Wim Gombert

CHAPTER 4. Reading and listening skills 65 RESULTS Preliminary assumption checking revealed that there was a linear relationship between Reading and Listening scores in each group, as assessed by means of a scatterplot. ere was no evidence of multicollinearity, as assessed by Pearson correlations: r < .9 (r = .713 for group 1 and r=.537 for group 2). We found one univariate outlier in group 1 and there were no multivariate outliers in the data, as assessed by Mahalanobis distance (p > .001, maximum value = 10.12 and critical value = 13.82). Reading and Listening scores were normally distributed for each group, as assessed by Shapiro-Wilk’s test with Bonferroni adjustment (Armstrong, 2014): p > .125). ere was homogeneity in our data, as assessed by Box’s test of equality of covariances matrices (p = .322) and homogeneity of variances could be assumed, as assessed by Levene’s Test of Homogeneity of Variance (p>.05). As most assumptions were thus met, except for one univariate outlier found in the SB group, and the number of participants in this group was rather large (55), we assumed the result would not be a ected and we decided to use a Hotelling’s T2 test for all data, including the outlier. e di erence between the programs on the combined dependent variables was statistically signi cant, F(2,125) = 13.389, p < .001; Wilks’ Λ = .824; partial η2 = .176. To compare each dependent variable, independent samples t-tests were performed on reading scores and listening scores separately. Table 9 shows the mean reading and listening scores per condition as well as the comparison results. . TABLE 9. Reading and listening scores after six years of instruction SB program N=55 DUB program N=73 Cohen’s d Significance (2-tailed) Mean (SD) Mean (SD) Reading scores (Max. 10) Listening scores (Max. 10) 6.36 (1.28) 5.96 (1.32) 6.67 (1.07) 7.06 (1.24) d = 0.26 d = 0.86 p = .143 p < .001*** Independent samples T-tests showed a signi cant di erence between both groups of students on listening in favor of the DUB students ((p <.001), with a large e ect size (d=0.86), and a non-signi cant di erence between both groups on reading skills (p = .143), with a small e ect size (d = 0.26).

RkJQdWJsaXNoZXIy MTk4NDMw