Skip to main content
SearchLoginLogin or Signup

Peeking into the Black Box of Offender Decision-Making: A Novel Approach to Testing Situational Action Theory’s Perception Choice Process

Published onJan 21, 2025
Peeking into the Black Box of Offender Decision-Making: A Novel Approach to Testing Situational Action Theory’s Perception Choice Process
·

Abstract

Situational Action Theory (SAT) and its perception choice process (PCP) explain how individual traits and setting characteristics interact to produce criminal behavior. It does so by reference to the cognitive mechanisms linking causally relevant variables and their interactions to behavior. Testing this model poses considerable methodological challenges, requiring data that captures who is engaging in what behavior in what kind of setting, and provides insights into the decision-making process. This study uses an evolution of the experimental vignette methodology, integrating data about person and setting characteristics as well as participants’ decision processes to address these challenges. A diverse sample of adults living in the UK (N = 595) was recruited via the online recruitment platform Prolific. The focal behavior under investigation was theft by finding. Our findings mostly support our predictions, while those that contradict the theory offer new insights into ongoing theoretical debates. The study further demonstrates the viability of the developed methodological approach.

1. Introduction

Why do people engage in acts of crime? To answer this question researchers in criminology have mostly concerned themselves with identifying a myriad of characteristics of people and places (and more rarely their interactions) that are associated with crime (Ellis, Farrington, and Hoskin 2019). Much less attention has been devoted to the mental mechanisms (processes) that link the influence of person and setting characteristics and their interactions to crime (Wikström et al. 2012; Wikström and Kroneberg 2022). An exception to this is Situational Action Theory (Wikström 2006; Wikström et al. 2012; Wikström and Treiber 2007). It emphasizes that acts of crime result from the convergence and interaction of a person’s criminal propensity (needs and sensitivities, personal morality and ability to exercise self-control) and the setting’s criminogeneity (opportunities, frictions, and the moral context comprising moral norms and their enforcement). An interplay among these factors determines whether the individual enters a motivational state conducive to crime, whether they perceive crime as an action alternative, and whether processes such as self-control or deterrence play a part in individuals’ decision-making and if so, whether these processes will succeed in preventing the individual from engaging in crime. Thus, SAT makes predictions not only about input-output associations, but also allows us to derive hypotheses about some of the key intermediate stages of the decision-making process.

Accordingly, testing such a theory requires data that captures individual differences, setting characteristics, behavioral outcomes of person-setting convergences as well as data about the decision processes (Herrmann 2024). However, the status quo of testing the PCP consists of modelling the relationships between person and setting factors, their interaction, and behavior, while ignoring the mental processes that give rise to action. In the cognitive and decision making sciences, it is widely accepted that adequate tests of process models require not only data about the inputs (i.e., person and setting characteristics) and outputs of decision making (i.e., behavior) but also data pertaining to the decision process itself (Ford et al. 1989; Johnson, Schulte-Mecklenbeck, and Willemsen 2008; Schulte-Mecklenbeck, Kühberger, and Johnson 2019; Schulte-Mecklenbeck, Kühberger, and Ranyard 2011). While some criminologists have begun developing ways to gain insights into criminal decision processes (e.g., Brauer and Tittle 2017; Eifler 2015; Herrmann 2024; Verhoef et al. 2021, 2022, 2023), these approaches are still in their infancy.

To advance this line of research, the present study introduces an evolution of the traditional experimental scenario method. Where prior research established which inputs (into the PCP) make criminal action (output) more likely, we try to sequence the PCP and examine its stages: if the offender deliberated or acted automatically, what action alternatives he or she perceived, and what the content of deliberation was. Only this way can we uncover the mechanisms leading individuals to engage in or abstain from acts of crime. Note that SAT makes concrete predictions about the mental mechanisms linking inputs to outputs, yet most applications of the theory omit all that happens between inputs and outputs (inside the decision-making process). The main advantage of the approach taken here is that it provides data about who makes what decision in what setting, while also offering insights into individuals’ decision processes, hence allowing for a more adequate test of the PCP.

2. Situational Action Theory’s Perception Choice Process

SAT’s predictions about the PCP detail how the coming together of person propensities and setting characteristics gives rise to a mental process, which ultimately moves individuals to engage in acts of crime. It consists of three consecutive steps; motivation, the perception of action alternatives, and automatic versus deliberative choice (see Figure 1).

2.1 Motivational states

Whether an individual enters a motivational state that is conducive to crime (temptation or provocation) depends on their desires and sensitivities and the potential opportunities and frictions the setting provides (the circumstance; Wikström 2006). For instance, if a setting offers an opportunity to acquire a good that is desired, a temptation will arise. These motivational states are best characterized as goal-directed attention (Wikström 2017; Wikström and Treiber 2015). In other words, a circumstance captures the attention of the individual because it is of relevance to them. Importantly, however, temptations and provocations alone neither prescribe a criminal action alternative nor do they automatically make the individual consider engaging in crime. Motivation is thus a necessary but ultimately not sufficient condition for crime to occur.

Figure 1. Situational Action Theory’s (SAT) Perception Choice Process (PCP), adapted from Herrmann (2024)

While SAT sees crime-relevant motivation a necessary condition for perceiving crime as an action alternative (it makes the goal salient), the theory remains somewhat vague with regards to the influence of the motivational strength on deliberative choice and does not make predictions regarding the perceptions of action alternatives. Are people more likely to perceive deviant action alternatives for satisfying a motivation when that motivation is stronger? Are deliberations about engaging in crime more likely to lead the individual to engage in crime when motivations are stronger? Apart from evidencing that the intensity of the motivational trigger increases the likelihood of crime engagement (Wikström et al. 2012), existing research has mostly ignored the influence of motivation in tests of the PCP, that is most often, motivation is taken for granted. Given this lack of theoretical precision and empirical work, this study explores what influence the strength of individuals’ motivations might have on the perception of action alternatives as well as its influence on deliberative choice.

2.2 The perception of action alternatives

A central concept in SAT is the perception of action alternatives. The theory states that only some individuals who are tempted/provoked will perceive crime as an action alternative. This largely depends on the strength of their personal morality (i.e., commitment to the official moral rule expressed in moral beliefs and moral emotions) and their perceptions of the strength of the moral norms of the context (Wikström et al. 2012; Wikström and Treiber 2007). SAT’s principle of moral correspondence states that strong law-relevant personal morality1 combined with strong (perceived) law-relevant moral norms2 will lower the likelihood that the individual will perceive crime as an action alternative (Wikström et al. 2012). In that case, they will not engage in crime and will also not deliberate about engaging in crime. Alternatively, when both the individual’s personal morality and the perceived moral norms of the setting in which they find themselves are weak, the individual will be more likely to perceive crime as the only action alternative and proceed to engage in crime without contemplating (deliberating about) crime engagement. A third option is that the individual deliberates about engaging in crime, that is they perceive crime as an action alternative, but they are not entirely sure whether it is the best course of action. According to the theory, deliberation about engaging in crime is most likely under conditions of moral incongruency/conflict, that is when either personal morality or the law-relevant perceived moral norms of the context are weak while the other is strong.

However, this formulation which focuses on moral conflict leaves unaddressed the question as to how intermediate levels of personal morality and the (perceived) strength of the moral norms relate to the perception of action alternatives as well as well as automatic and deliberative choice (De Buck and Pauwels 2022). We propose that deviations from congruent strong moral rule guidance (strong personal morality and strong perception of the moral norms) as well as deviations from weak congruent moral rule guidance (weak personal morality and weak perceptions of the moral norms) predict the engagement of deliberative crime-related processing (see Figure 2 for a comparison of the standard formulation and the more nuanced formulation). In other words, it is not the incongruency of the strength of personal morality and the perceived moral norms of the setting, but their (overall) weakening relative to strong congruent moral rule guidance, or their strengthening relative to weak congruent moral rule guidance, which increase the likelihood of deliberations about engaging in crime. This more nuanced formulation thus captures cases of weak and strong moral congruency, moral conflict/incongruency, but also intermediate values of personal morality and the (perceived) moral norms of the setting.

Figure 2. Two Formulations of the principle of moral correspondence

Note: red = only crime perceived; green = crime not perceived; yellow = deliberation

Beyond the influence of personal morality and the moral norms of the context, it may be plausible that the strength of the motivation, ceteris paribus, also increases the likelihood that the individual perceives crime as an action alternative. Consider a person who is desperate to get urgent medical treatment for a loved one. While the same person may not perceive bribing the doctors for a sooner treatment in most circumstances,3 given the dire need for treatment bribing the doctor now becomes an option. SAT merely states that the perception of alternatives refers to morally acceptable ways to respond to the particular motivation under the particular circumstance (Wikström, Treiber, and Roman 2024). However, it does not provide sufficiently specific and therefore testable predictions in that regard. Given this lack of theoretical precision and empirical work, this study explores what influence the strength of individuals’ motivations might have on people’s perceptions of action alternatives.

2.3 Self-control abilities, internal and external control processes

SAT’s principle of the conditional relevance of controls states that an individual’s self-control abilities matter for the decision to engage in crime only if the individual enters the deliberative stage . A person’s abilities to exercise self-control are grounded in their executive functioning. Executive functions allow the individual to create and maintain strong mental representations of their personal morality as well as the potential negative consequences of engaging in crime (including the likelihood and severity of formal and informal sanctions) when they deliberate (Treiber 2011, 2017; Wikström and Treiber 2007). If sufficiently strong, these mental representations may allow the individual to override any perceived criminal action alternative in favor of an alternative that is in line with their rule abiding internal moral standards (when exercising self-control) or the official moral rule (when being deterred).

