Revisiting the tautology problem in rational choice theory: What it is and how to move forward theoretically and empirically

Rational choice perspectives have been a mainstay of the criminological canon for over two centuries. Early “classical” iterations challenged dominant approaches to punishment at the time which were “often unfair, cruel and affected by religious superstition and prejudice” and argued that crime was largely the result of free-willed actors rationally choosing to maximize pleasure and minimize pain (Mooney, 2020: 29). Starting in the late 1960s, contemporary rational choice approaches merged classical thought with insights from economics and psychology, contending that crime is the result of rational decision-making processes to commit a crime based on assessments of perceived costs of criminal activity against perceived benefits, as situationally constrained as such assessments might be (Becker, 1968; Clarke and Cornish, 1985; Cornish and Clarke, 1986, 1987). The perspective has become the go-to for criminologists interested in criminal decision-making processes ever since.

Despite such popularity, rational choice theory has not been without detractors. ¹ Questions have been raised regarding the perspective's limited conception of rationality, deterministic view of human agency, and tendency to advance policies and programs rooted in punishment and deterrence (Cullen et al., 2002: 279; Hayward and Young, 2004; Tunnell, 2002). Further, scholars contend that, outside of vignette studies, empirical support of the perspective is weak (Pratt et al., 2006: 381–2; Tunnell, 2002). For their part, rational choice theorists have sought to address these criticisms. They deny, for instance, that the theory reduces humans to homo economicus—actors that make cost–benefit decisions for purely material or fiscal reasons (Paternoster et al., 2015). They also nuance the theory to consider a range of factors that may circumscribe rationality (e.g. Bouffard et al., 2000; Exum, 2002; Nagin and Paternoster, 1993; Paternoster and Piquero, 1995; Piquero and Tibbetts, 1996).

Though appreciative of the contributions rational choice has made to the field, this essay advances an additional critique of the perspective as problems persist that cannot be addressed through minor modifications. To this end, the current analysis details a fatal tautology resting at its heart—that if a crime occurs, then by definition the benefits must exceed the costs. We are not assaulting the idea of rational choice as an ontological position nor the idea that humans can act rationally. This analysis is also not an indictment of criminological approaches to criminal decision making more generally (e.g. situational action theory, structuration theory, cultural criminology, symbolic interactionism, etc.) (e.g. Ferrell et al., 2015; Giddens, 1984; Treiber, 2011; Tunnell, 2002; Wikstrӧm, 2006). Instead, we are critiquing the prevailing use of rational choice theory in criminology as a specific theory of crime causation which frames crime as a product of self-interest that is shaped in fundamental ways by the threat of punishment. While this may seem like a narrow band of theorizing to target, it is unfortunately the approach to criminal decision making most widely used in criminological research and taught in criminology courses.

We are certainly not the first to level the critique of tautology against rational choice theory (see, e.g. De Haan and Vos, 2003; Friedman and Hechter, 1988; Whitford, 2002). And since others have come before us, it may be tempting to assume that criminologists are already well aware of this criticism. The problem, however, is that the field at large does not seem to have gotten the message since the bulk of the contemporary rational choice/deterrence literature ² in criminology still sees the choice function (e.g. the weighing of costs and benefits) as the key predictor of criminal behavior (Nagin et al., 2018). It therefore seems like many criminological rational choice theorists are either unaware of the tautology problem or are engaging in a kind of “fetishistic disavowal,” a condition where “we can continue not wanting to know what we already know and therefore continue to live and act as if we don’t know” (Hall, 2012: 93). In other words, such rational choice theorists may be aware of the problem and choose to ignore it.

Accordingly, in the present paper we revisit the rational choice perspective's tautology problem with specific attention paid to its philosophical and empirical consequences. Our broader purpose is not necessarily to “rescue” the rational choice perspective in criminology—it does not need rescuing since a sizable portion of the field is committed to keeping it around regardless of whether it makes sense or provides a convincing explanation of criminal behavior (see, e.g. the discussion by Pratt, 2008). Instead, our purpose here is to lay out how this perspective could be altered—both theoretically and empirically—to be more useful for thinking about crime and its potential control. The concern is that carrying on “business as usual” with a tautological rational choice theory undermines the potential utility of such theory and, given its pervasive influence in criminology, the scientific legitimacy of the field as a whole.

