Exploring individual longitudinal development in a corpus of 'natural' disciplinary writing: what could it mean for teaching?

Two major research issues for longitudinal investigations are the 'naturalness' of the learner language being studied, and the extent to which the research design is truly longitudinal. Regarding the first issue, data for longitudinal studies are usually collected in language classrooms or exams, with the writing tasks/topics being tightly controlled. A major disadvantage is that tasks can be unrepresentative of the writing required in disciplinary content courses. The present study analyzes writing development as it occurs 'naturally' in university disciplinary content courses, using four complexity features (i.e., finite adverbial clauses, finite relative clauses, attributive adjectives, and nouns as pre-modifiers) with the goal of comparing the kinds of findings in a True-longitudinal approach, a Quasi-longitudinal approach, and a Case study approach. This close examination of writing development that controlled for discipline and task showed an increase in development as measured by these linguistic indicators.