Public perceptions of the seriousness of crime: Weighing the harm and the wrong

Introduction

The seriousness of crime, or ‘crime seriousness’, features prominently in at least four areas of contemporary criminal policies: sentencing, criminalization, crime control and prevention. In several jurisdictions (for example, Finland, Sweden, the UK and several US states), sentencing guidelines explicitly require the severity of the penalty to be proportionate to the seriousness of the offence, which, in turn, depends on a combination of the harm done or risked through the commission of the offence and the offender’s culpability (for example, Von Hirsch and Ashworth, 2005); in others (for example, France, Germany, Italy), maximum sentences reflect the presumed seriousness of the offence. Crime seriousness also constitutes an important criterion for deciding whether to criminalize particular conduct (Ashworth, 2006: 35–42). In addition, scholars and practitioners have promoted seriousness as a tool for allocating resources and identifying priorities in crime control and prevention (for example, Office for National Statistics, 2016; Sherman et al., 2016). At the same time, ‘serious crime’, which typically encompasses transnational organized crime, such as drug and human trafficking, has become an organizing framework for criminal policy in the European Union (EU) ¹ and elsewhere (Paoli et al., 2016). ²

To establish the seriousness of crime, policy-makers have turned to the considerable body of research on public perceptions of crime seriousness that has emerged from Sellin and Wolfgang’s 1964 landmark study, and demonstrated a consensus across social groups and countries in the ranking of crimes (see Stylianou, 2003, for a review). England and Wales, for example, accounted for public perceptions in the development of sentencing guidelines (see Sentencing Guidelines Council, 2004). In their turn, several Anglo-American law enforcement and policy-making agencies use sentencing guidelines or data as benchmarks for prioritizing prosecution and allocating resources (for example, UK Metropolitan Police Service in Tusikov, 2012) or for mapping crime trends (for example, Office for National Statistics, 2016). The Belgian police (Charlier et al., 2006) also factor public concern for different crimes into their long-term priorities.

Scholars who, in line with Sellin and Wolfgang (1964, 1991), view public perceptions of seriousness as an important measure of crime seriousness for policy-making support these policy developments, at least conceptually. In their theory of ‘intuitive justice’, Robinson and Darley (for example, 2007) argue that the public can give nuanced and sophisticated assessments of both the offence seriousness and the offender’s blameworthiness. ³ They contend that only such ‘empirical deserts’ can guarantee the legitimacy and moral credibility of criminal policy. In the extreme, Tonry (2015: 665), argues that the ‘literature on citizens’ rankings of the seriousness of crime tell us all we need to know for purposes of sentencing and corrections policy’.

By comparison, other scholars (for example, Roberts, 2011; Von Hirsch and Jareborg, 1991) stress the limitations and risks of relying solely on public opinion in sentencing, by raising the possibility of knowledge deficits, factual misjudgement, unprincipled attitudes and volatility of perceptions. In a first review of the literature on crime seriousness perceptions, Rossi and Henry (1980: 500) specifically warned that ‘without additional knowledge about the nature of crime seriousness and clear theoretical understanding of how seriousness judgments are made, it would be hazardous to run the criminal justice system by looking to crime seriousness as a guide’. Similarly, Stylianou (2003: 48), who provides the most recent review, finds that the ‘normative philosophical principles’ the public applies in its seriousness judgements constitutes an ‘issue that deserves more attention’.

The few studies (for example, Piquero et al., 2008) that have emerged more recently have not addressed these gaps; hence, this article pursues two aims. First, it explores the logic – or ‘normative philosophical principles’ (Stylianou, 2003: 48) – with which the public assesses crime seriousness and, second, it considers the extent to which that logic aligns with legal principles and policy requirements. In particular, it uses survey data and statistical methods to evaluate the relative weights of ‘consequentialist’ and ‘moralist’ concerns in the public’s assessments of crime seriousness. These two approaches map to the two main philosophies of criminal law (Duff, 2013) and to Warr’s (1989) specification of perceived ‘crime seriousness’, which entails a combination of perceived ‘harmfulness’ and ‘wrongfulness’.

In the following sections we examine criminological research and legal theory on crime seriousness and its perception; introduce our conceptualization of crime seriousness; and present our data and analysis. We expand on Warr’s specification by parsing harmfulness in terms of the severity and incidence of harms (Greenfield and Paoli, 2013), and distinguishing between ‘typical’ and ‘total’ harmfulness. Our analyses are based on data from a survey of public perceptions of seriousness that address wrongfulness, each component of harmfulness, and an overall measure of seriousness. The survey, which elicited 1278 valid responses, was conducted in Flanders (the Dutch-speaking part of Belgium) in 2014. The final section summarizes our findings and discusses their implications for criminalization, crime control and prevention prioritization, and sentencing.

