Abstract

This chapter focuses on the development of a listening test that will be used for placing students in intensive English classes or exempting them from further English language coursework. The test uses excerpts from actual university lectures and a short answer format to test more directly the listening skills students need to navigate in a U.S. university classroom. The chapter has three main sections: listening comprehension and lecture discourse theoretical considerations, a description of the format of the test based on such considerations, and statistical analysis of the performance of the test. Included in the statistics is the investigation of such issues as the effect of using both technical and non-technical lectures as listening stimuli, the effect of prior knowledge of topic on test performance, and performance of different proficiency levels on different types of questions. Recommendations are made for teaching and testing based on the findings.

Introduction

There have been two competing traditions in language testing, indirect and direct testing. Indirect tests “tap ‘true’ language performance obliquely or indirectly” (Henning 1987), predicting performance in language use situations. The discourse and tasks are designed to be generally accessible with a greater emphasis on skills and microskills. The tests are less natural, more contrived and are what many people call tests of general language proficiency. The emphasis is on reliability, getting the same results with different forms, different administrations, and being able to test any population.