An Experimental Test of How Selfies Change Social Judgments on Facebook

Abstract

Selfies are everywhere on social media. Research has focused only on who is posting selfies and has not addressed the audience members viewing selfies. This study aims to fill this gap by analyzing the judgments people make of selfies posted on Facebook. Using an online experiment, we test how including a selfie on a Facebook status update changes people's appraisals of narcissism, message appropriateness, and social attraction. We also consider how the valence and intimacy of the status update text interplay with the selfie to change social judgments. Participants rated posts with selfies as more narcissistic and inappropriate, and less socially attractive. Selfie evaluations also depended upon the valence and intimacy of the status update text. Gender of the selfie poster did not influence evaluation of posts. One implication from these results is that posting selfies on social media may lead to negative judgments about the poster.

Sharing selfies on social media has become a global phenomenon. In 2014, 55% of Millennials reported sharing a selfie.¹ A selfie is a self-taken self-portrait photograph, usually shared on social media.^2,3 Selfies help social media users present their identities.⁴ Photographs shared online preserve personal memories, help construct identities, and facilitate relationship building.⁵ Thus, selfie sharing has implications for people's identity and relationships.

Understanding how social media users interpret selfies is critical because of the invisible audience on social media. The complexities of audience management on social media mean that people often do not know who sees their posts.⁶ Not knowing who sees a post leaves ambiguity about impression management for posters. Despite the role audience play in online sharing, research primarily investigates selfie posters and not their audience.² Studying judgments people make about selfies posted on social media can determine the meaning audience members (e.g., friends, bosses, potential romantic partners, and family members) might attach to selfies. To determine the judgments people make about selfie posts on Facebook, we tested how judgments of Facebook status updates differ when selfies are included or not. We examined judgments of narcissism, appropriateness, and social attraction with potential moderators of message valence, message intimacy, and poster gender.

Selfies

Research on selfies tends to focus on their creators. Adolescent females post the most selfies on Instagram, but there is no difference in how men and women take selfies that are posted to Instagram.^7,8 Personality research on selfie posting has focused on narcissism as a predictor. Narcissism is defined as a “pervasive pattern of grandiosity, self-focus, and self-importance.”⁹ Scholars have theorized a connection between narcissism and selfie posting, as narcissism is associated with attention-seeking behavior and a fixation on physical attractiveness.^10,11 Numerous studies have found a positive correlation between narcissism and selfie posting on social media.^2,11–14 High levels of narcissism predicted greater involvement with responses to selfies.¹³

Initial research on selfie judgments found that people make personality judgments (e.g., extraversion) about selfie posters based on available physical cues.¹⁵ Research demonstrates that people are able to judge another person's narcissism level by viewing their photograph.¹⁰ We anticipate that posts containing selfies will signal to viewers that the poster focuses on their physical appearance more than posts without a selfie. We predict that people judge Facebook posts with selfies as more narcissistic than posts without selfies (H1).

Studies of disclosure interpretation frequently consider two other variables: (1) message appropriateness and (2) social attraction.¹⁶ Appropriate messages are those that fit the context in which they are shared.¹⁷ Social attraction is the degree to which an individual is perceived as likable.¹⁸ Drawing upon studies of self-disclosure interpretation in zero-history dyads, we review how including a selfie on social media can change perceived message appropriateness and social attraction.

Hess¹⁹ argues that sharing a selfie is considered intimate. Intimate self-disclosure is often viewed as appropriate in established relationships or other intimate situations, but frequently has negative repercussions outside personal relationships.²⁰ Intimate messages are considered inappropriate when shared through Facebook status updates.^21,22 A medium's social norms also influence the perceived appropriateness of a selfie. Research on photo sharing on Facebook suggests that posting positive pictures of friends is more appropriate than posting selfies.²³ Selfies are considered appropriate on Snapchat because the selfies are typically sent to strong ties and automatically deleted.^24,25 Facebook's norms suggest that audience members will consider selfies more inappropriate than other types of messages. Therefore, viewers should consider Facebook status updates that include selfies to be more inappropriate than status updates without selfies (H2).

