CSCL research has given rise to a plethora of analysis methods, all with specific analysis goals, units of analysis, and for specific types of data (chat, threaded discussions, etc.). This article describes some challenges of CSCL-analysis. The development of an analysis method for negotiation processes in multidisciplinary teams serves as an example of how these challenges occur in practice. Results reveal a number of tensions with regard to establishing reliable coding procedures without compromising validity, such as how to define codes, and rules for applying codes. The discussion offers some guidelines for content analysis of CSCL-data.