Reason Taking the Reins From Impulsivity

Abstract

Aspects of criminological theory are premised on the belief that criminals make poor decisions. There have been suggestions that individual differences in self-control, willpower, impulsivity, time orientation, or more recently, thoughtfully reflective decision making (TRDM) influence choices and, ultimately, deviant outcomes. While much of the literature suggests there are differences among these concepts, they are often used interchangeably or at least noted to share common ground. Using survey data collected from university undergraduate students, this article explores the conceptual and empirical overlap and areas of distinction between key theoretical constructs. Using hypotheses derived from a dual-systems model, findings suggest impulsivity, self-control, temptation, and TRDM are distinct but interrelated constructs. Impulsivity was positively related to intentions to drink and drive in a hypothetical scenario, but temptation, self-control, and TRDM had no significant effect on intentions. Consistent with a dual-systems conceptualization, we found impulsivity and self-control work in tandem, as the risk of drinking and driving was highest for those respondents who were highly impulsive and had low self-control. Future research may seek to use a dual-system model to further reconcile trait-based and decision-based models of criminological theory.

Keywords

self-control impulsivity temptation TRDM individual differences dual-systems model

Our impulses are too strong for our judgment sometimes.

—Thomas Hardy

Introduction

In recent years, criminological theory has seemed to move along two separate and very distinct paths. One path is represented by trait-based theories which posit that crime and a wide assortment of other self-harming behaviors are the product of stable-individual differences in some underlying propensity to commit these acts. The second path is associated with decision-based theories where crime and problem behaviors are understood to be a rational response to the balance of situational temptations and restraints, the estimated net utility of such behavior, or due to breakdowns in rational decision making such as poor decision-making strategies/styles, discounting of delayed events, or decision-making heuristics and shortcuts. Both theoretical models have existed and even thrived within criminology while virtually ignoring the other’s possible strengths and the seemingly obvious connection between the two on a number of points.¹ Just as troubling has been the confusion within the trait-based perspective as to exactly what time-stable individual trait or traits are important in understanding crime. Gottfredson and Hirschi (1990), for example, have argued that the individual-level trait that explains “all crime, at all times and, for that matter, many forms of behavior that are not sanctioned by the state” (p. 117) is what they term self-control. The definition of self-control has, however, been both all-inclusive at times and has changed over the years. For example, in the 1990 exposition, self-control appears to entail the individual’s capacity to resist impulsive behavior, yet it also seems to include impulsivity as one of its dimensions—that is, among other things, low self-control is impulsivity.

There is, however, a third possible path which can reconcile these divergent positions that has been given a great amount of attention in cognitive psychology and behavioral economics and that is the notion of dual-system decision making. There are different specific versions of dual-system theories, but the common ground is that they all assume that decision making consists of two distinct but related systems (Camerer, 2007; Kahneman, 2003, 2011; Loewenstein, Rick, & Cohen, 2008; Stanovich, 1999, 2011). One system is intuitive, fast, and is based on affect and the demands of the immediate situation. It might easily be called more impulsive decision making. The second system is slow, and is based on the collection and processing of information, and is, therefore, premised on contemplation and the rational consideration of the costs and benefits of action, particularly the long-term costs and gains of a given action. These are not completely different and unrelated processes, for they work together in that as Kahneman (2011) noted, “[o]ne of the functions of System 2 (the slower, more thoughtful system) is to monitor and control the actions ‘suggested’ by System 1” (p. 45). In other words, while our more emotive and intuitive system may impulsively suggest one line of action, the more thoughtful and slower part determines whether that is a good action to take and can, therefore, be expressed, or a not-so-good action that can and should be inhibited.

Dual-system models, then, hold out the promise of uniting the trait-based and decision-based models of criminological theory because they combine intuitive, strongly affective, and impulsive inputs to behavior as well as more rational, thoughtful, and deliberate inputs. The position of dual-systems theory is that risk-taking behavior such as crime, drug use, and promiscuous sex is not due solely to impulsive people acting impulsively, nor is it entirely due to the rational weighing of the costs and gains of such risks, but involves the operation of both. Before dual-systems theory can unite the trait-based and decision-based paths in criminology, however, there is a need to clear up the ambiguity present in current understanding of key criminological concepts like impulsivity and self-control because there seems to be little agreement as to what these terms mean, and perhaps most frequently they are treated as synonymous.

In what follows, we first present brief descriptions of trait-based and decision-based theories of crime. In the former, we also present what we think is a fundamental misunderstanding about self-control in the field and how this concept differs from impulsivity. We then move to a discussion as to how dual-system models can reconcile trait- and decision-based models and hypothesize that both systems are involved in understanding antisocial behavior like crime. We derive some hypotheses from the dual-systems model and test these with data collected from a sample of university students. We close by discussing the implications of our findings for criminological theory.

Trait-Based Theories of Crime

Trait theories, which presume that crime is based on time-stable individual differences in the propensity to engage in risky or harmful behaviors, are among the oldest in the discipline (Goddard, 1914; Hooton, 1939; Prichard, 1837). The essence of a trait-based approach is that the individual possesses a characteristic that makes them more likely to commit crimes and other self-risky behaviors. Sometimes, this trait is biological as in atavism (Lombroso, 1876), or it is psychological as in the psychopathic personality (Cleckley, 1941; Yochelson & Samenow, 1976); other times, it is the combination of the two (Eysenck, 1977; Eysenck & Gudjonsson, 1989; Wilson & Herrnstein, 1985); and at still others, there is a decidedly social origin (Gottfredson & Hirschi, 1990). Over the years, one of the most consistent traits that has been linked to crime and other self-harmful behaviors such as gambling, drug and alcohol addiction, and psychiatric disorders is impulsivity. Impulsivity has been variously defined but seems to involve at its core the notion of behavior that is committed without foresight or contemplation. Impulsive behavior consists of actions that one does without thinking about the actions or the consequences very much or at all. It is characterized as involving speedy decision making, a lack of planning, and a reliance on quick intuition or affect rather than cognition. While the trait of impulsivity has been linked to a wide assortment of antisocial and self-injurious behaviors, an impulsive decision is not in itself bad, and the way a decision is made should be differentiated from its outcome. For example, while acting on impulse can lead to unwise and unaffordable purchases, weight gain, smoking, drinking, and gambling problems, it also helps us swerve the car to avoid crashing into a deer that darted out into the road, respond to a fire alarm, or even seize a good financial opportunity. At times, then, impulsive behavior can be good.