Self-disclosure research has found that social attraction is determined by the perceived intimacy and appropriateness of a message. Offline, intimate disclosures predict more social attraction.^16,26 However, people tend to report less social attraction to disclosers of intimate messages on public social media, with perceived message appropriateness mediating this relationship.^20–22 If people interpret selfies in status updates as more intimate and inappropriate than nonselfie posts, then they should also report less social attraction to the poster. People should rate Facebook status updates with a selfie as less socially attractive than status updates without a selfie (H3).

Moderators of Selfie Judgments

Additional influences likely alter how people understand selfies. First, the text accompanying selfie posts must be considered.²⁷ Multimodality, or processing of both visual and textual information, complicates social judgments.²⁸ Research on self-disclosure interpretation has found that message intimacy and valence moderate judgments of the message and message discloser. For instance, college students rated posters as less socially attractive when they shared intimate and negative self-disclosures on Facebook, compared to when they shared nonintimate and positive disclosures.²¹ High levels of perceived disclosure intimacy and positivity on Facebook status updates also predicted greater feelings of conversational connectedness.²⁹ Therefore, there is reason to suspect that message intimacy and valence might interact with the presence of a selfie to change social judgments, prompting the following research question: How does the intimacy and valence of the self-disclosure accompanying a selfie change (a) narcissism, (b) appropriateness, and (c) social attraction judgments? (RQ1).

The gender of the selfie poster could impact selfie judgments. Previous studies have found that gender plays a critical role in selfie production; women post more selfies than men.⁸ Yet, narcissism predicts selfie posting on social media among men, but not women.^3,30 Given the findings about gender and selfies, gender may be important in understanding viewers' selfie judgments. We ask the following research question: Does the poster gender influence perceptions of (a) narcissism, (b) appropriateness, and (c) social attraction judgments of selfie posts? (RQ2).

Method

Undergraduate students at a large northeastern university participated in the study in exchange for extra credit in communication, information science, and computer science courses. Our sample (n = 115) was primarily female (70.4%) with a mean age of 20.8 (SD = 2.70). The majority were Caucasian (69.6%) or Asian (7.4%). Participants were required to have a Facebook account. Participants were 8.7% freshmen, 34.8% sophomores, 30.4% juniors, 18.3% seniors, 7.0% graduate, and 0.90% others. All procedures for this study were approved by the Institutional Review Board.

Piloting

Selfies were pilot tested for physical attractiveness and valence of facial expression before use as stimuli. A group of raters (n = 10) evaluated 30 selfies.³¹ Raters determined whether the photograph was a selfie and rated the physical attractiveness of the subject on a scale from −10 (not physically attractive) to 10 (physically attractive) and the facial expression in the photograph on a scale from −10 (strongly negative) to 10 (strongly positive). Analyses of these scores gave two male and two female selfies that were considered neutral in physical attraction and valence of facial expression.

Experiment Design

This study used a 2 × 2 × 2 design, with post intimacy as a between-subject factor and selfie condition and post valence as within-subject factors.³² Three variables were manipulated: selfie condition (selfie vs. no selfie), post valence (positive vs. negative), and post intimacy (high vs. low). Participants were presented with four different fictitious Facebook status updates that varied in valence and selfie presence. All status updates that participants saw were blocked on the same level of intimacy. In the selfie condition, participants saw a status update message that included one of the selfies selected from the pilot test. In the no selfie condition, participants saw the same status update without a selfie. Participants each saw status updates from two females and two males. Messages for status updates were pulled verbatim from Bazarova²¹; however, “Spring Break” was changed to “Fall Break” to fit with the cover story timeline. The stimuli were presented following a Latin square design, to randomize stimulus order and guarantee no participant saw the same experimental condition, profile, or message twice. Participants were told that the status updates were collected from students at a different university during Fall 2015.

Measures

Poster narcissism

Narcissism was operationalized using an adapted version of the narcissism subscale from the Dirty Dozen measure.³³ The four-question scale was rewritten to ask observers about the poster (e.g., “This person tends to want others to admire her” and “This person tends to want others to pay attention to him”). Ratings were collected on a five-point Likert scale (1 = Strongly Disagree; 5 = Strongly Agree), α = 0.77.

