Skip to main content An official website of the United States government Here's how you know Here's how you know Official websites use .gov A .gov website belongs to an official government organization in the United States. Secure .gov websites use HTTPS A lock ( Lock Locked padlock icon ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites. Search Log in Dashboard Publications Account settings Log out Search… Search NCBI Primary site navigation Search Logged in as: Dashboard Publications Account settings Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide PERMALINK Copy As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice BMC Psychol . 2026 Mar 9;14:524. doi: 10.1186/s40359-026-04265-3 Search in PMC Search in PubMed View in NLM Catalog Add to search Longitudinal trajectories of perceived stress during college transition: the protective roles of psychological resilience and social support Jing Han Jing Han 1 Institute of Developmental Psychology, Beijing Normal University, Beijing, China 2 Office of the Commission for Discipline Inspection, Jiangsu University of Technology, Changzhou, China Find articles by Jing Han 1, 2 , Meijing Chen Meijing Chen 1 Institute of Developmental Psychology, Beijing Normal University, Beijing, China Find articles by Meijing Chen 1 , Xiaoyi Wu Xiaoyi Wu 1 Institute of Developmental Psychology, Beijing Normal University, Beijing, China Find articles by Xiaoyi Wu 1 , Mingjun Xie Mingjun Xie 1 Institute of Developmental Psychology, Beijing Normal University, Beijing, China Find articles by Mingjun Xie 1 , Yu Bu Yu Bu 1 Institute of Developmental Psychology, Beijing Normal University, Beijing, China Find articles by Yu Bu 1 , Qinglin Xu Qinglin Xu 1 Institute of Developmental Psychology, Beijing Normal University, Beijing, China 3 Mental Health Education and Counseling Center, Beijing Wuzi University, Beijing, China Find articles by Qinglin Xu 1, 3 , Danhua Lin Danhua Lin 1 Institute of Developmental Psychology, Beijing Normal University, Beijing, China Find articles by Danhua Lin 1, ✉ Author information Article notes Copyright and License information 1 Institute of Developmental Psychology, Beijing Normal University, Beijing, China 2 Office of the Commission for Discipline Inspection, Jiangsu University of Technology, Changzhou, China 3 Mental Health Education and Counseling Center, Beijing Wuzi University, Beijing, China ✉ Corresponding author. Received 2025 Dec 19; Accepted 2026 Feb 25; Collection date 2026. © The Author(s) 2026 Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/ . PMC Copyright notice PMCID: PMC13077879 PMID: 41804003 Abstract Background The transition to college triggers significant stress that adversely affects student outcomes. Perceived stress during this period critically influences psychological well-being and physical health. However, longitudinal stress trajectories and their underlying protective factors remain poorly understood. Methods A four-wave longitudinal design tracked 2326 Chinese first-year college students ( M age = 18.16, SD = 0.57; 55.6% female) over two years. Validated measures assessed perceived stress, psychological resilience, social support, and adjustment outcomes. Distinct perceived stress trajectories were identified using latent class growth analysis. Predictors of trajectory membership were examined via multinomial logistic regression, adjusting for baseline covariates. Differences in distal outcomes across trajectories were tested using the Bolck-Croon-Hagenaars method, with baseline outcome levels statistically controlled. Results Latent class growth analysis revealed three stress trajectories: high-stable (60.19%), middle-decreasing-rebounding (29.58%), and low-decreasing-rebounding (10.23%). The latter predicted better adjustment and was associated with higher baseline resilience and social support. Crucially, a synergistic interaction showed that high resilience combined with high social support increased the odds of following this adaptive trajectory only among first-generation students. Conclusions Findings reveal heterogeneous stress pathways and identify a conditional resource synergy that buffers stress primarily for first-generation students, highlighting the need for integrated interventions targeting both internal and external resources. Supplementary Information The online version contains supplementary material available at 10.1186/s40359-026-04265-3. Keywords: College students, Perceived stress trajectories, Psychological resilience, Social support, Well-being Introduction Perceived stress occurs when individuals appraise environmental demands as exceeding their coping resources [ 1 ]. The transition to college epitomizes such a period, requiring students to navigate multiple salient developmental tasks across academic, independent living, and social domains [ 2 , 3 ]. These challenges are closely associated with heightened anxiety, depressive symptoms, compromised well-being, and poorer physical health among students [ 4 , 5 ]. However, research to date has predominantly relied on cross-sectional or short-term longitudinal designs (e.g., spanning one semester), limiting insight into the longitudinal dynamics of perceived stress across this transition [ 6 , 7 ]. Clarifying these dynamics is crucial, as perceived stress is a key predictor of subsequent academic and psychological adaptation [ 8 , 9 ]. Notably, sustained high stress or an increasing stress trajectory over time—as opposed to transient spikes at enrollment—is more strongly associated with severe academic risks and psychological distress [ 8 ]. The cognitive appraisal theory of stress frames this as a dynamic process arising from continuous person-environment interactions [ 1 ]. Similarly, developmental contextualism posits that stress fluctuations during transitions reflect ongoing adjustments between an individual’s coping resources (e.g., resilience, social support) and changing environmental demands [ 10 ]. Therefore, a systematic examination of college students’ longitudinal stress trajectories and individual differences therein is essential for understanding their adaptation. Moreover, identifying protective factors that shape these trajectories can inform interventions to reduce stress and promote healthy development during this transition. Although theory and research highlight the roles of environmental and individual antecedents [ 1 ], key empirical gaps persist. Evidence remains limited on how these factors interact to shape perceived stress trajectories during this transition, and whether their influences differ between first-generation college students (FGCS; neither parent holds a college degree) and continuing-generation college students (CGCS; at least one parent holds a college degree). To address these gaps, the present study aims to model the developmental trajectories of perceived stress during the college transition among Chinese undergraduates; identify protective factors that predict these trajectories; and examine potential differences in these relationships between FGCS and CGCS. Longitudinal development trajectories of perceived stress Increasing evidence indicates that perceived stress during the college transition is associated with critical developmental outcomes. This period involves critical shifts in both individual development (the transition to adulthood) and environment (the move from secondary to higher education), requiring students to establish independence while adapting to new demands—a process that can generate considerable stress [ 10 ]. Such transition-related stress impacts both short-term adjustment and long-term development. Students who struggle to manage this stress often report higher academic burnout and poorer physical and mental health [ 11 ]. However, most existing research relies on cross-sectional or short-term longitudinal data, thereby neglecting the possibility that stress follows distinct developmental trajectories over time, which may themselves be linked to different adaptive outcomes. The dynamic nature of stress during this period can be understood through two complementary theoretical lenses. The cognitive appraisal theory of stress [ 1 ] emphasizes that stress arises from ongoing person-environment transactions, while developmental contextualism [ 10 ] highlights how persistent stress emerges from a “stage-environment mismatch”—a disconnect between an individual’s developmental needs and environmental supports. The college transition epitomizes such a mismatch [ 2 ], as students’ cognitive appraisals and coping resources (e.g., psychological resilience, social support) undergo continuous change [ 12 , 13 ]. Consequently, mapping the longitudinal trajectory of perceived stress is essential for understanding how these experiences co-evolve in tandem with developmental tasks and coping resources. The protective role of social support and psychological resilience Identifying protective factors associated with perceived stress during the college transition is crucial for facilitating student adaptation and designing effective interventions. Developmental contextualism posits that resilience arises from a combination of external and internal resources [ 14 ], a view consistent with cognitive appraisal theory, which frames stress as a product of individuals’ dynamic evaluations of environmental demands relative to their coping capacities [ 1 ]. Social support, as a vital external