Listening Comprehension: Construct Validity and Trait Characteristics*

Abstract

It has been customary among both language teachers and testers to regard listening as a separate skill in language proficiency. However, the evidence for the existence of listening comprehension as a separate trait is contradictory. This paper reviews the conflictingevidence, and presents two studies that use the multitrait-multimethod (MTMM) methodology to examine the construct validity of the listening trait. These were carried out in Japan using college-level students of English (N=220 and 353). Both studies had two traits, listening and reading comprehension, with three and four methods, respectively. One study shows no significant trait effect, whereas the other shows a strong trait effect; indicating that there is a separate listening trait, but that this is not necessarily operationalized by oral input alone. This necessitates recognition of two types of listening test: (1) orally presented tests of general language comprehension, and (2) tests of the listeningtrait proper. The tests used in the two studies are compared and some tentative suggestions are made regarding those variables that may account for the separate nature of listening comprehension.