Message appropriateness

This measure consisted of four bipolar adjective scales about the participant's impression of the status update message.^21,34 Items were measured on a seven-point scale. Higher scores indicated greater perceived inappropriateness of sharing the message (e.g., “Appropriate: Inappropriate” and “Suitable for the situation: Unsuitable for the situation”), α = 0.89.

Poster social attractiveness

Participants reported their level of social attraction to the status update poster on a version of McCroskey and McCain's¹⁸ social attraction scale adapted for Facebook.²¹ Responses to questions (e.g., “I would like to have a friendly chat with this person” and “I would like to be this person's Facebook friend”) were on a seven-point Likert scale (1 = Strongly Disagree; 7 = Strongly Agree), α = 0.87.

Results

Analysis was conducted using multilevel modeling. Participants were modeled as random effects to control for the nonindependence in observations between the four profiles each participant rated.³⁵ Judgments of each profile were modeled as Level 1 variables. Participant-level characteristics (e.g., age) were modeled as Level 2 variables. Full models included main and all two- and three-way interaction effects for selfie condition, valence, and intimacy. We tested an interaction between profile gender and selfie condition. Our analysis controlled for participant age, gender, and Facebook use frequency because of other studies linking these variables to selfie posting.^9,30 Covariates are reported throughout the article when significant. Table 1 contains descriptive statistics for all dependent variables.

Table 1.

Correlations and Descriptive Statistics for Dependent Variables

	1	2	3	M (SD)
1. Inappropriateness	—			3.79 (1.54)
2. Narcissism	0.34^**	—		3.39 (0.76)
3. Social attraction	−0.58^**	−0.40^**	—	3.57 (1.16)

Numbers across the top row correspond with numbers for each variable in the first column (e.g., 1 = inappropriateness).

p < 0.01.

Manipulation check

A manipulation check found that participants correctly identified status updates with and without a photograph 98.5% of the time. Participants who failed the manipulation check were removed from analysis. Participants viewing negative status updates considered these posts more negative than positive status updates, F(1, 344) = 561.29, p < 0.001. Participants in the high intimacy condition rated messages as more intimate than messages in the low intimacy condition, F(1, 113) = 46.31, p < 0.001.

Poster narcissism

H1 predicted that participants would rate users who posted selfies as more narcissistic than those who did not post selfies. RQ1a questioned how the intimacy and valence of the post interacted with selfies to predict narcissism. To test H1 and answer RQ1a, we ran a model with all main effects as well as interaction effects between selfie condition, message valence, and message intimacy, ICC = 0.20, η² = 0.21. A significant main effect for selfie condition on narcissism ratings was found, F(1, 322.80) = 74.00, p < 0.001, η² = 0.19. Participants rated status updates with selfies as more narcissistic than the same posts without selfies (Table 2). Selfie condition did not interact with valence or intimacy to predict narcissism ratings, nor was there a three-way interaction. RQ2a asked how participants' narcissism judgments varied by poster gender. Profile gender was a significant predictor of narcissism rating, F(1, 322.97) = 8.19, p < 0.01. Female profiles (M = 3.45, SE = 0.06) were considered more narcissistic than male profiles (M = 3.29, SE = 0.06), p < 0.01, but there was no interaction between profile gender and selfie, F(1, 107.23) = 0.18, p > 0.05.

Table 2.

Main Effects for Selfie Posts

	Narcissism M (SE)	Inappropriateness M (SE)	Social attraction M (SE)
Selfie	3.62 (0.06)	4.06 (0.12)	3.34 (0.09)
No selfie	3.12 (0.06)	3.38 (0.12)	3.93 (0.09)

All means within each column are significantly different from each other at p < 0.001.

Message appropriateness

Consistent with H3, posts with selfies were considered more inappropriate than the same posts without selfies, F(1, 325.81) = 42.86, p < .001, η² = 0.09 (Table 2). We tested a model with interactions between selfie, valence, and intimacy to answer RQ1b, η² = 0.31, ICC = 0.28. No significant interaction for selfie was found. RQ2b addressed if gender of the profile would change perceived appropriateness, but no significant effect for profile gender was found, F(1, 325.84) = 1.32, p > 0. 05.

