Simulating relational event history data: why and how

Area of Science:

• Social Network Analysis
• Computational Social Science
• Statistical Modeling

Background:

• Social phenomena often involve repeated interactions over time, necessitating methods to analyze these dynamics.
• Understanding social interaction mechanisms requires statistical simulation techniques for fine-grained temporal network data.

Purpose of the Study:

• To present statistical frameworks for simulating relational event networks using dyadic and actor-oriented models.
• To demonstrate the utility of simulation in addressing key challenges in temporal social network analysis.
• To introduce the R package 'remulate' for implementing these simulation frameworks.

Main Methods:

• Development of statistical frameworks for relational event models.
• Implementation of these frameworks in the R package 'remulate'.
• Application of simulation techniques across five diverse case studies.

Main Results:

• The 'remulate' package provides tools for simulating relational event networks.
• Simulation aids in model assessment, social theory development (e.g., optimal distinctiveness), and understanding intervention effects.
• Simulation-based analysis enhances model sensitivity assessment and future relational dynamics prediction.

Conclusions:

• The presented simulation framework and 'remulate' package are valuable tools for researchers.
• These tools facilitate a deeper understanding of social interaction dynamics from real-life relational event data.
• Simulation is crucial for advancing temporal social network analysis.