6 Assessing productive and interactive skills Personal refl ection

TC

---

144
 Part 
II
straightforward to use and interpret as they involve a series of yes/no 
decisions. Their simplicity makes them particularly suitable for use in 
peer- and self-assessment where learners assess each other’s 
performance or review their own work. Learner understanding of 
success criteria can be increased if learners are included in the design 
process. They can discuss model responses and identify features that 
seem to characterise good performance. These features may then be 
brought together to form a checklist (see 
Table 6.2
). Completed 
checklists can provide immediate diagnostic feedback to the assessee, 
showing areas of success and areas in need of improvement.
The simplicity of checklists can also be considered a weakness. 
Successful communication is not readily reducible to yes/no choices 
and aggregating the checks to arrive at a score may, like error counting, 
misrepresent the relative communicative effectiveness of assessees.
Rating scales
The most widely adopted strategy for scoring performance on 
assessments of writing or speaking is to use rating scales. These are 
also known by various other names such as scoring rubrics and 
marking schemes.
Rating scales combine the advantage of checklists – defi ning aspects 
of performance to be attended to – with the advantages of impression 
scoring – scores that refl ect different degrees of success. Checklists can 
easily be converted to simple rating scales by replacing each checkbox 
with a numeric scale (see 
Tables 6.2
and 
6.3
). They more usually 
consist of graded descriptions intended to characterise different levels 
of performance. The descriptions guide the rater in making a decision 
about which level described on the scale best matches the sample of 
language being assessed (scale levels are sometimes also referred to as 
grades, bands or points). The development of a scale and the 
descriptors 
used to defi ne them is of critical importance for the validity of the 
assessment. As McNamara (1996) noted, the scale used in assessing 
performance tasks represents, either implicitly or explicitly, the 
theoretical foundation upon which the test is built.
Table 6.2 
A basic checklist for assessing short interview tasks
Speaks intelligibly
⌧
Opens and closes the conversation appropriately
Has suffi cient vocabulary to answer basic personal questions (age, job, family, etc.)
Produces simple personal questions accurately enough to be understood

---