resource, is a well-established buffer against transition-related stress. Meta-analytic confirm that support from peers, family, and institutions protects against psychological distress and promotes adaptation among college students [ 15 ]. It may be especially salient in China’s collectivist cultural context, where strong social bonds and interdependent values heighten the importance of support networks for stress management [ 16 ]. Empirically, perceiving high-quality social support is linked to reduced vulnerability to academic and social challenges, as well as lower levels of anxiety and depression. Psychological resilience, as a vital internal resource, may also moderate stress appraisals during this period. Defined as the capacity to maintain or regain adaptive functioning amid adversity, resilience encompasses emotional regulation, an optimistic explanatory style, and goal persistence [ 13 , 17 ]. Students high in resilience are more likely to appraise stressors as manageable challenges, employ effective coping strategies, and consequently report lower stress and higher well-being [ 18 ]. Beyond the independent effects, the interaction between resilience and social support requires investigation. Developmental contextualism suggests that internal and external resources may jointly influence development through cumulative or compensatory processes [ 19 ]. A synergistic effect is possible, where high levels of both resources yield optimal outcomes [ 20 ]. Alternatively, a compensatory effect may occur, where abundance in one resource offsets scarcity in the other [ 21 ]. For example, social support might be most protective for those with lower resilience, while high resilience could foster proactive support-seeking, creating a virtuous cycle. Thus, these factors likely interact to shape stress trajectories across the college transition. Differences between FGCS and CGCS Compared to their Western counterparts [ 22 – 25 ], Chinese undergraduates (FGCS and CGCS) face distinct stressors, including a highly competitive college entrance examination system, elevated parental expectations, and regional disparities in educational resources [ 26 – 28 ]. FGCS often report more acute stress, experiencing higher academic pressure and psychological distress, which suggests their stress experiences and developmental trajectories during the college transition may differ substantially from those of CGCS. Although China lacks a localized theoretical framework for this phenomenon, cultural mismatch theory provides a cross-cultural lens [ 23 ]. It posits that the implicit middle-class norms of university settings (e.g., valuing independent expression and proactive help-seeking) often conflict with the communal cultural schemas FGCS bring from their families and communities (e.g., deference to authority and reliance on familial support). This cultural mismatch may create cumulative disadvantages that impair stress adaptation through two primary pathways: by hindering the development of psychological resilience, and by constraining the establishment and effective use of social support networks [ 24 , 25 ]. However, the strong social bonds and interdependence valued in China’s collectivist culture also offer unique resources for mobilizing support [ 16 , 26 – 28 ]. It suggests that, despite structural barriers, Chinese FGCS may still navigate stress by leveraging collective support systems, highlighting the need to examine the specific role of social support within this context. Crucially, existing empirical evidence on how protective factors such as social support and psychological resilience operate across these two groups is both limited and inconsistent. Some studies find that social support is more protective for FGCS’s academic persistence [ 29 ], while others show psychological resilience as a stronger predictor for CGCS’s success [ 25 ]. This inconsistency may arise because prior research has typically examined these resources in isolation, failing to test how internal and external resources interact synergistically [ 19 ]. According to resource synergy theory, such resources can interact in gain-enhancing or compensatory modes [ 14 ]. For FGCS, who often face relative resource scarcity, these synergistic mechanisms may be particularly critical. In summary, by jointly considering internal (psychological resilience) and external (social support) resources, it is vital to investigate how these factors influence perceived stress trajectories across FGCS and CGCS during the college transition. This investigation will yield a more nuanced understanding of how developmental assets shape stress across subgroups, providing an empirical basis for designing targeted interventions—especially for FGCS, who are often at a resource disadvantage. The present study To address literature gaps, this study utilizes two years of four-wave longitudinal data to pursue three primary aims. First, this study aims to delineate the varied longitudinal patterns of perceived stress during the transition to college among Chinese students. Based on theories of individual differences, it was hypothesized that distinct perceived stress trajectories (i.e., increasing, decreasing, and stable) would emerge (Hypothesis 1a). Furthermore, students on low perceived stress trajectories exhibit more positive adjustment outcomes (specifically, lower levels of anxiety and depressive symptoms, higher life satisfaction, and better physical health) than those on high or middle perceived stress trajectories (Hypothesis 1b). Second, this study aims to explore whether trajectory class membership in perceived stress trajectories is predicted by the main and interactive effects of social support and psychological resilience. It was hypothesized that students with higher levels of social support and psychological resilience are more likely to belong to a more adaptive perceived stress trajectory (through a main or interaction effect) (Hypothesis 2). Third, this study aims to explore the interactive effects of social support and psychological resilience on perceived stress trajectory patterns, as well as differences in such interaction between FGCS and CGCS. It is hypothesized that a significant three-way interaction exists among social support, psychological resilience, and student group (FGCS and CGCS); furthermore, FGCS with both higher social support and higher psychological resilience are expected to be classified into perceived stress trajectories associated with more positive adjustment outcomes (Hypothesis 3). Given the exploratory nature of this comparative inquiry, other patterns regarding the differential protective role of these assets will also be examined. Methods Participants and procedure This longitudinal study tracked the psychosocial adaptation of Chinese college students during their transition to college. Participants were recruited from three public universities in China, selected to ensure regional diversity: two in Beijing and one in Changzhou, Jiangsu Province. Data were collected via online surveys at four time points: at university entry (T1: October 2023), after the first semester (T2: March 2024), at the beginning of their second academic year (T3: October 2024), and after the third semester (T4: March 2025). To ensure data quality, each survey included three attention-check items. Responses were excluded if participants failed more than one check or completed the survey in less than five minutes, a threshold established during pilot testing. At T1, 2,760 students provided valid responses ( M age = 18.18, SD = 0.58; 54.3% female). Valid response counts were 2,654 at T2, 2,496 at T3, and 2,326 at T4. The analytic sample for trajectory analyses comprised 2,326 participants with complete data across all four waves ( M age = 18.16, SD = 0.57; 55.6% female), including 1,888 from Beijing and 438 from Jiangsu in Table S1. Attrition analyses revealed no significant baseline differences in demographics or key variables between those retained in the sample and those lost to follow-up ( ps > 0.10). Outcome measurement design The timing of outcome assessments was designed to align with their distinct theoretical dynamics during the transition. To establish pre-transition baselines, depressive and anxiety symptoms, life satisfaction, and physical health were first measured at T1. Depressive and anxiety symptoms, given their acute sensitivity to transitional stress and adaptation, were reassessed at T4. In contrast, life satisfaction and physical health, expected to change more gradually, were tracked at T2 and T4 to observe their cumulative development. In all analyses, T1 scores for each outcome were included as control variables in the latent growth models. This analytical step was essential to statistically isolate the unique longitudinal association between stress trajectories and subsequent change in outcomes, independent of pre-existing individual differences at baseline. Ethical considerations The study was approved by the Institutional Review Board of Beijing Normal University (BNU202310200155) and conducted in accordance with the Declaration of Helsinki. All participants provided electronic informed consent after reviewing information about the study’s purpose, procedures, confidentiality protections, and their right to withdraw without penalty. Each survey took approximately 15–30 min to complete. Measures Demographic variables Demographic covariates were