Causal discourse analyzer: improving automated feedback on academic ESL writing

Expressing causal relations plays a central role in academic writing. While it is important that writing instructors assess and provide feedback on learners' causal discourse, it could be a very time-consuming task. In this respect, automated writing evaluation (AWE) tools may be helpful. However, to date, there have been no AWE tools capable of evaluating causal discourse. The authors of the present study attempt to fill in this gap by (1) developing an automated causal discourse analyzer and (2) investigating how accurately the analyzer processes learners' causal discourse in academic writing. The accuracy of the analyzer is evaluated on cause-and-effect essays written by 17 non-native undergraduate students. The results indicate precision of .93, recall of .71, and accuracy of .76, which is promising for pedagogical applications of the analyzer, that is, providing learners with automated formative feedback specific to causal discourse.