Using Coh-Metrix to Analyze Writing Cohesion in Introductory Courses With and Without Learning Assistants

Authors: Pauline Ho, Amanda Nili, Lynn C. Reimer
Presented at 2017 AERA

​Abstract: Learning Assistants provide one-on-one support for students enrolled in specific university classes. This study employed Coh-Metrix, a computational linguistic data-mining tool, to investigate the effects of having Learning Assistants (LAs) as a facet of instruction on students’ writing ability. Preliminary results suggest that students exposed to LAs display a higher level of textual cohesion in their writings. They also tended to write more paragraphs with shorter sentences in each paragraph, when compared to students enrolled in non-LA courses. However, preliminary results indicate that both groups are associated with similarly low levels of word concreteness, suggesting that exposure to LAs is not related to improvements in students’ concrete vocabulary.