Frederick (2002) has provided a good illustration of an impulsive act with his test of cognitive reflection. A subject is provided with the following: “A ball and a bat together cost US$1.10. The bat costs one dollar more than the ball. How much does the ball cost?” An immediate, effortless, intuitive, and unthinking answer would be to say that the ball costs 10¢, because the sum of the ball and the bat which costs US$1 more would sum to US$1.10. However quick, effortless, and intuitively appealing this answer may be, it is wrong, because if the ball costs 10¢ and the bat costs US$1 more, that would make the bat cost US$1.10, and the total cost would be US$1.20 not US$1.10. The correct answer, which most people do not get, and which requires effort, some careful thinking and the suppression of the intuitive answer one might leap to, is that the ball costs 5¢. Now the bat which costs US$1 more than the ball is US$1.05, which together with the 5¢ ball now totals to US$1.10. The immediate answer of 10¢ was done on impulse, because it intuitively seemed to connect with the US$1 bat to sum to what we wanted it to. The impulsive answer was quick and easy, but in this case, wrong, and would have been recognized to have been wrong had we not immediately given an answer but had been required to think for 10 minutes about the problem. That likely would have given us the time to inhibit our quick and easy answer, and put in the mental effort to think things out.

This simple test for cognitive reflection reveals a lot of what we want to understand about the concept of impulsivity or an impulsive action. It is based on intuition or emotion; as when we turn our head when we get frightened by a movie scene, it is effortless and quick; and it can at times be adaptive. Sometimes, however, impulsive actions can be harmful especially when they are relied on too frequently, when a person is unable to restrain impulses when they want to or should, or when we fail to use forethought and act impulsively on important matters. We have already alluded to the fact that people who are characterized as very impulsive are often involved in harmful and costly behaviors, and impulsivity, we have noted, has consistently been implicated in criminal behavior both in the near and distant past (Cleckley, 1941; Eysenck, 1977; Eysenck & Gudjonsson, 1989; Wilson & Herrnstein, 1985) and more recently (Lynam et al., 2000; Moffitt, 1993). It is probably not unreasonable to say that impulsivity has been the individual-level trait most frequently linked to crime.

In addition to impulsivity, another important time-stable individual-level trait that has been connected to the propensity to be involved in crime is the notion of low self-control, and in the criminological literature, self-control is synonymous with the work of Gottfredson and Hirschi (1990). When first discussed in their 1990 book, self-control was conceptualized as the tendency for persons to be tempted by quick and easy gratification and to eschew difficult, persistent activities. Those with low self-control loathe to consider the long-term costs of their actions, they are focused on the present and on the attainment of immediate and effortless satisfactions. For example, they smoke because the pleasures of smoking and the physiological lift from nicotine carry far more weight than the long-term risk of lung cancer, and for the same reasons, they may sell drugs or break into homes rather than pursue and endure what they would see as the drudgery of conventional employment. Involving as it does an inability to resist immediate and easy gratification and a tendency to behave without much forethought, it would appear that self-control as described by Gottfredson and Hirschi in 1990 shares a great deal of conceptual ground with impulsivity. Both low self-control and impulsivity are time-stable individual traits, and both involve the propensity to commit a diverse range of self-harmful behaviors, including crime, gambling, substance abuse, unemployment, divorce, smoking, unemployment, among other negative life outcomes. In fact, in its initial formulation, self-control was a much more inclusive concept containing six different dimensions, but one of these dimensions was explicitly impulsivity, and the others form a complex that would be difficult for anyone to distinguish from how an impulsive person would behave. Those with low self-control have a present orientation, they lack persistence and diligence, and they are risk seeking and self-absorbed. Finally, when self-control has been used in research and in theoretical discussions of the topic, it has been taken by criminologists to be virtually synonymous with impulsivity.

We argue here for a different conceptualization of self-control, one that more clearly defines what it is and how it is different from impulsivity but one that is related to impulsivity in that they work together. We also think our conceptualization of self-control is consistent with the main theoretical understanding of the term in Gottfredson and Hirschi’s general theory of crime. In our view, and the view consistent with how the concept is used in psychology, self-control is not acting impulsively but is the prevention of impulsive behaviors. Self-control, then, is likened to self-regulation, the capacity to exercise control over and override one’s impulses where the impulse is a behavioral tendency solicited by a tempting stimulus (Baumeister, 2002; Baumeister, Heatherton, & Tice, 1994; Baumeister & Tierney, 2012). To fully illustrate, consider this situation. I have just enjoyed a sumptuous dinner in a restaurant and feel fully sated. The desert cart is rolled over to my table and I see a chocolate mousse cake, a stimulus which greatly tempts me. I was not thinking about desert, however, until I saw the cake and felt temptation, and my first impulse in response to this temptation is to order the mousse cake and scarf it. If, however, I am able to stop and think about how eating the cake will upset my plan to lose weight, and is counter to the fact that I have been doing so well in the gym and am beginning to feel lighter, and the fact that I am pretty full already, I might be able to countermand the wishes of my impulse and just say no. In sum, a temptation brings about the possibility of acting impulsively which may or may not be acted upon depending on the strength of my self-control or ability to regulate myself. Viewed this way, self-control and impulsivity are not synonymous but are distinct processes that work together: impulsivity is the tendency to act without foresight and contemplation and to be concerned only with the pleasures of immediate gratification whereas self-control is the ability to override and regulate one’s impulses.² Self-control, then, requires the capacity to inject thoughtful and conscious concentration on all of the implications of an action into one’s decision making.³

