Navigating The Pandemic Online: How Social Media Coping Patterns Affect Mental Health

Critical Analysis of Existing Latent Class Applications

Winstone et al. (2022) conducted a rigorous person-centered analysis of social media use among UK adolescents during the pandemic, identifying four user types—High Communicators, Moderate Communicators, Broadcasters, and Minimal Users—and found that Broadcasters exhibited the poorest mental health outcomes over time. While its longitudinal design offers valuable insights into the stability of these classifications, several limitations restrict broader applicability and theoretical advancement. The exclusive focus on 13- to 14-year-olds limits generalization to adults, whose coping needs and social stressors differ substantially from adolescents driven by identity development and peer relationships. Additionally, the study assessed communication frequency and content sharing but did not examine coping motivations or stress regulation goals underlying these behaviors, thereby revealing what users do without clarifying why they do it. Finally, its reliance on social validation and attention-seeking frameworks, without integrating established stress and coping theories, constrains understanding of which user profiles might be more adaptive under varying crisis conditions, underscoring the need for theory-driven research linking behavioral patterns to psychological adjustment.

Methodological Gaps in Existing Research

Guelmami et al. (2022) applied latent class analysis to identify well-being profiles among Tunisian Facebook users, demonstrating the cross-cultural applicability of person-centered approaches (McCabe et al., 2024; Pi et al., 2024). However, their analysis focused on general well-being indicators rather than specific coping behaviors or crisis-related outcomes, limiting relevance for understanding optimal social media use during acute stressor periods.

The study's platform-specific focus on Facebook users also constrains generalizability to contemporary social media environments where individuals typically engage across multiple platforms with different affordances and content ecosystems (Primack et al., 2017). Modern social media coping likely involves complex cross-platform behaviors that single-platform analyses cannot adequately capture.

Theoretical and Empirical Contributions of the Present Study

The present research advances person-centered social media research through several critical innovations that address limitations in existing literature (Keum et al., 2023). First, this study integrates established stress and coping theory with person-centered methodology to identify user types based on theoretically meaningful coping strategy combinations rather than purely behavioral indicators. This approach enables examination of why different user patterns prove more or less adaptive during crisis conditions. Second, the research focuses specifically on adult populations during active crisis conditions, addressing the significant gap in understanding how working-age adults use social media for stress management when facing unprecedented disruptions to employment, family responsibilities, and social support systems (McKeighen et al., 2023). Third, the study examines coping strategy combinations across multiple theoretical domains (problem-focused, socioemotional, mental disengagement) rather than focusing on single behavioral categories, enabling identification of user types that employ complex, multidimensional coping approaches through digital platforms. Fourth, the research incorporates demographic moderation analyses to examine how age influences the effectiveness of different social media coping patterns, addressing critical questions about developmental differences in digital coping optimization that remain unresolved in existing literature (Thai et al., 2024).

Developmental Differences in Crisis Stressor Exposure

Adults experience pandemic-related stressors that fundamentally differ from adolescent concerns examined in previous research. Working-age adults faced unprecedented disruptions to career trajectories, income stability, and professional identity that required active problem-solving and information-seeking behaviors to navigate economic uncertainty and workplace modifications (Houston et al., 2015). Parents experienced additional stress from managing children's educational disruption, emotional needs, and safety concerns while potentially lacking traditional family and community support systems.

Older adults within the study age range faced heightened health vulnerability that created unique information needs regarding medical risk assessment, healthcare access, and safety decision-making. These stressors require different coping approaches than the social identity and peer relationship concerns that typically drive adolescent social media use, justifying separate examination of adult coping patterns.

Theoretical Framework for Age-Differentiated Analysis

Socioemotional selectivity theory suggests that older adults prioritize emotionally meaningful experiences and relationships over information acquisition or social network expansion. This developmental shift may influence how different age groups within the adult range utilize social media for coping purposes, with younger adults potentially benefiting more from information-seeking and social connection, while older adults may prefer selective engagement focused on maintaining close relationships (Oh et al., 2014).

The study’s age range of 18 to 70 years captures key developmental transitions in technology adoption, social network characteristics, and coping resource availability that may moderate the effectiveness of different social media use patterns. This age span enables examination of how digital nativity, life experience, and available offline resources influence optimal social media coping strategies during crisis periods.

