Intercultural interaction willingness: a PLS-PM approach to influencing factors and its mediation effect

International students constitute a significant proportion of college populations globally and face multiple acculturative challenges (linguistic, cultural, lifestyle, academic, and discriminatory). Social connectedness with host nationals benefits language development, cultural knowledge, adjustment, and academic outcomes. China has rapidly grown as a destination for international students, making interaction with Chinese peers critical for sociocultural and psychological adjustment. This study investigates factors influencing international students’ willingness to interact with Chinese students, its association with social connection, and whether willingness mediates effects of language proficiency, cultural knowledge, perceived racism/discrimination, and environment on social connection. The study proposes nine hypotheses (H1–H9) covering positive effects of language proficiency, cultural knowledge, and environment; negative effects of racism/discrimination on willingness; a positive effect of willingness on social connection; and mediation of these relationships by willingness.

Theoretical framework: The study adopts Ward et al.’s (2020) Model of the Acculturation Process, integrating culture learning and stress-coping approaches, distinguishing societal (macro) and individual (micro: personal and situational) predictors of psychological and sociocultural adjustment. Variables selected: language proficiency, culture, racism/discrimination, and environment as predictors; interaction willingness as mediator; social connection as outcome. Language proficiency: Essential for intercultural interaction; higher proficiency improves intercultural sensitivity and communication competence, reduces sociocultural adjustment difficulty; low proficiency hinders learning and engagement; accents can trigger status perceptions and stress. Culture: Cultural knowledge underpins effective, appropriate communication and sociocultural adaptation; greater cultural distance associates with social difficulty and stress; cultural differences can limit common ground and render interactions superficial. Racism and discrimination: Perceived discrimination elevates life stress and mental health issues, inhibits interaction with host nationals, fosters negative attitudes toward host country, and reduces satisfaction with life. Environment: Encompasses academic, social networks, housing, social media, and activities; shapes contact opportunities and social support. Housing configuration, academic stress, and social media use (e.g., WeChat) can facilitate or impede interaction. Prior work notes sporadic contact and limited attention to interaction willingness; this study addresses that gap.

Design and participants: Cross-sectional survey of international students at two universities in Shandong Province, China. Final sample N=184 from 13 countries (4 unspecified). Mean age 22.48 (SD=1.93; 6 missing); 48.37% female (N=89), 51.63% male (N=95). Of 178 who reported origin: Asia 51.69% (N=92), Europe 40.45% (N=72), Africa 3.37% (N=6), Middle East 4.49% (N=8). Length of stay (N=179): <1 year 10.61% (N=19), >1 year 13.41% (N=24), >2 years 45.81% (N=82), >3 years 30.17% (N=54). Procedure: Surveys conducted May–June 2022, March–April 2023, and January 2024 with institutional approval and informed consent. Paper questionnaires administered in dormitory halls and cafeterias; anonymity ensured. Likert scale from 1 (strongly disagree) to 5 (strongly agree). Example items: language proficiency (e.g., “I'm reluctant to talk to Chinese students because my Chinese is not good”), culture, racism/discrimination (e.g., “Chinese students show hatred toward me”), environment (e.g., “I meet new Chinese students via online chat”). Incentive: RMB 5 yuan. Of 191 completions, 190 retrieved, 6 excluded for missing/invalid data; final N=184. Instruments: Five-page questionnaire: (1) Demographics (age, gender, nationality, length of stay, self-reported Chinese/English proficiency—optional); (2) Interaction preferences (online/social media and language preferences, friendship networks, number of Chinese friends, club memberships, frequency of interaction); (3) 43 Likert items across constructs: language proficiency (LP), cultural knowledge (CL), racism/discrimination (RD), environment (ET), interaction willingness (WL), and social connection (SC). Bilingual (English/Chinese); some items reverse-coded. Data preparation: Missing values (0.85%) imputed by mean replacement. Reverse-coded items processed. Modeling approach: Reflective measurement models estimated with SmartPLS 4 using consistent PLS-SEM (PLSc) and bootstrapping (5000 subsamples; bias-corrected and accelerated, BCa, CIs). Mediation assessed via segmentation approach and Zhao et al. (2010) framework: test indirect and direct effects, then classify mediation type via product of coefficients. Measurement model evaluation: Indicator reliability (outer loadings mostly >0.70, all significant at 1%); composite reliability via ρA (all ≥0.770); convergent validity via AVE (all >0.50); discriminant validity via HTMT (<0.85 across constructs). Global fit: SRMR <0.08; SRMR, dULS, dG below 95% quantile benchmarks across models. Item reduction: 24 of 43 items removed due to low loadings or impaired reliability/validity; final 19 items retained (construct means and SDs reported).