A petit review of rational choice

The “Classical School” of criminology arose in the 1700s and is most notably associated with the Enlightenment philosophers Cesare Beccaria and Jeremy Bentham. While there were notable differences between the two (see Mooney, 2020), they shared the view that humans are rational beings imbued with free will. Consequently, crime is a relatively simple matter—humans will commit crime when an opportunity is present and the potential pleasures of offending outweigh the possible pains of punishment. Bentham argued such actors employ a “hedonistic calculus” where “human beings are presented as rational choice actors who are making calculated decisions [emphasis added]” (Mooney, 2020: 50). Because criminal motivation was largely taken for granted, these classical theorists dedicated more attention to penal reforms and the elements of successful deterrents than crime causation. The advantages or pleasures of crime were self-evident, and thus the state—to preserve the social contract—should enact predictable punishments for offending that would create disincentives toward criminal pursuits. For this reason, Beccaria (1764) argued that effective and just punishments are swift in execution, certain to occur following an infraction, and sufficiently severe to offset the potential gains of criminal activity, but not so harsh as to create greater social harm than the good achieved through crime prevention.

The importance of the Classical School cannot be overstated. It revolutionized criminal justice throughout the Western world. As an explanation of crime, however, the idea of the calculating rational actor fell by the wayside as positivistic approaches gained prominence in the 1800s and through the 20th century which considered the biological, psychological, and sociological determinants of crime (Mooney, 2020; Snipes et al., 2019). In the late-1960s, however, scholars like Gary Becker (1968) revisited the ideas of Beccaria and Bentham and merged them with insights from contemporary economics to create a “rational choice theory” of crime. Like Bentham, Becker argued that humans decide to commit crimes when the perceived benefits outweigh the costs. Following in his footsteps, Ronald Clarke and Derek Cornish (Clarke and Cornish, 1985; Cornish and Clarke, 1986, 1987) advanced the “reasoned offender” approach to crime (McCarthy, 2002: 418). This rational choice theory is considered “neo-classical” as it merges classical considerations of humans as rational actors with positivistic ideas about the importance of individual and situational differences for structuring decision making (Mooney, 2020: 70). Thus, contemporary rational choice theory argues that,

Offenders seek to benefit themselves by their criminal behavior; that this involves the making of decisions and of choice, however rudimentary on occasion these processes might be; and that these processes exhibit a measure of rationality, albeit constrained by limits of time and ability and the availability of relevant information. (Clarke and Cornish, 1985: 1)

Instead of criminality, rational choice theorists examine decision making. These include decisions involved in, but not limited to, commission, persistence, desistance, target selection, specialization, and frequency of offending (e.g., Clarke and Cornish, 1985; Hough, 1987; Thomas et al., 2020; Wright and Decker, 1994). Contemporary rational choice theory is also intractably connected to its fellow neo-classical approach to crime—perceptual deterrence theory (Paternoster, 1987; Waldo and Chiricos, 1972). While rational choice theories attempt to examine criminal evaluations of both the benefits and costs of engaging in criminal activity, perceptual deterrence theory focuses on the role of sanctions—particularly formal punishments—as potential deterrents to crime. Further, perceptual deterrence theory differs from classical approaches by arguing that subjective rather than objective experiences of swiftness, certainty, or severity of punishment are the most important for achieving deterrence. ³

Much of the empirical examinations of rational choice have stemmed from attempts to measure the impact of sanctions on criminal decision making conducted within the perceptual deterrence framework. The strongest quantitative associations in the perceptual deterrence literature are between people's “certainty estimates” and self-reported intentions to offend in vignette-based surveys (Pratt et al., 2006: 381–2; Tunnell, 2002). In effect, the propositions of the theory hold for individuals asked to make decisions in safe environments removed from actual criminal events, given a hypothetical scenario of a crime they are not actually committing, and then asked to consider the likelihood of punishment. It is unsurprising that individuals make decisions most consistent with the deterrence perspective under such sterile conditions. Yet when studies consider other measurements of criminal behavior and decision making, the predictors of deterrence and rational choice are weak (Paternoster, 1987; Pratt and Cullen, 2005; Pratt et al., 2006).