It is important to unambiguously define the concepts of the exercise of self-control and deterrence as they are used in the context of SAT. The theory submits that if an individual engages in internal moral deliberations and they adhere to their (strong) personal morality, they succeed in exercising self-control (Wikström and Treiber 2007). If this process fails to inhibit a criminal response this is referred to as a failure to exercise self-control, despite self-control processes being at play. If the individual considers the potential negative consequences of the criminal action alternative and manages to inhibit a perceived criminal action, the individual is said to have been deterred. In contrast, if they deliberate about negative consequences of the criminal action alternative, but engage in crime nevertheless, deterrence has failed despite deterrence related processes being relevant for the decision. By these definitions, an individual who does not perceive crime as an action alternative and therefore does not engage in crime has neither been deterred nor did they exercise self-control, since they did not deliberate. On the flipside, for someone who perceives crime as the only action alternative and engages in crime, neither deterrence nor a process of self-control is relevant, and therefore these individuals do not fail to exercise self-control and neither does deterrence fail for them. Control processes thus become relevant when the individual engages in some form of moral management whereby they assess which course of action is the most acceptable by internal and/or externally presented moral standards.

The theory further proposes that and individual’s perceptions of the likelihood and severity of negative consequences are only relevant when individuals perceive crime as an action alternative and deliberatively take into consideration potential negative consequences (Wikström 2006). Importantly, the theory distinguishes between the moral norms of the setting (relevant for the perception of action alternatives) and the deterrent threats (relevant provided the individual deliberates about negative consequences of engaging in crime). However, it is worth highlighting here, that while these concepts may be distinct in theory, they may often be closely related empirically (Hardie 2017; Hardie and Rose 2025). Norm cues such as what other people are doing (descriptive norms) and cues communicating what is considered acceptable (injunctive norm; Cialdini, Kallgren, and Reno 1991), may also impact individuals’ perceptions of the likelihood and severity of negative consequences (Bicchieri 2006; Fishbein and Ajzen 2010). If everyone is breaking the rule (descriptive norm), then this may also signal a low probability of getting caught, or that the consequences are unlikely to be severe. Similarly, if most people do not regard breaking the rule as reprehensible (weak injunctive norm), then why would potential observers or society at large (help) enforce the rule or impose informal sanctions? Further, classical deterrent cues (e.g., the presence of security cameras, police officers, and other individuals who may enforce the rules of the setting) are likely to raise the norm salience to individuals in their presence. The concept of norm salience refers to the prominence of a given norm in a specific setting (Cialdini et al. 1991). For instance, the presence of a police officer who represents the law and its authority not only influences the likelihood of facing negative consequences for breaking the law but also makes the moral rule explicit (salient) in the setting. This is an issue which we will return to in the methods section.

By SAT’s definitions, internal and external control processes may feature in an individual’s behavioral choice only if they perceive crime as an action alternative and they are unsure at first whether crime represents their best behavioral option (Wikström et al. 2012). The defining feature of the internal control process is an internal moral conflict (moral qualms or deliberations about what is by their own moral standard the right thing to do). The external control process features deliberations about negative consequences (the likelihood and severity of negative consequences).

Furthermore, among those who deliberate, the strength of personal morality and the strength of the perceived moral context will influence whether their deliberations will include an internal moral deliberation (process of self-control) or deliberations about the negative consequences of rule breaking (process of deterrence). While strong personal morality increases the chance of internal moral deliberations, strong perceptions of the moral context should increase the likelihood of deliberations about negative consequences amongst deliberators.

3. Key methodological considerations

What makes a theory testable is not only determined by how well the concepts are defined and how precise its predictions are, but also by the availability of adequate methodological tools. Whereas SAT provides ample opportunity to derive very specific empirical predictions (Wikström and Kroneberg 2022), the existing methodological landscape in criminology falls short of providing satisfactory tools to test many of its key predictions (for more detail see Herrmann 2024). This section discusses some of the most important methodological challenges, which include (a) situational analysis, (b) the limitations and weaknesses of the increasingly popular experimental vignette scenario method, and (c) the importance of gathering data pertaining to the decision process.

3.1 Situational analysis

A widely practiced approach to testing predictions of the PCP is based on cross-sectional data consisting of crime relevant individual differences, a proxy measure for criminogenic exposure (e.g., whether the individual has crime prone peers) and an aggregate of how many times in a given time period someone has engaged in criminal behavior (e.g., Hirtenlehner and Hardie 2016; Hirtenlehner and Leitgöb 2021; Kroneberg and Schulz 2018; Schepers and Reinecke 2018). In these studies, the unit of analysis is the individual. While researchers may be able to evidence that certain types of people who are likely to spend more time in certain types of settings (or have certain perceptions of the likelihood of getting caught for certain crimes and certain expectations of the severity of punishment) engage in crime more often, this insight cannot be taken for proof that certain types of people commit crime in certain types of settings (Wikström, Mann, and Hardie 2018). Reasoning of this sort constitutes an ecological fallacy whereby relationships evidenced at the aggregate level are expected to hold at the level of the individual units (situations) of which the aggregate is comprised (Hardie 2020). Individuals are (more or less) free to commit their crimes wherever they like. Even those who spend most of their time in strong moral contexts (or moral contexts they perceive to be strong) may venture into weak moral contexts to commit all their crimes. Hence, “before we even approach statistical estimation [of interaction effects between person and setting characteristics], the research design determines the extent to which our data capture how the convergence of individuals and settings lead to crime” (Wikström and Kroneberg 2022:190).

A theoretical model which explains why and how people choose to break or follow rules requires situational data. The situation in SAT is the result of the convergence of an individual, with their crime propensity, and the setting, with its criminogenic inducements. Situational data hence spatiotemporally links individual differences and setting characteristics (or the individual’s perceptions thereof) with the behavior their interaction presumably causes. In other words, situational data captures who is engaging in what action in what type of setting which allows for the testing of how these factors interact (Hardie, 2020).

The most popular method to capture such spatiotemporally linked data is the experimental vignette scenario method. Experimental scenarios are embedded in online surveys as text scenarios or more rarely as video scenarios (e.g., Craig 2019; De Buck and Pauwels 2022; Eifler 2015; Sattler et al. 2022; Wikström et al. 2012). The standard approach to testing the PCP using experimental vignette scenarios, involves three steps. In step one, researchers employ measurements of theoretically relevant personal characteristics. In step two, participants are randomly allocated to one experimental condition and are instructed to read or view one scenario permutation of the total scenario universe (all possible scenario permutations). The two or more experimental conditions and their corresponding scenario permutations reflect manipulated elements of the scenario such as the intensity of the potential motivator (e.g., Eifler, 2015; Wikström et al., 2012), the moral norm (e.g., Sattler et al., 2022), the likelihood or severity of negative consequences (e.g., Pauwels, 2018; Sattler et al., 2022), or the strength of the moral context (moral norms and their enforcement; Wikström et al. 2012). In step three, participants are asked how they would behave in such a scenario if this happened to them in real life. The gathered data is then used to test hypotheses reflecting predictions about what types of people commit crime in what types of settings.

3.2 Limitations and challenges of the vignette scenario approach

While existing vignette scenario studies allow researchers to capture situational data, the approach also has its weaknesses. Firstly, written scenarios are limited in the extent to which they provide a realistic account of highly complex and information-rich real-world environments (Bouffard and Niebuhr 2017). This may lead some participants to impute their own details about the scenario (e.g., the availability of an escape route), which weakens experimental control (Bouffard 2002; Exum and Bouffard 2010; Klepper and Nagin 1989). They may also feel compelled to form a behavioral intention based on insufficient information in the face of missing contextual cues (Exum and Bouffard 2010), which may result in high levels of uncertainty on behalf of participants when asked about their behavioral intentions. Researchers may attempt to counter the issue of missing contextual information by including a larger amount of information in the text description, however this may overburden participants’ attention span and imaginative abilities (van Gelder et al. 2019).

Another weakness of text vignettes is that they provide the same selection of contextual information to all participants, whereas in real-world contexts individuals perceive their environment differently because they selectively attend to personally relevant details (van Gelder et al. 2019). Relatedly, the manipulation of variables of interest by the researcher may produce unnatural saliency and priming effects (Bouffard 2002), whereby the participant’s perception of action alternatives or deliberative choice are influenced by the description of the presence of a factor in the context, whereas the same individual may not take notice of this factor in a real-world setting.

Moreover, in its current form the approach fails to provide much insight into the perception of action alternatives as well as decision styles (automatic vs. deliberative) and deliberative content (internal moral conflict vs. negative consequences). If a study subject chooses not to engage in crime, we cannot know whether it was due to non-perception of crime as an action alternative or due to a process of self-control or deterrence. Testing hypotheses about the intermediate stages of a decision process requires data pertaining to the decision process (Herrmann 2024).

3.3 The need for process data

What sets SAT and its PCP apart from most theories of crime is that it spells out the mental mechanism (process) that links person-setting interactions to criminal or non-criminal behavior. However, with few exceptions researchers have relied on a structural approach to testing the theory that largely ignores decision processes. This approach consists of deriving hypotheses that reflect the predictions about the relationship between input variables (person and setting characteristics) and their interactions on the output variable (behavioural intention; e.g., Craig 2019; De Buck and Pauwels 2022; Pauwels 2018a, 2018b; Wikström et al. 2012). This approach has two distinct but related weaknesses.

Firstly, input-output testing relies on imprecise hypotheses which to some extent obscure more nuanced theoretical predictions (Herrmann 2024). Consider the hypothesis that self-control abilities influence the decision to engage in crime only (or more) amongst individuals facing incongruent moral rule guidance (strong personal morality and weak moral norms of the setting or weak personal morality and strong moral norms of the setting). On theoretical grounds this hypothesis is not incorrect but rather imprecise. This is because the hypothesis is a compound of two separate predictions; 1) individuals with incongruent moral rule guidance are more likely to deliberate, and 2) self-control abilities matter for the decision to engage in crime only amongst those who deliberate. If the researcher intent on testing an input-output relationship finds that the data does not support their hypothesis they will be left to wonder which of the two (or both of the) more nuanced predictions is wrong. Moreover, a considerable number of individuals with incongruent moral rule guidance may not deliberate about engaging in crime despite being more likely to do so than individuals with congruent moral rule guidance4. This imperfect relationship between moral (in)congruency and deliberative processing is likely to significantly lower the effect size of the of the interaction between the (in)congruency of moral rule guidance and self-control abilities, hence lowering the ability of the researcher to find support for their compounded hypothesis. Therefore, researchers should strive to test the two predictions independently, which requires collecting data that distinguishes between those who deliberate about engaging in crime and those who do not.

