scieee AI-readable full text Open interactive document viewer

Social Foraging for Crime: A Simulation Study on Co-Offenders' Specialization

Klymentiev, Ruslan

Publisher
Zenodo
DOI
10.5281/zenodo.17588633
Repository
zenodo.org
Language
en

Abstract

Similar to animals foraging for food, offenders navigate environments in search of rewarding criminal opportunities, weighing risks, rewards, and effort. This foraging perspective offers a useful lens for understanding how offenders specialize in certain crime types and how co-offending may influence this process. While previous studies have explored the specialization of co-offending groups, less is known about how collaboration affects individual specialization over time. To address this gap, this study introduces a formal agent-based model in which offenders move through a spatial environment, encountering diverse opportunities and deciding whether to offend alone or with a partner. Decisions are shaped by skill levels, risk of punishment, and the potential for reward or skill exchange through co-offending. In line with previous theoretical research on co-offending, specifically social exchange theory, we expect to show that criminal collaboration enables offenders to learn new skills faster and enables the commitment of more complicated types of crimes. We aim to provide new insights into the dynamics of co-offending specialization, drawing parallels between offender decision-making and behavioral ecology.

Full text

Social Foraging for Crime A Simulation Study on Co-Offenders' Specialization Ruslan Klymentiev, Paul Jeffrey Brantingham, Rafael Prieto-Curiel, Luis Enrique Correa Rocha & Christophe Vandeviver Research Day Law and Criminology 2025 –Ghent –14/11/2025 Image credit: Amanda Lea Offending behavior can be seen as foraging Johnson, 2014; Vandeviver et al., 2021 Optimal Foraging Theory (OFT) was used to explain the spatio-temporal aspects of offending However, studies are missing the concept of co-offending Image credit: Amanda Lea Co-offending as a means of social exchange Weerman, 2003 •Access to diverse parts of criminal network leads to a greater generalization (McGloin and Piquero, 2010; Klymentiev et al., 2025) •Co-offending allows the commitment of more sophisticated crime types (Felson, 2003; Tremblay, 2017) •Co-offending groups involve offenders with different skills (Bright et al., 2024) Prevalence of co-offenses Klymentiev et al. (2025) Why is co-offense prevalence less than 5% in some environments, while in others it's greater than 50%? Prevalence of co-offenses Why is co-offense prevalence less than 5% in some environments, while in others it's greater than 50%? Prevalence of co-offenses Perhaps in some environments it is: •Hard to find a partner for co-offending (due to availability, not enough trustworthy or skillful partners) •Crimes are relatively easy to execute alone •Offenders do not even want to co-offend Weerman (2003) Research question OUTCOME •Offenders are "foragers" •Co-offending leads to generalization •Co-offending allows the commitment of more sophisticated crime types (Felson, 2003; Tremblay, 2017) •Varying prevalence of cooffending (mean = 30%) (Klymentiev et al., 2025) Preliminary results Scenario 0: Random (everyone can access everyone with no skill or trust preference) Preliminary results Scenario 0: Random (everyone can access everyone with no skill or trust preference) Preliminary results Scenario 0: Random (everyone can access everyone with no skill or trust preference) Preliminary results Scenario 0: Random (everyone can access everyone with no skill or trust preference) Preliminary results Scenario 1: Limited access to partners, moderate skill and trust preference Preliminary results Scenario 1: Limited access to partners, moderate skill and trust preference Preliminary results Scenario 1: Limited access to partners, moderate skill and trust preference Preliminary results Scenario 1: Limited access to partners, moderate skill and trust preference Preliminary results Scenario 1: Limited access to partners, moderate skill and trust preference Preliminary results Scenario 1: Limited access to partners, moderate skill and trust preference