One of the reasons the perspective endures despite empirical shortcomings is its practical utility. The arguments offered by the perspective are easily applied to policies that modify punishments to maximize alleged deterrence effects or limit criminal opportunities through situational crime prevention strategies (Clarke, 1997). Indeed, these approaches have been popular as they appeal to common sense and often conservative notions about crime and criminal justice. Although these policies often overlook the nuances that rational choice theorists consider, the result, as Pratt (2008: 44) explains, is that,

Policy makers from both sides of the political spectrum have consistently embraced policies that crank up sentences for more types of offenses (especially drug offenses) and for more types of offenders (especially nonviolent and youthful offenders) under the rubric of concern over public safety.

The tautology problem in rational choice theory

Empirical and policy issues aside, the most significant problem we see confronting rational choice theory is conceptual. Most significantly, the perspective is tautological as an explanation of crime. We explore this issue in depth herein not to belabor the point, but because it is vital that we make clear the nature of the tautology and its consequences before charting the pathway forward for the perspective. Even the most cursory summation notes that the perspective hinges on calculations of perceived costs against perceived benefits of offending. These assessments may be “intuitive and impulsive,” “imperfect,” or even “impaired,” but regardless of the “boundedness” of rationality, offenders are said to make their decisions using some form of this calculus (Farrell, 2010: 52). Therefore, according to the theory, if a crime occurs then it must be because the perceived benefits outweighed the costs. A conception of human rationality totally dictated by cost–benefit analyses introduces a vexing question: Can rational choice theorists envision any scenario where crime occurs when perceived risks outweighed rewards? Or, alternatively, can a situation exist where an opportunity was present, and the perceived benefits outweighed the costs, but no crime was committed? The answer is likely “no.” From this perspective, if a crime occurs then by definition the benefits outweighed the costs (however cursory the search for costs and benefits was by the offender) (Coleman, 1990). ⁴ Statistically, any time the cost–benefit ratio and offending are not correlated at 1.0, then it can be assumed to be a result of measurement error. In this sense, looking at the relationship between cost–benefit ratio and offending is useless and unfalsifiable. Proponents of rational choice will likely reject such an assertion. Like rebuttals against charges of tautology endemic to self-control theory, such authors would likely consider this criticism “a compliment, an assertion that we followed the path of logic in producing an internally consistent result” (Hirschi and Gottfredson, 1994: 8). Yet we are not convinced such issues can be dismissed so easily.

Whether intentionally or unintentionally, rational choice theorists appear to side-step this rigid and circular view of rationality by claiming humans are not perfectly rational creatures. Their capabilities to make proper cost–benefit analyses can be constrained or “bounded” in multiple ways. As Cornish and Clarke (1987: 933) explain:

This “rational choice” perspective on crime assumes that offenders seek to benefit themselves by their criminal behavior; that this involves the making of decisions and choices, however rudimentary on occasions these choices might be; and that these processes, constrained as they are by time, the offender's cognitive abilities, and by the availability of relevant information, exhibit limited rather than normal rationality. [emphasis added]

To this end, rational choice theorists have considered a variety of such constraints on rationality including individual perceptions and preferences (Bouffard, 2007), affect and emotions (Bouffard, 2014; Bouffard et al., 2000; Clarke, 2014; Jacobs and Cherbonneau, 2017, 2019; Kamerdze et al., 2014; Paternoster and Pogarsky, 2009), morality (Piquero et al., 2016), substance use (Clarke, 2014; Exum, 2002), prior experiences with crime and punishment (Paternoster and Piquero, 1995; Stafford and Warr, 1993), attachments to prosocial institutions (Paternoster, 1987), cognitive biases and heuristics (Collins and Loughran, 2018; Pogarsky et al., 2017; Schneider and Ervin, 1990), and mental strategies to reduce perceptions of danger (Jacobs and Cherbonneau, 2017). Individual differences in criminal propensity, like low self-control, have similarly been examined (Nagin and Paternoster, 1993; Piquero and Tibbetts, 1996). Rational choice researchers have surveyed the role of formal and informal sanctions as well as the various rewards that might be considered (material, social, and expressive) and alternative courses of action to achieve such rewards (Paternoster et al., 2015; Piliavin et al., 1986; Loughran et al., 2016). Variations in perceptions of risks and rewards by crime type among individuals have been noted (Thomas et al., 2020). Bayesian logic has been utilized to understand how offenders adjust their perceptions of punishment likelihood based on experience (Anwar and Loughran, 2011). The conclusion of such work is that what constitutes rational can vary significantly between individuals and is largely dependent on context and individual proclivities. Yet, despite these nuances, all decisions remain fundamentally about maximizing advantages over disadvantages.

