Stressful life events and non-suicidal self-injury among college students: chain mediating effect analysis of resilience and emotion dysregulation Authors: Juan Ni, Xiaoli Liao, Jia Liu, Qinghua Jiang Published: 2025 ABSTRACT BackgroundNon-suicidal self-injury (NSSI) behavior represents a worrisome mental health concern across the world, with stressful life events emerging as a critical proximal predictor of NSSI behavior among college students. The high prevalence and severe implications of NSSI behavior necessitate a thorough investigation of the underlying mechanisms linking stressful life events to such maladaptive behavior among this vulnerable population.ObjectiveThis study aims to delineate the psychological pathways through which stressful life events contribute to NSSI behaviors among college students, with a focus on the sequential mediating roles of resilience and emotion dysregulation.MethodsA cross-sectional convenience sample of 1,392 college students from Hunan Province, China, was recruited between September and December 2023. Data were collected via a structured online questionnaire, including the Ottawa Self-Injury Inventory (OSI), the Adolescent Self-Rating Life Events Checklist (ASLEC), the Difficulties in Emotion Regulation Scale (DERS), and the Resilience Scale for Chinese Adolescents (RSCA). Descriptive statistics, Pearson correlation analyses, hierarchical multiple regression analyses, and sequential mediation analysis were performed to evaluate the hypothesized relationships.ResultsThe prevalence of NSSI behavior was 7.54% (105/1392), lower than reported in Western populations but consistent with some regional Chinese estimates. Between-group comparisons revealed that the NSSI group demonstrated elevated total scores and scores across all dimensions on the ASLEC (p < 0.01) and DERS (p < 0.01), while exhibiting lower total scores and scores across all dimensions on the RSCA (p < 0.01). Correlation analysis indicated that NSSI behaviors were positively correlated with stressful life events (r = 0.408, p < 0.01) and emotion dysregulation (r = 0.430, p < 0.01), while negatively correlated with resilience (r = -0.592, p < 0.01). Sequential mediation analysis revealed that resilience and emotion dysregulation jointly mediated the effect of stressful life events on NSSI behaviors, highlighting a nuanced interplay of protective and vulnerability factors.ConclusionThese findings elucidate the intricate chain mediating effects of resilience and emotion dysregulation in the relationship between stressful life events and NSSI behaviors among college students. Given the cross-sectional design, causal inferences are constrained. Nonetheless, this study further provides valuable insights for the development and refinement of effective prevention strategies and intervention programs aimed at reducing NSSI risk. Introduction Self-injury refers to any act of intentional self-destruction or self-harm behavior regardless of the type of motivation or the degree of suicidal intent [1]. Non-suicidal self-injury (NSSI) specifically involves self-inflicted damage to body tissue without any suicidal intent, encompasses behaviors such as cutting, burning, and scratching [2–4]. The growing prevalence and increasing clinical relevance of NSSI behaviors have garnered substantial attention, culminating in its classification as a condition requiring further study in the Fifth Edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) [5]. This formal recognition accentuates the critical necessity for rigorous research endeavors designed to systematically unravel the complex mechanisms underlying NSSI behaviors. NSSI is a multifactorial phenomenon arising from the dynamic interplay of biological, psychological, and social factors [6]. The developmental transition from late adolescence to young adulthood is a critical period marked by significant physiological, cognitive, and psychosocial changes that increase susceptibility to maladaptive coping behavior [7, 8]. Epidemiological data indicate that NSSI behaviors typically emerge in early adolescence, peak in later adolescence, and persist into young adulthood [9, 10].The global prevalence of NSSI varies considerably, with studies reporting rates from 5.5% to 17.2% globally and 6.4% to 47.5% in China [11–13].The prevalence of NSSI has exhibited a concerning upward trend across multiple countries, highlighted by an increase in presentations at general hospitals and primary care settings [14–16]. These empirical findings underscore the urgent imperative to investigate the underlying risk factors, which is essential for developing intervention strategies that effectively address the complexities of NSSI among vulnerable populations. NSSI carries significant clinical and public health implications that extend well beyond the immediate physical harm, serving as an important indicator of psychological distress. NSSI has been consistently identified as a robust predictor of subsequent suicidal ideation, suicide attempts, and even completed suicides [17–19]. Theoretical frameworks and empirical research suggest that individuals engaging in NSSI behaviors are at an elevated risk for subsequent suicidal ideation, attempts, and completion [20–23]. These behaviors may serve as a gateway to more severe forms of self-injury on the suicide spectrum, habituating individuals to physical pain and potentially empowering suicidal actions [24]. Findings from both correlational and longitudinal studies indicate that a history of NSSI increases the risk of suicide attempts and completed suicides by 30- to 100-fold in both community populations and clinical samples [19–27]. Despite the increasing global interest in NSSI behaviors, research has been predominantly concentrated in Europe and North America, with relatively limited studies conducted in China [28, 29]. Existing research in China has primarily focused on coastal cities and developed regions, leaving significant gaps in understanding the broader epidemiology and psychosocial determinants of NSSI across diverse demographic and regional contexts [13]. Emerging adulthood represents a critical developmental phase characterized by heightened interpersonal sensitivity and vulnerability, influenced by both psychosocial and physiological transitions. NSSI behaviors emerge as significant challenges, imposing substantial strains on families, communities, and healthcare systems alike. There is a discernible shift in public discourse surrounding mental health, with an increasing emphasis on adolescent psychological wellbeing taking center stage in public health agendas. Among the various risk factors associated with NSSI behaviors, stressful life events have been consistently identified as a critical proximal trigger. Stressful life events refer to various kinds of negative events which occur in all aspects of daily life [30, 31]. Stressful life events occur in daily scenarios of the school environment, family context, and social situation, which affect, threaten, and damage the physical status and psychological conditions of adolescents. Stressful life events are considered proximal risk factors for NSSI behaviors, as individuals often engage in these behaviors as a coping mechanism to alleviate the physiological stress and emotional turmoil caused