This conceptualization of self-control and impulsivity allows each independent status yet requires their interdependence. Self-control as the capacity to self-regulate one’s impulses is also consistent with Gottfredson and Hirschi’s (1990) understanding as to how self-control originates. According to one of the psychology’s major proponents of self-regulation (now used interchangeably with self-control) Roy Baumeister (2002), effective control over one’s impulses requires three things: (a) standards, (b) a monitoring process, and (c) the capacity to alter one’s behavior. Standards include some ideal condition or state that one aspires to or honors (losing weight and staying fit). The monitoring process consists of some way to track the relationship between what one actually does and the ideal (recognizing that eating chocolate mousse cake is inconsistent with one’s ideal). The capacity to alter one’s behavior simply refers to the fact that self-regulation requires the ability to change one’s behavior and override the impulsive action that is counter to one’s standards (the capacity to “just say no” to the cake). Baumeister (2002) described these sources of self-regulation from the perspective of an adult; however, they reflect the same sources described by Gottfredson and Hirschi (1990) for the creation of self-control from the perspective of a child where parents must know deviant behavior when it occurs (standards), monitor the child’s behavior (monitoring), and correct the unwanted behavior when it occurs (altering behavior). The creation of self-regulation is for Gottfredson and Hirschi something that initially occurs externally as when caregivers have the standards, do the monitoring, and alter the behavior, but when effective, is eventually taken over by the child and internalized.

In sum, we can agree that there are two time-stable individual-level traits that are directly implicated in crime and other antisocial/self-injurious behaviors, impulsivity, and self-control that should be understood as being distinct processes that work together. Self-control is not impulsivity⁴ but is the acquired ability to regulate one’s behavior so that desired impulsive actions can be inhibited and another action substituted. Furthermore, both impulsivity and self-control/self-regulation can be differentiated from the temptation or stimulus to the impulsive action.

Decision-Based Theories of Crime

Progressing along a separate path from these trait-based theories of crime have been theories that stress the idea that crime is the product not of some time-stable propensity to offend but of a person’s rational decision to break the law in response to the anticipated benefits and costs of the action. There are many varieties of decision theories in criminology, including routine activities, deterrence, and rational choice, but as it is the most general, we will restrict our attention to rational choice theory. Crime as a rational choice can be traced back to the work of Jeremy Bentham (1789/1948) where he illustrated that people are self-interested and when making a decision, they seek to maximize pleasure and minimize pain, in which pain and pleasure could be moral, physical, political, or religious. An economist, Gary Becker (1968), further established choice in criminology by suggesting that the decision to commit crime could be studied using a cost–benefit analysis similar to the way economists had been viewing consumer decisions. Becker’s (1968) model, which would later be called the expected utility model, suggested that individuals choose to commit crime as long as their expected benefit is greater than their expected cost. Notably, the expected utility model makes some key assumptions, including that costs and benefits are subjective and determined or weighted by individuals, and that decisions are based on expected utility or what people expect that outcomes will be. This perspective takes into consideration that people are imperfect in their information gathering, storing, and processing such that they make decisions with realistic limitations in an attempt to maximize their benefits, this is referred to as “bounded rationality” (Simon, 1957).

Rational choice theory in criminology assumes that persons contemplating committing a criminal act weigh the anticipated gains from committing a crime against the anticipated costs. The costs would include the formal and informal sanctions that would accrue for the offense, such as forgone non-criminal opportunities, the expected physical dangers, and any shame I would feel for violating a law. The benefits would include such things as the monetary value of the crime, the emotional or physic rewards, and any enhanced social status or reputation I would receive. Finally, I would also weigh and consider the costs of having to forego any non-criminal opportunity. At least in the abstract, all of these costs and benefits would be considered and weighted, however, crudely before a decision to offend would be made.

Under this model of offending, if the expected or subjective benefits of crime outweigh the expected costs of crime plus the forgone opportunities of non-crime, in other words when the expected utility of crime is greater than the disutility, then a decision to commit a crime will likely be made. Because persons are presumed to be rational actors and act on the basis of forethought and deliberation, they will make the decision that promises the greatest utility. Without getting into a dispute about how much information about a decision is gathered and how much or how good the rational contemplation and deliberation is before the decision is made, it should be clear that the model of crime which posits that offending is the result of a conscious decision is different from one that assumes that crime is due to some underlying time-stable criminal propensity. Our characterization of rational choice theory here is also an apt description of other choice-based theories in criminology. Deterrence theory, for example, posits that rational actors are influenced by the anticipated formal and informal sanctions for crime. Routine activities theory hypothesizes that crime is a consequence of a decision based on a rational consideration of such things as how easy it would be to commit the act without being noticed (the absence of a guardian) or how easy it would be or how much gain it would net (how attractive a target is). Choice or decision-based models of offending in criminology assume (require) much greater conscious deliberation and rationality than trait-based models which may partially explain why their paths have, for the most part, been separate and have rarely converged.

Dual-Process Models of Crime

In recent years, there has developed in cognitive/social psychology and behavioral economics a model of behavior that holds the promise of reconciling the trait and decision-based models of criminology (Evans & Stanovich, 2013; Kahneman, 2003, 2011; Stanovich, 1999). Dual-process models argue that actions are based not on one but two distinct but interrelated systems of thinking. One kind of thinking is fast, easy, and based on intuition and affect, whereas the other is slow, effortful, and based on rational calculation and deliberation. As well as anyone else, Kahneman (2011) provided a good description of the two systems which he refers to as Systems 1 and 2:

System 1 operates automatically and quickly, with little or no effort and no sense of voluntary control . . . effortlessly originating impressions and feelings . . .

System 2 allocates attention to the effortful mental activities that demand it, including complex computations. The operations of System 2 are often associated with the subjective experience of agency, choice, and concentration. One of the tasks of System 2 is to overcome the impulses of System 1. In other words, System 2 is in charge of self-control. (pp. 20-26, emphasis added)

System 1 thinking, therefore, is intuitive and impulsive whereas System 2’s is rational, thoughtful, and involves deliberation. In System 1 thinking, there is little reliance on the collection or processing of information and as a result, decisions are made in terms of its immediate consequences rather than more long-term effects. System 1 decision making is clearly in the “here and now” with little regard for the future and can be characterized as reactive, intuitive, affective, and impulsive. System 2 thinking, in contrast, is the cognitive, rational component. System 2 decisions are made with deliberate forethought and require concentrated effort. If System 1 is impulsive, present oriented, and directed at immediate gratification, System 2 is reasoned, future-oriented, and directed at longer-term objectives. As Kahneman noted above, one of the tasks of System 2 thinking is to put the brakes on the impulsive demands of System 1 and enabling persons to engage in better decision-making processes (collecting information, weighing costs, and benefits). Therefore, System 2 is the source of self-regulation or self-control. To repeat an important point, these two systems of thinking do not represent two distinct minds,⁵ rather they work together interdependently to structure how decisions get made regarding which actions should be taken.