Indeed, it is likely that the tautology undermines otherwise robust theorizing on the part of such scholars. Consider the recent analysis by Loughran and colleagues (2016) which advances rational choice as a general theory of crime. Drawing from the Pathways to Desistance study, the authors examine participant assessments of perceived likelihood of apprehension, sanction (both formal and informal), and reward attainment (expressive, social, and pecuniary) across hypothetical scenarios against self-reported involvement in crime. According to the authors, the study supports the notion that juveniles—a population often thought less rational—are to some extent rational decision makers when involved in crime. While the sophistication of the analysis is laudable, it suggests that the findings are evidence of a fundamental rationality in offender decision making. Further, it asserts that any failure to explain all crime does not imperil the theory but, rather, indicates the theory is incomplete. As they explain:

We emphasize that often the failure of individuals to act in a fully rational manner is not necessarily an indictment of the entire theoretical paradigm of rational choice, but instead, it can provide important insight into heuristic devices and predictable deviations from rational behavior that may be useful in tweaking the theory. (Loughran et al., 2016: 107)

Such an approach is consistent with Coleman's (1990: 16–7) discussion of rationality in Foundations of Social Theory:

It is also important to answer the objection that individuals do not always act rationally. It will not dispute the point, for it is clear that persons sometimes act self-destructively and at other times act with questionable rationality. I will say this, however: Since social scientists take as their purpose the understanding of social organization that is derivative from actions of individuals and since understanding of an individual's action ordinarily means seeing the reasons behind the action, then the theoretical aim of social science must be to conceive of that action in a way that makes it rational from the point of view of the actor. Or put another way, much of what is ordinarily described as nonrational or irrational is merely so because the observers have not discovered the point of view of the actor, from which the action is rational. [emphasis added]

An additional problem with assuming human rationality and working backward is that the possible range of variables that may structure individual perceptions, preferences, beliefs, and values, as indicated previously, are multitudinous (Satz and Ferejohn, 1994: 72–4). Unfortunately for rational choice theorists, there exists a fine line between nuance and convolution. The theory buries its core mechanic of cost–benefit analysis under a mountain of extraneous variables. Each failing of the theory results in an adjustment, but the adjustments only seem to obscure the operations of cost versus benefit decision making rather than illuminate them. Such complexity begs the question, “at what point does decision making become irrational?” Considering the complexities of human decision making suggested by the theory, the rationality of human actors is not only tautological, it risks reification or falling subject to a “divine fallacy”—that the theory must be true because there are no circumstances by which we can imagine that humans act irrationally. Any irrationalities must simply be a result of a failure to consider additional variables and processes at play.

Some may argue that rational choice theorists do not argue that all decision making is based on a utilitarian calculus—that only some decisions may be made in such a manner. For instance, in a recent analysis, Thomas et al. (2020: 488) explain that the “mathematical” approach to rational choice theory states that “if the parameter estimate is negative and statistically significant for arrest risk and positive and statistically significant for rewards, then this is seen as evidence that individuals engage in some utility calculus before deciding to offend [emphasis added].” The implication, then, is that at least some variation in criminal decision making is based on non-utilitarian concerns regarding perceived risks versus rewards. No alternative decision-making criteria are provided, unfortunately. Such an omission reflects a broader pattern within rational choice theory.

Some theorists, however, do hint that not all offenders act based on rationality. Yet upon further inspection, such caveats are similarly absorbed into cost–benefit logics. For instance, Pogarsky (2002: 433) argues that there exist some offenders who are entirely undeterrable, “driven by biological and psychological urges” or otherwise have executive functioning deficits that compromise their abilities to make cost–benefit decisions. It is not clear that such offenders make choices at all—their rationality is so severely curtailed that they cannot consider one side of cost–benefit calculus. On the other hand, there are “acute conformists” who do not deviate from their moral convictions regardless of potential criminal rewards (Pogarsky, 2002: 433). For Paternoster and Simpson (1996: 574), such conformists are deontological, that “strongly held moral beliefs that proscribe a given action may so effectively inhibit conduct that considerations of cost and benefit are not even brought into play [emphasis added].” There is little elaboration, unfortunately, about how such decisions to abide are made. If they are not making decisions, then they are perfectly controlled by their moral code. If they are making a choice to abide, then what deliberation occurs? For Pogarsky (2002: 444), such acute conformists may evaluate “self- and social disapproval” stemming from moral violations thus absorbing deontological considerations back into utilitarian terms.

