Activity 2

Activity 7
The obvious things you would need to know are the type of degree programme these students are on, the expectations or learning outcomes of the course, their previous learning experiences, what the task was, how it fits into the course, what the marking culture is, the marking scales employed and, of course, some marking criteria. These things seem pretty obvious perhaps and the task even seems a bit absurd, yet it is not terribly different from the one our universities set incoming foreign language assistants every year, often with minimal induction and support. The chances are that if you just dived in and marked the French student's essay according to your own rough expectation of such a piece of work, your marks would be wholly inappropriate for the particular context.

The point is that assessment is not an absolute science, and so much depends on the context in which the assessment takes place. Even closer to home, the same essay set on one UK university first-year course might be set on another university's final year and the range of marks attained might be very similar. This is because, as Coleman (1996) discovered, the entry-level foreign language proficiency in certain universities exceeds the exit-level proficiency at some others. This does not mean, however, that in the respective contexts the two assessments cannot be valid instruments of assessment. They can. What matters is the existence of good criteria to ensure marking in each case is reliable and fair.

Activity 8
Clearly Letter 1 displays a number of linguistic weaknesses (tenses, adverbs, poor lexical choices, some orthographical problems). However, it employs an appropriate format, is well laid out, addresses the task, refers to the criteria for the job and does everything asked.

Letter 2, by contrast, is very accurate, and displays breadth of vocabulary and a command of complex sentence structure. However, it is badly organized, a little short, fails to address the question properly, uses inappropriate formulae and adopts the wrong register in places.

Depending on the criteria you employed and the weighting you gave to language versus content, completion of task and appropriateness of formats, these might both have come out with average to poor marks for the target level.

Activity 11
Holistic approaches are generally valid, highly practical, can ensure high reliability across markers and emphasize student strengths. While the latter may suggest some minimal sort of positive washback on teaching, feedback to students is very limited and the broad, undifferentiated categories do not provide much information to students or tutors on areas that need to be worked on, thus also limiting the transparency of the approach.

Analytic approaches are highly valid. The use of marking grids can make marking quite practical and a reasonably high marker reliability can be achieved through sample marking in teams and second- or double-marking procedures. By breaking performance up into its constituent elements, they make marking criteria explicit, allow specific, targeted feedback on weaknesses and provide tutors with key information on students' abilities in different areas, both of which ensure very positive washback on learning and teaching.

Objective marking is probably the least practical of the three owing to its time-consuming arithmetic and the need to link it to parallel analytic marking of, for example, organization and content. However, it is probably the most valid, at least on accuracy of student production, ensures very high marker reliability, is completely transparent and provides clear (if negative) feedback for students. Generally, its washback effect is positive.

Activity 12

1. This places an undue burden on imagination. Apart from diluting the main purpose, ie to assess a student's ability to write in the FL, this disadvantages the student with little imagination or the student whose mind goes blank in the exam room.
2. This favours those students with a certain amount of creativity and especially those who are able to develop arguments on almost any topic and to present them in a logical fashion. It might be argued that this is an element of the essay task, but it disadvantages those who are less creative. It is also a task most would struggle to perform in L1, let alone an FL.
3. This requires background knowledge of the topic. We are not interested in students' general knowledge but in their FL writing ability. The title might be appropriate if students were given some ideas to work with, which would ensure a slightly more level playing field.
4. This too depends on general knowledge but, worse than this, it requires in addition a knowledge of economic issues, which not all students may possess. If the topic has featured in coursework, it may be fair enough; otherwise avoid it.
5. Apart from the inauthentic nature of the task in providing spoken language in writing, this again relies a lot on creativity under pressure and is a test of script-writing ability as much as anything.

Activity 13
Criteria for a task such as this need to take account of how well the task has been performed, the relevance of the information provided and/or given, the appropriateness to the situation of the language used (lexis and register), as well as the usual criteria for oral assessment, such as accuracy of language and the quality of pronunciation and intonation. Criteria for role plays are difficult to write; often they need to include slight variations to recognize the demands of different role-play situations. Here they would have to address the distinctive challenges of the two positions students are placed in: one in which they might need to placate, explain and justify, and the other in which they might need to apply skills of persuasion.

Activity 19

Activity 20

Activity 21

Activity 24
The full text is:
Many able students are not very proficient at L1>FL translation. They discover that, although they have good understanding of the target language text, they simply cannot find the appropriate English in which to express the foreign language. The reasons for this can vary greatly, but it may simply be due to poor knowledge of L1. It seems to be particularly true of able learners who have learnt the FL naturalistically, with minimal reference to their mother tongue. In such cases, the FL is processed in ways more akin to the processing of the mother tongue.

There are several potential ambiguities here: item 1 could be a range of epithets ('excellent', 'outstanding', 'perfect', etc); one might argue for 'with' in item 3; item 5 might be 'yet', 'however', etc; while item 11 might also be 'circumstances'. This illustrates how careful one has to be in devising a cloze test. The one above would either need a comprehensive mark scheme or it would need editing to avoid the ambiguities.