Naming the threat: lay prototypes of organized crime in Italy and the US

ABSTRACT

Organized crime represents a pervasive global threat while posing unique challenges to law enforcement and policymakers. Yet, research about public understanding of this phenomenon remains limited. This research, comprising three studies conducted in Italy (N_tot = 477) and the United States (N_tot = 474), fills this crucial gap by investigating the lay prototypes of organized crime and advancing scholarly debates surrounding public understanding of criminal phenomena. Studies 1a-b identified the shared features of organized crime. In Studies 2a-b, participants rated the features’ centrality. In Italy, centrality ratings were underlain by an 8-dimensional latent structure and predicted by exposure to news. In the US, centrality ratings were underlain by a 7-dimensional structure and predicted by exposure to TV and movies. Studies 3a-b further underscored the significance of the features by demonstrating their diagnostic value. Participants identified criminal entities described using central (versus peripheral) features as more representative of organized crime, and perceived them as more dangerous. Despite a tendency among US participants towards a more romanticized understanding, the studies revealed complex and multifaceted prototypes of organized crime across contexts. These findings have important implications for policy and communication strategies and serve as a catalyst for further research in this critical area.

KEYWORDS:

• Organized crime
• prototype approach
• meaning
• lay perceptions

Organized crime represents a pervasive global threat, affecting communities and societies worldwide (Travaglino, Burgmer, et al., 2023; von Lampe, 2016). Due to the COVID-19 crisis, its impact is predicted to increase, as criminal organizations will find it easier to infiltrate the legal economy and financial markets (Europol, 2020). However, despite the strong concerns it elicits, the legal and academic communities have not yet reached a consensus on the definition of organized crime (Abadinsky, 2013; Finckenauer, 2005; Levi, 2016; Varese, 2017; von Lampe, 2016; Wheatley, 2011). Even less clear is how the public understands this phenomenon.

Investigating laypeople’s understanding of organized crime is crucial for several reasons. First, grasping what the public perceives as key characteristics of this phenomenon is essential for an accurate assessment of the social harm and impact caused by organized crime. (cf. Kirby et al., 2018). Moreover, given that organized crime is a deeply socially embedded phenomenon (Hobbs & Antonopoulos, 2013; Kleemans & van de Bunt, 1999; Travaglino, Friehs, et al., 2023), criminal groups cannot thrive without ties with local agents and, in some circumstances, a degree of legitimization from the community (Travaglino & Abrams, 2019). Addressing public understanding of organized crime may shed light on the processes leading to communities’ tendencies to legitimize criminal groups. Lastly, effective responses to organized crime require the involvement of the wider civil society. Therefore, it is vital to ascertain the activities and characteristics the general public associate with this phenomenon.

The prevailing wisdom among experts is that public understanding of organized crime is ‘inaccurate’ and heavily shaped by media representations (Finckenauer, 2005, 2007; Hobbs & Antonopoulos, 2013; Rawlinson, 1998). However, these propositions have not been empirically tested. In the present research, we employed the prototype approach (Mervis & Rosch, 1981) to address how the public defines organized crime. We contribute to research on individuals’ understanding of criminal phenomena and the literature on organized crime by investigating the features individuals associate with this concept in two countries – the US and Italy. Both countries are characterized by a similar long-standing presence of criminal organizations on the territory but different media discourses and legislative environments (Lavorgna et al., 2013). Considering the differences and similarities between these two contexts allowed us to gain more complex and nuanced insights into public understanding of organized crime.

Defining and construing organized crime

There is very little consensus among experts about what ‘organized crime’ is (Abadinsky, 2013; Allum et al., 2010; Finckenauer, 2005; Paoli & Vander Beken, 2014; Von Lampe, 2001). von Lampe (2022) compiled a list of more than two-hundred definitions, varying extensively in precision, characteristics, and number of features. To make sense of such heterogeneity, von Lampe (2016) proposed a framework encompassing three basic dimensions of organized crime, namely the way criminal activities are carried out (i.e. market-based, predatory or control-related), the structures linking criminals (i.e. the degree of formalization) and criminal organizations’ governance role (i.e. the way power is used and amassed).