included in the trajectory prediction models because of their documented links to perceived stress [ 7 , 30 ]. These comprised age, gender (0 = female, 1 = male), student group (FGCS and CGCS), and subjective socioeconomic status (SSS, measured using the MacArthur Scale of Subjective Social Status, 1–10). The geographic recruitment site (0 = Changzhou, Jiangsu; 1 = Beijing) was also controlled for. Family income and urbanicity were not included as separate covariates. This decision was methodologically grounded: (1) SSS serves as a well-validated, psychologically salient composite measure of socioeconomic background, and (2) the core FGCS/CGCS grouping variable inherently encapsulates key structural disparities in family resources and background; adding these specific covariates risked statistical overcontrol that would obscure the central social-structural divide under investigation. Furthermore, baseline (T1) scores of all outcome variables were also included as covariates to adjust for pre-existing differences. Perceived stress (T1-T4) Perceived stress was measured by the Chinese version of the 10-item Perceived Stress Scale (PSS; [ 31 ]). Participants rated how often they experienced each situation (e.g., “being upset by something unexpected”) in the last month on a 5-point Likert scale from 0 ( never ) to 4 ( very often ). After reverse-scoring four positive items, a total score was summed, with higher scores indicating greater stress. The scale demonstrated good internal consistency across waves (Cronbach’s α = 0.83–0.88). Predictors (T1) Psychological resilience Psychological resilience was assessed with the 10-item Connor-Davidson Resilience Scale (CD-RISC; [ 32 ]), which measures trait resilience. Participants rated their agreement with each statement (e.g., “I can adapt to change”) on a 5-point Likert scale from 1 ( never ) to 5 ( always ). A total score was summed, with higher scores indicating greater resilience (Cronbach’s α = 0.92). Social support Social support was measured with the 12-item Multidimensional Scale of Perceived Social Support (MSPSS; [ 33 ]). Participants rated their agreement with each statement (e.g., “There is a special person who is around when I am in need”) on a 7-point Likert scale from 1 ( strongly disagree ) to 7 ( strongly agree ). A total score was summed, with higher scores indicating greater social support (Cronbach’s α = 0.95). Outcomes (T4) Depressive symptoms Depressive symptoms were assessed using the 9-item Patient Health Questionnaire (PHQ-9; [ 34 ]), which measures symptom frequency over the past month. Items (e.g., “Feeling down, or hopeless”) were rated on a 4-point Likert scale (0 = not at all to 3 = nearly every day) . A total score (range: 0–27) was computed, with higher scores indicating greater severity. According to the diagnostic criteria reported by Ye et al. [ 35 ], a score of ≥ 9 indicates probable clinical depression in Chinese populations (Cronbach’s α = 0.92). Anxiety symptoms Anxiety symptoms were measured using the 7-item Generalized Anxiety Disorder scale (GAD-7; [ 36 ]) to assess symptom frequency over the past month. Items (e.g., “Feeling nervous, anxious, or on edge”) were rated on a 4-point Likert scale (0 = not at all to 3 = nearly every day ). A total score (range: 0–21) was computed, with higher scores reflecting greater severity. According to the diagnostic criteria reported by Ip et al. [ 37 ], a score of ≥ 7 indicates probable clinical anxiety in Chinese populations (Cronbach’s α = 0.93). Life satisfaction Life satisfaction was assessed with the 5-item Satisfaction with Life Scale (SWLS; [ 38 ]). Participants rated their agreement with each item (e.g., “In most ways my life is close to ideal”) on a 7-point Likert scale from 1 ( strongly disagree ) to 7 ( strongly agree ). Responses were summed to yield a total score, with higher scores reflecting greater life satisfaction (Cronbach’s α = 0.96). Physical health Physical health was measured by the 7-item Index of Somatic Symptoms (ISS; [ 39 ]). Participants rated symptom frequency (e.g., “headaches, fatigue”) during the previous week on a 5-point Likert scale (0 = not at all to 4 = extremely ). A total score was calculated by summing all items, with higher scores indicating greater somatic symptom severity (Cronbach’s α = 0.91). Data analytic strategy Descriptive analyses for all the main variables were first conducted using SPSS (Version 27.0). The trajectories of participants’ perceived stress were then examined through latent class growth analysis (LCGA) in Mplus (Version 8.3; [ 40 ]). LCGA is a well-established method for uncovering heterogeneity in longitudinal data by modeling finite mixture distributions within a sample [ 41 ]. Full information maximum likelihood estimation was used to handle missing data, as this has been demonstrated to offer more reliable standard errors than other methods (e.g., mean imputation, listwise deletion, or pairwise deletion) [ 42 , 43 ]. An unconditional latent growth curve model (LGCM) was first derived to explore the overall trajectory for all participants. An LCGA with two to five class solutions was then performed with the variance of all growth factors constrained to zero (i.e., intercept and linear slope). To account for differences in the time between assessments, we fixed the factor loadings of the slope estimate to 0, 1, 2, and 3 (corresponding to baseline, 6 months, 12 months, and 18 months, respectively) in both LGCM and LCGA. Decisions on the optimal solution of growth trajectories were based on theoretical interpretations and statistical considerations, including the Akaike information criterion (AIC), smallest Bayesian information criterion (BIC), sample-size adjusted Bayesian information criterion (Adj-BIC), bootstrapped likelihood ratio test (BLRT), adjusted Lo-Mendell-Rubin likelihood ratio test (Adj-LMR-LRT), entropy, and the smallest class size (i.e., a minimum of 5% or more in each class) [ 44 ]. As suggested by previous research, lower values of BIC and Adj-BIC indicate a more parsimonious model [ 41 ], a value of entropy closer to 1 reflects greater precision (i.e., above 0.8) [ 44 ], and a significant p value of BLRT and Adj-LMR-LRT suggests that k classes are superior to k − 1 classes [ 45 ]. To test for differences in developmental outcomes across the identified stress trajectory classes, the manual Bolck-Croon-Hagenaars (BCH) procedure in Mplus was utilized (see Fig. 1 ). This method is preferred as it accounts for classification error in latent class assignment, thereby preserving the stability of the classes when comparing continuous distal outcomes [ 46 , 47 ]. Mean differences at T4 in depressive symptoms, anxiety symptoms, life satisfaction, and physical health were tested. Critically, the baseline (T1) level of each respective outcome was included as a continuous covariate in all BCH models. This adjustment isolates the effect of trajectory membership on subsequent change, independent of baseline levels. Predicting Class Membership: The Three-Step Approach (R3STEP) Predictors of latent class membership were examined using the automatic three-step approach (R3STEP command in Mplus), which treats covariates and predictors as auxiliary variables. This approach prevents the predictors from influencing the initial formation of the latent classes, allowing for unbiased estimation of their effects on classification [ 46 , 48 ]. A sequence of three multinomial logistic regression models was tested: Model 1 examined the main effects of the key predictors—psychological resilience and social support—while controlling for demographic covariates (gender, location). Model 2 added the two-way interaction between psychological resilience and social support. Model 3 added the three-way interaction between resilience, social support, and FGCS status to test for differential protective mechanisms. To facilitate the interpretation of interaction effects, all continuous predictors (resilience, social support) were centered via z-standardization prior to creating interaction terms. In the case of a significant three-way interaction, follow-up simple-slope analyses were conducted to examine the nature of the conditional effects within each student group (FGCS and CGCS). Fig. 1. Open in a new tab Conceptual model of the current study. Note . PS: Perceived stress; PR: Psychological resilience; SS: Social support; FG: First-generation college student status; DS: Depressive symptoms; AS: Anxiety symptoms; LS: Life satisfaction; PH: Physical health Results Common method bias test Given that our key constructs (e.g., perceived stress, depressive symptoms) were measured via self-report, we assessed common method bias using Harman’s single-factor test. An exploratory factor analysis including all items from the core variables measured at T1 revealed that the largest single factor accounted for 28.64% of the total variance, which is below the conventional threshold of 50%, suggesting that common method bias is unlikely to pose a serious threat [ 49 ]. Descriptive analyses Table 1 presents descriptive statistics and correlations among the main variables. Psychological resilience and social support showed consistent negative correlations with perceived stress across all four waves ( ps < 0.001). Perceived stress was positively correlations with depressive symptoms and anxiety symptoms at T4 ( p < 0.001). Furthermore, higher perceived stress at wave T4 was