Methods. Using 2010 Taiwan Children and Youth Living Conditions Survey, a latent class analysis (LCA) was performed with a sample of 1,324 adolescents aged between 12 and 18. Six measures of delinquent behaviors—using drugs, stealing, running away from home, gangster involvement, skipping class/school, and bullying—were used as indicators to identify the delinquent patterns. Multi-group LCA was used to examine the possible gender and age differences. A regression model was conducted to investigate the influence of delinquent patterns on well-being when the individual characteristics (gender, age, education level), family SES (parent’s employment status and education level), and perceived structural conditions (e.g., society is trustworthy) were controlled in the model. All the analyses were conducted using SAS 9.3.
Results. One-class to five-class models were tested and the 2-class model, non-committed (91.83%) and crime-committed (8.17%) group, was selected because the fit indices were satisfactory (AIC = 80.06, BIC = 147.49, Entropy = 0.90). For the non-committed group, the probability of engaged in each delinquent behaviors was low (< 5%), with a slightly higher in bullying (8.6%); while the crime-committed group had higher probability in stealing and running away from home (close to 50%) and in skipping class and bullying (around 70%), with lower probability in gangster involvement (31.5%) and using drugs (11.9%). The multi-group LCA showed that the patterns of delinquent behaviors varied by gender and age. Females (7.2%) and late adolescence youth (2.70%) were less likely to commit crimes compared to their counterparts (males: 10.4%; early adolescence youth: 11.92%), but the probability of crime involvement varied a great deal. Females were more likely to run away from home or skip class, but males were more likely to steal, skip class, or bully. Early adolescence youth had higher probability of skipping class and bullying, but the late adolescence youth were more likely to involve in crime, such as using drugs, stealing, and skipping class. Regression model showed that, controlling for covariates, crime-committed group had worse well-being outcome compared to non-committed group (b = −0.36, p < .001).
Conclusions and Implications. This study suggests that two clusters of delinquent behaviors could be observed among Taiwanese adolescents. Further, delinquent patterns vary by gender and age. Among crime-committed group, prevention efforts should be paid attention to males and late adolescence youth because they are more likely to engage in severe delinquent behaviors, which in turn may deteriorate their well-being.