Workplace violence, legal knowledge, and support for nursing legislation among nurses in China: A multicenter cross-sectional study Authors: Xiaolong Wang, Yufei Xing, Yinghui Wu, Qiong Fang Published: 2026 ABSTRACT AbstractBackgroundWorkplace violence is a major occupational hazard for nurses, but its association with support for legislative protection remains unclear. This study examined the associations among workplace violence, nursing-related legal knowledge, and explicit support for nursing legislation.MethodsA multicenter cross-sectional survey was conducted among 4,391 nurses from 13 hospitals in Shanghai, China, in January 2024. Workplace violence was assessed using a weighted poly-victimization index based on eight violence types. Nursing-related legal knowledge and legislative support were assessed using 30-item and 14-item instruments, respectively. Spearman correlations and PROCESS Model 4 were used to estimate the indirect association between workplace violence and legislative support through nursing-related legal knowledge, adjusting for demographic and occupational covariates. Exploratory subgroup and sensitivity analyses examined department risk level, alternative workplace violence scores, age and work-tenure specifications, and hospital-level clustering.ResultsWorkplace violence was weakly negatively correlated with legal knowledge (r = -0.036, p = .017), whereas legal knowledge was positively correlated with legislative support (r = 0.131, p < .001). Workplace violence and legislative support were essentially uncorrelated in the unadjusted analysis (r = −0.001, p = .947). In the adjusted model, the total association was small (β = −0.046, 95% CI [−0.076, −0.016], p = .002), the direct association was not statistically significant (β = −0.025, 95% CI [−0.054, 0.004], p = .088), and the indirect association through legal knowledge was also small (β = −0.021, 95% bootstrap CI [−0.030, −0.012]). Subgroup analyses showed descriptive differences across department risk groups, and sensitivity analyses yielded similar coefficient estimates.ConclusionsNursing-related legal knowledge was associated with stronger explicit support for nursing legislation and contributed to a small indirect statistical association between workplace violence and legislative support. The small effect estimates and cross-sectional design preclude causal or strong policy conclusions. Longitudinal studies are needed to clarify temporal ordering and practical relevance.Keywords: Workplace violence, Nursing-related legal knowledge, Nursing legislation, Nurses, Occupational safety, Policy engagement What is already known •Workplace violence is a major occupational hazard for nurses and is associated with adverse psychological and organizational outcomes. Workplace violence is a major occupational hazard for nurses and is associated with adverse psychological and organizational outcomes. •Support for nursing legislation is relevant to occupational protection, professional rights, and workforce safety. Support for nursing legislation is relevant to occupational protection, professional rights, and workforce safety. •Nursing-related legal knowledge may shape how nurses understand professional responsibilities and statutory protections. Nursing-related legal knowledge may shape how nurses understand professional responsibilities and statutory protections. What this paper adds •This study examines support for nursing legislation as a policy-related outcome in workplace violence research. This study examines support for nursing legislation as a policy-related outcome in workplace violence research. •Nursing-related legal knowledge formed part of a small indirect statistical association between workplace violence and legislative support. Nursing-related legal knowledge formed part of a small indirect statistical association between workplace violence and legislative support. •Exploratory subgroup analyses showed descriptive differences across clinical settings, without establishing between-group differences. Exploratory subgroup analyses showed descriptive differences across clinical settings, without establishing between-group differences. Introduction Workplace violence (WPV) against healthcare professionals remains a major occupational and public health concern worldwide. Nurses are particularly vulnerable because of their sustained contact with patients and families in demanding clinical environments (Spector et al., 2014; Varghese et al., 2022). Such violence may involve multiple forms of aggression, including verbal abuse, psychological harassment, threats, physical assault, sexual harassment, and other boundary violations (Najafi et al., 2018; Wang et al., 2025; Zhang et al., 2025). Repeated or cumulative exposure has been associated with impaired well-being, burnout, turnover intentions, and risks to care quality and patient safety (Havaei, 2021; Lanctôt and Guay, 2014; O’Brien et al., 2024). In China, nurses’ legal protection has received increasing attention alongside broader efforts to strengthen nursing professionalization and practice environment (Wang et al., 2021). A dedicated legal framework for nursing practice has therefore become an important professional and policy issue. However, the factors associated with frontline nurses’ support for nursing legislation remain insufficiently understood (Wang et al., 2023). WPV research has focused mainly on psychological and organizational consequences, with less attention to nurses’ policy-oriented responses or support for statutory protection (de Raeve et al., 2023). Examining legislative support may therefore extend WPV research beyond conventional health and workforce outcomes (Havaei, 2021; Stafford et al., 2022). Cognitive mediational theory proposes