Analyzing network data from the worlds of sports and entertainment provide unique insights: athletes and performers are often in uniquely fragile occupations where contracts are short and payoffs are high. This variance increases the visibility of social mechanisms that are present but largely unobservable in more stable occupations.
For example, networks of fist-fights between players in the NHL reveal interesting patterns: mainly the emergence of a centralized "core-periphery" structure, even though fighting is allegedly part of a decentralized and honor-based system of social control. When interesting structures and patterns appear in network data sets - it is rarely by chance. These patterns are a clue to understanding what motivates and restricts interaction between pairs of individuals. I show how the core-periphery pattern in hockey is actually due to gradual specialization of fighting over time (the emergence of 'enforcers') and the need for these players to signal their own willingness to act violently on behalf of their teammates. A second paper I have written on the topic demonstrates how players specialize in this way over the course of their career.
In other studies currently in progress my colleagues and I use rich relational data sets to identify and explore similar mechanisms and puzzles. Instead of focusing on fighting relationships within hockey games, we focus on collaborative relationships within various sectors of the entertainment industry.