Table of Contents
Type: Literature Review | Subject: Psychology | Level: Undergraduate | Word Count: ~3,200 words
This model Literature Review was produced by an Essays UK specialist as reference material for learning purposes only. For support in this field, see our psychology writers.
Overview: This model literature review examines the relationship between social media use and anxiety in adolescents. It organises a body of conflicting empirical findings into three themes, critically appraises the methodological limitations that explain the inconsistency, and identifies a clear gap for future longitudinal and person-specific research.
Adolescence is a developmental period characterised by heightened sensitivity to peer evaluation, active identity formation and increased emotional reactivity. It is also the life stage during which anxiety disorders most commonly first emerge, with the majority of lifetime cases having their onset before early adulthood. Over the past two decades, the everyday lives of young people in the United Kingdom and comparable high-income countries have been reshaped by near-constant access to smartphones and social media platforms such as Instagram, TikTok, Snapchat and X (formerly Twitter). The convergence of a developmentally vulnerable window with an unprecedented shift in how adolescents communicate, compare themselves to others and spend their leisure time has generated sustained public and clinical concern that social media may be contributing to rising levels of adolescent anxiety.
This concern is not confined to the research community. Campaign reports, parliamentary discussion and press coverage have frequently presented social media as a primary driver of a youth mental health crisis. Yet the academic literature is markedly more equivocal. While some studies report meaningful associations between heavy or particular types of social media use and anxiety symptoms, others find associations so small as to be of questionable practical significance, and a number find no reliable relationship at all once confounding variables are controlled. The field is therefore characterised less by a clear answer than by a pattern of conflicting findings that resist simple summary.
Several theoretical mechanisms have been proposed to link social media use with anxiety, and these help to frame the empirical debate. The displacement hypothesis holds that time spent on social media displaces activities that protect mental health, notably sleep, physical activity and face-to-face interaction; sleep disruption in particular is a well-established correlate of anxiety. A longitudinal study by Boers, Afzali and Conrod (2019) reported within-person temporal associations between higher screen and social media use and increased anxiety symptoms across adolescence, a pattern consistent with such a displacement account. A contrasting social compensation perspective proposes that social media may benefit anxious or socially isolated young people by offering accessible channels for support and connection. These opposing predictions, harm through displacement versus benefit through compensation, recur throughout the literature and partly explain why findings differ according to which population and which type of use is examined.
The purpose of this review is to synthesise the empirical evidence on the relationship between social media use and adolescent anxiety, to make sense of why findings diverge, and to identify where the most important uncertainties remain. Rather than treating the literature as a single question with a single answer, the review adopts a thematic structure that groups studies according to the aspect of social media use they examine. Three themes are developed: the frequency and duration of use; the distinction between passive and active engagement, including social comparison processes; and the individual differences and contextual factors that appear to moderate any effect. Following the thematic synthesis, the review critically appraises the methodological features that account for the inconsistent evidence base before setting out a research gap that future work should address. For the purposes of this review, anxiety is treated broadly, encompassing both diagnosed anxiety disorders and self-reported subclinical anxiety symptoms, as the primary studies vary in how they operationalise the construct.
An illustrative search was conducted to identify peer-reviewed studies examining the relationship between social media use and anxiety or related internalising symptoms in adolescent and young adult samples. Three electronic databases were searched, supplemented by backward citation searching of relevant systematic reviews. The search combined terms relating to the exposure (social media, social networking sites, screen time), the population (adolescent, teenager, youth, young people) and the outcome (anxiety, internalising symptoms, psychological distress, well-being). Boolean operators and truncation were used to broaden retrieval, and results were limited to English-language sources published between 2012 and 2023 to reflect the period in which smartphone-based social media became dominant among young people. The parameters applied are summarised in the table below.
| Parameter | Detail |
|---|---|
| Databases searched | PsycINFO, PubMed, Web of Science |
| Example search string | (“social media” OR “social networking” OR “screen time”) AND (adolescen* OR teenager* OR youth) AND (anxiety OR “internalising” OR “psychological distress”) |
| Date range | January 2012 to December 2023 |
| Inclusion criteria | Peer-reviewed; quantitative or systematic review; mean sample age 11–25; validated anxiety or well-being measure |
| Exclusion criteria | Clinical samples only; non-social-media screen use (e.g. gaming); opinion pieces and editorials |
| Records retained for synthesis (illustrative) | 14 key sources, including two systematic reviews and one umbrella review |
The resulting set was deliberately weighted towards systematic reviews, meta-analyses and large-sample studies, as these provide the most reliable basis for a short thematic synthesis and allow the conflicting single studies they aggregate to be placed in context. The retained sources are not presented as an exhaustive mapping of the field but as a representative selection sufficient to illustrate the main currents of debate and the methodological tensions within them.