Crime seriousness and its perception in criminology and criminal law

The criminological research on crime seriousness and its perception has long been characterized by a ‘distressing lack of theoretical base’ (Rossi and Henry, 1980: 497; see Stylianou, 2003). Until the late 1980s many studies assumed that a definition of crime seriousness was unnecessary (for example, Rossi et al., 1974) and simply asked respondents how ‘serious’ they thought a certain crime was. Perceived crime seriousness was first explicitly conceptualized by Warr (1989), who identified the perceived ‘harmfulness’ and ‘wrongfulness’ of a criminal act as the two main dimensions of perceived seriousness. According to Warr, perceived harmfulness is established through a ‘factual assessment of the consequences of the offence upon the victim’. Perceived wrongfulness, instead, is the result of a ‘a normative evaluation’ of the ‘moral gravity of committing the act’ (ibid.: 796). ⁴ A Dallas survey (Warr, 1989) and two later studies (O’Connell and Whelan, 1996; Rosenmerkel, 2001) provide empirical support for this operationalization. ⁵

There is also no consensus in the literature on which dimension, if any, dominates perceptions of seriousness. Warr (1989) and Rosenmerkel (2001) show that, for most respondents, either wrongfulness or harmfulness can drive seriousness ratings, depending on the type of crime under consideration. O’Connell & Whelan (1996) and Alter et al. (2007) find that their respondents emphasize wrongfulness over harmfulness. In his literature review, Stylianou (2003: 42–3) argues that ‘the most important characteristic associated with perceived seriousness of an act is the act’s perceived consequences’, adding that ‘this conclusion has been confirmed by virtually all studies’. In his view, moral wrongfulness is relevant only for ‘victimless crimes’, which ‘can be judged as more or less serious based on the extent to which they violate the moral standards of society’. Likewise, Eisner et al. (2017: 63–64) conclude that the seriousness ranking mainly mirrors intuitions about the size of the inflicted harm”.

The idea of actual ‘crime seriousness’, as distinct from its perception, has also been conceptualized poorly (Paoli et al., 2016), but the underlying dimensions of seriousness, ‘harmfulness’ and ‘wrongfulness’, have received considerable attention in criminological reflections, empirical studies and criminal law.

‘Harm’, in particular, is gaining ground as a policy goal and analytical construct. Restorative justice and victimological studies (for example, Braithwaite, 2004; Ignatans and Pease, 2016; Walklate, 2007) centre on harm, and some scholars (for example, Rubin, 1999; Sherman, 2007) recommend harm minimization as the overall goal of criminal policies. The concept of harm has also been used to draw attention to, and promote the criminalization of, environmentally detrimental activities and human-rights violations (for example, Rothe and Kauzlarich, 2014; White, 2011). Some critical criminologists (for example, Hillyard and Tombs, 2007) even propose replacing ‘crime’ with ‘social harm’. Sherman et al. (2016), Ratcliffe (2014) and others have introduced crime harm indices and Greenfield and Paoli (2013) have developed a comprehensive ‘harm assessment framework’ that we draw from to conceptualize crime seriousness below.

Moreover, both harmfulness and wrongfulness are well represented in legal theory, where they map to ‘consequentialism’ and ‘moralism’, the two main schools of thought on criminalization, sentencing and punishment. Mill ([1859] 1978: 9) offers the first example of the former in his famous articulation of the Harm Principle: ‘the only purpose for which power can be rightfully exercised over any member of a civilized community, against his will, is to prevent harm to others.’ Legal moralism instead makes the immorality of the conduct and the moral culpability of the offender the focal concern of criminal law. Whereas some scholars still defend the original positions (for example, Moore, 1997), others have advanced less extreme versions of both consequentialism and moralism and variously combined harmfulness with wrongfulness for the purposes of criminalization (see Duff, 2013).

