Factors Associated with Non-Suicidal Self-Injury Among Middle and High School Students in China Using a Random Forest Classifier

Jiaqi Zhou, Mengyao Wang, Jiangboheng Shi, Xilong Cui*, Jingbo Gong*

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

Objectives: Previous research has linked NSSI to various psychological, social, and environmental factors, but the relative importance of these factors across different developmental stages remains unclear. This study investigates the multidimensional factors associated with NSSI among middle and high school students in China using a Random Forest Classifier to inform targeted interventions. Methods: A total of 2735 Chinese students provided self-reported data on NSSI behavior, functions of NSSI behavior, as well as a number of associated factors. The study utilized a Random Forest Classifier to identify key factors related to NSSI across different time periods (last month, past 6 months, and past year), and compared these factors between middle school and high school student cohorts. Results: The Random Forest models showed good discriminative ability for identifying NSSI (AUC = 0.84 for middle school; AUC = 0.79 for high school students), with high specificity (94.7-97.3%) but low sensitivity (27.6-42.8%). For middle school students, depressive and anxiety symptoms, social support, negative self-evaluation, and self-control were primary factors. For high school students, depression, anxiety, schemas related to shame, emotional abuse, and executive functioning deficits were found primary factors. Within the subgroup who reported NSSI, differing motivations and factors were observed across developmental stages. Conclusions: NSSI behavior in Chinese adolescents is influenced by a complex interplay of emotional, psychological, social, and cognitive factors, with clear developmental distinctions between middle and high school students. These findings underscore the need for tailored prevention and intervention strategies that are sensitive to the developmental stage of the adolescent population.

Original languageEnglish
JournalArchives of Suicide Research
DOIs
StateAccepted/In press - 2025

Keywords

  • Adolescents
  • developmental differences
  • non-suicidal self-injury
  • random forest classifier

Fingerprint

Dive into the research topics of 'Factors Associated with Non-Suicidal Self-Injury Among Middle and High School Students in China Using a Random Forest Classifier'. Together they form a unique fingerprint.

Cite this