Introduction

This chapter discusses the need for test scores that are meaningful and useful, and provides guidance for reporting and interpreting test scores. The need for useful test scores can be seen from two perspectives: (1) that of the test developer who needs to report meaningful information on test results and to provide guidance for their use, and (2) that of the test user who needs to be able to interpret and use scores from tests. In the discussion, I will focus on considerations and procedures for reporting scores in the belief that an understanding of these will also facilitate the interpretation of test scores.

Raw scores – or the scores actually obtained from a language test – may not be easily interpretable to many test users, particularly users who are not familiar with the exact content of the test, the number of questions, the scoring procedures, and so forth. In addition, the different types of decisions that are made on the basis of test scores need to be based on different types of score interpretations. For these reasons, we need to find ways to report our test results in ways that will be interpretable and useful to test users. There are two major approaches to doing this, corresponding to the types of selection decisions and frames of reference discussed in Chapter 1: (1) a criterion-referenced approach, which is particularly appropriate for absolute decisions, and (2) a norm-referenced approach, which is more appropriate for relative decisions. Within each of these general approaches, several specific procedures may be used, depending on the particular test use that is intended.