, alleg, violen, offen, incident, drunk, drinking game, disciplinary, controv and deaned.
Some of the above are stems rather than full words (so sexis matches sexism and sexist), and college epithets (e.g., Catz) were also included in it.
There is plenty to spot in the chart. For instance three times more articles about Wadham mention scandals than do not. Their JCR President, Jahni Emmanuel, responded: “I’d say the reason you’ve come up with this ‘data’ is because lots of Cherwell journalists are from Wadham… The way you’ve categorised scandalous stories is…based on the words your own journalists have used rather than…assessing what’s actually happened… I don’t think Wadham has an issue with its image at all.”
JCR President Alex Bartram revelled in his native Balliol receiving the most articles overall. “It doesn’t surprise me that Balliol registers so highly…this reflects how central the College and the JCR are to university life. Besides, I think most people in the JCR would agree… they’d rather go to a newsworthy college than one in which no scandals ever happen.”
At Brasenose the reason for their recognition might be more sinister: “We did discover that email@example.com was subscribed to our mailing list for several years, which may explain some of the issue”, wrote JCR President James Blythe.
While the matching itself was conducted carefully, a host of weaknesses with our approach are freely conceded to at this point. First, the article filtering is crude (though unlikely to show systematic bias towards particular colleges).
A second difficulty is in trying to make an association between high numbers of articles, and high rates of actual scandals. If scandals at certain colleges are more likely to reach the newsroom, as seems plausible, then the link can’t be made.
However you wouldn’t expect that effect to undermine conclusions drawn from the ratio of the two types of article at a particular college. Magdalen stands out in the results for attracting an unusually high proportion of non-negative stories.
Meanwhile, rather tragically, St Hilda’s does not appear on the chart at all. Our sample found only five articles in as many years about the college, though it is possible some stories evaded being correctly coded. Another possibility is that they never reply to Cherwell emails, as indeed was this case when researching this article.
With just 533 college articles matching the inclusion criteria, even a small run of outrage can change the order. At St Hugh’s, the Damien Shannon “wealth selection” story boosted them eight places higher than otherwise.
The Oxford Union, were it a college, would have come top of the charts with an astronishing 31 scandals, whilst OUCA and the Bullingdon Club also registered with 12 and 2 scandals apiece.