Now that we’ve righted those faux schooling percentages – that is, the ranges of school-attendance-reporting, whole numbers posing as fractions of 1 – of the achievers enrolled in the Elite Britain study’s spreadsheet –a few additional nuts and bolts need to be tightened in turn (and it goes without saying that the rectifications described in last week’s post work for the Profession by school type sheet in the workbook, too).
First, since we’ve moved past the percent misrepresentations, the field-heading (%) indicators dotting columns E through H need to go. Second, I’m not crazy about the Percentage known field; its informational contribution is trifling, because if necessary those percentages could be derived (Population known/Population size, after all), and a little design parsimony won’t impoverish the narrative.
One noteworthy stat suggests itself: the proportion of university attendees who walked through the ivied doors of Russell Group institutions. These “leading” schools – 24 in number (out of a couple of hundred in the country all told; how the global count tallies depends on designations of institutes and the like) – seemed to have commanded the lion’s share of elites’ tuitions; by simply dividing the summed Russell Group by Any University fields we arrive at 67.55% of all those in the survey who plied some manner of higher education, correlative evidence of an “elite” clustering. On the other hand, of course, correlations don’t prove a relation of necessity between A and B. In theory, at least, a modestly-backgrounded young person could rightfully earn a place at an estimable school, an attainment which by itself would not serve to clinch the “elite” thesis. Speaking in the abstract, a duly deserved seat somewhere in the Russell Group would simply and only validate the achiever’s just desert. An equitably framed reward system that succeeds in promoting the best and brightest and assembling these in better schools (however understood) could be said to be doing nothing but its job; but of course, our data can’t by themselves point to that ideal. Sociological robustness would call for research traveling back to the square ones – the economic class inceptions of the 4,200 biographies numbered here, before untrammelled merit could be properly ushered into the equation.
Of course, a purely methodological question jabs at all of the above: the disproportions of individual profession counts yanking that 67.55% in…proportion to their numbers. Residents of the house of Lords and Parliament contribute 31% of the survey complement, but in fact their Russell Group/Any university figure of 70% doesn’t badly skew the overall ratio. And if instead we divide all Russell Groupers by the entire Population known, the aggregate 55.85% (that’s sans the House numbers) compares most comparably with the Houses’ 57.73%; not much difference there at all.
And a final consideration might qualify as something of a non-finding, notable precisely in virtue of its apparent uneventfulness. As we have them, the data don’t seem, to fall back upon a perfectly dreadful modifier, particularly actionable. If anything, they resemble a pivot table, fait accompli – broken out, kind of, into their fields and items (and that’s a nasty equivocation to be sure) in a kind of done-deal finality. As such one could wonder where the data could be taken next, and I’m not quite prepared to signal a direction. Is there such a thing as a spreadsheet that has nothing, or nothing else, to teach us? Maybe, but only maybe, and the matter isn’t wholly captive to a smallness in the numbers that might stunt the analysis. One of the issues here asks after the prospects for aggregation that the data hold out, as the presently stand. One could, for example, perhaps dispatch a pivot table to group the professions and the associated numbers by their respective sizes, but I’m not sure what profit would accumulate thereby. Do university percentages respond to sector largeness or smallness? You’d have to look and hard before you’d put your investigative standing behind anything other than the null hypothesis here.
What might work is a coding scheme that would align sectors by some shared nature, and let the school numbers fan out along accordingly; but that plan too would have to be preceded by some deep thinking.
So here, then, is one of the larger questions: take a pivot table, the product of some concerted, reasonably well-devised intention, and treat it as an initial data set. Where do you take it next?
I think that’s a fair question – at least a reasonably well-devised one.