Despite these differences, consequentialists and many contemporary legal moralists agree that the seriousness of the crime reflects the harm done or risked through the commission of the act. Moralists, such as Duff (2014), acknowledge that the moral wrongfulness of a conduct is very frequently derived from its harmfulness. The retributivists, among them, emphasize that the assessment of harm should provide a gauge in sentencing in addition to the offender’s culpability (Ashworth, 2006; Von Hirsch, 2017). For the consequentialists, harm is the crucial dimension, with wrongfulness merely representing a side-constraint on the consideration of harmfulness (Hart, 1968). Whereas moralists focus on the harm generated by the offence, consequentialists might also consider the harm that can be prevented in the future through the criminalization or sentencing of a specific act (see Duff, 2013).

These two approaches – consequentialism and moralism – also relate to policy. According to Tonry (2011: 6) and others, ‘punishment theories, institutions, policies, and practices in the English-speaking countries were based largely on consequentialist ideas’ until the early 1980s and the ascendance of retributivism.

Our conceptualization of crime seriousness and hypotheses

Our conceptualization of crime seriousness acknowledges the two dimensions of harmfulness and wrongfulness and, building on Greenfield and Paoli (2013), extends the concept of harmfulness. Greenfield and Paoli’s framework draws on Feinberg’s (1984) and defines harm – and by extension harmfulness – as a setback to legitimate interests. It also emphasizes ‘total harm’ (MacCoun, 1998; Sherman, 2007) and differentiates between severity and incidence, the latter of which depends on both the incidence of the crime itself and the incidence of harm in relation to that crime. ⁶ On that basis, we define ‘crime seriousness’ in terms of two plus two constitutive elements or ‘components’: (a) the wrongfulness of a crime, defined as the severity of the violation of moral norms and sentiments implied by a criminal activity; (b) the severity of the harms of crime, defined as the gravity of the injury or damage that the crime inflicts; (c) the incidence of the crime, defined as the frequency with which this type of crime actually happens; and (d) the incidence of the harms of a crime, defined as the frequency with which a type or range of harms results from a crime event.

Whereas incidence was excluded from the original understanding of seriousness, ⁷ we also consider the incidence of crime and of harms in relation to crime because they – depending on the policy uses – also contribute to seriousness and provide necessary information for crime control and prevention prioritization. A crime may entail the most severe harms but, if it occurs only rarely, it might merit less attention than a crime involving less serious harms but occurring more frequently. The question is, therefore, whether the respondents assess the seriousness of a standard event of a certain crime or the seriousness of all occurrences of that crime by simultaneously considering crime and harm incidence in addition to harm severity. The former – which we call ‘typical’ seriousness – is relevant for sentencing. The latter – which we label ‘total’ – is important for crime control and prevention priorities, but it may also play a role in other phases of criminal policy. Feinberg (1984: 216) and Simester and Von Hirsch (2011: 45), for example, stress the importance of the probability of harm – a concept related to incidence – in making policy decisions about criminalization. Total seriousness might also occasionally affect the development of sentencing guidelines. The Sentencing Council of England and Wales (Sentencing Guidelines Council, 2004: 9) states that ‘exceptional local circumstances … may lead a court to decide that prevalence (of crime) should influence sentencing levels’. In addition, knowledge about the incidence of harms in relation to broad offence categories can assist in defining minimum and maximum sentences (Sentencing Guidelines Council, 2004). The idea of total seriousness also informs the newer, but related, concept of serious crime, which has become prominent in criminal policy debates at the EU level, and elsewhere, since the 1990s.

Despite the policy relevance of total seriousness, previous, albeit inconclusive, research suggests that seriousness ratings imply ‘a typification based on stereotypical perceptions of offenders and events’ (Miethe, 1982: 518). This hypothesis is reinforced by research on related topics. Several experiments in psychology have shown what Nordgren and McDonnell (2011) call ‘scope insensitivity,’ that is, people’s insensitivity to the magnitude of outcomes. In their experiment, respondents rated a fraud case as more severe when they were told that it affected 3 people rather than 30 (see also Kahneman, 2011).

We also expand on Warr (1989) by delineating different ‘bearers’ of harm (or claimants) and types of harm (Greenfield and Paoli, 2013; Von Hirsch and Jareborg, 1991). We distinguish four categories of bearers, namely, individuals, private sector entities, government and society, to include, for example, problems of social order. Further, we identify four types of harm, consisting of damage to functional integrity (that is, physical and psychological integrity in the case of individuals and operational integrity in the case of private sector and government entities and society), material interest, reputation and privacy. ⁸ This specification goes beyond most harm-based theories of sentencing and criminalization, which typically focus on harms to individuals (for example, Simester and Von Hirsch, 2011; Von Hirsch and Jareborg, 1991). ⁹

