While online students should take charge of their own learning and form collaborative learning communities, constructivist instructors should scaffold online learning without dominating course discussions. This research continues the longitudinal investigation of web-based courses at the Faculty of Education, University of Pretoria. The mixed methodological approach this investigation followed consisted predominantly of qualitative methods, augmented with quantitative approaches. I used two distinct online tools to explore student participation in an eight-week online Masters’-level course delivered via the WebCT™ platform. First, I reviewed the use of metaphors in the literature by a framework of requirements for successful online learning. The use of metaphor supports constructivism, facilitates course interaction, helps to avoid students’ initial inertia in online discussions, and contributes to the development of virtual learning communities. I researched how an explanatory metaphor as tool supported online participation and indicated that metaphors eased students’ communication of important and difficult issues. Secondly, I used the tool of a covert virtual student that also acted as an additional facilitator and course helper. I examined the ethical implications of the carefully concealed real identity of the mythical online helper, methical Jane. As she took part in all course activities and assignments, as well as providing her co-students with cognitive and technical support, the students accepted and integrated her presence in their virtual learning community. I consequently analysed students’ reactions to her identity after disclosure of her origin after the course. Although the exposure precipitated students’ shock, disbelief and dismay as she was a convincing virtual student, they did not object to the presence of a virtual student, but rather felt betrayed due to her hidden real identity. The benefits of this teaching intervention include experts supplying technical expertise, multiple faculty enriching the learning experience, and support and teaching assistants and tutors participating with smaller groups in large online classes. I further examined how frequency of course access, discussion postings, collaborative behaviour and integration into a virtual learning community relate to learning and course completion. Quantitative indices indicated highly significant differences between the stratifications of student performance. Absent and seldom-contributing students risked missing the benefits of the online learning community. Students were discontent with peers who rarely and insufficiently contributed to group assignments. Low participation varied from only reading, skimming, or deliberately harvesting others’ contributions, to high student contributions of little value. Conclusions on the formation of an online learning community indicate that the passport to membership of the community is quality participation, rather than prior peer acquaintance. I indicated that students’ learning benefited from contributing high quality inputs to online learning communities while students with poor participation did not benefit from the online learning community. Online facilitators contribute to students’ learning through the timeliness and quality of tailored scaffolding. Recommendations for future research include uncovering the reasons for students’ stressful experiences of online learning; the effect of online assessment on student course participation; the alignment of learning metaphors in multi-cultural learning environments; and the support of non-participating online students.