The nuances, addendums, and caveats made to rational choice theory are ad hoc adjustments to sustain a theory that is conceptually and empirically flawed. ⁵ Criminological rational choice scholarship constitutes “normal science” where observations and data are constantly tweaked to conform to the expectations of a paradigm—in this case, that all human action is based on a rational cost–benefit analysis (Kuhn, 1962). In this manner, efforts by rational choice scholars to nuance the theory appear to force observations to conform to underlying assumptions about human nature and cognition. What may be more useful, however, is to reconsider the core assumptions and propositions of the theory. With some adjustment, rational choice can remain a useful theory for policy while also a more robust explanation of criminal behavior.

Reevaluating the “rational” in “rational choice theory”

Pointing out the tautology that plagues the rational choice perspective does not mean that scholars must abandon the concepts of rationality and rational action. To quote Hayward (2012: 31), we “recognize that most human decision-making is (to a certain extent) the product of complex formations based around rational criteria.” Instead, the objective is to decouple rationality from the limited foundation of risk-reward assessments. Rational choice theory must rethink this core axiom and rebuild accordingly. Further, criminologists can understand that human behavior is “rational” in the sense that it can be examined and understood by the outside observer (Coleman, 1990). The rationality of concern, however, involves the level of rationality engaged in by decision makers themselves. In this context, rationality is perhaps best understood as deliberate and thoughtful consideration of available information and basing action on the results of such reasoning processes. Instead of asking “how do offenders assess the costs and benefits associated with offending?” this approach asks, “under what circumstances do offenders use rational decision-making processes?” The implication of this question is significant for rational choice theory. Rationality is no longer assumed but, instead, becomes itself a variable to be examined and analyzed (we will explore the empirical implications of this conceptualization of rationality in the following section).

This conceptualization of rationality is not new. Social theory has long considered the role of agency and rationality in decision making and action. Tunnell (2002: 270), for instance, suggests that symbolic interactionism could provide a promising avenue forward for rational choice theory to connect cognitive decision-making processes to “the determinism of social variables, the significance of social structure, the meanings given to decisions and action, and the unpredictability of human agency.” In other words, a symbolic interactionist approach involves transcending the role of cognition in human decision making to consider how “social life… is produced through rich, textured interactions and communications among people as they live their lives and as they come to give definitions of phenomena” (Tunnell, 2002: 271). From this view, the role of social variables and rules enter the equation. Involved is a recognition that some decisions are made with significant thought, others are made with little though, and many may be some amalgamation of both. Rather than centralize costs versus benefits, a symbolic interactionist approach privileges how decision-making criteria may be shaped by “one's understanding and knowledge filtered through their own personal subjective definitions of such” definitions co-produced through interactions with others (Tunnell, 2002: 271). Meaning thus becomes central to rationality more so than the quest to maximize benefits over costs. Fortunately, some notable efforts have been made to incorporate symbolic interactionist insights into rational choice models—a development we hope will continue (e.g. Matsueda et al., 2020)—though these hybrid approaches still prioritize the role of costs–benefit calculi in decision making.

Similarly, criminologists of all sorts have drawn inspiration from Katz’ (1988) Seductions of Crime, which argued that humans are certainly capable of acting rationally, but there also exists circumstances in which their behaviors become decidedly irrational, such as when they are “magically” seduced or repulsed toward criminal activity. He explains that “only rarely do we actually experience ourselves as subjects directing our conduct” and

We rest our subjectivity on rhythmic sensibilities, feelings for directions, and visions of unfolding patters, allowing esthetics to guide us. Self-reflexive postures in which one creates a distance between the self and the world and pointedly directs the self into the world, occur typically in an exceptional mood of recognizing a malapropism, after a misstep, or at the slip of the pen. With a slight shock, we recognize that it was not the things in themselves but our perspective that temporarily gave things outside of us the power to seduce or repel. (Katz, 1988: 5)

