Several studies have analysed and assessed online performance and discourse using quantitative and qualitative methods. Quantitative measures have typically included the analysis of participation rates and learning outcomes in terms of grades. Qualitative measures of postings, discussions and context features aim to give insights into the nature of learning activities taking place in the online learning environment. No matter what the particular method, one of the critical aspects in these analyses is the coherence between theory and methodology as well as the explicit focus on the unit(s) of analysis. Technology-enhanced learning (TEL) settings typically involve processes (e.g. cognitive or social) at different levels, and these processes are often intertwined with each other in a way that is not reducible to any one level only. In this article, a short review on the challenges of research-based evaluation of TEL, and computer-supported collaborative learning in particular, is made. With these challenges in the background, the integration of different levels of analysis is demonstrated with the aid of the method developed for analysing the level and reciprocity of discussions in a Web-based learning setting.