Other authors have attempted to identify a set of sufficient and necessary features of organized crime. Table 1 summarizes some of the most influential frameworks. Hagan (1983, 2006) and Maltz (1985) analyzed the content of prior definitions and proposed frameworks of eleven (subsequently reduced to seven) and nine dimensions, respectively. In his influential analysis, Abadinsky (2013) highlighted eight fundamental dimensions of organized crime. Finckenauer (2005) also proposed an eight-dimension framework aimed at overcoming the US-centered focus of earlier work. More recently, Varese (2017) performed a content analysis of one-hundred-fifteen definitions of organized crime employed between the years 1915 and 2009. The resulting features were summarized in two macro-dimensions, namely structures and activities. Interestingly, Varese’s analysis illustrates how the dimensions’ content has changed over the decades, stressing different facets of organized crime and, by extension, different ways of combatting it. Relatedly, Sergi (2017) proposed the existence of a theoretical continuum ranging from the concept of organized crime (a set of criminal activities) to that of mafia (a specific structure characterized by political influence). According to Sergi (2017), both poles of this continuum are characterized, to a different degree, by four distinct features.

Table 1. Expert Definitions of Organized Crime.

Authors	Features	Dimensions
Maltz (1985)	Corruption; Violence; sophistication; continuity; structure; discipline; multiple enterprises; legitimate businesses; bonding.	_
Hagan (2006)	Primary: Non-ideological; violence/threats of violence; illicit services/criminal enterprises, immunity, corruption, intimidation. Secondary: Structured hierarchy; rules/codes of secrecy; initiation/exclusive membership.	Activities; groups
Finckenauer (2005)	Ideology (or lack thereof); structure/organized hierarchy; continuity; violence/use of force or the threat of force; restricted membership/bonding; illegal enterprises; penetration of legitimate businesses; corruption	_
Varese (2017)	Structures: specialization; hierarchy; illegal enterprise; La Cosa Nostra; network and harm; Activities: monopoly; provision of illegal goods; illegal activities	Structure; activities
Abadinsky (2013)	Non-political goals; hierarchical; limited or exclusive membership; unique subculture; self-perpetuating; illegal violence; monopolistic; governed by explicit rules/regulations	_
von Lampe (2016)	___	Activities; structures; governance
Sergi (2017)	Control of the territory; ties with society; secrecy and trust; mobility	Structure; activities

Conceptual ambiguity notwithstanding, some scholars believe that defining organized crime remains a necessary endeavor (e.g. Finckenauer, 2005). Definitions can guide research by indicating which phenomena are relevant to the investigation. They can also contribute to shaping policy decisions and communicating with the public. Others, however, have questioned the value of adopting a single, all-encompassing definition of organized crime (van Duyne, 2000; von Lampe, 2016). For instance, von Lampe (2016) argues scholars should focus on specific characteristics that both experts and laypeople dynamically associate with the concept across contexts and situations. Notably, both approaches recognize the importance of addressing public understanding of organized crime.

However, despite the prominent role of the public in shaping policy discourses (Allum & Kostakos, 2010), there is little research empirically investigating how people understand organized crime. Research on public perception of organized crime has so far relied on measuring individuals’ attitudes towards specific criminal groups (Travaglino et al., 2016; but see Kirby et al., 2018). Alternatively, it has inferred public understanding from media representations, without addressing how such representations might have been received by people (Di Ronco & Lavorgna, 2018; Levi, 2005). Some authors have speculated that the public has a partial, even naïve view of organized crime. They have proposed that individuals’ understanding may be derived from movies and sensationalistic news reporting involving mafias. For instance, Finckenauer (2007, p. 33) argued that ‘the media have stamped a kind of “mafia myth” on the popular imagination’ (see also Larke, 2003). Similarly, Levi (2016, p. 4) suggested that mafia-type groups provide the image ‘that most graphically captures what many people think of when they use the term’. These propositions have yet to be subjected to empirical scrutiny.

