What Life Events are Disclosed on Social Media, How, When, and By Whom?

Title of the Paper

What Life Events are Disclosed on Social Media, How, When, and By Whom?

Paper Information

• Subject Areas: Human-Computer Interaction, Social Computing, Psychology
• Keywords: Social Media, Life Events, Self-Disclosure, Audience, Individual Differences, Surveys, Self-Presentation

Research Background and Questions

• Identified Problem: Although social media has become a common platform for sharing personal life events, there is still limited understanding of which life events users disclose on social media, how they disclose them, and the differences between individuals. This lack of understanding also limits social media platforms' ability to design features for managing sensitive life events.
• Significance: Life events (e.g., career transitions, health changes) can have profound psychological impacts, influencing individuals' social support, emotional connections, and mental health. Understanding the phenomenon of online disclosure of these events can help enhance the user experience on social media platforms.
• Research Motivation: There is a need to bridge the gap between social media data and traditional survey data (e.g., self-report questionnaires) to explore the factors influencing users' decisions to reveal or conceal their life events on social media.

Proposed Solution

• Methodology:
  • Developed a standardized coding scheme based on the PERI Life Events Scale to identify life event disclosures on social media.
  • Built regression models using Facebook data and related survey responses from 236 participants to analyze the relationship between life event disclosures and individual/event characteristics.
• Innovations:
  • Defined a systematic coding framework for the process of life event disclosure.
  • Combined social media data with self-report questionnaires to propose a theoretical framework explaining social media disclosures and their deviations.
• Implementation Steps:
  1. Data Collection: Full-year Facebook data (14,202 posts) and participants' self-report questionnaire data.
  2. Data Annotation: Manually annotated life events using the coding scheme.
  3. Model Construction: Used individual and event characteristics as covariates to explore life event disclosure patterns.
  4. Experiments and Analysis: Quantitatively compared life event records on social media and in surveys.

Research Findings

• Key Results:
  • Life events disclosed on social media primarily focus on positive and anticipated events, such as vacations and relationships, while major, recent, and intimate events are more likely to be disclosed via surveys.
  • Only 14% of Facebook posts explicitly expressed life events, whereas 100% of survey responses contained related records.
  • Social media users tend to share emotional and expressive content but avoid disclosing negative or sensitive events, such as health losses or work-related issues.
  • Certain life events (e.g., health improvements, family gatherings) are more likely to be disclosed on social media, while others (e.g., work stress, financial changes) are more likely to be reported through surveys.
• Advantages Over Existing Solutions:
  • Provided a comprehensive and fine-grained classification of life event disclosures.
  • Conducted in-depth analysis of how individual differences and specific event characteristics influence disclosure behaviors, addressing gaps in existing literature.
• Experimental Evaluation Results:
  • High annotation agreement (Fleiss κ=0.71).
  • Regression models revealed that personality traits (e.g., agreeableness and extraversion), event emotions (e.g., positivity), and event characteristics (e.g., intimacy) significantly influence disclosure behaviors.
• Limitations and Future Directions:
  • Data sources were limited to Facebook, which may not fully represent other social media platforms.
  • Lacked direct "ground truth" data as a benchmark.
  • Did not deeply explore why individuals choose to disclose or not disclose specific events. Future research could incorporate interviews to uncover motivations.
  • Focused only on public posts, excluding private messages or other data formats such as images.
  • Future work could explore multi-platform data integration and comparison, as well as include a broader user sample to reduce selection bias.

Conclusion

This study combines quantitative research models with natural language annotation to investigate the phenomenon of life event disclosure on social media. The research highlights the importance of individual and event characteristics in shaping disclosure behaviors and offers suggestions for improving social media design to enhance sensitivity to life events. It also emphasizes the need to address data privacy and ethical concerns.