Beyond Legal Deterrence: A Replication and Extension Study on Psychopathy,Self-Control,and Crime-Specific Willingness to Offend

Abstract

Objectives

This study explored the extent to which law-abiding behavior depends on legal punishment by replicating and extending Meldrum et al. who found that 18% of participants were willing to offend if all crimes were legal during a hypothetical day (i.e., “purging”) and that these individuals exhibited higher psychopathic traits and lower self-control. We sought to replicate their findings and extend them through methodological refinements.

Methods

A community convenience sample of 865 predominantly European participants completed an online survey. Participants responded to a single item asking whether they would offend in the absence of legal controls, followed by a 33-item questionnaire assessing willingness of committing specific crimes under the same conditions, and self-report measures of psychopathy and self-control.

Results

Thirty percent of participants reported willingness to offend on the single-item measure, and 76.7% endorsed at least one specific crime. Minor offenses (e.g., theft; ∼50%) were more frequently endorsed than severe ones (e.g., violent- or sex-crimes; 0.3%-4.3%). An exploratory factor analysis of the purge questionnaire identified seven crime categories with distinct endorsement levels. Regression analyses indicated that psychopathic traits and low self-control were positively related to purging, with psychopathy emerging as the stronger predictor of purging and most crime categories.

Conclusions

Legal punishment alone does not explain law-abiding behavior; individuals appear to refrain from severe crimes for extralegal reasons, whereas minor crimes may require legal deterrence. Psychopathic traits and low self-control represent robust psychological risk factors, underscoring the need for prevention strategies that target these predispositions alongside traditional legal deterrence.

Keywords

classical deterrence theory low self-control psychopathic traits crime-specific offending

Introduction

In the dystopian film series The Purge, a totalitarian government designates one day each year during which all crimes are legalized, purportedly to allow citizens to release pent-up aggression and reduce crime for the rest of the year. This event elicits a range of responses among citizens. Some individuals choose to abstain and shelter indoors, while others actively commit criminal acts, which are referred to as purging. These purging behaviors are driven by a variety of motives, including revenge, thrill-seeking, ideology, necessity, or opportunism. Although fictional, this scenario highlights a central question in criminology: to what extent does the threat of legal punishment deter individuals from offending?

Classical deterrence theory, a foundational framework in criminology, holds that the threat of legal punishment is a central mechanism for preventing crime (Becker 1968). Grounded in rational choice assumptions, it posits that individuals weigh the perceived costs and benefits of offending and are deterred when punishment is certain, swift, and proportionate (Nagin 2013). At the same time, this theory does not claim that legal threats are the sole reason individuals comply with the law. A substantial body of research indicates that law-abiding behavior is also shaped by social, situational, and psychological factors, which in some cases may outweigh deterrent effects (Paternoster 2010; Wikström, Tseloni, and Karlis 2011). The concept of differential deterrability captures this variability, positing that responsiveness to legal threats varies according to individuals’ personal and situational characteristics (Piquero et al. 2011). Accordingly, some individuals are highly sensitive to legal threats, others remain largely unaffected, and still others abstain from crime regardless of legal deterrents (Pogarsky 2002).

To examine the extent to which law-abiding behavior depends on legal controls and how individual traits relate to offending intentions, Meldrum, Lehmann, and Flexon (2021) drew on The Purge films and asked 500 U.S. participants: “If all crime—including murder—was legal one day every year, how likely would you be to commit crime(s) on that day?” (1,590). If punishment were the primary deterrent, high levels of willingness to offend would be expected. Instead, only 18% reported such intent, suggesting that law-abiding behavior is influenced by factors beyond the avoidance of punishment. Participants who reported a willingness to offend scored lower on self-control, which involves difficulties with impulse regulation and delaying gratification (Tangney, Baumeister, and Boone 2004), and higher on psychopathic traits, characterized by socially aversive characteristics such as callousness and egocentricity (Jones and Paulhus 2011). These findings are unsurprising, as both traits are well-established predictors of antisocial behavior (Gillespie, Jones, and Garofalo 2023; Vazsonyi, Mikuška, and Kelley 2017). However, they are noteworthy in that these associations have traditionally been attributed to reduced sensitivity to legal sanctions (Altikriti, Nedelec, and Silver 2021; Altikriti, Nedelec, and Barnes 2022; Gottfredson and Hirschi 1990). Yet, Meldrum, Lehmann, and Flexon’s (2021) findings suggest that offending among such individuals cannot be explained solely by a tendency to discount punishment.

Overall, Meldrum, Lehmann, and Flexon (2021) introduced a novel design by removing the threat of legal punishment altogether. This allowed the researchers to directly assess the extent to which law-abiding behavior persists in the absence of legal controls and how individual differences may influence offending intentions. Despite its innovation, the study also presented methodological limitations. The present research therefore sought to build on their work in two ways: first, by replicating their findings in light of growing concerns about replicability in the behavioral sciences (McNeeley and Warner 2015); and second, by refining and extending the methodology through the development and implementation of a new measure of purging behavior. The theoretical rationale for these extensions, as well as details of the new measure, are presented in the following section.

Replicating and Extending Meldrum et al.'s Study: A Theoretical Rationale

In their study, Meldrum, Lehmann, and Flexon (2021) measured willingness to offend with a single-item question (henceforth referred as the single purge question). Although innovative, the researchers acknowledged that explicitly including “murder” may have led participants to interpret the scenario in extreme terms, suppressing reports of less severe offenses and contributing to the relatively low endorsement rate (18%). Participants also may not have considered the full range of possible crimes when responding. To address these issues, the present study developed the purge questionnaire, a measure designed to capture willingness to offend across a wider variety of crimes, including theft, fraud, vandalism, kidnapping, sexual offenses, and murder. Examining intentions across multiple offenses also allows for a clearer assessment of whether deterrent processes differ by crime type. Indeed, Meldrum, Lehmann, and Flexon (2021) noted that purging is likely nuanced, with some individuals considering only minor offenses (e.g., theft or substance use) and others envisioning more violent acts. This is consistent with criminological evidence showing that legal punishment tends to be more salient in deterring low-severity offenses, such as property crimes, than more severe or violent crimes (Chambliss 1967; Dölling et al. 2009; Entorf and Spengler 2000; Nagin 2013). In fact, deterrability may vary even within the same individual, with sanctions influencing their decision to offend in some crime types but not in others (Bouffard, Exum, and Niebuhr 2018). Thus, in the absence of legal controls, offending intentions likely differs across crime types.

Moreover, Meldrum, Lehmann, and Flexon (2021) also found that psychopathic traits were more strongly associated with purging behavior than low self-control, but it remains unclear whether this holds across different crime types. Both traits occupy central roles in criminological theory (DeLisi 2009; Gottfredson and Hirschi 1990), and recent studies have begun testing them head-to-head to clarify their relative importance, yielding mixed results. In a correctional sample, low self-control emerged as the stronger predictor of violent and property delinquency (DeLisi et al. 2018), whereas community and student samples more often identify psychopathic traits as the stronger predictor of self-reported delinquency and aggression, frequently diminishing the effects of self-control (Flexon et al. 2016; Flexon and Meldrum 2013; Pechorro et al. 2022). One explanation is that low self-control may be more relevant for impulsive and opportunistic crimes, while psychopathic traits better capture severe and violent offending, especially in samples less likely to include individuals with extreme scores in these traits (DeLisi, Bouffard, and Miller 2022; Wright et al. 2017). Findings on substance use remain mixed, with some studies linking it primarily to low self-control and others to both traits (DeLisi et al. 2018; Flexon et al. 2016; Pechorro et al. 2022). Given the ongoing debate, it is important to examine both traits together and assess their unique and relative contributions across different types of crime. This study therefore explored their effects on willingness to commit distinct crimes in the absence of legal controls.

The Present Study

Classical deterrence theory posits that individuals weigh the perceived costs and benefits of offending, with legal punishment serving as a key deterrent. However, the extent to which legal costs explain law-abiding behavior remains an open question. Given the continued influence of deterrence theory in shaping crime prevention policy (Mears and Stafford 2024; Pratt 2008), clarifying the role of legal punishment in compliance is essential for developing evidence-based policies that are cost-effective and targeted toward the crimes and individuals most responsive to legal sanctions. Meldrum, Lehmann, and Flexon (2021) found that most individuals comply with the law even in the absence of punishment, suggesting that legal costs may not be the primary basis of compliance. In contrast, individuals with low self-control and high psychopathic traits expressed greater willingness to offend when legal controls were removed, highlighting their relevance as psychological risk factors. Identifying such predispositions in general community samples can provide valuable insights for both primary and secondary prevention efforts (Reidy et al. 2015).