that responses to stressful experiences depend partly on how individuals appraise the threat and the resources available to manage it (Folkman et al., 1986). Within this framework, nursing-related legal knowledge may be relevant to how nurses understand professional risks, rights, responsibilities, and statutory protection (Yassine et al., 2021). Legal knowledge has also been associated with professional judgment and awareness of legal obligations among healthcare professionals (Lim and Yi, 2021). However, little is known about the concurrent associations among cumulative WPV exposure, nursing-related legal knowledge, and support for nursing legislation (Vrablik et al., 2019). These associations may also vary across clinical environments. European multicountry evidence indicates that nurses’ exposure to and reporting of WPV differ across national and institutional contexts, while violence risk is often elevated in emergency and other high-intensity clinical settings (Babiarczyk et al., 2020; Çakır et al., 2025). This study therefore examined the associations among cumulative WPV exposure, nursing-related legal knowledge, and nurses’ explicit support for nursing legislation in China. Guided by cognitive mediational theory, we estimated an indirect association through legal knowledge and explored descriptive patterns across low-risk and high-risk clinical departments. The specified ordering was theory-informed and exploratory and was not intended to imply an established temporal sequence. By addressing these questions, this study extends WPV research beyond traditional health and organizational outcomes and provides evidence relevant to legal education, occupational protection, and nursing legislation. Methods Study design This cross-sectional survey was conducted between January 21 and January 31, 2024, among nurses from 13 hospitals in Shanghai, China. The number of participants contributed by each hospital and hospital-level characteristics are provided in Supplementary Table S1. The study was reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology checklist (Appendix 1) (von Elm et al., 2007). Participants A convenience cluster sampling strategy was used. Eligible participants were nurses aged 18 years or older who were currently engaged in clinical practice and provided informed consent. Nurses who were on leave or not actively working during the study period were excluded. The sample size was estimated using the events-per-variable principle, with a recommended ratio of 10∼20 participants per variable (Peduzzi et al., 1996). Because the initial survey included 56 variables, the minimum required sample size was estimated at 1,120 participants. After allowing for a potential 20% non-response or invalid response rate, the target sample size was 1,344. A total of 4,481 questionnaires were submitted. Ninety questionnaires were excluded because they did not meet the prespecified data-quality criteria, including implausibly short completion times or internally inconsistent responses. The final analytic sample therefore comprised 4,391 questionnaires, representing 98.0% of all submitted questionnaires. Variables All study variables were measured using a structured questionnaire developed for this study. The questionnaire was constructed based on national policy documents, existing literature, and multidisciplinary expert consultation. The development process included item generation, two rounds of expert review, and pilot testing. Detailed information on questionnaire item composition is provided in Appendix 2. Demographic and occupational characteristics Demographic and occupational information included gender, age, highest educational attainment, professional title, years of work, employment status, and current department. For subgroup analyses, departments were categorized into low-risk and high-risk settings according to clinical exposure intensity and prior literature on WPV risk. High-risk departments included emergency departments, intensive care units, operating rooms, and pediatrics; all other clinical departments were categorized as low-risk departments. Workplace violence exposure Workplace violence was assessed using a self-reported multiple-response item covering eight types of incidents encountered during clinical practice. The item did not specify a fixed recall period or restrict reported incidents to the participant’s current hospital. A study-specific weighted poly-victimization index was constructed to reflect both the number and perceived severity of violence types. Verbal insult was assigned 1 point; threats and damage to hospital or personal property, 2 points; psychological harassment, physical attack, and actual or threatened physical injury to the nurse or another person, 3 points; and sexual assault and unauthorized photographic or video recording, 4 points. The severity levels were determined through multidisciplinary expert consultation based on potential physical harm, psychological impact, violations of dignity or privacy, and implications for professional safety. Unauthorized photographic or video recording was classified as a severe digital privacy and reputational violation rather than physical violence. Scores ranged from 0 to 22, with higher scores indicating greater cumulative exposure and perceived severity. Because the weights were based primarily on expert judgement rather than externally validated severity coefficients, sensitivity analyses repeated the models using an unweighted PVI and a simplified two-level weighted PVI (Spector et al., 2014; Vento et al., 2020). Nursing-related legal knowledge Nursing-related legal knowledge was operationalized as participants’ factual knowledge of laws and regulations relevant to nursing practice. It was assessed