Pursuant to the literature (see Miethe, 1982; Pease, 1988; Stylianou, 2003; Warr, 1989) and our conceptualization of crime seriousness, we posit three hypotheses with regard to the first aim of the article:

Methods

Results

Table 1 summarizes the sample characteristics. Compared with the general population, our sample is representative for gender but not for age and education. Young people under the age of 34 are underrepresented (χ² = 15.360, df = 4, p < .05), whereas those above 50 are overrepresented. Similarly, people with lower educational levels are underrepresented, whereas people with higher educational levels are overrepresented (χ² = 2068.370, df = 3, p < .05). Respondents of Belgian origin accounted for 86 percent of the sample. About one-third had been direct victims of theft (36.1 percent) and vandalism (31.7 percent) followed by burglary (14.3 percent). The victimization rates for physical and sexual assault and for fraud were below 10 percent. Two-thirds of the respondents had had indirect victimization experiences of theft (66.7 percent) and burglary (63.5 percent) and about half of vandalism (49.4 percent), whereas the rates of indirect victimization of other crimes were all lower than one-third.

OPEN IN VIEWER

Table 1.

Socio-demographic characteristics and victimization experiences of the respondents.

Variable	N	Percent
Gender
Male	596	47.0
Female	673	53.0
Age
18–24	116	9.2
25–34	162	12.8
35–49	322	25.5
50–64	372	29.4
65+	292	23.1
Educational level
None	33	2.6
Primary education	116	9.2
Secondary education	563	44.9
College	391	31.2
University	131	10.4
Origin
Belgian	1077	86.4
Other	169	13.6
Direct victimization
Burglary	180	14.3
Theft	457	36.1
Physical assault	108	8.5
Sexual assault	54	4.3
Fraud	98	7.8
Vandalism	402	31.7
Indirect victimization
Burglary	796	63.5
Theft	841	66.7
Physical assault	344	27.2
Sexual assault	126	10.0
Fraud	156	12.3
Vandalism	623	49.4

Figure 1 gives an overview over the 95% confidence intervals for all 10 crimes and the four components, following the rank order of seriousness of crimes, which is located at the left-hand side of Figure 1. Table 2 gives an overview of mean values and the resulting rank ordering for seriousness and all four components. (Both Figure 1 and Table 2 thus include the three crimes with more detailed measurement.) Given the severity of the crimes that the public had to judge, both seriousness and wrongfulness are located at the higher end of the scale and within a comparatively small range of mean values between nearly 100 for murder and 81 (seriousness) and 84 (wrongfulness) for cannabis trade. Though also located at the higher end of the scale, assessments of the severity of harm have a wider range, from 94 for murder to 67 for cannabis trade. The incidence of harms as a result of the crimes covers a similar range, from 82 for sexual assault to 56 for vandalism. The spread for the incidence of crimes is wider, from 86 for burglary to 33 for terrorism, though this reflects the situation before the attacks on Brussels airport in 2016.

OPEN IN VIEWER

Figure 1.

Ratings of crime seriousness and its components for 10 crimes (confidence intervals (95%)).

OPEN IN VIEWER

Table 2.

Rank order of crime seriousness and its components for 10 crimes (mean, standard deviation and t-test for differences between adjacent ranks (p < .01)).

Seriousness		M	SD	P	Wrongfulness	M	SD	P	Harm severity	M	SD	P	Incidence of crime	M	SD	P	Incidence of crime harms	M	SD	P
1.	Murder	99.50	4.06		Murder	99.51	3.36		Murder	93.96	13.62		Burglary	86.14	16.99		Sexual assault	82.26	14.06
2.	Terrorism	97.99	7.39	**	Terrorism	98.55	6.55	**	Terrorism	93.06	16.45	*	Theft	86.14	17.14	ns	Burglary	77.83	12.35	**
3.	Sexual assault	96.31	8.32	**	Sexual assault	97.04	7.95	**	Sexual assault	87.76	14.74	**	Vandalism	79.71	19.16	**	Theft	75.48	12.08	**
4.	Physical assault	92.44	11.92	**	Physical assault	93.84	11.74	**	Corporate fraud	75.68	20.62	**	Cannabis trade	75.94	20.16	**	Terrorism	70.26	20.08	**
5.	Burglary	89.75	14.36	**	Burglary	93.18	12.11	ns	Vandalism	74.39	16.94	ns	Physical assault	74.31	19.41	ns	Physical assault	69.45	13.20	ns
6.	Theft	88.68	15.13	**	Theft	92.13	12.37	**	Cocaine trafficking	74.32	11.83	ns	Cocaine trafficking	70.78	20.52	**	Cocaine trafficking	69.16	12.68	ns
7.	Cocaine trafficking	88.28	16.81	ns	Cocaine trafficking	90.69	14.85	**	Theft	72.18	17.29	**	Sexual assault	69.69	19.36	ns.	Corporate fraud	62.19	16.32	**
8.	Vandalism	85.40	17.02	**	Vandalism	89.70	13.87	ns	Burglary	71.85	11.70	ns	Corporate fraud	67.69	20.60	*	Vandalism	55.75	16.90	**
9.	Corporate fraud	84.16	17.91	*	Corporate fraud	87.30	16.63	**	Physical assault	69.85	10.31	**	Murder	51.84	17.10	**
10.	Cannabis trade	81.39	23.13	**	Cannabis trade	84.45	21.13	**	Cannabis trade	66.64	21.54	**	Terrorism	33.48	16.56	**

