A form of assessment only slightly less contentious than self-assessment is peer assessment. Many reject it out of hand, claiming students want an authoritative verdict. Indeed, in a series of interviews with students and staff in modern languages, Schmidt and O'Dochartaigh (2001: 34-35) sought views on the role of peer assessment in the subject. Students seemed to have little understanding of the purpose or benefits of peer assessment. Most were reluctant to criticize fellow students in public and some objected to the whole idea, suggesting they wanted the view of an 'expert' on their work rather than that of a classmate.

However, there are clear advantages to be identified. When students get to grips with marking criteria and have to apply them to mark others' work, they are able to see into our minds a bit better and understand more clearly what we expect of them in terms of performance (Race, 2001a: 94). It is often said that trying to teach something is the best way to get a firm grasp of the topic. The same might be said of assessing someone's work, since in seeking to judge it, you develop and enrich your understanding of the topic, especially if you are involved in assessing several pieces on the same subject. Equally, especially in such accuracy-based subjects as languages, students can learn a lot from looking closely at work which is better than their own and thinking about why it is superior.

Peer assessment can work particularly well with class-based oral work. For example, students might be divided into pairs or small groups to perform a task (eg role play, information-gap activity, debate). While Pair or Group 1 and 2 perform the task, Student 3 or Group 3 observes the performance and uses a sheet prepared by the tutor to check if the task is being carried out properly, for example to see whether key information is mentioned or key language is being used, and to comment on their perception of each speaker's performance (pronunciation and intonation, independence, fluency).

Whole classes might also be involved in assessing (anonymously) individuals' performances in an oral presentation to the group, or else sub-groups might be asked to look at, for example, fellow students' written work and to assess it according to certain criteria. (I once did this with a post-'A' level first year group and their provisional marks were remarkably close to my own. There is in fact research evidence to support this claim for the reliability of peer marking - see Race, 2001b.) The mere process of having to check work against criteria and of comparing it informally with their own efforts, can help students to become clearer about expectations and to identify effective approaches to tasks, such as (in speaking) certain communicative strategies or (in writing) useful arguments, idioms or expressions.

In spite of this, one might well feel that peer assessment is more problematic for accuracy-based exercises in languages. Indeed in the survey mentioned earlier, staff felt peer assessment was better suited to tasks in which transferable skills such as presentation, negotiation or persuasion were involved. There was also a feeling that students' marking may need to be moderated in some way in case they are being too strict or too lenient. It was further suggested that if the marks awarded were too far removed from the tutor's own estimation, they should be altered, which rather begs the question: why bother with it in the first place? As Race (2001b: 13) notes, if tutors are too heavy-handed in moderating self- (or peer) assessment, students tend not to put their hearts into being objective. Language tutors generally felt that for peer assessment to be effective, tutors needed to:

• ensure clear goals and objectives;
• work out with students (or provide) assessment criteria;
• divide responsibilities: students assess presentational and communication skills, staff assess linguistic level and accuracy;
• average the marks awarded by the group to each piece of work (possibly eliminating the best and worst one or two marks in each case to avoid bias or extreme views);
• be prepared to invest considerable time and effort in setting up peer-assessment (Schmidt and O'Dochartaigh, 2001: 34-35).

With reference to this last point, it is popularly thought that peer assessment saves tutors time. Experience would suggest this is not always the case. It takes quite a bit of time to prepare assessment sheets to guide students and, with written work, to copy pieces of work for students to look at. It also takes quite a bit of time to prepare the group, to explain to them the reasons for doing it and to reassure them about the process. Finally, for the reasons mentioned above, if the final mark is to count towards assessment of the module, peer marks ought to be supplemented by a teacher mark. In the end, peer assessment can be just as time-consuming as traditional tutor assessment. Nevertheless, it does have clear benefits in terms of student engagement in and improved understanding of the assessment process; so, if only on occasions, it is worth including in your assessment procedures. (See also Module 9, section 9.4.3 on peer evaluation in writing.)