Secondly, if researchers do find the hypothesized input-output relationships, they tend to explain these findings in terms of the mental mechanisms that are presumed to have given rise to the observed outcomes in a post-hoc fashion. While these explanations may be theoretically plausible, the researcher cannot ascertain whether their observations came about by the mental processes set out in the theory. For instance, if a study finds that individuals who face law-abiding moral rule guidance (strong personal morality and strong moral norm of the setting) reliably indicate that they would not engage in crime in a scenario described to them, then this does not provide evidence that these individuals did not perceive crime as an action alternative as predicted by the theory. In the absence of evidence about how individuals make decisions, some may argue that a careful, deliberative weighing of the pros and cons of several perceived action alternatives could have produced the same outcome. Details of the mental process thus cannot be inferred from input variables and their relationship with the behavioral outcome alone.

To date, only two published experimental text vignette studies testing SAT’s predictions have attempted to gain measured insights into participant’s decision making processes (Brauer and Tittle 2017; Eifler 2015). Brauer and Tittle (2017) asked participants to imagine a scenario in which a stranger insults them and to subsequently estimate the likelihood that they would contemplate engaging in a number of aggressive behaviors in such a scenario. Similarly, in a theft by finding scenario Eifler (2015) asked participants to agree or disagree with the statement that one could feel tempted to take and keep the money in the scenario (“In dieser Situation kann man sich in Versuchung geführt fühlen, den Geldschein mitzunehmen und zu behalten”) (p. 241). In both studies researchers intended to measure whether individuals would perceive crime as an action alternative. While these studies recognize the need for capturing data pertaining to the decision process, their approach also reveals important theoretical and methodological weaknesses.

Firstly, Eifler’s study inquired into whether one (read anyone) could perceive crime as an action alternative in the scenario. This may have led participants to make judgements about whether they believe others as well as they themselves might perceive crime as an action alternative, which may therefore not accurately represent participants’ own perceptions of action alternatives. Moreover, both Eifler (2015) as well as Brauer and Tittle (2017) failed to further distinguish those who perceive crime as one of several action alternatives (deliberators) against those who perceive crime as the only action alternative. Since, SAT’s principle of moral correspondence refers to three rather than two categories (crime not perceived, deliberation about engaging in crime, and only crime perceived), these studies were unable to provide an adequate test of the principle.

4. The present study

This study sought to test key propositions of SAT’s PCP, with a special focus on testing predictions about the mental processes underlying rule breaking and rule abidance. The behavior of interest in this study was theft by finding. More specifically we were interested in whether given the opportunity to acquire a £50 note that someone drops, a witness of this event will decide to take and keep the money (as opposed to handing it back or not reacting at all). Based on Section 1 of the Theft Act 1968 of UK law, when finding property, the finder must undertake reasonable steps to reunite the owner of the property with their possessions. Hence, appropriating money when witnessing the owner of the money losing it and making no attempt to give the money back to the person who lost it constitutes a crime.5

4.1 Methods

The method of choice is an experimental video scenario technique comparable to DeBuck and Pauwels (2022) which overcomes many of the weaknesses of text scenarios. In addition, this study further develops existing attempts to gain measured insights into decision processes.

4.1.1 Sample

Participants were recruited via the online crowdsourcing platform Prolific Academic, which allows researchers to recruit participants for survey studies. The sample consisted of adults with their primary residence in the UK. In total N = 595 valid responses were collected. Participants were 46.9 percent male (including trans male), 47.9 percent female (including trans female), 3 percent non-binary, and 2.2 percent indicated that they would rather not say or the data was missing. Participants were on average 43 years old (SD = 14.24), and about 72 percent of participants who specified their personal income earned less than £30.000 a year. Further, 7.5 percent were unemployed, 17.3 percent were not in paid labor, 21.3 percent in part-time employment, 49.6 percent were in full-time employment and 4.2 percent in other employment arrangements.

4.1.2 Procedure

The study took about 5 minutes to complete, and participants received just over 70 pence for participating. Participants who consented to take part in the study filled in a set of questions assessing their monetary desire and levels of self-control ability. The self-control ability scale included two attention check measurements. Only six participants were excluded from the study for failing both attention checks. The high level of attentiveness observed may be attributed to participants needing to pass the researchers' attention checks to qualify for compensation. Participants were then randomly allocated to either a present or absent observer condition (whether an observer is present in the setting). After viewing the video scenario, they were asked questions about their behavioral intention, the perception of action alternatives and any deliberative content. Subsequently, they were asked about the ease/difficulty of making the decision as well as the confidence they had about the decision they made. They then answered questions about their perceptions of the moral context portrayed in the video and their crime-specific personal morality. They were thanked for their participation and redirected to the Prolific platform where they could claim reimbursement for participating.

4.1.3 Video scenarios

Videos were recorded using a point-of-view (POV) camera, which assumes the perspective of an individual who is present in the setting. Participants thus watched the video from a first-person viewpoint as though they are taking part in the setting. Two scenarios were recorded in a public space with benches, adjacent to a bus station in an urban UK context (Figure 3). The two scenarios were identical in all but one respect: the presence or absence of a potential observer of the crime. These scenario versions are henceforth referred to as the “present observer” or “absent observer” conditions. The videos begin with a still frame of the respective settings, which was overlayed with an introductory text, which served mainly to position the viewer and their position in the setting. The text states “Imagine you are out in town on your own. You have decided to take a short break.” In the absent observer condition the text further reads “In front of you there is a man sitting on a bench. Otherwise, there is no one else close-by.”, whereas in the present observer condition the text says “In front of you there are two men sitting on benches. Otherwise, there is no one else close-by”. The man who is about to lose £50 faces away from the POV position, the potential observer faces towards the POV standpoint. The potential observer is reading a newspaper.

Figure 3. Video stills of the weak (bottom; absent observer) and strong (top; present observer) moral context manipulation.

At this point the video stills convert into a real video. In both conditions, the man sitting on the left bench gets up and starts walking away. As he walks away, he reaches inside his trouser pocket and takes out his phone. As he takes out the phone, a £50 note slips out of his pocket and drops to the floor. To ensure that participants notice the man losing the money, the videos featured a brief text saying “he drops a £50 note… and walks away”. A few moments later the videos stop.

4.1.4 Predictor variables

We measured monetary desire with a single item: “How happy would it make you if a £50 note 'magically’ appeared in your letter box?” on a five-point Likert scale ranging (1 = not at all happy to 5 = extremely happy). For self-control ability we initially used the eight-item low self-control scale from Wikström et al. (2012), measured using a four-point Likert scale (1 = strongly disagree to 4 = strongly agree). One item was dropped due to an unsatisfactory factor loading such that we proceeded with a seven-item scale.6 Participants were asked to agree or disagree with a set of statements about themselves (e.g., “I often act on the spur of the moment without stopping to think.”). Item scores were inverted (1 = 4; 2 = 3; etc.), summed up and then averaged to arrive at the final score of self-control ability.

Since SAT’s situational model specifically refers to personal moral rules attached to specific contexts and behaviors (Barton-Crosby 2022; Wikström et al. 2012), we measured crime-specific personal morality (alpha = .77, factor loadings > .51) capturing participants’ moral beliefs with regards to theft by finding, their anticipated feeling of guilt in case of stealing the money and getting away with it, and anticipated satisfaction in case of handing the money back to the man who lost it (for precise wording see Appendix A).

This measure departs from the predominant existing measures of personal morality in several ways. Firstly, most existing studies have used a general measure of personal morality representing the moral beliefs and anticipated moral emotions attached to a range of rule violations, rather a crime-specific measurement (Kroneberg and Schulz 2018; Pauwels 2018a; Wikström et al. 2012).7 Since a person’s moral evaluations of different behaviors as well as the same behavior occurring in different contexts can vary, a crime-specific measurement should be more causally relevant for a specific behavior (Herman and Pogarsky 2024). Secondly, existing measures of personal morality tend to inquire into the anticipated moral emotion guilt, as well as shame of being caught (e.g., Wikström et al. 2012). We decided against using an item capturing the anticipated shame for being caught because of its conceptual overlap with the anticipation of negative consequences (i.e., the disapproval of others), conditional upon being caught. In our view, this overlap may contaminate the personal morality measure and did therefore not form part of our personal morality measure. Thirdly, to ensure that our measure of guilt is not confused for shame if caught, we added a description to our guilt measure specifying that “No one witnesses you, and you are sure that no one will find out that you took it”. Lastly, we added a measure of anticipated moral righteousness (satisfaction) they would derive if they decided to give the money back to the man, thus adding a further facet to the moral emotion dimension of personal morality. This measure acknowledges that individuals can derive a sense of satisfaction or moral pride from doing what they think is the right thing to do (Andreoni 1990; Hitlin 2007; Tangney, Stuewig, and Mashek 2007). The scores on the three items were summed up and then averaged to arrive at an overall personal morality score.

Next, we turn to our manipulation of the moral context. The moral context in SAT is defined as the moral norms and their enforcement. The presence or absence of an observer was intended to reflect a strong and weak moral context condition, respectively. The strong moral context manipulation was expected to signal opposition to rule breaking, whereas the weak moral context manipulation was expected to signal comparatively less opposition to rule breaking. We stress that this manipulation of the context may both influence the strength/salience of the moral norm of the context as well as the deterrent qualities (level of enforcement). Condition was thus coded as binary variable with absent observer as the reference category.

