Involuntary Embarrassing Exposures in Online Social Networks: A Replication Study

Mohammadreza Ebrahimi, Juan-Manuel MartínezAIS 2019 · DOI

Abstract

In this study, we carry out a methodological replication of the research done by Choi et al. (2015) published in Information Systems Research. In the original study, the authors integrate the privacy and teasing literatures under a social exchange framework to understand online involuntary exposures. The original study was conducted on students from Southeast Asia. Our study uses a significantly larger sample of college students in the United States. Our replication results show that whereas most of the hypotheses supported by the original results on behavioral responses replicate with high consistency (8 out of 12 hypotheses), the results that deal with the effects of network commonality on perceived privacy invasion and perceived relationship bonding did not replicate (4 out of 12 hypotheses). These results could stem from a failed manipulation of network commonality. We look into the possible rationales for this and show what would be an effective manipulation in our context. Further, we expand the original study by testing an additional embarrassing scenario catered to our subject pool. The results suggest that perceived privacy invasion and perceived relationship bonding affect individual’s behavioral responses to embarrassing exposures.

Introduction

Online Social Networks (OSNs) have fundamentally altered the landscape of social interaction, introducing new forms of interpersonal friction. Among these is "involuntary embarrassing exposure," a phenomenon where a user is depicted in a compromising or silly manner by a third party—for instance, a friend posting a photo of one sleeping in class or during a night of heavy drinking. To understand how individuals navigate these exposures, Choi et al. (2015) proposed a model rooted in Social Exchange Theory. This framework posits that users weigh the "costs" of such exposures, primarily perceived privacy invasion, against the "benefits," such as perceived relationship bonding or "friendly teasing."

The original study identified two primary drivers of these perceptions: Information Dissemination (operationalized by tagging) and Network Commonality (the degree of mutual friend overlap). However, the digital landscape has shifted significantly since the original data collection in Southeast Asia circa 2010. This replication study seeks to determine if these theoretical constructs remain valid in a contemporary Western context, specifically among college students in the United States, where social media norms and friend counts have evolved drastically.

Methodology

This research constitutes a methodological replication, maintaining the core experimental structure of the 2015 study while expanding the scope and sample size to ensure statistical power.

Experimental Design

The study employed a 2x2 factorial experiment. Participants were presented with mock-up Facebook posts where they were the subject of an embarrassing narrative. The researchers manipulated two variables:

  1. Information Dissemination: High (the target is tagged) vs. Low (no tag).
  2. Network Commonality: High (many mutual friends) vs. Low (few mutual friends).

To ensure cultural relevance, the authors utilized the original "sleeping in class" scenario and introduced a new "heavy drinking on a Friday night" scenario. The latter was hypothesized to be more evocative and embarrassing for the target demographic of US undergraduates.

Participants and Data Collection

The primary replication involved 266 business students from a large university in the American Southwest, a sample nearly triple the size of the original study (n=109). To investigate the failure of certain manipulations, a follow-up survey of 198 students established modern norms for Facebook friend counts. Finally, a separate group of 88 participants was used to evaluate "scaled" thresholds for network commonality.

Analytical Approach

ANOVA was used to assess how the experimental manipulations influenced perceptions of privacy and bonding. Subsequently, logistic regression mapped these perceptions to four distinct behavioral responses: Inaction, Transactional Avoidance (unfriending/blocking), Interpersonal Avoidance (reporting the post), and Approach (liking or commenting).

Key Insights and Results

The replication yielded a nuanced set of results, supporting 8 out of 12 original hypotheses.

The "Friendship Inflation" Phenomenon

The most significant divergence from the original study was the failure of the "Network Commonality" manipulation. In 2010, 65 mutual friends served as a robust signal for a "high" overlap. In the 2018 US context, however, 65 mutual friends were perceived as a relatively low number. Data from the follow-up surveys revealed that the average US student in the sample had approximately 588 friends. Consequently, the threshold required to trigger a sense of "high commonality" had shifted to approximately 250 mutual friends. This "friendship inflation" led to the failure of hypotheses H1 and H2, which linked mutual friend counts to perceptions of privacy and bonding.

Robustness of Behavioral Logic

Despite the failure of the specific network triggers, the internal logic of the social exchange model remained remarkably consistent. Hypotheses H3 and H4, which connect perceived privacy invasion and relationship bonding to behavioral outcomes, replicated with high consistency. When a user perceived a high degree of privacy invasion, they were significantly more likely to engage in avoidance behaviors. Conversely, when the exposure was perceived as a form of bonding, users were likely to "approach" the post. This suggests that while the digital "triggers" are moving targets, the underlying psychological framework of the theory is sound.

Cultural and Contextual Nuances

The study also highlighted the importance of scenario relevance. The "sleeping in class" scenario, which was effective in the original Southeast Asian sample, failed to elicit a strong emotional response among US students. The "drinking" scenario proved necessary to generate the variance required for measurement, suggesting that what constitutes "embarrassment" is culturally and temporally dependent. Furthermore, the replication found that Dispositional Privacy Concern—an individual's general tendency to value privacy—was a significant predictor of "Transactional Avoidance" (unfriending), a factor not emphasized in the original findings.

Conclusion and Future Work

This replication study underscores both the durability of social exchange logic in OSNs and the volatility of experimental manipulations in rapidly evolving digital environments. The findings suggest that researchers must regularly recalibrate objective measures (like friend counts) to match contemporary social norms.

Future research should move beyond simple friend counts as a proxy for network closeness. Utilizing more sophisticated metrics from graph theory, such as centrality or closeness, may provide a more stable measure of network commonality that is less susceptible to inflation. Additionally, as social media usage shifts toward platforms like Instagram and TikTok, the dynamics of tagging and audience reach will require further investigation. Finally, while scenario-based surveys provide valuable insights, future studies should aim to capture real-time behavioral data to bridge the gap between stated intent and actual social media behavior.

Summary updated

Abstract

In this study, we carry out a methodological replication of the research done by Choi et al. (2015) published in Information System Research. In the original study, the authors integrate the privacy and teasing literatures under a social exchange framework to understand online involuntary exposures. The original study was conducted on students from Southeast Asia. Our study uses a significantly larger sample of college students in the United States. Our replication results show that whereas most of the hypotheses supported by the original results on behavioral responses replicate with high consistency (8 out of 12 hypotheses), the results that deal with the effects of network commonality on perceived privacy invasion and perceived relationship bonding did not replicate (4 out of 12 hypotheses). These results could stem from a failed manipulation of network commonality. We look into the possible rationales for this and show what would be an effective manipulation in our context. Further, we expand the original study by testing an additional embarrassing scenario catered to our subject pool. The results suggest that perceived privacy invasion and perceived relationship bonding affect individual’s behavioral responses to embarrassing exposures.
BibTeX
@inproceedings{ebrahimi2019involuntary,
  author = {Mohammadreza Ebrahimi and Juan-Manuel Martínez},
  title = {{Involuntary Embarrassing Exposures in Online Social Networks: A Replication Study}},
  booktitle = {AIS},
  year = {2019},
  doi = {10.17705/1atrr.00039},
}
Figure