A Multivariate Supervised Learning Model as Predictor of the Learner’s Decision to Drop out or Persist in (L2) Classroom Courses

Published: 9 May 2018| Version 1 | DOI: 10.17632/wd39k4c2fr.1
Mohammed R. Dahman


Learners differ enormously in how successful they are in (L2) acquisition due to individual differences, which are varied not only in the speed of acquisition or the level of achievement but also in the decision to drop out or persist in a classroom learning courses. The purpose of this study was to determine whether the two groups of learners (persistence and dropouts) are different in their learner variables (demographic and affective variables). Quantitative data were collected from 278 learners who had dropped out or retained in Intermediate English Language Course (B1), which offered by a mass education center in Istanbul Turkey. The two groups showed statistical differences in the language proficiency test score (explaining roughly 34.6% of the variation); furthermore, in the perceptions of motivation (explaining 42% of the variation), attitude (explaining 40.2% of the variation) and anxiety (explaining 28.9% of the variation). It was also shown that the theoretical framework, which includes affective variables in addition to demographic variables, can predict learners’ decision to drop out or persist. The educational implications of these findings for understanding the learners’ decision to drop out or persist in (L2) learning classroom courses are discussed, along with suggestions for future research.



Education, Machine Learning, Decision Sciences