How to make methodological decisions when inferring social networks

Area of Science:

• Behavioral Ecology
• Network Science
• Animal Social Behavior

Background:

• Constructing social networks for new study systems presents challenges due to a lack of standardized guidance.
• Researchers may face multiple hypothesis testing issues when exploring various data collection and network construction methods.
• Existing methods lack a clear framework for making informed decisions in novel research contexts.

Purpose of the Study:

• To propose a standardized approach for constructing social networks in new study systems.
• To demonstrate how a priori knowledge can guide methodological decisions in social network analysis.
• To avoid the pitfall of multiple hypothesis testing when inferring social structures.

Main Methods:

• Evaluated different data collection strategies (e.g., feeder numbers) and association definitions.
• Utilized a colonial cooperatively breeding bird (sociable weaver) as a model system.
• Assessed methods based on their ability to recover known breeding groups and detect social differentiation.

Main Results:

• A priori knowledge of species behavior can independently inform the best edge definition for network construction.
• This approach successfully identified distinct breeding groups and differentiated social relationships in sociable weavers.
• The chosen method maximized the recovery of known social structures and detected significant social differentiation.

Conclusions:

• Decisions on network construction, particularly edge definition, should be based on existing biological knowledge.
• This strategy provides a robust framework for social network analysis in diverse study systems.
• Utilizing a priori knowledge minimizes methodological bias and enhances the reliability of social network research.