Subject Benchmarking

" broad statements which represent general expectations about standards for the award of honours degrees"

"[benchmarks] should be developed by the academic community itself"

" subject associations and professional bodies will play a role in developing benchmarks"

(QAA: Higher Quality issue 5, May 1999)



Working Group with members from

IMA + RSS + LMS + OR Society

+ HODOMS (2)

( John Erdos (Kings College) +

Stephen Ryrie (Univ. West of England) )

Chaired by Prof. Chris Robson (Leeds Univ.)

Brief: "To identify problems and stimulate discussion"

Degree Title

No. Institutions offering:


Mathematics, Mathematical Sciences, Mathematical Studies


Applied / Applicable Mathematics


"Business Mathematics" (various titles used)


Computing / Computational Mathematics


Engineering / Industrial Mathematics


Environmental Mathematics


European Mathematics


Experimental Mathematics


Financial Mathematics


Pure Mathematics


Mathematics in Society


Mathematics, Statistics & Computing




Applied Statistics


Business / Management Statistics


Economic Statistics


Mathematical Statistics


Medical / Health Statistics


Statistical Modelling


Social Statistics


Operational Research




Source: UCAS Handbook, 2000

Responses to Questionnaire

(Summary of about 25 responses)

1: Should the benchmark reflect the range of existing practice in departments, or should it set requirements which might involve changes in some degree programmes?

Consensus: The benchmark should reflect existing practice, though naturally it should require changes where existing programmes are inadequate in scope or standard.

2: Is it necessary for every student to have some appreciation of all the main areas of Mathematics, Statistics, and OR?

There was no consensus, but the majority were against this as a requirement. Several respondents were concerned about the future of degrees in pure mathematics.

Suggestion: that all courses should include all the material taught at A-level or before to a sufficient level that graduates would be easily able to teach mathematics in school.

3: Should we attempt to specify the mathematical content of the degree, given that QAA says that benchmarks are not about "listing specific knowledge"?

The majority was against this, but opinion was divided. Some would like a small core to be specified.

4: The following list is an attempt to cover the basic topics for a "Mathematics" degree. How would you change it?

The list of topics which followed was considered acceptable, but there was a lack of enthusiasm for specifying a core, consistent with the response to no. 3.

5: Could a similar list of essential topics be specified for degrees in Statistics or Operational Research?

Only 2 responses, each of which suggested a list essential topics for a statistics degree.

6: Any abilities and skills given in the benchmark will have to be demonstrably part of the student's development, but some may not lend themselves to formal assessment. We should not include items which are completely unverifiable. Is the following list realistic?

(There followed a list of subject-related skills and abilities and a list of general transferable skills)

In general, the list was seen as realistic. Some respondents expressed concern about the extra work involved in assessing a large number of skills and in using different methods of assessment.

7: If the benchmark includes project work, this will imply that all degree courses should have compulsory projects. At present many courses have optional projects in the final year; is it advisable to make it into a requirement for all students?

Almost all said it is not, although some implied that other ways should be found for giving students the same training that projects are supposed to give. Some said projects should be compulsory for statistics and OR even if they are not for mathematics.

8: According to QAA, the "typical" attainment is the main standard to be identified. Should it include some reference to content, i.e. topics which the typical student should have some knowledge of?

Opinion was divided, but there seemed a general feeling that it would be difficult to do this clearly enough to be of much use.

9: Do we want to define the "outstanding" student? QAA says that this is optional; the level they have in mind is probably higher than the traditional I class in Mathematics.

No (though not unanimously).

10: Mathematics is often studied as part of a Joint Honours or Combined Studies degree. One would expect a reduced mathematical content and level of experience from these courses. Is the reduction in breadth or in depth, and do we need a separate benchmark?

Most respondents thought the reduction was in breadth. There was a feeling that the whole situation regarding Joint Honours courses is more complicated than QAA seem to realise, also that a course such as Maths/Stats is as much a JH course as is Maths/Physics or Maths/Computing.



  1. Widespread cynicism about the value of the process and of its likely outcomes
  2. Benchmarks should broadly reflect existing practice.
  3. Majority against a specified list of topics, with a minority view otherwise. The opposite view is particularly strong in statistics.
  4. (N.B. Interesting to note that the draft Computing benchmarks do include a list of topics, which is nevertheless not intended to be prescriptive for all degrees)

  5. General agreement about inclusion of key skills.