Validity: An Exploration
 Assessment in School   Systems
 What do Items Really Test?
 Evolution in Action
 To See a Test in a Grain of   Sand...
 Analyzing Items and Tasks
 Designing an Alternative   Matrix
 Administration and   Alignment
 In a Time Far Far Away...

   

Related Weblinks

Evolution in Action

Both of us have held advisory positions on various committees for the Test of English as a Foreign Language, or TOEFL. This exam began with a conference in 1961, convened at the Center for Applied Linguistics by its then-director, Charles Ferguson. The genesis and early days of TOEFL are a tale well told by Bernard Spolsky in several publications, notably his book Measured Words (Oxford University Press, 1995). The TOEFL is used widely to determine English proficiency as part of the admission decision to English-speaking universities, particularly in North America.

Recently, the TOEFL has undergone extensive changes in its design, leading to the new 'iBT' version of the exam. The test has become decidedly oriented to academic language: both how we use language at university settings as well as the particular topics we might cover. For example, one new sample task concerns summarizing the opinion of another person without giving your own point of view.

We emphasize not only effect-driven testing in our book, but also an evolutionary vision for all test development. Tests change over time.

  • What do you think is the rationale behind the implementation of the new TOEFL summary item? Such an activity is very typical in academic settings, where a professor might ask a student to paraphrase the stance of another author or opinion-giver, using as input reading or listening materials. Why is such language use valued in higher education? Why do you think this particular summarizing task evolved at the TOEFL program? What made it a salient change over the last few years, as the new exam rolled out?

In our book, we discuss test specifications, which are generative documents from which many equivalent items or tasks can be produced. Specifications – or specs – contain two basic parts: sample items/tasks, and 'guiding language', which is everything else necessary to justify and explain a given test task, including the rationale. If you review a test task without extensive guiding language at your disposal, and if you induce more guiding language, then you are engaging in reverse-engineering: the process of figuring out a test spec when all you (really or mostly) have at hand are the items and tasks. In this case, you can review the TOEFL website and get help with some of the guiding rationale behind changes to the exam, such as the implementation of this summary task.

Copyright © 2006 Taylor & Francis Group, an informa business