Exam analysis

The opinion piece below was published by Times Higher Education (THE) on 19 June 2017. To download the article plus the notes that follow it, click here

Exam standards need standardisation

The wild variations in the proportions of top grades awarded by different departments and universities must be addressed

June 19, 2017

By Bernard Rivers

It’s no fun being a UK university student at this time of year; the torture of exams has been replaced by the torture of waiting for the results. The one consolation is that the chances of being awarded a first or upper second have never been higher.

Undoubtedly, this trend is caused in part by improvements in the calibre and diligence of students. But in certain subjects at certain universities the evidence suggests grade inflation – in other words, a lowering of standards – is also taking place. This must be addressed.

Within the University of Cambridge, there is something of a stand-off over the issue. Members of what one might call the “consider the students” group point to the glut of firsts and upper seconds being awarded by other universities. They note, for instance, that 53 per cent of final-year mathematics students at University College London (where entry standards are lower than at Cambridge) were awarded a first in 2016, compared with only 33 per cent at Cambridge. They worry that employers will no longer offer jobs to students who fail to earn at least an upper second, so they want upper seconds to be as easy to earn at Cambridge as elsewhere.

Meanwhile, members of what one might call the “maintain standards” group want a Cambridge first or upper second to still mean the same thing as it meant in the past. They point out that, in 1992, 32 per cent of final-year Cambridge students in English were awarded a lower second, compared with only 1 per cent in 2016 – making a lower second now akin to a failure.

In many Cambridge faculties, the “consider the students” group is dominant. In languages, for instance, the proportion of final-year students awarded a first went up from 20 per cent in 2000 to 41 per cent in 2016. In English, the percentage awarded at least an upper second rose from 75 per cent in 2000 to 99 per cent in 2016.

But in some faculties the “maintain standards” group is dominant, and lower seconds and thirds remain relatively common. As well as mathematics (in which 19 per cent of final-year students were awarded a lower second in 2016 and 6 per cent a third), another good example is engineering (14 and 4 per cent, respectively).

Percentages of final-year Cambridge students awarded a first or upper second in 2016Chart 1

Hence, exam standards vary wildly across the university. Firsts were awarded to 58 per cent of final-year linguistics students in 2016, compared with 41 per cent of languages students, 29 per cent of natural sciences students and 24 per cent of law students. External examiners and the university’s central administration have no authority to enforce consistency and, even more remarkably, the administration has not even publicly acknowledged the problem, let alone proposed any changes.

This situation is unfair to perfectly competent Cambridge students who are awarded lower degrees than peers of similar ability in other subjects or at other universities. At first sight, the obvious solution is for all faculties to award an upper second to virtually every student who fails to earn a first, as is already done not only in English, but also in history and languages. But what would that mean for the university’s reputation? And is the logical conclusion of such a policy the absurdity that, in 20 years, everyone will get a first?

An alternative approach would be for each faculty to predetermine the percentage of students to be awarded each exam classification, after discussion with other faculties and consideration of other universities’ statistics. This wouldn’t necessarily eliminate grade inflation, but it would at least enhance fairness.

But much better would be for such an exercise to be carried out at a sector-wide level. Whatever the incentives created by national league tables, universities need to agree to continue awarding degree classifications below the upper second if that classification is to maintain its meaning and if employers are to be able to continue distinguishing between graduates. A race to the bottom will benefit no one.

Bernard Rivers is a retired economist and was a visiting fellow at the University of Cambridge in 2013-2014. Further information is available at www.BernardRivers.com/exam-analysis.

Notes (not included in the Times Higher Education version):

  • For data and data sources for the above article, click here.

  • In the article, the term “subject area” is used for what is known within Cambridge as a “Tripos.”

  • There is no doubt that at Cambridge, higher percentages of firsts and upper seconds are awarded within the arts than within the sciences. Drawing upon a scoring system developed by Martin Baxter and Peter Tompkins (who each produce annual tables regarding Cambridge in which they assign five points for every first, three for every upper second, two for every lower second and one for every third), it is possible to calculate the average number of points across all students taking exams in each subject area. Clearly, a subject area in which only firsts were awarded would have an average “Baxter score” of 5.0 and one in which only thirds were awarded would have an average score of 1.0. For final-year students in 2015 and 2016, the 13 subject areas with the highest average Baxter scores (ranging from 3.47 to 4.01) were all arts subject areas. The six science subject areas all had average Baxter scores in the bottom half, between 3.26 and 3.46. For further details, see Table 3 in the document referenced in the first Note above.

  • It is sometimes suggested that faculties awarding higher percentages of firsts are justified in doing so because those faculties are the ones that get chosen by the students with the best entry qualifications. But a study of data from www.unistats.com shows no support for this hypothesis. For instance, even though far more firsts were awarded to final-year linguistics students and languages students than to natural sciences students and law students (as discussed in the article), the percentages of linguistics, languages and law students entering Cambridge with over 240 UCAS tariff points were all between 27% and 30%, whereas the percentage of natural sciences students entering Cambridge with over 240 UCAS tariff points was 55%.

  • For a further document containing charts showing historical classing statistics for each of the 29 Cambridge subject areas, individually and in relation to each other, click here. The charts can be summarized as follows:

    Percentages of final-year Cambridge students awarded a first or upper second in 2016: arts subjects in green, science subjects in redChart 2

Percentages of final-year Cambridge students awarded a first in 2016: arts subjects in green, science subjects in redChart 3
  • To receive word of the author’s next report on this topic, send an email to bernard.rivers@gmail.com with “Send next report” in the Subject line.

  • For an earlier report, “An analysis of how often “Firsts” and other classes are assigned in Cambridge University exams (Second edition)”, published by the author in February 2016, click here.