Similarly, practice theories—theories that examine the interplay between structure and agency—also provide one avenue forward (Ortner, 2006). These theories assume that decisions can be understood, but not that all decisions are made in a thoughtful and reflective manner. For instance, Pierre Bourdieu (1980: 56) contends that individuals act in contexts based on situational rules dictated by habitus or the “embodied history” structured by social circumstances which in turn structures current actions within a social setting (or “field”). Importantly, these rules become internalized and thus a person may not always consciously consider such rules before acting. Similarly, Antony Gidden's (1984) structuration theory contends that human action is constrained and influenced by social structural circumstances but that social structure itself is created and created through such actions. In these cases, social structure shapes behavior but humans can exercise their agency and choose to reinforce social structures or deviate from them. Individuals may, for example, choose to engage in crime but may give little thought to how their offending conforms to gendered social structural expectations (Messerschmidt, 1993). Others may consciously flout such gendered expectations. In short, the focus is on how structure permits, obstructs, encourages, discourages, and otherwise shapes the situational constraints of decision makers, the rules governing social behavior, and the criteria assessed by decision makers. Like with symbolic interactionism, rationality is understood to result from a dynamic interplay between structure and agency whereas rational choice theory tends to place the emphasis on individual cognition.

Criminologists have also advanced “situational action theory” as yet another perspective on criminal decision making which recognizes that individuals may act rationally but does not moor all actions in cost–benefit calculations (e.g. Wikstrӧm, 2006; Wikstrӧm et al., 2015; Wikstrӧm and Treiber, 2017). Admittedly, this is a rather complex and nuanced approach to decision making and thus fully capturing the perspective is beyond the scope of this analysis. Generally, however, situational action theory argues that humans are “rule-guided creatures” who make decisions in situations by choosing among perceived “action alternatives” (Wikstrӧm, 2019: 260). The theory requires an examination of the contexts in which decisions are made, charting how they may shape an actor's perceived possibilities for action, the permissibility of particular actions, and related matters. It also takes into account an individual propensity (e.g. motivations, self-control, etc.), which themselves interact with environmental context, a relationship the theory refers to as the “person-environment interaction” (ibid.). Crime is thus viewed as a “moral action” or one “guided by value-based and emotionally grounded rules of conduct about what is the right or wrong thing to do in particular circumstances” (ibid.: 261).

Scholars have also begun advancing dual-process or dual-systems approaches to decision making (e.g. Kahneman, 2011; van Gelder, 2017; Vazsonyi and Ksinan, 2017). While it is beyond the scope of the analysis to detail all the ins and outs of such approaches, in short, dual-process models argue that the kinds of decision-making processes involved in criminal activity may vary wildly depending on the amount of forethought, emotion, and impulsivity involved. In other words, some decisions may be careful and deliberate, others emotional and unreflective, and others out of habit (Thomas and McGloin, 2013)—and the kinds of processes involved in making such decisions may look different and result in different outcomes. Dual-system approaches, on the other hand, are similar but look less at the process and, instead, examine how different parts of the brain may be involved in certain kinds of decisions (Steinberg, 2010). Regardless, both approaches do not assume that all decisions are guided by self-interest.

These are only some examples of social theories which demonstrate that behavior can be rationally understood while also acknowledging that behavior might not itself be the result of a fully rational decision maker. The point is that rationality need not be oversimplified as cost–benefit assessments. There exists a bevy of alternatives for criminologists to consider. Interestingly, rational choice scholars themselves have advanced an alternative approach to rationality that could be useful for the task at hand with some modifications.

In their 2009 analysis, Paternoster and Pogarsky describe the concept of “thoughtfully reflective decision making” (TRDM)—a type of thinking thought correspond to “good choice-making” where outcomes from choices are “consistent with preferences.” Making choices based on TRDM requires that the decision maker:

(1) recognize that there are alternatives to attaining some goal, and that one must collect information about these alternatives and what the costs and benefits are to each, (2) consider and compare the costs and benefits of these alternatives, (3) make a decision as to which alternative one is going to choose based on that consideration, and (4) revisit that decision later to see if it could have been improved. (Paternoster and Pogarsky, 2009: 107)

Such decision making is objectively reasonable and the sort of thinking most likely to lead to desired short-term and long-term outcomes. For most rational choice approaches, the assumption is that any outcome chosen must have been selected because the benefits outweighed the costs. The approach advanced by Paternoster and Pogarsky (2009), however, only requires that such advantages and disadvantages are considered—the emphasis is thought on the degree of thoughtfulness given to a decision rather than the outcome necessarily. Certainly, individuals who engage in TRDM are likely to make decisions that maximize perceived benefits over costs, but it does not require such outcomes.