using a 30-item knowledge instrument developed based on the National Nursing Development Plan (2021–2025) and six statutory or regulatory documents relevant to nursing practice and healthcare safety in China. These documents included the Nursing Regulations, the Measures for the Administration of Registration of Practicing Nurses, the Measures for the Nurse Licensure Examination, the Law of the People’s Republic of China on Basic Medical and Health Care and the Promotion of Health, the Regulation on the Handling of Medical Accidents, and the Measures for the Administration of Hospital Infection. The item pool was refined through two rounds of multidisciplinary expert consultation. Content validity was assessed by seven experts using five-point item-importance ratings. Ratings of 4 or 5 were considered content-valid. The S-CVI/Ave was 0.919, indicating good overall content validity. Item-level results are provided in Supplementary Table S2. Each item was scored dichotomously, with 1 point assigned to a correct response and 0 points assigned to an incorrect or “unclear” response. Scores were summed to generate a total score ranging from 0 to 30, with higher scores indicating greater legal knowledge. The Kuder–Richardson Formula 20 coefficient was 0.797. The first principal component analysis explained 16.3% of the variance, indicating that the items did not form a strongly unidimensional scale. The summed score was therefore used as a pragmatic composite indicator of overall nursing-related legal knowledge rather than as a measure of a single latent construct. Legislative support Legislative support was operationalized as explicit endorsement of a range of proposed statutory provisions relevant to nursing practice. The 14 items covered several related legislative domains, including professional authority, protection against workplace violence, and legal provisions governing emergency nursing care. Content validity was assessed by seven experts using five-point item-importance ratings. Ratings of 4 or 5 were considered content-valid. The I-CVI values ranged from 0.714 to 1.000, and the S-CVI/Ave was 0.898. Items with lower I-CVI values were reviewed and revised following expert consultation. Item results are provided in Supplementary Table S3. The total score represented general explicit support for nursing-related legislation rather than support for a single legislative proposal, occupational protection alone, or actual policy-engagement behavior. Responses were recorded as “support”, “do not support”, or “unclear”. In the primary analysis, each item was recoded as an indicator of explicit support, with “support” coded as 1 and both “do not support” and “unclear” coded as 0. This coding distinguished explicit endorsement from responses that did not indicate explicit support; it did not assume that uncertainty was conceptually equivalent to opposition. Item scores were summed to generate an explicit legislative support score ranging from 0 to 14, with higher scores indicating stronger explicit support for nursing legislation. To examine the treatment of uncertainty, an alternative score coded “do not support” as 0, “unclear” as 1, and “support” as 2, treating uncertainty as intermediate between non-support and explicit support. The alternative score ranged from 0 to 28. The instrument showed excellent internal consistency, with a Cronbach’s α of 0.942. Principal component analysis supported the use of a total score, with a Kaiser–Meyer–Olkin value of 0.956, a significant Bartlett’s test of sphericity, χ² = 47,695.214, df = 91, p < .001, and 60.4% of the variance explained by the first component. Data collection Data were collected using Wenjuanxing (https://www.wjx.cn), a widely utilized online survey platform in China. Institutional permission was obtained from all participating hospitals before data collection. The survey link and QR code were distributed through nursing administrative networks. Nursing department directors disseminated the survey to frontline clinical nurses within their hospitals. The first page of the survey provided study information, including the study purpose, confidentiality statement, and voluntary nature of participation. Hospital and department administrators were responsible only for forwarding the survey invitation and did not have access to individual responses or participation records. The invitation stated that participation or non-participation would not affect employment, performance evaluation, or professional relationships. No names, employee numbers, or other direct personal identifiers were collected. Completion of the questionnaire was considered provision of informed consent. The survey was anonymous, and participants could withdraw at any time. All analytic items were required for questionnaire submission, and unsubmitted questionnaires were not recorded as completed responses. Each IP address was restricted to one submission. Submitted questionnaires were screened according to prespecified data-quality criteria, including completion time and internal response consistency. Questionnaires with implausibly short completion times or internally inconsistent responses were excluded before analysis. Data analysis All analyses were conducted using R version 4.3.2 and IBM SPSS Statistics version 26.0. All tests were two-tailed, and p < .05 was considered statistically significant. Sparse categories were collapsed before analysis: doctoral degrees and master’s degrees were combined, and professional titles were categorized as junior or below, intermediate, and senior. Age values outside 18–70 years and years-of-work values outside 0–60 years were treated as probable data-entry errors. Because the original entries could not be independently verified, implausible