The retained literature was organised into three themes. The first concerns the simplest and most widely studied variable, the amount of time spent using social media. The second moves beyond quantity to the manner of use, distinguishing passive consumption from active participation and foregrounding social comparison. The third addresses the growing recognition that effects are not uniform across individuals.
The earliest and most intuitive line of enquiry asks whether adolescents who spend more time on social media report more anxiety. A number of large cross-sectional studies have reported positive associations consistent with this expectation. Twenge and Campbell (2018), analysing a large population-level dataset, found that higher screen and social media use was associated with lower psychological well-being among children and adolescents, with heavier users reporting more frequent symptoms of low mood and worry. Vannucci, Flannery and Ohannessian (2017) similarly reported that greater daily social media use was associated with higher trait anxiety in a sample of emerging adults, while Primack et al. (2017) found that the use of a larger number of different platforms, rather than total time alone, predicted elevated symptoms of depression and anxiety.
The intuitive appeal of a simple dose–response relationship is also reflected in influential non-academic reports. The Royal Society for Public Health (2017), surveying young people about their experiences, concluded that several popular platforms were perceived to worsen anxiety, sleep and body image, and the report attracted considerable media attention. Such survey-based accounts capture genuine subjective experience and have done much to shape public debate, but they rely on self-perceived impact rather than controlled measurement, and they sit uneasily alongside the more cautious findings of the academic studies considered below.
However, the picture becomes considerably more complicated when the strength of these associations is examined rather than merely their direction. Orben and Przybylski (2019), using a rigorous analytic approach across several large datasets, demonstrated that the negative association between digital technology use and adolescent well-being, while statistically detectable, was extremely small, accounting for a trivial proportion of the variance in well-being and comparable in magnitude to factors such as wearing glasses. Huang (2017), in a meta-analysis of time spent on social networking sites, likewise reported only a weak negative relationship with well-being. These findings suggest that the quantity of use, treated as a single global measure, is a poor and possibly misleading predictor of anxiety. The implication drawn by several authors is that time spent is too blunt an instrument: it collapses qualitatively different activities, from supportive messaging with close friends to passive scrolling through idealised content, into a single number that obscures more than it reveals. This recognition motivates the second theme.
A substantial strand of research argues that how adolescents use social media matters more than how long they use it. Verduyn et al. (2017), in a critical review, drew a now widely cited distinction between passive use, defined as consuming others’ content without direct interaction, and active use, defined as directly communicating and contributing content. Their synthesis concluded that passive use tends to undermine subjective well-being, whereas active use, particularly direct exchanges with existing friends, can enhance it by strengthening social connection. The proposed mechanism for the negative effect of passive use is upward social comparison: by scrolling through curated, favourable depictions of peers’ lives, adolescents may conclude that their own experiences are comparatively inadequate, generating envy, self-doubt and anxious rumination.
Experimental and intensive longitudinal evidence lends some support to this account. Kross et al. (2013), using experience sampling with young adults, found that more frequent passive Facebook use predicted subsequent declines in moment-to-moment and overall well-being, a temporal pattern that is difficult to explain by well-being driving usage. Nesi and Prinstein (2015) focused specifically on comparison and feedback-seeking behaviours, reporting that adolescents who engaged more heavily in social comparison and sought reassurance through social media reported more depressive symptoms, a relationship that was stronger for girls and for less popular adolescents. This finding is important because it reframes the question: the concern is not the platform as such but a particular psychological process, social comparison, that social media affords and intensifies.
