High Utilizers of the Albemarle and Charlottesville Criminal Justice System; Understanding the Key Factors Influencing Young TikTok Users’ Addictive Usage Patterns Despite Algorithmic Awareness of Data-Driven Content Personalization

Larger societal systems and entities play a huge role in directing and deciding the mental health and wellbeing of society’s most vulnerable populations. Such populations often need vital services from society in order to heal, to be able to offer their best selves as positive contributors to society. Examples in today’s world include “high utilizer” prisoners who keep reoffending and are unable to fully reintegrate into society, social media users struggling with addiction and mental health, and historically marginalized socioeconomic groups that lack access to community services that they could benefit from. However, oftentimes, larger entities misallocate their resources, or even abuse their power, adversely affecting all those who use and rely on their services. Specifically in terms of high utilizers, every year nearly 428,000 people across the US recidivate, or reoffend, at least 3 times, cycling through law enforcement, courts, emergency departments, and other community services – demonstrating the magnitude of the issue (Prison Policy Initiative, 2019). In terms of social media addiction, an estimated 5-10% of Americans struggle with this, over 30 million people – and more and more of these are of the younger generation, on platforms driven by tech companies intentionally pushing addictive algorithms (Ricci, 2018). With the wellbeing of such large numbers of people suffering from a lack of vital services or digital addiction, it’s critical to understand how larger societal systems and entities can meet these needs, especially for the vulnerable populations that need them most. How can larger societal systems and entities more effectively use their power and resources to benefit society’s most vulnerable populations?

Like many communities across the United States, Charlottesville, VA includes a relatively small group of individuals who consume a disproportionate amount of public and private services. These services include criminal justice resources (such as police, courts, jail) and community services (including mental health counseling, EMT trips, hospital emergency room visits, and drug and alcohol rehabilitation). By learning as much as possible about these
“high utilizers,” the team could better inform local decision-makers about potential interventions that might lessen recidivism among them, improve their lives, and reduce costs to the community. High utilizers were defined as individuals with four or more arrests within a single year, and were identified using booking and related data provided by the ACRJ (AlbemarleCharlottesville Regional Jail) spanning a 34-month period. Analysis of this data and interviews with stakeholders resulted in a characteristic understanding of the differences between high utilizers and non-high utilizers, with high utilizers having 4.8 times more bookings per person, a flow chart built to illustrate the complex, multi-agency system that high utilizers step through each time they are booked, and a chart showing that larceny is the most common statute for initial arrests, but contempt of court and probation become more prevalent for subsequent arrests. This suggests a possible lack of transportation means after arrest to return to court, a lack of awareness of court summons, and frequent violations of release conditions, indicating potential points of intervention.

Most youth use social media simply for entertainment and socializing – to keep up with friends, because that is what they see as the norm, or even due to the difficulty of breaking an addictive habit. Still, many are oblivious to its drastic negative effects on their mental health, which often culminate in this addiction. All the while, big social media companies such as TikTok are profiting from this, using targeted, data-driven content personalization to keep their users on their platforms. Therefore, it’s important to understand the inner conflict faced by young
TikTok users’ about changing their usage patterns, despite their general algorithmic awareness of the platform’s content personalization – as identifying which factors (such as social groups, algorithm awareness, or habit psychology) are most influential towards their usage patterns shows where societal efforts should focus towards. A literature review of various algorithm awareness studies was conducted, searching Google Scholar for algorithm awareness studies from the past 5 years, using relevant keywords to narrow the search to a smaller sample size. Then, each study was parsed for its major findings, and a qualitative data analysis of these findings was performed, with each study’s most significant factors being grouped into repeating
“themes”. Social interaction/relationships (referenced in 3 papers, 5 times total), algorithm awareness (referenced in 4 papers, 4 times total), and content personalization (referenced in 3 papers, 3 times total) were the most commonly referenced factors, with many studies revealing a direct tug-of-war between social factors and algorithm awareness. Based on these findings, efforts should direct their resources towards understanding and deconstructing FOMO and other social interaction workings, greater societal algorithm awareness, and helping users to recognize and objectively view personalized content.

Overall, both projects were fruitful, producing valuable insights. While the technical project produced tangible statistics and areas for further investigation for local decision-makers, the STS project produced key factors that would be the highest impact areas for addiction reduction efforts. For the technical project, future steps could include numerical quantification of the financial and time aspects of the booking process, and an integrated analysis of the rehabilitation services most used by high utilizers. For the STS project, it could include a broader analysis of other relevant factors, such as the mental health symptoms and attention span of users.

School of Engineering and Applied Science

Bachelor of Science in Systems Engineering

Technical Advisor: Loreto Alonzi, Michael Smith, K. White

STS Advisor: Kent Wayland

Technical Team Members: Zakaria Afi, Sarah Bedal, Olivia Bernard, Mohini Gupta, Caroline Lee