Empirical Links Between Impulsivity, Self-Control, TRDM, and Crime

As mentioned, one of the oldest and most consistent empirical individual-level correlates of crime in the literature is impulsivity. The idea that criminal offenders lack the capacity to control their impulsive conduct was prominent among the earliest scientific (psychiatric) studies of crime (Rafter, 2004), and studies over the years have repeatedly found significant relationships between impulsivity and crime, drug and alcohol abuse, and delinquency (Farrington, Loeber, & Van Kammen, 1990; Loeber et al., 2012; Lynam et al., 2000; Moffitt, 1993; Moffitt, Caspi, Rutter, & Silva, 2001; Pulkinnen, 1986). Criminologists became re-interested in the role of impulsivity in crime with the work of Gottfredson and Hirschi (1990) for whom impulsivity was one of the six components of those with low self-control. Self-control has picked up where impulsivity might have left off in the criminological literature as one of the strongest and most consistent predictors of crime and delinquency (Hay, 2001; Pratt & Cullen, 2000). While Gottfredson and Hirschi were to eventually move away from the multi-dimensional conceptualization of self-control (Gottfredson, 2006; Hirschi, 2004), the development of the popular Grasmick 24-item Scale (Grasmick, Tittle, Bursik, & Arneklev, 1993) that provided an easy operational definition for self-control further cemented the convergence of self-control with impulsivity. Nevertheless, ignoring the original enumeration of the six dimensions of self-control and instead reading the theoretical descriptions about what self-control is supposed to do, it bears a closer resemblance to what psychologists have termed self-regulation.

The notion of self-control as the capacity to regulate one’s impulsive demands has had a long history in psychology. Consistently, individual differences in the ability to self-control or self-regulate have been shown to be related to a wide sweep of antisocial behaviors (Eisenberg & Fabes, 1992; Reynolds, 2006), both short-term and long-term. For example, failure to regulate impulses during childhood was related to drug use and delinquency both during adolescence (Block, Block, & Keyes, 1988; Henry, Caspi, Moffitt, & Silva, 1996) and young adulthood (Henry, Caspi, Moffitt, Harrington, & Silva, 1999). Caspi, Henry, McGee, Moffitt, and Silva (1995) conceptualized the failure to self-regulate as the “inability to modulate impulsive expression” (p. 59) and found low self-regulation measured at ages 3 and 5 was significantly related to teacher and parent ratings of externalizing problem behaviors (which included antisocial behavior such as aggression and delinquency) in early childhood and later on in adolescence.

TRDM is an agentic construct introduced into criminology by Paternoster and Pogarsky (2009) that describes the decision-making process. To Paternoster and Pogarsky (2009), those high in TRDM make good decisions, where good decisions are defined as choices that are consistent with one’s preferences. Using data from The National Longitudinal Study of Adolescent Health (“Add Health”), four items were used to conceptualize TRDM (p. 113): (a) collecting information pertaining to a problem that requires a decision, (b) thinking of alternative solutions to the problem, (c) systematically deliberating over how to determine which alternative might be best, and (d) retrospectively analyzing how good a problem solver one was in the situation. TRDM is conceptually distinct from self-control, in that it is operationalized differently, and has a more restricted conceptual boundary because TRDM describes the process an individual uses to make a decision (Paternoster & Pogarsky, 2009; Paternoster, Pogarsky, & Zimmerman, 2011). As theorized, TRDM was found to be correlated with self-control and to be independently related to outcomes of interest such that those high in TRDM were more likely to graduate from college and less likely to be involved in criminal activities as adults (Paternoster et al., 2011).

The parameters of a dual-system thinking theory of criminal behavior have only recently been discussed in the literature (Paternoster & Pogarsky, 2009; Paternoster et al., 2011; Steinberg, 2010). For example, Steinberg (2010) argued that both rational and impulsive thinking co-exists as forces in human behavior, but they develop along different temporal paths. He suggests that a more fully developed impulsive thinking system and a later developing rational system of self-regulation explains why antisocial and risky behaviors peak during adolescence and decline thereafter. Using a diverse sample of nearly 1,000 respondents that ranged in age from 10 to 30 years of age, he found that self-reported impulsivity steadily decreased from ages 10-11 to 26-30 whereas reward seeking peaked in adolescence and steadily declined to young adulthood.

Thomas and McGloin (2013) used a dual-systems approach to explain susceptibility to peer influence. Peer perspectives in criminology have suggested that peer influence is based both on normative influence which concerns more long-term consequences of behavior and the more immediate influences of unsupervised socializing with peers. Appealing to dual-systems thinking models, Thomas and McGloin join these two perspectives by arguing that both systems may be at work in adolescents’ decisions to offend. Using two different data sets, they found partial support for their dual-systems position: The normative influence of deviant peers had a stronger effect on delinquency for those adolescents who were low on impulsivity than for those higher in impulsivity. This study shows the promise of dual-systems thinking for criminologists both in terms of reconciling divergent empirical findings and suggesting important and interesting new lines of research.

Van Gelder and de Vries (2014) presented a dual-process theory of offending that included what they called a propensity account that consists of both a “cool” cognitive component to decision making where decisions are made with effort and deliberation and an impulsive, affect strong “hot” component where decisions are the result of intuition and fast and easy information processing.⁶ In two separate studies using responses to questionnaire items and hypothetical scenarios, they found that both the cool cognitive component and the hot emotional component were related to criminal decisions. Interestingly, in the second of their scenario results, they found, in support of the dual-process model, that criminal choice in response to the hypothetical scenario context occurred through two independent mechanisms, one through a negative emotional state (a “hot” link) and the other through the perceived risk of punishment (a “cool” more cognitive link).