Operational Definition of Psychological Adjustment

Before developing hypotheses about psychological adjustment outcomes, we establish a clear operational definition of this construct. Psychological adjustment refers to an individual’s ability to maintain emotional equilibrium, functional capacity, and adaptive coping responses when confronting significant stressors or life changes (Cruz et al., 2020). This construct encompasses both the absence of psychological distress symptoms and the presence of positive adaptation indicators such as emotional regulation, behavioral flexibility, and maintenance of daily functioning.

Psychological adjustment differs from related constructs such as life satisfaction or subjective well-being by specifically focusing on adaptive responses to stressor exposure rather than general happiness or contentment (Wei, 2022). In the context of the COVID-19 pandemic, psychological adjustment captures individuals’ capacity to maintain emotional stability and functional effectiveness despite unprecedented disruptions to normal life patterns, social relationships, and economic security.

This study operationalizes psychological adjustment through the Brief Adjustment Scale-6, which assesses participants’ self-reported ability to manage emotions, maintain daily routines, and cope effectively with stressful circumstances during the pandemic period. Higher scores indicate better psychological adjustment, reflecting greater adaptive capacity during crisis conditions.

Critical Analysis of Coping Strategy-User Type Relationships

The relationship between specific coping strategies and social media user types requires careful theoretical consideration that extends beyond simple correspondence assumptions. While stress and coping theory suggests that problem-focused strategies should characterize information-seeking users, empirical evidence presents more complex patterns that challenge straightforward predictions (Vinney, 2024).

Stress-Driven Social Media Use: Beyond Simple Activation Models

The relationship between perceived stress and social media coping strategy adoption extends beyond simple stress-activation models that assume higher stress automatically increases coping behavior engagement. Research during the COVID-19 pandemic reveals complex patterns where stress may simultaneously motivate coping behavior while compromising the cognitive resources necessary for effective strategy selection and implementation (Zhao & Zhou, 2020).

Empirical evidence supports the general principle that crisis-related stress increases social media engagement (Wolfers et al., 2024), but this relationship may be moderated by individual differences in stress tolerance, digital literacy, and available offline coping resources. Some individuals may respond to stress by increasing social media use across multiple domains, while others may selectively increase specific types of engagement based on their coping preferences (Dhir et al., 2018) and past experience with digital platforms.

User Type Differences in Psychological Adjustment: Moving Beyond Tautological Predictions

The prediction that different social media user types will demonstrate varying levels of psychological adjustment requires theoretical justification that extends beyond the tautological assumption that adaptive behaviors produce better outcomes. The complexity lies in determining which patterns of social media engagement prove most adaptive during crisis conditions, given contradictory evidence about the psychological effects of different platform uses.

Comprehensive Users might theoretically demonstrate superior adjustment due to their diverse coping repertoire and active engagement with multiple stress management strategies. Alternatively, they might show poorer adjustment due to information overload, cognitive strain from managing multiple platform activities simultaneously, or exposure to excessive crisis-related content across different engagement domains.

Socio-Recreational Users might benefit from their selective focus on social connection and positive emotional experiences while avoiding potentially distressing information consumption. However, they might also demonstrate poorer adjustment if their limited information seeking leaves them feeling unprepared or out of control during crisis conditions.

Minimal Users might show better adjustment due to reduced exposure to social comparison, misinformation, and algorithmic content designed to maximize engagement rather than support well-being. Conversely, they might demonstrate poorer adjustment due to reduced access to social support, information resources, and community connection during periods when traditional offline support systems were disrupted.

Direct Stress Effects on Psychological Adjustment

While the relationship between stress exposure and psychological adjustment appears self-evident, the specific mechanisms through which COVID-19 stress impacts adjustment during pandemic conditions require empirical verification. The unprecedented nature of pandemic stressors, including their duration, unpredictability, and global scope, may create adjustment challenges that differ qualitatively from responses to more typical life stressors.

Age Moderation: Theoretical Foundations and Developmental Considerations

The prediction that age moderates relationships between social media user types and psychological adjustment stems from multiple theoretical frameworks addressing developmental differences in technology use, social support systems, and coping resource availability across the lifespan.