associated with lower life satisfaction and greater physical health ( ps < 0.001). Table 1. Descriptive statistics and correlation among main variables ( N = 2326) 1 2 3 4 5 6 7 8 9 10 1. PS (T1) - - - - - - - - - - 2. PS (T2) 0.46 *** - - - - - - - - - 3. PS (T3) 0.40 *** 0.58 *** - - - - - - - - 4. PS (T4) 0.39 *** 0.51 *** 0.63 *** - - - - - - - 5. PR (T1) -0.43 *** -0.36 *** -0.33 *** -0.30 *** - - - - - - 6. SS (T1) -0.45 *** -0.36 *** -0.34 *** -0.31 *** 0.50 *** - - - - - 7. DS (T4) 0.25 *** 0.34 *** 0.42 *** 0.54 *** -0.21 *** -0.23 *** - - - - 8. AS (T4) 0.24 *** 0.32 *** 0.39 *** 0.50 *** -0.16 *** -0.21 *** 0.74 *** - - - 9. LS (T4) -0.28 *** -0.42 *** -0.51 *** -0.61 *** 0.33 *** 0.29 *** -0.44 *** -0.36 *** - - 10. PH (T4) 0.19 *** 0.28 *** 0.34 *** 0.42 *** -0.16 *** -0.21 *** 0.58 *** 0.55 *** -0.32 *** - 11. Gender 0.01 0.07 *** 0.08 *** 0.08 *** -0.11 *** -0.04 0.02 -0.01 -0.17 *** -0.01 12. Age 0.01 0.02 -0.01 0.02 -0.02 -0.01 0.01 -0.01 -0.02 -0.01 13. SSS -0.17 *** -0.13 *** -0.11*** -0.11 *** 0.21 *** 0.19 *** -0.07 ** -0.04 * 0.14 *** -0.06 ** M 17.09 15.72 15.41 14.94 47.37 19.65 3.78 2.54 21.15 8.17 SD 6.70 6.37 6.21 6.16 10.80 6.14 4.20 3.52 5.10 4.08 Open in a new tab PS Perceived stress, PR Psychological resilience, SS Social support, DS Depressive symptom, AS Anxiety symptoms, LS Life satisfaction, PH Physical health, SSS Subjective socioeconomic status; * p < 0.05, ** p < 0.01, *** p < 0.001 Distinct trajectories of perceived stress An unconditional LGCM was specified to model the change in perceived stress over time. Analyses of the unconditional LGCM for perceived stress indicated that a quadratic model provided a good fit to the data ( χ² = 11.53, df = 5, CFI = 0.99, TLI = 0.97, RMSEA = 0.06, SRMR = 0.01). A linear model was examined, but it did not fit the data well for perceived stress. The mean of the intercept ( b = 17.56, SE = 0.13, p < 0.001) and the slope ( b = -1.07, SE = 0.16, p < 0.001) were statistically significant, indicating that the participants, on average, reported high levels of perceived stress during the initial transition period, which tends to intensify over time. Furthermore, results showed that the variance of both the intercept ( b = 15.83, SE = 2.28, p < 0.001) and the slope ( b = 11.14, SE = 3.29, p < 0.001) were significantly different from zero, indicating that there were significant interindividual differences in the levels of participants’ perceived stress at baseline and in how their perceived stress changed over time. These results suggest possible heterogeneity in perceived stress trajectories, supporting the presence of distinct subgroups following different patterns of change. Table 2 displayed the fit indices for the unconditional LCGA with two to five classes of perceived stress trajectories. LCGA was subsequently used to identify unobserved subgroups. Preliminary growth mixture modeling (GMM) analyses were attempted but failed to converge [ 44 ], supporting the selection of LCGA as the more methodologically stable approach for these data in Tables S2-S3. Given their proven accuracy in identifying the true number of classes across varying sample sizes [ 44 ], BLRT and BIC were selected as the most reliable indicators for model selection. All fit indices suggested that models with more classes were statistically acceptable, with the five-class solution showing adequate fit. However, four- and five-class models were excluded because they contained classes with less than 5% of the sample. The three-class solution was therefore selected as optimal. It showed superior fit over the two-class model, as indicated by consistently lower AIC/BIC values and a significant Adj-LMR-LRT. While entropy was slightly lower than in the two-class model, it remained above 0.80, suggesting classification accuracy exceeded 90%, which fully aligns with the core objectives of this study (see Fig. 2 ). Class 1 contained 60.19% ( n = 1400) of the sample and described a high-stable trajectory, with high perceived stress at baseline that remained elevated over time (intercept: b = 19.40, SE = 0.19, p < 0.001; quadratic slope: b = − 0.55, SE = 0.06, p < 0.001). Class 2 contained 29.58% ( n = 688) of the sample and described a middle-decreasing-rebounding trajectory, with middle baseline perceived stress, showing an initial decline followed by a subsequent rebounding (intercept: b = 14.56, SE = 0.33, p < 0.001; quadratic slope: b =0.38, SE = 0.13, p < 0.001). Class 3 contained 10.23% ( n = 238) of the sample and described a low-decreasing-rebounding trajectory, with low baseline perceived stress, showing an initial decline followed by a subsequent rebounding (intercept: b = 10.51, SE = 0.74, p < 0.001; quadratic slope: b = 1.27, SE = 0.20, p < 0.001). Table 2. Model fit statistics for latent class growth analyses results ( N = 2326) Classes AIC BIC Adj-BIC Entropy Adj-LMR-LRT ( p value) BLRT ( p value) N (%) 1 60833.53 60873.79 60851.55 (N/A) (N/A) (N/A) 2326 (100%) 2 58195.28 58258.55 58223.60 0.86 < 0.001 < 0.001 673 (28.93%) 3 57773.39 57859.67 57812.01 0.80 < 0.001 < 0.001 238 (10.23%) 4 57642.26 57751.54 57691.18 0.81 < 0.001 < 0.001 69 (2.97%) 5 57488.98 57621.27 57548.20 0.82 < 0.01 < 0.001 89 (3.83%) Open in a new tab The final class solution is bolded. AIC Akaike information criterion, BIC Bayesian information criterion, Adj-BIC Sample-size adjusted Bayesian information criterion, Adj-LMR-LRT Adjusted Lo-Mendell-Rubin likelihood test, BLRT Bootstrapped likelihood ratio test Fig. 2. Open in a new tab Trajectories of perceived stress among college students over 18 months. Note . Three latent classes were identified: High-stable (60.19%, n = 1,400), maintaining elevated stress; Middle-decreasing-rebounding (29.58%, n = 688), showing an initial moderate level followed by a decrease and subsequent rebound; and Low-decreasing-rebounding (10.23%, n = 238), showing an initial low stress level that further decreased before rebounding. Points represent estimated marginal means at each measurement (0, 6, 12, 18 months) To further validate between- and within-group differences in perceived stress, supplementary analyses of variance (ANOVA) and paired-samples t -tests were conducted. In Table S4, there were significant differences between the three groups regarding perceived stress at each wave ( p < 0.001). The low-decreasing-rebounding trajectory reported significantly lower perceived stress at T1 than the middle-decreasing-rebounding and high-stable trajectories ( ps < 0.001). Stress levels decreased from T1 to T2/T3 ( ps < 0.001), followed by a slight T4 rebound that remained below T1 ( p < 0.001) and comparable to T2/T3 ( ps > 0.05). The middle-decreasing-rebounding trajectory also showed lower T1 perceived stress than the high-stable trajectory ( p < 0.001) and a significant decreasing trend over time ( p < 0.001). In contrast, the high-stable trajectory maintained consistently elevated perceived stress with no significant change ( p = 0.336). The trajectories of perceived stress differed significantly between FGCS and CGCS across the four time points. Independent-samples t -tests were used to compare perceived stress between FGCS and CGCS at each time point. No significant differences were found at T1 ( t (1417.41) = 1.68, p = 0.093) or T4 ( t (1434.20) = − 0.04, p = 0.967). However, FGCS reported significantly higher perceived stress than their continuing-generation peers at both T2 ( t (1422.29) = 2.90, p = 0.004) and T3 ( t (1393.18) = 2.31, p = 0.021). Distal outcomes associated with stress trajectories As presented in Table 3 , the BCH method revealed significant differences across trajectory classes in all outcome measures. The low-decreasing-rebounding trajectory demonstrated the most favorable outcomes, including the highest life satisfaction ( p < 0.001), the lowest levels of depressive ( p < 0.001) and anxiety symptoms ( p < 0.001), and the best physical health ( p < 0.001). In contrast, the high-stable trajectory exhibited the lowest life satisfaction ( p < 0.001), the highest depressive ( p < 0.001) and anxiety symptoms ( p < 0.001), and the poorest physical health ( p < 0.001). Table 3. Developmental outcomes at T4 by perceived stress trajectory groups Variables Perceived stress trajectory classes High-stable vs. Middle-decreasing-rebounding Middle-decreasing-rebounding vs. Low-decreasing-rebounding High-stable vs. Low-decreasing-rebounding M dif SE p M dif SE p M dif SE p DS_T4 -0.81 0.07 < 0.001 0.34 0.08 < 0.001 -1.15 0.06 < 0.001 AS_T4 -0.87 0.07 < 0.001 0.15 0.07 < 0.05 -1.02 0.05 < 0.001 LS_T4 0.86 0.06 < 0.001 -0.47 0.11 < 0.001 1.33 0.05 < 0.001 PH_T4 -0.54 0.08 < 0.001 0.282 0.13 0.09 -0.76 0.11 < 0.001 Open in a new tab DS Depressive symptom, AS Anxiety symptoms, LS Life Satisfaction, PH Physical health Psychological resilience and social support as antecedents of stress trajectories The automatic 3-step method (i.e., R3STEP command) was utilized to examine how psychological resilience and social support were associated with distinct trajectories of perceived stress among FGCS and CGCS college students. As shown in Table 4 , both psychological resilience and social support were significantly associated with the classification of the latent trajectory patterns. Model 1 suggested that after controlling for demographic characteristics (i.e., gender, age, participant group, urban location, and SSS), individuals with higher levels of psychological resilience