Therefore, the present study aimed to replicate and extend the work of Meldrum, Lehmann, and Flexon (2021). We employed both the original single purge question and a newly developed purge questionnaire designed to capture willingness to engage in a broader range of criminal behaviors, and we tested the predictive roles of psychopathic traits and low self-control across purging measures and crime categories. Two preregistered hypotheses were tested. First, we hypothesized that a proportion of participants comparable to that observed by Meldrum, Lehmann, and Flexon (2021) would report a willingness to offend when legal sanctions are removed, reflecting an endorsement rate of about 18% (Hypothesis 1 [H1]). We also expected the purge questionnaire to produce higher purging rates than the single purge question, as its broader scope may reduce ambiguity and underreporting. Furthermore, because Meldrum, Lehmann, and Flexon (2021) did not account for participants’ prior criminal behavior, willingness to purge in their study may not have reflected a genuine shift in intent among individuals already involved in crime. We therefore calculated “absolute purging” to capture willingness to offend among participants with no criminal history. In addition, we examined whether purging varied across offense types by conducting an exploratory factor analysis (EFA) of the purge questionnaire and assessing crime-specific endorsement rates. Although prior work suggests that legal sanctions are more influential for minor than severe offenses, we did not formulate specific hypotheses for these analyses given the novelty of applying the purge paradigm to crime types and the unknown factor structure of the questionnaire. Finally, we hypothesized that low self-control and high psychopathic traits would positively predict purging behavior while controlling for demographic covariates (Hypothesis 2 [H2]), and explored whether these traits showed differential associations across offense types.

Methods

Participants and Procedure

For this cross-sectional study, participants were recruited from the general community through convenience sampling. Undergraduate psychology students from various cohorts at Tilburg University (the Netherlands) distributed an online Qualtrics survey via social media and personal invitation. The survey, available in English and Dutch, included several psychosocial questionnaires and had a median duration of 35 min. Prior to participation, respondents read an information letter, were informed of their right to withdraw from the study at any time, and provided informed consent. Upon completion, participants were debriefed on the study's objective: examining the role of legal controls in deterrence. Respondents could choose to be compensated with five euros. All procedures were conducted according to the 1964 Helsinki declaration, the study was approved by the Ethics Review Board of Tilburg School of Social and Behavioral Sciences, with reference code TSB_RP807, and data protection was ensured according to the General Data Protection Regulation (GDPR).

In total, 955 participants were recruited, however, hypotheses were tested on three different subsamples because not all participants completed all questionnaires of the survey. Specifically, analyses were conducted with either 858 (H1), 847 (H2), or 865 participants (EFA and subsequent analyses). Additional details regarding sample size determination and response rate are provided in the Online Supplemental Materials. To assess potential bias arising from case exclusion, we compared included and excluded participants on demographic and study variables (Tables SA-1 and SA-2). The two groups did not differ significantly in age, self-control, psychopathy, SSES, purge question scores, sex, relationship status, number of children, or education level. Demographic information was calculated based on the total sample (N = 865). Participants’ age ranged from 18 to 86 years old (M = 26.3, SD = 10.9). Moreover, 87.3% of the sample was born in a European country (incl. the United Kingdom), with 24.7% being Dutch. Of the sample, 49.9% resided in The Netherlands. Moreover, 67.5% of participants were students, and 49.1% indicated that they held either a part- or full-time job. The Online Supplemental Materials provide additional details on the participants’ countries of origin and residence (Table SA-3) and their employment status (Table SA-4). Table 1 summarizes the major demographics of the sample. For details on how this information was obtained, see “Demographic Covariates” section.

Table 1.

Sociodemographic Information of the Sample.

Demographic variable	n	%	M (SD)	Median
Sex
Male	300	34.7
Female	555	64.2
Other	9	1.0
Relationship status
Single	435	50.3
Married/registered partnership	81	9.4
Divorced/separated	15	1.7
In a relationship^a	263	30.4
Living with partner (but not married)^a	65	7.5
Widowed^a	5	0.6
Number of children			1.2 (0.7)	1.0
No children	772	89.2
One child	23	2.7
Two children	39	4.5
Three children	23	2.7
Four or more children	8	0.9
Completed education
Early childhood education	1	0.1
Primary education	8	0.9
Secondary education	452	52.3
College level education	111	12.8
Bachelor's degree or equivalent	184	21.3
Master's degree or equivalent	103	11.9
Doctoral or equivalent level	6	0.7
Subjective socioeconomic status (SSES)			6.1 (1.6)	6.0

Note. N = 865. M = mean, SD = standard deviation. ^a To align with Meldrum, Lehmann, and Flexon’s (2021) design, the categories in a relationship, living with a partner (but not married), and widowed were combined into the dummy variable other relationship (n = 333, 38.5%).

Measures

The survey included self-report assessments following the procedure of Meldrum, Lehmann, and Flexon (2021). All participants first completed self-control and psychopathy questionnaires, followed by the single purge question and the newly developed purge questionnaire.

Low Self-Control

Trait self-control was measured with the eight-item version of the Brief Self-Control Scale (BSCS; de Ridder et al. 2011; Maloney, Grawitch, and Barber 2012; Tangney, Baumeister, and Boone 2004). Participants were asked to rate how much the statements reflected them using a four-point Likert scale (1 = not at all, 4 = very much). One example item includes “I often act without thinking through all the alternatives.” Mean scores were calculated, where higher scores indicated lower self-control. The scale demonstrated good reliability with a Cronbach's alpha of .78, and a Guttman's lambda-2 (λ²) of .78. The scale exhibited an average inter-item correlation (IIC) of .30, falling within the acceptable range of .15 to .50 (Clark and Watson 2019).

Psychopathic Traits

The Short Dark Triad inventory (SD3; Jones and Paulhus 2014) was utilized to measure psychopathic traits. This 27-item measure consists of three subscales—psychopathy, narcissism, and Machiavellianism—each with nine items. Since both Meldrum, Lehmann, and Flexon’s (2021) and the present study focus on psychopathy, only the psychopathy subscale was included. The scale underwent back-translation from English to Dutch by the authors in collaboration with undergraduate students. Statements were rated on a five-point Likert scale ranging from 1 = strongly disagree, to 5 = strongly agree. One example item included “I’ll say anything to get what I want.” Mean scores were calculated, and higher scores indicated more psychopathic traits. The scale's reliability was acceptable/good (α = .69, λ² = .70, IIC = .20).

Purging Behavior

Purging behavior was measured with two separate assessments. First, with the single purge question employed by Meldrum, Lehmann, and Flexon (2021), and then with the new purge questionnaire. All items were translated from English to Dutch by the present authors together with undergraduate students.

Purge Question. In line with Meldrum, Lehmann, and Flexon (2021), participants were asked “If all crime—including murder—was legal one day every year, how likely would you be to commit crime(s) on that day?” (1590). Response options ranged from 1 = very unlikely, to 4 = very likely. Higher scores indicated greater willingness to commit crimes in the absence of legal controls. In their study, Meldrum, Lehmann, and Flexon (2021) dichotomized the purge question because they found little variability in the responses. Following their procedure, we dichotomized the purge question. Participants who rated that they would be very unlikely (=1) or unlikely (=2) to purge were assigned a value of 0. Those who would likely (=3) or very likely (=4) purge were assigned a value of 1.

Purge Questionnaire. The purge questionnaire was developed for this study by aggregating crimes of varying severity sourced from online platforms (e.g., Mascolo 2023) and a risk assessment instrument (i.e., the Historical Clinical and Future-Revised; Spreen et al. 2014). A pilot study (N = 40) was conducted to ensure completeness, gathering feedback from participants on any types of crimes they felt were missing. The final questionnaire included 33 crimes (see Online Supplemental Materials). Respondents were prompted with the statement “Think of the previous scenario where all crimes are legal for 24 hours. In such a day, how likely would you be to…”. Given Meldrum, Lehmann, and Flexon’s (2021) aforementioned issues with the variability of responses, it was decided to employ a six-point Likert scale instead of a four-point scale, asking participants to rate each crime on a scale from 1 = not at all, to 6 = absolutely. This 33-item questionnaire exhibited excellent reliability (α = .93, λ² = .94, IIC = .32).

Consistent with Meldrum, Lehmann, and Flexon’s (2021) approach, responses to the items were dichotomized: participants who indicated that they would not at all (=1), very unlikely (=2), or unlikely (=3) commit the specific crime were coded as 0, and those who reported that they would likely (=4), very likely (=5), or absolutely (=6) commit the crime were coded as 1. A dichotomous variable was then created to indicate whether participants would commit at least one crime (1) or none (0). In addition, to leverage the greater variability provided by this questionnaire compared to the single purge question, exploratory analyses used a continuous score calculated by averaging item responses, with higher scores indicating greater willingness to commit crimes in the absence of legal controls. An EFA was also conducted to identify potential underlying categories of crime (i.e., factors) based on participants’ responses to the purge questionnaire (see “Results” section). The factors identified were subsequently used in analyses focusing on specific types of crime.

Absolute Purging. To identify individuals who reported a willingness to offend only in the absence of legal controls, we constructed a measure of absolute purging. Prior criminal behavior was assessed with a single item asking participants whether they had ever committed a crime, excluding traffic violations (0 = no, 1 = yes). Hence, this item captured self-reported offending regardless of whether participants had been apprehended or officially recorded by authorities. In total, 221 participants (25.8% of the sample) reported prior offending. Absolute purgers were defined as those who (1) reported no prior offending on this item and (2) nonetheless responded yes on either of the dichotomized purging measures (i.e., the single purge question or the purge questionnaire). This approach made it possible to identify individuals without a history of crime who nevertheless expressed a willingness to offend in the absence of legal sanctions. Absolute purgers were classified separately for each purging measure, and in sensitivity analyses those with prior offending histories were excluded so that analyses focused solely on this group.