Concerning the measurement of individuals’ perceptions of the strength of the moral context we opted for a measurement of the perceived deterrent qualities of the setting while omitting a measure of the perceptions of the strength/saliency of the moral norms of the setting (see Appendix B). This decision was made due to the difficulty of assessing participants’ perceptions of the strength/saliency of the moral norm of a setting. To the best of our knowledge no adequate measurements exists to date8. However, given the likely covariation of perceptions of the strength/saliency of a norm and individuals’ perceptions of the enforcement of the norm, we believed that an assessment of participants perceptions of the deterrent qualities of the setting to be a satisfactory (although perhaps not ideal) proxy for the moral context. The perceived moral context score was thus computed as the product of two averaged three-item scales; the perceived likelihood of negative consequences (alpha = .82, factor loadings > .58) and the perceived severity of negative consequences (alpha = .65, factor loadings > .53). Sample items for these scales are respectively “How likely is it that you would get in trouble if you took the money?” and “How annoying would you find the consequences for getting caught taking the money?”. Conceptualized in this way our measures of the perceptions of the moral context should play a role for the perception of action alternatives as well as the deliberative external control process.

4.1.5 Behavioral intention and decision process measurements

Most existing research using vignette scenarios employ a continuous measure of crime intention, such as the likelihood or certainty that the participant would engage in a specific criminal action alternative, or that they would act in the way that a fictional character in the scenario behaved (e.g., Craig 2019; van Gelder et al. 2022; e.g., Pauwels 2018a; Schoepfer and Piquero 2006). We decided against using such a continuous measure, and in favor of a binary yes versus no measurement of crime intention to more accurately reflect the discreet nature of behavior in the real world.

Intermediate scores on widely used continuous crime intention measures, are likely to tap into varying levels of decisions confidence or may represent participants’ experienced ease or fluency when deciding. We argue that both dimensions (confidence and ease/fluency) have theoretical significance within the tenets of the PCP. Decision ease/fluency as well as decision confidence have been argued and shown to be markers of automatic/intuitive and deliberative choice (Glöckner 2010; Thompson, Prowse Turner, and Pennycook 2011; Topolinski and Strack 2009c, 2009a, 2009b). Whereas automatic/intuitive decisions tend to be made confidently and with a sense of ease, deliberative decisions tend to be made less confidently and appear more difficult to make. Thus, those who do not perceive crime as an action alternative, and those who perceive crime as the only action alternative should make their decisions more easily, and more confidently than those who deliberate about engaging in crime. Therefore, in addition to asking participants about their behavioral intention captured as a discreet behavior, we asked participants how confident they were in the decision they made (“How confident are you about your choice?”; 1 = Not at all confident to 5 = very confident) as well as how easy or difficult they found making the decision (“How easy or difficult was it for you to decide how you would behave?”; 1 = very easy to 5 = very difficult). This way we can assign theoretical significance to would-be intermediate values of continuous crime intention measurements.

Another marker of automatic versus deliberative choice is response latency (decision times). Prior research has shown that deliberative decisions tend to take longer than decisions made in an intuitive/automatic fashion (Achtziger and Alós-Ferrer 2014; De Neys 2006). If we assume that individuals who would deliberate about engaging in crime in the real world are also more likely to deliberate about their crime intention in this study, then we should expect deliberators to take longer than non-deliberators to decide whether they would take and keep the money (Beier 2016). Hence, we measured how long it took participants to indicate whether they would engage in crime using a response latency function on the survey platform Qualtrics.

While decision ease, confidence and decision times may provide circumstantial evidence for automatic and deliberative decision styles (Herrmann 2024), we included further questions to gain more direct insights into participants’ perceived action alternatives, and any deliberative content (see Figure 4). Participants who indicated that they would engage in crime were subsequently asked which of three statements would best reflect their decision making. The statements were 1) “I would just take and keep the money, without a second thought” (only crime perceived), 2) “I would take and keep the money, but before taking it, I would make sure that I won’t get into trouble for it” (deliberation about deterrents), and 3) “I would take and keep the money, but I would be somewhat morally conflicted” (moral deliberation).

Conversely, the statements for individuals who indicated that they would not engage in crime were 1) “It would not even cross my mind to take and keep the money” (crime not perceived), 2) “I would consider taking the money, but decide not to take it because I would worry about getting in trouble” (deliberated about deterrents), and 3) “I would consider taking the money, but then decide not to do it because I would be morally conflicted” (moral deliberation). Subsequently, all participants were asked to describe in their own words how they would respond, in case they believed that none of the response options adequately captures their reaction.

Figure 4. Response classifications based on survey questions about the perception of action alternatives and deliberative content.

4.2 Hypotheses

In accordance with the theoretical propositions described above, we derived a series of hypotheses. Concerning our predictions about the perception of action alternatives (crime not perceived, crime perceived but initially unsure (deliberation), only crime perceived) we formulated the following hypotheses.

Hypothesis 1: Stronger personal morality and stronger perceptions of the strength of the moral context increase the likelihood that the individual will not perceive crime as an action alternative (as opposed to deliberate about engaging in crime and as opposed to perceiving crime as the only action alternative).

Hypothesis 2: Weaker personal morality and weaker perceptions of the strength of the moral context increase the likelihood that the individual will perceive crime as the only action alternative (as opposed to deliberate about engaging in crime or not perceiving crime as an action alternative).

With regards to the relationship between automatically/deliberative decision making and response times, decision difficulty and decision confidence we tested the following hypotheses:

Hypothesis 3: Compared to individuals who decide automatically/intuitively, those who indicate that they would deliberate about engaging in crime have slower response times.

Hypothesis 4: Compared to individuals who decide automatically/intuitively, those who indicate that they would deliberate about engaging in crime will be less confident in the decision they make.

Hypothesis 5: Compared to individuals who decide automatic/intuitively, those who indicate that they would deliberate about engaging in crime find making the decision more difficult.

Regarding the principle of the conditional relevance of self-control abilities we derived the following hypothesis:

Hypothesis 6: Self-control abilities inhibit the individual from engaging in crime only amongst individuals who indicate that they deliberated about engaging in crime.

In addition, concerning the content of individuals’ deliberations we tested the following hypotheses:

Hypothesis 7: Amongst those who deliberate, strong personal morality increases the likelihood that the individual engages in internal moral deliberations.

Hypothesis 8: Amongst those who deliberate, a strong (perceived) moral context increases the likelihood that the individual deliberates about the negative consequences of engaging in crime.

Given the remaining theoretical ambiguity concerning the role of motivation (i.e., desire for money), we further statistically explored to what extent individuals’ desire to acquire a £50 note influences their perceptions of action alternatives and deliberative choice.

We further intended to test whether our manipulation of the moral context had an influence on the perception of action alternatives and deliberation as well as the behavioral outcome via perceptions of the strength of the moral context. However, since our manipulation of the moral context had a rather ambiguous effect on participants (see results below), we were unable to adequately test these hypotheses.

4.3 Results

In what follows, we present the findings of our study beginning with the distribution of main variables of interest and their bivariate correlations. We then proceed to test the hypotheses related to the perception part of the PCP (H1-2: Under which conditions are individuals more likely to perceive crime as an action alternative?). Next, deliberative and automatic mode are compared against a set of markers: decision confidence and fluency, as well as recorded response times (H3-5: Do the markers of deliberation distinguish self-reported deliberators from automatic decision-makers?). We then test the conditional relevance of self-control abilities (H6: Do self-control abilities matter only among those who deliberate?). Among those who deliberate, we examine whether our variables can predict the content of deliberation (H7-8: When are individuals likely to deliberate about morals vs. about deterrents?). The final, and largely exploratory portion of the analysis is devoted to the role of desire at different stages of the PCP.

4.3.1. Descriptive analysis

We first inspect the distribution of the (unstandardized) variables of interest. As is often the case of samples drawn from the general population, measures of law-relevant personal morality are highly skewed towards the right with few people dismissing the prevailing norms against theft by finding. Self-control ability was distributed somewhat more normally. The monetary desire variable has high central tendency measures meaning that most participants would be sufficiently motivated by the prospect of earning £50. In terms of the likelihood and severity of negative consequences, most participants scored low on the overall scale.

Table 1. Distribution of causal variables

Variable

Min

Max

Mean

SD

Distribution

Personal morality (PM)

1.00

5.00

4.12

0.83

Self-control ability (SCA)

1.63

4.00

3.04

0.46

Desire

1.00

5.00

4.28

0.91

Perceived moral context (PMC)

1.00

25.00

8.37

4.17

Turning now to bivariate correlations, we observe that personal morality was moderately correlated with self-control abilities (r = .24, p < .001, which is a common finding (e.g. Wikström et al. 2012). Individuals high in law-relevant personal morality were also more likely to perceive a stronger moral context (r = .26, p < .001. This effect becomes most apparent by the fact that our sample did not include any individuals with weak personal morality who perceived the moral context to be strong, although the reverse (individuals with strong personal morals who perceive the moral context to be weak) was much more common.

Further, while condition affected participants’ perceived moral context score in the desired direction, its influence on the two facets of the perceived moral context (likelihood and severity of negative consequences) were rather ambiguous. We found that in the present observer condition, participants perceived a significantly higher likelihood of getting caught and facing negative consequences (t = -4.844, df = 593, p-value <.001), but a significantly lower severity of the consequences if caught (t = 2.045, df = 593, p-value = .041). A successful manipulation of the moral context should influence both the likelihood and severity in the same direction, or if it influences only one facet, then it should at least not influence the other facet in the opposite direction. We suspect that the presence of the observer may have led participants to focus on negative consequences such as disapproving looks of the observer whereas in the absent observer condition, participants may have taken into consideration the possibility of formal sanctions.

