Inference for influence over multiple degrees of separation on a social network

Area of Science:

• Social network analysis
• Sociology
• Network science

Background:

• Individual states and behaviors are often correlated within social networks.
• The 'three degrees of influence rule' by Christakis and Fowler attempts to quantify this dependence.
• Understanding the extent and nature of social influence is crucial in various fields.

Purpose of the Study:

• To critically examine three distinct interpretations of the 'three degrees of influence rule'.
• To differentiate between association, actual causation (contagion/social influence), and direct effects.
• To outline analytical procedures and methodological challenges for each interpretation.

Main Methods:

• Conceptual analysis of social influence models.
• Discussion of statistical and analytical approaches for network data.
• Identification of methodological hurdles in inferring causation from correlation.

Main Results:

• The paper distinguishes between correlation-based influence (associations), causal influence (contagion), and direct effects.
• It highlights that evidence for each interpretation requires different analytical rigor.
• Methodological challenges escalate with each successive interpretation, moving from association to causation.

Conclusions:

• The interpretation of social influence rules requires careful consideration of the underlying mechanisms.
• Distinguishing between association and causation is paramount in social network research.
• Future research must address the significant methodological complexities in studying social influence.