3.5 Marking Procedures and Moderation
1. All examinations are subject to the University’s Anonymous Marking procedure, with candidates retaining anonymity until the Institute/Departmental Examination Board. At that stage, the recommendations of Institute/Departmental Special Circumstances Panels are also considered to take account of medical or other special circumstances which have been reported by students under the University’s Special Circumstances procedure. A similar policy applies to written coursework, subject to exemptions approved by Institutes where anonymity is impractical or undesirable.
2. The expectation of the UK Quality Code (Chapter B6) is that higher education institutions will have in place transparent and fair systems for marking and moderation. The University needs to be assured that robust, effective and consistent internal moderation processes are being applied in all Departments across all Institutes. The details of these processes are likely to vary according to local circumstances and professional body requirements, but all Institutes should work to the definitions and minimum requirements set out below in applying their own internal moderation processes.
(i) Internal moderation of assessed work is the process of ensuring that assessment criteria are applied consistently by examiners, that students are being treated fairly through the assessment process, and that there is a shared understanding of the academic standards students are expected to achieve. Moderation is the process of ensuring that the marks awarded for an assessment task within a module are within reasonable limits, in the context of the criteria against which students’ work is being assessed. Note that separate assessment criteria should be in place for each different component of assessment within a module. Moderation may be limited to sampling and second marking a representative number of pieces of assessed work across the marking range from a cohort of students; or it may involve second marking the work of the whole cohort (double marking); or it may involve scaling of marks for a component of assessment.
(ii) Second marking is the process in which a second allocation of marks is given to a piece of work by a second internal examiner. This process may either be carried out blind (where the second examiner does not have access to the marks and comments of the first marker) or sighted (where the second examiner can view the marks and comments of the first marker, and adds their own).
(iii) Scaling can be applied to any assessment component mark, not just an examination mark (refer to paragraph 4.3 for procedure). There will be circumstances where failures in the assessment process will make scaling obligatory. Examples of these might be a misprinted examination paper, the interruption of an examination or, in a science laboratory, an instrumental malfunction not obvious at the time of the experiment. Such circumstances are very rare but marks may be scaled when they happen.
Scaling may also be appropriate when a particular assessment component has produced marks for the class as a whole at a different level to their marks elsewhere. Normally this is an examination which, especially in numerical subjects, may be much harder or easier than intended. This is usually detected because the overall mean mark is very different from that on other examinations. There are also cases where, because a threshold of knowledge or skill is necessary to achieve a reasonable mark, the marks of poorer candidates are more affected.
In some cases the same module causes concern annually for at least part of its assessment. In such cases the assessment practices of the module should be revised or the weighting of the contentious items reduced.
It is not expected that scaling would be used in instances of a module returning a high failure rate or unusually low marks.
(iv) Simple addition scaling: a notional percentage is added to every mark. The advantage of this is that, on the module and overall, all students’ classes have the same chance of being changed since the same number is added for everyone. The disadvantage is that, across the whole spread of marks, poorer students benefit more in percentage terms than stronger students if the mark is raised; suffer more if it is lowered.
(v) Multiplication by a factor:every mark is multiplied by the same factor. A good student has a greater chance of a class change since they will lose or gain more marks. This applies within the module and on the overall spread.
(vi) Piecewise linear interpolation: the mark achieved, in this case on a module exam, is plotted for each student against their average mark on all other examinations.
This particular example was scaled by raising fixed point marks of 0 to 0, 25 to 35, 65 to 70 and 100 to 100 and using linear interpolation between the fixed points. The aim is to get the points to lie as well-distributed as possible in relation to the diagonal line.
This is by far the most flexible method of scaling and the initial plotting provides a means of assessing deviation from a norm.
3. All assessed work submitted for the award of University credit at all levels must be subject to a process of internal moderation, consistent with the Institute or Department’s moderation policy. This applies to all modes and all levels of assessment in all delivery locations. The only exception to this will be where assessment methods are automated (i.e. the answers are machine or optically read), or in quantitative assessments in which model answers are provided to the marker. Where assessment is not in written form, every effort should be made to apply an appropriate form of moderation. Practice-based assessments must also be subject to an appropriate process of internal moderation.