The Current Study

The dual-systems process model of decision making can reconcile the distance between trait-based and decision-based models in criminology as well as offer some clear hypotheses to test. The dual-systems approach would suggest that decisions to offend can be influenced by both impulsive and rational decision making. The former may likely have a direct effect on behavior whereas the latter may be more likely to have an indirect effect, via considerations of costs and benefits. In the empirical analyses that follow with a convenience sample of undergraduate students, we wish to provide only a very rudimentary test of a dual-systems model with some very simple initial hypotheses with our efforts hopefully spurring the field on to undertake more ambitious tests. The dual-systems model we examine in this article is illustrated in Figure 1.

Figure 1.

Illustration of dual-process model of offending.

At the outset, it would seem critical to determine whether there is an empirical distinction among concepts from the trait-based and decision-based models—temptation, impulsivity, and self-control/regulation. Our hypothesis would be as follows:

Hypothesis 1: Using factor analysis, there are separate though related latent factors for temptation, impulsivity, and self-control/regulation.

If impulsivity can be distinguished from self-control and both in turn distinguishable from temptation, then it would be useful to use these in conjunction with a measure of rational deliberation into a preliminary dual-process model and hypothesize the following:

Hypothesis 2: Measures of temptation, impulsivity, self-control/regulation, deliberative thinking (TRDM), and the perceived certainty of punishment will be directly related to the willingness to commit an offense. Each of these independent sources should have an independent effect on the willingness to offend (WTO).

Hypothesis 3: Measures of thoughtful thinking will also be indirectly related to a WTO through a measure of the perceived certainty of punishment. As a measure of rational deliberation and long-term consequence assessment, thoughtful reflection should be related to the perceived risk of punishment for a behavior, which in turn is related to a WTO.

Hypothesis 4: Self-control will act to restrain impulsive behavior, such that offending will be least frequent when impulsivity is low and self-control is high and most frequent when impulsivity is high and self-control is low.

Method

Sample

Our sample is a convenience sample of 141 college undergraduates taken from an introductory social science class at a major state university. The data collection instrument was passed out to all students present on the day of the administration and were asked to complete the questionnaire. There were no refusals, and missing data were limited. The sample ranged in age from 18 to 28 with a mean and median of 20 (SD = 1.42), about 47% of the sample was male, 63% Caucasian, 18% African American, 11% Hispanic, and 7% Asian or Other race. While we certainly realize the drawbacks in using a student sample, we think that given the preliminary nature of our empirical analysis, it seems prudent to begin simply.

Measures

We measured several constructs that would be relevant to a dual-process decision theory of crime. A two-item temptation summated scale was created with responses to the following questions: “I am able to resist temptation when I know there is work to be done” and “Meeting tomorrow’s deadlines and doing other necessary work comes before tonight’s play.” These items capture a trait of proneness to temptation rather than a situation-based measure of temptation felt at the moment. Impulsivity was measured with three items that frequently appear on psychological scales of the same concept: “I do things without thinking about them first,” “I make my mind up quickly,” and “I act on impulse.” Each of these items at face value seem to capture the notion of impulsivity as quick and easy action that is taken without much forethought. Self-regulation or self-control as we understand it as the capacity to impede an impulsive action is measured with three items: “It is easy for me to concentrate on things like reading a book,” “I find it hard to sit still for long periods of time,” and “I concentrate easily.” We operationalize self-control in this way because based on the work of Mischel (2014; see also Kahneman, 2011), concentration is needed to virtually break the spell that an impulse creates. In the various “marshmallow experiments” conducted by him and numerous others, those who are able to concentrate on other things/tasks/visual images are more likely to be able to resist the impulse of immediate gratification. We measure rationally deliberative thinking with the four items that comprise TRDM (Paternoster & Pogarsky, 2009): “When I have a problem to solve, one of the first things I do is get as many facts about the problem as possible”; “When I am attempting to find a solution to a problem, I usually try to think of as many different approaches to the problem as possible”; “When making decisions, I generally use a systematic method for judging and comparing alternatives”; and “After carrying out a solution to a problem, I usually try to analyze when went right and what went wrong.”

As part of the data collection instrument, we asked the students to respond to a hypothetical scenario:

Now, please imagine a hypothetical scenario in which you drove by yourself one night to meet some friends at a bar. The bar is located approximately 10 miles from your apartment. You have been casually drinking throughout the evening, and now you are ready to leave. You also remember that you have to be at work early the next morning, and your boss will have a fit if you are late. You can either drive home yourself or find another ride. However, if you find another ride tonight and leave your car, you will have to return early the next morning before work to pick it up.

In response, they were asked several questions that are relevant to the present study. First, as our measure of perceived certainty (Risk), they were asked to estimate on a scale from 0 (no chance) to 100 (a 100% chance) how likely it was that would be “pulled over” if they were to drive under the described condition. We also created a measure of the WTO (drive while intoxicated) by asking each student to estimate the probability (on the same 0-100 scale) how likely it was that they would drive their own car home under the circumstances described in the scenario.

Results

The first hypothesis was that the constructs of temptation, impulsivity, and self-control would comprise separate, though related, latent theoretical constructs. To examine this issue, we conducted a factor analysis of all eight individual items that comprise these three theoretical concepts. We estimated, using maximum likelihood with oblique oblimin rotation, a factor analysis. The results indicated that there were three distinct underlying latent constructs with the three impulsivity items loading on one factor (the factor loadings were .694, .514, and .441), the three self-control items loading on a second (loadings .569, .357, and .930), the two temptation items loading on the third (loadings .982 and .504). As expected, the three factors were interrelated: the correlation between impulsivity and self-control was −.38, between impulsivity and temptation was .21, and that between self-control and temptation was −.22.⁷ We then estimated a LISREL measurement model for all four theoretical constructs relevant to the dual-systems model under investigation here: self-control, impulsivity, temptation, and TRDM. The results (standardized solution) are reported in Figure 2. All of the factor loadings for the latent constructs are significant as are the correlations among the latent constructs. These results verify that impulsivity and self-regulation/control are conceptually and empirically distinguishable, consistent with Hypothesis 1 and with our contention from dual-systems theory (Kahneman, 2011) that self-control is not synonymous with impulsivity. It also is consistent with our conjecture that trait-based factors and decision-based factors are distinct though related constructs.