values were corrected using a deterministic consistency rule based on the participant’s other plausible value and the education-specific median workforce-entry age. In total, 59 records were corrected, including 51 age values and eight years-of-work values. All retained questionnaires contained complete data for the analytic variables because these items were mandatory for submission; therefore, no statistical imputation was required. Categorical variables are presented as frequencies and percentages, and continuous variables as means and standard deviations. Group differences were examined using Welch’s independent-samples t test for characteristics with two groups and Welch’s one-way analysis of variance for characteristics with three or more groups. Spearman’s rank correlations were used to assess bivariate associations among age, years of work, the weighted PVI, legal knowledge, and legislative support. For the self-developed instruments, content validity was assessed using item- and scale-level content validity indices. Internal consistency was evaluated using the Kuder–Richardson Formula 20 for the legal knowledge instrument and Cronbach’s α for the legislative support instrument. Principal component analysis was used to assess the suitability of total scores. PROCESS Model 4 was used to estimate the indirect association of workplace violence with explicit legislative support through nursing-related legal knowledge. Because all three variables were measured concurrently, the specified ordering was treated as a theory-informed exploratory framework, and the model was interpreted as an associational decomposition rather than evidence of temporal or causal mediation. The primary models were adjusted for gender, education, professional title, employment status, department, and age. Because age and years of work were strongly correlated, multicollinearity was assessed using variance inflation factors and tolerance statistics; age was retained in the primary model, while sensitivity models replaced age with years of work or included both variables. Indirect associations were estimated using 5,000 bootstrap resamples. Exploratory subgroup models were estimated separately for high-risk and low-risk departments using the same covariates except department; these analyses did not formally test between-group differences. Sensitivity analyses repeated the primary model using an unweighted PVI, a simplified two-level weighted PVI, and an alternative legislative-support score in which “do not support,” “unclear,” and “support” were coded 0, 1, and 2, respectively. To address within-hospital dependence, the respondent-level hospital identifier was linked to the analytic dataset using the unique questionnaire sequence number. Hospital-level intraclass correlation coefficients were calculated, and the primary models were re-estimated with hospital fixed effects and hospital-clustered robust standard errors using a finite-cluster correction. The indirect association was additionally evaluated using 5,000 hospital-level cluster bootstrap resamples. Results Participant characteristics A total of 4,391 nurses were included in the analysis. As shown in Table 1, most participants were female (95.4%), held a college diploma (73.0%) or bachelor’s degree (22.5%), and had junior professional titles or below (70.1%). The largest age group was 31–40 years (44.2%), and 41.2% of participants had 6–15 years of clinical experience. Overall, 1,356 participants (30.9%) worked in high-risk departments. Workplace violence was common in the sample. Mean weighted PVI scores were similar between low-risk and high-risk departments, although the unadjusted mean was numerically lower in the high-risk group (7.5 vs 7.7, p = .363). Based on the eight predefined WPV types, 729 nurses (16.6%) reported no WPV exposure, 562 (12.8%) reported one type of WPV, and 3,100 (70.6%) reported two or more types of WPV, indicating poly-victimization. The weighted PVI ranged from 0 to 22, with a mean score of 7.62 and a median score of 7.00. Table 1Participant characteristics and comparisons of study variables (n = 4,391).CharacteristicsN (%)Weighted PVI, Mean (SD)p valueNursing-related legal knowledge, Mean (SD)p valueExplicit legislative support, Mean (SD)p valueGender0.002<.0010.039Female4,187 (95.4)7.5 (5.8)18.5 (4.9)12.3 (3.1)Male204 (4.6)9.0 (6.7)15.1 (5.8)11.7 (4.3)Age<.001<.0010.005≤30 years1,201 (27.4)6.9 (6.3)17.6 (5.3)12.1 (3.5)31–40 years1,940 (44.2)7.6 (5.8)18.5 (4.8)12.3 (3.2)≥41 years1,250 (28.5)8.3 (5.5)18.9 (4.8)12.5 (2.8)Education0.132<.001<.001Master’s degree or above83 (1.9)8.0 (6.2)16.3 (5.3)11.2 (3.8)Bachelor’s degree986 (22.5)7.2 (6.3)17.4 (5.2)12.1 (3.4)College diploma3,202 (72.9)7.7 (5.7)18.7 (4.9)12.4 (3.1)Secondary school120 (2.7)7.5 (5.6)19.5 (4.7)12.8 (2.8)Professional title<.001<.001<.001Junior or below3,075 (70.0)7.4 (6.0)18.0 (5.1)12.1 (3.4)Intermediate1,200 (27.3)8.3 (5.4)19.3 (4.7)12.7 (2.6)Senior, associate or full116 (2.6)8.2 (5.5)19.7 (4.5)12.8 (2.6)Years of work<.001<.0010.057≤5 years1,085 (24.7)6.8 (6.3)17.6 (5.2)12.1 (3.4)6–15 years1,809 (41.2)7.5 (5.8)18.5 (4.9)12.3 (3.2)≥16 years1,497 (34.1)8.3 (5.6)18.8 (4.8)12.4 (3.0)Employment status<.0010.0050.019Formally established post2,506 (57.1)8.0 (5.7)18.5 (4.9)12.4 (3.1)Non-established post1,885 (42.9)7.1 (6.1)18.1 (5.1)12.2 (3.4)Department risk level0.363<.0010.071Low-risk departments3,035 (69.1)7.7 (5.7)18.6 (4.8)12.4 (3.1)High-risk departments1,356 (30.9)7.5 (6.2)17.8 (5.3)12.2 (3.4)Note: Categorical characteristics are presented as N (%), and study variables as mean (SD). Exact two-sided p values are reported. Welch’s independent-samples t test was used for characteristics with two groups, and Welch’s one-way analysis of variance was used for