Network analysis of moral courage, moral resilience, and moral sensitivity among nursing students: a cross sectional survey.

Background: Nursing students frequently encounter ethically challenging situations during clinical internships. Although moral sensitivity, moral courage, and moral resilience are interrelated moral attributes, their component-level conditional associations remain insufficiently understood. Objectives: To examine the component-level network structure of moral sensitivity, moral courage, and moral resilience among nursing students, identify components with the highest expected influence, and evaluate network accuracy and stability. Design: A cross-sectional survey. Setting: A medical university in southeastern China and its affiliated clinical internship settings. Participants: A total of 302 fourth-year undergraduate nursing students undertaking clinical internships participated in the study. Their mean age was 21.0 years (SD = 1.6), and 262 participants (86.75%) were female. Methods: Moral courage, moral sensitivity, and moral resilience were assessed using the Nurses' Moral Courage Scale, the Moral Sensitivity Questionnaire for Nursing Students, and the Chinese version of the Rushton Moral Resilience Scale, respectively. Eleven subscale scores were included as nodes in a regularized Gaussian graphical model estimated using graphical LASSO with Extended Bayesian Information Criterion model selection. Expected influence was used to assess node centrality. Network accuracy and stability were examined using non-parametric and case-dropping bootstrapping with 1,000 iterations. Results: The strongest edge in the network was observed between coping with moral adversity and relationship integrity (r = 0.506). Within moral courage, the compassion and true presence component was strongly associated with moral responsibility (r = 0.453). Moral responsibility within moral courage had the highest expected influence (z = 1.418), whereas moral responsibility within moral sensitivity had the lowest expected influence (z = -1.329). The network showed strong stability, with correlation stability coefficients of 0.672 for expected influence and 0.748 for edge weights. Conclusions: Action-oriented moral responsibility occupied a central structural position in the estimated network, whereas awareness-oriented moral responsibility was less strongly connected. The findings identify component-level associations that may inform future longitudinal and educational research, but they should not be interpreted as evidence of causal or directional relationships.