In the present research, we investigated lay prototypes of organized crime. Specifically, we extended prior research assessing institutional perceptions (Caliendo & Scotto Di Carlo, 2019; Sergi, 2017) and media representations (e.g. Di Ronco & Lavorgna, 2018; Young & Allum, 2012) by directly examining individuals’ shared representations of this phenomenon. The public plays an extremely crucial role in the social construction of organized crime because criminal groups’ activities are widely discussed and debated (Hobbs & Antonopoulos, 2013; Massari, 2003; see also Surette, 2011). For instance, by embracing (or rejecting) specific narratives about what constitutes ‘organized crime’, the public may contribute to shaping policy agendas, law enforcement strategies, and political countermeasures. Moreover, the systematic study of the features individuals associate with organized crime allows us to test the assumption, seemingly widespread among scholars, that the public has a naïve view of what organized crime is. Lay and expert definitions may be compared to reveal convergences or divergences between lay and academic discourses, and correct misconceptions in public narratives.

Public understanding of organized crime is likely to vary across countries. Caliendo and Scotto Di Carlo (2019) and Whithorn (2014) have highlighted the existence of country-specific ‘culturemes’ in the institutional and legal discourses about organized crime. Culturemes refer to phenomena and concepts that, although denoted by the same term, acquire specific connotations only in a given cultural framework (Katan, 2009). Culturemes are akin to shared, country-specific prototypes of organized crime. Their existence at the institutional level suggests that individuals from different countries may attribute distinct sets of meanings to the phenomenon of organized crime.

Two especially relevant contexts are Italy and the US (Lavorgna et al., 2013; Sergi, 2015). Italy hosts ‘traditional’ mafia-type groups, which exercise political influence over territories (Travaglino & Abrams, 2019). The presence of mafias could make it harder for the public in Italy to consider other aspects of organized crime that are not directly linked to these groups. The US is characterized by a very influential entertainment industry, which could have shaped the nation’s discourse about organized crime, binding the concept to the Italian American mafia and contributing to generating moral panic (Whithorn, 2014; Woodiwiss & Hobbs, 2009). Thus, by examining shared prototypes of organized crime in these two countries, we can observe which features are more likely to generalize and which, instead, are specific to each context.

The prototype approach

In the present research, we addressed how individuals understand organized crime using the prototype approach (Mervis & Rosch, 1981; Rosch, 1975). This approach is designed to address how individuals use concepts in their everyday life. For instance, prior research has employed the prototype approach to explore concepts that are difficult to define using a single set of necessary and sufficient features, including the emotions of love (Fehr & Russell, 1991) and disillusionment (Maher et al., 2020), cultural constructs such as honor (Cross et al., 2014), figures such as heroes (Kinsella et al., 2015) and crimes such as rape (Krahé, 1991).

The prototype approach is especially suitable to address the question of how the concept of organized crime is understood by the public. A prototype is defined as ‘a collection of the most typical or highly related features associated with a category’ (Cantor & Mischel, 1977). The prototype approach is rooted in the assumption that categories have fuzzy boundaries, which tend to vary among people. Rather than individuating sufficient and necessary features to define a concept, the approach enables researchers to identify what features are more frequently used by individuals to represent a category, and the perceived centrality of each feature to define the concept.

Study 1a and 1b: the features of organised crime

The aim of Study 1 was to identify the common features individuals associate with the concept of organized crime in Italy (1a) and the US (1b). We focussed on the national level of analysis because we were interested in exploring public prototypes among people that share the same language and institutional culturemes. Notably, organized crime may affect the public discourse and people’s attitudes beyond communities directly impacted by the threat (Travaglino, Burgmer, et al., 2023). Moreover, institutional and legal responses and policy strategies are typically set at the national level (Sergi, 2022). It is thus crucial to address how people in specific national contexts understand the organized crime phenomenon.

