This paper presents a systemic intervention approach as a means to overcome the methodological challenges involved in research into computer-supported collaborative learning applied to the promotion of mathematical problem-solving (CSCL-MPS) skills in schools. These challenges include how to develop an integrated analysis of several aspects of the learning process; and how to reflect on learning purposes, the context of application and participants' identities. The focus of systemic intervention is on processes for thinking through whose views and what issues and values should be considered pertinent in an analysis. Systemic intervention also advocates mixing methods from different traditions to address the purposes of multiple stakeholders. Consequently, a design for CSCL-MPS research is presented that includes several methods. This methodological design is used to analyse and reflect upon both a CSCL-MPS project with Colombian schools, and the identities of the participants in that project. Copyright © 2013 John Wiley & Sons, Ltd.Keywords computer-supported collaborative learning; mathematical problem-solving; critical systems thinking; systemic intervention; methodological design INTRODUCTION In today's world, the use of information and communication technologies (ICTs) to assist learning processes is on the increase. One of the main purposes of ICT use in learning processes is to facilitate collaboration and, therefore, improve learning by means of sharing and distributing knowledge. This purpose has been pursued through a new educational paradigm called 'computersupported collaborative learning' (CSCL) (Lipponen, 2002;Lehtinen, 2003), which has raised fresh theoretical and methodological issues for researchers Systems Research and Behavioral Science Syst. Res. 32, 86-105 (2015) Published online 20 September 2013 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/sres.2220 because of the complexity of analysing different factors in this type of learning process. For example, research has to consider analyses at both the individual and group levels; the nature of interactions between learners; students' attitudes; technological aspects; and the school's context. Considering these factors in a coherent and reflective manner is a methodological challenge for CSCL research (Dillenbourg, 1999;Daradoumis et al., 2006;De Laat et al., 2006;Stahl et al., 2006;Arbaugh et al., 2010, Arjava, 2011Strijbos, 2011).Currently, CSCL researchers are proposing several different methodologies to analyse CSCL processes, such as the triangulation of quantitative methods (e.g. social network analysis, statistics and surveys) with qualitative methods (e.g. interviews and content analysis) (Daradoumis et al., 2006;De Laat et al., 2006;Cho et al., 2007;Rienties et al., 2009;Gress et al., 2010). Also, some analysis of the context of CSCL activities is beginning to be included alongside the analyses of processes (Arnseth and Ludvigsen, 2006;Arjava, 2011). Although these methodological developments are improving the way CSCL ...