The (almost really) Complete Works of Lewis Carroll

From Michaelmas 1873 to Michaelmas 1881

The unpreceded phenomenon of last Term—when more than half the Candidates for Responsions failed to obtain “testamurs”—was a matter of surprise, not to say of consternation, alike to teachers and to taught. While enquiring into the causes of this apparent catastrophe, I was led to collect and tabulate the statistics of previous Examinations, and the results seem to be of sufficient interest to be printed and circulated. The following Tables, while exonerating the Examiners of last Term from the charge, to which at first sight they appeared liable, of extraordinary severity, yet furnish evidence that on other occasions the standard employed by the Examiners has been liable to violent fluctuations, which has sometimes led to great injustice being unintentionally done to the Candidates. Considering that the Candidates outnumber the Examiners in the ratio of at least 40 to 1, it may be fairly assumed that these fluctuations are due to the varying “personal equations” of the Examiners, rather than to sudden changes in the average quality of the Candidates.

In estimating the percentage who passed in Responsions last Term, it will be necessary to take into consideration the fact that this Examination was, for the first time, preceded by an Examination “in lieu of Responsions,” held by the same Examiners. Hence we have to estimate how many Candidates there would have been, and how many of these would have passed, if the previous Examinations had not taken place. To get the total number of Candidates, we add together the numbers in the two Lists ($321+308=629$) and deduct those (81) common to both Lists, who would otherwise be counted twice, and also those (55) who appear in the first List but had not matriculated before the Responsion-List was drawn up ($629-81-55=493$): thus the total number is 493. To get the total number of those who passed, we add together those who passed in the two Lists ($176+151=327$), and deduct those (15) who passed in the first List but had not matriculated before the Responsion-List was drawn up: the remainder is 312. Thus the real percentage who passed is 63.

The Analysis begins with Michaelmas 1873, as that was the first occasion when a List was published of those “qui quaestionibus Magistrorum Scholarum in Parviso pro forma responderunt.” I have made separate Tables for the three Terms, because they seem to be governed by different laws—the Michaelmas Term yielding the highest percentage of Candidates who passed, and the Act Term the lowest. These phenomena are probably due to the number of Honour-men who began to reside in October, and to the exceptional temptations to idleness afforded by the Summer Term.

Further, I have tabulated separately the two halves into which the List is always divided for viva voce (each half being under 3 only of the 6 Examiners), because I observe that the percentage of those who pass sometimes differs widely in the two halves of the same List.

These Tables give a grand total of 4474 who have passed, out of 6952 Candidates, i. e. an average percentage of 64.

Wherever the percentage on the “whole List” for some one Examination differs much from the general percentage on the Table in which it occurs, while the two halves of the List yield nearly equal percentages, it seems fair to conclude that all 6 Examiners were using too high or too low a standard, so that probably the papers were too hard or too easy: e. g. they were probably too hard in Hilary 1874, Act 1874, and Michaelmas 1874, and too easy in Act 1877, Hilary 1878, and in the entirely abnormal Examination of Michaelmas 1879.

Wherever the percentage on the two halves of a List differ much from each other, it is probable that, in the viva voce Examination, 3 of the Examiners were using a different standard from the other 3. Instances may be noted in Hilary 1877, Act 1878, Michaelmas 1878, and Act 1879.