Note: ns: not significant at p < .01 level, * p < .01, ** p < .001.

Respondents rank violent crimes as the most serious, followed by burglary and theft, which are the most common property crimes. Crimes often related to organized crime, such as cocaine trafficking and cannabis trade, or to powerful actors, as in the case of corporate fraud, obtain low ranks for seriousness, along with vandalism.

The assessments of the wrongfulness of the crime deliver exactly the same ranking. In contrast, assessments of harm severity follow this ‘logic’ only for the three violent crimes that respondents have also deemed most serious and wrongful. Corporate fraud, cocaine trafficking and vandalism are seen as more harmful than the two property crimes of burglary and theft, and physical assault and cannabis trade are seen as the least harmful ones. The severity of harms mirrors neither the moral judgement nor the overall perceived seriousness of crime. Harm severity presumably reflects the ‘potential’ for large-scale harm that corporate fraud and cocaine trafficking have with many victims (or dependent users), whereas many of the property crimes or even assaults are deemed to be crime incidents that cause only little or minor harm.

When assessing the incidence or frequency of crime, respondents seem to rely largely on local knowledge from neighbourhoods and communities. Property crimes including damage to property – as in vandalism – are realistically seen as the most frequent crimes, followed by cannabis trade and physical assault, where presumably the minor incidents are kept in mind. Cocaine trafficking ranks sixth. Severe violent crimes such as sexual assault, murder and terrorism are correctly deemed much less frequent, as is corporate fraud, which might be less visible because it primarily affects higher social strata. The latest Belgian victimization survey, conducted in 2008, confirms a realistic ranking in the assessment of the incidence of four of the selected crimes by the Flemish public. Victimization rates for the last 12 months among the participants of the Belgian Safety Monitor were highest for the property crimes of burglary and theft (6.08 percent and 4.48 percent respectively). Physical assault had a lower victimization rate than property crimes (2.72 percent), sexual assault figured with 1.19 percent as the lowest victimization rate among the participants (Federale Politie, 2009).

The assessment of the incidence of harms from crime covered all crimes except murder and cannabis trade and, even with those exceptions, spanned a range of very harmful and presumably rather less harmful crime. Sexual assault is seen as most frequently entailing harms for individual victims, followed by the property crimes of burglary and theft. Around an average of 70 we find terrorism, physical assault and cocaine trafficking. Corporate fraud and vandalism are perceived as the crimes that least frequently entail harms for individual victims, as they might be seen to generate harm for communities and the government rather than for individuals.

Comparing across the different components as shown in Figure 1, the public’s overall assessments of the seriousness of these crimes are highly correlated with their moral judgements on these crimes (between a low of .636 for physical assault and a high of .903 for cocaine trafficking; all p < .001), and coincide with those judgements on these crimes. In contrast, the components that gauge the harmful consequences of crimes – in terms of both severity and frequency – follow different logics of assessment. These components represent a consequentialist view of crimes and seem to be independent of the general public’s assessments of seriousness and wrongfulness and the resulting rank orders. The ‘wrong’ and the ‘harm’ thus represent different rationales.

We also tested whether the means and the resulting ranks represented significant differences notwithstanding the compression of mean values in the respective assessments of the crimes. Because respondents had not been asked to compare pairs of crimes (Kwan et al., 2000, 2002), it is essential that the rank ordering represents meaningful distinctions. The results of pairwise t-tests for all consecutive back-to-back crimes in the respective rank order are presented in Table 2, indicating significant differences from one rank to the next higher one (p < .01).

