6 Assessing productive and interactive skills Personal refl ection

rating scale
, sometimes referred to as a 
scoring rubric
.
Example 6.1: Speaking task
You are a purchasing agent for a wholesale catering business which 
sells eggs to restaurants.
You need at least 240 large eggs every month, but you could buy up 
to 50 more as your business is expanding.
The current price of a dozen eggs is typically 97 cents, but your 
director has told you that you must fi nd a lower price. 
Delivery should be made within two days.
Call three suppliers of eggs. Negotiate a discount. 
Place your order with the cheapest supplier who can meet your needs.
Extended performances are inevitably less predictable than responses 
to multiple choice items and the assessment designer has much less 
control over the response. This limits the reliability of any resulting 
scores and threatens their interpretation. There are steps that designers 
can take to impose control, such as standardising the tasks. In 
Example 6.1, genuine suppliers could be replaced by recorded 
questions or by actors, each with a partly scripted role to play, so that 
every assessee faces a similar challenge. However, as with adaptations 
of input for listening and reading, attempts to control the task 
inevitably threaten authenticity. Actors may not respond in the same 
ways as real suppliers and may not respond in the same way to each 
individual assessee. Recordings can be used to standardise the input, 
but cannot be engaged in a negotiation. Even where tasks are 
standardised, there may be aspects of the specifi c task selected – the 
topic of an essay, the personality or vocal characteristics of an 
interviewer, the familiarity of the assessee with the task type – that 
affect individual assessees in different ways.
When it comes to rating the performance of the assessee, McNamara 
(1996) made a distinction between a strong and a weak sense of 
performance assessment. In the strong sense of the term, the focus is 

---