characteristics with three or more groups. PVI = poly-victimization index. Participant characteristics and comparisons of study variables (n = 4,391). Note: Categorical characteristics are presented as N (%), and study variables as mean (SD). Exact two-sided p values are reported. Welch’s independent-samples t test was used for characteristics with two groups, and Welch’s one-way analysis of variance was used for characteristics with three or more groups. PVI = poly-victimization index. Bivariate correlations among core variables Workplace violence was weakly negatively correlated with legal knowledge (r = -0.036, p = .017), whereas legal knowledge was positively correlated with explicit legislative support (r = 0.131, p < .001). Workplace violence and explicit legislative support were essentially uncorrelated (r = -0.001, p = .947). Following correction of 59 implausible age or years-of-work entries, age and years of work remained strongly correlated (Spearman’s ρ = 0.929, p < .001). When both variables were included in the same adjusted model, the variance inflation factors were 9.817 for age and 9.295 for years of work, with corresponding tolerance values of 0.102 and 0.108. Age was therefore retained in the primary adjusted model, while alternative specifications were examined in sensitivity analyses. The correlation matrix is shown in Supplementary Fig. S1. Indirect-association analysis Although the unadjusted Spearman correlation between workplace violence and legislative support was close to zero, the covariate-adjusted standardized regression coefficient was small and negative. As shown in Fig. 1 and Table 2, the total association between workplace violence and explicit legislative support was β = -0.046 (SE = 0.015, p = .002, 95% CI [−0.076, −0.016]). After legal knowledge was included in the model, the direct association was attenuated and was not statistically significant (β = -0.025, SE = 0.015, p = .088, 95% CI [−0.054, 0.004]). Workplace violence was weakly associated with lower legal knowledge (path a: β = −0.077, SE = 0.015, p < .001), while legal knowledge was positively associated with explicit legislative support after adjustment for workplace violence (path b: β = 0.275, SE = 0.015, p < .001). The adjusted models explained 6.4% of the variance in nursing-related legal knowledge and 8.9% of the variance in explicit legislative support. Complete regression parameters are provided in Supplementary Table S4. Hypothesized indirect-association model involving workplace violence, nursing-related legal knowledge, and legislative support. Note: Values are standardized regression coefficients from the primary covariate-adjusted model. The model was adjusted for department, professional title, education, gender, employment status, and age. The indirect statistical association was β = −0.021, with a 95% bootstrap confidence interval of [−0.030, −0.012]. The coefficients represent cross-sectional associations and do not establish temporal or causal mediation. PVI = poly-victimization index. Table 2Total, direct, and indirect effects of workplace violence on legislative support (n = 4391).PathβSEtp value95% CITotal association (c)−0.0460.015−3.0360.002[−0.076, −0.016]Direct association (c′)−0.0250.015−1.7070.088[−0.054, 0.004]Indirect association (ab)−0.0210.004——[−0.030, −0.012]Path a (PVI → Knowledge)−0.0770.015−5.158<.001[−0.106, −0.048]Path b (Knowledge → Support)0.2750.01518.447<.001[0.246, 0.304]Note: β = standardized regression coefficient; SE = standard error; c = total cross-sectional association of workplace violence with explicit legislative support; c′ = direct cross-sectional association after adjustment for nursing-related legal knowledge; a = association of workplace violence with legal knowledge; b = association of legal knowledge with explicit legislative support after adjustment for workplace violence; ab = indirect statistical association estimated as a × b; CI = confidence interval; PVI = poly-victimization index. Total, direct, and indirect effects of workplace violence on legislative support (n = 4391). Note: β = standardized regression coefficient; SE = standard error; c = total cross-sectional association of workplace violence with explicit legislative support; c′ = direct cross-sectional association after adjustment for nursing-related legal knowledge; a = association of workplace violence with legal knowledge; b = association of legal knowledge with explicit legislative support after adjustment for workplace violence; ab = indirect statistical association estimated as a × b; CI = confidence interval; PVI = poly-victimization index. Exploratory subgroup analysis across clinical environments To explore potential heterogeneity across clinical environments, the indirect-association model was re-estimated separately in low-risk and high-risk departments. Low-risk departments included general clinical departments, whereas high-risk departments included emergency departments, ICU, operating rooms, and pediatrics. Full regression parameters are provided in Supplementary Tables S5 & S6. A summary of the core associations is presented in Table 3 and Fig. 2. Small indirect associations through nursing-related legal knowledge were observed in both department groups. In low-risk departments, the indirect effect was β = −0.019, with a 95% bootstrap CI of [−0.029, −0.010]. In high-risk departments, the indirect effect was β = −0.017, with a 95% bootstrap CI of [−0.036, −0.001]. Table 3Exploratory indirect-association models by department risk level.PathDirectionLow-Risk, n=3035High-Risk, n=1356Path aWorkplace violence → Legal knowledge−0.074, p < .001−0.054, p = .042Path bLegal knowledge → Legislative support0.256, p < .0010.318, p < .001Direct c'Workplace violence → Legislative