Overall seriousness ratings all differ significantly with one exception: theft and cocaine trafficking. This implies that the assessment of the seriousness of the eight other crimes represents a perspective of distinct differences and a clear hierarchy of crimes. In the case of moral wrongfulness, there is no significant distinction between burglary and physical assault, possibly owing to the attack on the integrity of both body and home and on the general security of the victim. In addition, the difference between cocaine trafficking and vandalism is not significant. These rankings make a clear statement on the importance of individual victimization for the perceptions of moral wrongfulness and overall seriousness, as the crimes with no immediately identifiable victim rank at the bottom of the two rankings.

For the three harm-related components of seriousness, we find less distinct hierarchies. For harm severity, both property crimes are rated similarly. However, they are both set apart from the preceding group of crimes where no significant differences in terms of harm severity are found, namely corporate fraud, vandalism and cocaine trafficking. The latter group is particularly interesting because it seems to involve more ‘distant’ victims. For corporate fraud and cocaine trafficking, the victims and harmed groups might be socially distant, because they either belong to the higher strata of society or are seen as marginalized drug addicts. Vandalism is a crime where victims often are not individuals but the government or communities, and the damage is done to public goods. As regards harm severity, these crimes are situated in the middle between the most serious and violent crimes against individual victims, and the crimes for which many incidents result in only minor damage.

With regard to the incidence of crime events, the two property crimes (burglary and theft) are rated similarly. Cannabis trade and physical assault represent another group of similar assessments of incidence, as do cocaine trafficking and sexual assault. Consequently, the assessment of crime incidence comes in a much less clear hierarchical order than the moral judgements or the seriousness of crime.

As for the incidence of crime harms (without murder and cannabis trade), there is a clear ranking at the top, with sexual assault ranking first, followed by the two property crimes (burglary and theft) ranking second and third. Terrorism, physical assault and cocaine trafficking, in contrast, do not differ significantly from each other.

In sum, we can conclude that moral judgement produces a more distinct and clearer hierarchy of crimes than the ‘consequentialist’ components of severity and incidence. Though reasonably realistic in its judgements on severity and incidence, the less distinct hierarchies might indicate that the public is less confident in establishing crime rankings among these components. The coincidence between seriousness and moral wrongfulness ratings suggests that moral judgements might largely capture the public stance on crime seriousness. We explored this issue in the last step when we used regression analysis to test how moralistic and consequentialist logic informs and shapes the overall seriousness ratings. ¹⁴ Table 3 presents both the restricted model (Model 1) and the full model (Model 2). We first present the six crimes – burglary, theft, physical assault, sexual assault, vandalism and corporate fraud – for which we have measurements of all components and all controls, including victimization. We then consider murder, terrorism, cocaine trafficking and cannabis trade without controls for victimization, and for murder and cannabis trade without measures of crime harms.

OPEN IN VIEWER

Table 3.

Predictors of crime seriousness ratings for 10 crimes (standardized coefficients).

	Burglary(n = 683)		Theft(n = 717)		Physical assault (n = 701)		Sexual assault(n = 721)		Vandalism(n = 715)		Corporate fraud(n = 730)
	Model 1	Model 2	Model 1	Model 2	Model 1	Model 2	Model 1	Model 2	Model 1	Model 2	Model 1	Model 2
Components seriousness
Wrongfulness	.58***	.56***	.63***	.60***	.61***	.58***	.66***	.67***	.57***	.55***	.76***	.73***
Severity of crime harms	.24***	.22***	.15***	.15***	.17***	.16***	.16***	.16***	.25***	.24***	.12***	.09***
Incidence of harms	−.03	−.03	.04	.01	−.02	−.03	.03	.05	.11***	.09***	.03	.03
Incidence of crime	.05	.06*	−.04	−.01	−.05	−.01	.01	.01	−.05	−.03	.01	.01
Control variables
Gender (men = 0)		−.00		.01		.05		−.01		.04		.03
Age		.11***		.14***		.10**		.05*		.09***		.10***
SES		−.04		−.06*		−.08*		−.04		−.05		−.08***
Education (0 = none/ primary)		−.03		.00		.02		.01		−.00		.00
Origin (0 = Belgian)		.02		.06*		.03		.02		.01		−.01
Religion (0 = none)		.01		.06*		.04		−.01		.02		.00
Victimization (0 = no)
Direct victim		.06*		.05		−.04		.00		.01		.03
Indirect victim		−.04		−.02		−.01		−.01		−.01		.00
Adjusted R²	.46	.48	.51	.54	.43	.45	.57	.57	.60	.61	.67	.68
F change	146.50***	3.63***	184.00***	6.63***	132.69***	3.92***	237.03***	1.08	267.29***	2.94**	366.63***	5.69***

