Future Signals of Adolescent Delinquency Using Social Big Data: Focusing on the Causes and Types of Adolescent Delinquency in South Korea

Background/Purpose:

This study aims to detect keywords, as well as analyze current and future trends in adolescent delinquency, with the ultimate goal of suggesting preemptive measures to address this social issue. Social big data is a valuable source of information for understanding the latest developments and future trends in adolescent delinquency, and further can be used for developing countermeasures and policies. Despite its potential benefits, the collection and analysis of social big data present difficulties that have limited its wide use in previous research. To overcome these challenges, this study used social big data to detect relevant keywords and gain a better understanding of the current and evolving nature of adolescent delinquency in South Korea.

Methods:

This study collected social big data on terms related to adolescent delinquency. From January 1, 2018, to June 30, 2019, a total of 1,144,616 documents were collected from 127 news sites, social media networks, and online forums. Based on a pre-developed ontology and thematic analysis, twenty-five delinquency-cause keywords (e.g., family relationship strain, attachment) and thirty-six delinquency-type keywords (e.g., violence, drug use) were classified. A future signal approach was examined to verify current and future trends in adolescent delinquency issues. Keywords were categorized based on average term frequency, average document frequency, and their rate of increase: Latent signal (low frequency, low rate of increase), Weak signal (low frequency, high rate of increase), Strong signal (high frequency, high rate of increases), Strong but low increasing signal. Social trends progress from Latent signal, Weak signal, Strong signal, to Strong but low increasing signal. Next, association rules analysis was conducted to uncover patterns among delinquency-related keywords.

Results:

Keywords were classified into four signals. Keywords such as violence, school academic strain, low self-control, and belief were identified as Strong signals or Strong but low increasing signals, indicating that they are currently the subject of social interest. Drug use, theft, gambling, and self-control were identified as Weak signals, which have the potential of transforming into Strong signals over time. Association rules analysis revealed that adolescent drug use and theft co-appear with keywords such as media and low social status. Finally, running away from home, smoking, drinking, and school relational strain were categorized as Latent signals.

Conclusions and Implications:

Study results suggest that violence is the most significant issue concerning adolescent delinquency in South Korea. From 2018 to 2019, school violence and cyber violence were considered crucial issues. In 2023, South Korea is confronted with an emergent adolescent drug problem, which validates the progression of past weak signals into future strong signals. This study also highlights the significant role of media in shaping the issue of drug use among adolescents. Lastly, although there is a relatively lower social interest in latent signals such as teenage pregnancy and sexual exploitation, these keywords may evolve into weak signals over time. These findings underscore the importance of employing the future signal approach to better understand trends in adolescent delinquency and develop proactive approaches for intervention.