How Does Digital Connection Shape Cultural Adaptation? Differentiated Social Media Use,Intercultural Communication Self-Efficacy,and Adaptation Among International Students in China

Abstract

Evidence regarding the adaptive role of social media in acculturation remains fragmented, often overlooking behavioral nuances and contextual constraints. This study examined a moderated mediation model disentangling how three types of social media use—passive, active social, and active non-social—are associated with psychological and sociocultural adaptation through intercultural communication self-efficacy. Participants were 2,478 international students in China (M = 23.67 years, SD = 3.82 years) who completed a multi-language survey. Structural equation modeling indicated that active social and active non-social use were positively associated with adaptation outcomes via enhanced intercultural communication self-efficacy, whereas passive use showed no significant association with efficacy beliefs. Furthermore, perceived cultural distance significantly moderated these indirect pathways; the adaptive associations between active engagement and self-efficacy were substantially attenuated for students perceiving high cultural disparity. These findings challenge universalist assumptions of digital empowerment, identifying intercultural communication self-efficacy as a core cognitive engine and perceived cultural distance as a critical boundary condition in digital acculturation.

Keywords

social media use types intercultural communication self-efficacy perceived cultural distance psychological adaptation sociocultural adaptation international students

Introduction

The global landscape of higher education is increasingly shaped by sustained international student mobility, a demographic pattern that brings both developmental opportunities and heightened psychological vulnerability (Hofhuis et al., 2023; Maharaj et al., 2025; Zhai et al., 2025). While digital technologies are often framed as resources that can mitigate stressors associated with cultural transition, a central debate persists regarding whether social media primarily functions as a “digital bridge” that facilitates host-society participation or as a “digital cocoon” that reinforces withdrawal and isolation (Hofhuis et al., 2023; Zhou & Yin, 2024). Despite the proliferation of research in this domain, a key gap concerns how the qualitative features of digital engagement (e.g., communication-oriented activities, specific interaction patterns, or problematic use), rather than mere frequency or time spent, relate to adjustment and well-being (Pang & Ruan, 2024; Pang & Wang, 2025; Yang et al., 2025; Zhou & Yin, 2024). This issue is particularly salient for international students in China, where a unique digital ecosystem and vast cultural disparities challenge existing universalist assumptions of digital empowerment (Cao et al., 2023; Cao & Jia, 2023; Mao, 2025; Zhang & Ting, 2025). By integrating Social Cognitive Theory with the Cultural Fit Hypothesis, this study challenges the simplistic “technological optimism” of previous literature and introduces a more rigorous conceptual framework to examine the boundary conditions of digital intervention. Drawing on Social Cognitive Theory’s emphasis on efficacy beliefs as proximal determinants of behavioral regulation and adaptation (Bandura, 1977), this study examines whether differentiated social media use types are associated with adaptation outcomes through intercultural communication self-efficacy, while accounting for the constraining role of perceived cultural distance.

Differentiated Social Media Use and Intercultural Communication Self-Efficacy

The functional perspective of digital engagement suggests that psychological utility depends on the form of use rather than use per se. Prior work has repeatedly indicated that broad indicators (e.g., time spent) often yield near-zero average associations, whereas effects become more interpretable when specific activities and experiences are distinguished (Zhou & Yin, 2024). Accordingly, literature that does not separate passive consumption from active participation risks obscuring mechanisms through which online behavior may (or may not) support resource acquisition, efficacy development, and subsequent adaptation (Cao & Jia, 2023; Hofhuis et al., 2023). In this study, digital engagement is operationalized using the Passive and Active Use Measure (PAUM), which differentiates active social use (direct interpersonal interaction), active non-social use (broadcasting or content production without direct interaction), and passive use (consuming others’ content without interaction; Gerson et al., 2017). Evidence from the broader social media literature indicates that active social engagement is more consistently related to perceived social support and well-being, whereas passive consumption is more often linked to social comparison and poorer affective outcomes (Pang, 2023; Verduyn et al., 2015). In cross-cultural contexts, online interaction may also provide an additional channel for language practice, norm learning, and relationship maintenance beyond offline encounters, suggesting that functional differences in engagement may matter for adaptation (Pang & Wang, 2020; Rui & Wang, 2015). According to the social affordances framework, active social use and active non-social use provide international students with opportunities for “mastery experiences” and “vicarious learning,” which are essential precursors to building robust efficacy beliefs (Pang & Wang, 2020; Xin et al., 2025). In contrast, passive use—characterized by the non-interactive monitoring of content—frequently correlates with increased social comparison rather than competence building. Based on this theoretical grounding, we propose Hypothesis 1: Active social use and active non-social use are positively associated with intercultural communication self-efficacy, whereas passive use exhibits a non-significant or weaker association with this construct.

The Mediating Role of Intercultural Communication Self-Efficacy in Adaptation

From an acculturation perspective, adaptation reflects how effectively sojourners manage everyday demands and develop competent participation in the host society (Berry, 1997; Boski et al., 2025; Ward & Geeraert, 2016). We propose that intercultural communication self-efficacy (ICSE) is a plausible psychological mechanism linking digital engagement to adaptation. Self-efficacy refers to beliefs about one’s capability to organize and execute actions required to manage prospective situations (Bandura, 1977). In intercultural contexts, higher ICSE is associated with greater cultural competence and better adjustment outcomes (Kabir & Sponseller, 2020; Wilson et al., 2013). While social media platforms offer a low-stakes environment for cultural experimentation, the adaptive value of these interactions is realized through the enhancement of an individual’s internal belief system regarding their communicative success (Pang & Wang, 2020; Xin et al., 2025, 2026). Students who report higher levels of intercultural communication self-efficacy are likely to exhibit better psychological adjustment and more effective sociocultural functioning (Jurasek & Wawrosz, 2023; Kabir & Sponseller, 2020). Consequently, we propose Hypothesis 2: Intercultural communication self-efficacy mediates the associations between active social media use types (social and non-social) and both psychological and sociocultural adaptation outcomes.

Perceived Cultural Distance as a Boundary Condition

The adaptive efficiency of digital tools is not inherently universal but is instead embedded within the broader socio-cultural context of the host environment. The Cultural Fit Hypothesis suggests that the psychological and functional load of acculturation increases proportionally with the magnitude of the gap between home and host cultures (Ward & Geeraert, 2016). For international students perceiving significant cultural distance, the “efficacy boost” derived from social media may be significantly attenuated by the overwhelming complexity of real-world interactions (Hofhuis et al., 2023; Malay et al., 2024; Rui & Wang, 2015). Under conditions of profound cultural disparity, the psychological resources garnered online may be insufficient to overcome the stress of navigating radically different social norms and communicative scripts (Lu et al., 2025). Therefore, we propose Hypothesis 3: Perceived cultural distance moderates the indirect associations between social media use types and adaptation outcomes, such that these indirect pathways through intercultural communication self-efficacy are significantly attenuated when perceived cultural distance is high.

In summary, current theoretical frameworks regarding the digital acculturation of international students remain fragmented, often failing to account for the qualitative diversity of digital behavior or the contextual gravity of cultural distance. This study addresses these theoretical gaps by providing three distinct contributions to the field of cross-cultural psychology. First, it moves beyond monolithic measures of social media use by validating a differentiated behavioral model, thereby challenging existing assumptions about the universal benefits of digital connectivity. Second, it elevates the role of intercultural communication self-efficacy as a central cognitive engine in the acculturation process, bridging the gap between virtual activity and real-world adjustment. Finally, by examining the moderating influence of perceived cultural distance, this research introduces an essential environmental perspective that highlights the limits of digital empowerment in the face of deep-seated cultural divides. This research not only clarifies the standing of digital behaviors within the broader academic dialogue on acculturation but also provides a more solid foundation for institutional support strategies in an increasingly digitized global academy.