or social support were more likely to be classified into the low-decreasing-rebounding trajectory than into either the high-stable trajectory or the middle-decreasing-rebounding trajectory. Model 2 indicated that psychological resilience and social support did not interact significantly to differentiate trajectory class membership. Table 4. Multinomial logistic regression odds ratios for baseline predictors Ref. High-stable Ref. Moderate-decreasing-rebounding Moderate-decreasing-rebounding Low-decreasing-rebounding Low-decreasing-rebounding Coeff (SE) OR Coeff (SE) OR Coeff (SE) OR Model 1 Gender(ref.girls) 0.28(0.19) 1.32 0.31(0.16)† 1.37 0.59(0.13) *** 1.81 Age -0.05(0.08) 0.95 0.05(0.08) 1.02 -0.03(0.05) 0.97 SSS 0.25(0.09) ** 1.28 0.47(0.08) *** 0.63 -0.22(0.06) *** 0.80 BeiJing(ref.Jiang Su) 0.87(0.37) * 2.38 -1.52(0.33) *** 0.22 -0.65(0.17) *** 0.52 Model2 Gender(ref.girls) -0.62(0.13) *** 0.54 -0.38(0.08)† 0.69 0.24(0.21) 1.27 FG(ref.CG) 0.23(0.15) 1.26 0.01(0.07) 1.01 -0.22(0.21) 0.8 Age 0.04(0.05) 1.05 0.04(0.08) 0.96 0.00(0.00) 0.92 SSS 0.09(0.17) 1.1 0.17(0.1)† 1.18 0.07(0.1) 1.07 BeiJing(ref.Jiang Su) 0.68(0.18) *** 1.98 1.32(0.32) *** 3.73 0.63(0.34)† 1.88 PR 0.51(0.08) *** 1.66 0.92(0.15) *** 2.51 0.41(0.15) ** 1.51 SS 0.59(0.09) *** 1.8 1.29(0.16) *** 3.63 0.7(0.16) *** 2.02 Model3 Gender(ref.girls) -0.62(0.13) *** 0.54 -0.38(0.2)† 0.69 0.24(0.21) 1.27 FG (ref.CG) 0.23(0.12) 1.26 0.001(0.02) 1.01 -0.22(0.21) 0.8 Age 0.05(0.07) 1.05 -0.04(0.08) 0.96 -0.08(0.08) 0.92 SSS 0.1(0.07) 1.11 0.1(0.07)† 1.19 0.07(0.11) 1.08 BeiJing(ref.Jiang Su) 0.69(0.18) *** 1.99 1.31 (0.33) *** 3.69 0.62(0.37)† 1.85 PR 0.5(0.1) *** 1.64 1.01(0.56)† 2.73 0.07(0.11) 0.61 SS 0.58(0.09) *** 1.78 1.37(0.51) ** 3.93 0.51(0.61) 0.56 PR * SS 0.05(0.09) 1.05 -0.07(0.57) 0.93 -0.12(0.6) 0.95 Model4 Gender (ref.girls) -0.6(0.14) *** 0.55 -0.4 (0.2) * 0.67 0.2(0.21) 1.22 FG (ref.CG) 0.19(0.51) 1.21 0.29(0.39) 1.33 0.1(0.42) 1.1 Age 0.04(0.05) 1.04 -0.030(0.09) 0.97 -0.07(0.09) 0.93 SSS 0.1(0.07) 1.11 0.16(0.1) 1.17 0.05(0.1) 1.05 BeiJing(ref.Jiang Su) 0.66 (0.18) *** 1.94 1.42(0.33) *** 4.13 0.76(0.36) * 2.13 PR 0.48 (0.14) *** 1.61 1.24(0.29) *** 3.47 0.77(0.31) * 2.15 SS 0.5 (0.14) *** 1.64 1.53(0.36) *** 4.61 1.03(0.37) ** 2.8 PR * SS 0.1(0.17) 1.1 -0.76(0.34) * 0.47 -0.85(0.38) * 0.43 FG * PR(ref. CG * PR) 0.18(0.18) 1.2 -0.42(0.42) 0.66 -0.6(0.43) 0.55 FG * SS(ref. CG * SS) -0.01(0.12) 0.99 -0.36(0.27) 0.7 -0.35(0.26) 0.71 FG * PR * SS(ref. CG * PR * SS) -0.01(0.16) 1.0 0.91(0.31) ** 2.48 0.91(0.32) ** 2.49 Open in a new tab OR Odds ratio, PR Psychological resilience, SS Social Support, SSS Subjective socioeconomic status. † p < 0.10; * p < 0.05; ** p < 0.01; *** p < 0.001 Model 4 indicated a significant three-way interaction among psychological resilience, social support, and student group (FGCS and CGCS) when comparing the high-stable or middle-decreasing-rebounding trajectory with the low-decreasing-rebounding trajectory. Following previous research [ 50 ], two separate models estimating the two-way interaction between the two protective factors were conducted for FGCS and CGCS. As presented in Table S5, when comparing the high-stable trajectory with the low-decreasing-rebounding trajectory, a significant two-way interaction between resilience and social support was found only among FGCS, but not among CGCS. Simple slope analyses of the significant two-way interaction revealed that, among FGCS with higher levels of psychological resilience (i.e., 1 SD above the mean), greater social support significantly predicted membership in the low-decreasing-rebounding trajectory relative to the high-stable trajectory (OR = 1.47, p = 0.032), and marginally predicted membership relative to the middle-decreasing-rebounding trajectory. However, this predictive relationship was not observed among FGCS with lower levels of psychological resilience (i.e., 1 SD below the mean; OR = 1.40, p = 0.053). The results of the simple slope analyses are presented in Fig. S1. This pattern of conditional synergy, where social support significantly amplifies the benefits of resilience only among higher-resilience FGCS, provides a nuanced empirical illustration of resource gain theory, which posits that resources are most potent when they converge under conditions of structural disadvantage. This finding sets the stage for a deeper discussion of the differential protective mechanisms at play for FGCS and CGCS. Sensitivity analysis Since the participants were recruited from two different cities, the city was dummy-coded and included as a covariate in the three-step approach to control for potential regional differences. The results showed that, after adjusting for city effects, the main and interaction effects of psychological resilience and social support on predicting perceived stress trajectory membership remained significant ( p < 0.05). Additionally, the associations between patterns of perceived stress trajectories and their predictors and outcomes were re-examined in SPSS Version 27.0 using multinomial regression analysis and a series of ANOVAs. The key findings retained their statistical significance, supporting the robustness and reliability of the current conclusions. Discussion The transition into college represents a dual challenge for students, as they navigate developmental changes from late adolescence into emerging adulthood while simultaneously adapting to increased academic demands. Prior research has established that perceived stress during this critical period carries lasting implications for students’ mental and physical well-being [ 7 ]. Yet key questions persist regarding the temporal patterns of stress, their differential outcomes, and the factors that shape distinct developmental pathways. Using a longitudinal design and a person-centered approach, the current study identified three perceived stress trajectories: high-stable, middle-decreasing-rebounding, and low-decreasing-rebounding. Additionally, psychological resilience and social support, as measured at the start of the initial two years of college, contributed to variations in trajectory. Students in the low-decreasing-rebounding trajectory exhibited better well-being and physical health than those in the other trajectories. These results advance understanding of stress dynamics across the first two years of college and underscore the value of person-centered approaches in capturing meaningful heterogeneity. The study also clarifies how resilience and social support interact to influence stress pathways, particularly in distinguishing outcomes between FGCS and CGCS. Implications for tailored support and future research are highlighted. Longitudinal trajectories of perceived stress Substantial and heterogeneous changes in students’ perceived stress levels were observed over the two years following enrollment; that is, they reported consistently high stress levels that remained stable over time. This pattern aligns with the dynamic perspective of transition cycle theory, where adaptation unfolds through distinct phases over time [ 51 ]. The change in stress followed a nonlinear rather than a linear trajectory [ 52 , 53 ], and significant individual differences were found in both initial stress levels and their rates of change, supporting the use of longitudinal designs to capture heterogeneous adaptation pathways [ 54 ]. Three distinct perceived stress trajectories were identified during the first two years following college entry: high-stable, middle-decreasing-rebounding, and low-decreasing-rebounding (supporting Hypothesis 1a). The majority of students (68.31%) followed a high-stable trajectory, exhibiting chronic stress characteristics with significantly elevated stress levels at enrollment (PSS scores > 16 at all waves) and minimal decline over the tracking period. This prevalence substantially exceeds levels observed in pre-pandemic cohorts [ 7 ], challenging the assumption that stress naturally attenuates with adaptation. It suggests that the contemporary college transition may constitute a chronic stress context [ 55 ], in which cumulative pressures sustain high stress. This finding refines the developmental assets framework [ 14 ] by identifying a stable-high pathway not merely as a transient mismatch outcome, but as a long-term dominant adaptation within high-pressure ecological contexts. The observed U-shaped trajectories provide empirical substantiation for the cognitive appraisal theory of stress, framing it as a dynamic, nonlinear process of person-environment transaction [ 1 ]. Specifically, the modest rebound in the low-declining-rebounding trajectory suggests a dynamic recalibration process, in which stress fluctuations signify active engagement with novel developmental challenges after initial adaptation, rather than maladjustment. This pattern points to a capacity for adaptive stress utilization [ 56 ], wherein transient increases are integrated into growth. Methodologically, the longitudinal design shifts the focus from static endpoint comparisons to a dynamic assessment of the adaptation process [ 57 ]. This lens reveals a critical process inequality: although FGCS and CGCS reached comparable endpoint levels, FGCS traversed a pathway marked by a pronounced mid-transition stress spike, reflecting a journey of greater hardship. Thus, the study makes a dual theoretical contribution: (1) at the micro level, it delineates the temporal dynamics of adaptive stress