Related Weblinks

What do Items Really Test?

Units A3, B3, and C3 address the question: what is really assessed in our language tests?

Effect-driven testing suggests that if you think about test impact from the very beginning of your test developments, the "best results will follow" (to quote Helen Parkhurst, 1922, Education on the Dalton Plan, New York – E.P. Dutton, pg 225). We believe this to be true regardless of the particular model of language that supports a given test. That said, in our book we explore in somewhat great detail one particularly influential model of language ability, a model you could have in mind as an effect-driven test is built. This is the famous Canale and Swain model, presented in two publications (1980 and 1983). Alla Anisimova has a general summary of the key features of the model at a site maintained by TESOL/Greece.

Studying the Canale and Swain model means reviewing and strengthening knowledge of linguistics. Their model makes ample use of terms and concepts from that field. Here are some websites that are helpful:

• www.wikipedia.org (a user-created encyclopedia; just enter a particular linguistics term to begin exploring its meaning and to uncover related links, such as 'sociolinguistics' or 'grammar', both terms used in the Canale and Swain model.)
• Nicole McBride has produced a nice overview of teaching linguistics using the web, which – in turn – is a good way to review the entire field.
• The Linguist List is a worldwide forum for discussing matters related to linguistics, in all guises.
• You can also learn a lot about linguistics – and review your knowledge – by study of the websites of university linguistics departments. The organization of these departments, the reports of faculty activities, the announcement of conferences all illustrate how this science is structured and is evolving. Here are a few university departments from which such a search can start: Stanford, the University of Edinburgh, the University of Hong Kong, and the University of Pennsylvania (which claims to be the oldest linguistics department in the USA).