Method

Research Design

This study employed a cross-sectional survey design to examine the associations among social media use types, intercultural communication self-efficacy, perceived cultural distance, and cross-cultural adaptation outcomes (Figure 1). Cross-sectional designs are commonly used in cross-cultural psychology research to examine complex multivariate relationships when the primary goal is establishing patterns of association rather than causal sequences (Demes & Geeraert, 2014). Consistent with methodological recommendations regarding causal inference limitations in observational research (Maxwell & Cole, 2007), we interpret associations as indirect effects rather than causal mediation effects. This epistemological positioning acknowledges that temporal ordering cannot be established in cross-sectional data while still permitting theoretically meaningful examination of hypothesized pathways.

Figure 1.

Theoretical moderated mediation model of social media use and adaptation.

Participants and Sampling

Participants were 2,478 international students enrolled in higher education institutions across Guangxi Province, China. Eligibility criteria included: (a) non-Chinese nationality, (b) current residence in Guangxi for academic study, and (c) sufficient proficiency in at least one of the offered survey languages (English, French, Vietnamese, Thai, or Indonesian) to complete the questionnaire. Participants were recruited through stratified convenience sampling across eight universities, with stratification based on institution type (comprehensive university, specialized institution) and geographic location within Guangxi.

Of the total sample, 54.6% were male (n = 1,353) and 45.4% were female (n = 1,125). Participants ranged in age from 18 to 42 years (M = 23.67, SD = 3.82). Length of residence in China ranged from 6 to 72 months (M = 18.45, SD = 10.23). Participants originated from 68 countries, with the largest representations from Vietnam (18.4%), Thailand (15.2%), Indonesia (12.6%), Pakistan (9.8%), and various African nations (26.3%). Self-rated Chinese language proficiency on a 5-point scale (1 = very poor to 5 = excellent) indicated moderate proficiency (M = 2.89, SD = 1.12). Approximately 31.2% of participants reported prior overseas study or living experience before coming to China.

Sample size was determined through a priori power analysis for structural equation modeling. Following recommendations by Kline (2023), we targeted a minimum of 200 cases or 10 to 20 cases per estimated parameter. Given the complexity of our conditional process model with approximately 45 estimated parameters, a minimum sample of 450 to 900 was required. Monte Carlo simulation using Mplus indicated that a sample size of approximately 800 would provide power of 0.80 to detect small-to-medium indirect effects (β = .10) at α = .05. The obtained sample of 2,478 participants substantially exceeded these requirements.

Procedure

Data were collected via an anonymous online questionnaire administered through Wenjuanxing (a Chinese survey platform widely used in academic research). Recruitment occurred through multiple channels: international student office announcements, class-based invitations with instructor permission, and social media posts in international student community groups. Data collection took place between September and December 2023.

To address the well-documented ambiguity regarding which social media platform participants report on (Gerson et al., 2017), we implemented a platform-anchoring procedure. Participants first indicated their most frequently used social media platform in the past month from a provided list (including WeChat, Weibo, TikTok/Douyin, Facebook, Instagram, WhatsApp, Twitter/X, Telegram, and Other). All subsequent social media use items were then explicitly anchored to “the platform you use most often.” This platform-agnostic approach preserves theoretical generalizability regarding use-type behaviors while clarifying the behavioral referent for measurement.

Several procedural remedies were implemented to minimize common method bias (Podsakoff et al., 2003). First, we assured participants of response anonymity and emphasized that there were no right or wrong answers to reduce evaluation apprehension and social desirability bias. Second, predictor and criterion variables were placed in separate sections with transitional instructions and filler items between them to create psychological separation. Third, different response scale formats and anchors were used across constructs where feasible (e.g., frequency scales for social media use, agreement scales for self-efficacy, difficulty scales for sociocultural adaptation). Fourth, three attention-check items were embedded throughout the questionnaire (e.g., “Please select ‘Strongly Agree’ for this item”), and participants who failed more than one attention check were excluded from analysis (n = 127 exclusions). The study received ethical approval from the first author’s institutional review board, and all participants provided informed consent prior to participation.

Measures

Social Media Use Types

Social media use was assessed using an adapted version of the Passive and Active Facebook Use Measure (PAUM; Gerson et al., 2017). Following the original instrument’s tripartite structure, the measure assesses three distinct use types: passive use (e.g., “Look at friends’ photos”), active social use (e.g., “Post on friends’ walls/timelines”), and active non-social use (e.g., “Find useful information for my studies or work”). Item wording was adapted by replacing “Facebook” with “your primary social media platform” throughout. Participants responded on a 5-point scale (1 = never to 5 = very frequently). In the present sample, Cronbach’s alpha coefficients were .88 for passive use, .86 for active social use, and .84 for active non-social use.

Intercultural Communication Self-Efficacy

Intercultural communication self-efficacy was measured using the Self-Efficacy in Intercultural Communication Scale-Short Form (SEIC-SF; Kabir & Sponseller, 2020). This 10-item instrument assesses perceived competence in handling intercultural communication situations (e.g., “I can communicate effectively with people from different cultures”; “I am confident in my ability to understand people from different cultural backgrounds”). Participants responded on a 6-point scale (1 = strongly disagree to 6 = strongly agree). Higher scores indicate greater perceived intercultural communication capability. Cronbach’s alpha in the present sample was .92.

Perceived Cultural Distance

Perceived cultural distance was assessed using the 12-item Brief Perceived Cultural Distance Scale (BPCDS; Demes & Geeraert, 2014). This measure captures subjective appraisals of how different various aspects of the host culture (e.g., values, climate, everyday customs, food) are from one’s home culture (e.g., “How different is the climate in your current country of residence from that of your home country?”). Items were anchored to Chinese culture as experienced in daily life. Participants responded on a 7-point scale (1 = very similar to 7 = very different). Higher scores reflect greater perceived distance. Cronbach’s alpha was .90.

Psychological Adaptation

Psychological adaptation was measured using the 10-item Brief Psychological Adaptation Scale (BPAS; Demes & Geeraert, 2014). This measure assesses emotional/affective adjustment to the host environment (e.g., “Excited about being in China”; “Homesick” [reverse-scored]). Participants rated how often they experienced each feeling in the past 2 weeks on a 7-point scale (1 = never to 7 = always). After reverse-scoring negative items, higher mean scores indicate better psychological adaptation. Cronbach’s alpha was .87.

Sociocultural Adaptation

Sociocultural adaptation was assessed using the 12-item Brief Sociocultural Adaptation Scale (BSAS; Demes & Geeraert, 2014). This measure captures functional/behavioral adjustment to everyday life demands in the host culture (e.g., “Building and maintaining relationships”; “Adapting to local etiquette”). Participants rated the difficulty of each behavior on a 7-point scale (1 = very difficult to 7 = very easy). Higher mean scores indicate better sociocultural adaptation. Cronbach’s alpha was .89.

Covariates

Based on prior cross-cultural adaptation research (Demes & Geeraert, 2014; Galchenko & van de Vijver, 2007), we included five covariates theoretically linked to adaptation outcomes: age (years), gender (0 = male, 1 = female), length of residence in China (months), self-rated Chinese language proficiency (1–5 scale), and prior overseas experience (0 = no, 1 = yes). Age and length of residence were included because developmental maturity and acculturation exposure time influence adaptation trajectories. Gender was included given documented differences in social media use patterns and coping styles. Chinese language proficiency was included as a key resource for intercultural communication and daily functioning. Prior overseas experience was included as an indicator of previous cross-cultural exposure that may facilitate adaptation.