Social attractiveness of poster

We ran a final model to test how the experimental manipulations and interactions influenced perceived social attraction, η² = 0.24, ICC = 0.25. Supporting our prediction that posts with selfies would be less socially attractive to viewers (H3), there was a main effect for selfie condition on social attractiveness, F(1, 324.76) = 49.87, p < 0.001, η² = 0.11. Selfie posters were rated as less socially attractive than posters without selfies (Table 2). There was a significant interaction between selfie and valence (RQ1c), F(1, 324.67) = 4.67, p < 0.05. Positive posts without selfies received the highest attractiveness ratings. Negative messages with selfies received the lowest attractiveness ratings (Table 3). RQ2c asked how gender moderated social attraction. Male profiles were rated as less socially attractive than female profiles, F(1, 324.85) = 3.92, p < 0.05 (M_male = 3.713, SE = 0.09, M_female = 3.55, SE = 0.09), but there was no significant interaction between profile gender and selfie condition, F(1, 107.96) = 0.06, p > 0.05.

Table 3.

Social Attraction to Selfie Posters

	Selfie M (SE)	No selfie M (SE)
Negative post	3.15 (0.11)	3.56 (0.11)
Positive post	3.53 (0.11)	4.29 (0.11)

All means within each column are significantly different from each other at p < 0.001.

Discussion

This study expands research on how selfies are interpreted. Our findings indicate that including a selfie with a Facebook post alters interpretations of message appropriateness, perceived poster narcissism, and social attractiveness. Gender of the selfie poster did not influence judgments in our sample.

Our results contribute to research about the relationship between selfies and narcissism. While other studies have considered the role of narcissism in producing selfies,^12,13 we demonstrated that people are more likely to judge Facebook selfie posters to be narcissists than those who post messages without selfies. This finding held regardless of message valence or intimacy. Selfies may signal to audience members that the poster is self-focused and attention seeking.

Judgments of posters' social attractiveness were predicted by an interaction effect between post valence and selfie condition. Including a selfie on a post negatively impacted the social attractiveness of a poster. Posts that were positive and did not contain a selfie were considered the most socially attractive among our participants. Selfies with positive messages had social attraction ratings similar to negative posts without selfies. As in previous public intimacy research, people rate sharing negative intimate information on highly visible media as unattractive.²² Whether or not selfie posting reflects negative personality traits is currently a debate within selfie production research.^2,36 We demonstrate that zero-history audience members are more likely to view selfie posters negatively. Our narcissism and social attraction findings are curious considering that psychology research has demonstrated that narcissistic individuals tend to make positive first impressions.⁹ Thus, selfies may be judged erroneously by people lacking knowledge of the poster. These findings suggest that people looking to make friends or romantic partners online may want to avoid posting selfies, as this may lead to less social attraction.

Although selfies did not substantively change the message of the posts in our experiment, messages with selfies were interpreted as more inappropriate than those without selfies. A selfie poster may wish to connect with their audience or clarify their self-concept by sharing the photograph,⁴ but the message interpreted from their selfie may not match their intention. This misinterpretation could result in their post being perceived inappropriate to their audience. Selfie posts can be viewed by unintended audiences, and context collapse in social media amplifies the difficultly of managing message appropriateness among audiences.³⁷ Developers may consider implementing more advanced privacy settings for selfies to reduce exposure to negative evaluations. Algorithms could suggest altering privacy settings when selfies are included on Facebook.

RQ1 addressed multimodality's impact on perceptions of self-disclosures. We found that text self-disclosures on Facebook status updates were influenced by the inclusion of a selfie, predicting social attractiveness ratings. Attraction results suggest that studies of text or visual online content alone are insufficient because text interpretation content depends on images. Selfie judgments for narcissism and appropriateness were not dependent upon the message. One potential reason for the nonsignificant interactions is that people pay more attention to photographs compared to text on Facebook posts.³⁸ Participants' judgments may have been determined by the attention devoted to the selfie and not the status update message. Eye tracking may be necessary to determine how participants process multimodality.