Consequently, we were unable to test any hypotheses concerning the influence of the moral context on the perception of action alternatives, deliberation and action intention, mediated by perceptions of the strength of the moral context. Moreover, our manipulation had no significant effect on participants’ intention to take and keep the money (χ2 = 0.77929, df = 1, p-value = 0.38). Given the failed manipulation of the moral context we decided to run all statistical tests on the entire sample while including condition as a control variable.

Table 2. Zero-order correlations

PM

SCA

Desire

PMC

Condition

PM

1

SCA

.24***

1

Desire

.05

-.04

1

PMC

.26***

.00

.03

1

Condition

-.05

-.07

-.03

.1*

1

Note: * p≤ .05; **p ≤ .01; ***p ≤ .001. PM – personal morality;

SCA – self-control ability; PMC – perceived moral context.

Across the entire sample, 8.4% of participants (n = 49) indicated that they would take and keep the money, while the remaining 91.6% stated otherwise. Since follow-up questions about potential deliberation about engaging in crime were asked to both, we report the respective frequencies in Table 3. Most non-offending participants did not perceive theft as an option, and most offenders chose to keep the money upon deliberating about either deterrents or morals (in equal proportion). Conversely, deliberating but non-offending participants considered their personal morals overwhelmingly more often than deterrents. Incidentally, a two-proportion z-test shows that deliberating about morality is associated with greater rule compliance than deliberating about sanction threats (z = −5.72, p-value < .001).

Only few participants indicated in the open response box that the response options provided to them did not match how they would respond. Five of these suggested that participants may have misunderstood the scenario and were therefore excluded from the data set. Three participants did not provide an indication of their perception of action alternatives or deliberative content, but their open-response answers allowed us to recode them into the category “would not perceive crime as an action alternative.”

Table 3. Offending and type of deliberation

Crime not perceived

Deliberated about

Only crime perceived

Total

deterrents

morals

No crime

347

27

163

537

Crime

20

20

9

49

Total

347

47

183

9

4.3.2. The perception of action alternatives

To test our alternative formulation of the principle of moral correspondence, we first divided our samples into those who perceive crime as the only option (n = 9), those who deliberate (were initially unsure about engaging in crime; n = 239), and those who do not perceived crime as an option (n = 347). Since there were only 9 individuals who said they would perceive crime as the only action alternative we were unable to statistically compare this group against the two other groups using multinomial logistic regression. Hence, we decided to conduct one logistic regression comparing those who do not perceive crime as an action alternative against deliberators. Further, we decided to graphically examine the relationship between the two key predictors personal morality and perceptions of the moral context and individuals’ tendency to perceive crime as the only action alternative (versus deliberate versus do not perceive crime as an action alternative).

Figure 5. Plotted β coefficients of logistic regression model predicting not perceiving crime as an action alternative. Reference category is deliberation. McFadden’s Pseudo-R2 is .26. The coefficients in this and other plots were not standardized to avoid complicating the interpretation of product and quadratic terms.

The results of our binary logistic regression (Figure 5) provide answers to H1. When comparing those who would not perceive crime as an action alternative with those who would deliberate, we found that individuals with stronger personal morality (+1SD) also had five times greater odds not to perceive crime as an option (OR = 5.02, p-value < .001). This means that as personal morality weakens (compared to strong personal morality), individuals became more likely to indicate that they would deliberate about engaging in crime. Further, higher perceptions of the strength of the moral context were associated with greater odds of deliberating about engaging in crime (as opposed to not perceiving crime as an action alternative; 1OR\frac{1}{OR} = 1.53, p-value < .001), which would contradict expectations. However, the relationship between perceptions of the moral context and the log-odds of the dependent variable was non-linear, as demonstrated by a significant Box-Tidwell test. Hence, to circumvent issues of non-linearity, we included a squared term of perceptions of the moral context in our model. The coefficient and significance of the squared term (p-value = .029) indicates a curvilinear relationship whereby at lower values of perceived moral context the effect on the dependent variable was positive (individuals became more likely not to perceive crime as an action alternative), as predicted, while at greater values (past +1SD), and against expectations, its relationship was negative (individuals became more likely to deliberate). Thus, we found only partial support for H1.

Our graphical examination of the relationship between personal morality, perceptions of the strength of the moral context and the perception of action alternatives are shown in Figure 6. We plotted participants’ personal morality against their perceptions of the strength of the moral context. Each data point was then color-coded according to whether they perceived crime as the only action alternative (red), deliberated about engaging in crime (yellow) or did not perceive crime as an action alternative (green). We then subdivided the scatterplot into 36 equally sized quadrants and shaded each according to the dominant color of the data points belonging in each quadrant.

Figure 6. Predicting the perception of action alternatives and deliberation using personal morality and perceptions of the moral context. The quadrants are colored according to the dominant category of perceived action alternatives/deliberation.

With regards to deviations from weak moral congruency (H2: weak personal morality and weak perceived moral context) we find that the few individuals who reported that they would perceive crime as the only option scored consistently low on measures of personal morality and perceptions of the moral context. This finding lends support for H2.

This illustration also provides further insight into our predictions and statistical results concerning H1. When it comes to deviations from strong moral congruency (strong personal morality and strong perceived moral context), Figure 6 suggests that individuals with strong personal morality predominantly did not perceive crime as an action alternative regardless of their perceptions of the strength of the moral context. This may imply that in many cases personal morality ‘blocks’ the perception of crime as an action alternative without much influence of individuals’ perceptions of the moral context.

4.3.3. Automatic or deliberative choice

Turning to the distinction between automatic and deliberative decision-making (H3-5), we compared the respective subgroups in terms of recorded response times, confidence in decision and self-reported ease/difficulty of deciding. Self-reported deliberators took significantly longer to reach a decision (7.54s vs 5.94s; t = -3.91, df = 327.11, p-value < .001). They were less confident about the chosen action alternative (t = 13.80, df = 283.52, p-value < .001), and found the choice to be more difficult to make (t = -14.07, df = 265.7, p-value < .001) than those who reported only considering one action alternative (only crime or no crime). These findings support H3-5 and provide converging evidence about the distinction between automatic and deliberative decision-making based on participants’ self-reports.

Since we hypothesized that deliberation is more likely in cases of deviations from moral congruency (strong personal morality and strong moral context or weak personal morality and weak moral context), we mapped our three markers of deliberation (decision speed, confidence, and difficulty) onto a matrix of personal morality and perceived moral context (Figure 7). These illustrations show that deviations from strong personal morality are associated with higher levels of decision difficulty and lower decision confidence, thus lending further credence to our results regarding H1. Concerning deviations from weak, congruent moral rule guidance the illustration is less clear-cut. Nevertheless, the graph hints at a trend that deviations from weak congruent moral rule guidance are associated with lower decision confidence and higher decision difficulty, thus also underlining our findings regarding H2. The mapping of response times did not show any clear patterns, perhaps in part because of the small difference in response times between deliberators and non-deliberators that we found.

Figure 7. Markers of automatic and deliberative choice.

4.3.4. Conditional relevance of self-control abilities

Next, we examined whether self-control ability influences the decision to engage in crime only among those who indicated that they would deliberate about engaging in crime. This we do using a linear probability model (LPM; Figure 8) including a product term of self-control ability and deliberation (R2 = .33).9 The coefficient estimates plotted in Figure 8 show that the effect of self-control is neither significant in those who reached a decision automatically (main effect) nor among those who would deliberate about engaging in crime. While these results support predictions to the extent that self-control abilities do not matter for the decision to engage in crime amongst those who do not deliberate, they contradict expectations amongst deliberators.

Figure 8. Plotted β coefficients of a linear probability model predicting crime intention amongst deliberators. Confidence intervals are calculated using robust standard errors. The relationship of personal morality with the outcome variable initially violated the linearity assumption. The inclusion of a quadratic term improved the fit of the model (F [1, 577] = 54.42, p < 0.001).

In an exploratory analysis we decided to further break down the deliberating subgroup into those who deliberated about morals and those who deliberated about deterrents (Figure 9). In this second LPM (Figure 9; R2 = .39), self-control had a negative effect on crime amongst those who deliberated about deterrents (p-value = .0003), but not amongst those engaging in moral deliberation (p-value = .82). Amongst individuals deliberating about deterrents, those with higher self-control (+1SD) were 13.9pp less likely to decide that they would take and keep the money.10 This still lends only partial support to H6, according to which self-control abilities should enable all deliberators to resist the temptation, regardless of deliberative content.

Figure 9. Plotted β coefficients of a linear probability model predicting crime intention amongst deliberators. Confidence intervals are calculated using robust standard errors. A quadratic term of personal morality was included due to an initial violation of the assumption of linearity.

4.3.5. Predicting the content of deliberation

Can we also predict the content of deliberation? In the final regression model (Figure 10), we pitted participants who indicated that they would deliberate about deterrents (reference category) against those who indicated that they would be morally conflicted (McFadden’s Pseudo-R2 = .12). As expected, personal morality was strongly associated with the moral deliberation process (OR = 21.08, p-value = .003). However, the relationship violated the linearity assumption. We thus included a quadratic term. The coefficient of the quadratic term suggests that the relationship between personal morality and moral deliberation, while strongly positive at low and medium values of personal morality, reverses around +1SD value of personal morality. Further, the perceived strength of the moral context predicted deliberation about deterrents (OR = 0.54, p-value = .002). We were therefore able to conditionally confirm hypothesis 7 and fully confirm hypothesis 8.

Figure 10. Plotted β coefficients of a logistic regression model predicting moral deliberation among those who deliberated. Reference category is “deliberation about deterrents”.

4.3.6. Exploring the influence of desire

The model reported in Figure 5 above also partially answers our exploratory question about the effects of desire on the perception of action alternatives and deliberation. Those with stronger desire for £50 were more likely to at least consider taking it as part of a deliberative choice (as opposed to not perceiving it as an action alternative; 1OR\frac{1}{OR} = 1.76, p-value < .001). Due to the low number of individuals who perceive crime as the only action alternative, we were unable to estimate the effect of desire on this outcome.