Translation Procedures and Measurement Equivalence

The questionnaire was developed in English and translated into French, Vietnamese, Thai, and Indonesian using standardized back-translation procedures (Brislin, 1970). For each language, an initial translation was produced by a bilingual translator with expertise in cross-cultural psychology, followed by independent back-translation by a second bilingual individual. Discrepancies between the original and back-translated versions were resolved through discussion with a third bilingual expert. Face validity of translated versions was further evaluated through cognitive interviews with pilot samples of 10 to 15 international students per language group, ensuring that items were understood as intended. While these procedures, together with the subsequent measurement invariance testing reported below, established functional equivalence across the five language versions used in this study, we acknowledge in the Discussion section that formal psychometric validation of each individual scale within independent monolingual samples for each represented language remains a task for future research.

Data Analysis Strategy

Analyses proceeded in four stages. First, we examined descriptive statistics, bivariate correlations, and multicollinearity diagnostics (variance inflation factors; VIFs) among all study variables. Second, we assessed common method variance using both Harman’s single-factor test and confirmatory factor analysis with an unmeasured latent method factor (Podsakoff et al., 2003). To provide a more rigorous assessment, we compared substantive path coefficients between models with and without the latent method factor to determine whether conclusions were affected by common method variance. Third, we tested measurement invariance across the five language groups using multi-group confirmatory factor analysis (MGCFA), following the sequential approach recommended by Chen (2007). Configural, metric, and scalar invariance were assessed using changes in CFI (ΔCFI ≤−0.010) and RMSEA (ΔRMSEA ≤0.015) as decision criteria, given the sensitivity of chi-square difference tests to large sample sizes.

Fourth, we estimated the hypothesized conditional process model using path analysis with observed scale scores. This approach was selected to prioritize substantive theory testing while retaining a parsimonious model specification. The model specified: (a) direct paths from each social media use type to both adaptation outcomes; (b) indirect paths through intercultural communication self-efficacy; and (c) moderation by perceived cultural distance on the paths from social media use types to intercultural communication self-efficacy (i.e., first-stage conditional process model). Both psychological adaptation and sociocultural adaptation were modeled as parallel outcomes to examine whether the hypothesized pathways operated similarly across distinct adaptation dimensions.

Analyses were conducted using Mplus 8.8 (Muthén & Muthén, 2017) with maximum likelihood (ML) estimation. Missing data were handled using full information maximum likelihood (FIML), which produces unbiased and efficient estimates under the missing at random assumption (Enders, 2022). Indirect effects and conditional indirect effects were evaluated using bias-corrected bootstrap confidence intervals with 10,000 resamples (Preacher & Hayes, 2008). The index of moderated mediation (Hayes, 2015, 2017) was computed to formally test whether the indirect effect varied as a linear function of the moderator. Conditional indirect effects were probed at low (−1 SD), mean, and high (+1 SD) levels of perceived cultural distance. All continuous predictors were mean-centered prior to creating interaction terms to reduce nonessential multicollinearity.

Given that participants were nested within eight universities, we conducted a sensitivity analysis using TYPE = COMPLEX with CLUSTER = university to obtain robust standard errors that account for potential nonindependence of observations within institutions (Muthén & Muthén, 2017). This analysis examines whether conclusions are robust to potential clustering effects at the university level.

Model fit was evaluated using standard indices: chi-square statistic, comparative fit index (CFI ≥0.95 preferred), Tucker-Lewis index (TLI ≥0.95 preferred), root mean square error of approximation (RMSEA ≤0.06 preferred), and standardized root mean square residual (SRMR ≤0.08 preferred; Hu & Bentler, 1999). Effect sizes for path coefficients were interpreted following established guidelines (small: |β| = .10–.19; medium: |β| = .20–.29; large: |β| ≥ .30; Keith, 2019). For indirect effects, we additionally computed the ratio of the indirect effect to the total effect (PM) as a measure of the proportion of the total association accounted for by the indirect path.

Results

Preliminary Analyses

Prior to hypothesis testing, data were screened for missing values, outliers, and distributional properties. Missing data were minimal (<2.1% on any single variable) and appeared to be missing completely at random (Little’s MCAR test: χ² = 234.56, df = 198, p =.078). The percentage of missing data ranged from 0.4% (age) to 2.1% (perceived cultural distance). Full information maximum likelihood estimation was used to handle missing data, which produces unbiased estimates under the missing at random assumption (Enders, 2022). Univariate distributions were examined for normality; skewness values ranged from −0.58 to 0.47 and kurtosis values ranged from −0.39 to 0.72, all within acceptable limits (|skewness| < 2, |kurtosis| < 7; Kline, 2023). Multivariate outliers were examined using Mahalanobis distance; 19 cases exceeded the critical value (p < .001) and were retained following sensitivity analyses confirming that parameter estimates and conclusions remained unchanged when these cases were excluded.

Regarding primary social media platform use, the most commonly reported platforms were WeChat (42.3%), Facebook (18.7%), Instagram (12.4%), WhatsApp (11.2%), TikTok/Douyin (8.6%), and other platforms (6.8%). This distribution reflects the mixed composition of the sample, with domestic Chinese platforms predominating overall but Western platforms common among participants from Southeast Asia, Africa, and Europe. Given the platform-anchoring procedure employed, we conducted sensitivity analyses including primary platform type as a covariate, with results reported below.

Multicollinearity was assessed via variance inflation factors (VIFs). All VIFs were below 2.5 (range: 1.18–2.34), well below the conventional threshold of concern (VIF >5; Hair et al., 2019), indicating that multicollinearity did not pose a threat to parameter estimation. The highest VIF (2.34) was observed for the active social use × perceived cultural distance interaction term, which is expected for product terms even after mean-centering.

Descriptive Statistics and Correlations

Means, standard deviations, and bivariate correlations among all study variables are presented in Table 1. The three social media use types showed moderate positive intercorrelations (rs = .36–.54), consistent with prior research using the PAUM (Gerson et al., 2017) and supporting their conceptualization as related but distinct behavioral patterns. Notably, passive use showed weaker associations with intercultural communication self-efficacy (r = .06, p = .003) compared to active social use (r = .34, p < .001) and active non-social use (r = .36, p < .001), providing preliminary support for the differentiated effects of use types.

Table 1.

Means, Standard Deviations, and Bivariate Correlations Among Study Variables.

Variable	M	SD	1	2	3	4	5	6	7	8	9
1. Passive use	3.42	0.87	—
2. Active social use	2.94	0.92	.54***	—
3. Active non-social use	3.31	0.79	.36***	.43***	—
4. ICSE	4.15	0.91	.06**	.34***	.36***	—
5. PCD	4.28	1.14	.09***	−.05*	−.03	−.24***	—
6. Psychological adapt.	4.52	1.06	−.07**	.26***	.28***	.52***	−.31***	—
7. Sociocultural adapt.	4.74	0.96	−.04*	.31***	.33***	.58***	−.35***	.68***	—
8. Age	23.67	3.82	−.09***	−.06**	.03	.15***	−.07**	.10***	.12***	—
9. Length of residence	18.45	10.23	.05*	.08***	.10***	.26***	−.14***	.21***	.29***	.45***	—
10. Chinese proficiency	2.89	1.12	.07**	.14***	.17***	.41***	−.21***	.34***	.45***	.20***	.37***

Note. N = 2,478. ICSE = intercultural communication self-efficacy; PCD = perceived cultural distance; adapt. = adaptation.

p < .05. **p < .01. ***p < .001.