Justification for Latent Class Analysis

The choice of latent class analysis (LCA) over alternative methods reflects both methodological and theoretical considerations aligned with this study’s objectives. Unlike variable-centered approaches such as regression or structural equation modeling, which assume uniform relationships across populations, LCA identifies naturally occurring subgroups of users employing distinct coping strategy combinations. It offers key advantages: its model-based classification provides statistical criteria for determining class number and assessing uncertainty, producing more reliable and interpretable groupings than distance-based clustering methods like k-means. LCA also handles categorical and ordinal indicators, such as Likert-scale items, more appropriately and generates conditional probability estimates that clarify the likelihood of specific behaviors within each class. Compared to latent profile analysis, which assumes continuous indicators, LCA better captures social media coping as categorical response patterns, while its focus on behavioral patterns rather than latent trait distributions distinguishes it from mixture modeling. Importantly, its person-centered orientation aligns with coping theories emphasizing coherent behavioral profiles over isolated actions, enabling identification of user types whose engagement patterns differentially relate to psychological outcomes—insights unattainable through variable-centered analyses.

Sampling Framework and Recruitment

Data were collected from 479 adults aged 18 to 70 residing in urban areas who experienced social restrictions during the COVID-19 pandemic. The sampling strategy employed stratified random sampling to ensure adequate representation across multiple demographic dimensions. The stratification framework incorporated three primary variables: age group (18–30, 31–59, 60–70 years), gender (male, female, non-binary), and geographic regions (Northeast, Southeast, Midwest, Southwest, West Coast) based on major metropolitan areas within each region.

Participants were recruited through multiple channels to enhance representativeness and reduce selection bias. Primary recruitment occurred through online research panels maintained by Qualtrics, supplemented by social media advertising targeted to specific demographic groups to ensure adequate representation of underrepresented populations. Additional recruitment utilized community partnerships with organizations serving diverse populations, including cultural centers, community colleges, and local health organizations.

Demographic Representativeness and Minority Group Inclusion

The final sample achieved reasonable demographic diversity across several key characteristics. Gender distribution included 52.4% female participants, 45.8% male participants, and 1.8% non-binary or other gender identities. Age distribution reflected the stratified sampling approach with 34.2% young adults (18–30 years), 43.0% middle-aged adults (31–59 years), and 22.8% older adults (60–70 years).

Racial and ethnic composition included 68.1% White participants, 14.2% Hispanic or Latino participants, 12.3% Black or African American participants, 3.8% Asian participants, 1.0% Native American participants, and 0.6% participants identifying as other races or multiple racial categories. Educational attainment varied across the sample, with 15.2% holding high school education or less, 28.4% having some college education, 31.7% holding bachelor's degrees, and 24.7% possessing graduate or professional degrees.

While the sample achieved reasonable diversity, several limitations regarding representativeness should be acknowledged. The sample slightly overrepresented college-educated individuals compared to national demographics, reflecting the online recruitment methodology and digital literacy requirements for participation. Geographic representation favored urban areas as specified in the inclusion criteria, potentially limiting generalizability to rural populations. Additionally, the requirement for regular social media access may have systematically excluded individuals with limited technology access or digital literacy skills.

Power Analysis and Sample Size Justification

The final sample of 479 participants was determined through power analysis tailored to latent class analysis (LCA). Monte Carlo simulations by Nylund et al. (2007) indicate that LCA requires larger samples than traditional methods due to the complexity of estimating class membership probabilities and within-class item response patterns, with 300 to 500 participants sufficient to detect 3–5 latent classes under moderate to strong separation. For this study, power calculations based on pilot data and prior research using the Inventory of Online Coping suggested that 450 to 500 participants would provide 80% power to detect 3–4 classes with medium effect sizes (Cohen’s w = 0.30) at α = .05. The achieved sample of 479 met these requirements while allowing for exclusions due to incomplete responses. This size also supports stable estimation of class parameters and meets recommended minimums of 25 to 50 cases per class, ensuring adequate representation for meaningful interpretation and subsequent multinomial logistic regression analyses predicting class membership.

Inclusion and Exclusion Criteria

Inclusion criteria required participants to meet several conditions ensuring relevance to the research questions. Participants needed to be at least 18 years of age to provide informed consent and possess cognitive capacity for survey completion. Regular access to social media platforms was defined as using at least one major platform (Facebook, Twitter, Instagram, TikTok, LinkedIn, or similar) at least weekly during the pandemic period. Experience with social restrictions required participants to have lived in areas with mandatory or recommended lockdown measures, quarantine requirements, or social distancing mandates during any phase of the COVID-19 pandemic.

