This paper explores how to identify and model the depth and quality of student participation in online discussion forums by analyzing message content through two frameworks: the ICAP and Community of Inquiry (CoI). By training predictive classifiers on the same dataset, the study aims to compare these frameworks' perspectives on learning. The findings indicate that deeper and higher-quality participation is linked to longer, more complex messages and the importance of threaded replies over temporal order. However, there are notable differences in how each framework handles messages of affirmation.