Risk Patterning for Adolescent Weapon Carrying: A Mixture Model Approach

BACKGROUND: Despite downward trends in crime in the United States, weapon violence, particularly gun violence, continues to be a substantial problem among youth. Firearms are now the leading cause of death among youth (CDC, 2021), and over 13% of high school students reported carrying a weapon in the past month (CDC, 2020). Many potential risk factors have been considered (Shetgiri et al., 2016), and perhaps most publicly mental health concerns, but research until this point has largely examined these as linear associations. This approach may miss important subgroups of youth whose risk profiles relate to differential rates of weapon carrying. The present study aims to provide a more nuanced understanding of which youth may be highest risk for weapon carrying by examining the patterning of established risk factors among the general youth population. Findings hold implications for tailored violence prevention efforts.

METHODS: Data come from the 2018 Washington State Healthy Youth Survey, a state representative behavioral health survey of students in the 8^th, 10^th, and 12^th grades (n=11,057). Latent class analyses identified distinct subgroups of youth using 10 risk factors associated with weapon carrying: internalizing symptoms, suicidality, maltreatment history, peer victimization, hopelessness, emotional or learning disability, history of fighting, history of dating violence, gang membership, and lack of adult support. Models were compared using relative fit indices (likelihood-based indices, VLMR Likelihood Ratio Tests), interpretation, and parsimony (Collins & Lanza, 2010) and adjusted for school-level clustering. Chi-squared tests examined overall and between-class differences in the prevalence of weapon carrying, accounting for classification error between latent classes.

RESULTS: Approximately 3.4% of youth reported past-year weapons carrying, which was more prominent among males (4.8%) than females (1.9%). Initial results indicate that the best-fitting and most parsimonious model was a 4-class solution. Classes were identified as Low Behavioral Health Risk (60.9%; low likelihood of each risk factor), Emotional Health Risk (16.4%; high likelihood of emotional health risk and slightly elevated maltreatment risk), Maltreatment/Aggression Risk (13.8%; elevated maltreatment and violence risks), and High Behavioral Health Risk (8.9%; high likelihood of each risk factor). Prevalence of weapon carrying differed significantly across classes: 1.1% for the Low Behavioral Health Risk class, 0.4% for the Emotional Health Risk class, 9.1% for the Maltreatment/Aggression Risk class, and 12.4% for the High Behavioral Health Risk class.

DISCUSSION: Despite public concerns regarding mental health problems and weapons use, our findings suggest a more nuanced explanation of these associations. The class of youth experiencing emotional health risk alone had the lowest prevalence of weapon carrying, while the greatest observed prevalence was among classes of youth with pronounced maltreatment, peer victimization, and fighting histories, both with and without mental health symptoms. This may align with prior research suggesting that victimized youth may be motivated to carry weapons as a form of self-protection (Lewis et al., 2007). Social work researchers and practitioners should continue to advocate for a more nuanced understanding of the association between mental health and weapons while attending to the needs of youth exposed to maltreatment and other forms of violence.