A Structural Equation Model and Intervention Study of Individual Differences, Willingness to Communicate, and L2 Use in an EFL Classroom

Abstract

In this study I investigated foreign language learners' Willingness to Communicate, frequency of L2 communication, and eight individual difference variables hypothesized to influence them: L2 learning anxiety, L2 learning motivation, integrativeness, international posture, ought-to L2 self, ideal L2 self, L2 linguistic self-confidence, and valuing of global English. Based on the concept of possible selves (Markus & Nurius, 1986), Dörnyei (2005) proposed the concept of the ideal L2 self: an idealized self-image involving future linguistic proficiency and professional success through mastery of an L2. In this study, Dörnyei's (2005) hypothesis that Willingness to Communicate is primarily determined by linguistic self-confidence and the ideal L2 self is tested using a structural equation model. A second purpose of this study, tested by comparing alternative structural equation models, was to confirm whether students' self-reported Willingness to Communicate best predicts foreign language use in the classroom. In addition, gender differences in L2 WTC and the ideal L2 self, and the effects of visualization and goal-setting activities on the enhancement of Willingness to Communicate were investigated using multivariate statistical techniques. A total of 662 Japanese university students participated in the study, 373 as core participants and 289 for cross validation. A model was hypothesized based on the WTC model (MacIntyre, 1994), the socioeducational model (Gardner, 1985), and the concept of the L2 Motivational Self System (Dörnyei, 2005), and tested using questionnaire data collected at the beginning of the university semester. The hypothesized model showed marginal fit to the data (CFI = .902, RMSEA = .081). The path from ideal L2 self to L2 WTC, tested for the first time in this study, was the most substantial predictor of L2 WTC in the model with a path weight of .51. It was also confirmed that self-reported estimates of WTC directly predicted observed L2 use in the classroom, while Motivation and Ideal L2 Self did not. A model specifying a direct path from WTC to L2 Use and indirect paths via WTC for Motivation and Ideal L2 Self showed good fit to the data (CFI = .962; RMSEA = .083). Regarding gender differences, female participants scored higher than males in both L2 WTC and Ideal L2 Self. Concerning whether L2 WTC can be enhanced by classroom tasks such as visualization and goal-setting, the results suggested that the visualization treatment alone was not effective in enhancing learners' L2 WTC over the non-treatment group. The increase in learners' L2 WTC was significantly greater for the When visualization and goal-setting group compared with the visualization group and the non-treatment group. The first implication of this study is that considering the strong impact of ideal L2 self on L2 WTC, there is significant potential for enhancing L2 WTC by applying motivational strategies that enhance or develop second language learners' ideal L2 self. Second, considering the importance of L2 output for developing communicative proficiency, the finding that self-reported L2 WTC predicted actual L2 use in the classroom lends additional credence to such motivational approaches. That finding also supports the validity of other studies that have relied on self-report for measures of L2 WTC. A third implication is that because females generally exhibited higher measures for L2 WTC and Ideal L2 Self, gender diversity is preferable to promote active classroom communication. Finally, for researchers and practitioners interested in designing activities to enhance learners' L2 WTC, connecting the proximal goals in the class to future distal goals (Miller & Brickman, 2004) could be an important aspect for the success of the activities).