The first LPM (Figure 8) also sheds light on the role of desire amongst those who deliberate. Results show that within this subgroup, a 1SD increase in desire predicted 2pp higher probability of crime intention (p-value = .02).

5. Discussion

The status quo of the research on offender decision making consists of the identification of input-output associations but as such largely fails to provide much evidence about how people make decisions (Herrmann 2024). To advance the field, the present study offers a novel and rigorous methodological approach to understanding offender decision-making, which allowed us to gain measured insights into the decision processes. We applied the method we developed to a test of SAT’s PCP, which exemplifies a mechanism-based account of how putative causes and their interplay lead to behavior.

Based on a diverse sample of adults living in the United Kingdom, we tested SAT’s key propositions pertaining to the PCP using video scenarios featuring an opportunity to unlawfully acquire £50. Overall, our results provide support for many of our predictions and raise important theoretical questions wherever they contradicted expectations.

Descriptively, there were some important findings worth highlighting. Our results showed that most people decided that they would not take and keep the money because they would not even perceive it as an action alternative. This means their decision not to steal the money is neither thoughtful nor reflective (i.e., deliberative), which largely contradicts some theorists claims that good decisions (i.e., the decision not to engage in crime) are made in a thoughtful and reflective manner (e.g., Paternoster and Pogarsky 2009). Further, amongst those who perceive crime as an action alternative, only a very small share of respondents indicated that they would habitually engage in crime without first deliberating about engaging in crime. We speculate that this may in part be due to the unfamiliarity with the decision context, and absence of high arousal emotions, each of which could be favoring deliberative over automatic processing (Pachur and Spaar 2015; Van Gelder 2013). Further, a considerable share of respondents indicated that they would perceive taking and keeping the money as an alternative, but that their decision would require some level of deliberation characterized by either moral conflict or an assessment of the risk of negative consequences.

Regarding our hypotheses and in line with expectations, we found that when personal morality and perceptions of the moral context are weakest, individuals are most likely to engage in crime in a habitual/automatic manner, that is they perceive crime as the only option. As both personal morality and perceptions of the moral context strengthen, people become more likely to deliberate about engaging in crime, either contemplating the risk of negative consequences or being morally conflicted. Whether internal (moral conflict) or external (fear of negative consequences) control processes characterize individuals’ deliberations depends on the strength of people’s personal morality and perceptions of the strength of the moral context. The stronger each is the more likely it will dominate the content of their deliberations.

As personal morality further increases, so do the chances that individuals do not even perceive crime as an action alternative. However, counter expectations, we found that the highest levels of the perceived strength of the moral context were associated with deliberation rather than not perceiving crime as an action alternative. Future research should inquire if this rather surprising finding replicates, and if so, what possible explanation this might have.

Our graphical exploratory analysis highlighted another interesting finding concerning the perception of action alternatives and deliberation. Personal morality, when sufficiently strong appeared to “block” the perception of crime as an action alternative for most participants, regardless of the perceived moral context. This is in line with Brauer and Tittle (2017), who speculate that strong personal morality may act as a “moral imperative” which may categorically preclude the perception of crime as an action alternative. Hence, this would mean that personal morality alone, if sufficiently strong, may be able to filter out criminal action alternatives. Similar ideas have been proposed and demonstrated by scholars researching differential deterrability (Herman and Pogarsky 2022; Paternoster and Simpson 1996; Pogarsky 2002). For instance, Pogarsky (2002) argued that “acute conformists” follow the law because of their strong moral identity and therefore their perceived risk and severity of sanctions play no role in determining their behavior. Given our results and the existing body of literature documenting the effect, we suggest that future studies should test this hypothesis vis à vis SAT’s principle of moral correspondence. However, we add that perhaps a very weak moral context, which we were unable to create in our scenario (e.g., strong peer pressure), may counter the blocking effect of strong personal morality and therefore lead to deliberations even among individuals with a very strong moral compass.

To provide evidence supporting the distinction between deliberators and non-deliberators, we triangulated participant responses about whether they deliberated about engaging in crime with three markers of automatic vs. deliberative choice; decision confidence, ease/difficulty and decision time. As we hypothesized, deliberators were less confident, found it more difficult to decide what to do, and took longer to decide than non-deliberators. These results firmly square with the proposed use of so-called process tracing techniques that seek to gather cognitive or behavioral data pertaining to individuals’ decision processes (Beier 2016; Herrmann 2024). While Herrmann discusses their use in relation to real-time decision-making in virtual reality scenarios (i.e., making a decision as the scenario unfolds), our study suggests that measures of decision confidence, ease/difficulty and decision time may with some modification also be informative in survey studies using video scenarios.

Being able to distinguish deliberators from non-deliberators is crucial for testing whether self-control abilities are relevant only for those who deliberate. Existing research, based on input-output testing, was unable to converge on a clear picture regarding the role of self-control (e.g., Kroneberg and Schulz 2018; Pauwels 2018b). While our results support the claim that self-control abilities do not influence the decision to engage in crime amongst those who do not deliberate, its role in predicting who engages in crime amongst those who deliberate requires further nuance. Rather than showing that self-control abilities decrease the likelihood of crime engagement amongst all who deliberate, our exploratory analyses suggest that self-control abilities matter only for those who deliberate about deterrents, but not for those who engage in moral deliberations. One reason for this may be that the self-control ability measure we employed is not sufficiently attuned to capturing the abilities required for individuals to adhere to their moral compass when tempted to do otherwise (Hasselhorn et al. 2024; Pauwels, Svensson, and Hirtenlehner 2018), and more attuned to capturing the abilities required to mentally represent the risk of negative consequences.

Prior studies have placed greater emphasis on the potential role of self-control abilities when individuals are expected to engage in moral deliberations (e.g., Eifler 2015; Hirtenlehner and Hardie 2016). They have often overlooked the predictions that the same abilities that would allow an individual to act in line with their personal morality, when tempted or pressured to do otherwise, might also underly the ability to consider deterrents in a deliberative fashion and thus inhibit a perceived criminal action alternative (Wikström and Treiber 2007). Our study suggests that self-control abilities may very well be influential for the deterrence-related deliberative process (external control process).

Further, our study is the first to empirically distinguish between control processes (self-control and deterrence) and the ability (self-control ability) required for these control processes to inhibit a perceived criminal action alternative. The conceptual and empirical distinction between these concepts provides a much-needed step towards more rigorous and adequate tests of the theory.

Regarding the role of desires for the perception of action alternatives, our exploratory analyses suggest that desire increases the likelihood that individuals deliberate about engaging in crime relative to not perceiving as an action alternative. However, given our rather small pool of individuals who indicated that they would perceive crime as the only action alternative, we were unable to analyze whether greater desire also makes individuals more likely to perceive crime as the only action alternative. Moreover, amongst deliberators, stronger desire for money increases the chances that internal or external control processes fail, and the individual chooses to engage in crime. These findings underline the important role motivational aspects play in SAT, and call for more precise theorizing and inclusion in future studies testing the PCP.

6. Limitations and directions for future research

A key limitation of the study is its correlational design and hence our limited ability to make causal claims. Our manipulation of the strength of the moral context did not have the desired effect on participants’ perceptions of the moral context and behavior. Hence, we were unable to test its influence on intermediate stages of the decision process and behavior via perceptions of the moral context. Future research should implement more effective manipulations of the moral context, so that it can be shown that differences in the moral context lead to differences in behavior via individuals’ perceptions of the moral context. In addition, and for the same reasons, we suggest future studies should also experimentally manipulate the strength of the motivational trigger.

Further, we argue that manipulations of the moral context should try to go beyond the presence and absence of observers. If the strength of the moral context is conceived of as a spectrum from discouraging to encouraging of rule breaking, then the mere absence of an observer may not be sufficient to create a crime-encouraging (weak) moral context. Peer pressure, overt rule breaking by others in the setting, or a bystander crowd that cheers on rule breaking may be much more likely to create a very weak moral context.

Concerning the measurement of individuals’ perceptions of the moral context we advise future studies to develop and employ measures of both the perceived deterrent qualities of the setting as well as the perceptions of the saliency and strength of the moral norms of the context in which the crime occurs. This way the two conceptually distinct, but empirically related constructs (perceived norms and deterrents) can be measured independently. Using these measures researchers should be able to offer a more nuanced test of the principle of moral correspondence, and the conditional relevance of perceptions of the strength of deterrent cues.

We further stress that regarding the measurement of self-control abilities, we think it would be advantageous to employ scales which better reflects those abilities required for the exercise of self-control and deterrence, respectively. This may allow researchers to more adequately test the conditional relevance of self-control abilities (for the self-control process) and deterrence related abilities (for the deterrence process).

Moreover, our novel methodological approach to testing SAT enabled us to test several new hypotheses, subjecting the theory to a more rigorous and adequate test than previous studies. Our results should be replicated using the same or similar methods in studies using different crime types, as well as different or more representative samples. We further advice researchers to inquire more deeply into the idea that strong personal morality can function as a moral imperative that “blocks” the perception of crime as an action alternative, even when perceptions of the strength of the moral norms of the setting are weak.

7. Conclusion

We have argued that existing studies using the dominant experimental vignette approach have been limited in their attempts to adequately test the PCP. Specifically, their inability to gain measured insights into the perception choice process including the perception of action alternatives, and deliberative and automatic decision making has hindered the testing and further development of the theory. For the most part, references to intermediate stages of decision-making have been made post hoc, that is after predictor variables and their interactions showed the hypothesized effects.

Our study offers what most existing research has so far failed to achieve; a credible distinction between individuals who do not perceive crime as an action alternative, those who perceive crime as an action alternative but engage in deliberations (characterized by an internal moral conflict or fear of negative consequences), and those who perceive crime as the only action alternative. This distinction is crucial for a test of key predictions of SAT, including predictions about the perception of action alternatives, automatic and deliberative choice, and the conditional relevance of controls.

