Park, Patrick S., Joshua E. Blumenstock, and Michael W. Macy. 2018. “The Strength of Long-Range Ties in Population-Scale Social Networks.” Science 362 (6421): 1410–13. https://doi.org/10.1126/science.aau9735.

Notes

Questions:

• Isn't long-range ties in a way not contesting the idea of homophily? How do we compare the effects of different variables on a connection?
• Is there a psychological aspect to the strength of long-range ties?

Summary

The study "The Strength of Long-Range Ties in Population-Scale Social Networks" challenges long-standing assumptions about the relationship between social tie strength and network distance1. Analyzing 11 population-scale communication networks (56 million Twitter users across eight countries and 58 million mobile phone subscribers in three nations), the researchers found that long-range social ties are nearly as strong as embedded ties within close-knit circles, overturning the traditional "diversity-bandwidth trade-off" hypothesis.

Key Findings

1. U-shaped relationship between tie strength and range:
   • Tie strength initially decreases as range (network distance) increases from 2 to 4, but then strength rebounds at longer ranges (6+).
   • In phone networks, long-range ties (range ≥6) were comparable in strength to embedded ties with one common neighbor.
2. Network "wormholes":
   • These rare but strong long-range connections (0.46% of ties in Singapore's Twitter network) act as high-bandwidth shortcuts, substantially reducing path lengths between distant network regions.
   • Counterfactual analysis showed wormholes decrease average path length by 26% and accelerate information diffusion.
3. Robust cross-platform patterns:
   • Observed in both Twitter (public, multilateral) and phone networks (private, dyadic)
   • Consistent across diverse cultures (US, Europe, Asia, Africa)
   • Validated through multiple strength metrics: call duration/frequency, message reciprocity, emotional affect

Methodology

• Tie range: Defined as second-shortest path length between nodes if direct connection were removed
• Strength measurement:
  • Twitter: Reciprocated @mentions frequency and sentiment analysis
  • Phones: Call duration and frequency
• Analyzed within-individual (comparing same person's ties) and between-individual patterns

Implications

• Information diffusion: Challenges hub-centric models, showing peripheral nodes connected by wormholes can drive spread
• Social integration: Strong long-distance ties may facilitate cultural exchange and emotional contagion
• Network theory: Reveals limitations of small-world models based purely on weak ties

Unexplained Phenomena

1. Formation mechanisms:
   • May connect low-degree peripheral nodes
   • Could persist through social mobility (strong ties surviving network changes)
   • Potentially represent multidimensional relationships (e.g., religious/political ties separate from professional networks)
2. Content patterns:
   • Associated with social/recreational topics rather than work-related communication
   • More active during non-working hours
     The findings fundamentally reshape our understanding of social network dynamics, demonstrating that strong ties can simultaneously bridge large network distances while maintaining high bandwidth - a phenomenon with significant implications for predicting information spread and designing network interventions.

In-text annotations