The effect of automated fluency-focused feedback on text production

Abstract

This article presents a new intervention for improving first-language writing fluency and reports an empirical study investigating the effects of this intervention on process and product measures of writing. The intervention explicitly encourages fluent text production by providing automated real-time feedback to the writer. Participants were twenty native-English-speaking undergraduate students at a large Midwestern university in the United States, all of whom were proficient writers. Each participant composed two texts (one in each of the control and the intervention condition) in an online text editor with embedded keystroke logging capabilities. Quantitative data consisted of product and process measures obtained from texts produced by participants in the control and the intervention condition, and qualitative data included participants’ responses to an openended questionnaire. Linear mixed-effects regression models were fit to the quantitative data to assess differences between conditions. Findings demonstrated that there were significant differences between the intervention and the control condition in terms of both the product and the process of writing. Specifically, participants wrote more text, expressed more ideas, and produced higher-quality texts in the fluency-focused intervention condition. Qualitative findings from questionnaire responses are also discussed.