Figure 2.

Measurement model for dual-systems theory of offending.

The second and third hypotheses concern the relationship between the latent constructs for the dual-systems model—impulsivity, self-control, temptation, TRDM—and the decision to offend, in our case, the self-reported willingness to drive while possibly drunk in response to a hypothetical scenario. Hypothesis 2 states that each construct should have a direct effect on the WTO whereas the third hypothesis states that the effect of the rational component, TRDM, should also be indirect, through its effect on the estimated perceived certainty of getting pulled over for driving drunk. The results for these hypothesis tests are shown in Figure 3. There is only partial support for the dual-systems model here. The only constructs that have a direct and statistically significant effect on the willingness to drive home drunk (WTO) are impulsivity and perceived certainty of apprehension (Risk). As expected, those respondents who reported being more impulsive were significantly more likely to report that they would drive while intoxicated. Self-control, temptation, and TRDM had no significant effect on intentions to drink and drive. Moreover, contrary to Hypothesis 3, TRDM did not indirectly affect the decision to offend via its effect on the perceived certainty of punishment. Consistent with much of the deterrence and rational choice literature, however, the perceived risk of getting pulled over did have a significant inverse effect of the willingness to drink and drive—those who thought it was more likely that they would be pulled over were less likely to report that they would drive. At least with respect to one type of stable-individual trait, impulsivity, and one component of the decision-based model, perceived risk, the presumption from dual-process theory that these factors work together to influence risky behavior such as drinking and driving has provided at least some support.

Figure 3.

Structural model for WTO (drink and drive).

The fourth and final hypothesis takes seriously the notion that self-control is not synonymous with impulsivity but instead acts to regulate and override one’s impulsive actions when necessary. If so, then these two factors would interact such that self-control would lead to a reduced level of involvement in risky behavior when impulsivity is high but would be much less salient when impulsivity is low. To examine this issue, we dichotomized both the impulsivity and self-control scales⁸ and from this created a four-category taxonomy: low impulsivity/high self-control, low impulsivity/low self-control, high impulsivity/high self-control, and high impulsivity/low self-control. We would predict that involvement in risky actions would be highest in the first of these where there is a highly impulsive person with low self-control and lowest in the last where there is a person with low impulsivity and high self-control. Predictions for the middle two categories are difficult to make a priori and would likely depend upon the magnitude of the relationship to the risky behavior of each. We test this hypothesis using our measure of self-reported intentions to drink and drive.

We test this hypothesis with an ANOVA, first using all respondents and then as a sensitivity test, we ran a second ANOVA where we excluded respondents who reported a zero probability of drinking and driving in response to the scenario. The mean values for the willingness to drink and drive are reported in Table 1 (along with the F test), and Figures 4 and 5 provide a mean plot for the values. First, when all respondents are included, results are consistent with Hypothesis 4. The highest mean reported probability of offending is highest for persons who were high in impulsivity and low in self-control (36.08). For these respondents, there is very weak self-regulation and as a result, the probability of risky behavior such as drinking and driving is very high. The lowest mean probability of offending was for those respondents that were low in impulsivity and high in self-control (17.05), and the difference between these two means is significantly different from zero. This pattern is exactly duplicated when we excluded those who said that there was a zero probability of drinking and driving in response to the described scenario conditions. Suggestive of the strong effect that impulse can have on risky behaviors, among all respondents and just those with a non-zero probability of offending, there was a higher mean probability of offending for the high impulsivity/high self-control group compared with those low in both impulsivity and self-control. For both all respondents and the zeros excluded, when impulsivity was high, risky behavior such as drinking and driving was less likely when self-control was also high and could offer some resistance, but because of our small sample sizes, this difference was not statistically significant. What is quite clear, however, is that a personal trait like a propensity to act on impulse matters a great deal for risky behavior and that one’s ability to self-control or self-regulate helps to curb that propensity.

Table 1.

Results of Interaction Between Impulsivity and Self-Control on the Reported Willingness to Drink and Drive.

	n	$\bar{X}$	SD	F
All respondents
High impulsivity/low self-control	38	36.08^a	35.66	2.858*
Low impulsivity/low self-control	28	21.89	26.46
High impulsivity/high self-control	34	26.18	29.26
Low impulsivity/high self-control	40	17.05^a	25.10
Zeros excluded
High impulsivity/low self-control	27	50.78^a	32.16	3.107*
Low impulsivity/low self-control	20	30.65	26.69
High impulsivity/high self-control	23	38.70	27.89
Low impulsivity/high self-control	24	28.42^a	27.05

Pair of means that are significantly different from each other in a Tukey’s Honest Significant Difference Test.

p < .05.

Figure 4.

Mean plot for intensions to drink and drive—All respondents.

Figure 5.

Mean plot for intensions to drink and drive—Zeros excluded.

Discussion

The history of criminological theory has been a movement along two separate and distinct paths. One path consists of trait-based theories which posit that crime is due to time-stable between-individual differences in some underlying propensity to offend. For years, one specific form this propensity has taken is impulsivity—offenders were presumed to be oriented to the present and seduced by the immediate temptations of criminal actions while unmindful of their long-term costs. Recently, the impulsivity position was superseded by Gottfredson and Hirschi’s (1990) concept of low self-control. In fact, in perusing the criminological literature since the publication of their book, it seems clear that low control has been taken to be synonymous with impulsivity both conceptually and empirically. The second path has been represented by choice-based theories such as deterrence, routine activities, and rational choice which claims that behavior is not driven by impulse but by the rational and thoughtful calculation of the costs and gains of crime. In this article, we suggest a reconciliation of these two positions through a dual-process model of behavior that has been well-described in psychology and economics (Camerer, 2007; Kahneman, 2011; Stanovich, 2011). In dual-process models, behavior is argued to be the result of both quick, easy, and impulsive decision and more rational, thoughtful, and effortful cognitive decisions. In other words, rather than opposites, impulse, affect, intuition, and deliberate thought processes act together. Furthermore, in dual-process models, low self-control is not the presence of impulsivity, rather it is taken as the capacity to override impulsive actions in large measure by a more careful consideration of the consequences of the action, both short-term and long-term (Baumeister, 2002). In dual-process thinking, then, self-control is synonymous with self-regulation not impulsivity.

