AUTHENTIC SELF-EXPRESSION ON SOCIAL MEDIA IS ASSOCIATED WITH GREATER SUBJECTIVE WELL-BEING

The study examines whether expressing oneself authentically versus in a self-idealized manner on social media influences subjective well-being. Building on debates that social media can promote idealized, unrealistic self-presentations versus authentic extensions of offline identities, the authors focus on personality as a core aspect of the self. They hypothesize that greater authentic self-expression—operationalized as alignment between self-reported Big Five personality and observed social media behavior—will be positively associated with Life Satisfaction. They also explore whether this relationship depends on how socially desirable an individual’s personality profile is, testing the proposition that people with more socially desirable traits might benefit more from authenticity.

Two contrasting perspectives frame prior research: (1) a self-idealization view suggesting social media fosters idealized, curated, and exaggerated self-presentations, and (2) an authenticity view proposing social media, particularly platforms with offline ties like Facebook, reflect users’ actual personalities. Prior work often relies on self-reported authenticity (state feelings or judgments of honesty/consistency), which are vulnerable to valence and social desirability biases. The authors argue for a behavioral measure, proposing that authenticity can be quantified as the alignment between self-view and observable self-expression, and note that deviations from one’s self-view on social media are typically in the direction of self-enhancement rather than self-deprecation, though desirable directions may be idiosyncratic to one’s network norms.

Study 1 (Observational): Data from 10,560 Facebook users who completed Big Five personality assessments and Life Satisfaction via the myPersonality app (2007–2012). Authenticity was quantified by comparing self-reported Big Five traits with computer-predicted traits from two observable sources: Facebook Likes (N ≈ 9273) and Facebook status update language (N ≈ 3215). Personality predictions used established machine-learning approaches (e.g., 10-fold cross-validated LASSO for Likes; cross-validated regression using language features including words, n-grams, and topics). Quantified Authenticity was primarily computed as the inverse Euclidean distance between the five self-reported and predicted trait scores; robustness was examined using alternative distance/similarity metrics (Manhattan, Euclidean, correlation similarity, cosine similarity). Regressions predicted Life Satisfaction from Quantified Authenticity with and without controls. Controls included self-reported Big Five traits and a personality extremeness index (sum of absolute z-scores across traits), with robustness checks including age and gender. Additional analyses separated normative self-enhancement (rating oneself higher than behavior indicates on socially desirable traits) from self-deprecation.

Study 2 (Pre-registered longitudinal experiment): 90 university students/social media users (Mage ≈ 22.98; 22.72% female) completed a two-week crossover intervention with randomized order. Each participant completed two 7-day posting periods: one instructed to post authentically and one to post in a self-idealized way. Instructions were personalized using feedback from a personality pre-screen (IPIP/BFI-25) and asked participants to plan specific posting strategies aligned with either their actual personality (authentic) or desired impression (idealized). Subjective well-being was measured at baseline (t0), after week 1 (t1), and after week 2 (t2): Life Satisfaction (SWLS), positive and negative affect, and general mood. Between-subject comparisons at t1 examined condition differences; within-subject comparisons contrasted each participant’s authentic vs. idealized week outcomes. Manipulation checks assessed perceived posting authenticity per week.

Study 1: Quantified Authenticity was positively associated with Life Satisfaction in models without controls and remained significant with controls for self-reported Big Five and personality extremeness. The association replicated across independent sources (Likes and language) and across multiple distance/similarity metrics, reaching significance in 11 of 16 model specifications. Effects did not consistently depend on personality profile, providing little support that individuals with more socially desirable traits benefited disproportionately. Analyses distinguishing discrepancy direction suggested normative self-enhancement was negatively related to well-being, whereas normative self-deprecation showed no effect.

Study 2: Between-subjects comparisons after week 1 showed no significant differences between authentic and self-idealized conditions on well-being outcomes. Within-subjects comparisons indicated better well-being during the authentic week: higher mood (mean difference = 0.19 [0.003, 0.374], t = 2.02, d = 0.43, p = 0.046), higher positive affect (mean difference = 0.17 [0.012, 0.318], t = 2.14, d = 0.45, p = 0.035), and marginally lower negative affect (mean difference = -0.20 [-0.419, 0.016], t = -1.84, d = -0.39, p = 0.069). No significant within-subject effect was found for Life Satisfaction (mean difference = 0.09, p = 0.342). Manipulation checks suggested between-subject instructions alone did not markedly shift perceived authenticity after week 1, but within-person contrasts showed participants felt more authentic in the authentic-posting week (mean difference = 0.30 [0.044, 0.556], t = 2.33, d = 0.49, p = 0.022).

Findings support the hypothesis that more authentic self-expression on social media is linked to greater subjective well-being. Behaviorally quantified authenticity—alignment between self-view and observable expression—correlates with higher Life Satisfaction in a large observational dataset and causally improves short-term affective states (mood, positive affect, lower negative affect) in a randomized crossover experiment. The benefits of authenticity did not reliably vary by personality profile, suggesting that authenticity may aid well-being broadly rather than primarily for those with socially desirable traits. These results inform debates on social media’s impact on well-being by emphasizing that effects depend on how people use social platforms; authentic expression appears beneficial, whereas self-enhancing deviations from one’s self-view relate to poorer well-being. The experimental results indicate that affective components of well-being are more sensitive to short-term changes in posting strategies than the more stable cognitive evaluation of Life Satisfaction.

The paper introduces a behavioral metric of authenticity—Quantified Authenticity—as the alignment between self-reported and externally inferred personality from social media behavior. Across a large-scale observational study and a pre-registered longitudinal experiment, the authors show that authentic self-expression is associated with higher Life Satisfaction and causally improves short-term mood and affect relative to self-idealized posting. The work advances understanding of when social media use may be beneficial and suggests that encouraging authentic expression can improve users’ well-being. Future research should compare authentic use to no use, investigate motives and awareness underlying authenticity deviations, and further quantify contextual moderators that influence the authenticity–well-being link.

• Generalizability: Study 1 focuses solely on effects among social media users and cannot determine whether authentic social media use is better than abstaining from social media altogether.
• Mechanisms: The studies do not disentangle why people deviate from authenticity (e.g., lack of self-insight vs. intentional impression management), which may moderate well-being effects.
• Effect sizes: Observational effects are small but robust; while experimental effects are larger for affective outcomes, overall magnitudes are modest relative to major life factors (e.g., health, income).
• Experimental constraints: The pre-registered sample size was reduced due to COVID-19 lab shutdown; several effects were marginal, and between-subject manipulation in week 1 did not reliably shift behavior or perceived authenticity, limiting between-group inferences.
• Measurement nuances: Life Satisfaction, a stable cognitive evaluation, may be less responsive to short-term interventions than affective measures, potentially underestimating impacts on broader well-being.