Developmental trajectories of firearm carrying among Justice-involved youth: effects of socioeconomic status and gang membership

ABSTRACT

There is limited understanding of how socioeconomic status, gang membership, and their interaction predicts the development of firearm carrying. The Pathways to Desistance data were analyzed. Group-based trajectory modeling identified heterogeneity in patterns of firearm carrying. Multinomial logistic regression examined socioeconomic status and gang membership as predictive of development. A seven-group firearm carrying model best fit the data. Gang membership predicted increased risk of assignment to the Low Chronic, Adolescence Limited, and High Chronic groups. Socioeconomic status did not exert a significant direct effect on group assignment. The interaction between socioeconomic status and gang membership was significant and negative for the High Chronic group. Gang members with higher socioeconomic status presented the highest risk of high chronic firearm carrying.

KEYWORDS:

• Firearm carrying
• Justice-involved youth
• group-based trajectory modeling
• gangs
• socioeconomic status

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. Stata/MP 16.1 had issues estimating the unaltered gun carrying frequency variable for trajectory analyses. This appeared to be because of the wide range of values between the lowest and highest scores and high outliers on this variable. This then necessitated a recoding of the variable so that the software was able to analyze the data. This study took the approach described in Wojciechowski’s (2020b) research to address this issue. This entailed a top-coding of the variable the highest frequency count possible in order to maintain as much of the original variance on the outcome as possible while also reducing the range between the highest and lowest scores so that the data could be analyzed. As such, 27 was identified as the ceiling category for this variable and all scores above this were recoded as 27 at each wave. This was the highest frequency score at which analyses were possible to be carried out.