by such events [3]. Longitudinal research involving 1,482 adolescents in Switzerland demonstrated that exposure to stressful life events significantly predicts engagement in NSSI behaviors [32]. Similarly, a longitudinal study of 279 migrant children in China indicated that stressful life events prospectively predict NSSI behaviors [33], providing convergent evidence across culturally and demographically distinct populations. Evidence from both cross-sectional and longitudinal studies consistently supports the link between stressful life events and NSSI behaviors, suggesting that the experience of stressful life events could increase the risk of NSSI behaviors [34, 35]. Despite consistent empirical support for the link between stressful life events and NSSI behaviors [34, 35], the precise psychological and behavioral pathways mediating this association remain insufficiently elucidated. Clarifying these pathways is essential for identifying potential mediating and moderating mechanisms that govern the translation of stress exposure into self-injurious behavior. The present study draws upon the Process Model of Emotion Dysregulation [36] and the Protective Factor Model of Resilience as foundational theoretical frameworks [37–39]. Resilience referred to the ability to overcome adversity events and obtain positive adaptations, thereby attenuating the deleterious effects of stressful life events [37]. The Protective Factor Model of Resilience posits that resilience functions as a psychological buffer, enabling individuals to mobilize internal resources and external supports to counteract emotional distress and psychological strain induced by stressors [38]. Higher levels of resilience equip individuals with greater potential to mitigate the impact of stressful experiences and reduce the risk of engaging in NSSI behaviors. Emotion regulation refers to the ability to monitor, evaluate, and modulate emotional responses to adverse experiences, acting as a crucial mechanism linking stress exposure to maladaptive behavioral outcomes. According to the Emotional Cascade Model, engagement in NSSI behaviors serves as a temporary strategy to relieve overwhelming negative emotions, illustrating how deficits in emotion regulation can translate stress into self-injurious behavior [40]. Thus, the present study aims to delineate the sequential psychological mechanisms through which stressful life events influence NSSI behaviors among Chinese college students. The study proposes a chain mediation model in which resilience serves as the initial protective factor that buffers stress effects, subsequently shaping emotion regulation capacities, which in turn determine the likelihood of NSSI engagement. Investigating this sequential pathway holds substantial theoretical and practical significance, offering a nuanced understanding of the dynamic interplay between protective and vulnerability factors in NSSI behaviors and providing evidence to guide the development of targeted interventions that strengthen resilience and enhance adaptive emotion regulation among at-risk populations. Our study addresses the following research questions: (1) How do stressful life events influence NSSI behaviors among Chinese college students?(2) Do resilience and emotion dysregulation sequentially mediate the relationship between stressful life events and NSSI behaviors? Correspondingly, we formulated the following hypotheses: H1: Stressful life events are positively associated with NSSI behaviors among college students. H2: Emotion dysregulation mediates the relationship between stressful life events and NSSI behaviors, such that greater exposure to stressful life events is associated with increased emotion dysregulation, which in turn heightens the risk of NSSI behaviors. H3: Resilience serves as an initial protective factor that mitigates the impact of stressful life events on emotion dysregulation, establishing a sequential mediation pathway: stressful life events → resilience → emotion dysregulation → NSSI behaviors. H4: This sequential mediation pathway accounts for a significant proportion of variance in NSSI behaviors, highlighting the interplay between protective and vulnerability factors in shaping maladaptive behavioral outcomes. Method Study design This investigation employed a cross-sectional design using a convenience sampling strategy to recruit college students from public universities in Hunan Province, China, between September and December 2023. The cross-sectional approach enabled efficient assessment of self-reported NSSI behaviors, exposure to stressful life events, resilience, and emotion dysregulation within a well-defined population of emerging adults. Study participants A total of 1,426 college students were invited to participate. Following the exclusion of 34 invalid responses , the final analytical sample comprised 1,392 participants, resulting in a valid response rate of 97.61%. Eligible participants were aged 16 to 22 years and free from severe psychiatric conditions such as schizophrenia or bipolar disorder. Exclusion criteria included students from foreign countries or regions outside mainland China (including Hong Kong, Macau, and Taiwan) and those on academic leave during the data collection period. Ethical considerations This study was conducted in strict accordance with the principles outlined in the Declaration of Helsinki and received ethical approval from the Ethics Committee of the sampled University (Approval No.: YXLL20230905). Participants received detailed information about the study’s objectives, procedures, and potential implications, allowing them to make an informed decision regarding their participation. Informed consent was obtained from all participants, who were assured that their involvement was entirely voluntary and that they could withdraw from the study at any time without any consequences. Participants were assured of strict confidentiality regarding their responses and were explicitly informed of their right to withdraw from the study at any time without consequences. Recruitment Data collection was conducted via the online survey platform Wenjuanxing, widely validated and utilized in Chinese research contexts. The survey comprised three sequential modules: (1) eligibility screening according to predefined inclusion and exclusion criteria, (2) electronic informed consent with detailed ethical safeguards, and (3) administration of standardized psychometric instruments assessing the primary constructs of interest. Recruitment employed a targeted, network-based approach, whereby research team members collaborated with student counselors across multiple universities via WeChat. Counselors received structured briefings on eligibility criteria, ethical considerations, and standardized procedures for survey dissemination. Survey links were distributed to eligible students only after procedural comprehension was confirmed. Participants completed the survey independently in a quiet, distraction-free environment, with counselors providing procedural guidance without influencing responses. All responses were collected anonymously, and survey completion was interpreted as evidence of informed consent. Measures General information form A general information form was designed to capture key sociodemographic