support0.004, p = .813−0.083, p = .001Indirect abWorkplace violence → Legal knowledge → Legislative support−0.019, 95% CI [−0.029, −0.010]−0.017, 95% CI [−0.036, −0.001]Note: Standardized coefficients are presented. Models were estimated separately by department risk level and adjusted for professional title, education, gender, employment status, and age. Bootstrap confidence intervals were based on 5,000 resamples. These subgroup analyses were exploratory and did not constitute formal tests of between-group differences. CI = confidence interval. Exploratory indirect-association models by department risk level. Note: Standardized coefficients are presented. Models were estimated separately by department risk level and adjusted for professional title, education, gender, employment status, and age. Bootstrap confidence intervals were based on 5,000 resamples. These subgroup analyses were exploratory and did not constitute formal tests of between-group differences. CI = confidence interval. Exploratory subgroup indirect-association models by department risk level. Note: Values are standardized regression coefficients from separately estimated subgroup models. Models were adjusted for gender, education, professional title, employment status, and age. The subgroup analyses were exploratory and did not formally test between-group differences. Sensitivity analyses Sensitivity analyses produced broadly consistent findings across alternative model specifications. Replacing the four-level weighted PVI with either an unweighted PVI or a simplified two-level weighted PVI yielded similarly small indirect associations through nursing-related legal knowledge, with bootstrap confidence intervals excluding zero, while the corresponding direct effects remained non-significant (Supplementary Table S7 and S8). The results were also stable when age was replaced by years of work or when both variables were included simultaneously (Supplementary Table S9). Hospital-level ICCs were low, ranging from 0.002 to 0.028. After hospital fixed effects and hospital-clustered robust standard errors were applied, the core coefficient estimates remained similar to those in the primary model. The hospital-level cluster-bootstrap estimate of the indirect association was β = −0.021 (95% CI [−0.031, −0.011]), although the small total association was no longer statistically significant under clustered standard errors (p = .109; Supplementary Table S10). “Unclear” responses accounted for 6.2%–14.6% of responses across the 14 legislative propositions. Treating “unclear” as intermediate between “do not support” and “support” produced a similar indirect association (β = −0.022, 95% bootstrap CI [−0.031, −0.013]; Supplementary Tables S11 and S12). Discussion This study examined the concurrent associations among workplace violence, nursing-related legal knowledge, and explicit support for nursing legislation. Three findings were identified. First, workplace violence was weakly associated with lower legal knowledge, whereas legal knowledge was more clearly associated with stronger legislative support. Second, the adjusted model identified a small indirect statistical association through legal knowledge, although workplace violence and legislative support were not correlated at the bivariate level. Third, exploratory subgroup analyses showed different descriptive patterns across clinical environments. Because the study was cross-sectional, these findings do not establish that workplace violence preceded changes in legal knowledge or legislative support (Folkman et al., 1986). The observed associations may be considered in relation to the cognitive mediational theory of stress and coping, which proposes that responses to stressful experiences depend partly on appraisal and available resources. Within this framework, nursing-related legal knowledge may be relevant to how nurses understand occupational risks, professional rights, and available statutory protections. In this study, greater workplace violence exposure was weakly associated with lower legal knowledge, whereas higher legal knowledge was positively associated with stronger explicit support for nursing legislation. This pattern is broadly consistent with evidence linking workplace violence to adverse psychological and occupational outcomes among nurses, including distress (Lanctôt and Guay, 2014), burnout (Lambert et al., 2024), and turnover intention (Makarem et al., 2025). It also accords with studies showing that legal knowledge and law-related consciousness are associated with professional judgment, practice-related attitudes, and awareness of legal obligations among nurses and other health professionals (Willmott et al., 2020; Yassine et al., 2021). Qualitative evidence from China has similarly identified perceived gaps in legal clarity, legal education, and legislative support in nursing practice (Wang et al., 2025). However, the observed pattern does not confirm the proposed temporal ordering. Nurses with greater pre-existing legal knowledge may be more likely to recognize or report workplace events as violence and may simultaneously express stronger support for legislative protection (Kimura et al., 2023; Lim and Yi, 2021). Reverse, reciprocal, and common-cause explanations therefore remain plausible. Workplace violence and legislative support were essentially uncorrelated in the bivariate analysis, whereas the covariate-adjusted total effect was small and negative. These estimates are not directly equivalent because the former was an unadjusted Spearman correlation and the latter was obtained from a standardized regression model controlling for demographic and occupational characteristics. The difference may therefore reflect covariate adjustment and the different