4. The process by which marks for assessed work will be allocated should be clearly communicated to students via scheme and module handbooks, along with clear assessment criteria.
5. Documentary evidence that an internal moderation process has taken place, including evidence of any scaling applied and the rationale for selecting any given method, must be available for scrutiny by external examiners (see below) and other interested parties.
6. The requirement for establishing robust internal moderation procedures applies equally to collaborative schemes leading to Aberystwyth University awards. Arrangements for internal moderation, which must involve at least one member of University staff, should be agreed with the partner organisation and clearly set out in the Operations Manual.
7. Marking of coursework should be completed within an appropriate timescale in order to allow for the timely return of provisional marks and feedback for assessment components to students. Marks remain provisional until confirmed by Examination Boards at the end of each semester.
8. The internal moderation process should always be completed prior to the presentation of marks at Examination Boards. Final moderated marks must always be entered onto AStRA by the deadline provided by the Academic Quality & Records Office ahead of each Senate Examination Board.
9. Module marks that have not been confirmed by an Examination Board should not be issued to students. Assessed work for a component of a module that has been through the internal marking process may be returned to students prior to the Examination Board, on condition that students are clearly informed that the mark/grade given remains subject to confirmation.
10. All completed assessments should be first marked independently by appropriately experienced members of staff. Evidence of marking and an indication of how the marks have been allocated should be shown on all assessments.
11. For non-written forms of assessment, e.g. oral examinations, presentations, or recitals, at least two internal examiners should normally be involved in first marking the assessment and agreeing the final mark for each piece of work. The external examiner should have access to the agreed comments of the assessors, which should be provided as feedback to the student.
12. If the internal moderation process for the module is based on second marking, all assessments should then be second marked.
13. For modules employing a sampling approach to moderation, the internal moderator for the module (a member of academic staff other than the first marker/s) should then either review the marking or second mark a sample of completed assessments. Samples should:
(i) be representative of every delivery location, and every mode of study (note that this provision is aimed at delivery where there is a cohort studying at a different location, or via a different study mode, rather than at individual students taking internal or external repeat modules on a part-time basis);
(ii) be drawn from, and reflect, the full range of marks, including borderline cases and fail grades;
(iii) be of an appropriate size with respect to the size of the cohort (at least 10% and a minimum of 5);
(iv) include all components of the assessment for the module.
If there is clear evidence from the sample selected that there are serious discrepancies in the marks being awarded, the Module or Scheme Coordinator (or equivalent) should arrange for all the assignments affected (either within a specified grade band, or the whole cohort) either to be discussed by the markers to determine how to proceed or to be re-marked, as appropriate.
14. Although the University does not require blind second marking of all dissertations (or equivalent), this is considered to be good practice.
15. Internal moderation policies must be clear about the procedure to be followed in order to resolve any disagreement between first and second markers and assign a final mark for a piece of work.
16. If the internal moderation process for the module is based on scaling, the internal marker and moderator should record, as a minimum, the following for approval at the Institute/Departmental Examination Board:
(i) The method of scaling selected;
(ii) The reason for scaling;
(iii) The evidence considered in arriving at the decision;
(iv) The justification for the scaling applied;
(v) The relationship between the original marks and the proposed scaled marks;
(vi) An explanation of the consequences of scaling the marks;
(vii) An explanation of how this issue will be rectified in future years so that scaling will not be necessary.
17. Students should be provided with a single confirmed mark on their assessed work, as agreed by the internal examiners, and the feedback given on their performance in the assessed work must be consistent with the final assigned mark.
18. Where a sampling approach to internal moderation is adopted, the sample of work that is moderated may be the same sample sent to the external examiner. If the sample that is sent to the external examiner does not include any of the work that has been sampled through the internal moderation process (for example, where a random sample is selected from across the grade bands), they should be provided with additional information about the internal moderation process that has been followed. Note that the role of external examiner is to verify that internal marking and moderation are being properly conducted, not to mark work. Thus it is essential that external examiners can see evidence that the internal moderation process is working satisfactorily. For a full description of the role of the External Examiner at Aberystwyth University, please refer to chapter 5 of the Quality Code.