CHAPTER 7. Speaking skills 113 directly associated with di erent aspects of language performance: a er one year, the strong CLT group outperformed the weak CLT group on language accuracy measures, but this di erence disappeared in the second year. From the very start, the strong CLT group was more e ective in terms of oral complexity and uency measures. In another, three-year longitudinal classroom study (with participants aged 12-15), comparing an SB and a DUB program, Rousse-Malpat et al. (2022) conducted a large cohort study speci cally targeting the development of oral and writing pro ciency of French as a foreign language at several Dutch secondary schools. e DUB program was found to be more e ective than its SB counterpart in both speaking and writing. e studies constitute longitudinal (2-3 years) investigations, use free response measures, and are conducted in existing classrooms, making them reliable and highly suitable in comparing the instructional e ectiveness of di erent teaching programs. As far as we know, there has not been a study so far to compare speaking skills in SB and DUB learners a er the full six years of pre-university secondary language instruction. Based on the ndings by Rousse-Malpat et al. (2022) and on the fact that DUB approaches focus on developing speaking skills, it is to be expected that DUBtaught students outperform their explicitly SB taught peers, but this prediction has never been empirically tested up until now. THIS STUDY e aim of this study is to compare oral pro ciency skills a er six years of instruction in two conditions. e rst group received a so-called “weak” CLT version using an SB teaching approach. e second group received a “strong” CLT version on the basis of a DUB approach. is study seeks to answer the following research question: How do the weak CLT (SB) and strong CLT (DUB) groups compare in terms of their L2 French oral pro ciency a er six years of instruction? Based on earlier ndings that a “strong” version of CLT is e ective in L2 productive skills as assessed in previous classroom research, we expected that the DUB learners would outperform their SB peers especially in overall oral pro ciency as well as in the listening comprehension sub-component of the SOPA test, but it will would be interesting to see if there are di erences in the sub-components as, for example, vocabulary and grammar were more focused on in the SB method than uency and listening comprehension.