# Componential modeling of argumentative essay writing from multiple online sources: a Bayesian network approach

## TL;DR

This study explores how college students write argumentative essays using multiple sources, revealing that the process relies on higher-order thinking and often falters due to poor counterargument framing.

## Contribution

The paper introduces a novel Bayesian network approach to model the cognitive processes involved in writing argumentative essays from multiple sources.

## Key findings

- The argumentative essay task is initiated and sustained by higher-order integration components.
- Students struggle with framing counterarguments and performing optimal critical analysis.
- The Bayesian network model supports the stage-based Integrated Framework of Multiple Texts.

## Abstract

Writing argumentative essays using multiple sources is a critical skill for college students, yet it remains a significant challenge. Despite previous research acknowledging this difficulty, the specific dynamics of the argumentative essay writing process and where breakdowns occur remain unclear.

College students wrote argumentative essays on a controversial topic after reading multiple documents. The data were fitted to two competing theory-based Bayesian networks, a method highly suited to the modeling of cognitive processes identified with argumentative writing.

The best-fitting model showed that the argumentative essay task is both initiated and sustained by higher-order integration components. This model lends support to the description of the process of argumentation writing from multiple documents put forth by the stage-based Integrated Framework of Multiple Texts. Further, we found that the process of argumentation falters due to students' inability to frame counterarguments and their non-optimal critical analysis.

This research not only enriches our understanding of the mechanics of argumentative writing from multiple sources, but the innovative Bayesian approach could lead to further refinement of the model by future researchers.