Detection of Compliance Violations using Management Information and Monitoring Systems
Communication has shown to be crucial for decision-making in very different contexts. For example, experimental research has repeatedly demonstrated that between-subject communication is a powerful driver of collusion in various market games. Yet, in the beginnings of experimental research, only little was known about the specific way in which between-subject communication affects individual decision-making. Therefore, recently, the role and analysis of natural language communication has gained attention in experimental research.
One important question is whether certain types of communication affect decisions differently than others. In this regard, Houser and Xiao (2011) present an approach for the classification of natural language messages, using a coordination game with external evaluators. This objective classification procedure allows to infer deeper insights from communication data. The primary limitation of this approach is its application to large data sets. Hence, I present a new approach based on the coordination game by Houser and Xiao, combining their procedure with a machine learning text analysis component. Thus, I am able to analyse larger data sets, based on a small training data set classified beforehand by human evaluators. Data reported by Charness and Dufwenberg (2006) is used in order to test the approach. I am able to substantially replicate the classification results of the original classification given by Charness and Dufwenberg themselves.