Collective intelligence at Enron during the California energy crisis : uncovering collaborative innovation networks using social network analysis

Abstract

As interaction takes place between individuals, relationships are formed and collaboration and innovation emerge. In this thesis I have applied Coolfarming (Gloor, 201 lb), a social network analysis method using Condor, a software tool to quantify communication patterns based on various data sources. I analyzed the Enron email archive to see if communication patterns of convicted employees differ from ordinary ones. Toward that goal, I compared the dynamic semantic social network metrics of 17 Enron employees convicted in the criminal trial following Enron's implosion with a control group of ordinary employees. I focused on 17 mailboxes of 24 Enron executives that were convicted. Identifying criminals based on email behaviors is possible depending on the sampling strategy. When sampling based on employees with comparable total emails, the statistical analysis of the Contribution Index (Ci) metric revealed that criminals were less active. When sampling based on employees with comparable total influence, the statistical analysis of Betweenness Centrality Oscillation (Bco) and Degree Centrality (Bc) metrics revealed that criminals were less connected to others and less creative.