Detecting phase transitions in small group conversations through entropy methods

Document Type

Conference Proceeding

Publication Date



Entropy methods (also known as mutual information methods) are often used to analyze time series categorical data, such as results from qualitative coding of small group discussions. A maximum in the calculated entropy is indicative of a phase transition. This paper presents a Bayesian-inspired method to distinguish between random fluctuations that result in a local maximum value of entropy and a maximum value of entropy that indicates a transition in the system. Using probabilities derived from the data on one side of the local maximum as a prior and the probabilities derived from the data on the other side as a likelihood, posteriors are calculated. Through bootstrapping and chi-squared tests, significant differences between priors and posteriors can be determined. Applications to real world data include the ability to detect phase changes in group interactions.


Presented at the 7th Annual Conference of the Upstate New York Chapters of the American Statistical Association in Rochester, New York, April 21, 2018.

This document is currently not available here.

Additional Files