In the studies, we instructed participants to list words and short sentences that came to their minds when thinking about organized crime (e.g. Kinsella et al., 2015). The characteristics generated by the two samples were subsequently coded in broader features, allowing us to examine what aspects appeared more frequently in the sample.

Methods

Participants

One-hundred twenty-two participants from Italy (59 women, 62 men, 2 unreported; M_age = 30.21 SD_age = 11.51) took part in Study 1a. One-hundred twenty-one participants from the United States (60 women, 61 men; M_age = 33.12 SD_age = 9.78) took part in Study 1b. In both studies, sample size was predetermined at one-hundred and twenty. The additional participants were oversampled by the data collection platform. According to previous research (Cross et al., 2014; Kinsella et al., 2015), this sample size is sufficient to generate a broad but manageable number of features and characteristics. Participants participated in the study via the Qualtrics survey software and the website ‘Prolific’.

Procedure

Participants were invited to participate in a survey ‘about the definition of words and concepts’. Materials were in Italian (Study 1a) or English (Study 1b). After completing a consent form, participants read the following instructions (Gregg et al., 2008): ‘This research is part of a larger study about what people have in mind when they hear and use words. For example, someone may list words such as ‘freedom’ and ‘elections’ as characteristics of ‘democracy’ and phrases such as ‘orders others around’ and ‘takes charge’ as characteristics of a ‘dominant person’’.

Next, participants read: ‘In this study, we are interested in the definition of organized crime. Use the space below to respond to the following questions: what comes to your mind when you think about what organized crime is? What are the characteristics that you associate with organized crime?’. Participants were informed that they could use single words or short sentences and list all words and sentences that came to their mind. Participants were also reminded that there were no right or wrong answers. A text-entry box was employed to record the characteristics generated by participants. Participants were given no minimum or maximum amount of time to complete the task. Participants were subsequently asked to provide demographic information (age, gender). Finally, they were thanked, debriefed, and compensated for their time. The studies were approved by the ethics review board of the University of Kent.

Results and discussion

In both countries, participants generated a similar number of elements. Italian participants generated M = 7.93 (SD = 4.46) features each, whereas American participants generated M = 8.12 (SD = 4) features. The difference between the two samples was not statistically significant, t(239) = −0.34, p = .74, suggesting participants from both countries experience similar levels of ease in retrieving elements describing organized crime.

We proceeded with the analyzes by reducing the number of elements generated by participants in broader features. The coding was conducted separately for each sample and independently by two coders fluent in Italian and English. The coding involved two main steps (e.g. Fehr & Russell, 1991). In the first step, complex linguistic units were divided into simpler elements. Units were broken down if they could reasonably stand alone as separate elements while preserving their meaning. For instance, the unit ‘groups of people who agree to make money and gain power’ was broken down into two simpler elements ‘groups of people who agree to make money’ and ‘groups of people who agree to gain power’. Those units that could stand alone in their original formulation were kept untouched, for instance ‘hard to combat’.

In the second step, we reduced the larger number of elements to a smaller set of features. We merged elements that expressed similar meanings while aiming to preserve small differences in meaning among terms. Two independent coders assigned each element to a feature if i) they were grammatical variations of the same word (e.g. ‘conspired’ and ‘conspiracy’); ii) the elements expressed very similar meaning (‘trafficking of weapons’ and ‘trafficking of arms’); iii) the elements represented different instances of the same phenomenon (e.g. ‘pyramid scheme’ and ‘various frauds’ were both categorized as ‘fraud’). The names of different criminal groups (e.g. Yakuza, Mafia) were kept separated because of the relevance that these groups have in experts’ definitions of organized crime (see Table 1). When participants used different words to express the same general idea (e.g. ‘illegal actions’, ‘actions that are criminal’), the elements were coded in the same category. However, different forms of illegal actions were kept separated when expressed in a more specific form (e.g. corruption, murder) by participants. Inter-coder agreement was 67.32% for the Italian sample and 79.08% for the American sample. A third coder checked the final features, provided suggestions, and helped resolve discrepancies in the categorization of the elements.