Intercultural communication self-efficacy was positively associated with both psychological adaptation (r = .52, p < .001) and sociocultural adaptation (r = .58, p < .001), representing large effect sizes and supporting the theoretical centrality of capability beliefs in adaptation. Perceived cultural distance showed negative associations with intercultural communication self-efficacy (r = −.24, p < .001) and both adaptation outcomes (rs = −.31 to −.35), consistent with theoretical expectations that greater subjective cultural distance poses challenges for adaptation.

Common Method Variance Assessment

Two complementary approaches were used to assess common method variance. First, Harman’s single-factor test indicated that when all items were loaded onto a single factor in exploratory factor analysis, this factor explained only 24.8% of the total variance, well below the 50% threshold suggesting problematic common method variance. Second, we employed a more rigorous latent method factor approach in confirmatory factor analysis, comparing model fit and path coefficients between models with and without an unmeasured latent method factor (Podsakoff et al., 2003). The addition of a common method factor produced minimal improvement in model fit (ΔCFI = 0.006; ΔRMSEA = 0.002), and the average variance extracted by the method factor was only 2.8%. Critically, when comparing path coefficients between the baseline model and the method-factor-controlled model, substantive coefficients changed by an average of only 0.02 (range: 0.00–0.04), and no conclusions regarding statistical significance changed. Together, these results suggest that common method variance does not pose a serious threat to the validity of the study findings.

Measurement Invariance Across Language Groups

Multi-group confirmatory factor analysis was conducted to assess measurement invariance of the six multi-item scales across the five language groups (English, n = 743; French, n = 412; Vietnamese, n = 523; Thai, n = 456; Indonesian, n = 344). Results are presented in Table 2. The configural model, which allowed factor loadings and intercepts to vary freely across groups while constraining factor structures to be equivalent, demonstrated acceptable fit (CFI = 0.951, TLI = 0.945, RMSEA = 0.044), indicating equivalent factor structures across groups. Moving from configural to metric invariance (constraining factor loadings equal across groups) produced minimal decrease in CFI (ΔCFI = −0.005) and negligible increase in RMSEA (ΔRMSEA = 0.002), supporting the equivalence of factor loadings and indicating that the constructs have the same meaning across language groups. The scalar invariance model (additionally constraining item intercepts equal) also demonstrated acceptable fit degradation (ΔCFI = −0.007; ΔRMSEA = 0.003), supporting the equivalence of item intercepts. Full scalar invariance permits meaningful cross-group comparisons of latent means and structural paths.

Table 2.

Measurement Invariance Testing Across Language Groups.

Model	χ²	df	CFI	TLI	RMSEA	ΔCFI	ΔRMSEA
Configural	2,756.43	1,520	0.951	0.945	0.044	—	—
Metric	2,918.27	1,648	0.946	0.941	0.046	−0.005	0.002
Scalar	3,124.58	1,776	0.939	0.935	0.049	−0.007	0.003

Note. Δ indicates change from previous (less constrained) model. Criteria for supporting invariance: ΔCFI ≥−0.010 and ΔRMSEA ≤0.015 (Chen, 2007). CFI = comparative fit index; TLI = Tucker-Lewis index; RMSEA = root mean square error of approximation.

Structural Model and Hypothesis Testing

The hypothesized conditional process model demonstrated good fit to the data: χ²(24) = 58.67, p < .001, CFI = 0.984, TLI = 0.958, RMSEA = 0.024 (90% CI [0.016, 0.032]), SRMR = 0.016. The significant chi-square is expected given the large sample size and is not indicative of model misfit when alternative fit indices suggest good fit (Kline, 2023). The model explained 38.6% of the variance in intercultural communication self-efficacy (R² = .386), 41.2% of the variance in psychological adaptation (R² = .412), and 49.8% of the variance in sociocultural adaptation (R² = .498). These variance-explained values represent substantial explanatory power, exceeding benchmarks for cross-cultural adaptation research (Wilson et al., 2013). Standardized path coefficients and their significance tests are presented in Table 3.

Table 3.

Standardized Path Coefficients for the Conditional Process Model.

Path	β	SE	t	95% CI
Paths to intercultural communication self-efficacy
Passive use → ICSE	.02	0.02	0.82	[−0.02, 0.06]
Active social use → ICSE	.26***	0.02	12.45	[0.22, 0.30]
Active non-social use → ICSE	.24***	0.02	11.32	[0.20, 0.28]
PCD → ICSE	−.17***	0.02	−8.34	[−0.21, −0.13]
Passive use × PCD → ICSE	−.10**	0.02	−4.12	[−0.14, −0.05]
Active social use × PCD → ICSE	−.13***	0.02	−5.68	[−0.17, −0.08]
Active non-social use × PCD → ICSE	−.09**	0.02	−3.78	[−0.14, −0.04]
Paths to psychological adaptation
Passive use → Psych. adapt.	−.11***	0.02	−5.23	[−0.15, −0.07]
Active social use → Psych. adapt.	.12***	0.02	5.67	[0.08, 0.16]
Active non-social use → Psych. adapt.	.13***	0.02	6.12	[0.09, 0.17]
ICSE → Psych. adapt.	.42***	0.02	20.34	[0.38, 0.46]
PCD → Psych. adapt.	−.21***	0.02	−10.12	[−0.25, −0.17]
Paths to sociocultural adaptation
Passive use → Sociocult. adapt.	−.07**	0.02	−3.24	[−0.11, −0.03]
Active social use → Sociocult. adapt.	.14***	0.02	6.54	[0.10, 0.18]
Active non-social use → Sociocult. adapt.	.16***	0.02	7.45	[0.12, 0.20]
ICSE → Sociocult. adapt.	.46***	0.02	22.67	[0.42, 0.50]
PCD → Sociocult. adapt.	−.24***	0.02	−11.56	[−0.28, −0.20]

Note. Model controlled for age, gender, length of residence, Chinese proficiency, and prior overseas experience. All VIFs <2.5. ICSE = intercultural communication self-efficacy; PCD = perceived cultural distance; Psych. = psychological; Sociocult. = sociocultural; adapt. = adaptation.

p < .01. ***p < .001.

Direct Effects

Results for direct effects are presented in Table 3. Regarding paths from social media use types to the mediator (intercultural communication self-efficacy), the three use types showed markedly different patterns. Active social use showed a medium-sized positive association with intercultural communication self-efficacy (β = .26, p < .001, 95% CI [0.22, 0.30]). Similarly, active non-social use was positively associated with intercultural communication self-efficacy (β = .24, p < .001, 95% CI [0.20, 0.28]). In contrast, passive use showed a near-zero and nonsignificant association with intercultural communication self-efficacy (β = .02, p = .412, 95% CI [−0.02, 0.06]). The narrow confidence interval around this estimate, spanning only 0.08 units and including zero, indicates that even if an association exists in the population, its magnitude is likely very small. These results strongly support H1, demonstrating that the type of social media engagement matters: active forms of use that involve social interaction or purposeful information seeking are associated with enhanced intercultural communication capability beliefs, whereas passive consumption is not.