regulation, demonstrating how nonlinear trajectories can represent recalibration; and (2) at the macro level, it maps divergent group pathways, offering empirical support for cultural mismatch theory. Critically, the finding that group differences emerged only at T2/T3—while outcomes converged again at T4—provides a powerful empirical demonstration of process inequality being masked by outcome equality. This underscores the necessity of dynamic, longitudinal designs over static comparisons to uncover such hidden disparities. The mid-transition peak in stress for FGCS precisely captures the period when dissonance between institutional norms and student cultural schemas most strongly impacts adjustment [ 58 , 59 ]. Practically, this pinpoints the mid-transition (T2/T3) as the critical window for intervention. The findings translate into an actionable three-tiered support framework for FGCS: (1) proactive screening via brief surveys for early risk identification; (2) integrated interventions (e.g., Resilience Mentorship Circles, University Navigation Workshops) that concurrently build internal and external resources; and (3) a centralized FGCS Success Hub to streamline access to support services. This provides a clear pathway from evidence to campus policy. Distal outcomes associated with stress trajectories The current study also examines the long-term outcomes associated with college students’ mental and physical health. Specifically, students in the low-decreasing-rebounding trajectory group exhibited the most favorable outcomes across both mental and physical domains. Conversely, students in the high-stable trajectory group exhibited the most severe depressive and anxiety symptoms, the lowest life satisfaction, and the poorest physical health (supporting Hypothesis 1b). These findings are consistent with prior research and theories underscoring the detrimental impact of stress on holistic well-being [ 7 , 54 , 60 ]. The current findings also support and extend contemporary research on stress and health [ 61 ] by revealing the multifaceted effects of distinct stress trajectories on the subsequent development of college students during the transition period—a time that presents both developmental opportunities and risks. Notably, the levels of depressive and anxiety symptoms reported by students in the high-stable trajectory group reached the clinical cutoff for the Chinese population, suggesting that these students may require clinical assessment. As reflected in the high-stable pattern, high-density stressful experiences during such a critical period may further deteriorate students’ mental and physical health [ 62 , 63 ]. Consequently, these findings re-emphasize the critical importance of monitoring perceived stress and identifying at-risk groups during the early stages of college life. The roles of psychological resilience and social support: differences between FGCS and CGCS Our findings support the developmental assets framework by showing that social support and psychological resilience each independently predict membership in distinct perceived stress trajectories. Specifically, higher levels of either factor increase the likelihood of following the low-decreasing-rebounding trajectory rather than the chronic high-stable or middle-decreasing-rebounding trajectories (supporting Hypothesis 2). These effects, however, are not merely additive. A significant interaction emerged exclusively among FGCS: for those with higher baseline resilience, greater social support substantially increased the probability of belonging to the adaptive trajectory (supporting Hypothesis 3). This aligns with resource gain theory [ 14 ], which posits that resources are most potent when they combine under structural disadvantage. These results address a gap in prior research that examined external or internal assets in isolation among FGCS and CGCS [ 25 , 29 ]. In the Chinese context, FGCS face compounded stressors during the college transition, such as Gaokao competition, high family expectations, and potential cultural mismatch [ 27 , 64 ], which contribute to their elevated initial stress levels [ 65 ]. Consequently, they often require the simultaneous mobilization of internal psychological resources and external support to disrupt chronic high stress. Particularly within this subgroup, social support appears to function not merely as a supplementary resource but as a critical enabling scaffold—translating psychological resilience into concrete adaptive coping strategies. This suggests that for students facing structural disadvantages, reliable external support may be a necessary condition for activating their inherent strengths. When the external environment (e.g., family, peers, institutions) provides adequate support, psychological resilience can be more effectively activated and translated into adaptive coping strategies, facilitating a positive shift in stress trajectories. This observation aligns with Western research highlighting the critical role of resource synergy for disadvantaged student groups [ 54 ]. In summary, this study suggests that for FGCS, interventions focusing solely on a single resource (e.g., resilience training or support provision) may be insufficient. Effective strategies should promote the synergistic development of psychological capital and the empowerment of supportive ecologies. Such integrated approaches can more effectively intervene in the chronicity of high-stress pathways. For CGCS, psychological resilience and social support exerted significant, yet independent and additive, influences on their perceived stress trajectories. This pattern aligns with their relatively resource-replete ecological contexts in China, characterized by greater access to established support networks and more embedded positive value systems [ 66 ]. Within such contexts, a single robust resource (whether high resilience or strong support) can be sufficiently effective in mitigating stress, without necessitating a strong synergistic interaction between them. In contrast, the protective mechanism for FGCS was found to be conditional and synergistic, as detailed in the preceding paragraph. In summary, accounting for the heterogeneity between these student groups allows for a more precise identification of how developmental assets differentially protect against maladaptive stress trajectories. This differentiation provides a foundation for designing targeted interventions that address the distinct resource-related needs of each subgroup. Contributions, limitations, and future directions This study contributes to the literature in several integrated domains. Firstly, it advances a dynamic, person-centered perspective by delineating distinct perceived stress trajectories during the college transition, particularly within the post-pandemic context, which appears to foster a prevalent chronic high-stable stress pattern [ 67 ]. This finding challenges adaptation assumptions and refines transition theories. Furthermore, by identifying conditional synergy between resilience and social support (specific to higher-resilience FGCS) and a process inequality in which FGCS endure a mid-transition stress spike despite similar endpoints, the study highlights subgroup specificity and boundary conditions within person-context transaction and resource-gain models. Methodologically, the research demonstrates the indispensable value of a multi-wave longitudinal design not only for capturing heterogeneity but, more critically, for unveiling process inequalities that are completely obscured in cross-sectional or endpoint analyses. Practically, the results translate into an actionable, stratified, and sequence-sensitive intervention framework. It moves beyond risk profiling to specify concrete strategies: universal early screening, integrated skill-building for moderate-risk students, stabilization-first intensive support for high-risk students, and the precise timing of interventions around the critical mid-transition window, especially for FGCS. Several limitations suggest avenues for future work. First, the two-year timeframe, while covering a critical transition, does not encompass the full undergraduate journey. Longer follow-ups are needed. Second, measuring resilience and social support only at baseline limits insights into their dynamic co-evolution with stress. Future studies require repeated assessments. Third, despite using validated scales, reliance on self-report remains a limitation, particularly given our findings of clinical-level distress in the high-stable group. Future research should integrate objective indicators (e.g., academic records, clinical assessments, physiological markers like cortisol) to strengthen validity. Fourth, the regional specificity of our sample (Beijing and Jiangsu) limits generalizability across China’s diverse socioeconomic spectrum. Nationally representative samples are needed. Finally, cross-cultural replications are essential to test the generalizability of the observed patterns and mechanisms across different educational and cultural contexts. Conclusion Moving beyond variable-centered approaches, this longitudinal study charts a person-centered map of stress adaptation during the college transition. It reveals a prevalent chronic high-stress pathway among contemporary students, with profound implications for their well-being. Crucially, the analysis demonstrates that the protective power of psychological