At the end of the categorization, we obtained a measure of the frequency with which each feature appeared in the dataset. Following previous research (e.g. Cross et al., 2014), features that appeared only once were deemed idiosyncratic and discarded from the final list. Of the total number of features, in the Italian sample, 105 (54.68%) were idiosyncratic; in the American sample, 122 (49.10%) were idiosyncratic. There was a total of 87 remaining features in the Italian sample and 91 in the American sample.

Table 2a and b summarize the non-idiosyncratic feature in the two samples. There were similarities across the two countries in the most frequent features generated by participants. The most frequent features included specific crime groups (the Mob or the Mafia), activities (drug trafficking), characteristics (hierarchical structures, violence) and objectives (financial gain). Interestingly, whilst in the US sample, 21.5% of the participants mentioned ‘Movies and TV Shows’, in the Italian sample, only 4.13% did. A two-proportion z-test indicated that the two proportions were statistically different, z = −3.471, p < .001, suggesting that features related to entertainment media were more prevalent in the US sample.

Table 2. a. Features Obtained in Study 1a (Italy). b. Features Obtained in Study 1b (US).

Category (a: Italy)	Frequency	Centrality	Category (a: Italy)	Frequency	Centrality
Mafia	78	6.88	Damage to society	9	6.31
Collective undertaking	60	4.96	Planned and premeditated activities	9	5.1
Drug trafficking	39	6.46	Secrecy	9	5.49
Illegal activities	38	6.54	Southern Italy	9	5.36
Violence	33	6.26	Gangs	8	5.64
Corruption of police, politicians and/or other institutions	32	6.33	Boss (e.g. Riina, Provenzano etc)	7	6.77
Camorra	29	6.83	Dangerous	7	5.82
Hierarchical structure	27	5.95	Power	7	6.02
Financial gain	26	5.94	Money laundering	7	6.55
Extorsion	24	6.32	Terrorism	7	5.27
Murders	24	6.14	Prison	6	5.61
Ndrangheta	24	6.8	Antimafia heroes	6	6.08
Guns	22	6.11	Ignorance	6	5.03
Collusion with police, politicians and/or other institutions	20	6.12	Robberies	6	5.03
Evil	17	5.68	Merciless	5	6
Omertà	16	6.25	Control	5	5.59
Business	14	6.28	Blackmail	5	6.13
Control of the territory	14	6.28	Movies and TV shows	5	5.2
Fear	14	5.81	To harm others	5	6.01
Criminals	13	6.28	Politics	5	5.38
Prevarication	13	5.54	Exploitation	5	5.8
Family	11	4.9	Use of intimidation	5	6.49
Prostitution	11	5.41	Police	4	4.64
Arms trafficking	4	6.22	Ability to achieve objectives	2	4
Fraud	4	5.27	Casalesi	2	6.34
Pain	4	5.1	Connivance	2	5.14
Judicial bodies	4	4.92	Ethnic connotation	2	3.21
Death	4	5.75	Crime control	2	5.62
Enduring structures	4	4.32	Disgust	2	5.4
Human trafficking	3	5.67	Ingrained phenomenon	2	5.98
Revenge	3	5.68	Thief	2	4.47
Decay	3	5.53	Mafia Capitale	2	6.53
Honor	3	4.75	Pride	2	4.41
Presence in the world	3	5.2	Poverty	2	4.42
Sacra Corona Unita	3	6.02	Social problem	2	6.27
Bribes	3	6.17	Environmental crimes	2	5.58
Smuggling	2	6.17	Recruitment	2	5.1
Cosa Nostra	2	6.75	Blood	2	5.18
Theft	2	4.92	Submission of the weak	2	5.87
Italians	2	5.27	Massacres	2	5.84
Loyalty	2	4.17	Victim	2	5.17
Widespread	2	5.21
