Multi‐componential Models of L2 Listening

Abstract

This study evaluated four hypothesized L2 listening models to examine underling structures of L2 listening. These listening models represented different underlying structures of thirteen listening ability variables derived from the scores of a listening section of the TOEFL (N = 450). Model 1 posited a single factor structure model of general L2 listening; Model 2 a two‐factor structure of short and long textual lengths; Model 3 a three‐factor structure of a general language attribute and two types of comprehension; and Model 4 a four‐factor structure of two types of language knowledge and the same two comprehension factors of Model 3. Using confirmatory factor analyses with LISREL, the study compared the overall model fit indices of Chi‐Square (χ² _M) Test, Root Mean Square Error of Approximation (RMSEA), Comparative Fit Index (CFI), and Root Mean Square Residuals (RMSR) and concluded that Model 4 was the best fit model of listening. In the selected four‐factor model, two language factors represented organizational and functional knowledge, whereas the other two suggested varying degrees of comprehension complexities. The selected model empirically supported the framework of cognitive models of language ability and other theoretical descriptions of language knowledge and comprehension processes.