Here are suggested answers to the Think About It items in this chapter of the Online Learning Center website.

1. The type of tests we give learners has ramifications for what instructors do in the classroom. This is called the washback effect. Give an example of how a test might have a positive washback effect. Give an example of how a test might have a negative washback effect.



Teacher’s Note: Response should be similar to the following:


• A test has a positive washback effect if its format pushes teachers to conduct class in a way that is in tune with the institution’s teaching approach and curricular goals.
• A test has a negative washback effect if it encourages teachers to teach in a way that contradicts the institution’s accepted teaching approach. For example, if the advocated teaching approach is communicative but the test is full of drills, then teachers may feel the need to do lots of drills in the classroom, taking time away from communicative activities.

2. Suppose you have been given the task of writing all the exams for your language program. Explain what you would do to make sure that your tests have inter-rater reliability and are bias free.



Teacher’s Note: Response should be similar to the following:


• To ensure inter-rater reliability, I would have at least two teachers grade the same exam to see if they come up with a similar score.
• To reduce bias in a test, I would make sure that the elicitation and test evaluation procedures are the same for all test takers. For example, for an aural comprehension task, I would provide all instructors giving this test with a tape-recorded passage rather than have each individual instructor read the passage. In this way, the voice and rate in which the passage is read will be the same for all learners.

3. Some instructors incorporate pair interactions into their courses and evaluate them as speaking events. Which type of rating scale would be appropriate to evaluate this type of speech event, a holistic or componential scale? Why?



Teacher’s Note: Response should be similar to the following:


• A componential scale is better. A componential scale allows instructors to separate different components of communicative competence and to adjust the weights of each component as necessary.
• Holistic scales give the impression that language is a unitary ability which it is not. Furthermore, this type of scale makes it easy to misassign a level of proficiency.