Demographic Covariates

In keeping with Meldrum, Lehmann, and Flexon (2021), several demographic variables were measured. These included age (measured in whole years), sex (0 = female, 1 = male),¹ highest level of completed education (1 = early childhood education, 2 = primary education, 3 = secondary education [high school], 4 = college level education, 5 = Bachelor's or equivalent degree, 6 = Master's or equivalent degree, 7 = Doctoral level or equivalent), and number of children (1 = no children, 2 = one child, 3 = two children, 4 = three children, 5 = four or more children). Relationship status was measured and categorized into four dummy variables: single, married, divorced or separated, and other (which included living with a partner but not married, in a relationship, and widowed). These three categories were combined under other to align with the design employed by Meldrum, Lehmann, and Flexon (2021). Unlike Meldrum, Lehmann, and Flexon (2021) who assessed household income, we did not include this variable due to concerns about participant discomfort and potential social desirability bias. Instead, we measured subjective socioeconomic status (SSES) using the MacArthur Scale of Subjective Social Status – Adult Version (Adler et al. 2000), which asks participants to place themselves on a ladder ranging from 1 (worst off) to 10 (best off) to reflect their perceived standing in society. Additionally, we did not include race or region of origin, as these categories are less commonly used or clearly defined in European contexts, where racial and ethnic identities are often more fluid and not easily captured using discrete labels.

Data Analysis

All analyses were pre-registered (https://aspredicted.org/FC2_ZF9), with deviations noted in the Online Supplemental Materials. Analyses were conducted with IBM SPSS (Version 27) and R (v.4.2.3; R Core Team 2023). Descriptives and Pearson's correlations were first computed for all target variables. Then, hypotheses were tested following Meldrum, Lehmann, and Flexon’s (2021) procedure. To test H1, we calculated frequencies for both the dichotomized single purge question and the dichotomized purge questionnaire. Absolute purging was then calculated by identifying participants who reported no prior criminal behavior but indicated a willingness to commit crimes on either purging measure. To test H2, we conducted two stepwise logistic regressions, using the dichotomized single purge question and the dichotomized purge questionnaire as dependent variables. In each analysis, demographic covariates were entered at Step 1, followed by either low self-control (Step 2a) or psychopathic traits (Step 2b). In Step 3, both traits were entered simultaneously to assess their independent contributions. Statistical significance was evaluated using an alpha level of .05, and odds ratios (ORs) were reported.

Then, an EFA was conducted on the 33 items of the purge questionnaire to identify potential latent variables. Using the psych (Revelle 2022) and tidyverse packages in R (Wickham et al. 2019), principal axis factoring with oblimin rotation was employed to allow for correlated factors (Howitt and Cramer 2011). Prior to analysis, data suitability was assessed using Kaiser's measure of sampling adequacy (MSA), the Kaiser–Meyer–Olkin (KMO) test,² and Bartlett's test of sphericity (α < .05; Bartlett 1951; Cerny and Kaiser 1977). The number of factors was determined using eigenvalues larger than 1, scree plots, and parallel analysis. Model fit was evaluated using the root mean square residual (RMSR), Tucker–Lewis Index (TLI), root mean square error of approximation (RMSEA), and Bayesian Information Criterion (BIC).³ Items with factor loadings greater than .40 were assigned to their respective factors, while those loading between .30 and .40 were retained only if they demonstrated clear theoretical alignment with the conceptual meaning of the corresponding factor and its associated items. Items failing to load onto any factor were removed, and the EFA was repeated iteratively until all items were assigned to a factor. Reliability and descriptive statistics were then calculated for each resulting factor. Subsequently, separate multiple linear regressions (α < .05, bootstrapped with 5,000 resamples) were conducted to examine whether low self-control and psychopathic traits predicted each crime category (i.e., factors’ mean scores). Power analyses indicated that the sample size was sufficient to conduct the analyses. Details are provided in the Online Supplemental Materials.

To evaluate the robustness of the findings, several sensitivity analyses were conducted, which are detailed in the Online Supplemental Materials, and briefly presented in the “Results” section. First, whereas Meldrum, Lehmann, and Flexon (2021) dichotomized purging due to limited variability in responses, the purge questionnaire in our study provided sufficient variability to construct a continuous score by averaging responses across items. This score reflects the overall degree of willingness to offend across crime types, offering a more nuanced measure than a dichotomous classification. As a sensitivity analysis, this continuous score was used as the dependent variable in multiple linear regressions testing the effects of self-control and psychopathy while controlling for demographics. Second, we excluded participants who self-reported prior criminal behavior and re-ran the stepwise logistic regressions to examine whether the predictors remained significant among those willing to offend only in the absence of legal controls. Third, participants who had watched The Purge might have been influenced in their responses. To account for this, we measured participants’ familiarity with the film (see the Online Supplemental Materials for details), examined whether familiarity was associated with willingness to purge, and controlled for it in analyses testing the effects of self-control, psychopathy, and demographics on purging and crime categories. Fourth, to assess sex differences in purging rates, we compared males and females on overall purging endorsement, individual purge items, and crime categories. Fifth, simple linear regressions were used to examine the independent effects of psychopathy and low self-control on the crime categories. Sixth, to control for potential confounders, multiple regressions were repeated with demographic covariates, low self-control and psychopathic traits as independent variables and the crime categories as outcomes.

Results

Descriptives and Bivariate Correlations

Table 2 presents descriptive statistics and zero-order correlations among purge variables, psychopathic traits, low self-control, demographic covariates, and prior criminal behavior. Psychopathic traits showed moderate positive correlations with the single purge question (r = .30) and low self-control (r = .34), a strong correlation with the continuous purge questionnaire (r = .52), and a weak correlation with the dichotomized purge questionnaire (r = .28). Low self-control was weakly and positively correlated with the purge question (r = .27), the dichotomized purge questionnaire (r = .29), and the continuous purge questionnaire (r = .29). Age was weakly and negatively related to the purge question (r = −.20), dichotomized purge questionnaire (r = −.27), continuous purge questionnaire (r = −.20), low self-control (r = −.24), and psychopathic traits (r = −.17). All of these associations were statistically significant (p < .001). In addition, male sex was positively associated with all purge variables, psychopathic traits, and prior criminal behavior. Prior criminal behavior was positively linked to purge variables, low self-control, psychopathic traits, and negatively correlated with age.

Table 2.

Descriptives and Bivariate Correlations Between Purging, Low Self-Control, Psychopathic Traits, Demographics, and Prior Criminal Behavior.

	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15
1. Purge question	-
2. Purge questionnaire (dichotomized)	. 35***	-
3. Purge questionnaire (continuous)	.51***	.47***	-
4. Low self-control	.27***	.29***	.29***	-
5. Psychopathic traits	.30***	.28***	.52***	.34***	-
6. Age	−.20***	−.27***	−.20***	−.24***	−.17***	-
7. Sex	.16***	.11**	.25***	.02	.21***	−.02	-
8. SSES	−.06	−.02	−.07*	−.11**	.05	.04	.02	-
9. Education	−.12***	−.16***	−.12**	−.19***	−.11**	.46***	−.03	.06	-
10. Number of children	−.17***	−.22***	−.16***	−.20***	−.13***	.77***	−.02	−.02	.25***	-
11. Married	−.14***	−.19***	−.16***	−.16***	−.14***	.66***	−.02	.01	.27***	.64***	-
12. Divorced	−.09*	−.13***	−.09*	−.13***	−.07*	.32***	−.06	.08*	.18***	.30***	−.04	-
13. Other relationship	.01	.05	−.00	.03	.00	−.12***	.01	−.06	.02	−.15***	−.26***	−.10**	-
14. Single	.09**	.09**	.11**	.10**	.09**	−.35***	.02	.04	−.22***	−.31***	−.32***	−.13***	−.80***	-
15. Prior criminal behavior	.25***	.23***	.33***	.26***	.31***	−.12***	.16***	.03	−.12***	−.10**	−.09**	−.08*	.04	.04	-
Mean	-	-	1.7	2.3	2.2	26.3	-	6.1	-	1.2	-	-	-	-	-
Standard deviation	-	-	0.6	0.5	0.6	10.9	-	1.6	-	0.7	-	-	-	-	-
Possible range	-	-	1–6	1–4	1–5	18+	-	1–10	-	1–5	-	-	-	-	-
Range in current sample	-	-	1.0–5.0	1.0–3.9	1.0–4.1	18–86	-	1.0–10.0	-	1.0–5.0	-	-	-	-	-

Note. N = 858. Sex: female = 0, male = 1. Prior criminal behavior: no = 0, yes = 1. SSES = subjective socioeconomic status. Significant correlations are depicted in bold.

*p < .05, **p < .01, ***p < .001.

How Many Participants Would Offend in the Absence of Legal Controls?

To test H1, frequency analyses were conducted on the full sample (N = 858) to determine whether the proportion willing to purge was comparable to that reported by Meldrum et al. (i.e., 18%; 2021). Responses to the single purge question showed that 40.6% were very unlikely (n = 348), 29.0% unlikely (n = 249), 20.2% likely (n = 173), and 10.3% very likely (n = 88) to offend. After dichotomizing responses, 30.4% (n = 261) indicated a willingness to purge. Among these, 111 also reported prior criminal behavior, resulting in an absolute purging rate of 17.5% (n = 150). In contrast, responses to the purge questionnaire indicated that 76.7% (n = 658) of participants would commit at least one crime in the absence of legal controls. Of these, 205 reported prior criminal behavior, yielding an absolute purging rate of 52.8% (n = 453). The most frequently endorsed offenses included theft of basic necessities (50.0%), luxury items (48.5%), large sums of money (44.6%), and trespassing (44.1%). Conversely, fewer participants endorsed sexual offenses against minors (0.3%–0.6%) or adults (1.2%–1.5%), forced enslavement (1.3%), indecent exposure (1.4%), hate crimes (1.5%), kidnapping (1.9%), or murder (4.3%). Other offenses, such as traffic violations (33.1%) and substance use (28.3%), were endorsed at moderate levels. Full item-level frequencies are reported in Table SA-5.

