Comprehensive Study Guide: Social Media, Body Image, and Disordered Eating Behaviors (Sanzari et al. 2023)

Bibliographic Information and Research Context

• Full Title: The impact of social media use on body image and disordered eating behaviors: Content matters more than duration of exposure.

• Authors: Christina M. Sanzari, Sasha Gorrell, Lisa M. Anderson, Erin E. Reilly, Martha A. Niemiec, Natalia C. Orlo, Drew A. Anderson, and Julia M. Hormes.

• Publication Details: Published in Eating Behaviors (Eat Behav.), April 4, 2023; Vol 49:101722.

• Institutional Affiliations:
    - National Library of Medicine (NLM)/National Institutes of Health (NIH).
    - University in the northeastern United States.

• Identifiers: DOI: $10.1016/j.eatbeh.2023.101722$, PMCID: $PMC10363994$, NIHMSID: $NIHMS1914473$, PMID: $37060807$.

• Funding Acknowledgments:
    - National Science Foundation Graduate Research Fellowship [Grant No. $1645421$] awarded to Christina Sanzari.
    - National Institutes of Mental Health (NIMH) grants: $K23MH126201$ (Dr. Gorrell), $K23MH123910$ (Dr. Anderson), and $K23MH131871$ (Dr. Reilly).

Introduction and Theoretical Framework

• Expansion of Social Media (SM): SM use has become ubiquitous. In $2021$, $72\,\%$ of U.S. adults used at least one SM site, compared to $60\,\%$ in $2014$. By $2021$, Americans subscribed to an average of $7$ SM accounts.

• The Influence of COVID-19: The pandemic accelerated SM use, particularly among young adults, due to social distancing requirements.

• Mental Health Associations: While SM provides social interaction and peer support, it is linked to negative mental health outcomes, including body image disturbances and disordered eating (ED).

• Previous Research Discrepancies:
    - Studies by Ferguson et al. ($2014$) and Meier & Gray ($2014$) found no association between duration of SM use and eating pathology.
    - Meta-analyses by Saiphoo & Vahedi ($2018$) noted that the "valence" of the relationship between SM and body image dissatisfaction is unclear due to inconsistent operationalization of SM use (e.g., measuring duration versus intensity).

• Impact of Content Specifics: Research suggests that engagement with appearance-focused content (e.g., "thinspiration" and "fitspiration") has significant adverse effects. These types of content often promote objectification, restrictive eating, and weight loss.

• The Shift to Image-Based Platforms: Use has shifted toward Instagram, Snapchat, and TikTok. TikTok, in particular, became the world's most accessed domain in $2021$, with approximately $136.5$ million American users in $2022$.

Body Positivity and Body Neutrality

• Body Positivity (#BoPo): A movement aimed at challenging the "thin ideal" by encouraging body acceptance across all shapes and sizes. Studies like Stevens & Griffiths ($2020$) suggest exposure to body-positive content may increase body satisfaction.

• Body Neutrality: Promotes the perspective that the body is worthy of respect regardless of its appearance. Unlike body positivity, there is limited research on its relationship with disordered eating.

• Concerns with Content: Content analyses (e.g., Lazuka et al., $2020$) show that most individuals in "body positive" posts still conform to mainstream beauty ideals, potentially harming viewer body image despite the intended message.

Study Aims and Hypotheses

• Aim 1: Compare patterns in SM use, body image, and ED behaviors between two undergraduate cohorts ($2015$ vs. $2022$), including the impact of COVID-19.
    - Hypothesis 1a: The $2022$ sample will report longer daily SM use across more accounts compared to $2015$.
    - Hypothesis 1b: The $2022$ sample will attribute SM use changes to the COVID-19 pandemic.
    - Hypothesis 1c: The $2022$ sample will show greater body image dissatisfaction and ED behaviors.

• Aim 2: Examine if the association between SM use and outcomes is moderated by the specific content consumed (weight loss vs. body positivity/neutrality).
    - Hypothesis 2a: Weight loss content will moderate the association between SM duration/platform count and adverse outcomes.
    - Hypothesis 2b: Body positivity/neutrality content will have a protective effect, moderating SM use to result in fewer adverse outcomes.

Methodology and Materials

• Participants: Undergraduate students at a large university in the northeastern U.S. recruitment occurred in two waves: $2015$ ($n = 359\text{ to }362$, via SurveyMonkey) and $2022$ ($n = 267\text{ to }280$, via Qualtrics).

