Inferring norms from numbers: Boomerang effects of online virality metrics on normative perceptions and behavioral intention

https://doi.org/10.1016/j.tele.2019.101279Get rights and content

Highlights

  • We examined the effects of online virality metrics on behavioral intentions.

  • We conducted an online experiment involving young adults aged 18–35.

  • Virality metrics decreased injunctive norms supporting condom use and HIV testing.

  • Virality metrics may reduce the persuasive impact of online health messages.

Abstract

Online virality metrics have been shown to influence information processing and behavioral outcomes. This study examined how virality metrics, in terms of the number of online shares, impacted the persuasiveness of a health campaign message in influencing viewers’ evaluation of the message, their perceived injunctive norms, descriptive norms, and preventative behavioral intentions. We conducted a 4 by 2 factorial online experiment (N = 621), featuring an HIV prevention video accompanied by a virality metric (i.e., high, medium, low, and no metric) and a video poster (i.e., demographic similarity versus dissimilarity). Results showed that virality metrics did not influence message evaluation. However, the presence of metrics decreased viewers’ perceived injunctive norms supporting condom use and HIV testing, which further decreased the corresponding behavioral intentions. The similarity between the viewer and the poster did not moderate the effects of virality metrics. These findings suggest a boomerang effect of virality metrics in reducing the persuasive impact of health messages. Theoretical explanations and practical implications are discussed.

Introduction

Social media are increasingly playing a vital role in promoting public health and positive behavior changes (Chou et al., 2009, Korda and Itani, 2013, Kreps and Neuhauser, 2010). Compared to the large volume of research examining content features (Dunn et al., 2017, Lyson et al., 2018, Surian et al., 2016), theoretical and empirical works on understanding the persuasive effects of online social informational cues (e.g., shares, likes, comments) surrounding health messages are rather limited. One promising line of research has started to examine virality metrics, the aggregated counts of online shares for a piece of information (Alhabash et al., 2015, Walther and Jang, 2012). A few studies have shown messages with a large number of online shares are perceived to be of higher quality and credibility (Chung, 2017, Knobloch-Westerwick et al., 2005, Lee and Sundar, 2013, Messing and Westwood, 2014, Sundar et al., 2007). Health messages accompanied by high numbers of online shares have also been shown to be more persuasive (Chung, 2019, Kim, 2018b, Lee-Won et al., 2016, Lee-Won et al., 2017).

The core underlying theoretical processes concern how people infer others’ opinions and behaviors from their shared online messages. Similar to how people can develop normative perceptions by observing others’ opinions and behaviors in offline settings, they may also be able to infer such norms by interpreting others’ motivations behind sharing a certain message online. Social normative perceptions are categorized into injunctive norm, prescribing others’ approval or disapproval of a behavior, and descriptive norm, reflecting the perception of whether others perform the behavior (Cialdini et al., 1991, Rimal and Real, 2005). Injunctive and descriptive norms are often related but can show different predictions on behavioral intention (Smith et al., 2012). For instance, regarding socially stigmatized topics, people’s expressed opinions toward a behavior (e.g., supporting HIV testing) may not align with their actual behavior (e.g., getting tested for HIV). So far, only two studies showed empirical evidence suggesting high virality metrics could lead to increased injunctive norm, specifically for supporting a health behavior (Chung, 2019, Lee-Won et al., 2016). It remains underexplored whether high virality metrics can also translate into the related normative perception that many other people are practicing the behavior. The central question motivating this research is thus whether online virality metrics are interpreted to signal the sharers’ own behaviors in addition to their approval, and thus lead to simultaneous changes in the viewers’ descriptive norm and injunctive norm, and subsequent behavioral intentions.

In addition, we consider the potential role of source similarity – the similarity between the viewer and the message poster – in moderating these effects. Research has shown that source similarity leads to greater engagement with the message (Kim et al., 2016) and can elicit stronger normative influence (Centola, 2011, Centola, 2013), thus it is possible that the strengths of social inferences about others will be dependent upon the extent of source similarity. Given the dual presence of visible sources and anonymous virality metrics on social media messages, it is important to explore whether and how they interactively influence viewers’ judgement of the message and their related behavioral determinants in a health context.

We focus on HIV-risk reduction for young adults as the research context because young adults are experiencing an increasing rate of HIV infection (Centers for Disease Control and Prevention, 2018), and are the most avid users of social media (Perrin and Jiang, 2018). Understanding how they process virality metrics has practical implications for developing effective strategies to reduce HIV risks using social media messaging.

In the following sections, we first review previous literature on theorizing virality metrics as a heuristic cue and a persuasive cue through normative inferences, as well as theoretical predictions regarding source similarity. We then report our methods, describing the online experiment, data collection, and analytical procedure. Next, we report all results according to research hypotheses and questions. Lastly, we discuss the findings’ theoretical and practical implications, along with limitations and directions for future research.

Section snippets

Virality metrics as heuristic cues

Typically displayed as numbers alongside online information, virality metrics indicate the real-time popularity of the information (Metzger et al., 2010, Sundar, 2008). Theoretically, virality metrics are a specific type of online social informational cues that signify the aggregated numbers of other users’ reactions to a piece of information (e.g., shares, likes, comments) (Walther and Jang, 2012). Since the advent of online technology that can aggregate and show collective ratings, scholars

Study design and manipulations

This study was reviewed and approved by the authors’ institutional review board. An online experiment was conducted through a website developed by the researchers. The study employed a 4 (high, medium, low, or no metrics) by 2 (source similarity or dissimilarity) between-subject design.