characteristics for the purpose of sample description and between-group comparisons. The items included gender (male, female), ethnicity (Han, other), age, being left-behind children (yes, no), living in a dormitory (yes, no), and being the only-child (yes, no), family economic status (poor, fair, good), family structure (single parent/recombined family, living with grandparents, living with parents, living with other relatives), parents’ educational levels (junior high school or below, senior high school/technical school, college or above), and parenting style (strict, pamper, neglect, often changed, open-minded). An “only child” is defined as a child who has no siblings, meaning that they are the sole offspring of their parents. This status is a direct consequence of the only child policy implemented in China, which typically permits each couple to have only one child. Only children often experience unique familial dynamics and expectations that can influence their emotional and social development. A “left-behind child” is defined as a child who remains in their hometown while one or both parents migrate to different regions, typically transitioning from rural to urban areas, in search of employment opportunities. These children are usually cared for by grandparents or other relatives, and they generally have limited contact with their parents, often seeing them infrequently, typically only during significant holidays such as the Spring Festival. This separation can lead to emotional and psychological challenges, as well as impact their educational experiences and overall wellbeing. Ottawa self-injury inventory (OSI) Non-suicidal self-injury (NSSI) was assessed using the Ottawa Self-Injury Inventory (OSI), a rigorously validated instrument specifically designed to identify deliberate self-inflicted harm in the absence of suicidal intent [41, 42]. In the present study, an abbreviated 3-item version was employed to assess the occurrence of NSSI behaviors over the preceding 1 month, 6 months, and 12 months. Participants rated the frequency of NSSI behaviors on a 4-point ordinal scale (0 = 0 times, 1 = 1 time, 2 = 2 times, 3 = 3 times or more). Participants reporting any intentional self-harm without concurrent suicidal ideation during these periods were classified as exhibiting NSSI behaviors. The OSI has demonstrated robust psychometric validity and reliability across diverse adolescent populations, effectively distinguishing NSSI behaviors from suicidal self-harm behaviors [43, 44]. The abbreviated 3-item version employed here has been used in previous studies to provide a concise and reliable measure of NSSI behaviors [45, 46]. (Example Item: In the past month, how often have you intentionally engaged in self-injurious behaviors without suicidal intent? Such as cutting, biting, burning, carving, scratching, hitting or banging yourself, or rubbing your skin against rough surfaces.) Adolescent self-rating life events check list (ASLEC) The Adolescent Self-Rating Life Events Checklist (ASLEC) is a psychometrically sound instrument developed within the Chinese cultural milieu to systematically quantify exposure to stressful life events among adolescents [47, 48]. In the present cohort, the ASLEC exhibited excellent internal consistency (Cronbach’s α = 0.94). The instrument comprises 27 items spanning six dimensions: interpersonal relationships, academic pressure, punishment, bereavement, health adjustment, and others. Respondents rate the perceived impact of each event on a 5-point Likert scale (1 = no influence, 5 = severe influence), enabling precise quantification of both objective exposure and subjective stress appraisal. Elevated cumulative scores signify greater stress burden and heightened susceptibility to adverse psychological outcomes. Owing to its rigorous validation and cultural specificity, the ASLEC has been extensively utilized in epidemiological and clinical research to elucidate the influence of life stressors on adolescent mental health. (Example Item: Over the past year, I have felt misunderstood by others.) Difficulties in emotion dysregulation scale (DERS) Emotional dysregulation was assessed using the 16-item Difficulties in Emotion Regulation Scale (DERS), a psychometrically validated instrument for Chinese adolescent populations [49, 50]. In the present sample, the scale demonstrated excellent internal consistency (Cronbach’s α = 0.93). The DERS-16 evaluates five core domains of emotion regulation difficulties: difficulty in emotional awareness, difficulty in impulse control, difficulty in goal-directed behavior, difficulty in accepting emotional responses, and difficulty in effectively applying emotion regulation strategies. Each item is rated on a 5-point Likert scale ranging from 1 (“never”) to 5 (“always”), with higher total scores indicating greater deficits in emotion regulation and reduced regulatory capacity. The DERS thus constitutes a comprehensive and empirically validated instrument for the rigorous evaluation of emotion dysregulation in adolescent populations. (Example Item: I have difficulty understanding my own feelings.) Resilience scale of Chinese adolescent (RSCA) Resilience was assessed using the Resilience Scale for Chinese Adolescents (RSCA), a culturally validated instrument designed to provide a multidimensional evaluation of resilience among adolescents [51, 52]. In the present study, the RSCA exhibited high internal consistency (Cronbach’s α = 0.92). The scale comprises 27 items organized into two primary dimensions: personal strength and supportive capacity. The personal strength domain evaluates attributes such as goal focus, emotion control, and positive cognition, whereas the supportive capacity domain captures family support and interpersonal assistance. Responses are recorded on a 5-point Likert scale (1 = complete inconsistency, 5 = full consistency), with higher scores reflecting greater resilience. The RSCA represents a psychometrically robust and culturally pertinent instrument for the systematic assessment of resilience in Chinese adolescent populations. (Example Item: I have clear goals in life.) Data analysis All analyses were conducted using SPSS version 27.0 and the PROCESS macro for SPSS version 4.1, with a two-tailed significance threshold of p < 0.05. Continuous variables were summarized as means ± standard deviations, while categorical variables were reported as frequencies and percentages. Between-group differences in continuous variables were examined using independent-sample t-tests, whereas categorical variables were analyzed with Chi-squared or Mann-Whitney U tests for ordinal or non-normally distributed data. Assumptions of regression analysis, including linearity, independence of residuals, and homoscedasticity, were confirmed, and multicollinearity was evaluated using Variance Inflation Factors, all below 5, indicating no substantive multicollinearity concerns. Pearson correlation analyses assessed bivariate relationships among stressful life events, NSSI behaviors, resilience, and emotion dysregulation. Hierarchical multiple regression analyses were conducted to identify predictors of NSSI behaviors. Demographic variables were entered in Model 1, followed by the inclusion of stressful life events in Model 2, emotion dysregulation in Model 