Contributions

Christoph Herrmann: Developed the experimental video scenario method including all materials for this paper, compiled the literature review, led the hypothesizing, and plan for data analysis, and wrote the discussion and conclusion sections. He was involved alongside author 2 in the analysis of the results, the visualization of results and writing of the result section. He further directed and filmed the video scenarios used in this paper. Andrzej Uhl: Contributed to the data analysis, the visualizations of results, the writing of the results section, and reviewed the writing and study materials. He was also involved in the production of the video materials. Kyle Treiber: Provided guidance on study materials, offered theoretical guidance, reviewed the manuscript, and contributed to writing in revision.

Correspondence: Christoph Herrmann, CB2 1TL Cambridge, United Kingdom, email: [email protected].

Funding: This work was supported by the Cambridge Trust.

Ethical Approval: The study has been approved by the Ethics Committee of the Institute of Criminology, University of Cambridge.

Declaration of interest

The authors report there are no competing interests to declare.

References

Achtziger, Anja, and Carlos Alós-Ferrer. 2014. ‘Fast or Rational? A Response-Times Study of Bayesian Updating’. Management Science 60(4):923–38. doi: 10.1287/mnsc.2013.1793.

Ai, Chunrong, and Edward C. Norton. 2003. ‘Interaction Terms in Logit and Probit Models’. Economics Letters 80(1):123–29. doi: 10.1016/S0165-1765(03)00032-6.

Andreoni, James. 1990. ‘Impure Altruism and Donations to Public Goods: A Theory of Warm-Glow Giving’. The Economic Journal 100(401):464–77. doi: 10.2307/2234133.

Barton-Crosby, Jennifer. 2022. ‘The Nature and Role of Morality in Situational Action Theory’. European Journal of Criminology 19(6):1421–37. doi: 10.1177/1477370820977099.

Beier, Harald. 2016. ‘Wie wirken “Subkulturen der Gewalt”? Das Zusammenspiel von Internalisierung und Verbreitung gewaltlegitimierender Normen in der Erklärung von Jugendgewalt / How do “Subcultures of Violence” lead to violence? The interplay of internalization and prevalence of norms legitimizing violence in the explanation of youth violence’. Kölner Zeitschrift für Soziologie und Sozialpsychologie 68(3):457–85. doi: 10.1007/s11577-016-0369-2.

Bicchieri, Cristina. 2006. The Grammar of Society: The Nature and Dynamics of Social Norms. Cambridge: Cambridge University Press.

Bouffard, Jeffrey A. 2002. ‘Methodological and Theoretical Implications of Using Subject Generated Consequences in Tests of Rational Choice Theory’. Justice Quarterly 19(4):747–71. doi: 10.1080/07418820200095411.

Bouffard, Jeffrey A., and Nicole Niebuhr. 2017. ‘Experimental Designs in the Study of Offender Decision Making’. Pp. 479–97 in The Oxford Handbook of Offender Decision Making, edited by W. Bernasco, J.-L. van Gelder, and H. Elffers. Oxford University Press.

Brauer, Jonathan R., and Charles R. Tittle. 2017. ‘When Crime Is Not an Option: Inspecting the Moral Filtering of Criminal Action Alternatives’. Justice Quarterly 34(5):818–46. doi: 10.1080/07418825.2016.1226937.

Cialdini, Robert B., Carl A. Kallgren, and Raymond R. Reno. 1991. ‘A Focus Theory of Normative Conduct: A Theoretical Refinement and Reevaluation of the Role of Norms in Human Behavior’. Pp. 201–34 in Advances in Experimental Social Psychology. Vol. 24, edited by M. P. Zanna. Academic Press.

Craig, Jessica M. 2019. ‘Extending Situational Action Theory to White-Collar Crime’. Deviant Behavior 40(2):171–86. doi: 10.1080/01639625.2017.1420444.

De Buck, Ann, and Lieven J. R. Pauwels. 2022. ‘Moral Correspondence, Moral Conflict, and Minor Rule-Breaking in a Randomized Visual Scenario Study: Lessons Learned from a Partial Test of Situational Action Theory’. International Criminology 2(2):128–42. doi: 10.1007/s43576-022-00058-x.

De Neys, Wim. 2006. ‘Automatic–Heuristic and Executive–Analytic Processing during Reasoning: Chronometric and Dual-Task Considerations’. Quarterly Journal of Experimental Psychology 59(6):1070–1100. doi: 10.1080/02724980543000123.

Eifler, Stefanie. 2015. ‘Situation und Kontrolle: Eine Anwendung der Situational Action Theory auf Gelegenheiten zur Fundunterschlagung’. Monatsschrift für Kriminologie und Strafrechtsreform 98(3):227–56. doi: 10.1515/mks-2015-980305.

Ellis, Lee, David P. Farrington, and Anthony W. Hoskin, eds. 2019. Handbook of Crime Correlates. San Diego: Academic Press.

Exum, M. Lyn, and Jeffrey A. Bouffard. 2010. ‘Testing Theories of Criminal Decision Making: Some Empiricla Questions About Hypothetical Scenarios’. Pp. 581–94 in Handbook of Quantitative Criminology, edited by A. Piquero and D. Weisburd. New York, NY: Springer.

Fishbein, Martin, and Icek Ajzen. 2010. Predicting and Changing Behavior: The Reasoned Action Approach. 1st ed. Oxford: Oxford: Psychology Press.

Ford, J. Kevin, Neal Schmitt, Susan L. Schechtman, Brian M. Hults, and Mary L. Doherty. 1989. ‘Process Tracing Methods: Contributions, Problems, and Neglected Research Questions’. Organizational Behavior and Human Decision Processes 43(1):75–117. doi: 10.1016/0749-5978(89)90059-9.

van Gelder, Jean-Louis, Reinout E. de Vries, Andrew Demetriou, Iris van Sintemaartensdijk, and Tara Donker. 2019. ‘The Virtual Reality Scenario Method: Moving from Imagination to Immersion in Criminal Decision-Making Research’. Journal of Research in Crime and Delinquency 56(3):451–80. doi: 10.1177/0022427818819696.

van Gelder, Jean-Louis, Reinout E. de Vries, Iris van Sintemaartensdijk, and Tara Donker. 2022. ‘Personality Pathways to Aggression: Testing a Trait-State Model Using Immersive Technology’. Criminology 60(3):406–28. doi: 10.1111/1745-9125.12305.

Glöckner, Andreas. 2010. ‘Multiple Measure Strategy Classification’. Pp. 83–105 in Foundations for Tracing Intuition: Challenges and Methods, edited by A Glöckner and C. L. M. Witteman. London: Psychology Press.

Hardie, Beth. 2017. ‘Why Monitoring Doesn’t Always Matter:’ University of Cambridge, Cambridge.

Hardie, Beth. 2020. Studying Situational Interaction: Explaining Behaviour by Analysing Person-Environment Convergence. Cham: Springer International Publishing AG.

Hardie, Beth, and Chae Rose. 2025. ‘What next for Tests of the Situational Model of Situational Action Theory? Recommendations from a Systematic Review’. European Journal of Criminology 14773708241306945. doi: 10.1177/14773708241306945.

Hasselhorn, Fabian, Sebastian Sattler, Clemens Kroneberg, and D. Seddig. 2024. ‘The Self-Control Ability Scale: Measuring a Key Construct of Situational Action Theory’. Justice Quarterly. doi: 10.1080/07418825.2024.2413584.

Herman, Shaina, and Greg Pogarsky. 2022. ‘Morality, Deterrability, and Offender Decision Making’. Justice Quarterly 39(1):1–25. doi: 10.1080/07418825.2019.1709884.

Herman, Shaina, and Greg Pogarsky. 2024. ‘Moral Salience, Situational Moral Evaluations, and Criminal Choice’. Crime & Delinquency 00111287241259474. doi: 10.1177/00111287241259474.

Herrmann, Christoph. 2024. ‘A Framework for Testing Theories of Criminal Decision Making Using VR Scenarios and Process Tracing and Its Application to Situational Action Theory’. doi: 10.31235/osf.io/km3ph

Hirtenlehner, Helmut, and Beth Hardie. 2016. ‘On the Conditional Relevance of Controls: An Application of Situational Action Theory to Shoplifting’. Deviant Behavior 37(3):315–31. doi: 10.1080/01639625.2015.1026764.

Hirtenlehner, Helmut, and Franziska Kunz. 2016. ‘The Interaction between Self-Control and Morality in Crime Causation among Older Adults’. European Journal of Criminology 13(3):393–409. doi: 10.1177/1477370815623567.

Hirtenlehner, Helmut, and Heinz Leitgöb. 2021. ‘Differential Self-Control Effects: Moral Filtering and the Subsidiary Relevance of Self-Control’. International Criminology 1(2):91–106. doi: 10.1007/s43576-021-00012-3.

Hitlin, Steven. 2007. ‘Doing Good, Feeling Good: Values and the Self ’s Moral Center’. The Journal of Positive Psychology 2(4):249–59. doi: 10.1080/17439760701552352.

Johnson, Eric. J., Michael Schulte-Mecklenbeck, and Martijn. C. Willemsen. 2008. ‘Process Models Deserve Process Data: Comment on Brandstätter, Gigerenzer, and Hertwig (2006).’ Psychological Review 115(1):263–72. doi: 10.1037/0033-295X.115.1.263.

Kammigan, Ilka, and Ruth Linssen. 2012. ‘Korruption als »Situational Action«. Eine theoretisch-integrative Erklärung korrupten Verhaltens auf Basis der »Situational Action Theory«’. Monatsschrift für Kriminologie und Strafrechtsreform 95(5):331–47. doi: 10.1515/mks-2012-950502.

Klepper, Steven, and Daniel Nagin. 1989. ‘The Deterrent Effect of Perceived Certainty and Severity of Punishment Revisited’. Criminology 27(4):721–46.

