Spectral clustering of ten thousand causes of action and some management implications for the future

A fascinating article concerns an analysis of causes of action in some 2,500 civil cases filed in Federal courts filed between 2000 and 2008.  The research was by Christina Boyd, et al, in “Building a Taxonomy of Litigation: Clusters of Causes of Action in Federal Complaints”, which is available electronically.  The four authors used a quantitative procedure known as cluster analysis, specifically spectral clustering, to determine from 11,439 causes of action eight primary clusters.    The article nicely explains all the research and draws solid conclusions.


From the standpoint of those who manage corporate litigation, this research has much potential.  If there were data for these cases on costs paid by the defendants as well as for durations of the cases, it would help predict case outcomes, case lengths, costs, and forms of resolution.  A law department could analyze its portfolio of pending litigation by cause of action and perhaps determine where proactive research makes sense, knowledge management have more payoff, form documents make a contribution, and choice of law firm.  The authors take mostly a scholarly approach to their work, but they do write (at 26), “the more systematic information that case actors have about their case and how it compares to others, the more able they should be to make strategically wise decisions.”   This is another example of the deep potential for evidence-based management.

We welcome comments

Your email address will not be published. Required fields are marked *