With these ideas as our framework, we derived and tested a series of hypotheses with an admittedly convenience sample of college undergraduates and found partial support for a dual-process model of criminal decision making. First, in a factor analysis, we found that impulsivity and self-control are distinct but interrelated constructs that are separate from each other. We also found that these were not only independent from but also related to temptation and TRDM. We had hypothesized that these four constructs would have a direct effect on risky behavior, in this case, the intention to drink and drive in response to a hypothetical scenario, and that as part of the more rational and cognitive component of dual-process theory that TRDM would have an indirect effect on risky intentions via its effect on the perceived certainty of getting caught for drinking and driving. We found only partial support here. Impulsivity was positive and significantly related to intentions to drink and drive, but temptation, self-control, and TRDM had no effect on intentions. The perceived risk of getting pulled over for drinking and driving was also significantly related to intentions, but TRDM had no relationship with perceived risk. It is not surprising perhaps that self-control had no effect on intentions because as we have conceptualized it, self-control reflects the ability to concentrate and forestall impulsive actions, as such it should work in tandem with impulsivity. With respect to this, we found that the risk of drinking and driving was highest for those respondents who were highly impulsive and had low self-control and lowest for those who were low in impulsivity and high in self-control. Having high self-control did act to diminish the probability of drinking and driving intentions when impulsivity was also high, but because of a small sample size, this difference was not statistically significant. In this sense, we do find evidence for dual-process models, in that there is an effect for both impulsivity and perceived risk on WTO, which is consistent with the argument put forth by van Gelder and de Vries (2014).

We hope the results of our study encourage interest in dual-process theory in criminology. We think it holds promise as a way to reconcile what in the past have been competing perspectives on the causes of crime. Clearly, we need to know a great deal more about how the two systems of thinking work and interact together. The notion of dual cognitive processes is as old as Plato’s allegory of the human soul in his dialog Phaedrus as a charioteer driving two horses. In this metaphor, one horse represents the soul’s good passions or moral instincts whereas the other represents the soul’s demanding appetites and passions. The charioteer represents reason which must guide the horses, careful to keep the dark horse in check. Within recent years, important research has done a great deal to illustrate the processes behind intuitive or passion-driven thinking and more thoughtful thinking, even identifying their neurobiological sources. Criminologists can make their own contributions to this literature by examining how dual-process thinking works for antisocial behavior. We also hope that we can get the field to begin to think differently about the notion of self-control as the capacity to restrain impulsive behavior. For too long, impulsivity has been used as simply another word for impulsive behavior. Although the role of impulsivity in risky and criminal behavior is now well-known, and well-documented, criminologists have had very little to say about how these impulses can be curbed. We have suggested that self-control should be thought of as self-regulation, the ability to short-circuit impulsive demands for immediate gratification through concerted thoughts about the long-term consequences of one’s behavior or as in the marshmallow experiments concentrating on other things. It seems to us time for criminologists to conduct some marshmallow-like experiments of their own.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

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

Notes

Author Biographies

Chae Mamayek is a doctoral student in the Department of Criminology and Criminal Justice at the University of Maryland. Her research interests include offender decision-making, quantitative methodology, deterrence and public policy.

Tom Loughran is an associate professor in the Department of Criminology and Criminal Justice at the University of Maryland. His research interests include offender decision-making and deterrence, illegal market participation and quantitative methods.

Ray Paternoster is professor in the Department of Criminology and Criminal Justice at the University of Maryland. His research interests include rational choice theory, offender decision-making, and issues related to capital punishment.

References

Albert

Steinberg

(2011). Judgment and decision making in adolescence. Journal of Research on Adolescence, 21, 211-224.

Baumeister

R. F.

(2002). Yielding to temptation: Self-control failure, impulsive purchasing, and consumer behavior. Journal of Consumer Research, 28, 670-676.

Baumeister

R. F.

Heatherton

T. F.

Tice

D. M.

(1994). Losing control: How and why people fail at self-regulation. San Diego, CA: Academic Press.

Baumeister

R. F.

Tierney

(2012). Willpower: Rediscovering the Greatest Human Strength. New York, NY: Penguin.

Becker

(1968). Crime and punishment: An economic approach. Journal of Political Economy, 76, 169-217.

Bentham

(1948). An introduction to the principles of morals and legislation. New York, NY: Hafner. (Original work published 1789)

Block

J. H.

Keyes

(1988). Longitudinally foretelling drug usage in adolescence: Early childhood personality and environmental predictors. Child Development, 59, 336-355.

Camerer

(2007). Neuroeconomics: Using neuroscience to make economic predictions. The Economic Journal, 117(519), C26-C42.

Caspi

Henry

McGee

R. O.

Moffitt

T. E.

Silva

P. A.

(1995). Temperamental origins of child and adolescent behavior problems: From age three to age fifteen. Child development, 66, 55-68.

10.

Cleckley

H. M.

(1941). The mask of sanity. St. Louis, MO: Mosby.

11.

Eisenberg

Fabes

R. A.

(1992). Emotion and its regulation in early development. San Francisco, CA: Jossey-Bass.

12.

Evans

J. S. B.

Stanovich

K. E.

(2013). Dual-process theories of higher cognition advancing the debate. Perspectives on Psychological Science, 8, 223-241.

13.

Eysenck

H. J.

(1977). Crime and personality. London, England: Routledge.

14.

Eysenck

H. J.

Gudjonsson

G. H.

(1989). The causes and cures of criminality. New York, NY: Plenum.

15.

Farrington

D. P.

Loeber

Van Kammen

W. B.

(1990). Long-term criminal outcomes of hyperactivity-impulsivity-attention deficit and conduct problems in childhood. In Robins

L. N.

Rutter

(Eds.), Straight and devious pathways from childhood to adulthood (pp. 62-81). New York, NY: Cambridge University Press.

16.

