This dataset contains 99 persuasive texts produced by L2 undergraduate students at a U.S. university. It also includes the writing process information for these texts which was recorded using Inputlog 7 (Leijten & Van Waes, 2015).
The persuasive texts were annotated using an argumentative rubric for classifying discourse elements found in Crossley et al. (2022). The rubric comprised five categories as the building blocks of the argumentation framework: final claim, primary claim, counterclaim, rebuttal, and data.
This repository contains the 99 persuasive texts, the corresponding keystroke logs (in .idfx format), the annotation results, the keystroke measures for each argument element in the texts, and the demographic information for the L2 writers. In addition, we also include two R scripts: one was used to calculate inter-annotator agreement (IAA) for the dataset and the other was used to build the MCMCglmm model for the statistical analysis.
Please refer to the paper below for a detailed description of the dataset and the context of its construction: