Crime, Choice, and Context

This special issue is the result of a workshop held on 22 October 2021 at the Max Planck Institute for the Study of Crime, Security and Law in Freiburg, Germany. The reason for organizing the workshop was that in spite of its ubiquitous influence on human judgment and decision making, the study of context has not yet reached center stage in research on criminal choice. The workshop addressed this hiatus and set the stage for novel research that revisits the multidisciplinary roots of the study of criminal decision making and expands its rational choice foundations. To this end, participants to the workshop were invited to examine how context shapes criminal decision processes. More specifically, they were encouraged to do so in ways that move beyond the axiomatic construct of rational decision making underlying neoclassical economics to incorporate findings and theoretical perspectives from several decades of research in criminology, behavioral economics, and psychology.


Lessons From Psychology and Behavioral Economics for the Study of Context and Crime
As a point of departure, consider the optical illusion displayed in Figure 1. This illusion by German psychologist Hermann Ebbinghaus  shows two orange circles of exactly the same size. The orange circle on the left, which is surrounded by larger circles, appears smaller even though it is the same size as the circle on the right. The reason the center circle on the right appears larger is because the circles surrounding it, which provide the context, are smaller. This optical illusion, which is known as the Ebbinghaus Illusion, illustrates a broader phenomenoncontext affects perceptions. Research by cognitive psychologists since has demonstrated the ubiquitous influence of context not just on how we perceive things, but also on our memory, how we process information, make judgments, and, ultimately, on how we behave (e.g., Gibson 1979;Kirshner and Whitson 1997;Neisser 2014;Smith, Glenberg, and Bjork 1978).
Violations of the axiomatic assumptions of rational choice due to context are also well documented in the behavioral economics literature (e.g., Kelman, Rottenstreich, and Tversky 1996;Loewenstein 1999). For example, simply framing the same decision problem in terms of either gains or losses may result in preference reversals, both in hypothetical and in real situations (Tversky and Kahneman 1981). The frame provides contextual information that, as several decades of research bear out, predictably affects how the decision problem is evaluated.
The assumptions of rationality also dictate that preferences between different choice alternatives should be unaffected by availability or quality of other options, an assumption known as intransitivity of preferences (Chang, Gershman, and Cikara 2019). That is, if Option A is preferred over Option B, and Option B is preferred over Option C, then Option A should also always be preferred over Option C. Again, research shows this is not always the case. One example of an exception is the decoy effect, where the introduction of an inferior third alternative, the decoy, changes the preferences between the (superior) alternatives of a set (Slaughter, Sinar, and Highhouse 1999).
In their contribution to this special issue, Nguyen, Loughran, and Topalli identify an instance in which the assumptions of rational choice models are violated in the domain of criminal choice. The authors point out that rational models of consumption assume the interchangeability, or fungibility, of money: there should be no difference between proceeds obtained illegally versus proceeds obtained legally in terms of future spending. They question this assumption of interchangeability as it applies to licit and illicit income. Using a mixed-method approach including a quantitative survey of incarcerated offenders and a qualitative, interview-based study of active offenders, the authors address two important questions 1) Do offenders perceive money earned by legal versus illegal income generating activities the same way? 2) How do consumption patterns (spending and saving) differ between legal versus illegal income generating activities? Their findings show differential patterns of legal and illegal money expenditures by offenders, with illegal funds more likely to be spent on partying and conspicuous consumption and legal earnings more often being earmarked for housing and essential bills.
Considered in the light of the current special issue, the results of this study, on the one hand, emphasize how contextual factors even preceding the actual moment of decision making can impact criminal choice, which begs a reconsideration of choice models and their often restricted portrayal of the choice process as a discrete, temporally confined, and one-off decision event. On the other hand, the results testify to the importance of studying the meaning of money in criminology. This topic, as indicated by Nguyen, Loughran, and Topalli, has been curiously absent in criminal decisionmaking research, in spite of money being the primary motivator for acquisitive crimes and the backbone of the illicit economy.

Context as Social Influence
Context is commonly referred to as the physical and spatial properties of the immediate environment. Situational and opportunity perspectives are, for example, premised on the assumption that whether crime takes place or not is contingent on the characteristics of a situation. Factors such as street lighting, neighborhood disorder, and the presence of CCTV have each been shown to affect crime. However, as Barnum and Pogarsky (2022: 536) argue, the immediate context for crime consists not only of environmental factors, people too contribute to it, for example as instigators, co-offenders, bystanders, or peers.
Peer effects are commonly divided into socialization effects in which norms and motives conducive to delinquency develop over longer timeframes, on the one hand, and situational influences that operate in the immediate situation, on the other. Hoeben and Thomas (2019: 759) observe that there is increasing evidence that the presence of peers can affect an actor's decision calculus concerning the anticipated consequences of crime, for example, by lowering expectations of risk or by making individuals more tolerant of such risk. Yet, peer studies often include confounds and establishing the causal effect of a factor has traditionally plagued research in this domain. As Warr (2002) noted in his classic book Companions in Crime, strong associations between the behavior of adolescents and that of their friends by themselves cannot distinguish peer influence from homophily, i.e., a "birds of a feather flocking together" phenomenon.
One way to overcome this challenge to causal inference is by using experimental research designs. Capitalizing on the affordances offered by the controlled conditions of a lab experiment, in his contribution to this special issue, Engel examines the effect of the number of peers violating an arbitrary and unenforced rule on the decision to violate the rule oneself. Drawing from criminology's rich tradition of peer research, Engel both sheds new light on this domain and addresses the confound problem with a novel experiment. In the experiment, context is stripped to the bare minimum to rule out any potential confounds. The controlled environment of the research lab therefore enables the identification of the causal effect of knowing how many peers violate a(n) (arbitrary and unenforced) rule on the propensity to violate this same rule. The experiment finds that the more peers violate the rule, the more participants are likely to do so as well, and the more severe the violation. This main finding replicates in a vignette study that is part of the same contribution. The Engel study thus provides important experimental evidence buttressing a large body of nonexperimental criminological evidence on peer effects.

Morality as a Contextual Cue
Another way in which crime researchers have tried to capture the influence of context is through the use of written scenarios or vignettes. The vignette method was pioneered in the 1970s by social science researchers (e.g., Alexander and Becker 1978;Nosanchuk 1972) who argued that questionnaires and interviews were unsuitable for the study of attitudes and behavior because they tend to elicit biased and unreliable responses. The lack of contextual detail in these methods leads people to respond in terms of their own mental picture of the task before them (Alexander and Becker 1978: 93). The situation in criminology mirrored that of the social sciences. Early perceptual deterrence studies asked respondents to estimate the risk of arrest for specific crimes without reference to circumstances, leading them to impute the circumstances of the crime themselves (Nagin 1998). However, as Barnum, Nagin, and Pogarsky (2021: 216) observe: "because risk perceptions are highly dependent on circumstances, questions about risk without details on context are ill posed." In the late 1980s and early 1990s, deterrence researchers, who had hitherto relied on decontextualized surveys to assess sanction risk, started using vignettes to provide context surrounding (hypothetical) offenses thought to be relevant for people's decisions (e.g., Klepper andNagin 1989, Nagin andPaternoster 1994). By providing such information and locating the choice process within the situation, these studies were able to capture the situation-specific nature of offending decisions (Pogarsky 2004).
Prior work using vignettes to operationalize context has examined a variety of factors, such as sanction severity, probability and fairness, and also looked at emotions (e.g., Van Gelder et al. 2013). In their contribution to this special issue, Herman and Pogarsky take the next step in this research tradition by examining morality as a contextual factor. Specifically, they examine situational moral dynamics in offending decisions, decoupling the more stable individual morality, or moral identity, from situational moral evaluations of crime opportunities, and examine interrelations between the two.
Herman and Pogarsky observe that research and theorizing on criminal choice often conceive of morality as a time-stable individual trait, thus bypassing the possibility that it may, at least in part, also be contextdependent. Using a survey containing randomized experiments administered to a nationwide USA sample of respondents, the authors provide descriptions of different criminal behaviors (property damage, physical assault, and theft), occurring under varying circumstances. The experimental conditions contained additional content in the form of a rationalization or justification for the offense (e.g., denial of injury or responsibility). The results indicate that, in line with expectations, moral evaluations of crime opportunities depend on situational context, and these situational moral evaluations, in turn, influence offending decisions. Furthermore, circumstances conducive to rationalization increased the reported moral acceptability of crime opportunities.
Two things make the research by Herman and Pogarsky of particular interest when considered in the larger context of the study of criminal choice. Aside from adding a novel, i.e., the moral, dimension to vignettebased research on criminal decision making, a key contribution of this work is how it flips the narrative. Rather than seeing morality strictly as a trait-like property of individuals, the authors offer a view of morality that is, at least in part, state-like, and grounded in context. This trait-state approach to morality builds on and extends other recent work that ties timestable properties of individuals to factors that operate in the moment of decision making (e.g., Van Gelder et al. 2022).

Conclusion
The three empirical contributions that emerged from the workshop and that make up this special issue exemplify the ambition to advance the theoretical and methodological toolkit for the study of criminal choice processes in different ways. First, each of the contributions addresses an important contextual aspect influencing the framing of choice, namely peers, morals, and the (il)legal context in which money is acquired. Together, they illustrate some of the variety and breadth of contextual factors that affect criminal choice. Second, each of the contributions uniquely demonstrates how advances in various scientific disciplines can advance the state of the art in the study of criminal choice. Cognizant of the interdisciplinary nature of the field and in recognition of the fact that breakthrough insights in the field of criminology have regularly trickled down from scientific discovery in adjacent fields, the contributions draw from psychological research (Herman and Pogarsky), sociology (Nguyen, Loughran, and Topalli) and economics (Engel; Nguyen, Loughran, and Topalli). Third, and finally, each of the contributions extends existing choice models to explain offender decision making in context in a unique way using different qualitative and quantitative research methods including lab experiments, vignettes, and offender interviews.
While the topics of these three papers are quite distinct-how the source of income effects it use, how situational features shape moral judgments, and how number of peer and effects rule adherence, they share the common thread of demonstrating that context matters and how it does so. We conclude that the role of context in information processing, judgment, and decision making is pervasive and deserves careful empirical study in future criminological research.

Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD
Jean-Louis van Gelder https://orcid.org/0000-0002-7928-5066 behaviors over the life course, the deterrent effect of criminal and non-criminal penalties on illegal behaviors, and the development of statistical methods for analyzing longitudinal data.
van Gelder and Nagin