3, and resilience in Model 4. Standardized coefficients (β), t-values, and significance levels were reported for each predictor, allowing the assessment of incremental contributions of psychosocial and regulatory factors. Sequential mediation analysis was performed using Hayes’ PROCESS macro (Model 6), specifying stressful life events as the independent variable (X), NSSI behaviors as the dependent variable (Y), and resilience (M1) and emotion dysregulation (M2) as sequential mediators. Both direct and indirect effects were estimated, with statistical significance of indirect effects determined using 5,000 bootstrap resamples to generate 95% confidence intervals, and mediation effects were considered significant if intervals excluded zero. All variables were standardized to z-scores prior to analysis to obtain interpretable standardized coefficients (β). Results A total of 1,426 participants were invited to complete the online questionnaire, of which 1,392 valid responses were obtained, yielding an effective response rate of 97.61%. Descriptive analysis of demographic characteristics Table 1 summarizes the demographic characteristics of the total sample and provides comparisons between participants with and without NSSI behaviors. The mean age of participants was 18.70 years (SD = 2.86). The sample was predominantly female (90.1%) and of Han ethnicity (82.8%). Participants in the NSSI group (n = 105, 7.54%) exhibited a significantly higher prevalence of left-behind children (χ² = 7.381, p = 0.007) and only-child status (χ² = 13.290, p < 0.001) compared with the non-NSSI group (n = 1,287). No other demographic characteristics differed significantly between groups. Table 1Demographic characteristics of participants (n = 1,392)FeaturesAll (n = 1392)Non-NSSI (n = 1287)NSSI(n = 105)Statistic (χ²/t) P Gender (N/%)0.2290.632 Male138(9.9)129(10.0)9(8.6) Female1254(90.1)1158(90.0)96(91.4) Ethnicity (N/%) 0.6060.436 Han1153(82.8)1068(83.0)85(80.9) Other239(17.2)219(17.0)20(19.1) Age(M ± SD) 16.70 ± 2.8616.71 ± 2.9616.57 ± 0.990.4930.506Being left-behind children (N/%)7.3810.007 Yes473(34.0)450(35.0)23(21.9) No919(66.0)837(65.0)82(78.1)Living in dormitory (N/%)1.1010.294 Yes1290(92.7)1190(92.5)100(95.2) No102(7.3)97(7.5)5(4.8)Being only children (N/%)13.290< 0.001 Yes353(25.4)342(26.6)11(10.5) No1039(74.6)945(73.4)94(89.5)Family economic status (N/%)3.7170.156 Poor499(35.9)470(36.5)29(27.6) Fair815(58.6)747(58.0)68(64.8) Good78(5.6)70(5.5)8(7.6)Family structure(N/%)0.4250.935 Single parent/recombine family123(8.8)113(8.8)10(9.5) Living with grandparents243(17.5)226(17.6)17(16.2) Living with parents996(71.6)921(71.6)75(71.4) Living with other relatives30(2.2)27(2.1)3(2.9)Father’s educational level(N/%)1.9100.385 Junior high school or below160(11.5)152(11.8)8(7.6) Senior high school/technical school369(26.5)342(26.6)27(25.7) College or above863(62.0)793(61.6)70(66.7)Mother’s educational level(N/%)2.5540.279 Junior high school or below105(7.5)101(7.8)4(3.8) Senior high school or technical school304(21.8)278(21.6)26(24.8) College or above983(70.6)908(70.6)75(71.4)Father’s parenting style(N/%)3.3470.502 Strict241(17.3)219(17.0)22(21.0) Pamper311(22.3)292(22.7)19(18.1) Neglect124(8.9)116(9.0)8(7.6) Often changed293(21.0)274(21.3)19(18.1) Open-minded423(30.4)386(30.0)37(35.2)Mother’s parenting style(N/%)2.5190.641 Strict319(22.9)291(22.6)28(26.7) Pamper459(33.0)425(33.0)34(32.4) Neglect55(4.0)52(4.0)3(2.9) Often changed302(21.7)284(22.1)18(17.1) Open-minded257(18.5)235(18.3)22(21.0) NSSI Non-Suicidal Self-Injury, M Mean, SD Standard Deviation χ² = Chi-square value; t = t-value from independent samples t-test P-values in bold indicate statistical significance (p < 0.05) Table 1Demographic characteristics of participants (n = 1,392)FeaturesAll (n = 1392)Non-NSSI (n = 1287)NSSI(n = 105)Statistic (χ²/t) P Gender (N/%)0.2290.632 Male138(9.9)129(10.0)9(8.6) Female1254(90.1)1158(90.0)96(91.4) Ethnicity (N/%) 0.6060.436 Han1153(82.8)1068(83.0)85(80.9) Other239(17.2)219(17.0)20(19.1) Age(M ± SD) 16.70 ± 2.8616.71 ± 2.9616.57 ± 0.990.4930.506Being left-behind children (N/%)7.3810.007 Yes473(34.0)450(35.0)23(21.9) No919(66.0)837(65.0)82(78.1)Living in dormitory (N/%)1.1010.294 Yes1290(92.7)1190(92.5)100(95.2) No102(7.3)97(7.5)5(4.8)Being only children (N/%)13.290< 0.001 Yes353(25.4)342(26.6)11(10.5) No1039(74.6)945(73.4)94(89.5)Family economic status (N/%)3.7170.156 Poor499(35.9)470(36.5)29(27.6) Fair815(58.6)747(58.0)68(64.8) Good78(5.6)70(5.5)8(7.6)Family structure(N/%)0.4250.935 Single parent/recombine family123(8.8)113(8.8)10(9.5) Living with grandparents243(17.5)226(17.6)17(16.2) Living with parents996(71.6)921(71.6)75(71.4) Living with other relatives30(2.2)27(2.1)3(2.9)Father’s educational level(N/%)1.9100.385 Junior high school or below160(11.5)152(11.8)8(7.6) Senior high school/technical school369(26.5)342(26.6)27(25.7) College or above863(62.0)793(61.6)70(66.7)Mother’s educational level(N/%)2.5540.279 Junior high school or below105(7.5)101(7.8)4(3.8) Senior high school or technical school304(21.8)278(21.6)26(24.8) College or above983(70.6)908(70.6)75(71.4)Father’s parenting style(N/%)3.3470.502 Strict241(17.3)219(17.0)22(21.0) Pamper311(22.3)292(22.7)19(18.1) Neglect124(8.9)116(9.0)8(7.6) Often changed293(21.0)274(21.3)19(18.1) Open-minded423(30.4)386(30.0)37(35.2)Mother’s parenting style(N/%)2.5190.641 Strict319(22.9)291(22.6)28(26.7) Pamper459(33.0)425(33.0)34(32.4) Neglect55(4.0)52(4.0)3(2.9) Often changed302(21.7)284(22.1)18(17.1) Open-minded257(18.5)235(18.3)22(21.0) NSSI Non-Suicidal Self-Injury, M Mean, SD Standard Deviation χ² = Chi-square value; t = t-value from independent samples t-test P-values in bold indicate statistical significance (p < 0.05) Demographic characteristics of participants (n = 1,392) NSSI Non-Suicidal Self-Injury, M Mean, SD Standard Deviation χ² = Chi-square value; t = t-value from independent samples t-test P-values in bold indicate statistical significance (p < 0.05) Descriptive analysis of main variables Table 2 presents descriptive statistics for stressful life events, emotion dysregulation, and resilience. Specifically, the NSSI group demonstrated significantly elevated total scores and scores across all dimensions on the ASLEC (p < 0.01) and DERS (p < 0.01). Conversely, the NSSI group obtained significantly lower total scores and scores across all dimensions on the RSCA (p < 0.01). Table 2The level of stressful life events, emotion dysregulation, and resilienceScaleAll (n = 1392)Non-NSSI (n =1287)NSSI (n =105)t P ASLEC 27–62Total score20.47 ± 11.6916.39 ± 4.4838.28 ± 8.59−24.028< 0.001Interpersonal relationship5.83 ± 4.454.47 ± 3.3411.73 ± 3.89−21.154< 0.001Academic pressure7.25 ± 4.695.91 ± 3.7613.12 ± 3.77−20.653< 0.001Punishment2.61 ± 2.791.82 ± 1.976.04 ± 3.24−19.640< 0.001Bereavement0.68 ± 1.440.55 ± 1.311.28 ± 1.82−7.329< 0.001Health adjustment1.81 ± 2.491.51 ± 2.213.14 ± 3.16−7.557< 0.001Other2.29 ± 2.132.13 ± 2.032.96 ± 2.40−4.863< 0.001 DERS 16–74Total score31.31 ± 10.5927.38 ± 7.4344.76 ± 9.07−23.147< 0.001Difficulty in emotional awareness4.00 ± 2.484.00 ± 1.615.00 ± 2.63−18.530< 0.001Difficulty in impulse control6.00 ± 4.815.00 ± 2.648.00 ± 5.26−19.636< 0.001Difficulty in goal orientation6.00 ± 5.886.00 ± 3.239.00 ± 5.49−20.621< 0.001Difficulty in accepting emotional responses6.00 ± 