RQ2 asked whether the selfie poster's gender would change social judgments of the message or the poster. Although there were significant differences for profile gender for narcissism and intimacy, neither interacted with selfie presence to alter evaluations. Research on selfie creation has found that selfie posting differs by gender,^8,11,30 but we found that evaluations of selfies did not differ by gender.

Limitations and future directions

Although this study contributes to research on selfies, we are aware of a few limitations. First, our experiment did not include a control photo for comparison against selfies. Although photographs were piloted to ensure that they were identifiable as selfies, we cannot say with complete certainty whether selfies change social judgments or if it is enough to include any photograph with a Facebook post. Second, message viewers were outside of the poster's social network. Perceptions would likely be different if the viewers were Facebook friends, with additional contextual information about the poster. Lack of knowledge about the selfie poster may have made the selfie seem out of place, prompting the negative appraisals. Third, judgments of strangers' selfies may be influenced by individuals' own selfie posting behavior; however, this study did not account for this potential moderator, and we cannot offer definitive conclusions about how participant selfie posting habits influenced the results. Finally, our study only focused on Facebook. Thus, our findings can only be generalized to social media sites with similar affordances and social norms. Applications that require photos in posts, such as Instagram, may yield different results.

This work is an initial step toward understanding perceptions of selfies. Further work needs to consider specific characteristics of selfies, such as differences in physical attractiveness, facial expressions, or solo versus group selfies. Nonselfie images (e.g., scenery and objects) could be compared to selfies to differentiate the effects of selfies on social judgments. Future research could investigate the between-channels differences for selfie judgments and how people make judgments based on selfie-posting patterns (i.e., frequency) rather than single posts. These findings suggest that there is still much to learn about how people perceive selfies.

Footnotes

Acknowledgment

This work is supported, in part, by the National Science Foundation through IIS-1405634.

Author Disclosure Statement

No competing financial interests exist.

References

Taylor

. More than half of Millennials have shared a “selfie.” Pew Research Center 2014. www.pewresearch.org/fact-tank/2014/03/04/more-than-half-of-millennials-have-shared-a-selfie

Fox

, Rooney

. The Dark Triad and trait self-objectification as predictors of men's use and self-presentation behaviors on social networking sites. Personality and Individual Differences, 2015; 76:161–165.

Sorokowski

, Sorokowska

, Oleszkiewicz

, et al. Selfie posting behaviors are associated with narcissism among men. Personality and Individual Differences, 2015; 85:123–127.

Van House

. (2007) Flickr and public image-sharing: distant closeness and photo exhibition. In CHI’07 extended abstracts on human factors in computing systems. New York: ACM Press, pp. 2717–2722.

Van House

. Personal photography, digital technologies and the uses of the visual. Visual Studies, 2011; 26:125–134.

Litt

. Knock, knock. Who's there? The imagined audience. Journal of Broadcasting and Electronic Media, 2012; 56:330–345.

Bruno

, Bertamini

, Protti

. Selfie and the city: a world-wide, large, and ecologically valid database reveals a two-pronged side bias in naïve self-portraits. PLoS One, 2015; 10:e0124999.

Souza

, de las Casas

, Flores

. (2015) Dawn of the selfie era: the whos, wheres, and hows of selfies on Instagram. In Proceedings of the 2015 ACM Conference on online social networks. New York: ACM Press, pp. 221–231.

Back

, Schmukle

, Egloff

. Why are narcissists so charming at first sight? Decoding the narcissism-popularity link at zero acquaintance. Journal of Personality and Social Psychology, 2010; 98:132–145.

10.

Vazire

, Naumann

, Rentfrow

, et al. Portrait of a narcissist: manifestations of narcissism in physical appearance. Journal of Research in Personality, 2008; 42:1439–1447.

11.

Weiser

. #me: narcissism and its facets as predictors of selfie-posting frequency. Personality and Individual Differences, 2015; 86:477–481.

12.

Kim

, Lee

, Sung

, et al. Predicting selfie-posting behavior on social networking sites: an extension of theory of planned behavior. Computers in Human Behavior, 2016; 62:116–123.

13.

Lee

, Sung

. Hide-and-seek: narcissism and “selfie”-related behavior. Cyberpsychology, Behavior, and Social Networking, 2016; 19:347–351.

14.