• Path directions: Chinese language proficiency (LP), cultural knowledge (CL), and environment (ET) positively predict interaction willingness (WL) and social connection (SC); racism and discrimination (RD) negatively predict WL and SC.
• Effect of WL on SC: WL→SC significant and positive across models: 0.566 (p<0.001, M1); 0.426 (p<0.01, M2); 0.605 (p<0.001, M3); 0.466 (p<0.001, M4).
• Mediation (specific indirect effects via WL): • LP→WL→SC = 0.322 (t=5.184, p<0.001; 95% BCa CI lower bound 0.229) • CL→WL→SC = 0.333 (t=2.752, p<0.01; 95% BCa CI lower bound 0.145) • RD→WL→SC = −0.458 (t=3.829, p<0.001; 95% BCa CI lower bound −0.690) • ET→WL→SC = 0.314 (t=3.603, p<0.001; 95% BCa CI lower bound 0.199) All indirect effects significant, indicating mediation by WL. Because direct effects were also significant, mediation type is complementary (product of direct and indirect coefficients positive) in all models.
• Measurement quality: ρA reliability ~0.770–0.870; AVE >0.50 for all constructs; HTMT <0.85; SRMR well below 0.08, indicating satisfactory reliability, validity, and model fit.
• Descriptive highlights: High self-perceived Chinese proficiency (89.67% reported reasonably good); low mean scores for perceived discrimination items (e.g., hatred 1.80, safety concern 1.86); friendly environment and high social media integration (WeChat) with 87.50% reporting exclusive use for interactions.

Findings align with literature linking higher host-language proficiency to improved intercultural adjustment and social ties. In this sample, most students reported reasonably good Chinese proficiency, which reduced language barriers, increased interaction willingness, and facilitated rapport and friendships with Chinese peers. Cultural knowledge likewise supported willingness and connectedness, likely aided by longer residence (about 76% had >2 years in China) and friendships with host nationals. Environment, particularly via online platforms (notably WeChat), provided frequent, low-friction contact opportunities that fostered relational intimacy, satisfaction, and network expansion, thereby enhancing social support and buffering acculturative stress. Conversely, higher perceived racism/discrimination reduced willingness; however, reported discrimination was generally low in this sample, consistent with a friendly sociocultural context that encourages broader engagement with the host community. Overall, interaction willingness emerged as a key mechanism linking individual and environmental factors to social connection outcomes, underscoring the attitudinal pathway in intercultural adaptation.

The study elucidates how language proficiency, cultural knowledge, perceived racism/discrimination, and environment influence international students’ social connection with Chinese peers, with interaction willingness functioning as a central mediator. Results highlight the importance of multilingualism, length of stay, empathetic host attitudes, and supportive, digitally enabled environments (e.g., WeChat) in reducing barriers and strengthening social ties. The validated measurement models perform well in the Chinese cultural context among students with >2 years of residence. Future research should test the models across diverse cultural contexts, at different sojourn stages (including new arrivals), and broaden environmental factors (administrative processes, faculty interactions, academic/financial stress, housing, and leisure activities).

• Sample size is relatively small and concentrated (majority from Bangladesh and Russia), limiting generalizability across nationalities and cultural backgrounds.
• Cross-sectional design with most participants having resided in China for at least two years; length of stay may bias findings, and results may differ for recent arrivals with greater social difficulties.
• Environment construct scope was limited mainly to online social media and on-campus activities; other relevant environmental dimensions (administrative support, faculty, academic and financial stress, housing, leisure) were not fully captured.