5.265.00 ± 2.669.00 ± 6.01−20.033< 0.001Difficulty in effectively using regulatory strategies10.00 ± 12.669.00 ± 6.6114.00 ± 10.16−22.015< 0.001 RSCA 42–98Total Score77.33 ± 9.6781.00 ± 5.4961.88 ± 8.72−24.954< 0.001Goal focus14.00 ± 12.0515.00 ± 9.3010.00 ± 8.72−18.182< 0.001Emotion control18.00 ± 10.7618.00 ± 6.6514.00 ± 16.90−15.058< 0.001Positive cognition11.00 ± 7.5011.00 ± 4.928.00 ± 11.56−17.453< 0.001Family support18.00 ± 6.2818.00 ± 2.5915.00 ± 11.16−17.793< 0.001Interpersonal assistance18.00 ± 9.0319.00 ± 4.8815.00 ± 12.59−17.545< 0.001 NSSI Non-Suicidal Self-Injury, t t-value from independent samples t-test P-values in bold indicate statistical significance (p < 0.05) Table 2The level of stressful life events, emotion dysregulation, and resilienceScaleAll (n = 1392)Non-NSSI (n =1287)NSSI (n =105)t P ASLEC 27–62Total score20.47 ± 11.6916.39 ± 4.4838.28 ± 8.59−24.028< 0.001Interpersonal relationship5.83 ± 4.454.47 ± 3.3411.73 ± 3.89−21.154< 0.001Academic pressure7.25 ± 4.695.91 ± 3.7613.12 ± 3.77−20.653< 0.001Punishment2.61 ± 2.791.82 ± 1.976.04 ± 3.24−19.640< 0.001Bereavement0.68 ± 1.440.55 ± 1.311.28 ± 1.82−7.329< 0.001Health adjustment1.81 ± 2.491.51 ± 2.213.14 ± 3.16−7.557< 0.001Other2.29 ± 2.132.13 ± 2.032.96 ± 2.40−4.863< 0.001 DERS 16–74Total score31.31 ± 10.5927.38 ± 7.4344.76 ± 9.07−23.147< 0.001Difficulty in emotional awareness4.00 ± 2.484.00 ± 1.615.00 ± 2.63−18.530< 0.001Difficulty in impulse control6.00 ± 4.815.00 ± 2.648.00 ± 5.26−19.636< 0.001Difficulty in goal orientation6.00 ± 5.886.00 ± 3.239.00 ± 5.49−20.621< 0.001Difficulty in accepting emotional responses6.00 ± 5.265.00 ± 2.669.00 ± 6.01−20.033< 0.001Difficulty in effectively using regulatory strategies10.00 ± 12.669.00 ± 6.6114.00 ± 10.16−22.015< 0.001 RSCA 42–98Total Score77.33 ± 9.6781.00 ± 5.4961.88 ± 8.72−24.954< 0.001Goal focus14.00 ± 12.0515.00 ± 9.3010.00 ± 8.72−18.182< 0.001Emotion control18.00 ± 10.7618.00 ± 6.6514.00 ± 16.90−15.058< 0.001Positive cognition11.00 ± 7.5011.00 ± 4.928.00 ± 11.56−17.453< 0.001Family support18.00 ± 6.2818.00 ± 2.5915.00 ± 11.16−17.793< 0.001Interpersonal assistance18.00 ± 9.0319.00 ± 4.8815.00 ± 12.59−17.545< 0.001 NSSI Non-Suicidal Self-Injury, t t-value from independent samples t-test P-values in bold indicate statistical significance (p < 0.05) The level of stressful life events, emotion dysregulation, and resilience NSSI Non-Suicidal Self-Injury, t t-value from independent samples t-test P-values in bold indicate statistical significance (p < 0.05) Correlation analysis Table 3 presents the correlation analysis of NSSI behaviors, stressful life events, emotion dysregulation, and resilience. NSSI behaviors demonstrated significant positive correlations with stressful life events (r = 0.408, p < 0.01) and emotion dysregulation (r = 0.430, p < 0.01), and a significant negative correlation with resilience (r = −0.472, p < 0.01). Stressful life events were strongly negatively correlated with resilience (r = −0.742, p < 0.01) and positively correlated with emotion dysregulation (r = 0.592, p < 0.01). Resilience and emotion dysregulation also exhibited a strong negative correlation (r = −0.716, p < 0.01). Table 3The correlation analysis of NSSI behaviors, stressful life events, emotion dysregulation, and resilienceVariables1234 1. NSSI behaviors1- 2.Stressful life events0.408**1 3.Resilience−0.472**−0.742**1 4.Emotion dysregulation0.430**0.592**−0.716**1 * P < 0.05; **P < 0.01; ***P < 0.001 Table 3The correlation analysis of NSSI behaviors, stressful life events, emotion dysregulation, and resilienceVariables1234 1. NSSI behaviors1- 2.Stressful life events0.408**1 3.Resilience−0.472**−0.742**1 4.Emotion dysregulation0.430**0.592**−0.716**1 * P < 0.05; **P < 0.01; ***P < 0.001 The correlation analysis of NSSI behaviors, stressful life events, emotion dysregulation, and resilience * P < 0.05; **P < 0.01; ***P < 0.001 Multiple regression analysis of all variables Table 4 presents hierarchical multiple regression analyses were conducted to examine the determinants of NSSI behaviors. In Model 1, demographic variables including gender, ethnicity, age, being left-behind children, living in dormitory, being only children, family economic status, family structure, parental education, and parenting styles were entered. Among these, being left-behind children (β = 0.08, t = 2.76, p = 0.01) and being only children (β = 0.10, t = 3.58, p < 0.001) emerged as significant predictors. In Model 2, stressful life events were added, showing a strong positive association with NSSI behaviors (β = 0.41, t = 16.19, p < 0.001). In Model 3, emotion dysregulation was included, revealing that both stressful life events (β = 0.24, t = 8.14, p < 0.001) and emotion dysregulation (β = 0.29, t = 9.70, p < 0.001) independently predicted NSSI behaviors, suggesting that difficulties in emotion regulation partially mediate the effects of stressful experiences. In Model 4, resilience was incorporated, demonstrating that while stressful life events (β = 0.11, t = 3.19, p < 0.001) and emotion dysregulation (β = 0.17, t = 5.12, p < 0.001) remained significant, resilience exerted a robust protective effect (β = −0.27, t = −6.55, p < 0.001), attenuating the risk of NSSI behaviors. Table 4Hierarchical multiple regression analysis of factors predicting NSSI behaviorsModelUnstandardizedCoefficientStandardized Coefficientt p BSEBeta 1 Constant−1.270.76−1.670.10Being left-behind children0.330.120.082.760.01Being only children0.460.130.103.580.00 2 Constant−2.240.70−3.190.00Being left-behind children0.300.110.072.700.01Stressful life events0.070.000.4116.190.00 3 Constant−3.290.69−4.780.00Being left-behind children0.260.110.062.390.02Stressful life events0.040.010.248.140.00Emotion dysregulation0.060.010.299.700.00 4 Constant2.401.102.180.03Being left-behind children0.230.110.052.140.03Stressful life events0.020.010.113.190.00Emotion dysregulation0.030.010.175.120.00Resilience−0.060.01−0.27−6.550.00 Table 4Hierarchical multiple regression analysis of factors predicting NSSI behaviorsModelUnstandardizedCoefficientStandardized Coefficientt p BSEBeta 1 Constant−1.270.76−1.670.10Being left-behind children0.330.120.082.760.01Being only children0.460.130.103.580.00 2 Constant−2.240.70−3.190.00Being left-behind children0.300.110.072.700.01Stressful life events0.070.000.4116.190.00 3 Constant−3.290.69−4.780.00Being left-behind children0.260.110.062.390.02Stressful life events0.040.010.248.140.00Emotion dysregulation0.060.010.299.700.00 4 Constant2.401.102.180.03Being left-behind children0.230.110.052.140.03Stressful life events0.020.010.113.190.00Emotion dysregulation0.030.010.175.120.00Resilience−0.060.01−0.27−6.550.00 Hierarchical multiple regression analysis of factors predicting NSSI behaviors Multiple regression analysis of core psychological variables A separate multiple regression analysis was conducted with NSSI behaviors as the dependent variable and stressful life events, resilience, and emotion dysregulation as independent predictors. The model was statistically significant, F(3,1388) = 187.25, p < 0.001, and explained 28.8% of the variance in NSSI behaviors (R² =0.288). All three variables were significant unique predictors: stressful life events (β = 0.11, p < 0.001), resilience (β = −0.27, p < 0.001), and emotion dysregulation (β = 0.17, p < 0.001). This confirms that, collectively, these psychological constructs account for a substantial and significant portion of the variance in NSSI behaviors. Chain mediating effect analysis Sequential mediation was tested using Model 6 of the PROCESS macro for SPSS, with all variables standardized (z-scores) to facilitate interpretation of standardized coefficients (β) (as shown in Tables 5 and 6).Table 5Regression coefficients for the paths in the chain mediation modelPredictorResilience (M1)Emotion dysregulation(M2)NSSI behaviors (Y)Constant90.205***80.427***3.619***Stressful life events (X)−0.609***0.117***0.019**Resilience (M1)−0.678***−0.057***Emotion dysregulation (M2)0.034***R20.7420.7220.495Adjusted R20.5510.5220.245F statistic852.365***504.361***112.763***R²: indicates the proportion of variance in the dependent variable explained by the independent variable(s) and mediator(s) collectively in the chained mediation modelAdjusted R²: accounts for the number of predictors in the model to provide a more accurate measure of explanatory power in the mediation contextF statistic: tests the overall significance of the chained mediation regression model *p <0 .05, **p <0 .01, ***p <0 .001Table 6Bootstrap analysis of the direct and indirect effects in the chain mediation modelEffect TypeEffectBoot LLCIBoot ULCIEffect Contribution (%)Total0.0720.0630.081100Direct0.0190.0070.03126.39Indirect0.0530.0370.07373.61Stressful life events → Resilience→ NSSI behaviors0.0350.0180.05948.61Stressful life events →Emotion dysregulation → NSSI behaviors0.00400.015.56Stressful life events→ Resilience → Emotion dysregulation → NSSI behaviors0.0140.0090.01919.44CI Confidence Interval, Boot LLCI Lower limit of the bootstrap confidence interval, Boot ULCI Upper limit of the bootstrap confidence interval *p <0 .05, **p <0 .01, ***p <0 .001 Table 5Regression coefficients for the paths in the chain mediation modelPredictorResilience (M1)Emotion dysregulation(M2)NSSI behaviors (Y)Constant90.205***80.427***3.619***Stressful life events (X)−0.609***0.117***0.019**Resilience (M1)−0.678***−0.057***Emotion dysregulation (M2)0.034***R20.7420.7220.495Adjusted R20.5510.5220.245F statistic852.365***504.361***112.763***R²: indicates the proportion of variance in the dependent variable explained by the independent variable(s) and mediator(s) collectively in the chained mediation modelAdjusted R²: accounts for the number of predictors in the model to provide a more accurate measure of explanatory power in the mediation contextF statistic: tests the overall significance of the chained mediation regression model *p <0 .05, **p <0 .01, ***p <0 .001 Regression coefficients for the paths in the chain mediation model R²: indicates the proportion of variance in the dependent variable explained by the independent variable(s) and mediator(s) collectively in the chained mediation model Adjusted R²: accounts for the number of predictors in the model to provide a more accurate measure of explanatory power in the mediation context F statistic: tests the overall significance of the chained mediation regression model *p <0 .05, **p <0 .01, ***p <0 .001 Table 6Bootstrap analysis of the direct and indirect effects in the chain mediation modelEffect TypeEffectBoot LLCIBoot ULCIEffect Contribution (%)Total0.0720.0630.081100Direct0.0190.0070.03126.39Indirect0.0530.0370.07373.61Stressful life events → Resilience→ NSSI behaviors0.0350.0180.05948.61Stressful life events →Emotion dysregulation → NSSI behaviors0.00400.015.56Stressful life events→ Resilience → Emotion dysregulation → NSSI behaviors0.0140.0090.01919.44CI Confidence Interval, Boot LLCI Lower limit of the bootstrap confidence interval, Boot ULCI Upper limit of the bootstrap confidence interval *p <0 .05, **p <0 .01, ***p <0 .001 Bootstrap analysis of the direct and indirect effects in the chain mediation model *p <0 .05, **p <0 .01, ***p <0 .001 The model examined the indirect effects of stressful life events (X) on NSSI behaviors (Y) through resilience (M1) and emotion dysregulation (M2) sequentially. Regression coefficients for each path are summarized in Table 5. Stressful life events significantly predicted lower resilience (β = −0.609, p < 0.001) and higher emotion dysregulation (β = 0.117, p < 0.001). Resilience, in turn, significantly predicted reduced emotion dysregulation (β = −0.678, p < 0.001) and lower NSSI behaviors (β = −0.057, p < 0.001), while emotion dysregulation positively predicted NSSI behaviors (β = 0.034, p < 0.001). Bootstrapped indirect effect analysis using 5,000 resamples further confirmed the sequential mediation effects Table 6. The total indirect effect of stressful life events on NSSI behaviors through the mediators was significant (Effect = 0.053, Boot SE = 0.009, 95% BootCI [0.037, 0.073]), accounting for 73.61% of the total effect. Three specific indirect pathways reached significance Table 6: ASLEC → Resilience → NSSI behaviors (Effect = 0.035, 95% BootCI [0.018, 0.059]); ASLEC → Emotion Dysregulation → NSSI behaviors (Effect = 0.004, 95% BootCI [0.000, 0.010]); and ASLEC → Resilience → Emotion Dysregulation → NSSI behaviors (Effect = 0.014, 95% BootCI [0.009, 0.019]). The direct effect of stressful life events on NSSI behaviors remained significant (Effect = 0.019, 95% BootCI [0.007, 0.031]), supporting a model of partial mediation. The sequential mediation model accounted for 49.5% of the variance in NSSI behaviors (R² = 0.495), reflecting strong theoretical coherence and the precision afforded by the large sample. A conceptual diagram illustrating the chain mediation model is presented in Fig. 1.Fig. 1The chain mediation model of resilience and emotion dysregulation between stressful life events and NSSI The chain mediation model of resilience and emotion dysregulation between stressful life events and NSSI Discussion The prevalence of NSSI behaviors The prevalence of NSSI behaviors in this study (7.54%) is substantially lower than that reported in Western populations, such as India (22.4%) and Belgium (10.3% in year 1 and 6.0% in year 2) [45, 63]. Cultural differences in the manifestation and reporting of self-injurious behaviors and methodological differences, including study design and measurement instruments, may account for this variation. For example, the elevated prevalence observed in the Indian cohort may reflect heightened socioeconomic inequalities and academic pressures, whereas the decreasing trend reported in the Belgian longitudinal study may indicate the effectiveness of institutional mental health interventions over time [45, 63]. Conversely, the markedly higher prevalence (29%) identified in a rural Chinese sample highlights the potential impact of regional disparities and limited access to mental health resources in under-resourced settings [53]. Analysis of our provincial college student sample provides further granularity, illuminating heterogeneity within sub-populations in China. Differences in operational definitions of NSSI behaviors, sample composition, and timing of data collection relative to academic cycles may also contribute to variability across studies. Beyond cultural and methodological explanations, temporal factors such as the timing of data collection relative to academic cycles (e.g., examination periods or semester transitions) may