Given the disproportionate number of females in the sample, sensitivity analyses were conducted to assess sex differences in purging and to determine whether males accounted for the high purging rates (see the Online Supplemental Materials). Chi-square tests showed that males were significantly more likely than females to report a willingness to purge across both measures of purging behavior. Males were also more likely to endorse most specific crime items, with the exception of stealing basic or luxury items, committing fraud against a private person, stalking, purchasing or consuming illegal substances, and touching or groping a minor, where no significant sex differences were found (Table SB-4).

Predicting Purging Behavior: The Role of Low Self-Control and Psychopathic Traits

Stepwise logistic regressions were conducted on a sample of 847 participants to test whether low self-control and high psychopathic traits were positively associated with purging behavior (H2). Tables 3 and 4 present the results using the dichotomized single purge question and purge questionnaire as dependent variables, respectively. Sensitivity analyses that excluded participants with prior criminal behavior (Tables SB-1 and SB-2), employed continuous purge questionnaire scores as the dependent variable (Table SB-3), and controlled for familiarity with The Purge films are reported in the Online Supplemental Materials. Low self-control and high psychopathic traits consistently predicted higher levels of purging, with these effects remaining consistent across sensitivity analyses. Both traits remained significant whether entered independently or jointly, with psychopathic traits showing stronger effects when crime versatility was considered (continuous purge scores). In contrast, results for demographic covariates were mixed across analyses. Age and SSES were generally negatively associated with purging, and male sex was positively associated, although significance varied across models, suggesting less robust effects. Other demographic variables did not show consistent associations with purging.

Table 3.

Logistic Regressions with the Dichotomized Single Purge Question as Dependent Variable.

Variable	Step 1				Step 2a				Step 2b				Step 3
Variable	B	SE	OR	LL, UL	B	SE	OR	LL, UL	B	SE	OR	LL, UL	B	SE	OR	LL, UL
Age	−0.03	0.02	0.97	0.93, 1.00	−0.03	0.02	0.97	0.94, 1.01	−0.03	0.02	0.97	0.94, 1.01	−0.02	0.02	0.98	0.94, 1.02
Male	0.68***	0.16	1.98	1.46, 2.70	0.72***	0.16	2.05	1.49, 2.82	0.47**	0.17	1.60	1.16, 2.22	0.54**	0.17	1.71	1.23, 2.38
SSES	−0.09	0.05	0.92	0.83, 1.01	−0.06	0.05	0.94	0.85, 1.04	−0.12*	0.05	0.89	0.80, 0.98	−0.09	0.05	0.92	0.83, 1.02
Education level	−0.06	0.08	0.94	0.80, 1.11	−0.01	0.09	0.99	0.84, 1.17	−0.03	0.09	0.97	0.82, 1.15	−0.00	0.09	0.99	0.84, 1.18
Number of children	−0.29	0.28	0.75	0.43, 1.30	−0.24	0.28	0.79	0.45, 1.38	−0.38	0.30	0.69	0.38, 1.23	−0.33	0.29	0.72	0.40, 1.29
Married	0.27	0.56	1.32	0.44, 3.93	0.31	0.56	1.37	0.45, 1.38	0.21	0.57	1.24	0.40, 3.78	0.28	0.57	1.33	0.43, 4.08
Divorced^a	18.69	10,632.32	130,980,180.7	0.00, 0.00	18.51	10,520.70	109,210,095.6	0.00, 0.00	18.47	10,388.16	105,105,397.1	0.00, 0.00	18.43	10,326.84	100,977,233.5	0.00, 0.00
Other relationship	0.10	0.16	1.10	0.81, 1.52	0.10	0.16	1.10	0.80, 1.53	0.08	0.17	1.08	0.78, 1.50	0.08	0.17	1.08	0.77, 1.51
Low self-control	-	-	-	-	1.01***	0.16	2.75	2.02, 3.74	-	-	-	-	0.74***	0.17	2.10	1.52, 2.91
Psychopathic traits	-	-	-	-	-	-	-	-	1.15***	0.16	3.16	2.33, 4.29	0.96***	0.16	2.60	1.89, 3.58
Nagelkerke R²	.11				.18				.20				.23

Note. N = 847. OR = odds ratio; LL = lower limit 95% confidence interval; UL = upper limit 95% confidence interval; SSES = subjective socioeconomic status. Step 1 was a significant improvement from the null model (X²(8) = 67.97, p < .001), 69.3% of participants were correctly classified, and the Hosmer–Lemeshow test (Hosmer and Lemeshow 2013) indicated good fit (X²(8) = 15.78, p > .05). Step 2a was a significant improvement from Step 1 (X²(1) = 44.16, p < .001), correctly classified 70.5% of participants, and exhibited good fit (X²(8) = 8.50, p > .05). Step 2b was a significant improvement from Step 1 (X²(1) = 60.18, p < .001), correctly classified 72.6% of the sample, and exhibited good fit (X²(8) = 5.51, p > .05). Step 3 was a significant improvement from Step 2a and Step 2b (X²(1) = 36.70, p < .001), correctly classified 73.1% of participants, and exhibited good fit (X²(8) = 11.38, p > .05). Results depicted in bold were statistically significant.

These extremely large effects may be present because only 14 participants were divorced and none of these participants would purge.

*p < .05, **p < .01, ***p < .001.

Table 4.

Logistic Regressions with the Dichotomized Purge Questionnaire as Dependent Variable.

Variable	Step 1				Step 2a				Step 2b				Step 3
Variable	B	SE	OR	LL, UL	B	SE	OR	LL, UL	B	SE	OR	LL, UL	B	SE	OR	LL, UL
Age	−0.04**	0.01	0.96	0.94, 0.99	−0.03*	0.01	0.97	0.94, 0.99	−0.03*	0.01	0.97	0.94, 0.99	−0.03*	0.01	0.97	0.94, 0.99
Male	0.59**	0.19	1.80	1.24, 2.62	0.59**	0.20	1.81	1.23, 2.65	0.31	0.20	1.36	0.92, 2.02	0.37	0.20	1.45	0.97, 2.15
SSES	−0.01	0.06	0.98	0.89, 1.10	0.03	0.06	1.03	0.92, 1.14	−0.03	0.06	0.97	0.87, 1.08	−0.00	0.06	1.00	0.89, 1.12
Education level	−0.09	0.08	0.91	0.77, 1.06	−0.05	0.09	0.95	0.80, 1.12	−0.09	0.08	0.91	0.77, 1.08	−0.05	0.09	0.95	0.80, 1.12
Number of children	−0.08	0.17	0.92	0.67, 1.28	−0.02	0.17	0.98	0.70, 1.37	−0.10	0.18	0.90	0.64, 1.28	−0.06	0.18	0.94	0.66, 1.34
Married	0.16	0.40	1.17	0.54, 2.57	0.12	0.42	1.13	0.50, 2.57	−0.04	0.43	0.96	0.42, 2.22	−0.05	0.44	0.95	0.40, 2.26
Divorced	0.46	0.68	1.59	0.42, 6.03	0.24	0.70	1.27	0.32, 5.04	0.28	0.71	1.33	0.33, 5.35	0.12	0.73	1.13	0.27, 4.68
Other relationship	−0.06	0.19	0.94	0.64, 1.37	−0.06	0.20	0.94	0.64, 1.38	−0.13	0.20	0.88	0.60, 1.30	−0.09	0.20	0.91	0.61, 1.35
Low self-control	-	-	-	-	1.31***	0.20	3.72	2.54, 5.45	-	-	-	-	1.03***	0.20	2.81	1.88, 4.18
Psychopathic traits	-	-	-	-	-	-	-	-	1.28***	0.19	3.60	2.49, 5.21	1.01***	0.20	2.74	1.87, 4.02
Nagelkerke R²	.12				.20				.20				.25

Note. N = 847. LL = lower limit 95% confidence interval; UL = upper limit 95% confidence interval; SSES = subjective socioeconomic status. Step 1 was a significant improvement from the null model (X²(8) = 69.06, p < .001), classified 77.9% of the participants correctly, and exhibited good fit (X²(8) = 6.70, p > .05). Step 2a was a significant improvement from Step 1 (X²(1) = 51.69, p < .001), correctly classified 78.5% of the sample, and had good fit (X²(8) = 13.60, p > .05). Step 2b was a significant improvement from Step 1 (X²(1) = 52.82, p < .001), correctly classified 79.5% of participants, and exhibited good fit (X²(8) = 5.49, p > .05). Step 3 was a significant improvement from Steps 2a and 2b (X²(1) = 27.77, p < .001), correctly classified 80.6% of the sample, and exhibited good fit (X²(8) = 4.94, p > .05). Results depicted in bold were statistically significant.

