This research investigates the ability of predictive measures to differentiate level of language proficiency among learners across languages, language categories, and learning contexts. It fills a gap in the literature pertaining to language categorization and demonstrates differential predictive ability of language learning aptitude measures depending on the language being learned. In addition, it challenges a default assumption that aptitude and other individual difference measures ought to be context independent. This is done through an analysis of the effects of context on the predictive ability of individual difference measures where results show the differing predictive patterns between a foreign language classroom, a domestic intensive instruction setting, and a study abroad program. Finally, several individual difference measures that have shown some past success in differentiating foreign language outcomes for learners are examined to analyze incremental predictive validity. Measures that demonstrate incremental predictive validity are useful in developing selection protocols for language learning programs. Additionally, measures that show differential incremental predictive validity across language categories and contexts may indicate a potential for aligning learners within a category and context to benefit learner outcomes.

This research provides evidence to support claims that suggest an interactive role between the learner and context leading to differential learning outcomes based on individual differences. It highlights the fact that predictive models of proficiency are not consistent within language category, nor are they consistent across language category boundaries. It shows that a measure of general cognitive memory may be the best indicator of long term language learning success across languages. Finally, it replicates earlier findings that the Defense Language Aptitude Battery (DLAB) provides incremental predictive validity in the face of other individual difference measures indicating that it remains a useful predictor of language learning performance.