Sorokowska

, Oleszkiewicz

, Frackowiak

, et al. Selfies and personality: who posts self-portrait photographs?. Personality and Individual Differences, 2016; 90:119–123.

15.

Qiu

, Lu

, Yan

, et al. What does your selfie say about you?. Computers in Human Behavior, 2015; 52:443–449.

16.

Altman

, Taylor

. (1973). Social penetration: the development of interpersonal relationships. New York, NY: Holt, Rinehart and Winston.

17.

Spitzberg

. Preliminary development of a model and measure of computer-mediated communication (CMC) competence. Journal of Computer-Mediated Communication, 2006; 11:629–666.

18.

McCroskey

, McCain

. The measurement of interpersonal attraction. Speech Monographs, 1974; 41:261–266.

19.

Hess

. The selfie assemblage. International Journal of Communication, 2015; 9:1629–1646.

20.

Baruh

, Cemalcılar

. Rubbernecking effect of intimate information on Twitter: when getting attention works against interpersonal attraction. Cyberpsychology, Behavior, and Social Networking, 2015; 18:506–513.

21.

Bazarova

. Public intimacy: Disclosure interpretation and social judgments on Facebook. Journal of Communication, 2012; 62:815–832.

22.

Orrego

, Smith

, Mitchell

, et al. (2000). Disclosure and privacy issues on television talk shows. In Petronio

, ed. Balancing the secrets of private disclosures. Mahwah, NJ: Erlbaum, pp. 249–259.

23.

McLaughlin

, Vitak

. Norm evolution and violation on Facebook. New Media and Society, 2012; 14:299–315.

24.

David

. (2015). All what we send is selfie: images in the age of immediate reproduction. In Carvalheiro

, Telleria

, eds. Mobile and digital communication: approaches to public and private. Covilha, Portugal: LabCom Books, pp. 79–100.

25.

, Chang

, Welker

, et al. (2016). Automatic archiving versus default deletion: what Snapchat tells us about ephemerality in design. In Proceedings of CSCW’ 16. New York: ACM Press, pp. 1662–1675.

26.

Collins

, Miller

. Self-disclosure and liking: a meta-analytic review. Psychology Bulletin, 1994; 3:457–475.

27.

Mitchell

WJT

. There are no visual media. Journal of Visual Culture, 2005; 4:257–266.

28.

Giese

, Baden

. Putting the image back into the frame: modeling the linkage between visual communication and frame-processing theory. Communication Theory, 2015; 25:46–69.

29.

Utz

. The function of self-disclosure on social network sites: not only intimate, but also positive and entertaining self-disclosures increase the feeling of connection. Computers in Human Behavior, 2015; 45:1–10.

30.

Sorokowski

, Sorokowska

, Frackowiak

, et al. Sex differences in online selfie posting behavior predict histrionic personality scores among men but not women. Computers in Human Behavior, 2016; 59:368–373.

31.

Walther

, Van Der Heide

, Kim

S-Y

, et al. The role of friends' appearance and behavior on evaluations of individuals on Facebook: are we known by the company we keep?. Human Communication Research, 2008; 34:28–49.

32.

Jackson

, Jacobs

. Generalizing about messages: suggestions for design and analysis of experiments. Human Communication Research, 1983; 9:169–191.

33.

Jonason

, Webster

. The dirty dozen: a concise measure of the dark triad. Psychological Assessment, 2010; 22:420–432.

34.

Canary

, Spitzberg

. Appropriateness and effectiveness perceptions of conflict strategies. Human Communication Research, 1987; 14:93–118.

35.

Hayes

. A primer on multilevel modeling. Human Communication Research, 2006; 32:385–410.

36.

Senft

, Baym

. What does the selfie say? Investigating a global phenomenon. Introduction. International Journal of Communication, 2015; 9:1588–1606.

37.

Marwick

, Boyd

. Networked privacy: how teenagers negotiate context in social media. New Media and Society, 2014; 16:1051–1067.

38.

Lim

, Fussell

. Understanding how people attend to and engage with foreign language posts in multilingual newsfeeds. Proceedings of the Eleventh International AAAI Conference on Web and Social Media 2017; 588–591.