Good Fences Make Good Neighbors: What’s Acceptable (and What’s Not) in Neighborhood Video Surveillance

Paper Title

Good Fences Make Good Neighbors: What's Acceptable (and What's Not) in Neighborhood Video Surveillance

Publication Info

• Topic area: Privacy norms and management in neighborhood video surveillance.
• Keywords: Privacy, Contextual Integrity Theory, video doorbells, group privacy, neighborhood surveillance, social norms, smart home devices, information flows, security, community dynamics.

Background and Problem

• Problem / challenge: Existing research on video surveillance largely focuses on the privacy of primary users, neglecting the broader implications for neighbors, visitors, and bystanders. Privacy norms in neighborhood contexts remain underexplored, particularly regarding the sharing of video doorbell footage.
• Significance: Understanding privacy norms in neighborhoods is critical for designing technologies and policies that respect collective privacy and security needs, especially as smart home devices become more prevalent.
• Motivation and related work: Prior work has explored privacy as individual control over data flows but fails to address collective contexts like neighborhoods. Traditional models do not account for the networked nature of smart home devices, where privacy is shaped by multiple actors. This study applies Contextual Integrity Theory to analyze group-level privacy norms in residential neighborhoods.

Solution

• Proposed approach: A combinatorial matrix survey grounded in Contextual Integrity Theory to systematically analyze the acceptability of video footage sharing in neighborhood contexts.
• Novelty:
  1. Identification of informational norms specific to neighborhoods, including variations by individual and neighborhood factors.
  2. Extension of Contextual Integrity Theory to group-level contexts, emphasizing collective privacy management.
  3. Adaptation of the combinatorial matrix survey method to quantify variations in privacy norms across diverse scenarios.
  4. Policy and design recommendations for safeguarding privacy beyond device owners.
• Procedure and key techniques:
  • Developed a survey using a combinatorial matrix approach to evaluate 342 unique information flow combinations.
  • Participants rated the acceptability of sharing video footage across different subjects, recipients, and transmission principles.
  • Conducted ordinal logistic regression and interaction analyses to identify key factors influencing acceptability.

Results

• Concrete findings:
  • Transmission principles (e.g., safety, gossip) were the strongest predictors of acceptability, explaining 18.3% of the variance.
  • Sharing footage for safety purposes was most acceptable (7.9x more likely to be rated acceptable), while gossip and entertainment were least acceptable.
  • Recipients like the police were viewed more favorably, while sharing to Facebook groups was generally unacceptable.
  • Subjects like delivery drivers and bystanders were more surveillable, while social class distinctions influenced acceptability in nuanced ways.
• Advantage over baselines: The study highlights the contextual and socially situated nature of privacy norms, offering a more nuanced understanding than traditional individual-focused models.
• Experiments / evaluation:
  • Sample: 247 UK participants (mean age 40, 43.7% video doorbell owners).
  • Metrics: Acceptability of information flows rated on a 5-point Likert scale.
  • Analysis: Ordinal logistic regression with fixed effects and interaction terms.
• Limitations and future work:
  • Limited generalizability due to UK-specific sample and recruitment via Prolific.
  • Did not explore three-way interactions or real-world sharing behaviors.
  • Future work should include larger, more diverse samples and investigate additional scenarios and factors.

Summary

This study investigates privacy norms surrounding video doorbell footage sharing in residential neighborhoods using Contextual Integrity Theory and a combinatorial matrix survey. It finds that privacy management is socially situated, with transmission principles (e.g., safety, gossip) being the strongest determinants of acceptability. Individual and neighborhood factors, such as social class and community closeness, further shape these norms. The findings extend privacy theory to group-level contexts and provide actionable recommendations for device design, neighborhood social media moderation, and policy. This work underscores the importance of considering collective privacy in networked environments.