Various factors affecting language learning are introduced, including demographic variables, and learners’ L1, cultural background and the context of language use, noting that the analysis of learner corpora can enable the exploration of language-learning processes during SLA and across different contexts. Practical challenges involved in building extensive learner corpora, especially spoken learner corpora, are discussed (e.g. variable constraints, scale of data, availability of data). The Trinity Lancaster Corpus (TLC), a spoken corpus based upon a language proficiency test, and two other corpora, are then introduced. The chapter then discusses MDA and its adaptation for short texts (short-text MDA). The chapter describes the challenges of analysing short texts within corpora and explains how short-text MDA may make it possible to explore discourse at both the micro-structural (turn) and macro-structural (discourse units) levels. The chapter concludes by noting that this exploration will lead to a deeper analysis of narrative structures as a result of the findings from the corpora studied in the book using short-text MDA.