A Brief Examination of Degree Classifications

Grade Points:

A or A+ 5.0
A- 4.5
B+ 4.0
B 3.5
B- 3.0
C+ 2.5
C 2.0
D+ 1.5
D 1.0
F 0.0


From given figures:

3.5% First Class* => 4.50 & Honours Project A-
12% Second (Upper) Class* 4.00 - 4.49
65% Second (Lower) Class** 3.50 - 3.99
19.5% Third Class & Other** 3.20 - 3.49 (3rd Class)

* From previously quoted PDF. These are somewhat conservative estimates, but assuming the Deputy Vice-Chancellor knew what he was talking about, these should be pretty close to the actual. The higher end of estimates should be about 5% for First Class and 15% for Second (Upper). N.B. Quotes are from the Engineering faculty.

** Estimated figures, relatively low guess for third class honours based on http://en.wikipedia.org/wiki/British_undergraduate_degree_classification and the assertion that the "majority" get Second (Lower). (Quote: "Few third class degrees are in fact awarded: this is because the candidate must average a Third, but avoid failing too many parts of the course... It is therefore rare for a graduating class to include more than a small handful of Thirds.)

In the NUS context, a student must have a minimum CAP of 2.0 to graduate but will be dismissed if he has CAP < 1.5 in two consecutive semesters, or CAP < 2.0 in three, i.e. a straight C student will likely not survive.

I regret that more justification for the percentages cannot be given, due to the lack of information in the public domain on NUS's retention rates/cohort graduation rates. Anyone with reliable information in these respects and willing to share will be greatly appreciated.

My analysis of available statistics from http://www.nus.edu.sg/registrar/statistics.htm did not yield useful data, other than the fact that the percentage of students who failed to obtain a degree should be pretty low. This is assuming students take an average of four years to complete their course, coupled with the fact that the number of first degree grads corresponds very closely to a quarter of the total undergrad enrolment each year. Example data in Appendix A.



OBSERVATION I

While the project A- requirement for First Class Honours may be a limiting factor on the number awarded each year (something like the 9:45 timing for 2.4 for the IPPT Gold), it can still be deduced that only about 15% - 20% of the cohort graduate with a CAP of 4.00 or higher.

Thus, only 15% - 20% of students have a B+ or higher *average*.



OBSERVATION II

While the CAP range for Second (Lower) Class Honours is only 0.5 (3.50 to 3.99), by all accounts it contains at least half of each cohort, and probably more. This suggests that the corresponding average grade (B) is very common.

In fact, "B" is likely to be given to about half the candidates sitting for an exam, as otherwise it would be difficult to keep so many students in that particular Honours band.

This distribution should hold even for higher-level modules, where grades are rumoured to be generally lower, since it is such a large proportion of the whole cohort.



OBSERVATION III

While no official statistics on the cohort graduation rate could be found, it should be well above 90% (compare with say, average US universities) from the figures in Appendix A. Also, NUS is part of Universitas 21 Global, where the student retention rate is over 95%. Presumably the offline version has similar rates. http://www.u21global.edu.sg/portal/corporate/html/press-2005-09-07.html

This indicates that the lower failing grades (C and below) should be rather difficult to obtain, requiring wholly inadequate understanding or damn powerful competition.



BELL CURVE HYPOTHESIS

http://en.wikipedia.org/wiki/Grading_in_the_United_States
Informal conversation with a tutor revealed that modules with enrolment above a certain number (30 or 50 if i recall rightly) have to be marked on a Bell (normal) curve.

http://en.wikipedia.org/wiki/Bell_curve_grading
Very nice graphic of the curve here. Without more concrete data, it seems reasonable that A grades are given to scores better than one standard deviation from the average (~ top 16%). Similarly, B grades are probably given to scores within one standard deviation, and C and below grades to scores worse than one standard deviation. This would produce final graduation results close to that shown above.

It is said to be generally easier to obtain better grades in lower level modules, but the average module level a student takes in his university career should be balanced. Therefore, any shifts in the grade curve in lower levelled modules have to be countered but shifts in the opposite direction when higher levels are reached.

Tellingly, New Paper (25 Dec) had an article on "Confessions of a University Lecturer" (Page 26) who says "almost all my students get Bs for their exams as I'm a lenient marker." She also earns about "$2500/mth" for four hours work (2 hours lecture, 2 hours tutorial) a week, but that's not the point :P

The point is that perhaps lecturers do have some influence on the final grade, which may indeed involve much more complicated calculations than a simple curve. But until one of us becomes a lecturer I guess we'll never know for sure.



APPENDIX A

Total Undergrad Enrolment

2004/05 - 22751
2003/04 - 23092
2002/03 - 23137
2001/02 - 23086

First Degree Grads

2004/05 - 5796
2003/04 - 6067
2002/03 - 6201
2001/02 - 6433