Network Communication Experiments

While narratives have shaped cognition and cultures for centuries, digital media and online social networks have introduced new narrative phenomena. With increased narrative agency, networked groups of individuals can directly contribute to and steer narratives that center our collective discussions of politics, science, and morality. Online network experiments provide a unique window to empirically study how networked group communication shapes narrative interaction. However, previous research has relied on relatively simplistic interactions to assess how a group's network structure impacts collective outcomes, with limited generalizability to the interaction complexity of real-world online environments (e.g., coordinating hashtags to facilitate narrative spread). This is a repository for experimetnal and data analysis software for running online netowrk experiments, and data and scripts from experiments our team has previously ran. Please conslut these articles for more information on our findings:

• Priniski, J. H., Linford, B., Krishna, S., Hirschman, A., Rodriguez, N., Morstatter, F., Brantingham, J., & Lu, H. (2024). Cognitive complexity and neighborhood structure shape narrative interaction in online social networks. (In prep).
• Priniski, J. H., Linford, B., Krishna, S., Morstatter, F., Brantingham, J., & Lu, H. (2024). Online network topology shapes personal narratives and hashtag generation. Proceedings of the Cognitive Science Society. Article PDF

What is in this repository?

Repository for experimental software, statistical analysis, computational modeling, and behavioral data used in ongoing network experiments on group communication. This README.md provides a roadmap for the software and data stored in this repository. Directories will have additional README.md files to help with using the software.

Statistical Analyses

This directory contains Bayesian GLMs and ggplot figures that visualize and model group-level behaviors and narrative shifts resulting from interactions under different network conditions. These GLMs include predictors that test hypotheses on how the cognitive complexity of interacting over various digital media mediates the impact of a group's network structure on its dynamics.

Experimental Data

Raw and processed data from all experimental runs is in this directory. Data carpentry scripts are available as Jupyter Notebooks, which transform the OTree dataframes (one per experimental run) into long format and concatenate them across all runs for statistical analysis and hypothesis testing.

Network Experiment Software

This directory contains the Python (OTree) and JavaScript software for running a network experiment on group communication. There are two folders, face_experiment and hashtag_experiment, which ask participants to communicate and align responses using different digital media. The README.md in this directory provides instructions for setting up a server to host the experiment, which you can customize and extend with minimal Python and JavaScript expertise. Below is a high-level overview of the experimental procedure. The pre and postinteraction phases were administered using a Qualtrics survey, which linked participants to the network server (OTree python code). The code for each of these experimental phases can be found in this directory.