substantially influence stress exposure and behavioral outcomes. Furthermore, contextual factors, including regional disparities in mental health resources and the availability of institutional support systems, may also contribute to cross-study differences. These findings affirm that NSSI behaviors represent a significant mental health concern among college students and underscore the necessity of identifying both risk and protective factors to inform prevention and intervention strategies during this critical developmental period. The impact of only-child and left-behind status Both left-behind children and only children exhibited a disproportionately high prevalence of NSSI behaviors, underscoring the complex interplay of sociodemographic factors in shaping vulnerability to self-injurious behaviors within the Chinese context. Elevated NSSI behaviors among left-behind students likely reflect prolonged parental absence, which can disrupt the formation of secure attachment, limit exposure to effective emotional scaffolding, and constrain opportunities to acquire adaptive coping strategies in response to stress [54, 55]. These deficits may predispose individuals to reliance on maladaptive regulatory mechanisms, with NSSI behaviors functioning as an immediate yet maladaptive strategy to alleviate psychological distress triggered by stressful life events. The heightened prevalence of NSSI behaviors among only children indicates that concentrated parental attention and resource abundance do not uniformly confer resilience. Intensified parental expectations and heightened self-imposed performance pressures may exacerbate emotional dysregulation, increasing susceptibility to self-injurious behaviors. This observation aligns with emerging evidence suggesting that only children, despite potential advantages in familial support and material resources, encounter distinctive psychosocial stressors that elevate the risk of NSSI behaviors [56, 57]. These findings indicate that both parental absence and concentrated familial expectations constitute critical, yet distinct, risk factors for NSSI behaviors among Chinese college students. Prevention and intervention efforts should be strategically tailored to these subpopulations, integrating structured peer support networks, school-based mentoring programs, and accessible mental health services. Interventions should prioritize the enhancement of psychological resilience, promotion of adaptive emotion regulation, and establishment of alternative channels for emotional expression. Such strategies address individual vulnerabilities while emphasizing the necessity of culturally sensitive and contextually informed frameworks for mitigating NSSI behaviors in Chinese collegiate populations. The predict role of stressful life events Exposure to stressful life events markedly increases the probability of engagement in NSSI behaviors. This finding is consistent with an extensive body of literature documenting a positive association between stress exposure and self-injurious behaviors across diverse populations [34, 58]. A prospective longitudinal study of 1,482 college students revealed that stressful events related to academic life, peer interactions, intimate relationships, and family dynamics were associated with increased engagement in NSSI behaviors [32]. The Experiential Avoidance Model posits that self-injurious behaviors serve as strategies to alleviate emotional distress and psychosocial strain in response to adverse events [3]. Students encountering stressful circumstances may engage in NSSI behaviors as a coping mechanism to manage overwhelming emotional and psychological challenges. Although the association between stressful life events and NSSI behaviors has been firmly established, the psychological mechanisms mediating this relationship remain insufficiently clarified. The present study addresses this gap through a sequential mediation model, revealing how resilience and emotion dysregulation transform stress exposure into NSSI behaviors. This approach provides a clear framework linking stress experiences to specific psychological pathways that drive maladaptive behavior. Early identification of stressors and the application of targeted interventions are critical for reducing risk, while coordinated support from mental health professionals, educators, and families can foster adaptive coping strategies and decrease vulnerability to NSSI behaviors. The chain mediating effect of resilience and emotion dysregulation Sequential mediation analysis elucidates the central roles of resilience and emotion dysregulation, highlighting the dynamic interplay between protective and vulnerability factors in shaping NSSI behaviors. While the individual path coefficients in the model (Fig. 1) may appear small, it is important to interpret them within the context of a multiple mediation model. The total effect of stressful life events on NSSI is partitioned into direct and several indirect pathways. The significance of these paths, confirmed by bootstrapping, indicates robust mediating mechanisms. The model collectively explains a substantial proportion of variance in NSSI (R² =0 0.495), underscoring the clinical relevance of this pathway. Resilience serves as the primary mediator, reflecting the capacity to maintain adaptive functioning under adversity. Elevated resilience confers a repertoire of psychological resources, encompassing effective coping strategies, refined emotion regulation, and robust social support networks, collectively mitigating emotional distress and reducing susceptibility to NSSI. Moreover, resilience exerts a downstream influence on emotion regulation processes, establishing a foundational mechanism for preventing maladaptive coping behaviors. Empirical evidence demonstrates that resilience attenuates the detrimental psychological impact of stressful life events [37, 59], while studies among Chinese college students indicate that it buffers the effects of depressive symptoms under stress, underscoring its protective function against NSSI behaviors [58]. The current findings extend this understanding by positioning resilience not merely as a moderator but as a critical mediator that mitigates subsequent emotion dysregulation. Emotion dysregulation emerges as the proximal mediator linking stress exposure to NSSI behaviors. Inefficient management of negative affect promotes maladaptive coping strategies, increasing the likelihood of self-injury. The emotional cascade model posits that NSSI serves as a transient mechanism for alleviating overwhelming negative emotions [40]. The present study demonstrates that emotion dysregulation mediates the effect of resilience on NSSI, integrating prior findings into a coherent sequential framework. Meta-analytic evidence confirms a robust association between emotion dysregulation and NSSI across diverse populations [60], and dysregulated emotional processing mediates the influence of personality traits, such as neuroticism, on self-injurious behaviors [61]. Collectively, these findings underscore the critical importance of interventions that concurrently enhance resilience and develop adaptive emotion regulation skills to effectively reduce engagement in NSSI behaviors.Furthermore, socio-demographic factors (e.g., only-child, left-behind