Further, this approach recognizes that “not all decision making involves calculation and reasoning and that a great deal of the decisions we make in life involve intuition, habit, emotional and moral factors” (Paternoster and Pogarsky, 2009: 121). From this view, rational actors still consider costs and benefits, but actors are not always assumed to be acting rationally. Rationality is no longer restricted to the pursuit of costs and benefits but, instead, to the degree of reflective thoughtfulness given to the consideration to a decision. For example, the authors argue that

Even among generally thoughtful persons many decisions may be made reflectively or impulsively as in deciding where to eat on the basis of the closest restaurant (immediate gratification), other decisions may be made on the basis of habit (what route to take to work), or emotional urges. (Paternoster and Pogarsky, 2009: 110)

Before moving on, it should be noted that research within the rational choice tradition has incorporated TRDM and similar approaches into a “dual-process thinking” model of criminal decision making (Pogarsky and Herman, 2019). This approach differentiates between decisions made as a result of conscious consideration of costs and benefits against those that are made in a rasher manner or on the basis of intuition (Paternoster et al., 2011). Unfortunately, the development of such dual-process models among rational choice theorists remains grounded in assessments of risk, costs, and benefits (Pogarsky and Herman, 2019). For instance, Loughran (2019) gives a deep consideration of the value of behavioral economics for the study of crime, noting how the field accounts for seemingly irrational behavior (or “anomalies”) through nuanced models of decision making. Still, the author claims “a useful theory must include how potential offenders will respond to both disincentives (i.e. sanctions) and incentives from crime” (Loughran, 2019: 742). Paternoster (2017) similarly argues for the role of agency in criminological research and attempts to create a nuanced portrayal of humans as rational actors. Though the effort is laudable, he defines an “agent” as a person who “makes things happen on their behalf first by contemplation and rumination, making decisions, formulating plans, and then acting to bring those plans about” (Paternoster, 2017: 351). By including “on their behalf” in this definition, however, such an approach to agency remains grounded in self-maximization. We think this is a mistake as it means that an otherwise promising development in rational choice theory devolves back into tautology. Instead, rational choice theorists interested in dual-process models should reject the assumption that all decisions can be reduced to hedonistic reasoning, following the trajectory of other criminologists considering pursuing dual-process or -structure approaches to decision making (e.g. van Gelder, 2017; Vazsonyi and Ksinan, 2017).

Empirical implications for testing the rational choice perspective

Common practice in empirical research in the rational choice‒crime tradition—whether it comes from criminologists or sociologists or economists—tends to go in one of two ways. First, rational choice/deterrence scholars have long embraced the “vignette” method of collecting data (Pratt et al., 2006). In the vignette method, survey respondents (often captive college students) read a brief description of a person committing a crime. Respondents then offer their judgments about the chances that they might be caught if they were to commit the offense described in the vignette (a “certainty estimate”), how severe they think their punishment might be (a “severity estimate”), and the odds that—based on their perceptions of the punishment they were just asked about—how likely they would actually commit the offense themselves (Barnum et al., 2021; Barnum and Pogarsky, 2022). The findings from these studies are mixed at best, yet the entire endeavor has been criticized for: (1) forcing respondents to think about potential penalties when in their actual daily life they might not, and (2) removing criminal behavior from its “real life” context (one often characterized by high emotion and chemical impairment) and instead intellectualizing it with a pen and paper from the calm and comfort of a classroom (Pratt and Turanovic, 2018).

To get around this problem of the unreliability of people's subjective judgments of whether they would or would not engage in particular offenses, or what they think the penalties might be if they did, the second most common type of study in the rational choice‒deterrence tradition is where scholars will measure survey respondents’ perceptions of the potential “costs” of crime (usually their perceptions of the certainty and severity of punishment should they engage in offending behavior, see Paternoster, 1987; Pratt et al., 2006). Next, they will see if those measures are associated with either respondents’ past criminal behavior or their projections as to whether they would engage in some form of criminal behavior in the future (Apel, 2019; Hirtenlehner and Wikstrӧm, 2017; Thomas et al., 2020). Finally, they will control statistically for other potential criminological factors (e.g. self-control, social bonds, deviant peer influences), as if they are a nuisance that needs to be accounted for, to say that the effects of the perceptions of costs and benefits on crime are not spurious (Nagin, 2018; Nagin and Paternoster, 1993; Paternoster, 2018).