After indicating consent on the study website, participants completed a pre-survey assessing demographic backgrounds, then viewed the experimental webpage, and lastly completed a post-survey. The webpage mimicked

Sample characteristics

The majority of participants were female (58.8%), with a mean age of 27.9 (SD = 4.3). About 74.9% were White, 8.9% were Asian, 8.4% were Black, 7.7% were Latino, and the remaining 0.2% were Indian. About 31.4% were married and 53.5% obtained a bachelor’s degree or higher. About 50.7% of the participants identified as having liberal political views, and roughly 58.1% of participants reported having any religious beliefs.

Manipulation checks

To ensure internal validity, we only included participants who correctly

Discussion

This study aims to examine the causal effects of virality metrics in impacting the persuasiveness of health messages in influencing viewers’ evaluation of the message, their perceived injunctive norms, descriptive norms, and subsequent behavioral intentions. This study adds new empirical findings and theoretical discussions to the line of research on virality metrics, and more broadly, on online social informational cues.

Conclusion

Our study systematically tested the effects of virality metrics on message evaluation and related behavioral determinants in the context of HIV risk reduction. Our findings added to previous literature by showing that virality metrics did not impact message evaluation and perceived descriptive norm. The presence of virality metrics, regardless of magnitude, may reduce young viewers’ perceived injunctive norms in support of HIV preventive behaviors, suggesting that health practitioners may need

Funding statement

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgement

We thank The Elizabeth Taylor AIDS Foundation for providing the campaign video for this study.

References (70)

  • S. Alhabash et al.

    From clicks to behaviors: the mediating effect of intentions to like, share, and comment on the relationship between message evaluations and offline behavioral intentions

    J. Interact. Advertis.

    (2015)
  • H. Allcott et al.

    Trends in the diffusion of misinformation on social media

    Res. Polit.

    (2019)
  • R. Aswani et al.

    Exploring content virality in Facebook: a semantic based approach

  • R. Aswani et al.

    Detection of spammers in Twitter marketing: a hybrid approach using social media analytics and bio inspired computing

    Informat. Syst. Front.

    (2018)
  • J.N. Bailenson et al.

    Facial similarity between voters and candidates causes influence

    Public Opin. Quart.

    (2008)
  • F.M. Bass

    A new product growth for model consumer durables

    Manage. Sci.

    (1969)
  • T.C. Brock

    Communicator-recipient similarity and decision change

    J. Pers. Soc. Psychol.

    (1965)
  • M. Buhrmester et al.

    Amazon's Mechanical Turk: a new source of inexpensive, yet high-quality, data?

    Perspect. Psychol. Sci.

    (2011)
  • Centers for Disease Control and Prevention. 2018. Estimated HIV incidence and prevalence in the United States,...
  • D. Centola

    An experimental study of homophily in the adoption of health behavior

    Science

    (2011)
  • D. Centola

    Social media and the science of health behavior

    Circulation

    (2013)
  • S. Chaiken

    Heuristic versus systematic information processing and the use of source versus message cues in persuasion

    J. Pers. Soc. Psychol.

    (1980)
  • Z. Chen et al.

    When, why, and how controversy causes conversation

    Journal of Consumer Research

    (2013)
  • W.-Y.S. Chou et al.

    Addressing health-related misinformation on social media

    JAMA

    (2018)
  • W.Y.S. Chou et al.

    Social media use in the United States: implications for health communication

    J. Med. Inter. Res.

    (2009)
  • A.H. Eagly et al.

    The Psychology of Attitudes

    (1993)
  • M. Fishbein et al.

    Predicting and Changing Behavior: The Reasoned Action Approach

    (2011)
  • A.J. Flanagin et al.

    Connecting with the user-generated Web: how group identification impacts online information sharing and evaluation

    Inform. Communic. Soc.

    (2014)
  • A.F. Hayes

    Introduction to Mediation, Moderation, and Conditional Process Analysis: A Regression-Based Approach

    (2017)
  • G.A. Heeren et al.

    Theory-based predictors of condom use among university students in the United States and South Africa

    AIDS Educ. Prev.

    (2007)
  • W. Hong et al.

    Does animation attract online users' attention? The effects of flash on information search performance and perceptions

    Inform. Syst. Res.

    (2004)
  • J.B. Jemmott et al.

    Predicting intentions to use condoms among African-American adolescents: the theory of planned behavior as a model of HIV risk-associated behavior

    Ethn. Dis.

    (1992)
  • J.W. Kim

    Rumor has it: the effects of virality metrics on rumor believability and transmission on Twitter

    New Med. Soc.

    (2018)
  • M. Kim et al.

    Effect of character-audience similarity on the perceived effectiveness of anti-smoking PSAs via engagement

    Health Commun.

    (2016)
  • S. Knobloch-Westerwick et al.

    Impact of popularity indications on readers' selective exposure to online news

    J. Broadcast. Electron. Media

    (2005)
  • Cited by (3)

    • Who else likes it? Perceived gender of social endorsers predicts gender equality support

      2021, Computers in Human Behavior
      Citation Excerpt :

      Higher social endorsement metrics on social media messages about bone marrow donation also increased the perceived norm that others would think highly of joining a bone marrow donation registry, which then contributed to greater intentions to actually join that registry (Lee-Won et al., 2016). While Calabrese and Zhang (2019) found that endorsement manipulations attached to a health video produced null or opposite of expected effects, they suggest that this may be due to the strength of the video overpowering the heuristic or the endorsed message not being advocated comparably by close peers. Though this work demonstrates how social endorsement metrics can influence norms and subsequent actions, it typically does not distinguish between ingroup and outgroup norms.

    1

    The two authors had equal contributions to this manuscript.

    View full text