Intercultural communication self-efficacy showed strong positive associations with both adaptation outcomes. The path to psychological adaptation was large and significant (β = .42, p < .001, 95% CI [0.38, 0.46]), as was the path to sociocultural adaptation (β = .46, p < .001, 95% CI [0.42, 0.50]). These large effects underscore the central role of capability beliefs in cross-cultural functioning, consistent with self-efficacy theory (Bandura, 1977) and prior research on intercultural competence (Kabir & Sponseller, 2020).

Regarding direct effects on adaptation outcomes (controlling for the indirect pathway through the mediator), the three social media use types again showed differentiated patterns. Passive use showed a small but significant negative direct effect on psychological adaptation (β = −.11, p < .001, 95% CI [−0.15, −0.07]) and a smaller negative effect on sociocultural adaptation (β = −.07, p = .001, 95% CI [−0.11, −0.03]). This pattern suggests that passive social media consumption may have direct detrimental associations with well-being beyond its non-relationship with self-efficacy, consistent with prior findings linking passive use to social comparison and reduced affect (Verduyn et al., 2015).

Active social use showed positive direct effects on both psychological adaptation (β = .12, p < .001, 95% CI [0.08, 0.16]) and sociocultural adaptation (β = .14, p < .001, 95% CI [0.10, 0.18]). Active non-social use similarly showed positive direct effects on psychological adaptation (β = .13, p < .001, 95% CI [0.09, 0.17]) and a somewhat larger effect on sociocultural adaptation (β = .16, p < .001, 95% CI [0.12, 0.20]). These direct effects suggest that active use types contribute to adaptation through pathways beyond self-efficacy enhancement, potentially including direct social support receipt, practical information acquisition, and maintenance of home-culture ties.

Indirect Effects Through Intercultural Communication Self-Efficacy

Bootstrap analysis results for indirect effects are presented in Table 4. Supporting H2, active social use showed significant indirect effects on both psychological adaptation (β = .11, 95% CI [0.09, 0.14]) and sociocultural adaptation (β = .12, 95% CI [0.10, 0.15]) through intercultural communication self-efficacy. These indirect effects represent a substantial proportion of the total effect: for psychological adaptation, the indirect effect accounted for 47.8% of the total effect (PM = 0.478); for sociocultural adaptation, 46.2% (PM = 0.462).

Table 4.

Indirect Effects of Social Media Use Types on Adaptation Through Intercultural Communication Self-Efficacy.

Indirect effect path	β	SE	95% Boot CI	PM
Psychological adaptation outcome
Passive use → ICSE → Psych. adapt.	.01	0.01	[−0.01, 0.03]	—
Active social use → ICSE → Psych. adapt.	.11	0.01	[0.09, 0.14]	0.478
Active non-social use → ICSE → Psych. adapt.	.10	0.01	[0.08, 0.13]	0.435
Sociocultural adaptation outcome
Passive use → ICSE → Sociocult. adapt.	.01	0.01	[−0.01, 0.03]	—
Active social use → ICSE → Sociocult. adapt.	.12	0.01	[0.10, 0.15]	0.462
Active non-social use → ICSE → Sociocult. adapt.	.11	0.01	[0.09, 0.14]	0.407

Note. Bootstrap confidence intervals based on 10,000 resamples. Effects with CIs not containing zero are significant. ICSE = intercultural communication self-efficacy; Psych. = psychological; Sociocult. = sociocultural; adapt. = adaptation; PM = proportion of the total association accounted for by the indirect path (indirect/total).

Similarly, active non-social use showed significant indirect effects on psychological adaptation (β = .10, 95% CI [0.08, 0.13]) and sociocultural adaptation (β = .11, 95% CI [0.09, 0.14]). The proportion of the total association accounted for by the indirect path was 43.5% for psychological adaptation and 40.7% for sociocultural adaptation. The indirect effects were stable across bootstrap replications, with narrow confidence intervals excluding zero, supporting the substantive importance of intercultural communication self-efficacy as a mechanism linking active social media engagement to cross-cultural adaptation in the present sample.

In contrast, passive use did not show significant indirect effects on either psychological adaptation (β = .01, 95% CI [−0.01, 0.03]) or sociocultural adaptation (β = .01, 95% CI [−0.01, 0.03]). This null finding is a direct consequence of passive use’s near-zero association with intercultural communication self-efficacy and reinforces the theoretical distinction between consumption-oriented and engagement-oriented social media behaviors.

Conditional Indirect Effects: Moderation by Perceived Cultural Distance

To test H3, we examined whether perceived cultural distance moderated the indirect effects. The indices of moderated mediation, which quantify whether indirect effects vary linearly with the moderator (Hayes, 2015, 2017), are presented in Table 5. For psychological adaptation, the indices of moderated mediation were significant for all three use types: passive use (index = −0.04, 95% CI [−0.07, −0.02]), active social use (index = −0.05, 95% CI [−0.08, −0.03]), and active non-social use (index = −0.04, 95% CI [−0.06, −0.02]). The negative indices indicate that the indirect effects become weaker (more attenuated) as perceived cultural distance increases. For sociocultural adaptation, the indices of moderated mediation were also significant for all three use types: passive use (index = −0.05, 95% CI [−0.08, −0.02]), active social use (index = −0.06, 95% CI [−0.09, −0.04]), and active non-social use (index = −0.04, 95% CI [−0.07, −0.02]).

Table 5.

Conditional Indirect Effects at Different Levels of Perceived Cultural Distance.

Effect	PCD level	β	SE	95% CI
Panel A: Psychological adaptation
Passive use → ICSE → Psych. adapt.	Low (−1SD)	.06	0.02	[0.03, 0.09]
	Mean	.01	0.01	[−0.01, 0.03]
	High (+1SD)	−.04	0.02	[−0.07, −0.01]
Index of moderated mediation		−.04	0.01	[−0.07, −0.02]
Active social use → ICSE → Psych. adapt.	Low (−1SD)	.16	0.02	[0.13, 0.20]
	Mean	.11	0.01	[0.09, 0.14]
	High (+1SD)	.06	0.02	[0.03, 0.09]
Index of moderated mediation		−.05	0.01	[−0.08, −0.03]
Active non-social use → ICSE → Psych. adapt.	Low (−1SD)	.14	0.02	[0.11, 0.17]
	Mean	.10	0.01	[0.08, 0.13]
	High (+1SD)	.06	0.02	[0.03, 0.09]
Index of moderated mediation		−.04	0.01	[−0.06, −0.02]
Panel B: Sociocultural adaptation
Passive use → ICSE → Sociocult. adapt.	Low (−1SD)	.07	0.02	[0.04, 0.11]
	Mean	.01	0.01	[−0.01, 0.03]
	High (+1SD)	−.05	0.02	[−0.08, −0.02]
Index of moderated mediation		−.05	0.02	[−0.08, −0.02]
Active social use → ICSE → Sociocult. adapt.	Low (−1SD)	.18	0.02	[0.14, 0.22]
	Mean	.12	0.01	[0.10, 0.15]
	High (+1SD)	.06	0.02	[0.03, 0.10]
Index of moderated mediation		−.06	0.01	[−0.09, −0.04]
Active non-social use → ICSE → Sociocult. adapt.	Low (−1SD)	.15	0.02	[0.12, 0.19]
	Mean	.11	0.01	[0.09, 0.14]
	High (+1SD)	.07	0.02	[0.04, 0.10]
Index of moderated mediation		−.04	0.01	[−0.07, −0.02]

Note. Bootstrap confidence intervals based on 10,000 resamples. ICSE = intercultural communication self-efficacy; PCD = perceived cultural distance; Psych. = psychological; Sociocult. = sociocultural; adapt. = adaptation. Low = 1 SD below the mean; High = 1 SD above the mean.