resilience and social support is not universal but contingent—manifesting as a potent synergy specifically for first-generation students entering college with higher internal resources. This pattern of conditional protection underscores the fundamentally divergent pathways characterizing the transition for different student groups. Collectively, by mapping heterogeneous stress trajectories and their conditional protective pathways, this work deepens the theoretical understanding of dynamic person-environment interactions. The findings carry clear implications for promoting equity in higher education: integrated interventions that simultaneously foster psychological resilience and strengthen social support systems—particularly when tailored for first-generation and other vulnerable student populations—constitute a promising strategy for disrupting maladaptive stress cycles and facilitating a successful college transition. Supplementary Information Supplementary Material 1. (116.4KB, docx) Acknowledgements We extend our sincere gratitude to the university students who participated in this study and to the research assistants for their invaluable support in data collection. Authors’ contributions J.H. participated in conceptualizing the study, conducting statistical analyses, and drafting the manuscript; M.C. helped to interpret the data and edited the manuscript; X.W. participated in data collection, and helped to conduct statistical analyses; M.X. helped to revise and edit the manuscript; Y.B. helped with the visualization of the results and helped to revise and edit the manuscript; Q.X. participated in the design and coordination of the study, provided critical reviews of the manuscript, and contributed to funding acquisition. D.L. participated in the design and coordination of the study, provided critical reviews of the manuscript, and contributed to funding acquisition. All authors read and approved the final manuscript. Funding This work was partly supported by the General Project of Philosophy and Social Science Research in Jiangsu Universities (No. 2020SJA1182) and the Jiangsu Education Science Planning Project (No. C/2023/01/35). Data availability The data supporting this research are not publicly accessible but can be provided by the corresponding author upon justified request. Declarations Ethics approval and consent to participate This study was approved by the Institutional Review Board of Beijing Normal University (Approval No.: BNU202310200155). All methods were performed in accordance with the relevant guidelines and regulations, including the Declaration of Helsinki. Informed consent was obtained from all participants prior to their involvement in the study. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References 1. Lazarus RS, Folkman S. Stress, appraisal, and coping. Springer Pub. Co.; 1984. 2. Conley CS, Kirsch AC, Dickson DA, Bryant FB. Negotiating the transition to college: Developmental trajectories and gender differences in psychological functioning, cognitive-affective strategies, and social well-being. Emerging Adulthood. (Thousand Oaks CA). 2014;2(3):195–210. 10.1177/2167696814521808. [ Google Scholar ] 3. Arnett JJ. College students as emerging adults: The developmental implications of the college context. Emerg Adulthood. 2016;4(3):219–22. 10.1177/2167696815587422. [ Google Scholar ] 4. Denovan, A., Macaskill. Stress and subjective well-being among first year UK undergraduate students. J Happiness Stud. 18;505–525. 10.1007/s10902-016-9736-y. 5. Liu Y, Zhang N, Bao G, Huang Y, Ji B, Wu Y, Liu C, Li G. Predictors of depressive symptoms in college students: A systematic review and meta-analysis of cohort studies. J Affect Disord. 2019;244:196–208. 10.1016/j.jad.2018.10.084. [ DOI ] [ PubMed ] [ Google Scholar ] 6. Bettis AH, Coiro MJ, England J, Murphy LK, Zelkowitz RL, Dejardins L, Eskridge R, Adery LH, Yarboi J, Pardo D, Compas BE. Comparison of two approaches to prevention of mental health problems in college students: Enhancing coping and executive function skills. J Am Coll Health: JACH. 2017;65(5):313–22. 10.1080/07448481.2017.1312411. [ DOI ] [ PubMed ] [ Google Scholar ] 7. Zhao S, Zhang Y, Yu C, Zhang H, Xie M, Chen P, Lin D. Trajectories of perceived stress among students in transition to college: Mindset antecedents and adjustment outcomes. J Youth Adolesc. 2023;52(8):1873–1886. 10.1007/s10964-023-01788-5. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 8. Conley CS, Shapiro JB, Huguenel BM, Kirsch AC. Navigating the college years: Developmental trajectories and gender differences in psychological functioning, cognitive-affective strategies, and social well-being. Emerging Adulthood. (Thousand Oaks CA). 2020;8(2):103–17. 10.1177/2167696818791603. [ Google Scholar ] 9. Epel ES, Crosswell AD, Mayer SE, Prather AA, Slavich GM, Puterman E, Mendes WB. More than a feeling: A unified view of stress measurement for population science. Front Neuroendocr. 2018;49:146–69. 10.1016/j.yfrne.2018.03.001. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 10. Eccles JS, Midgley C, Wigfield A, Buchanan CM, Reuman D, Flanagan C, Iver M, D. Development during adolescence: The impact of stage-environment fit on young adolescents’ experiences in schools and in families. Am Psychol. 1993;48(2):90–101. 10.1037/0003-066X.48.2.90. [ DOI ] [ PubMed ] [ Google Scholar ] 11. Browman AS, Destin M, Carswell KL, Svoboda RC. Perceptions of socioeconomic mobility influence academic persistence among low socioeconomic status students. J Exp Soc Psychol. 2017;72:45–52. 10.1016/j.jesp.2017.03.006. [ Google Scholar ] 12. Zhang L, Wang L, Liu Y, Zhang J, Zhang X, Zhao J. Resilience predicts the trajectories of college students’ daily emotions during COVID-19: A latent growth mixture model. Front Psychol. 2021;12:648368. 10.3389/fpsyg.2021.648368. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 13. Masten AS. Ordinary magic: Resilience in development. Guilford Press; 2014. 14. Benson PL. Developmental assets: An overview of theory, research, and practice. In R. K. 2007. 15. Barrera M. Social support research in community psychology. In: Rappaport J, Seidman E, editors. Handbook of Community Psychology. Boston, MA: Springer; 2000. 10.1007/978-1-4615-4193-6_10. [ Google Scholar ] 16. Kim HS, Sherman DK, Taylor SE. Culture and social support. Am Psychol. 2008;63(6):518–26. 10.1037/0003-066X. [ DOI ] [ PubMed ] [ Google Scholar ] 17. Connor KM, Davidson JRT. Development of a new resilience scale: The connor-davidson resilience scale (CD-RISC). Depress Anxiety. 2003;18(2):76–82. 10.1002/da.10113. [ DOI ] [ PubMed ] [ Google Scholar ] 18. Southwick SM, Bonanno GA, Masten AS, Panter-Brick C, Yehuda R. Resilience: The science of mastering life’s greatest challenges. 2nd ed. Cambridge University Press; 2014. 19. Gerard JM, Booth MZ. Family and school influences on adolescents’ adjustment: The moderating role of youth hopefulness and aspirations for the future. J Adolescence (London England). 2015;44:1–16. 10.1016/j.adolescence.2015.06.003. [ DOI ] [ PubMed ] [ Google Scholar ] 20. Hanimoğlu E. A mediation and moderation model for life satisfaction: The role of social support, psychological resilience, and gender. Acta Psychol. 2025;259:105360. 10.1016/j.actpsy.2025.105360. [ DOI ] [ PubMed ] [ Google Scholar ] 21. Niitsu K, Houfek JF, Rice MJ, Stoltenberg SF, Kupzyk K, Barron C. Analyzing the relationships among psychosocial variables associated with resilience. Arch Psychiatr Nurs. 2024;53:33–41. 10.1016/j.apnu.2024.09.006. [ DOI ] [ PubMed ] [ Google Scholar ] 22. Redford J, Battle D, Bielick S. Homeschooling in the United States: 2012. Washington, DC: National Center for Education Statistics, Institute of Education Sciences, U.S. Department of Education. 2017. https://nces.ed.gov/pubsearch/pubsinfo.asp?pubid=2016096rev . Retrieved August 1, 2017, from. (NCES 2016-096.REV). 23. Stephens NM, Fryberg SA, Markus HR, Johnson CS, Covarrubias R. Unseen disadvantage: How American universities’ focus on independence undermines the academic performance of first-generation college students. J Personal Soc Psychol. 2012;102(6):1178–97. 10.1037/a0027143. [ DOI ] [ PubMed ] [ Google Scholar ] 24. Jack AA. (Harm in asking: Class, acquired cultural capital, and academic engagement at an elite university. Sociol Educ. 2016;89(1):1–19. 10.1177/0038040715614913. [ Google Scholar ] 25. Soria KM, Stebleton MJ. First-generation students’ academic engagement and retention. Teach High Educ. 2012;17(6):673–85. 10.1080/13562517.2012.666735. [ Google Scholar ] 26. Wang ZX. The breakthrough of students from humble backgrounds: A review of first-generation college student research at home and abroad. China Youth Study. 2020;0194–104. 10.19633/j.cnki.11-2579/d.2020.0014. 27. Tian J, Yu X. From a deficit perspective to a strengths perspective: A review of research on first-generation college students. Chongqing High Educ Res. 2021;9(5):106–18. 10.15998/j.cnki.issn1673-8012.2021.05.010. [ Google Scholar ] 28. Chen T, Zhang Y, Xue HP. Toward reshaping: The dual effects of intergenerational transmission in family education of first-generation college students. J East China Normal Univ. 2023;41(4):111–26. 10.16382/j.cnki.1000-5560.2023.04.009. [ Google Scholar ] 29. Stephens NM, Hamedani MG, Destin M. Closing the social-class achievement gap: A difference-education intervention improves first-generation students’ academic performance and all students’ college transition. Psychol Sci. 2014;25(4):943–53. 