Kidnapping	2	5.45
Network	2	5.14
Yakuza	2	5.91
41bis	2	5.75
Category (b: US)	Frequency	Centrality	Category (b: US)	Frequency	Centrality
The mob	64	6.15	Evil	9	5.31
Planned and premeditated activities	58	5.94	Cartels	9	6.03
Collective undertaking	48	5.18	Smuggling	9	5.85
Hierarchical structures	46	5.63	Extortion	9	5.86
Drug trafficking	36	6.02	Fraud	8	5.8
Financial gain	36	5.73	Old school mafia	8	5.89
Gangs	30	5.88	Hard to defeat	7	5.06
Violence	29	5.99	Boss (e.g. Al Capone, Baby Face, Dillinger)	7	6.15
Movies and TV shows	26	4.93	Terrorism	7	4.82
Corruption of police, politicians and/or other institutions	22	5.73	Power	7	5.73
Large scale acts	21	4.96	Prostitution	7	4.64
Illegal activities	19	6.32	Intentional harm	6	5.59
Guns	16	5.76	Merciless	6	5.26
Murders	15	5.77	Russians	6	3.88
Theft	15	5.52	Human trafficking	6	5.6
Money laundering	15	6.09	Suits	6	3.78
Italians	14	4	Loyalty	6	4.95
Secrecy	13	5.62	Yakuza	6	4.96
Criminals	13	6.23	Blackmail	5	5.73
Protection racket	11	5.26	Clever	5	4.63
Dangerous	11	6.03	Gambling	5	5
Mafia	10	6.33	Collusion with police, politicians, and/or other institutions	5	5.57
Family	9	4.09	Tax evasion	4	5.6
Kidnapping	4	5.16	Wealth	3	5.18
Police	4	4.64	Network	3	5.34
Threat	4	5.77	Revenge	3	5.2
Underground	4	5.4	Habitual action	2	4.45
Manipulation	4	5.6	Professionals	2	4.66
Prison	4	5.27	Taking advantage of others	2	5.77
Black market	4	5.78	Greed	2	5.78
Sicilian	4	4.13	Execution	2	4.99
Hitmen	4	5.7	Loan sharks	2	5.24
Cybercrime	4	4.89	Systemic crime	2	5.77
Damage to society	4	5.67	Coordination	2	5.36
Brotherhood	3	4.65	Alcohol	2	4.39
Control	3	5.41	White people	2	3.96
Enforcers	3	5.38	Sunglasses	2	2.58
Betrayal	3	4.84	Lies	2	5.51
White collar	3	4.1	Respect	2	3.89
Men	3	4.82	New York	2	4.35
Front business	3	5.43	Cosa nostra	2	4.62
RICO	3	4.42	Opposing authority	2	5.24
Cohesion	3	4.74	Undercover	2	4.91
Arms trafficking	3	5.83
Business	3	4.82
FBI	3	4.57
Escaping authorities	3	5.52
Widespread	3	4.36

Note. The features are listed in the order of frequency. Features shared across countries are bolded.

Finally, we examined the number of shared features across samples. Approximately half the features (N = 45) were shared (see bolded features in Table 2a-b). Thus, a substantial number of features and characteristics associated with organized crime were idiosyncratic to each context. This result is consistent with the existence of different institutional organized crime culturemes in the two countries (Whithorn, 2014). Although participants from Italy and the US may use a similar label to indicate organized criminal activities, our data suggest the existence of substantial variation in the meaning and core features linked to the phenomenon. The features generated exclusively in the American sample included ‘suits’, ‘white men’, ‘Russians’. Whereas ‘environmental crimes’ emerged only in the Italian sample, ‘cybercrime’ emerged only in the American one. Although both samples generated a ‘control’ feature, ‘control of territories’ and ‘control of other criminals’ were generated only in the Italian sample, suggesting greater awareness of the political role of organized crime.

