Wim Gombert

CHAPTER 7. Speaking skills 119 With regard to the analysis of scores on the four sub-facets of oral pro ciency, some assumptions were violated, and a non-parametric test was necessary for the analysis of these scores. A visual inspection of the distributions of scores on all four aspects revealed a similar score shape for both groups. Henceforth a Mann-Whitney U test was selected but, as the scores were not normally distributed, median scores had to be used for comparing both groups following Field (2013). As can be seen in Table 19, all median scores were found to be higher for DUB students than for SB students. TABLE 19. Scores on four aspects of oral proficiency compared between two groups Fluency Grammar Vocabulary Listening Median (Mean / SD) Median (Mean / SD) Median (Mean / SD) Median (Mean / SD) SB program (N=55) DUB program (N=73) Mann Whitney U Stand. score z Cohen’s d Asymptotic Sign. (2-tailed) 5.0 (4.8 / 1.2) 7.0 (6.9 / 0.9) 3673 8.2 d = 2.01 p < .001 5.0 (4.6 / 1.1) 6.0 (6.3 / 1.2) 3322 6.5 d = 1.31 p < .001 4.0 (4.5 / 1.1) 7.0 (6.5 / 1.2) 3533 7.5 d = 1.71 p < .001 4.0 (4.2 / 1.2) 7.0 (6.9 / 1.1) 3749 8.5 d = 2.21 p = .000 Subsequent Mann Whitney U tests revealed these di erences to be signi cant for all of the four subtests of the SOPA measure, yielding substantial e ect sizes as well (Cohen, 1988) ranging from d = 1.31 to d = 2.21 on all four aspects of speaking. In short, on the basis of the SOPA scores of oral French L2 pro ciency, the DUBtaught students signi cantly outperformed the SB-taught students, not only on the total score but also on the sub-scores, albeit with some subtle di erences in e ect sizes. DISCUSSION AND CONCLUSION e aim of this study was to present long-term classroom research where students are learning to achieve communicative ability rather than “just learning for the sake of the experiment” (DeKeyser, 2003; page 337) in developing their L2 French. We speci cally targeted the development of oral L2 French pro ciency and compared two groups of students a er six years of CLT instruction—one taught by means of an SB-inspired program and the other on the basis of a DUB program, which can be viewed as a weak versus strong format of the CLT approach, respectively. e students in the SB program received explicit grammar and vocabulary teaching with minimal authentic L2 exposure. is program is commonly adopted in secondary schools in the Netherlands. e students in the DUB program, on the other hand, were part of a program that