Exclusion criteria eliminated participants whose responses might compromise data quality or study validity. Incomplete survey responses were excluded if participants completed fewer than 80% of key measures, while responses showing obvious response patterns (such as selecting the same response option for all items) were excluded after manual review. Participants who reported no social media use during the pandemic period were excluded as they could not provide meaningful responses about social media coping strategies.

Social Media Coping Strategies

Social media coping strategies were assessed using a modified version of the Inventory of Online Coping (van Ingen et al., 2016). The 14-item instrument measured three distinct coping strategies:

Participants rated each item on a 5-point Likert scale ranging from 1 (does not apply to me at all) to 5 (applies to me very much).

Psychological Adjustment

Psychological adjustment was measured using the Brief Adjustment Scale-6 (Cruz et al., 2020). This 6-item scale (α = .93) assesses participants’ ability to maintain psychological well-being during stressful circumstances. Items were scored on a 5-point scale from 1 (not at all) to 5 (extremely), with higher scores indicating better psychological adjustment after reverse-coding appropriate items.

Perceived COVID-19 Stress

Perceived stress related to COVID-19 was assessed using an 8-item checklist adapted from previous pandemic research (Main et al., 2011; Li et al., 2020). Participants indicated whether they had experienced specific stressors (e.g., “confirmed or suspected infection,”“significant reduction in family income”) by selecting either “yes” (1) or “no” (0). The total score ranged from 0 to 8, with higher scores indicating greater perceived stress.

Demographic Variables

Demographic information collected included age, gender, education level, geographic location, and daily social media use duration. Age was categorized into three groups for analysis: young adults (18–30 years), middle-aged adults (31–59 years), and older adults (60–70 years).

Digital Addiction Assessment Limitation

This study did not include validated measures to assess digital addiction, social media addiction, or problematic internet use patterns among participants. While demographic information regarding daily social media use duration was collected, standardized instruments for detecting addictive or problematic social media use behaviors, such as the Bergen Social Media Addiction Scale (Andreassen et al., 2012) or the Problematic Internet Use Questionnaire (Demetrovics et al., 2008), were not administered. This methodological limitation must be acknowledged when interpreting psychological adjustment differences across latent user classes, particularly regarding the Comprehensive Users group who demonstrated extensive engagement across all measured coping strategies. Our 8-item checklist measured stressor exposure (confirmed or suspected infection, income reduction, social isolation) rather than the multidimensional symptom profile that characterizes stress responses. This methodological approach does not permit differentiation between individuals who may experience predominantly psychological versus somatic stress manifestations, nor does it capture the intensity or duration of stress symptoms across these dimensions. This limitation must be considered when interpreting the relationships between perceived COVID-19 stress, social media coping strategies, and psychological adjustment outcomes.

Data Preprocessing

Prior to analysis, data were screened for missing values, outliers, and violations of statistical assumptions. Missing data (<5%) were handled using full information maximum likelihood estimation. All analyses were conducted using R (version 4.2.0) with the `poLCA' package for latent class analysis and `lavaan' for structural equation modeling.

Latent Class Analysis

A series of latent class models were estimated with an increasing number of classes (1–6). The following statistical criteria were used to determine the optimal number of classes:

Indicators for the LCA included the 14 items from the Inventory of Online Coping. Item responses were treated as ordinal variables.

The LCA modeling approach followed established best practices for person-centered analysis. Model estimation employed full information maximum likelihood with robust standard errors to handle missing data and non-normality. Multiple random starts were utilized for each model to ensure identification of global rather than local maxima, with 500 random start values and 50 optimizations retained for final model selection. Convergence criteria required log-likelihood improvement of less than 1e-6 across iterations, ensuring stable parameter estimates across all estimated models.

Analysis of Class Differences

After identifying the optimal latent class solution, differences between classes were examined using:

Moderation Analysis

To examine the potential moderating effect of age on the relationship between class membership and psychological adjustment, moderation analyses were conducted using a structural equation modeling framework. Interaction terms were created between age group (categorical) and class membership (using most likely class assignment).

