Abstract: This paper is a report on the findings of an action research study conducted in a middle school on eighth grade students who were participating in a blended, self-directed Biology course. Various types of data analysis techniques were performed in order to investigate the students' personal learning networks and how these were affected by this technology enhanced self-directed learning and a blended learning environment. Findings indicate that the students who performed better than expected in the class had the largest and most robust learning networks. The converse was also true; students who did less well than expected were shown to have small and less robust learning networks. This study develops a methodology for applying Connectivist learning theory to a blended learning environment.