The anonymity of candidates is not sufficiently protected by the substitution of numbers for names, says Roger Watson

Anonymous marking has become almost universal in ÖйúAƬ in the past ten years. As a result, allegations of discrimination on the grounds of race or gender in assessed work have become largely a thing of the past.

I do not wish to raise any doubts about the principles underlying anonymous marking. Rather, I want to reflect on my experience of examining in several institutions, which has pointed up serious problems with the systems being run in many of them. It does not seem that systems for anonymous marking are being reviewed.

The most common anonymous marking system is one which uses unique numbers to label examination scripts and course papers. Frequently the students' matriculation or registration numbers are used and these may also be used to return marks via noticeboards (also helping to spare the blushes of low achievers). However, lists of matriculation numbers and names are frequently held by several people, including course administrators and personal tutors. As a result, this system of anonymous marking is little better than the old system of named marking, under which conscientious lecturers would turn back the first page of an examination script or assignment, claiming not to look at the names on the scripts until the exercise of marking was complete.

ÖйúAƬ

Voluntary anonymity is hopeless and a system using widely published student identification numbers is no better than a voluntary system. I have seen enough evidence of examiners in several institutions priding themselves on their ability to subvert the anonymous system and to identify the students they are examining. Normally these are the same people who fatuously claim that they are quite capable of marking in an unbiased fashion, even if they know the identity of the students.

An obvious improvement on the use of registration numbers is the additional allocation of another unique identifier to be used only for assessment purposes. These numbers should never be attached to student names for the benefit of examiners. However, such a system would require additional administration, and well briefed non-academic staff to manage it. Its best feature is that examination scripts and course work would reach markers with no names attached. But there are some, potentially serious, problems. Within one academic year students are liable to have several assessed pieces of work. In the preparation of interim lists of results from course papers or class examinations, any outlying students can be easily identified by anyone, such as a personal tutor, who necessarily holds lists of names. It is at the point where students get vital feedback and any follow-up from individual members of staff that this system is most vulnerable and it is the students most easily identifiable - ie very high or very low outliers - who can be spotted. Identification per se as a result of a very poor performance on one assessment is not really the issue; the problem arises at subsequent assessments and it is ironic that anonymity really offers no protection.

ÖйúAƬ

Registration numbers normally follow students throughout their institutional career and, in my experience, where examination numbers are used these do likewise. Provided that names are never attached to numbers then there is no problem. However, it is common custom at some point during the examination board procedure to reveal the names of students. This is normally done by the chair of the board who has the privilege of holding the master list of names, registration and examination numbers. Unscrupulous examiners, at the point of revelation, have been seen frantically scribbling down names against numbers.

But, as decisions often remain to be made about borderline cases and conflicts remain to be resolved between examiners, it is surely better practice to maintain anonymity as long as possible. Whenever names are attached to results it is open for individual examiners to influence outcomes for the wrong reasons. Although I can almost hear the moans of exasperation from colleagues, there is absolutely no reason why names should ever be revealed at an examination board.

Designing a foolproof anonymous examination system is onerous and the more foolproof the safeguards the more unwieldy it becomes. Nevertheless, if we are truly striving for anonymity then more hard work is required. After all, while there are unscrupulous examiners there are also unscrupulous students. What is to stop an arrangement whereby an examination number is revealed to examiners by a student in order to impress upon them the student's identity and his/her desire for higher marks?

Clearly, and I have yet to encounter this in all its glory, the only truly anonymous system of marking is one where a unique number is ascribed to the student on each occasion that an examination is taken or a piece of course work is submitted. This could most easily be handled by an institutional or departmental office where students submit work fully labelled with a detachable page. In the process of handing the papers over to the examiners, a new number would be assigned to each piece of work.

ÖйúAƬ

In this way the examiners would never know whose work they were looking at and would not, even anonymously, be able to compare a piece of work in front of them with previous submissions by any student.

No single system of anonymous marking will obviate the difficulties inherent in teaching small classes, negotiating assignment topics and late submissions - all of which compromise anonymity. There is also the thorny issue of the viva voce, especially where this sits alongside traditionally marked work - as it often does in master's courses.

Nevertheless, these situations should not be seen as the thin end of the wedge; rather, they should be seen as unusual circumstances where additional steps such as double marking or special indications to external examiners may be introduced to offer additional protection to students.

Anonymous marking should not become a ritual, it should be a procedure. Rituals are rarely open to question but procedures require constant improvement. Anonymous marking, in my experience, is a ritual which is waiting to become a procedure.

ÖйúAƬ

Roger Watson is a seniorlecturer in the department of nursing studies at the University of Edinburgh.