Simple slope analyses illuminated the nature of the moderation effects. For active social use, the conditional indirect effect on psychological adaptation was substantially stronger at low perceived cultural distance (β = .16, 95% CI [0.13, 0.20]) than at the mean (β = .11, 95% CI [0.09, 0.14]) or at high perceived cultural distance (β = .06, 95% CI [0.03, 0.09]). Notably, even at high levels of perceived cultural distance, the indirect effect remained positive and significant, indicating that the mechanism is attenuated but not eliminated under conditions of high cultural distance. A similar attenuation pattern was observed for sociocultural adaptation, where the conditional indirect effect decreased from β = .18 at low PCD to β = .06 at high PCD.

For passive use, an intriguing pattern emerged: the conditional indirect effect shifted from positive at low cultural distance (psychological: β = .06, 95% CI [0.03, 0.09]; sociocultural: β = .07, 95% CI [0.04, 0.11]) to negative at high cultural distance (psychological: β = −.04, 95% CI [−0.07, −0.01]; sociocultural: β = −.05, 95% CI [−0.08, −0.02]). This sign reversal warrants careful interpretation in light of the concurrent significant negative direct effects of passive use on adaptation (β = −.11 and −.07 for psychological and sociocultural adaptation, respectively). The total effect of passive use on adaptation outcomes combines these indirect effects with the direct effects. At low cultural distance, the small positive indirect effect partially offsets the negative direct effect, whereas at high cultural distance, both the indirect and direct effects operate in the same negative direction, compounding potential detrimental associations. This pattern suggests that under conditions of high perceived cultural distance, passive social media use may be particularly maladaptive, with neither the self-efficacy pathway nor other unmeasured pathways providing protective buffering.

The practical implication of this pattern is noteworthy: at low perceived cultural distance, a 1-SD increase in active social use was indirectly associated with a .16-SD increase in psychological adaptation through enhanced self-efficacy—a medium effect. At high perceived cultural distance, this indirect association was reduced to .06 SD—still significant but small. These findings support H3: perceived cultural distance acts as a boundary condition that constrains the magnitude of the indirect associations between social media use and adaptation through intercultural communication self-efficacy.

Supplementary Analyses

Four supplementary analyses were conducted to assess the robustness of findings. First, to account for potential nonindependence arising from students being nested within eight universities, we re-estimated the structural model using TYPE = COMPLEX with CLUSTER = university to obtain robust standard errors (Muthén & Muthén, 2017). Results were highly consistent with the primary analysis: all previously significant paths remained significant with robust standard errors, and coefficient estimates changed by an average of only .01 (range: .00–.02). The design effect (DEFF) for most variables was close to 1.0 (range: 1.02–1.18), indicating minimal clustering effects at the university level. This robustness check confirms that the key conclusions are not artifacts of the nested data structure.

Second, we re-estimated the model including primary social media platform type (dummy-coded with WeChat as the reference category) as additional covariates. Results were substantively unchanged: all significant paths remained significant with coefficient changes averaging .01 (range: .00–.03), and the pattern of conditional indirect effects was identical to the main analysis. This suggests that the findings generalize across platform types despite differences in platform affordances.

Third, we compared the magnitudes of indirect effects on psychological versus sociocultural adaptation by computing contrast tests. The indirect effect of active social use on sociocultural adaptation was marginally larger than its effect on psychological adaptation (Δβ = .01, 95% CI [−0.01, 0.03]), but this difference was not statistically significant. Similarly, the indirect effects of active non-social use did not differ significantly across outcomes (Δβ = .01, 95% CI [−0.01, 0.03]). This suggests that the indirect associations through intercultural communication self-efficacy operate with comparable strength across the two adaptation dimensions, consistent with self-efficacy’s theorized role as a general adaptive resource.

Fourth, to address potential concerns about the positive correlation between passive use and perceived cultural distance (r = .09), we examined whether this association might reflect a coping-driven pattern whereby students who perceive greater cultural distance increase their passive social media consumption. Post-hoc regression analyses supported this interpretation: perceived cultural distance was significantly associated with passive use (β = .11, p < .001) even after controlling for demographics. This pattern aligns with prior findings suggesting that passive social media use may function as an avoidance-based coping response under stress (Verduyn et al., 2015). The joint occurrence of high passive use and high cultural distance may thus represent a maladaptive cycle: cultural distance stress prompts passive consumption, which fails to enhance intercultural competence and may directly undermine well-being.

Discussion

The increasing complexity of global student mobility in a hyper-connected era has necessitated a more granular understanding of how digital behaviors intersect with psychological adjustment. While previous research has frequently treated digital engagement as a uniform construct, significant gaps remain in understanding how specific patterns of use relate to the cognitive mechanisms underlying cultural transition. This study addressed these gaps by examining a conditional process model that disentangles the associations between three types of social media use—passive, active social, and active non-social—and adaptation outcomes through the mediating pathway of intercultural communication self-efficacy. By identifying perceived cultural distance as a critical boundary condition, this research contributes a sophisticated theoretical framework that highlights both the integrative potential and the contextual limitations of digital tools in the cross-cultural experience.

Differentiated Pathways of Digital Engagement

The findings of this study provide empirical support for the functional differentiation of digital behaviors, suggesting that the qualitative nature of social media engagement is more relevant to adaptation than its mere frequency. Active social use and active non-social use both demonstrated significant positive associations with intercultural communication self-efficacy, whereas passive use showed a near-zero relationship. This pattern aligns with the social affordances perspective, which suggests that digital platforms provide distinct opportunities for cultural learning and agency that are only realized through purposeful interaction (Pang & Wang, 2020). Consistent with recent evidence, active engagement facilitates a “digital bridge” that correlates with better adjustment by allowing students to test cultural hypotheses and receive real-time feedback (Hofhuis et al., 2023; Pang & Wang, 2025; Rui & Wang, 2015; Yang et al., 2025). In contrast, the lack of association between passive consumption and efficacy beliefs suggests that the mere observation of host-culture content without active participation does not provide the mastery experiences necessary to build confidence in navigating a new environment.

Intercultural Communication Self-Efficacy as a Cognitive Engine

A central contribution of this research is the identification of intercultural communication self-efficacy as a consistent mediating pathway linking active social media engagement to both psychological adaptation and sociocultural adaptation. These results suggest that the benefits of social media are largely channeled through the enhancement of an individual’s internal capability beliefs. Recent studies have emphasized that for international students, the development of communicative confidence is a prerequisite for functional success in the host country (Hofhuis et al., 2023; Rui & Wang, 2015). Our findings extend this by suggesting that active social media use serves as a contemporary platform for cultural learning, where students can develop and refine their communicative scripts in a relatively low-risk environment. This cognitive empowerment appears to be a robust predictor of better emotional well-being and social functioning, supporting the theoretical centrality of self-efficacy in the acculturation process.

The Boundary Condition of Perceived Cultural Distance

The most theoretically salient discovery of this study is the moderating role of perceived cultural distance, which functions as a boundary condition that attenuates the adaptive associations of digital engagement. As the perceived gap between the home and host cultures increases, the positive indirect associations between active social media use and adaptation outcomes through intercultural communication self-efficacy are significantly weakened. This finding is consistent with the cultural fit hypothesis, which posits that the magnitude of cultural disparity determines the cognitive load and stress levels experienced by the sojourner (Ward & Geeraert, 2016). When international students perceive profound differences in values and social norms, the efficacy-building potential of online interaction appears to be “damped” by the overwhelming complexity of real-world cultural demands.

