Summary

The chapter began by defining what a test item is and by providing some of the technical vocabulary that has developed for naming the parts of selected-response and constructed-response items. Next, general guidelines (organized around Grice’s maxims) were provided for avoiding linguistic and format confoundings in writing items. The next section of the chapter covered the advantages and disadvantages of each item format and guidelines for writing effective items in that format for the following item formats: selected-response tests (binary-choice, multiple-choice, and matching formats), constructed-response tests (fill-in, short-answer, and performance formats), and personal-response assessments (conference, portfolio, and self-assessment formats) in considerable detail (as summarized in Tables 3.3, 3.4, and 3.5).

Next the use of test and item specifications for Criterion-referenced test development were discussed. Test specifications included overall test descriptors and specific test descriptors, while item specifications included a general description, sample item, prompt attributes, response attributes, and specification supplement.

Finally, item quality and content analysis (including content congruence and content applicability) were explored. These quality and content issues were not approached in abstract terms, but rather in common sense and political terms, which related to the very real problems of test use in actual language programs or research agendas.