This study addresses the problem of information overload and disorganization in Massive Open Online Course (MOOC) discussion forums by developing a model to classify and identify threads based on their relevance to course content. The study defines content-related posts as those that provide or seek help on course material or share/comment on relevant resources. A linguistic model was created using manually-coded posts from a statistics MOOC and tested on posts from the same and another statistics course. The results indicated that content-related posts had unique linguistic features, and the model demonstrated strong cross-course reliability with high recall and precision. The number of views and votes did not contribute to effective classification.