• Demographics Variables: Age, gender identity (Female, Male, Other), race/ethnicity, and BMI ($\text{weight in kg} / (\text{height in m})^2$).

• Primary Measures:
    1. Body Appreciation Scale (BAS): $13$-item unidimensional scale; $1 = \text{never}$ to $5 = \text{always}$. Internal consistency (Cronbach's $\alpha$): $0.94$ ($2015$) and $0.95$ ($2022$).
    2. Fear of Negative Appearance Evaluation Scale (FNAES): $6$-item factor measuring concerns about negative evaluation. Cronbach's $\alpha$: $0.94$ ($2015$) and $0.96$ ($2022$).
    3. Eating Disorder Examination – Questionnaire (EDE-Q): Individual items assessing frequency (days out of the last $28$) of binge eating, vomiting, and laxative use.

• Exclusion Criteria: Impossible BMI values (< 10\,kg/m^2) or SM usage reports (> 24\,h/day).

Results: Cohort Comparison (Aim 1)

• Demographic Consistency: No significant differences in mean age ($19.38\,yrs$ in $2015$ vs. $19.05\,yrs$ in $2022$, $p = .06$) or BMI ($24.10\,kg/m^2$ vs. $24.78\,kg/m^2$, $p = .06$).

• Social Media Patterns:
    - Daily Minutes: Increased from $M = 199.14\,min$ ($2015$) to $M = 244.06\,min$ ($2022$) ($F(1, 610) = 15.89$, p < .001).
    - Platform Use Shift:
      - Facebook: $92.8\,\%$ ($2015$) to $53.8\,\%$ ($2022$).
      - Snapchat: $23.4\,\%$ ($2015$) to $89.5\,\%$ ($2022$).
      - TikTok: Reported by $82.9\,\%$ in $2022$ (not recorded in $2015$).
      - YouTube: $57.9\,\%$ ($2015$) to $84.4\,\%$ ($2022$).

• COVID-19 Impact: $64.2\,\%$ of the $2022$ sample reported using SM more since the pandemic; $11.0\,\%$ reported a decrease.

• Outcome Differences:
    - BAS Scores: Lower in $2022$ ($M = 4.88$) than $2015$ ($M = 5.14$).
    - FNAES Scores: Higher in $2022$ ($M = 2.76$) than $2015$ ($M = 2.40$).
    - Purging Behaviors: Frequency of vomiting ($F = 4.09$, p < .04) and laxative use ($F = 5.16$, p < .02) was significantly higher in the $2022$ cohort.

Results: Content Moderation (Aim 2 - 2022 Sample)

• Exposure Prevalence:
    - Weight loss (WL) content: Seek by $22.9\,\%$ ($n = 63$).
    - Body positivity (BP) content: Seek by $31.6\,\%$ ($n = 87$).
    - Females were significantly more likely to seek both WL and BP content (p < .05).

• Impact of Weight Loss Content: Exposure predicted lower body appreciation ($t = -3.37$, p < .003), higher fear of negative evaluation ($t = 4.02$, p < .003), and more frequent binge eating ($t = 3.58$, p < .003).

• The Paradox of Body Positivity/Neutrality: No significant protective effects found. Neither BAS, FNAES, nor binge eating frequency was significantly improved by consuming body-positive content.

• Key Interaction: Binge eating frequency was high only when respondents reported BOTH high SM duration AND exposure to weight loss content.

Discussion and Practical Implications

• Content vs. Duration: The study concludes that type of content matters more than total time spent on SM. High duration was mostly predictive of binge eating only when interactive with weight loss content.

• Intervention Targets: Interventions should focus on the quality of content engagement and literacy rather than just limiting screen time.

• The Failure of Body Positivity Content: The lack of benefit suggests that even positive content may trigger appearance comparisons, or that individuals with high body acceptance pay more attention to idealized images.

• Limitations:
    - Cross-sectional design prevents causal claims (do people with EDs seek weight loss content, or does the content cause the ED?).
    - Body positivity and body neutrality were measured together, though they may have different psychological impacts.
    - Sample homogeneity (mostly White and female) limits generalizability.

• Data Skewness: EDE-Q behavioral count data (vomiting, laxatives) was heavily skewed toward zero, requiring cautious interpretation of those specific frequencies.