*p < .05, **p < .01, ***p < .001.

EFA on Purge Questionnaire Items

An EFA was conducted on the 33 purge questionnaire items to identify latent factors (N = 865). Scree and parallel plots suggested testing solutions with three to eight factors. Seven factors had eigenvalues greater than one, and the seven-factor model demonstrated the best fit. However, four items—stalking (item 15), indecent exposure (item 27), hate crime (item 16), and fraud against a private person (item 9)—did not load onto any factor (i.e., factor loadings < .30). Thus, the EFA was conducted again excluding these items. The remaining 29 items exhibited good sampling adequacy with KMO values greater than .79, and an overall MSA of .91. Bartlett's test of sphericity also indicated good fit (X²(406) = 16,636.7, p < .001). Among models with three to eight factors, the seven-factor solution again provided the best fit (RMSR = .02, TLI = .90, RMSEA = .07 [0.06, 0.07], BIC = −414.18). Most items exhibited factor loadings above .40. Items with loadings between .30 and .40 were retained only if they demonstrated theoretical coherence with corresponding factor and its related items. For example, trespassing and entering (.35) and traffic violations (.34) were retained under Factor 7 (misdemeanors) due to their conceptual alignment with the other misdemeanors items (e.g., substance use and sale, reckless public behavior). Initial EFA solutions are presented in Tables SA-6 and SA-7, and eigenvalues, factor loadings, and inter-factor correlations for the final model appear in Table SA-8.

Reliability and Descriptive Statistics of Crime Categories

All factors demonstrated acceptable to excellent internal reliability, as indicated by Cronbach's alpha and Guttman's lambda-2 coefficients. Factor 1 (violent crimes) consists of seven items (α = .90, λ² = .91), factor 2 (white-collar crimes) includes four items (α = .85, λ² = .85), factor 3 (sex crimes against children) comprises of three items (α = .88, λ² = .88), factor 4 (theft) includes three items (α = .90, λ² = .90), factor 5 (destruction and arson) consists of four items (α = .86, λ² = .86), factor 6 (sex and coercive crimes against adults) comprises three items (α = .86, λ² = .86), and factor 7 (misdemeanors) includes five items (α = .78, λ² = .78). Theft showed the highest mean level of willingness to offend (M = 3.3, SD = 1.5), followed by misdemeanors (M = 2.3, SD = 1.1), white-collar crimes (M = 2.0, SD = 1.2), destruction and arson (M = 1.4, SD = 0.8), violent crimes (M = 1.3, SD = 0.7), sex and coercive crimes against adults (M = 1.1, SD = 0.5), and sex crimes against children (M = 1.0, SD = 0.3). Variability was lowest for sex crimes against children (0.07) and sex and coercive crimes against adults (0.25), and highest for theft (2.35) and misdemeanors (1.18). Sensitivity analyses (Table SB-5) showed that males were significantly more likely than females to report willingness to offend across all crime categories, except theft, where no significant difference was observed. Additionally, familiarity with The Purge films was positively associated with all purge variables and crime categories, except for sex offenses against adults and minors (see the Online Supplemental Materials).

Predicting Crime Categories with Low Self-Control and Psychopathic Traits

To examine whether low self-control and high psychopathic traits predicted the seven crime categories, multiple regressions were conducted (Table 5). High psychopathic traits were significantly and positively associated with all crime categories. In contrast, low self-control was positively related to white-collar crimes, theft, and misdemeanors, but negatively associated with violent crimes. This pattern remained consistent after controlling for demographic covariates (Table SB-7), and for familiarity with The Purge (Table SB-9). To assess whether the associations of low self-control with the crime categories, including its non-significant effects for some categories and its negative association with violent crimes, were attributable to the inclusion of psychopathy in the model, we conducted sensitivity analyses. Specifically, Pearson's correlations (Table SB-6) and simple linear regressions (Table SB-8) were performed. When examined separately, psychopathic traits were positively associated with all crime categories, whereas low self-control was positively associated with all except sex crimes against children. Across all models, standardized beta coefficients consistently indicated that psychopathic traits were more strongly associated with crime categories than low self-control.

Table 5.

Predicting Crime Categories: Associations with Low Self-Control and Psychopathic Traits.

Factor	R	R ²	F(2, 862)	β	B ^a	SE ^a	t(862)	95% CI^a
Factor	R	R ²	F(2, 862)	β	B ^a	SE ^a	t(862)	LL	UL
Factor 1: Violent Crimes
Low self-control	.42	.18	93.50***	−.07*	−0.09*	0.04	−2.20	−0.17	−0.01
Psychopathic traits	.42	.18	93.50***	.44***	0.52***	0.06	13.45	0.41	0.63
Factor 2: White-Collar Crimes
Low self-control	.43	.19	100.11***	.12***	0.26***	0.07	3.77	0.12	0.41
Psychopathic traits	.43	.19	100.11***	.38***	0.77***	0.07	11.56	0.63	0.91
Factor 3: Sex Crimes Against Children
Low self-control	.15	.02	10.20***	−.03	−0.02	0.02	−0.87	−0.05	0.02
Psychopathic traits	.15	.02	10.20***	.16***	0.08**	0.02	4.47	0.03	0.12
Factor 4: Theft
Low self-control	.44	.19	194.07***	.24***	0.70***	0.10	7.50	0.51	0.89
Psychopathic traits	.44	.19	194.07***	.29***	0.79***	0.10	8.89	0.59	0.97
Factor 5: Destruction of Property and Arson
Low self-control	.39	.16	79.27***	−.01	−0.01	0.05	−0.26	−0.10	0.08
Psychopathic traits	.39	.16	79.27***	.40***	0.52***	0.06	11.93	0.41	0.64
Factor 6: Sex and Coercive Crimes Against Adults
Low self-control	.25	.06	27.49***	.01	0.01	0.03	0.16	−0.05	0.07
Psychopathic traits	.25	.06	27.49***	.24***	0.21***	0.04	6.92	0.13	0.31
Factor 7: Misdemeanors
Low self-control	.49	.24	137.32***	.21***	0.42***	0.07	6.63	0.29	0.55
Psychopathic traits	.49	.24	137.32***	.38***	0.73***	0.07	12.05	0.60	0.86

Note. N = 865. β = standardized beta coefficient; B = unstandardized beta coefficient; SE = standard error; CI = confidence interval; LL = lower limit; UL = upper limit.

Bootstrapped coefficients.

*p < .05, **p < .01, ***p < .001.

Discussion

The present study aimed to replicate and extend the findings of Meldrum, Lehmann, and Flexon (2021), who reported that 18% of their U.S. sample would offend in the absence of legal controls, with these individuals exhibiting higher psychopathic traits and lower self-control. In our sample, 30.4% indicated they would offend if all crimes, including murder, were legally permitted. When presented with a list of specific offenses, endorsement increased to 76.7%. However, after excluding participants with a self-reported history of criminal behavior, these rates dropped to 17.5% and 52.8%, respectively. Thus, H1 was supported only when examining Meldrum, Lehmann, and Flexon’s (2021) single purge question among respondents without prior offending history. These findings highlight two methodological limitations in the original study. First, reliance on a single question likely underestimates willingness to offend, either due to the salience of “murder” or because participants did not consider a broad range of possible crimes. Second, failing to account for participants’ prior offending conflates those who would only offend in the absence of legal sanctions with those who already do so despite existing laws, thereby inflating purge rates.

Although more participants than expected reported a willingness to commit at least one crime in the absence of legal controls, this varied markedly by crime type. Endorsement was relatively high for non-violent, low-severity crimes such as theft and trespassing (over 40%), but very low for violent crimes like murder (4.3%) and sexual assault (under 1.5%). Mean endorsement across the seven crime categories identified through EFA followed this same pattern, with minor offenses (theft, misdemeanors) receiving higher support than violent or sexual crimes. These findings align with research indicating that legal punishment is more salient in deterring minor offenses than severe crimes (e.g., Dölling et al. 2009). From a rational choice perspective, minor offenses likely carry relatively low costs beyond legal sanctions and offer appealing benefits, like financial gain or gratification. For instance, stealing luxury goods can provide substantial rewards while eliciting only moderate social condemnation. In contrast, severe crimes such as violent or sexual offenses may offer fewer perceived benefits and evoke strong social condemnation, creating powerful extralegal deterrents that diminish the relative influence of legal punishment. It is also plausible that many individuals do not even consider committing severe crimes, rendering legal punishment irrelevant in their decision-making. According to Situational Action Theory (SAT; Wikström 2007), morality functions as a filter that excludes behaviors deemed unacceptable, so legal punishment influences behavior only when a crime passes this filter and is considered a viable option (Piquero et al. 2011; Pogarsky 2002; Wikström, Tseloni, and Karlis 2011). Consequently, legal punishment likely plays a central role in discouraging minor crimes, where costs and benefits are actively weighed, but is less critical for severe crimes, which are more often precluded by internalized moral standards and social condemnation.

Yet, the results also suggest that legal punishment remains an important deterrent for severe crimes among the subset of individuals who are sensitive to legal consequences. For example, 4.3% of participants reported they would commit murder, 7.5% would engage in physical fights, and 1.9% would commit kidnapping. While small, this is a meaningful minority, particularly given that the sample was largely composed of highly educated students and females, groups typically associated with lower offending risk. Taken together, these findings suggest that although legal punishment may not serve as the primary deterrent for most individuals, it continues to discourage minor offenses and functions as a critical safeguard against severe crimes among those responsive to legal consequences.