Kroneberg, Clemens, and Sonja Schulz. 2018. ‘Revisiting the Role of Self-Control in Situational Action Theory’. European Journal of Criminology 15(1):56–76. doi: 10.1177/1477370817732189.

Pachur, Thorsten, and Melanie Spaar. 2015. ‘Domain-Specific Preferences for Intuition and Deliberation in Decision Making’. Journal of Applied Research in Memory and Cognition 4(3):303–11. doi: 10.1016/j.jarmac.2015.07.006.

Paternoster, Ray, and Greg 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(2):103–27. doi: 10.1007/s10940-009-9065-y.

Paternoster, Raymond, and Sally Simpson. 1996. ‘Sanction Threats and Appeals to Morality: Testing a Rational Choice Model of Corporate Crime’. Law & Society Review 30(3):549–83. doi: 10.2307/3054128.

Pauwels, Lieven J. R. 2018a. ‘Analysing the Perception–Choice Process in Situational Action Theory. A Randomized Scenario Study’. European Journal of Criminology 15(1):130–47. doi: 10.1177/1477370817732195.

Pauwels, Lieven J. R. 2018b. ‘The Conditional Effects of Self-Control in Situational Action Theory. A Preliminary Test in a Randomized Scenario Study’. Deviant Behavior 39(11):1450–66. doi: 10.1080/01639625.2018.1479920.

Pauwels, Lieven J. R., Robert Svensson, and Helmut Hirtenlehner. 2018. ‘Testing Situational Action Theory: A Narrative Review of Studies Published between 2006 and 2015’. European Journal of Criminology 15(1):32–55. doi: 10.1177/1477370817732185.

Pogarsky, Greg. 2002. ‘Identifying “Deterrable” Offenders: Implications for Research on Deterrence’. Justice Quarterly 19(3):431–52. doi: 10.1080/07418820200095301.

Sattler, Sebastian, Floris van Veen, Fabian Hasselhorn, Guido Mehlkop, and Carsten Sauer. 2022. ‘An Experimental Test of Situational Action Theory of Crime Causation: Investigating the Perception-Choice Process’. Social Science Research 106:102693. doi: 10.1016/j.ssresearch.2021.102693.

Schepers, Debbie, and Jost Reinecke. 2018. ‘Conditional Relevance of Controls: A Simultaneous Test of the Influences of Self-Control and Deterrence on Criminal Behaviour in the Context of Situational Action Theory’. European Journal of Criminology 15(1):77–92. doi: 10.1177/1477370817732191.

Schoepfer, Andrea, and Alex R. Piquero. 2006. ‘Self-Control, Moral Beliefs, and Criminal Activity’. Deviant Behavior 27(1):51–71. doi: 10.1080/016396290968326.

Schulte-Mecklenbeck, Michael, Anton Kühberger, and J. G. Johnson, eds. 2019. A Handbook of Process Tracing Methods. 2nd ed. New York: Routledge.

Schulte-Mecklenbeck, Michael, Anton Kühberger, and Rob Ranyard. 2011. ‘The Role of Process Data in the Development and Testing of Process Models of Judgment and Decision Making’. Judgment and Decision Making 6(8):733–39.

Tangney, June Price, Jeff Stuewig, and Debra J. Mashek. 2007. ‘Moral Emotions and Moral Behavior’. Annual Review of Psychology 58:345–72. doi: 10.1146/annurev.psych.56.091103.070145.

Thompson, Valerie A., Jamie A. Prowse Turner, and Gordon Pennycook. 2011. ‘Intuition, Reason, and Metacognition’. Cognitive Psychology 63(3):107–40. doi: 10.1016/j.cogpsych.2011.06.001.

Topolinski, Sascha, and Fritz Strack. 2009a. ‘Scanning the “Fringe” of Consciousness: What Is Felt and What Is Not Felt in Intuitions about Semantic Coherence’. Consciousness and Cognition 18(3):608–18. doi: 10.1016/j.concog.2008.06.002.

Topolinski, Sascha, and Fritz Strack. 2009b. ‘The Analysis of Intuition: Processing Fluency and Affect in Judgements of Semantic Coherence’. Cognition and Emotion 23(8):1465–1503. doi: 10.1080/02699930802420745.

Topolinski, Sascha, and Fritz Strack. 2009c. ‘The Architecture of Intuition: Fluency and Affect Determine Intuitive Judgments of Semantic and Visual Coherence and Judgments of Grammaticality in Artificial Grammar Learning’. Journal of Experimental Psychology: General 138(1):39–63. doi: 10.1037/a0014678.

Treiber, Kyle. 2011. ‘The Neuroscientific Basis of Situational Action Theory’. Pp. 213–46 in The Ashgate Research Companion to Biosocial Theories of Crime, edited by K. M. Beaver and A. Walsh. Farnham, England: Ashgate.

Treiber, Kyle. 2017. Biosocial Criminology and Models of Criminal Decision Making. Vol. 1. edited by W. Bernasco, J.-L. Van Gelder, and H. Elffers. Oxford University Press.

Uhl, Andrzej, Christoph Herrmann, and Kyle Treiber. forthcoming. ‘Corruption: A Situational Action View’. Theory and Society.

Van Gelder, Jean-Louis. 2013. ‘Beyond Rational Choice: The Hot/Cool Perspective of Criminal Decision Making’. Psychology, Crime & Law 19(9):745–63. doi: 10.1080/1068316X.2012.660153.

Verhoef, Rogier E. J., Anouk Dijk, Esmée E. Verhulp, and Bram O. Castro. 2021. ‘Interactive Virtual Reality Assessment of Aggressive Social Information Processing in Boys with Behaviour Problems: A Pilot Study’. Clinical Psychology & Psychotherapy 28(3):489–99. doi: 10.1002/cpp.2620.

Verhoef, Rogier E. J., Anouk Van Dijk, Sander Thomaes, Esmée E. Verhulp, Maaike M. Van Rest, and Bram O. De Castro. 2023. ‘Detecting Social Information Processing Profiles of Boys with Aggressive Behavior Problems: An Interactive Virtual Reality Approach’. Development and Psychopathology 35(4):1843–55. doi: 10.1017/S0954579422000505.

Verhoef, Rogier E. J., Esmée E. Verhulp, Anouk van Dijk, and Bram O. de Castro. 2022. ‘Interactive Virtual Reality Versus Vignette-Based Assessment of Children’s Aggressive Social Information Processing’. Research on Child and Adolescent Psychopathology 50(5):621–36. doi: 10.1007/s10802-021-00879-w.

Wikström, Per-Olof H. 2006. ‘Individuals, Settings and Crime: Situational Mechanisms and the Explanation of Crime’. in The Explanation of Crime: Context, Mechanisms and Development, edited by P.-O. H. Wikström and R. J. Sampson. Cambridge University Press.

Wikström, Per-Olof H. 2017. ‘Character, Circumstances, and the Causes of Crime: Towards an Analytical Criminology’. Pp. 501–21 in The Oxford Handbook of Criminology. Oxford University Press.

Wikström, Per-Olof H., and Clemens Kroneberg. 2022. ‘Analytic Criminology: Mechanisms and Methods in the Explanation of Crime and Its Causes’. Annual Review of Criminology 5(1):179–203. doi: 10.1146/annurev-criminol-030920-091320.

Wikström, Per-Olof H., Richard Mann, and Beth Hardie. 2018. ‘Young People’s Differential Vulnerability to Criminogenic Exposure: Bridging the Gap Between People and Place Oriented Approaches in the Study of Crime Causation’. European Journal of Criminology 15:10–31. doi: 10.1177/1477370817732477.

Wikström, Per-Olof H., Dietrich Oberwittler, Kyle Treiber, and Beth Hardie. 2012. Breaking Rules: The Social and Situational Dynamics of Young People’s Urban Crime. Oxford: Oxford University Press.

Wikström, Per-Olof H., and Kyle Treiber. 2007. ‘The Role of Self-Control in Crime Causation: Beyond Gottfredson and Hirschi’s General Theory of Crime’. European Journal of Criminology 4(2):237–64. doi: 10.1177/1477370807074858.

Wikström, Per-Olof H., and Kyle Treiber. 2015. ‘Situational Theory: The Importance of Interactions and Action Mechanisms in the Explanation of Crime’. in The Handbook of Criminological Theory, edited by A. R. Piquero. Newark, US: John Wiley & Sons, Incorporated.

Wikström, Per-Olof H., Kyle Treiber, and Gabriela D. Roman. 2024. Character, Circumstances, and Criminal Careers: Towards a Dynamic Developmental and Life-Course Criminology/ Per-Olof H. Wikström, Kyle Treiber, and Gabriela D. Roman. First edition. Oxford, England: Oxford University Press, 2024.

Appendices

  1. Personal morality scale (crime specific)

Please imagine a situation in which a man loses a £50 note walking right in front of you. You pick up the money and leave. No one witnesses you, and you are sure that no one will find out that you took it. (Likert scales 1 = not at all to 5 = extremely)

  1. How guilty would you feel?

  2. How wrong would it be for you take and keep the money?

Now imagine that instead of taking the money, you hand the money back to the man who lost it. (Likert scale 1 = not at all to 5 = extremely)

  1. How satisfied would it make you feel to give the money back to the man?

  1. Perceived strength of the moral context scale (scenario specific)

Please imagine what it would be like to take and keep the money in the scenario you just watched. (Likert scales 1 = not at all to 5 = extremely)

  1. How likely is it that someone would notice if you took the money?

  2. How likely is it that you would get in trouble if you took the money?

  3. How likely is it that someone might see you taking the money?

  4. How uncomfortable would you be if someone found out that you took the money?

  5. How severe would you find the consequences if you got caught for taking the money?

  6. How annoying would you find the consequences for getting caught taking the money?

Note: items 1-3 reflect the likelihood of negative consequences; items 4-6 reflect the severity of negative consequences.

Comments
0
comment
No comments here
Why not start the discussion?