We have demonstrated that this approach is problematic theoretically—that one's perceptions of the costs/benefits of crime cannot truly be separated from the criminal behavior itself. All is not lost, however. There is a promising avenue that scholars can—and should—explore empirically: treat the perceptions of the costs and benefits of crime as the dependent variable of interest. If done in this way, scholars could use survey instruments to ask the rational choice/deterrence research question in several theoretically and empirically important ways. For example, what are the factors (e.g. low self-control, poor prosocial attachments, exposure to deviant peers) that influence why some people perceive the costs of crime to be prohibitively high while others do not? Alternatively, what is the role of those same factors in determining why some people are attracted to the benefits of crime while others fail to see the upside of it? In short, why do some people seem to give more consideration to decisions to commit crime than others regardless of whether those decisions result in actual or perceived utility? The important point here is that the key variable in most of the body of literature on rational choice and perceptual deterrence (one's perception of the costs and benefits of crime) should be the outcome of interest moving forward—not an independent or mediating or moderating variable.

What we are thus proposing is that all of those theoretically derived factors that scholars have typically treated as control variables (e.g. self-control, social bonds [to school, marriage, work], deviant peers) need to come back into the empirical equation—not as mere controls but as the key theoretical variables of interest. Scholars have already begun to do this (e.g. Altikriti et al., 2022; Lee et al., 2018; Piquero and Tibbetts, 1996). Most recently, using the Pathways to Desistance data, Thomas et al. (2022) found that both community structural characteristics (e.g. concentrated disadvantage) and individual differences (e.g. impulsivity and IQ) influenced people's perceptions of the risks and rewards of criminal behavior. Similarly, situational action theorists and scholars working within the dual-process and dual-systems approaches have already laid significant groundwork for perspectives that treat rationality itself as a variable in decision making rather than take it as a given (e.g. van Gelder, 2017; Vazsonyi and Ksinan, 2017; Wikstrӧm, 2006; Wikstrӧm et al., 2015; Wikstrӧm and Treiber, 2017).

These are all promising developments. But thinking about (and studying) the rational choice model in this way has not exactly “caught on” with the rational choice faithful at large. Instead, rational choice enthusiasts have complained that their pet perspective has been sold short—that the reason rational choice has not been bestowed the status of a “general” theory is because researchers have focused too hard on the costs of crime and have neglected perceptions of the benefits (Loughran et al., 2016). We disagree. Instead, we argue that if the rational choice folks want their theory to be the centerpiece at the criminological table, then they should embrace the very criminogenic factors that they typically give lip service to. Indeed, factors like community levels of disadvantage and disorder, along with individual characteristics like self-control, social bonds, and deviant peer associations, should not be treated as mere statistical controls that need to be included in the model so that the “rationality” variables can be reliably estimated as if they were isolated from everything else. Instead, scholars should start developing a knowledge base about why people vary in how they perceive the costs and benefits of criminal behavior and the extent that they employ rational criteria at all.

Conclusion

Regardless of the path that rational choice theorists choose, it is clear that sustaining the theory on the back of a tautology is not a productive way forward. Instead, theorists should consider the possibility of variations in the types of calculi used by decision makers and the degree of forethought given to decisions to commit crime as the lynchpin of their theory. Such an approach may allow for empirically and conceptually sound assessments of criminal decision making while also avoiding ad nauseum ad hoc adjustments to models to bend observations to narrow conceptions of rationality. It is worth repeating, however, that this critique does not mean that there is no room for analyses that consider cost–benefit analyses. The contention, here, is that reducing decision making to perceptions of risks and rewards not only transforms humans into two-dimensional actors but perpetuates theory founded on tautology. Breaking this cycle also has greater disciplinary implications—it means that rational choice theory will need to separate itself from the yoke of behavioral economics, a discipline moored in cost–benefit analyses. It also means that the theory will have to decouple itself from a politics of punishment—which not all rational choice theorists support—that assumes all criminals are hedonistic actors that can be deterred by properly implemented sanctions, a politics that has directly contributed to punitive regimes and mass incarceration. While the rational choice theory has certainly been a multidisciplinary perspective, drawing ideas and variables from other disciplines, allowing rationality itself to be the object of scrutiny opens the door to a truly interdisciplinary perspective more consistent with the criminological enterprise.