Psychopathic Traits, Self-Control, and Offending Beyond Legal Deterrence

Consistent with differential deterrability and Meldrum, Lehmann, and Flexon’s (2021) findings, individuals with low self-control and high psychopathic traits were more likely to report offending when legal controls were absent (H2). In convenience community samples such as ours, this implies that legal controls play a significant role in restraining these individuals. Yet evidence that these traits are linked to diminished sensitivity to legal costs, heightened reward seeking, and overrepresentation in the justice system indicates that legal punishment alone is unlikely to deter individuals at the extreme end of these traits (Altikriti, Nedelec, and Silver 2021; Fox and DeLisi 2019; Tharshini et al. 2021). Legal controls may be sufficient for individuals with moderate-to-high levels of these traits, as reflected in our sample, but reducing offending among individuals with very high, pathological scores likely requires complementary approaches beyond legal deterrence.

Our findings also showed that psychopathic traits were the strongest and most consistent predictor of purging and most crime types, exceeding low self-control and replicating prior work in community samples (e.g., Flexon et al. 2016; Flexon and Meldrum 2013; Meldrum, Lehmann, and Flexon 2021; Pechorro et al. 2022). Low self-control showed only slightly stronger effects with the dichotomized purge questionnaire and matched psychopathic traits in predicting theft. After controlling for psychopathic traits, low self-control remained positively associated only with minor crimes (i.e., theft, misdemeanors, and white-collar crimes) and was negatively associated with violent offenses.⁴ These results suggest that low self-control is more closely tied to minor offending, whereas the callousness and egocentricity captured by psychopathy may be central to severe offending intentions in community samples (DeLisi, Bouffard, and Miller 2022; Wright et al. 2017). Consistent with DeLisi's (2009) unified theory of crime, low self-control may function as a diluted form of psychopathy, with psychopathic traits accounting for much of its links to offending. At the same time, evidence that low self-control predicts criminality more strongly in correctional samples (DeLisi et al. 2018), that both traits correspond to different offending trajectories, and that the most chronic and severe offenders score high on both (Altikriti, Theocharidou, and Sullivan 2020; Wright et al. 2017), underscores the importance of continued research on their overlap, their distinct and combined contributions across offense types, and the influence of sampling and measurement choices.

Limitations and Future Research Directions

This study has several methodological limitations that should be considered when interpreting the findings. First, the use of snowball sampling limited generalizability and produced a sample that overrepresented young, European, female university students with relatively high SSES. This demographic profile likely influenced purging rates in opposing directions: the younger age of our participants compared to Meldrum, Lehmann, and Flexon’s (2021; M_age = 46.3) would be expected to elevate offending (Sampson and Laub 2003), whereas their higher education and predominantly female composition would typically suppress it. Supplemental analyses showed that males, though underrepresented, were more likely to endorse purging and most crime types, suggesting a disproportionate impact on overall rates. Still, the elevated purging despite protective demographic factors suggests that age-related risk may outweigh them or that purging is less constrained by sex and education than conventional crimes. In other words, females and highly educated individuals may be strongly constrained by legal punishment, and without it, they may be more willing to offend than expected. To strengthen generalizability, future research should recruit more diverse samples, particularly higher-risk groups, and move beyond convenience sampling toward field-based recruitment to capture greater variability in severe offending intentions and criminogenic traits.

A second limitation concerns the measurement of absolute purging, which relied on participants’ self-reports of whether they had ever committed a crime (excluding traffic violations). This approach depended on subjective judgments of what qualifies as a crime, likely introducing inconsistencies. Future research should use validated instruments with behaviorally specific items to better assess prior offending. Moreover, reliance on a hypothetical scenario limits ecological validity, as responses may be shaped by social desirability, self-awareness, or concerns about anonymity. Virtual reality simulations could provide a more reliable and valid method for examining antisocial behavior across different conditions of legal deterrents. In addition, over half of participants were familiar with The Purge (57.3%), and familiarity was positively associated with purging and most crime categories. Many of those willing to purge reported prior exposure to the film, raising the possibility that sensationalized portrayals inflated responses or that individuals inclined toward crime are drawn to such media. Although controlling for familiarity did not alter the main findings, future studies should consider less recognizable framings to reduce potential priming effects. Finally, four purge items were excluded due to low factor loadings, highlighting the need to refine the measure (see the Online Supplemental Materials), and future work should adopt a confirmatory approach with a priori hypotheses informed by these exploratory findings.

The findings also highlight several avenues for future research. Although some participants indicated they would not offend in the absence of legal controls, the reasons for this remain unclear. Drawing on SAT, one explanation is that individuals refrain because they perceive certain acts as morally unacceptable. Future studies should assess the moral acceptability of specific offenses and test whether such evaluations account for compliance, as well as how they differ across crime types and personality traits. Future research should also compare the effects of factors that influence both offending and compliance in the absence of legal controls, including moral filters (Mann et al. 2016; Wikström 2007), social bonds such as attachment to institutions (Gottfredson and Hirschi 1990), perceived rewards of offending, and broader structural conditions like inequality and education (Entorf and Spengler 2000). Examining their relative influence within and between individuals across different crime types could clarify mechanisms of compliance and inform prevention strategies that strengthen non-legal forms of social control.

Moreover, explanations for the link between low self-control, psychopathic traits, and offending often emphasize insensitivity to legal costs (Altikriti, Nedelec, and Silver 2021; Gottfredson and Hirschi 1990), yet our findings and those of Meldrum, Lehmann, and Flexon (2021) suggest these associations extend beyond cost considerations. Future research should examine alternative mechanisms—such as heightened motivation to offend, reduced concern for others’ harm (Cima, Tonnaer, and Hauser 2010), and increased reward sensitivity (Altikriti, Nedelec, and Barnes 2022)—to clarify how these traits contribute to offending and identify potential targets for intervention. Work is also needed to clarify the conceptual overlap between psychopathic traits and low self-control. A promising direction is the application and further psychometric validation of integrated instruments like the recently developed Low Self-Control Psychopathic Traits Scale (LSCPTS; Pechorro et al. 2025), which combines two factors from each construct into a unified four-factor measure. In addition, person-centered approaches like latent profile analysis could examine how different configurations of psychopathic and self-regulatory facets predict distinct forms of antisocial behavior and offending trajectories.

Implications

Policymakers often advocate for severe punishment to prevent crime (Mears and Stafford 2024; Pratt 2008). Yet research shows that many individuals comply with the law without the threat of legal sanctions (Brauer and Tittle 2017; Mann et al. 2016; Paternoster 2010; Pogarsky 2002). This is especially true for severe crimes, where most individuals are deterred by extralegal factors rather than legal punishment. Paradoxically, severe crimes often attract the strongest calls for harsher punishment (Mears and Stafford 2024; Paternoster 2010). Such demands are typically rooted in retributive logic, aimed at imposing punishment proportionate to the harm caused. While retribution is a legitimate function of the criminal justice system, prevention requires policies grounded in scientific evidence and cost-effectiveness (Welsh, Zane, and Mears 2024). In this regard, investing heavily in legal controls that target behaviors already constrained by extralegal factors may be inefficient (Hofer 2024). Thus, although legal controls play an important role in reinforcing moral norms (Mulder 2018), deterring minor offenses, and preventing serious crimes among those responsive to legal consequences, lasting reductions in crime are more likely when such measures are complemented by policies that strengthen the social and psychological conditions that foster compliance and mitigate risk factors before offending occurs.

In this context, the present findings highlight low self-control and psychopathic traits as robust predictors of offending intentions, with psychopathic traits especially salient for severe crimes, making both traits important targets for primary and secondary prevention (Reidy et al. 2015). Because individuals high in these traits are not only more prone to offending but also less responsive to the legal costs of crime (Altikriti, Nedelec, and Silver 2021), prevention efforts must extend beyond legal punishment. Although skepticism remains regarding the malleability of such traits, evidence indicates they can be addressed, particularly in youth, through interventions that target callous-unemotional traits, strengthen emotional functioning and self-regulation, and reduce criminogenic thinking (Piquero et al. 2016; Salekin, Worley, and Grimes 2010). Promising results may also come from targeting mediating mechanisms that link these traits to offending (Altikriti, Nedelec, and Silver 2021). In parallel, prevention efforts should aim to reduce the perceived rewards of crime and remove structural barriers to prosocial alternatives (see Mears and Stafford 2024), especially for reward-sensitive individuals like those with elevated psychopathic traits (Altikriti, Nedelec, and Barnes 2022). A comprehensive crime prevention strategy should therefore combine legal deterrence with evidence-based public health approaches that address underlying risk factors and emphasize primary and secondary prevention, with the goal of sustainably reducing violence and crime.

Conclusion

This study replicated and extended work of Meldrum, Lehmann, and Flexon (2021), addressing prior methodological limitations and contributing to ongoing replication efforts in the behavioral sciences. Results showed that while many individuals reported a willingness to offend in the absence of legal controls, minor offenses were endorsed more frequently than severe crimes. These findings indicate that legal punishment is not the sole explanation for law-abiding behavior, particularly in relation to more serious offenses. Individuals with low self-control and high psychopathic traits were more likely to endorse nearly all crime types, underscoring their importance as psychological risk factors. Although legal sanctions remain essential—deterring minor offenses and constraining certain individuals from more severe ones—sustainable crime reduction likely requires complementing punitive measures with evidence-based prevention strategies. In particular, primary and secondary prevention efforts that address underlying risk factors, such as self-control deficits and psychopathic traits, may be critical for reducing violence, victimization, and societal harm.

