Performance assessment: Existing literature and directions for research

Abstract

The paper begins by distinguishing between task-based testing and performance testing. It then provides a short historical overview of the development of performance assessment. Next, a preliminary overview of the performance testing literature sketches out some of the general trends in that literature including the fact that that it already contains many (a) overviews and discussions of language performance testing, (b) many papers and books providing guidelines for developing and doing performance testing, and (c) numerous articles on the place of performance testing in language curriculum. The paper then restricts itself to looking at the most productive recent areas of research on performance testing with the goal of addressing the following key questions (that also serve as main headings in the paper): 1. What happens when actual performance tests are developed and validated? 2. What are the characteristics of performance tests? 3. What are the benefits of performance testing (why bother?)? 4. What factors are involved in performance task difficulty? 5. How should we best go about scoring performance tests (in terms of scoring criteria and rater effects)? 6. Are performance tests reliable? 7. Are performance tests valid? 8. Are there other issues about performance tests that need to be researched? The conclusions section then extends each of the main questions into a number of as yet unanswered sub-questions that might prove useful for future research on performance testing.