120 CHAPTER 7 focused on L2 use, frequency of exposure and repetition, very much in line with the stronger version of a CLT program, and the L2 was spoken almost exclusively. To avoid bias, a valid and reliable test for assessing oral pro ciency had to be used. An existing instrument, the Student Oral Pro ciency Assessment (SOPA) was chosen as an appropriate test for this study, but as the SOPA was originally intended for use with younger children, it was adapted, and more academic tasks were developed and validated. To prepare for the nal oral pro ciency test (SOPA), the students in both teaching programs received the same number of instructional hours on seven academic topics. e activities included videos, listening exercises, vocabulary exercises, writing activities and speaking activities, all of which aimed at a mastery of language items necessary to talk about the topics. In line with expectations based on Rousse-Malpat et al (2021), we found that the strong CLT approach was more e ective in oral skill development than its weaker counterpart a er six years of instruction. DUB students obtained SOPA scores that were signi cantly higher than those of SB students with a large e ect size even though there were some subtle di erences in e ect sizes on the sub-components of the SOPA test. In light of DUB theory and common assumptions of strong versions of CLT, including focus on form approaches, these ndings are not surprising. Language is rst of all learned through exposure and active use, which includes problem solving skills and parsing sentences within a natural discourse, as suggested by Shintani (2013, 2015). Within the DUB L2 teaching program, homework- and classroom activities re ect both key elements. Homework activities through controlled online learning systems provide a massive amount of exposure to the L2, leaving time for classroom activities that mostly target productive skills. In addition, the strict application of the “French only” rule during all classroom activities adds a great deal of extra L2 practice and exposure. Such a program is likely to be more e ective than an SB-oriented L2 teaching program, which more prominently focuses on meta-linguistic issues and typically makes use of the L1 as the language of instruction. Rather than emphasizing exposure to authentic language, the weak CLT approach tends to focus on language structure, accuracy and reading and listening strategies, with less time being available for authentic and meaningful L2 use. e SOPA results indicated that the DUB students performed better on all four subcomponents. However, there were subtle di erences in e ect sizes between Oral Fluency and Listening Comprehension (very large e ect sizes) on the one hand, and Vocabulary and Grammar (large e ect sizes) on the other hand. We would like to speculate on the reasons for these minor di erences. Oral uency and listening comprehension can be characterized as two more general, skill-based dimensions of oral pro ciency, and are likely to bene t more from the enormous amount of exposure to and use of the L2, as is common in a DUB approach: input frequency and repeated use lead to the entrenchment and conventionalization of FUMMs (Rousse-Malpat & Verspoor, 2018), which implies that more language is
RkJQdWJsaXNoZXIy MTk4NDMw