Supplemental Material

sj-docx-1-10.1177_00224278251394278 - Supplemental material for Beyond Legal Deterrence: A Replication and Extension Study on Psychopathy, Self-Control, and Crime-Specific Willingness to Offend

Supplemental material, sj-docx-1-10.1177_00224278251394278 for Beyond Legal Deterrence: A Replication and Extension Study on Psychopathy, Self-Control, and Crime-Specific Willingness to Offend by Valentina Macías-Vasileff, Stefan Bogaerts, Elien De Caluwé, Yasemin Erbaş and Carlo Garofalo in Journal of Research in Crime and Delinquency

Footnotes

Acknowledgments

The authors would like to thank the psychology undergraduate students for their assistance with data collection, and all participants for their time and contribution to the study.

Author Contributions

All authors contributed to the conception and design of the study. Material preparation and data collection were led by Valentina Macías-Vasileff, with the collaboration and supervision of all co-authors. Valentina Macías-Vasileff conducted the analyses and drafted the manuscript. All co-authors provided critical feedback, reviewed, and edited manuscript drafts. All authors read and approved the final version of the manuscript.

Data Availability

Data will be provided upon request.

Declaration of Conflicting Interests

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

Ethical Approval

The study was conducted in accordance with the ethical standards outlined in the 1964 Declaration of Helsinki and its subsequent amendments. Ethical approval was granted, and all participants provided informed consent.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Valentina Macías-Vasileff

Stefan Bogaerts

Elien De Caluwé

Yasemin Erbaş

Carlo Garofalo

Notes

Supplemental Material

Supplemental material for this article is available online.

Correction (December 2025):

The references has been alphabetically updated in the article since its original publication.

Author Biographies

Valentina Macías-Vasileff is a PhD candidate at Tilburg University, The Netherlands. Her research examines how individual differences in moral functioning, personality traits, and psychopathy contribute to antisocial and criminal behavior.

Stefan Bogaerts is a full professor at Tilburg University, The Netherlands. His research interests include forensic and developmental psychopathology, psychopathy, sexual and violent offending, and the development and evaluation of treatment programs for offenders.

Elien De Caluwé is an associate professor of forensic psychology at Tilburg University, the Netherlands. Her research interests include the intersection between personality, (developmental) psychopathology and personality pathology.

Yasemin Erbaş is an assistant professor at Tilburg University, The Netherlands. Her research concerns emotion dynamics, individual differences, and their links to mental health, often using experience sampling methods.

Carlo Garofalo is senior assistant professor at the University of Perugia (Italy) and board member of the Professional Council of Psychologists - Region Umbria. His research interests concern the assessment, development, and manifestation of externalizing psychopathology and behavior across the lifespan.

References

Adler

N. E.

Epel

E. S.

Castellazzo

Ickovics

J. R.

. 2000. “Relationship of Subjective and Objective Social Status with Psychological and Physiological Functioning: Preliminary Data in Healthy White Women.” Health Psychology 19(6):586–92. https://doi.org/10.1037/0278-6133.19.6.586

Altikriti

Nedelec

J. L.

Silver

. 2021. “The Role of Arrest Risk Perception Formation in the Association Between Psychopathy and Aggressive Offending.” Youth Violence and Juvenile Justice 19(4):402–22. https://doi.org/10.1177/15412040211029991

Altikriti

Nedelec

J. L.

Barnes

J. C.

. 2022. “The Influence of Individual Differences on the Formation of Perceptions of Risk, Social Cost, and Rewards of Crime: A Meta-Analysis.” Journal of Criminal Justice 82:101962. https://doi.org/10.1016/j.jcrimjus.2022.101962

Altikriti

Theocharidou

Sullivan

C. J.

. 2020. “Specific Theories of Crime? A Longitudinal Assessment of the Competing Effects of Psychopathy and Self-Control.” Journal of Crime and Justice 43(5):547–67. https://doi.org/10.1080/0735648X.2020.1727765

Awang

2012. A Handbook on SEM Structural Equation Modelling: SEM Using AMOS Graphic. 5th ed. Kota Baru: Universiti Teknologi Mara Kelantan.

Bartlett

M. S.

1951. “The Effect of Standardization on a Chi-Square Approximation in Factor Analysis.” Biometrika 38(3-4):337–44. https://doi.org/10.1093/biomet/38.3-4.337

Becker

G. S.

1968. “Crime and Punishment: An Economic Approach.” Journal of Political Economy 76(2):169–217. https://www.jstor.org/stable/1830482

Bouffard

J. A.

Exum

M. L.

Niebuhr

. 2018. “Examining the Stability and Predictors of Deterrability Across Multiple Offense Types Within a Sample of Convicted Felons.” Journal of Criminal Justice 57:76–88. https://doi.org/10.1016/j.jcrimjus.2018.04.005

Brauer

J. R.

Tittle

C. R.

. 2017. “When Crime Is Not an Option: Inspecting the Moral Filtering of Criminal Action Alternatives.” Justice Quarterly 34(5):818–46. https://doi.org/10.1080/07418825.2016.1226937

10.

Cerny

B. A.

Kaiser

H. F.

. 1977. “A Study of a Measure of Sampling Adequacy for Factor-Analytic Correlation Matrices.” Multivariate Behavioral Research 12(1):43–7. https://doi.org/10.1207/s15327906mbr1201_3

11.

Chambliss

W. J.

1967. “Types of Deviance and the Effectiveness of Legal Sanctions”. Wisconsin Law Review 703(3): 703–19.

12.

Cima

Tonnaer

Hauser

M. D.

. 2010. “Psychopaths Know Right from Wrong but Don’t Care.” Social Cognitive and Affective Neuroscience 5(1):59–67. https://doi.org/10.1093/scan/nsp051

13.

Clark

L. A.

Watson

. 2019. “Constructing Validity: New Developments in Creating Objective Measuring Instruments.” Psychological Assessment 31(12):1412–27. https://doi.org/10.1037/pas0000626

14.

de Ridder

D. T. D.

de Boer

B. J.

Lugtig

Bakker

A. B.

van Hooft

E. A. J.

. 2011. “Not Doing bad Things Is Not Equivalent to Doing the Right Thing: Distinguishing Between Inhibitory and Initiatory Self-Control.” Personality and Individual Differences 50(7):1006–11. https://doi.org/10.1016/j.paid.2011.01.015

15.

DeLisi

2009. “Psychopathy Is the Unified Theory of Crime.” Youth Violence and Juvenile Justice 7(3):256–73. https://doi.org/10.1177/1541204009333834

16.

DeLisi

Tostlebe

Burgason

Heirigs

Vaughn

. 2018. “Self-Control Versus Psychopathy: A Head-to-Head Test of General Theories of Antisociality.” Youth Violence and Juvenile Justice 16(1):53–76. https://doi.org/10.1177/1541204016682998

17.

DeLisi

Bouffard

J. A.

Miller

H. A.

. 2022. “Another Look at the Self-Control vs. Psychopathy Debate: A Study Assessing Sexual Aggression, Aggression, and Substance Abuse.” American Journal of Criminal Justice 47(1):23–40. https://doi.org/10.1007/s12103-020-09582-w

18.

Dölling

Entorf

Hermann

Rupp

. 2009. “Is Deterrence Effective? Results of a Meta-Analysis of Punishment.” European Journal on Criminal Policy and Research 15:201–24. https://doi.org/10.1007/s10610-008-9097-0

19.

Entorf

Spengler

. 2000. “Socioeconomic and Demographic Factors of Crime in Germany: Evidence from Panel Data of the German States.” International Review of Law and Economics 20:75–106. https://doi.org/10.1016/S0144-8188(00)00022-3

20.

Flexon

J. L.

Meldrum

R. C.

Young

J. T. N.

Lehmann

P. S.

. 2016. “Low Self-Control and the Dark Triad: Disentangling the Predictive Power of Personality Traits on Young Adult Substance Use, Offending and Victimization.” Journal of Criminal Justice 46:159–69. https://doi.org/10.1016/j.jcrimjus.2016.05.006

21.

Flexon

J. L.

Meldrum

R. C.

. 2013. “Adolescent Psychopathic Traits and Violent Delinquency: Additive and Nonadditive Effects with Key Criminological Variables.” Youth Violence and Juvenile Justice 11(4):349–69. https://doi.org/10.1177/1541204012470850

22.

Fox

DeLisi

. 2019. “Psychopathic Killers: A Meta-Analytic Review of the Psychopathy-Homicide Nexus.” Aggression and Violent Behavior 44:67–79. https://doi.org/10.1016/j.avb.2018.11.005

23.

Gillespie

S. M.

Jones

Garofalo

. 2023. “Psychopathy and Dangerousness: An Umbrella Review and Meta-Analysis.” Clinical Psychology Review 100:102240. https://doi.org/10.1016/j.cpr.2022.102240

24.

Gottfredson

M. R.

Hirschi

. 1990. A General Theory of Crime. Stanford, California: Stanford University Press.