Model Selection

We estimated a series of latent class models with 1-6 classes using the 14 items from the Inventory of Online Coping. Table 1 presents the fit indices for each model.

OPEN IN VIEWER

Table 1.

Fit Indices for Latent Class Models (1–6 Classes).

Classes	Log-likelihood	AIC	BIC	SABIC	Entropy	LMR-LRT p-value	BLRT p-value
1	−6,892.45	13,854.90	13,974.32	13,875.17	—	—	—
2	−6,479.36	13,080.72	13,324.96	13,122.45	0.82	<.001	<.001
3	−6,271.18	12,716.36	13,085.41	12,779.54	0.86	<.01	<.001
4	−6,187.54	12,601.08	13,094.95	12,685.71	0.84	.07	<.001
5	−6,135.27	12,548.54	13,167.22	12,654.62	0.83	.23	<.001
6	−6,101.65	12,533.30	13,276.80	12,660.83	0.81	.42	.08

Based on the statistical criteria, the 3-class solution was selected as optimal. The 3-class model showed significant improvement over the 2-class model (LMR-LRT p < .01), while the 4-class model did not show significant improvement over the 3-class model (LMR-LRT p = .07). The 3-class model had good classification quality (entropy = 0.86) and exhibited interpretable class patterns.

Model Fit Indices for Latent Class Models (1–6 Classes)

The Figure 2. illustrates the AIC, BIC, and SABIC values across different latent class models (1–6 classes), providing insight into model fit. As the number of classes increases, AIC, BIC, and SABIC values generally decrease, indicating an improved model fit. However, BIC reaches its lowest point at 3–4 classes, suggesting that these models offer the best balance between fit and complexity. Additionally, LMR-LRT p-values indicate that a 3-class solution is statistically significant, whereas adding more classes results in diminishing returns, making 3–4 classes the most optimal solution.

OPEN IN VIEWER

Figure 2.

Model fit indices for latent class models (1–6 classes).

Latent Class Profiles

The three identified classes represented distinct patterns of social media coping strategies. Figure 3. displays the mean scores for each coping strategy across the three classes. The classes were labeled based on their distinctive patterns:

OPEN IN VIEWER

Figure 3.

Mean scores of social media coping strategies by latent class.

Table 2 presents the conditional probabilities of endorsing high levels (i.e., responses of 4 or 5 on the 5-point scale) of each coping strategy item by latent class.

OPEN IN VIEWER

Table 2.

Conditional Probabilities of High Endorsement for Each Coping Strategy Item by Latent Class.

Item	Comprehensive users	Socio-recreational users	Minimal users
Problem-focused coping
PFC1	0.83	0.21	0.15
PFC2	0.78	0.18	0.12
PFC3	0.74	0.24	0.11
PFC4	0.81	0.25	0.14
PFC5	0.72	0.22	0.09
PFC6	0.76	0.19	0.13
Socioemotional coping
SEC1	0.79	0.74	0.21
SEC2	0.82	0.78	0.23
SEC3	0.77	0.75	0.18
SEC4	0.81	0.73	0.19
SEC5	0.78	0.77	0.24
SEC6	0.84	0.81	0.20
Mental disengagement
MD1	0.89	0.83	0.26
MD2	0.86	0.79	0.22

Predictors of Class Membership

Multinomial logistic regression was conducted to examine how perceived COVID-19 stress and demographic variables predicted class membership. Table 3 presents the results, with the Minimal Users class as the reference category.

OPEN IN VIEWER

Table 3.

Multinomial Logistic Regression Results for Class Membership (Reference: Minimal Users).

Predictor	Comprehensive users	Socio-recreational users
	OR (95% CI)	OR (95% CI)
Perceived COVID-19 stress	1.42** (1.15–1.74)	1.29* (1.06–1.58)
Age group (ref: 18–30)
31–59	0.67* (0.48–0.94)	0.72* (0.53–0.98)
60–70	0.38** (0.21–0.69)	0.54* (0.32–0.91)
Gender (ref: Male)
Female	1.27 (0.91–1.78)	1.31* (1.02–1.68)
Education level	0.92 (0.79–1.07)	0.96 (0.83–1.12)
Daily social media use	1.34** (1.12–1.61)	1.29** (1.08–1.54)

*

p < .05, **p < .01.