Furthermore, the observation that passive use correlates negatively with adaptation outcomes under conditions of high perceived cultural distance warrants careful consideration. In high-stress environments characterized by significant cultural disparity, students may turn to passive scrolling as a form of “digital cocooning” or avoidance-based coping (Pang & Wang, 2020; Xin et al., 2025; Zhou & Yin, 2024). This pattern reinforces a maladaptive cycle where digital retreat is associated with lower efficacy and increased psychological distress, highlighting that the “digital dual-edged sword” is most pronounced when cultural barriers are highest.

Contextual Considerations in the Chinese Digital Environment

The unique characteristics of the Chinese digital landscape, dominated by multifunctional platforms like WeChat, also play a role in these associations. Recent research among international students in China has indicated that active use of domestic platforms is associated with higher levels of social capital and smoother functional adjustment (Cao et al., 2023). The high degree of integration within these “super-apps” may explain why active non-social use—such as seeking practical information—is such a strong correlate of sociocultural adaptation in this specific sample. Unlike the more fragmented digital environments in many Western contexts, the all-encompassing nature of the Chinese digital ecosystem requires students to develop specific technical and cultural competencies that are directly linked to their daily survival and adaptation (Cao & Jia, 2023).

Limitations and Future Research

Despite the strengths of this research, including its large sample size and rigorous measurement invariance testing, several limitations should be noted. First, the cross-sectional survey design means that the identified pathways must be interpreted as associations rather than causal sequences. It remains possible that students who are already well-adapted and confident are more inclined to engage actively on social media. Future research should employ longitudinal or experience-sampling methods to examine the dynamic, reciprocal links between digital behaviors and adjustment over time. Second, although we implemented Brislin's (1970) standardized back-translation procedure, supplemented this with cognitive interviews for each language version, and established full scalar invariance across the five language groups via MGCFA, an important caveat warrants explicit acknowledgement: none of the focal scales (PAUM, SEIC-SF, Brief Perceived Cultural Distance Scale, BPAS, BSAS) has been formally psychometrically validated in independent Vietnamese, Thai, Indonesian, or French samples in the published cross-cultural literature. While our invariance testing supports the comparability of factor structures, loadings, and intercepts across the present language groups, this functional equivalence does not preclude the possibility of subtle semantic or pragmatic shifts at the item level that translation procedures alone cannot fully detect (Boehnke, 2022; Boer et al., 2018). Future research should pursue dedicated psychometric validation studies of these instruments in each linguistic and cultural community, and where feasible, complement self-report measures with behavioral or observational indicators that are less dependent on linguistic equivalence. Third, while this study accounted for perceived cultural distance, other individual difference variables, such as cultural intelligence or personality traits, might further moderate these relationships. Finally, although we controlled for platform types, future investigations could more explicitly examine how the specific technological affordances of different platforms (e.g., short-video apps vs. messaging apps) uniquely relate to the development of intercultural communication self-efficacy.

Conclusion

While the digital landscape has become an integral part of the international student experience, empirical clarity regarding the psychological mechanisms that convert online engagement into successful cross-cultural transition remained elusive, particularly concerning the boundary conditions imposed by cultural disparity. This study addressed these critical gaps by disentangling the differentiated associations of passive and active social media use through the mediating lens of intercultural communication self-efficacy among a large-scale sample of international students in China. The findings demonstrate that active social and active non-social use are positively associated with adaptation outcomes specifically through the enhancement of individuals’ internal capability beliefs, whereas passive consumption lacks this adaptive utility. Most importantly, the identification of perceived cultural distance as a significant moderator highlights that the psychological benefits of digital engagement are substantially attenuated when students encounter profound symbolic and functional gaps between their native and host cultures.

These findings offer critical practical implications for higher education institutions, suggesting that student support services should prioritize the cultivation of active digital competencies and targeted intercultural communication training rather than assuming universal digital empowerment. By validating intercultural communication self-efficacy as a core cognitive engine for adaptation, this research establishes that digital tools function most effectively when they serve as platforms for agency and cultural mastery. Ultimately, this study underscores that successful cross-cultural adjustment in the digital age is not merely a product of connectivity, but is fundamentally constrained by the subjective appraisal of cultural distance, positioning psychological capability as the primary determinant of a sojourner’s success in an increasingly digitized global academy.

Footnotes

Acknowledgements

We thank the participants and the university staff who facilitated data collection.

ORCID iD

Yang Xin

Ethical Considerations

This study was approved by the Ethics Committee of Guangxi University of Finance and Economics (No. 20240321). All procedures were performed in accordance with relevant guidelines and regulations.

Consent to Participate

Participation was voluntary, and all participants provided written informed consent prior to data collection.

Author Contributions

Y.X.: Conceptualization, Methodology, Formal analysis, Writing—original draft. H.W.N.: Data curation, Investigation, Writing—review & editing. D.S.H.: Investigation, Resources, Writing—review & editing. D.Y.: Supervision, Funding acquisition, Project administration, Writing—review & editing. All authors read and approved the final manuscript.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Guangxi Colleges and Universities Humanities and Social Sciences Key Research Base Fund Grant (Project No. 22JDB03), and the Guangxi Zhuang Autonomous Region New Medical Research and Practice Project (Project No. XYK202318). The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

De-identified data supporting the findings of this study are available from the corresponding author upon reasonable request, subject to ethics approval and data-protection considerations.*

References

Bandura

(1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84(2), 191–215. https://doi.org/10.1037//0033-295x.84.2.191

Berry

J. W.

(1997). Immigration, acculturation, and adaptation. Applied Psychology, 46(1), 5–34. https://doi.org/10.1111/j.1464-0597.1997.tb01087.x

Boehnke

(2022). Let’s compare apples and oranges! A plea to demystify measurement equivalence. American Psychologist, 77(9), 1160–1168. https://doi.org/10.1037/amp0001080

Boer

Hanke

(2018). On detecting systematic measurement error in cross-cultural research: A review and critical reflection on equivalence and invariance tests. Journal of Cross-Cultural Psychology, 49(5), 713–734. https://doi.org/10.1177/0022022117749042

Boski

Sharmin

Tariq

Ospanova

Kurapov

(2025). Bicultural competencies and identities in acculturation and intercultural relations. Journal of Cross-Cultural Psychology, 56(6), 629–662. https://doi.org/10.1177/00220221251333123

Brislin

R. W.

(1970). Back-translation for cross-cultural research. Journal of Cross-Cultural Psychology, 1(3), 185–216. https://doi.org/10.1177/135910457000100301

Cao

Jia

(2023). International students’ social use of WeChat and sociocultural adjustment: Coping self-efficacy as a longitudinal mediator. Education and Information Technologies, 29(5), 5189–5209. https://doi.org/10.1007/s10639-023-12044-5

Cao

Meng

Zhang

(2023). A longitudinal examination of WeChat usage intensity, behavioral engagement, and cross-cultural adjustment among international students in China. Higher Education, 87(3), 661–683. https://doi.org/10.1007/s10734-023-01029-5

Chen

F. F.

(2007). Sensitivity of goodness of fit indexes to lack of measurement invariance. Structural Equation Modeling: A Multidisciplinary Journal, 14(3), 464–504. https://doi.org/10.1080/10705510701301834

10.

Demes

K. A.

Geeraert

(2014). Measures matter: Scales for adaptation, cultural distance, and acculturation orientation revisited. Journal of Cross-Cultural Psychology, 45(1), 91–109. https://doi.org/10.1177/0022022113487590

11.

