But why do better Cambridge colleges spend more on wine, really?
By S.V.P. | LONDON
This post has been corrected.
YESTERDAY a colleague covered some (not entirely serious) research hinting that Cambridge colleges that spent more on wine also had the best academic results (see chart above). While cheering to those of us who enjoy a glass or two, the data left this correspondent thirsty for more analysis. If you discount the tempting hypothesis that more boozing leads to more braininess, what may be some of the factors at play? A bit of data scraping off the internet (including—caveat emptor—from Wikipedia) and back-of-the-beermat calculations give some clues.
The original observation is that a college’s total wine consumption is correlated to its students’ top grades, generating an r-squared score of 61%. As a reminder, an r-squared figure captures how closely a dataset matches a linear relationship; an r-squared of 100% means there is a tight linear relationship between the two while one of 0% means there is no relationship. [Correction: This paragraph initially stated that an r-squared of 100% means complete correlation, a score of -100% is perfect inverse correlation, and a score of 0% means no link either way. That is true of the correlation coefficient. The r-squared term is the square of the correlation coefficient.] My own data crunching gives 63%; I think it takes academic data from different years. But surely colleges with lots of people are likely to drink more, without each student getting more sozzled. Looking at wine consumption per student, the correlation falls a few points, to 56%. That’s still a lot, actually.
Lots of things correlate to wine consumption, it avers. The most obvious one is the wealth of the college. Cambridge as a whole had an endowment in 2007 of £4.1 billion, split unevenly between its roughly 30 colleges. Students speak of “rich colleges”, such as Trinity (£892m in fixed assets) and “poor colleges”, often the new and less prestigious places with out-of-town campuses. Sure enough, there is a 64% correlation between wealth and wine budget. And because the older colleges are more prestigious and therefore richer, you also find a 61% link between the age of a college and its boozing per head. The older colleges aren’t all that much bigger, meaning that they are not only richer in absolute terms, but also have more assets per student.
A few other things spring up, not all of them necessarily revealing. Male-dominated colleges drink more (27% correlation between undergrad male population and wine budget/head). There’s a negative correlation between wine drunk and percentage of students from “maintained” backgrounds, i.e. those who didn’t go to fee-paying schools. So we know that old colleges are richer, and attract posher students.
Of course, there’s a lot the data don’t tell you. Who is drinking the wine? Not all colleges give free booze to their students; many reserve it for fellows (and some colleges have separate budgets for fellows). And what of quantities of wine? It may be Homerton is spending its measly £23 per student per year to buy copious amounts of plonk, while Pembroke uses its £237 to fill them with a few sips each of Puligny Montrachet.
And any conclusions based on the data will ultimately rest on deeply suspect premises. Big, prestigious, central colleges use their facilities to host events such as weddings and conferences, which requires them to invest in wine. That goes a long way towards explaining the link between wealth and alcohol-drinking. Looking at the data without any context, you might conclude that rich, old, posh colleges spend a lot on wine because they can afford it. A more prosaic explanation is that old, posh colleges are rich because they spend a lot on wine for cash-generating events.