Related constructs extend this mechanistic account. The fear of missing out, an apprehension that others are enjoying rewarding experiences from which one is absent, has been linked to more intensive and more anxious patterns of social media engagement, and it offers a plausible pathway from constant connectivity to heightened worry. Reassurance-seeking provides a further route: adolescents who post content and then repeatedly check for likes and comments may experience transient relief followed by renewed anxiety, a pattern reminiscent of the maintenance cycles described in cognitive models of anxiety disorders. These mechanisms share a common feature in that they locate the risk not in exposure to a screen as such but in the appraisal and evaluative processes that particular uses of social media provoke.
Even here, however, the evidence is not uniform. The passive–active distinction has been criticised for oversimplifying a continuum of behaviours, and some studies have failed to replicate a clean separation between harmful passive use and beneficial active use. Certain forms of active use, such as posting content and then anxiously monitoring the quantity of likes and comments it attracts, plausibly heighten rather than relieve anxiety, blurring the supposed boundary. The theme therefore refines, rather than resolves, the debate: it shifts attention productively towards mechanisms, but the mechanisms themselves appear to operate differently depending on who is using the platform and in what emotional state, which leads directly to the third theme.
The most recent and arguably most important development in the literature is the recognition that average effects conceal substantial variation between individuals. Beyens et al. (2020) conducted an intensive longitudinal study in which the same measures were taken repeatedly from each adolescent, allowing within-person effects to be estimated separately for each participant. They found that the effect of social media use on well-being was positive for some adolescents, negative for others and negligible for a third group. Crucially, these divergent individual patterns largely cancelled out when averaged, producing the small overall effects reported by studies such as Orben and Przybylski (2019). This result offers a compelling reconciliation of the conflicting literature: the modest average associations are not evidence that social media is harmless but may instead reflect the aggregation of genuinely harmful effects for a vulnerable minority with neutral or beneficial effects for others.
Consistent with this, several moderators have been proposed. Gender appears relevant, with a number of studies, including Nesi and Prinstein (2015), reporting stronger associations between comparison-oriented use and internalising symptoms among adolescent girls. Pre-existing vulnerability also matters: adolescents already prone to anxiety or with low self-esteem may be more likely both to engage in upward comparison and to be adversely affected by it, raising the possibility of a reinforcing cycle. Valkenburg, Meier and Beyens (2022), in an umbrella review synthesising prior reviews, concluded that the overall evidence points to small and highly heterogeneous effects, and argued explicitly that the field should move away from asking whether social media is good or bad for adolescents in general towards asking for whom, under what conditions and through which mechanisms effects occur. Keles, McCrae and Grealish (2020), in a systematic review focused on depression, anxiety and distress, reached a broadly compatible conclusion, noting that associations were observed but that study quality and measurement varied widely, limiting firm inference.
A further complication concerns the direction of influence. Several authors have argued that the relationship is likely to be reciprocal rather than unidirectional: anxious adolescents may gravitate towards passive, comparison-heavy use as a way of monitoring perceived social threats, and this use may in turn deepen their anxiety, establishing a self-reinforcing cycle. If such bidirectional dynamics are common, they would further undermine any causal reading of simple cross-sectional correlations and reinforce the argument that only repeated-measures designs can disentangle cause from consequence. The heterogeneity documented by person-specific studies and the likelihood of reciprocal effects together suggest that the long-standing search for a single average effect may, in retrospect, have been the wrong objective.
Taken together, the three themes reveal an evidence base that is extensive but methodologically constrained in ways that explain much of its apparent inconsistency. The most pervasive limitation is reliance on cross-sectional, correlational designs. The majority of the associations reported, including those in several of the studies reviewed above, are measured at a single point in time and therefore cannot establish temporal order, let alone causation. A positive correlation between social media use and anxiety is equally compatible with three interpretations: that use increases anxiety; that anxious adolescents use social media more, perhaps as a form of avoidance or reassurance-seeking; or that a third variable, such as poor sleep, family conflict or underlying temperament, drives both. The studies that do employ longitudinal or experience-sampling designs, notably Kross et al. (2013) and Beyens et al. (2020), are therefore disproportionately valuable, yet they remain comparatively few.
A second limitation concerns measurement. Social media use is most often assessed through self-reported estimates of time spent, which correlate only weakly with objectively logged usage and are subject to recall bias. Anxiety, likewise, is measured inconsistently across studies, ranging from validated clinical scales to single-item well-being questions, which undermines direct comparison and may partly explain divergent findings. Odgers and Jensen (2020) drew specific attention to these measurement weaknesses, cautioning that strong public claims about harm have outpaced the quality of the underlying data. A third issue is the heavy reliance on Western, and frequently North American, samples, which limits confidence that findings generalise to other cultural contexts, including the specific media ecology of the United Kingdom.