status) were significant predictors in initial models is well-noted. These factors represent crucial distal risk markers that shape an individual’s developmental context. However, the full mediation model (Table 4, Model 4) suggests that their influence may be largely channeled through proximal psychological mechanisms—namely, by depleting resilience and exacerbating emotion dysregulation. This delineation is vital for intervention, as it identifies resilience and emotion regulation as modifiable treatment targets that can mitigate the risk conveyed by fixed or background sociodemographic factors. Implications Clarification of the sequential mediation pathway offers a robust framework for designing interventions aimed at mitigating NSSI behaviors in college populations. Structured programs that integrate NSSI awareness, resilience enhancement, and emotion regulation training within student health services facilitate proactive identification, timely assessment, and personalized psychosocial support. Coordination among mental health professionals, educators, and families cultivates environments that systematically reinforce resilience and foster adaptive emotion regulation, thereby attenuating vulnerability to self-injurious behaviors. Targeted interventions emphasizing problem-solving acumen, stress management competencies, and sophisticated emotion regulation strategies equip students with psychological flexibility to navigate adversity. Systematic evaluation of students exhibiting low resilience or deficits in emotion regulation enables precise, tailored nursing interventions that maximize preventive efficacy. Nursing practice should incorporate comprehensive assessments that extend beyond external stressors to encompass intrinsic psychological capacities, including resilience, coping repertoires, and emotion regulation proficiency. Prioritizing these protective factors establishes an evidence-informed foundation for the prevention of NSSI behaviors and the promotion of enduring mental health within collegiate populations. Limitation Several limitations should be considered when interpreting the present findings. First, the cross-sectional design inherently constrains the ability to infer causal relationships or establish temporal precedence in the proposed sequential mediation pathways. Although the sequential mediation model delineates theoretically plausible links between stressful life events, resilience, emotion dysregulation, and NSSI behaviors, the absence of temporal ordering precludes empirical validation of the directionality among these constructs. Therefore, the proposed mechanisms must be viewed as hypothetical and future longitudinal or experimental investigations are indispensable to substantiate causal claims and elucidate the directionality of these psychological processes. Second, the reliance on a convenience sample drawn from a limited subset of public universities in Hunan Province restricts the broader applicability of the findings. Given the inherent sociocultural heterogeneity within the Chinese college student population, the patterns observed in this sample may not fully represent national trends. Replications in multi-site studies or among nationally representative cohorts are therefore essential to establish the external validity of these conclusions. Third, all constructs were assessed using self-report instruments, which are inherently prone to biases, such as retrospective recall errors, social desirability bias, and common method variance. Furthermore, self-report measures may fail to capture the full clinical complexity of NSSI behaviors, potentially leading to an underestimation of prevalence or an oversimplification of behavioral patterns. Specifically, the abbreviated three-item version of the Ottawa Self-Injury Inventory (OSI), while efficient for large-scale screening, may inadequately encompass the full phenomenological range of NSSI, thus potentially obscuring subtle behavioral variations. Fourth, although the psychometric tools utilized in this study have been validated in Chinese populations, the cultural context surrounding constructs such as resilience and emotion dysregulation may introduce subtle measurement biases, influencing the validity of these constructs. Future studies should focus on the cross-cultural refinement and further validation of these instruments to enhance their construct validity and ensure interpretative accuracy across diverse cultural contexts. Finally, while the relatively large sample size provides sufficient statistical power, the model accounts for only approximately half of the variance in NSSI behaviors, suggesting that other significant factors, such as genetic predispositions, trauma history, or co-occurring psychiatric comorbidities, were not included and may play crucial roles in the development of NSSI behaviors. Thus, future research should explore these additional variables to further elucidate the multifactorial etiology of NSSI behaviors. Conclusion Guided by theoretical models of resilience and emotion dysregulation [39, 62], this cross-sectional study provides evidence for a sequential mediation pathway in which resilience functions as an initial mediator that buffers the impact of stressful life events, modulating subsequent emotion dysregulation that directly influences engagement in NSSI behaviors among college students. The delineation of this chain of modifiable psychological mechanisms suggests a potential conceptual framework for understanding how environmental stressors translate into NSSI behaviors within the college population. An analysis of sociodemographic characteristics reveals that college students engaging in NSSI behaviors exhibit a disproportionate prevalence of being left-behind or only children compared to their peers. This pattern suggests the significance of individual psychological traits and sociocultural influences in modulating vulnerability to NSSI behaviors, though further replication studies across diverse populations are necessary to establish the generalizability of these findings. Unmet mental health needs among Chinese college students remain pronounced, with many individuals who engage in NSSI behaviors remaining untreated [63]. The findings underscore the urgent need to develop and implement more accessible mental health services specifically tailored to this population. While national health initiatives such as Healthy China 2030 provide an opportunity to address these gaps, their impact on mental health outcomes requires rigorous evaluation to ensure their effectiveness and scalability. Future research should focus on evaluating university-based mental health interventions aimed at enhancing resilience and emotion regulation, which may serve as effective strategies for reducing the frequency and severity of NSSI behaviors. However, the specific elements of these interventions that drive effectiveness remain unclear and warrant further exploration. The study contributes to the understanding of the sequential psychological processes linking stressful life events to NSSI behaviors, while highlighting the heightened vulnerability of specific sociodemographic subgroups. These findings may inform the development of targeted and effective preventive interventions within college populations.