Purpose This study rigorously explores how social media functions as a supportive mental health resource and a potential risk factor, focusing on diverse and marginalized populations. Drawing on three theoretical frameworks – the technology acceptance model (TAM), self-determination theory (SDT) and social influence theory – the research examines how algorithmic biases, platform features and cultural norms intersect to shape anxiety, depression and self-esteem. Through this lens, this study aims to inform evidence-based strategies that address digital literacy, foster ethical artificial intelligence (AI) implementation and accommodate global cultural nuances. Ultimately, the goal is to promote healthier online ecosystems that balance social media’s benefits with necessary safeguards. Design/methodology/approach Based on a mixed-methods approach, the study combined quantitative and qualitative data. A survey of 500 participants (aged 18–30) used standardized instruments (GAD-7, PHQ-9 and Rosenberg Self-Esteem) to measure anxiety, depression and self-esteem, with hierarchical regression revealing key correlations. Concurrently, 50 semi-structured interviews were purposively sampled for demographic diversity, and probed experiences of social comparison, emotional validation and algorithmic influences. The thematic analysis of transcripts and quantitative findings enabled methodological triangulation, enhancing robustness and interpretive depth. This design follows Creswell’s mixed-methods guidelines, providing statistically significant trends and richly contextualized insights into how social media engagement affects mental health across various populations. Findings Quantitative results showed that individuals who spent more than three daily hours on social media were significantly more likely to report anxiety and depression, with social comparison explaining 65% of the variance in self-esteem scores. Women, LGBTQ+ users and those from low-income regions faced compounded mental health risks because of intensified algorithmic biases. Qualitative interviews underscored themes of idealized self-presentation, privacy concerns and cultural pressures linked to body image and identity. Nonetheless, many participants also described social media as a vital source of community support. These findings highlight social media’s complex, bidirectional influence on mental well-being. Research limitations/implications Although the mixed-methods design strengthened the study’s validity through triangulation, certain limitations remain. The reliance on self-reported measures introduces potential response biases, including social desirability and recall inaccuracies. In addition, as the survey was conducted online, the sample may not fully represent individuals with limited digital access, which could exclude some at-risk or marginalized populations. The cross-sectional nature of the research also limits causal interpretation, as it captures a snapshot in time rather than longitudinal trends. Future research should incorporate longitudinal or experimental designs, and expand the sample to include older adults, neurodivergent individuals and participants from underrepresented regions. Nevertheless, the study offers evidence-based implications for designing targeted digital literacy programs, improving algorithmic transparency and enhancing psychosocial support features on social media platforms. Practical implications The findings support implementing culturally contextualized digital literacy programmes to help users recognize algorithmic manipulation, manage social comparison and mitigate cyberbullying. Educational curricula could integrate mental health awareness, equipping students with coping strategies and media literacy skills. Regulatory bodies and technology firms should collaborate to refine AI policies, minimizing biases that disproportionately affect marginalized groups. Platforms can promote a healthier online environment by fostering transparent design principles and user-friendly mental health resources. Governments, NGOs and tech companies could invest in initiatives – such as self-help apps and moderated support communities – empowering individuals to engage with social media more constructively. Social implications Addressing social media’s dual impact can substantially reduce mental health disparities and enhance societal cohesion. By acknowledging cultural norms and intersectional vulnerabilities, interventions can promote inclusivity, ensuring that women, LGBTQ+ communities and low-income users receive tailored support. Community-led campaigns may bolster resilience by normalizing open discussions on mental health, thereby weakening stigma and encouraging help-seeking. In addition, transparent algorithms and equitable moderation can restore user trust. Strengthening digital citizenship fosters empathy and responsible engagement, cultivating environments where diverse identities can thrive. A collective commitment to ethical design and inclusive policy can reshape online spaces into supportive communal resources. Originality/value This study distinguishes itself by integrating TAM, SDT and social influence theory to holistically evaluate social media’s mental health implications across multiple cultural and demographic contexts. By pairing quantitative measures of anxiety, depression and self-esteem with thematic insights into users’ lived experiences, it transcends one-dimensional analyses. The unique emphasis on marginalized groups – affected by both algorithmic biases and cultural pressures – underscores the necessity of intersectional, ethical approaches. Consequently, this research offers a robust framework for policymakers, educators and platform developers seeking to balance innovation with user well-being, setting a benchmark for future global studies on digital platforms and mental health.
Read full abstract