A further and more subtle problem concerns analytic flexibility. Large datasets of the kind used to study this question contain many possible measures of both social media use and well-being, and many defensible ways of combining them. Orben and Przybylski (2019) illustrated this directly by showing that the apparent strength of the association could be made to vary substantially depending on which analytic choices were made, implying that selective reporting of favourable specifications could inflate published effects. This concern, combined with the tendency of striking positive findings to attract publication and media coverage more readily than null results, suggests that the literature may collectively overstate the association even as individual small studies remain inconclusive.
Finally, the field has been shaped by small effect sizes that are nonetheless of potential public health significance. As Orben and Przybylski (2019) acknowledged, even a small average effect, when distributed across an entire adolescent population of heavy users, could correspond to a non-trivial number of affected individuals. The challenge of interpreting small, heterogeneous effects, rather than the absence of effects, is arguably the central methodological puzzle of this literature, and it cautions against both the alarmist and the dismissive readings that have characterised public debate.
The synthesis and appraisal above point to a coherent and well-defined gap. The dominant questions in the early literature, concerning whether social media use is associated with anxiety and whether more use is worse, have been substantially answered: associations exist but are small, inconsistent and poorly captured by crude measures of time. The pressing and under-addressed questions are now about direction, mechanism and heterogeneity. Specifically, there is a shortage of adequately powered longitudinal and intensive within-person studies capable of establishing whether social media use precedes increases in anxiety, of isolating the specific behaviours, such as upward social comparison, that mediate any effect, and of identifying the characteristics that distinguish adolescents who are harmed from those who are unaffected or helped.
Three features would strengthen future research. First, designs should prioritise within-person, longitudinal measurement over between-person snapshots, as the work of Beyens et al. (2020) demonstrates the explanatory power this offers. Second, exposure should be measured more precisely, ideally combining objectively logged platform data with assessments of the type and emotional quality of use rather than relying solely on estimated time. Third, research should be conducted with United Kingdom and more culturally diverse adolescent samples, given the preponderance of North American data and the possibility that platform norms and mental health contexts differ across settings. Addressing this gap would move the field beyond the unproductive question of whether social media is simply good or bad and towards the more clinically useful goal of identifying which adolescents are at risk and why.
Equally, future studies would benefit from clearer conceptual and measurement work on the anxiety outcome itself. Much existing research conflates transient worry, social anxiety and generalised anxiety, yet these may relate to social media use through different mechanisms; social anxiety, for example, might be expected to show a particularly close relationship with the comparison-based and evaluative features of social platforms. Disaggregating the outcome, alongside improving the exposure measure, would allow more precise hypotheses to be tested and would help to translate broad population-level associations into findings of genuine clinical and preventive value.
This review set out to synthesise the conflicting evidence on social media use and adolescent anxiety and to locate the most important remaining uncertainties. Organised thematically, the literature shows a clear trajectory. Early studies focusing on the quantity of use reported positive but very small associations with anxiety, and more rigorous analyses revealed these effects to be weaker than public concern implies. Attention consequently shifted to the manner of use, where the distinction between passive consumption and active engagement, and in particular the role of upward social comparison, offered a more promising account of how harm might arise. Most recently, person-specific research has shown that average effects mask substantial individual variation, suggesting that social media is genuinely detrimental to a vulnerable minority even where its population-level effect appears negligible.
The central message of the evidence is therefore one of heterogeneity rather than uniform harm or harmlessness. The principal barriers to firmer conclusions are methodological: a reliance on cross-sectional designs, inconsistent and self-reported measurement, and limited cultural range. The clearest priority for future work is a programme of longitudinal, within-person studies using precise exposure measures and diverse samples, capable of answering not whether social media affects adolescent anxiety but for whom, through which mechanisms and under what conditions. Until such evidence accumulates, measured interpretation, attentive to individual differences, remains more defensible than the confident generalisations that have dominated public discussion.
You May Also Like
All work is written by human writers. 100% AI free, guaranteed.
100% money back guarantee if you find plagiarism in our work.
COMPANY DETAILS