Perceived COVID-19 stress was significantly associated with higher odds of belonging to both the Comprehensive Users (OR = 1.42, p < .01) and Socio-Recreational Users (OR = 1.29, p < .05) classes compared to the Minimal Users class. This indicates that individuals experiencing higher levels of pandemic-related stress were more likely to utilize social media for coping.

Age was a significant predictor of class membership, with older adults (60–70) having lower odds of being Comprehensive Users (OR = 0.38, p < .01) or Socio-Recreational Users (OR = 0.54, p < .05) compared to young adults (18–30). Daily social media use duration was positively associated with membership in both the Comprehensive Users (OR = 1.34, p < .01) and Socio-Recreational Users (OR = 1.29, p < .01) classes.

The significant association between perceived COVID-19 stress and social media coping strategy adoption (H2 supported) must be interpreted within the constraints of our stress measurement approach. The binary checklist methodology assessed stressor exposure but did not differentiate between psychological and somatic dimensions of stress responses, limiting conclusions about specific symptom-coping strategy relationships. Future research employing multidimensional stress assessment instruments would provide more nuanced understanding of how different aspects of stress experience influence social media coping behavior selection.

Class Differences in Psychological Adjustment

One-way ANOVA was conducted to examine differences in psychological adjustment across the three latent classes. Table 4 presents the mean psychological adjustment scores for each class.

OPEN IN VIEWER

Table 4.

Mean Psychological Adjustment Scores by Latent Class.

Class	M	SD	F	p	η2
Comprehensive users	3.12	0.72	16.83	<0.001	0.14
Socio-recreational users	3.65	0.68
Minimal users	3.24	0.75

The ANOVA results indicated significant differences in psychological adjustment across the three classes (F(2, 476) = 16.83, p < .001, η² = 0.14). Post-hoc comparisons using Tukey's HSD test revealed that Socio-Recreational Users reported significantly higher psychological adjustment scores than both Comprehensive Users (mean difference = 0.53, p < .001) and Minimal Users (mean difference = 0.41, p < .001). There was no significant difference between Comprehensive Users and Minimal Users (mean difference = 0.12, p = .28).

Moderation Analysis: Age as a Moderator

To examine whether age moderated the relationship between class membership and psychological adjustment, a multigroup analysis was conducted. Table 5 presents the mean psychological adjustment scores by latent class and age group.

OPEN IN VIEWER

Table 5.

Mean Psychological Adjustment Scores by Latent Class and Age Group.

Class	18–30 years	31–59 years	60–70 years
	M (SD)	M (SD)	M (SD)
Comprehensive users	3.05 (0.73)	3.15 (0.72)	3.24 (0.68)
Socio-recreational users	3.78 (0.65)	3.62 (0.70)	3.41 (0.71)
Minimal users	3.14 (0.77)	3.26 (0.74)	3.35 (0.72)

A significant interaction effect was found between class membership and age group (F(4, 470) = 3.86, p < .01, η² = 0.03). The positive association between Socio-Recreational Users class membership and psychological adjustment was strongest among young adults (18–30 years) and progressively weaker in older age groups. This suggests that using social media primarily for socioemotional coping and mental disengagement was most beneficial for younger individuals.

A regression analysis with interaction terms further confirmed this moderation effect. The interaction term for Socio-Recreational Users class × Age Group was significant (β = −.12, p < .05), while the interaction term for Comprehensive Users class × Age Group was not significant (β = .07, p = .18).

Three distinct classes of social media coping were identified: Comprehensive Users (35.1%), Socio-Recreational Users (40.9%), and Minimal Users (24.0%). Higher perceived COVID-19 stress was associated with increased likelihood of using social media for coping (both comprehensive and socio-recreational patterns). Socio-Recreational Users demonstrated significantly higher psychological adjustment compared to both Comprehensive Users and Minimal Users. Age moderated the relationship between class membership and psychological adjustment, with younger adults benefiting more from socio-recreational social media use than older adults. These findings suggest that using social media primarily for socioemotional support and mental disengagement, rather than problem-focused coping, may be more beneficial for psychological adjustment during stressful periods, particularly for younger individuals.