estimands rather than the emergence of a strong underlying association. The indirect estimate through legal knowledge was also small. Accordingly, the findings should not be interpreted as showing that workplace violence meaningfully determines legislative support. Instead, they indicate a limited conditional association that warrants confirmation in longitudinal studies (Arnetz et al., 2025; Duan et al., 2019; Havaei, 2021). The subgroup findings further suggest that clinical environment may be relevant to these associations. The indirect association through legal knowledge was observed in both low-risk and high-risk departments, although the effect was small in both groups. However, the component associations showed different descriptive patterns. The association between legal knowledge and legislative support was numerically stronger in high-risk departments, and the direct association between workplace violence and legislative support was negative only in this subgroup (Stafford et al., 2022; Vrablik et al., 2019). These results should not be interpreted as formal evidence that department risk level moderated the indirect association, because the subgroup analyses were based on separate models rather than interaction tests. They instead indicate descriptive heterogeneity that requires confirmation using formal interaction or moderated-mediation analyses (Liu et al., 2022). The negative direct association observed in high-risk departments warrants cautious interpretation. One possible explanation is that repeated or severe workplace violence in high-risk settings may be accompanied by emotional exhaustion, reduced organizational attachment, or lower confidence in institutional responsiveness, which may weaken explicit support for legal or policy solutions even after legal knowledge is considered (Copeland and Henry, 2017; Huang et al., 2019). Another possibility is that unmeasured organizational or psychological factors, such as emotional exhaustion, organizational cynicism, or perceived ineffectiveness of formal reporting and protection systems, may partly account for this residual association. Measurement issues may also be relevant, especially because legislative support was assessed as explicit endorsement and “unclear” responses were not treated as support in the primary analysis. Although the alternative coding of “unclear” responses produced similar estimates, uncertainty may still reflect heterogeneous attitudes, including insufficient knowledge, ambivalence, or reluctance to express a position. Therefore, this finding should be viewed as hypothesis-generating rather than confirmatory. Another notable finding was that nurses in high-risk departments had lower mean nursing-related legal knowledge than those in low-risk departments. This result is not entirely intuitive, because nurses in more hazardous environments might be expected to have greater awareness of legal protection. However, exposure to acute workload, time pressure, and frequent interpersonal conflict may not necessarily translate into stronger legal knowledge (Potter et al., 2005). Instead, high clinical demand may reduce opportunities for structured legal learning or engagement with regulatory content. This interpretation is consistent with studies emphasizing that legal knowledge requires deliberate education and reinforcement rather than being an incidental by-product of clinical exposure (Lee et al., 2017; Yassine et al., 2021). However, because department groups may differ in multiple unmeasured characteristics, this result should not be interpreted as evidence that high-risk work reduces legal knowledge. It instead supports further examination of access to legal education and institutional guidance across clinical settings (Wang et al., 2024). Several implications may be considered from these findings. First, legal education for nurses should not be limited to abstract legal knowledge, but should be connected to workplace violence, professional rights, documentation, reporting, and institutional protection procedures. Second, legal education should be integrated into broader occupational safety strategies that include violence prevention, accessible reporting systems, managerial accountability, psychological support, and institutional protection. Third, frontline nurses may also be engaged in the development and communication of nursing-related legislation as professional stakeholders rather than solely as recipients of protection. Given the small effect estimates, these implications should not be interpreted as evidence that legal education alone would meaningfully change responses to workplace violence or legislative attitudes. Although our study focused on hospital nurses, the implications of strengthening occupational protection may extend across different nursing practice settings. Recent international multicentre evidence suggests that workplace violence is also prevalent among nurses working in community and home-care services, where its determinants may differ from those observed in hospital environments. Organizational factors, including workload, burnout, staffing adequacy, and managerial support, appear to play an important role in shaping violence risk across care settings (Cesare et al., 2026). Therefore, regulatory and legislative approaches to protecting nurses should consider the diverse contexts in which nursing care is delivered rather than focusing exclusively on hospital-based practice. However, whether the associations observed in the present study apply to non-hospital settings requires further investigation. Several limitations should be considered. First, the cross-sectional design and concurrent measurement preclude temporal or causal interpretation; reverse, reciprocal, and common-cause