25.

Hofer

2024. “Red Light Game Identifies Ineffective Criminal Deterrence.” The American Journal of Economics and Sociology 83(3):673–82. https://doi.org/10.1111/ajes.12572

26.

Hosmer

D. W.

Lemeshow

Sturdivant

R. X.

. 2013. Applied Logistic Regression. 3rd ed. Hoboken, NJ: Wiley and Sons.

27.

Howitt

Cramer

. 2011 “Chapter 30: Factor Analysis: Simplifying Complex Data.” Pp. 362–79 in Introduction to Statistics in Psychology. 5th ed Harlow, UK: Pearson.

28.

L. T.

Bentler

P. M.

. 1999. “Cutoff Criteria for Fit Indexes in Covariance Structure Analysis: Conventional Criteria Versus New Alternatives.” Structural Equation Modeling: A Multidisciplinary Journal 6:1–55. https://doi.org/10.1080/10705519909540118

29.

Jones

D. N.

Paulhus

D. L.

. 2011 “Differentiating the Dark Triad Within the Interpersonal Circumplex.” Pp. 249–69 in Handbook of Interpersonal Psychology: Theory, Research, Assessment, and Therapeutic Interventions Horowitz

L. M.

Strack

New York: Wiley & Sons.

30.

Jones

D. N.

Paulhus

D. L.

. 2014. “Introducing the Short Dark Triad (SD3): A Brief Measure of Dark Personality Traits.” Assessment 21(1):28–41. https://doi.org/10.1177/1073191113514105

31.

Maloney

P. W.

Grawitch

M. J.

Barber

L. K.

. 2012. “The Multi-Factor Structure of the Brief Self-Control Scale: Discriminant Validity of Restraint and Impulsivity.” Journal of Research in Personality 46(1):111–5. https://doi.org/10.1016/j.jrp.2011.10.001

32.

Mann

Garcia-Rada

Hornuf

Tafurt

. 2016. “What Deters Crime? Comparing the Effectiveness of Legal, Social, and Internal Sanctions Across Countries.” Frontiers in Psychology 7:Article 85. https://doi.org/10.3389/fpsyg.2016.00085

33.

Mascolo

2023. “List of Criminal Charges A to Z.” FindLaw. https://www.findlaw.com/criminal/criminal-charges/view-all-criminal-charges.html.

34.

McNeeley

Warner

J. J.

. 2015. “Replication in Criminology: A Necessary Practice.” European Journal of Criminology 12(5):581–97. https://doi.org/10.1177/1477370815578197

35.

Mears

D. P.

Stafford

M. C.

. 2024. “A Theoretical Critique of Deterrence-Based Policy.” Journal of Criminal Justice 95:102305. https://doi.org/10.1016/j.jcrimjus.2024.102305

36.

Meldrum

R. C.

Lehmann

P. S.

Flexon

J. L.

. 2021. “Who Would ‘Purge’? Low Self-Control, Psychopathy, and Offending in the Absence of Legal Controls.” Crime & Delinquency 67(10):1582–613. https://doi.org/10.1177/0011128720940953

37.

Mulder

L. B.

2018. “When Sanctions Convey Moral Norms.” European Journal of Law and Economics 46(3):331–42. https://doi.org/10.1007/s10657-016-9532-5

38.

Nagin

D. S.

2013. “Deterrence in the Twenty-First Century.” Crime and Justice 42(1):199–263. https://doi.org/10.1086/670398

39.

Paternoster

2010. “How Much Do We Really Know About Criminal Deterrence.” Journal of Criminal Law and Criminology 100(3):765–824.

40.

Pechorro

Curtis

DeLisi

Maroco

Nunes

. 2022. “Dark Triad Psychopathy Outperforms Self-Control in Predicting Antisocial Outcomes: A Structural Equation Modeling Approach.” European Journal of Investigation in Health, Psychology and Education 12(6):549–62. https://doi.org/10.3390/ejihpe12060041

41.

Pechorro

Cordeiro

Maroco

Simões

M. R.

DeLisi

. 2025. “The Antisocial Core: A Chimeric Measure of Low Self-Control and Psychopathic Traits.” Justice Quarterly: 1–21. https://doi.org/10.1080/07418825.2025.2479155

42.

Piquero

A. R.

Paternoster

Pogarsky

Loughran

T. A.

. 2011. “Elaborating the Individual Difference Component in Deterrence Theory.” Annual Review of Law and Social Science 7:335–60. https://doi.org/10.1146/annurev-lawsocsci-102510-105404

43.

Piquero

A. R.

Jennings

W. G.

Farrington

D. P.

Diamond

Gonzalez

J. M. R.

. 2016. “A Meta-Analysis Update on the Effectiveness of Early Self-Control Improvement Programs to Improve Self-Control and Reduce Delinquency.” Journal of Experimental Criminology 12(2):249–64. https://doi.org/10.1007/s11292-016-9257-z

44.

Pogarsky

2002. “Identifying “Deterrable” Offenders: Implications for Research on Deterrence.” Justice Quarterly 19(3):431–52. https://doi.org/10.1080/07418820200095301

45.

Pratt

T. C.

2008. “Rational Choice Theory, Crime Control Policy, and Criminological Relevance.” Criminology and Public Policy 7(1):43–52. https://doi.org/10.1111/j.1745-9133.2008.00489.x

46.

R Core Team. 2023. “R: A Language and Environment for Statistical Computing.” In R Foundation for Statistical Computing. https://www.R-project.org/.

47.

Reidy

D. E.

Kearns

M. C.

DeGue

Lilienfeld

S. O.

Massetti

Kiehl

K. A.

. 2015. “Why Psychopathy Matters: Implications for Public Health and Violence Prevention.” Aggression and Violent Behavior 24:214–25. https://doi.org/10.1016/j.avb.2015.05.018

48.

Revelle

2022. “psych: Procedures for Personality and Psychological Research.” In (Version 2.2.5). https://CRAN.R-project.org/package=psych.

49.

Salekin

R. T.

Worley

Grimes

R. D.

. 2010. “Treatment of Psychopathy: A Review and Brief Introduction to the Mental Model Approach for Psychopathy.” Behavioral Sciences & the Law 28(2):235–66. https://doi.org/10.1002/bsl.928

50.

Sampson

R. J.

Laub

J. H.

. 2003. “Life-Course Desisters? Trajectories of Crime among Delinquent Boys Followed to Age 70.” Criminology; An Interdisciplinary Journal 41(3):555–92. https://doi.org/10.1111/j.1745-9125.2003.tb00997.x

51.

Spreen

Brand

Ter Horst

Bogaerts

. 2014. Handleiding en Methodologische Verantwoording HKT-R, Historisch, Klinische en Toekomstige - Revisie [Guidelines and Methodological Research of the HKT-R, Historical, Clinical and Future – Revision]. Dr. van Mesdag kliniek.

52.

Tangney

J. P.

Baumeister

R. F.

Boone

A. L.

. 2004. “High Self-Control Predicts Good Adjustment, Less Pathology, Better Grades, and Interpersonal Success.” Journal of Personality 72(2):271–324. https://doi.org/10.1111/j.0022-3506.2004.00263.x

53.

Tharshini

N. K.

Ibrahim

Kamaluddin

M. R.

Rathakrishnan

Che Mohd Nasir

. 2021. “The Link Between Individual Personality Traits and Criminality: A Systematic Review.” International Journal of Environmental Research and Public Health 18(16):8663. https://doi.org/10.3390/ijerph18168663

54.

Vazsonyi

A. T.

Mikuška

Kelley

E. L.

. 2017. “It's Time: A Meta-Analysis on the Self-Control-Deviance Link.” Journal of Criminal Justice 48:48–63. https://doi.org/10.1016/j.jcrimjus.2016.10.001

55.

Welsh

B. C.

Zane

S. N.

Mears

D. P.

. 2024. “Evidence-Based Policy in a New Era of Crime and Violence Prevention and Social Justice.” Aggression and Violent Behavior 77:101940. https://doi.org/10.1016/j.avb.2024.101940

56.

Wikström

P. H.

2007. “Deterrence and Deterrence Experiences: Preventing Crime Through the Threat of Punishment.” Pp. 362-76 in International Handbook of Penology and Criminal Justice Shoran

Beck

Kett

New York: Routledge.

57.

Wikström

P. H.

Tseloni

Karlis

. 2011. “Do People Comply with the Law Because They Fear Getting Caught?” European Journal of Criminology 8(5):401–20. https://doi.org/10.1177/1477370811416415

58.

Wickham

Averick

Bryan

Chang

D’Agostino McGowan

François

Grolemund

Hayes

Henry

Hester

Kuhn

Pedersen

T. L.

Miller

Bache

S. M.

Müller

Ooms

Robinson

Seidel

D. P.

Spinu

Yutani

. 2019. “Welcome to the Tidyverse.” Journal of Open Source Software 4(43):1686. https://doi.org/10.21105/joss.01686

59.

Wright

J. P.

Morgan

M. A.

Almeida

P. R.

Almosaed

N. F.

Moghrabi

S. S.

Bashatah

F. S.

. 2017. “Malevolent Forces: Self-Control, the Dark Triad, and Crime.” Youth Violence and Juvenile Justice 15(2):191–215. https://doi.org/10.1177/1541204016667995

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.17 MB