Enders

C. K.

(2022). Applied missing data analysis. Guilford Publications.

12.

Galchenko

van de Vijver

F. J. R.

(2007). The role of perceived cultural distance in the acculturation of exchange students in Russia. International Journal of Intercultural Relations, 31(2), 181–197. https://doi.org/10.1016/j.ijintrel.2006.03.004

13.

Gerson

Plagnol

A. C.

Corr

P. J.

(2017). Passive and Active Facebook Use Measure (PAUM): Validation and relationship to reinforcement sensitivity theory. Personality and Individual Differences, 117, 81–90. https://doi.org/10.1016/j.paid.2017.05.034

14.

Hair

J. F.

Black

W. C.

Babin

B. J.

Anderson

R. E.

(2019). Multivariate data analysis (8th ed.). Cengage Learning EMEA.

15.

Hayes

A. F.

(2015). An index and test of linear moderated mediation. Multivariate Behavioral Research, 50(1), 1–22. https://doi.org/10.1080/00273171.2014.962683

16.

Hayes

A. F.

(2017). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. Guilford Publications.

17.

Hofhuis

van Egmond

M. C.

Lutz

F. E.

von Reventlow

Rohmann

(2023). The effect of social network sites on international students’ acculturation, adaptation, and wellbeing. Frontiers in Communication, 8, Article 1186527. https://doi.org/10.3389/fcomm.2023.1186527

18.

L. T.

Bentler

P. M.

(1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1–55. https://doi.org/10.1080/10705519909540118

19.

Jurasek

Wawrosz

(2023). How does self-efficacy in communication affect the relationship between intercultural experience, language skills, and cultural intelligence. Sage Open, 13(4), Article 21582440231211687. https://doi.org/10.1177/21582440231211687

20.

Kabir

R. S.

Sponseller

A. C.

(2020). Interacting with competence: A validation study of the Self-Efficacy in Intercultural Communication Scale–Short Form. Frontiers in Psychology, 11, Article 2086. https://doi.org/10.3389/fpsyg.2020.02086

21.

Keith

T. Z.

(2019). Multiple regression and beyond: An introduction to multiple regression and structural equation modeling. Routledge.

22.

Kline

R. B.

(2023). Principles and practice of structural equation modeling. Guilford Publications.

23.

Zhu

Huang

Zhou

Shen

Qiu

Zong

Guo

Yashengjiang

Man

(2025). Institutional cultural norms, cultural motives for college and adjustment: Cultural mismatch theory in the Chinese context. Journal of Cross-Cultural Psychology, 56(2), 150–174. https://doi.org/10.1177/00220221241305094

24.

Maharaj

Ndwiga

Chutiyami

(2025). Mental health and wellbeing of international students in Australia: A systematic review. Journal of Mental Health, 34(4), 431–449. https://doi.org/10.1080/09638237.2024.2390393

25.

Malay

E. D.

Coelen

R. J.

Otten

(2024). The dynamics in the relationship between perceived cultural distance, cultural intelligence and adjustment of international students. International Journal of Intercultural Relations, 102, Article 102016. https://doi.org/10.1016/j.ijintrel.2024.102016

26.

Mao

(2025). Exploring factors associated with international students’ academic adjustment in mainland China: A systematic review from 2005 to 2023. Journal of Cross-Cultural Psychology, 56(5), 558–581. https://doi.org/10.1177/00220221251320348

27.

Maxwell

S. E.

Cole

D. A.

(2007). Bias in cross-sectional analyses of longitudinal mediation. Psychological Methods, 12(1), 23–44. https://doi.org/10.1037/1082-989x.12.1.23

28.

Muthén

L. K.

Muthén

B. O.

(2017). Mplus user’s guide (8th ed.).

29.

Pang

(2023). Determining the influence of depressive mood and self-disclosure on problematic mobile app use and declined educational attainment: Insight from stressor-strain-outcome perspective. Education and Information Technologies, 29(4), 4635–4656. https://doi.org/10.1007/s10639-023-12018-7

30.

Pang

Ruan

(2024). Disentangling composite influences of social connectivity and system interactivity on continuance intention in mobile short video applications: The pivotal moderation of user-perceived benefits. Journal of Retailing and Consumer Services, 80, Article 103923. https://doi.org/10.1016/j.jretconser.2024.103923

31.

Pang

Wang

(2020). Promoting or prohibiting: Understanding the influence of social media on international students’ acculturation process, coping strategies, and psychological consequences. Telematics and Informatics, 54, Article 101454. https://doi.org/10.1016/j.tele.2020.101454

32.

Pang

Wang

(2025). Deciphering dynamic effects of mobile app addiction, privacy concern and cognitive overload on subjective well-being and academic expectancy: The pivotal function of perceived technostress. Technology in Society, 81, Article 102861. https://doi.org/10.1016/j.techsoc.2025.102861

33.

Podsakoff

P. M.

MacKenzie

S. B.

Lee

J.-Y.

Podsakoff

N. P.

(2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88(5), 879–903. https://doi.org/10.1037/0021-9010.88.5.879

34.

Preacher

K. J.

Hayes

A. F.

(2008). Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods, 40(3), 879–891. https://doi.org/10.3758/brm.40.3.879

35.

Rui

J. R.

Wang

(2015). Social network sites and international students’ cross-cultural adaptation. Computers in Human Behavior, 49, 400–411. https://doi.org/10.1016/j.chb.2015.03.041

36.

Verduyn

Lee

D. S.

Park

Shablack

Orvell

Bayer

Ybarra

Jonides

Kross

(2015). Passive Facebook usage undermines affective well-being: Experimental and longitudinal evidence. Journal of Experimental Psychology: General, 144(2), 480–488. https://doi.org/10.1037/xge0000057

37.

Ward

Geeraert

(2016). Advancing acculturation theory and research: The acculturation process in its ecological context. Current Opinion in Psychology, 8, 98–104. https://doi.org/10.1016/j.copsyc.2015.09.021

38.

Wilson

Ward

Fischer

(2013). Beyond culture learning theory: What can personality tell us about cultural competence? Journal of Cross-Cultural Psychology, 44(6), 900–927. https://doi.org/10.1177/0022022113492889

39.

Xin

Shusheng

Liuheng

(2026). Pro-environmental volunteering in Chinese university youth: Peer selection, influence, and the roles of social media and social self-efficacy. Journal of Youth and Adolescence, 55(5), 1202–1218. https://doi.org/10.1007/s10964-026-02319-8

40.

Xin

Shusheng

Weina

Yan

(2025). How does digital connection shape cultural adaptation? The impact of social media use on the cross-cultural adaptation of international students in China. Journal of International Students, 15(9), 1–26.

41.

Yang

Deng

(2025). Associations between social media use and academic adjustment among university students: An integrated variable-centered and person-centered analytical framework. Problems of Education in the 21st Century, 83(3), 383–402.

42.

Zhai

Xiong

Almaawali

Tian

(2025). National trends of mental health and service utilisation among international students in the USA, 2015–2024. General Psychiatry, 38(5), Article e102124. https://doi.org/10.1136/gpsych-2025-102124

43.

Zhang

Ting

S.-H.

(2025). Mediating and moderating effects of WeChat use on cross-cultural adaptation among international students in China. International Journal of Intercultural Relations, 106, Article 102160. https://doi.org/10.1016/j.ijintrel.2025.102160

44.

Zhou

Yin

(2024). International students’ social media use: An integrative review of research over a decade. Journal of Studies in International Education, 29(1), 42–63. https://doi.org/10.1177/10283153241275037