Individual profiling of second language learners through word association.

This thesis explores the organisation of second language learners' mental lexicons through the use of word association tests; a reliable measure of which would complement established measures of lexicon size. Following studies with native speakers (Russell & Jenkins, 1954; Ervin, 1961), research with second language learners began in the late 1950's (Lambert. 1956) although much of the methodology and theory had been developed decades before by clinical psychologists (Jung, 1918). Unlike the LI studies, the L2 studies have been plagued by inconsistent findings, leading some to conclude that the use of word associations to assess L2 learners is unfeasible (Kruse et al., 1987). In an attempt to realise the potential that word association tests have as a method of measuring the organisation of learner lexicons, a series of experiments was conducted. The initial experiment was a replication of Wolter (2001) using a traditional classification system. This was followed by five more experiments that centred around a quite different methodology and approach to data analysis put forward by Fitzpatrick (2006; 2007). The reliability of Fitzpatrick's individual profiling approach was tested using various kinds of stimulus words. The results indicate that the word class and frequency of the stimuli have little effect on the reliability of the response profiles generated, Improvements to the methodology and issues that arose during the experiments are discussed. The experiments were all conducted in Japan, with college aged learners between early 2007 and mid-2012. In that six year period, over 20,000 responses were elicited from 213 learners involved in the pilot tests and main experiments