	Murder(n = 729)		Terrorism(n = 725)		Cocaine trafficking(n = 684)		Cannabis trade(n = 748)
	Model 1	Model 2	Model 1	Model 2	Model 1	Model 2	Model 1	Model 2
Components seriousness
Wrongfulness	.59***	.58***	.67***	.67***	.83***	.82***	.83***	.81***
Harm severity	.12***	.12***	.02	.02	.08***	.07**	.13***	.13***
Incidence of crime harms			.05	.04	−.01	−.00
Incidence of crime	−.02	−.02	−.02	−.02	.04	.03	.00	.01
Control variables
Gender (men = 0)		.01		.04		.02		.00
Age		.00		−.02		.07***		.07***
SES		.01		−.01		−.03		−.03
Education (0 = none/primary)		−.02		−.04		−.03		−.01
Origin (0 = Belgian)		.02		.06*		.02		.02
Religion (0 = none)		−.00		−.03		−.00		−.00
Adjusted R²	.38	.38	.47	.47	.77	.77	.84	.84
F Change	151.24***	0.17	160.64***	0.01	571.11***	3.41**	1272.16***	4.18***

Note: ns.: not significant at p < .05 level, * p < .05, ** p < .01, *** p < .001.

Table 3 demonstrates that across all crimes the components that shape and inform seriousness ratings have a consistent and similar impact; that is, different crimes are mostly judged by the same yardstick, even if for a particular crime additional information might be taken into account. The models explain between 38 percent (murder) and 84 percent (cannabis trade) of the variance, which are sufficiently high levels given that no other attitudes or beliefs are included in the model. The models demonstrate the weight of the moral wrongfulness and, to a much lesser extent, harm severity, with controls having little or rare impact on seriousness.

For all crimes, seriousness ratings are determined first and foremost by moral wrongfulness and moral judgement, rather than by the potential harmful consequences or frequency of these crimes. The coefficients of wrongfulness by far exceed the weight of harm severity, the next most influential component. ¹⁵ The incidence of harms as a consequence of a crime has no impact on seriousness ratings, with vandalism being the exception. This is in line with the results of Innes and Fielding (2002) that vandalism is a crime with a considerable ‘social harm footprint’ of high scale and high intensity and is experienced as ‘incivility’ in public space. The incidence of crime does not inform public assessments of the seriousness of crime in a significant or substantial way. The exception is burglary, where crime incidence has a small impact, presumably owing to experiences in local neighbourhoods and communities.

With regard to the control variables, they mostly do not increase the explained variance substantially and significantly. Only age has a consistent and significant impact, with the exception of the most serious crimes of murder and terrorism; older people generally rate crimes as more serious. Higher social status affects seriousness ratings for only a small number of crimes; this applies to physical assault but not to sexual assault, to theft but not to burglary, and to corporate fraud, where it reduces seriousness in all cases. Level of education is not related to seriousness ratings at all. Citizens of non-Belgian origin – either respondents themselves or their parents – do not differ from others in their seriousness ratings, with two exceptions where they find the crimes more serious: terrorism and theft. Being aligned to a faith does not change seriousness ratings either, with the exception of theft.

Indirect experiences of victimization are not related to seriousness at all. In our analysis, direct and own victimization changes assessments of the seriousness of crime only if the person has been a victim of a burglary; victims rate burglary as significantly more serious than do non-victims. Otherwise the experience of victimization does not change the assessment of the seriousness of a specific crime.

Discussion and conclusion

We specified two aims at the outset: first, to explore the logic – or normative philosophical principles – that the public uses to assess the seriousness of crime and, second, to consider how that logic aligns with legal principles and policy requirements. Our results are based on a small number of crimes (10) and a regional survey (Flanders), a relatively low response rate, a sample with some biases (age, education), subject to known methodological challenges (Kwan et al., 2002), and might raise questions of generalizability, but, with these caveats in mind, we address each aim in turn.

With regard to the first aim, the data confirm our first two hypotheses and, with some caveats, also the third. The public, as manifest in our sample of 1278 respondents, ranks crimes involving harm to individuals’ physical and psychological integrity as most serious, property crimes as second, and crimes with no immediate individual bearers of harms as the least serious (Hypothesis 1).

