An analysis of complexity in primary school L2 English learning

In the context of a Complex Systems (CS) approach to language acquisition and teaching research, this paper will use Geometric Data Analysis (GDA, l’Analyse géométrique des données: Benzécri 1973; Le Roux & Rouanet 2010) to explore data collected from beginning learners of English (n = 54) in a primary school context in France. The study was initially designed along classical “group comparison” lines (six-year-olds vs. eight-year-olds), and not from the start for CS analysis; participants performed four different tasks in English at two times during their first year of English; they also completed a variety of psychometric tests and questionnaires measuring various cognitive, verbal, affective et conative characteristics. We will begin by looking at classic nonparametric comparisons of the two groups of learners and correlations between the psychometric tests and performance on the English tasks. These univariate statistics do not actually turn up very much structure in the data set, and we will use multivariate GDA for a closer look at the complex relationships between individual features and early L2 skill.