10.1177/0956797613518349. [ DOI ] [ PubMed ] [ Google Scholar ] 30. Liu X, Zhang Y, Gao W, Cao X. Developmental trajectories of depression, anxiety, and stress among college students: a piecewise growth mixture model analysis. Humanit Social Sci Commun. 2023;10(1). 10.1057/s41599-023-02252-2. Article 736. 31. Cohen S, Kamarck T, Mermelstein R. A global measure of perceived stress. J Health Soc Behav. 1983;24(4):385–96. 10.2307/2136404. [ PubMed ] [ Google Scholar ] 32. Connor KM, Davidson JR. Development of a new resilience scale: The connor-davidson resilience scale (CD-RISC). Depress Anxiety. 2003;18(2):76–82. 10.1002/da.10113. [ DOI ] [ PubMed ] [ Google Scholar ] 33. Zimet GD, Dahlem NW, Zimet SG, Farley GK. The multidimensional scale of perceived social support. J Pers Assess. 1988;52(1):30–41. 10.1207/s15327752jpa5201_2. [ Google Scholar ] 34. Kroenke K, Spitzer RL, Williams JB. The PHQ-9: Validity of a brief depression severity measure. J Gen Intern Med. 2001;16(9):606–13. 10.1046/j.1525-1497.2001.016009606.x. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 35. Ye Z, Yang X, Zeng C, Wang Y, Shen Z, Li X, Lin D. Resilience, social support, and coping as mediators between COVID-19-related stressful experiences and acute stress disorder among college students in China. Appl Psychology: Health Well-Being. 2020;12(4):1074–94. 10.1111/aphw.12211. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 36. Spitzer RL, Kroenke K, Williams JBW, Löwe B. A brief measure for assessing generalized anxiety disorder: The GAD-7. Arch Intern Med. 2006;166(10):1092–7. 10.1001/archinte.166.10.1092. [ DOI ] [ PubMed ] [ Google Scholar ] 37. Ip H, Suen YN, Hui CLM, Wong SMY, Chan SKW, Lee EHM, Wong MTH, Chen EYH. Assessing anxiety among adolescents in Hong Kong: psychometric properties and validity of the Generalised Anxiety Disorder-7 (GAD-7) in an epidemiological community sample. BMC Psych. 2022;22:703. 10.1186/s12888-022-04329-9. [ DOI ] [ PMC free article ] [ PubMed ] 38. Diener E, Emmons RA, Larsen RJ, Griffin S. The satisfaction with life scale. J Pers Assess. 1985;49(1):71–5. 10.1207/s15327752jpa4901_13. [ DOI ] [ PubMed ] [ Google Scholar ] 39. Kellner R. A symptom questionnaire. J Clin Psychiatry. 1987;48(7):268–74. [ PubMed ] [ Google Scholar ] 40. Muthén LK, Muthén BO. Mplus user’s guide (8th ed.). Muthén & Muthén. (1998–2017). https://www.statmodel.com/ . 41. Jung T, Wickrama KAS. An introduction to latent class growth analysis and growth mixture modeling. Soc Pers Psychol Compass. 2008;2(1):302–17. 10.1111/j.1751-9004.2007.00054.x. [ Google Scholar ] 42. Little RJ, Rubin DB. Statistical analysis with missing data. 2nd ed. Wiley; 2002. 43. Schafer JL, Graham JW. Missing data: Our view of the state of the art. Psychol Methods. 2002;7(2):147–77. 10.1037/1082-989X.7.2.147. [ PubMed ] [ Google Scholar ] 44. Nylund KL, Asparouhov T, Muthén BO. Deciding on the number of classes in latent class analysis and growth mixture modeling: A Monte Carlo simulation study. Struct Equation Modeling: Multidisciplinary J. 2007;14(4):535–69. 10.1080/10705510701575396. [ Google Scholar ] 45. Lo Y, Mendell NR, Rubin DB. Testing the number of components in a normal mixture. Biometrika. 2001;88(3):767–78. 10.1093/biomet/88.3.767. [ Google Scholar ] 46. Asparouhov T, Muthén B. Auxiliary variables in mixture modeling: Three-step approaches using Mplus. Struct Equation Modeling: Multidisciplinary J. 2014;21(3):329–41. 10.1080/10705511.2014.915181. [ Google Scholar ] 47. Nylund-Gibson K, Grimm RP, Masyn KE. Prediction from latent classes: A demonstration of different approaches to include distal outcomes in mixture models. Struct Equation Modeling: A Multidisciplinary J. 2019;26(6):967–985. 10.1080/10705511.2019.1590146. 48. Vermunt JK. Latent class modeling with covariates: Two improved three-step approaches. Political Anal. 2017;18(4):450–69. 10.1093/pan/mpq025. [ Google Scholar ] 49. Podsakoff PM, MacKenzie SB, Lee J-Y, Podsakoff NP. Common method biases in behavioral research: A critical review of the literature and recommended remedies. J Appl Psychol. 2003;88(5):879–903. 10.1037/0021-9010.88.5.879. [ DOI ] [ PubMed ] [ Google Scholar ] 50. Hayes AF, Matthes J. Computational procedures for probing interactions in OLS and logistic regression: SPSS and SAS implementations. Behav Res Methods. 2009;41(3):924–36. 10.3758/BRM.41.3.924. [ DOI ] [ PubMed ] [ Google Scholar ] 51. De Clercq M, Roland N, Brunelle M, Galand B, Frenay M. The delicate balance to adjustment: A qualitativeapproach of student's transition to the first year at university. Psychologica Belgica. 2018;58(1):67–90. 10.5334/pb.409. [ DOI ] [ PMC free article ] [ PubMed ] 52. Garett R, Liu S, Young SD. A longitudinal analysis of stress among incoming college freshmen. J Am College Health. 2017;65(5):331–338. 10.1080/07448481.2017.1312413. [ DOI ] [ PMC free article ] [ PubMed ] 53. Hatun O, Kurtça TT. Perceived stress, hope, and life satisfaction among college students: A two-wave, half-longitudinal study from turkey. Int J Ment Health Addiction. 2026;24:240–261. 10.1007/s11469-024-01411-z. 54. Chen S, Liao J, Ran F, et al. Longitudinal Associations between Future Time Perspective, Sleep Problems, and Depressive Symptoms among Chinese College Students: Between‐ and within‐Person Effects. J Youth Adolesc. 2025;54:480–492. 10.1007/s10964-024-02078-4. [ DOI ] [ PubMed ] 55. Karyotaki E, Cuijpers P, Albor Y, Alonso J, Auerbach RP, Bantjes J, Bruffaerts R, Ebert DD, Hasking P, Kiekens G, Lee S, McLafferty M, Mak A, Mortier P, Sampson NA, Stein DJ, Vilagut G, Kessler RC. Sources of stress and their associations with mental disorders among college students: Results. Front Psychol. 2020;11:1759. 10.3389/fpsyg.2020.01759. of the World Health Organization World Mental Health Surveys International College Student Initiative. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 56. Crum AJ, Salovey P, Achor S. Rethinking stress: The role of mindsets in determining the stress response. J Personal Soc Psychol. 2013;104(4):716–33. 10.1037/a0031201. [ DOI ] [ PubMed ] [ Google Scholar ] 57. Diemer MA, Mistry RS, Wadsworth ME, López I, Reimers F. Best practices in conceptualizing and measuring social class in psychological research. Analyses Social Issues Public Policy. 2013;13(1):77–113. 10.1111/asap.12001. [ Google Scholar ] 58. Osafo AB, Ohene-Darko S, Agormedah EK, Omane-Adjekum C. Navigating the transition to higher education: understanding the experience among first-year university students. J Tertiary Educ Learning. 2025;3(1):51–64. 59. Coleman JS. Foundations of social theory. Harvard University Press; 2023. 60. Selye H. The stress of life (Rev. ed.). McGraw-Hill; 1978. 61. Chiang JJ, Lam PH, Chen E, Miller GE. Psychological stress during childhood and adolescence and its association with inflammation across the lifespan: A critical review and meta-analysis. Psychol Bulletin. 2022;148(1-2):27–66. 10.1037/bul0000351. [ DOI ] [ PMC free article ] [ PubMed ] 62. Beníítez-Agudelo JC, Restrepo D, Navarro-Jimenez E, Clemente-Suáárez VJ. Longitudinal effects of stress in an academic context on psychological well-being, physiological markers, health behaviors, and academic performance in university students. BMC Psychol. 2025;13:753. 10.1186/s40359-025-03041-z. [ DOI ] [ PMC free article ] [ PubMed ] 63. Al Hadid LA, Al-Rajabi O, Al Barmawi M, Alhadidi M, Jaradat AM. Measuring the prevalence of psychologicalsymptoms and the predictors associated with seeking counseling and psychological help amonguniversity students. J Am College Health. 2023;71(2):389–395. 10.1080/07448481.2021.1891918. [ DOI ] [ PubMed ] 64. Fang SX, Lu Y. The different college experiences of first-generation students in China and the United States. Fudan Educ Forum. 2022;20(1):31–40. 10.13397/j.cnki.fef.2022.01.004. [ Google Scholar ] 65. Li YS, Zhang K, Bi YL, Zhang BS. Psychological challenge and its explanation of first-generation college students: A perspective from cultural mismatch theory. Adv Psychol Sci. 2022;30(10):2338–55. 10.3724/SP.J.1042.2022.02338. [ Google Scholar ] 66. Liao Y, Chen J, Liu Y, et al. The longitudinal relationship between social support and university adjustment: The roles of basic psychological needs satisfaction and intentional self-regulation. Psychol Sci. 2025;48(1):86–96. 10.16719/j.cnki.1671-6981.20250109. [ Google Scholar ] 67. Wu J. Psychological issues and mechanism construction among college students in the post-pandemic era (Review). China J Health Psychol. 2023;31(8):1229–34. 10.13342/j.cnki.cjhp.2023.08.022. [ Google Scholar ] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Supplementary Materials Supplementary Material 1. (116.4KB, docx) Data Availability Statement The data supporting this research are not publicly accessible but can be provided by the corresponding author upon justified request. Articles from BMC Psychology are provided here courtesy of BMC ACTIONS View on publisher site PDF (1.4 MB) Cite Collections Permalink PERMALINK Copy RESOURCES Similar articles Cited by other articles Links to NCBI Databases Cite Copy Download .nbib .nbib Format: AMA APA MLA NLM Add to Collections Create a new collection Add to an existing collection Name your collection * Choose a collection Unable to load your collection due to an error Please try again Add Cancel Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top