﻿ Wim Gombert – Page 81

# Wim Gombert

CHAPTER 5. Writing skills 79 STATISTICAL DESIGN Holistic ratings, morphosyntactic pro ling scores, and CAF measures were inserted in SPSS (Version 27) and a er assumptions of normal distribution and homogeneity of variance were checked, Independent Samples t-tests were conducted (p<.05) on all variables. ere were no outliers in the holistic and morpho syntactic pro ling scores: the scores of participants for each teaching program were normally distributed, as assessed by Shapiro-Wilk’s test (p>.05), and there was homogeneity of variances, as assessed by Levene’s test for equality of variances (p>.05). e mean score and the Standard Deviation were computed, and the e ect size was calculated, using Cohen’s d (Cohen, 1988). As to the CAF measures, there were minor outliers in the data. For most measures, the scores were normally distributed as assessed by Shapiro-Wilk’s test (p>.05) and there was homogeneity of variances, as assessed by Levene’s test for equality of variances (p>05). For Average Sentence Length (ASL), there were only minor violations in the SB group (one outlier and no normal distribution). Hence, a non-parametric test was not assumed necessary for analyzing these data and we decided to use only the regular parametric test. For text length, however, there were outliers in both groups (one higher and three lower) and both assumptions of normal distribution and homogeneity of variance were violated. erefore, a er having observed symmetry between the shape of distribution of both groups, we used the Mann-Whitney U test for analyzing scores on text length. For the interpretation of e ect sizes of both holistic and analytical measures Plonsky and Oswald’s (2014) SLA eld-speci c benchmarks were used: small (d = 0.4), medium (d = 0.7), and large (d = 1.0).

RkJQdWJsaXNoZXIy MTk4NDMw