In assessing crime seriousness, the public shows ‘scope insensitivity’ (Nordgren and McDonnell, 2011) and makes stereotypical assessments of the seriousness of crimes. Among the three components that pertain to harmfulness, only the severity of harm – and thus typical harmfulness – is taken into account. The incidence of crime and crime harms, and thus total harmfulness, are ignored – even when the public makes realistic estimates of the incidence of crime (Hypothesis 2). Moreover, the public adheres to moralism more than to consequentialism: seriousness ratings are defined primarily by the perceived moral wrongfulness. The latter’s relatively strong predictive power confirms O’Connell and Whelan’s (1996) and Alter et al.’s (2007) findings on the primacy of wrongfulness, but it partially contradicts Warr’s (1989) and Rosenmerkel’s (2001) results as well as Stylianou’s (2003) and Eisner et al.’s (2017) summaries. Beyond the specific literature on perceived crime seriousness, it also empirically supports Wikström et al.’s (2012) conceptualization of crime as a breach of moral rules.

Concerning the third hypothesis, the data show that socio-demographics are generally not relevant, with the major exception of age. The same is generally true for victimization, in line with previous findings that crime victims rank crime seriousness similarly to non-victims (Pease, 1988) and without apparent inflation (Ignatans and Pease, 2016: 187).

As regards the second aim, the data suggest that the Flemish public’s ratings are irreconcilable with pure consequentialist theories of criminalization (for example, Mill, 1978), which only acknowledge harmfulness, but that they are not wholly incompatible with contemporary theories and practices of criminalization that assign a role to wrongfulness and harmfulness. However, given the limited weight that the public accords to harm in its perceptions of seriousness, the public’s ratings might not be as useful as is sometimes claimed for the purposes of criminalization, crime control and prevention priority-setting, unless a society follows a predominantly moralistic path in these policy areas.

If a society accepts that policy deliberations on both criminalization and/or priority-setting should also take into account the consequences of crime – namely, its overall harm or ‘total harm’ – then perceptions alone cannot provide adequate guidance, because those consequences would fade from view. Illustrative of this disconnect, the Flemish public’s approach to seriousness, which treats wrongfulness as paramount, deviates markedly from the principles set by EU policy documents (for example, TFEU, 2008), which advance harm-based criteria for the purposes of criminalization and priority-setting in the control of serious crime. Driven by a moralistic logic, the public’s ranking of the selected crimes also undermines the EU’s list of serious crimes. With the exception of terrorism, the crimes that figure prominently in the EU’s policies against serious crime, such as cocaine trafficking and corporate fraud, occupy bottom ranks in terms of overall seriousness and wrongfulness, even if they have a mid-position in terms of harm severity. Cannabis trade, which also figures prominently in the EU’s list of serious crimes, is consistently ranked by the public as the least serious, least wrongful and least harmful.

A further question is whether citizens’ seriousness ratings should be taken into account for sentencing guidelines and specific sentencing decisions (Tonry, 2015). The answer, again, depends on the sentencing theory under consideration. Pending future assessments for other crimes and contexts, our findings would not back a consequentialist approach to sentencing, if actual seriousness is reduced to its public perceptions. To the extent that the moral wrongfulness of the act is separate from the offender’s culpability, as many retributivist scholars advocate (for example, Simester and Von Hirsch, 2011: 23), they might even not support retributivism, because the latter understands the seriousness of the offence as a combination of offender’s culpability and the harm done or risked by the act (for example, Von Hirsch, 2017).

Andrew Ashworth (2006: 39), a leading retributivist, notes that ‘the task of assessing the seriousness of the offence is … as complex and problematic as it is unavoidable and fundamental’. While we agree with Ashworth, our findings imply that it might be risky to assess such seriousness only through surveys of public perceptions and to make such perceptions the main benchmark of criminal policies. On the other hand, our findings might suggest a possible division of labour among ‘experts’ and the public. Whereas experts might be called to provide an ‘objective’, systematic assessment of the actual harmfulness of criminal and not-yet-criminalized activities, the public as a whole might be charged with assessing the wrongfulness of such activities on the basis of ‘the aims and values of … [the] polity’ (Duff, 2007: 142). However, to the extent that most judgements about the moral wrongfulness of crimes depend on their harmfulness, as Duff (2014) himself and others suggest, even such judgements would be stronger if they were informed by an empirical assessment of crime harms.

Supplementary Material