Clustering research questions

Research clusters
A really nice idea for clustering research questions as a basis for a workshop activity in a session at the OER meeting at Monterey facilitated by Lisa Petrides and Amee Godwin. See this cloud for a summary. Worked really well, story has any one else done similar things? I can see how this could be adapted in a range of ways to brainstorm, cluster and refine ideas.

5 Responses to “Clustering research questions”

  1. Will Says:

    We’ve used this for software development on a number of projects and I think it’s used quite widely for some of the larger EU projects. I’ve used it with projects involving EADTU and it works very well. We have used a version of this for refining requirements and based on Volere templates.

  2. liz thackray Says:

    I haven’t done this in a research context, but I can remember doing similar things in development context in 80’s. Was useful for refining and prioritising development ideas in NGO I worked in at the time.

    If I remember right, we also did something similar at one of the OpenCETL conferences, but didn’t take it beyond the gathering people with similar interests.

  3. Paul Sweeney Says:

    I’ve come across it - and used it many a time- as a standard technique in project management for brainstorming the key components of a project leading to what they call the Work Breakdown Structure (WBS). Good for collecting and categorising group knowledge as well as generating agreement about what to do next. Following a quick google, adapting i.e. decommercialising many of the techniques mentioned here and other places may be useful
    http://www.mindtools.com/pages/main/newMN_CT.htm

  4. GrĂ¡inne Says:

    Thanks all! Good to get this feedback! We also used a variant of this at the XDelia kick off meeting Eileen Scanlon and I have just been at.

  5. Robin Gower Says:

    I’ve done this in the more general context of clustering ideas into themes (called Affinity analysis on the link Paul provided). I find that it’s best done unilaterally by a domain expert/ lateral thinker rather than in group debate. The approach is useful but falls at the following hurdles:
    - causal/ interdependent relations across themes makes it hard to distinguish separate thematic areas of enquiry (the WBS and it’s twin the Product Breakdown Structure go some way to addressing this by combining discrete intermediate products into final products)
    - differing interpretations over the relevance of an idea (an idea can, quite rightly, belong in more than one cluster)
    - contradictory interpretations of the implication of an idea (again belonging in more than one cluster)
    - minority concerns tend to get overlooked (small awkward clusters are often ignored).

    I’ve found that networked dialogue maps (e.g. Compendium) are a more effective (if slightly more cumbersome) way of interpreting research questions.

Leave a Reply