explanations remain possible. Second, WPV was self-reported without a fixed recall period or restriction to the current hospital. Therefore, participants may have considered different exposure windows when responding, with some recalling recent incidents and others reporting experiences accumulated over longer periods of clinical practice. This variability may have reduced the comparability of workplace violence exposure across participants and introduced potential recall-related measurement error. Third, recruitment through administrative networks and mandatory item completion without a “prefer not to answer” option may have affected perceived voluntariness, selection, or response quality. Fourth, although hospital fixed-effects and cluster-robust analyses produced similar coefficient estimates, the study included only 13 hospitals from one metropolitan area, limiting generalizability and precise estimation of hospital-level variance estimation. Although the observed intraclass correlation coefficients were low, hospital-level variance estimates should still be interpreted cautiously given the limited number of participating institutions. Fifth, a small number of implausible age and years-of-work values were corrected using a deterministic consistency rule because the original entries could not be independently verified. Age and years of work also remained strongly correlated; although age was retained in the primary model and alternative specifications produced similar estimates, some measurement and model uncertainty may remain. Sixth, the legal-knowledge instrument was not strongly unidimensional, and the legislative-support score combined several policy domains and reflected stated attitudes rather than actual policy engagement or advocacy behavior. Finally, residual confounding from prior legal training, organizational culture, institutional support, managerial practices, and reporting climate cannot be excluded. Conclusion Nursing-related legal knowledge was associated with stronger explicit support for nursing legislation and contributed to a small indirect association between workplace violence and legislative support. However, the small effect estimates, cross-sectional design, and measurement limitations preclude causal or strong policy conclusions. Longitudinal studies across broader healthcare settings are needed to clarify the temporal ordering and practical relevance of these associations. Ethics approval and consent to participate This study was approved by the Institutional Review Board of the School of Public Health and Nursing, Shanghai Jiao Tong University (No. SJUPN-202017). All methods were performed in accordance with relevant guidelines and regulations. The first page of the online questionnaire provided information about the study purpose, voluntary participation, confidentiality, and the right to withdraw. Electronic informed consent was obtained from all participants before questionnaire completion. This study was approved by the relevant institutional review board (approval number blinded for review multiple time given inn source we are added singal heading or added apara pls check and correct asper msp"?>). Availability of data and materials The datasets generated and analyzed during the current study are not publicly available because they contain sensitive occupational information and are subject to ethical restrictions related to participant privacy. Competing interests The authors declare that they have no competing interests. Ethics declaration Written informed consent to take part in the study and to publish the article has been obtained from all participants or their legal representatives. The privacy rights of participants have been observed. This study was performed in compliance with relevant laws, regulatory frameworks and guidelines where the research took place. This study was approved by the Institutional Review Board of the School of Public Health and Nursing, Shanghai Jiao Tong University. (Approval No. SJUPN-202017) Funding This work was supported by the National Natural Science Foundation of China (No. 71904123). The funder had no role in the study design, data collection, data analysis, interpretation of data, writing of the manuscript, or decision to submit the manuscript for publication. CRediT authorship contribution statement Xiaolong Wang: Writing – review & editing, Writing – original draft, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Yufei Xing: Writing – review & editing, Writing – original draft, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Yinghui Wu: Writing – review & editing, Supervision, Methodology, Conceptualization. Qiong Fang: Writing – review & editing, Supervision, Project administration, Methodology, Conceptualization. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements We extend our deepest gratitude to all the frontline clinical nurses who participated in this study. Their willingness to share their invaluable perspectives and experiences, despite demanding clinical schedules, made this research possible. We are also sincerely thankful to the nursing directors and administrators across the participating hospitals for their invaluable support and coordination during the data collection process. Footnotes Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.ijnsa.2026.100668. Contributor Information Yinghui Wu, Email: yinghuiwu@shsmu.edu.cn. Qiong Fang, Email: jonesfang@sjtu.edu.cn. Supplementary materials References Associated Data Supplementary Materials Data Availability Statement The datasets generated and analyzed during the current study are not publicly available because they contain sensitive occupational information and are subject to ethical restrictions related to participant privacy.