Senate
Subject: Assessment
of students
Origin: Learning
and Teaching Committee; Programme Quality Team
As part of its work in
preparation for the forthcoming QAA institutional audit, the Audit Steering
Group (ASG) on 16 August 2007 considered the revised version of Section 6 of
the QAA Code of Practice, Assessment of
students. The ASG drew the attention
of the PQ Team to a number of issues arising from its discussions, with
proposals for action.
Areas under discussion included
double/second marking and moderation, as defined in the QAA Code. The ASG reviewed the current description of
University assessment procedures as set out in our Academic Quality Procedures
Handbook and proposed a number of amendments
(a) to address the fact that the University’s
current procedures on double marking/internal moderation only covered
examination scripts, and
(b) to bring the explanation of University procedures in
line with the QAA definitions.
The amendments agreed by the
Programme Quality Team and recommended to LTC were amended by LTC. The revised version of the procedures now
recommended to Senate by LTC is set out below.
The relevant QAA definitions
appear at the end of the document.
Academic Quality Procedures Handbook Section 3
3.6
Assessment process
Current version
(vii) Examination scripts are double-marked
by selective sampling based on the following criteria:
·
Borderline cases
·
Students showing
widely fluctuating marks within scripts
·
Other unusual
cases
·
Papers marked by
probationary staff
·
Exemplars clearly
demonstrating performance within a given class
Anonymous marking of
examination scripts is practised University-wide. Scripts are identified only
by ID number and desk number until marks are transferred to the student record.
Departments may specify that coursework for named modules will be marked
anonymously. This shall be clearly announced to students. It is expected that
staff will not take steps to break anonymity until the marking and double
marking process has been completed. There is no University requirement for
blind marking (where a second examiner marks a piece of work without seeing the
marks and comments made by the first examiner). For any module contributing to
the award of a degree assessed by coursework alone, a sample of the work must
be seen by the External Examiner. Coursework moderation requirements are under
review.
Proposed revision as amended by LTC
(vii) All examination scripts are subject to
internal moderation: i.e. a sample of scripts is scrutinised by a second internal
examiner to check that the assessment criteria have been applied consistently
(and where applicable that the marking scheme has been followed) and the
outcomes of the assessment are fair and reliable. The sample should include examples of scripts
from across the mark range and borderline cases.
(viii) All projects and dissertations are subject to
double/second marking: i.e. every piece of work is independently assessed by
more than one internal examiner. Each
marker keeps a record of all marks awarded, together with his/her rationale for
awarding each mark. In some cases,
second markers have sight of the first marker’s marks and/or
comments. Where this is not the case
(sometimes called ‘double blind marking’), marking sheets may be
used to ensure that the marks given by the first marker do not influence the
second marker’s judgement. The two
markers subsequently confer to arrive at a set of agreed marks.
(ix) Pieces of coursework, other than project
reports or dissertations, that contribute 50% or more
of the overall mark for the module, are subject to internal moderation by
selective sampling (cf examination scripts). This applies whether the module is assessed
by a mixture of written examination and coursework, or by coursework alone.
(x) In cases where there is a difference between
the marks of different internal examiners that cannot be resolved between them,
the opinion of a third internal examiner should be sought.
(xi) The involvement of external programme
assessors and external examiners in the assessment process is explained in the
‘Code of practice for external examining for taught undergraduate and
postgraduate programmes’. This
includes a requirement that they see a sample of work for modules assessed by
coursework alone.
(xii) All examination scripts are anonymously
marked: i.e. the scripts are identified only by student ID number and desk
number and the identity of students is not revealed to the markers. Departments may at their discretion determine
that coursework for named modules will also be marked anonymously: this will be
made clear to the students concerned.
Former (viii) becomes (xiii) et
seq
Relevant QAA definitions (QAA Code of Practice,
Section 6 - Assessment of Students, September
2006)
Anonymous marking: The identity of students is not revealed to markers and/or
to the assessment panel or examination board. There may be a point towards the
end of the assessment process where anonymity ends.
Double/second marking (also referred to as 'internal verification'): Student work
is independently assessed by more than one marker. Each marker normally keeps a
record of all marks awarded, together with his/her rationale for awarding each
mark. In some cases, second markers have the first marker's comments and/or marks/grades.
Where this is not the case, the use of marking sheets or similar
procedures for written work is sometimes used to ensure that the marks
given by the first marker do not influence the second marker's judgement.
Markers' notes enable discussions to take place, after initial marking, about
the reasons for individuals' decisions if there is a significant difference
between the markers' judgements. It is useful to define 'significant' in this
respect.
Moderation: A process intended to assure that an assessment outcome is fair and
reliable and that assessment criteria have been applied consistently. Forms of
moderation include:
Sampling
is most commonly used in the process of moderation (see above). It normally
involves internal or external examiners scrutinising a sample of work from a
student cohort. Sampling may be based on the desirability of checking
borderline marks of any kind, or to test that assessment criteria have been
applied consistently across the assessment of students in the cohort.