We applied a discourse analysis (DA) to the electronic chat room discussion
s of a 16-week, Internet-based section of a class in statistical methods in
psychology. This analysis revealed that across the semester, several DA ca
tegories (e.g., total number of student comments) were correlated with fina
l grade in the class. An additional analysis involving only the chat room d
iscussion of Week 3 revealed that 2 DA categories (i.e., student response t
o a problem or example given in lecture and total number of student comment
s) correlated with final grade in the class. We discuss the pedagogical imp
lication of these results with regard to an instructor's ability to identif
y early warning predictors of student performance in the virtual classroom.