Frederick

(2002). Cognitive reflection and decision making. Journal of Economic Perspectives, 19(4), 25-42.

17.

Goddard

H. H.

(1914). Feeblemindedness: Its causes and consequences. New York, NY: Macmillan.

18.

Gottfredson

(2006). The empirical status of control theory in criminology. In Cullen

F. T.

Wright

J. P.

Blevins

K. R.

(Eds.), Taking stock: The status of criminological theory (pp. 77-100). New Brunswick, NJ: Transaction.

19.

Gottfredson

Hirschi

(1990). A general theory of crime. Stanford, CA: Stanford University Press.

20.

Grasmick

H. G.

Tittle

C. R.

Bursik

R. J.

Arneklev

B. J.

(1993). Testing the core implications of Gottfredson and Hirschi’s general theory of crime. Journal of Research in Crime & Delinquency, 30, 5-29.

21.

Hay

(2001). Parenting, low self-control, and delinquency: A test of self-control theory. Criminology, 39, 707-736.

22.

Henry

Caspi

Moffitt

T. E.

Harrington

Silva

P. A.

(1999). Staying in school protects boys with poor self-regulation in childhood from later crime: A longitudinal study. International Journal of Behavioral Development, 23, 1049-1073.

23.

Henry

Caspi

Moffitt

T. E.

Silva

P. A.

(1996). Temperamental and familial predictors of violent and nonviolent criminal convictions: Age 3 to age 18. Developmental Psychology, 32, 614-623.

24.

Hirschi

(2004). Self-control and crime. In Baumeister

R. F.

Voss

K. D.

(Eds.), Handbook for self-regulation: Research, theory, and applications (pp. 537-552). New York, NY: Guildford Press.

25.

Hirschi

Gottfredson

(1994). The generality of deviance. In Hirschi

Gottfredson

M. R.

(Eds.), The generality of deviance (pp. 1-22). New Brunswick, NJ: Transaction.

26.

Hooton

(1939). Crime and the man. Cambridge, MA: Harvard University Press.

27.

Kahneman

(2003). A perspective on judgment and choice: Mapping bounded rationality. American Psychologist, 58, 697-720.

28.

Kahneman

(2011). Thinking, fast and slow. New York, NY: Farrar, Straus and Giroux.

29.

Loeber

Menting

Lynam

D. R.

Moffitt

T. E.

Stouthamer-Loeber

Stallings

. . .Pardini

(2012). Findings from the Pittsburgh youth study: Cognitive impulsivity and intelligence as predictors of age-crime curve. Journal of the American Academy of Child & Adolescent Psychiatry, 51, 1136-1149.

30.

Loewenstein

Rick

Cohen

J. D.

(2008). Neuroeconomics. Annual Review of Psychology, 59, 647-672.

31.

Lombroso

(1876). The criminal man (L’uomo Delinquente). First Edition. Milan: Hoepli.

32.

Lynam

Wikström

Caspi

Moffitt

T. E.

Loeber

Novak

S. P.

(2000). The interaction between impulsivity and neighborhood context on offending: The effects of impulsivity are stronger in poorer neighborhoods. Journal of Abnormal Psychology, 109, 563-574.

33.

Mischel

(2014). The Marshmallow Test: Mastering self-control. New York, NY: Little Brown.

34.

Moffitt

T. E.

(1993). Adolescence-limited and life-course persistent antisocial behavior: A developmental taxonomy. Psychological Review, 100, 674-701.

35.

Moffitt

T. E.

Caspi

Rutter

Silva

P. A.

(2001). Sex differences in antisocial behavior. New York, NY: Cambridge University Press.

36.

Paternoster

Pogarsky

(2009). Rational choice, agency, and thoughtfully reflective decision making: The short and long-term consequences of making good choices. Journal of Quantitative Criminology, 25, 103-127.

37.

Paternoster

Pogarsky

Zimmerman

(2011). Thoughtfully reflective decision making and the accumulation of capital: Bringing choice back in. Journal of Quantitative Criminology, 27, 1-26.

38.

Pratt

T. C.

Cullen

F. T.

(2000). The empirical status of Gottfredson and Hirschi’s general theory of crime: A meta-analysis. Criminology, 38, 931-964.

39.

Prichard

J. C.

(1837). Treatise on insanity and other disorders affecting the mind. London, England: Sherwood, Gilbert, and Piper.

40.

Pulkinnen

(1986). The role of impulse control in the development of antisocial and prosocial behavior. In Olweus

Block

Radke-Yarrow

(Eds.), Development of antisocial and prosocial behaviors: Theory, research, and issues (pp. 149-175). New York, NY: Academic Press.

41.

Rafter

(2004). The unrepentant horse-slasher: Moral insanity and the origins of criminological thought. Criminology, 42(4), 979-1008.

42.

Reynolds

(2006). A review of delay-discounting research with humans: Relations to drug use and gambling. Behavioral Pharmacology, 17, 651-667.

43.

Simon

(1957). Models of man. New York, NY: John Wiley.

44.

Stanovich

K. E.

(1999). Who is rational? Studies of individual differences in reasoning. Mahwah, NJ: Lawrence Erlbaum.

45.

Stanovich

K. E.

(2011). Rationality and the reflective mind. New York, NY: Oxford University Press.

46.

Steinberg

(2010). A dual-systems model of adolescent risk-taking. Developmental Psychobiology, 52, 216-224.

47.

Thomas

K. J.

McGloin

J. M.

(2013). A dual-systems approach for understanding differential susceptibility to processes of peer influence. Criminology, 51, 435-473.

48.

Van Gelder

de Vries

R. E

. (2012). Traits and states: Integrating personality and affect into a model of criminal decision making. Criminology, 50, 637-671.

49.

van Gelder

de Vries

R. E

. (2014). Rational misbehavior? Evaluating an integrated dual-process model of criminal decision making. Journal of Quantitative Criminology, 30, 1-27.

50.

Wilson

J. Q.

Herrnstein

R. J.

(1985). Crime and human nature. New York, NY: Touchstone.

51.

Yochelson

Samenow

S. E.

(1976). The criminal personality. Volume I: A profile for change. Lanham, MD: Rowman & Littlefield.