Visual Representation of Age Moderation Effects

To clarify the nature of the interaction between social media user class and age group on psychological adjustment, Figure 4 presents the interaction pattern graphically. This plot illustrates how the relationship between class membership and psychological adjustment varies across the three age groups.

OPEN IN VIEWER

Figure 4.

Age Moderation of the Relationship Between Social Media User Class and Psychological Adjustment.

The interaction plot reveals several important patterns that clarify the moderation effects. Among young adults aged 18 to 30 years, Socio-Recreational Users demonstrate substantially higher psychological adjustment (M = 3.78) compared to both Comprehensive Users (M = 3.05) and Minimal Users (M = 3.14), creating a clear differentiation between user types. This pattern indicates that selective social media engagement focused on emotional support and entertainment proves particularly beneficial for younger adults.

The age moderation effect becomes evident in the convergence of adjustment scores across user types as age increases. Among middle-aged adults (31–59 years), the advantage of Socio-Recreational Users diminishes (M = 3.62), while Comprehensive Users show slight improvement (M = 3.15) and Minimal Users maintain similar levels (M = 3.26). This convergence suggests that the benefits of selective social media engagement become less pronounced with increasing age.

Among older adults (60–70 years), the differences between user types become minimal, with Socio-Recreational Users showing the smallest advantage (M = 3.41) compared to other age groups, while Comprehensive Users (M = 3.24) and Minimal Users (M = 3.35) demonstrate relatively similar adjustment levels. This pattern supports the hypothesis that older adults rely less heavily on social media for psychological support and therefore show less variation in adjustment based on their digital coping patterns.

The graphical representation clarifies that the age moderation effect primarily reflects the diminishing advantage of socio-recreational social media use among older age groups rather than differential effectiveness of comprehensive or minimal use patterns across age groups. This finding has important implications for age-targeted interventions promoting optimal social media use during crisis periods.

Mental Health Professionals

Clinical practitioners should implement structured social media assessment protocols during crisis periods that evaluate coping pattern types rather than overall usage duration. Specifically, practitioners should assess whether clients engage in information seeking behaviors through social media and help them develop strategies for limiting crisis-related content consumption while maintaining social connection features. Treatment protocols should include digital literacy components that teach clients to recognize algorithmic content curation and develop skills for curating supportive digital environments.

Therapeutic interventions should incorporate specific techniques for transitioning Comprehensive Users toward more selective engagement patterns. This includes developing implementation intentions for social media use, such as designated times for information checking versus social connection activities, and teaching clients to recognize emotional states that trigger excessive information seeking behaviors. Mental health organizations should develop standardized screening tools that identify high-risk social media coping patterns and create referral pathways for digital wellness interventions.

Policymakers and Public Health Agencies

Public health communications during crisis periods should include explicit guidance about optimal social media engagement patterns, moving beyond generic recommendations for reduced screen time toward specific behavioral strategies. Policy frameworks should require social media platforms to provide users with transparent data about their information consumption patterns, particularly exposure to crisis-related content, enabling more informed self-regulation decisions.

Educational institutions should integrate digital wellness curricula that teach evidence-based social media coping strategies, with particular emphasis on selective engagement techniques and recognition of algorithm-driven content delivery. These programs should target different age groups with developmentally appropriate interventions, given the observed age moderation effects.

Regulatory considerations should address platform design features that promote comprehensive rather than selective engagement, such as infinite scroll mechanisms and cross-domain content mixing that combine crisis information with social content in single feeds.

Social Media Platform Development

Technology companies should implement user interface modifications that support selective engagement patterns identified as beneficial in this research. Specific recommendations include developing separate interfaces for information seeking versus social connection activities, allowing users to compartmentalize different types of social media engagement. Platforms should create user-controlled filtering mechanisms that enable individuals to limit exposure to crisis-related content while maintaining access to social support features.

Algorithmic modifications should prioritize supportive content over engaging content during declared public health emergencies, with transparent user controls for adjusting content prioritization based on individual coping preferences. Platforms should implement periodic digital wellness check-ins that assess user psychological state and provide personalized recommendations for adjusting engagement patterns based on empirical evidence from studies such as this research.

Platform analytics should provide users with detailed insights into their content consumption patterns, including ratios of crisis information versus social content, temporal patterns of usage, and mood correlations with different types of engagement. These features would enable users to develop self-awareness about their digital coping patterns and make informed adjustments.