Studies 2a and 2b: investigating the centrality of the features

The objectives of Studies 2a-b were threefold. First, we examined the perceived centrality of the features generated in Studies 1a-b. Whereas the frequency index obtained in Studies 1a-b indicates how easy it is to retrieve a feature from memory, centrality ratings depend on a more systematic process of appraisal of the information (Hassebrauck, 1997). Thus, centrality ratings inform us about the relevance of the features to the definition of organized crime.

The second aim of Studies 2a-b was to empirically identify latent structures underlying the features associated with organized crime in the two countries (cf. Cross et al., 2014). Such latent structures indicate the existence of macro-categorises summarizing what features co-vary together in terms of centrality. We investigated the existence of latent structures by employing Exploratory Graph Analysis (EGA; Golino et al., 2020; Golino & Epskamp, 2017). EGA is a relatively new approach derived from network psychometrics. Recent studies have shown that EGA performs equally as well as more traditional methods (e.g. parallel analysis) in identifying latent structures (Golino et al., 2020). Moreover, research indicates that EGA is especially suited for data with a larger factor structure and high correlations among dimensions (Golino et al., 2020), as was expected in the current study. Importantly, EGA also yields a graphic representation of the data structure that allows researchers to examine the relationships among features which do not belong to the same network.

The third aim of the study was to explore the associations between exposure to organized crime-related media content and the centrality ratings of the features. Specifically, we investigated how individuals’ attention to two types of entertainment media (video games and TV shows/feature films) and news content predict the extent to which features are perceived as central. Media are especially important in distributing knowledge about crime-related phenomena (e.g. Intravia et al., 2017; Mastrorocco & Minale, 2018; Pfeiffer et al., 2005; Surette, 2011). Little research has so far addressed the role of media in shaping individuals’ understanding of organized crime (e.g. Di Ronco & Lavorgna, 2018; Larke, 2003; Leiva & Bright, 2015), and none has directly examined the association between individuals’ attention to specific media content and their perceptions of features of organized crime. As discussed earlier, several authors have pointed out the emergence among the public of a ‘mafia myth’ created by the entertainment industry (e.g. Finckenauer, 2005; Larke, 2003). Therefore, we hypothesized that individuals’ centrality ratings should be predicted by their interest in entertainment media, and to a lesser extent, in the news. We tested our hypotheses using a measure of attention in the media because previous research suggests that attention is a more precise indicator of individuals’ tendency to consume specific content compared to the frequency of use (O’Keefe & Reid-Nash, 1987). In the analyses, we controlled for the features’ frequency index (derived from Studies 1a-b), and individuals’ socio-demographic characteristics of age, gender and subjective socio-economic status.

This study’s data structure involved a set of features separately nested within participants and type of feature. Such hierarchical data structure requires the use of multilevel models because the ratings of the features are not independent (e.g. the same individual rates multiple features). Non-independence violates a crucial assumption of regression, which may result in biased standard errors and an increased likelihood of false-positive or negative results (Brauer & Curtin, 2018; Hox et al., 2017). By considering the existence of hierarchical data structures, multilevel models are instead equipped to deal with the issue of non-independence. We, therefore, estimated a multilevel model where the fixed effects of media attention, gender, age, subjective social status, and frequency of the features (level 2) predicted the centrality ratings of the features (level 1).

Methods

Participants

Two-hundred fifty-five participants from Italy (128 women, 126 men, 1 unreported; M_age = 26.35 SD_age = 7.27) took part in Study 2a. Two-hundred fifty-two participants from the United States took part in Study 2b (125 women, 124 men, 3 other; M_age = 33.87 SD_age = 12.59). Sample size was predetermined at two-hundred fifty in each country. The remaining participants were oversampled by the platform. Participants took part in the study via the Qualtrics survey software and the website ‘Prolific’.

Procedure

Participants were invited to take part in a study about ‘How to define concepts’. They first read a brief introduction: ‘This research examines how people define concepts. Specifically, we are interested in the definition of Organized Crime. On the next page, there is a list of attributes about organized crime. These attributes can be single nouns, small sentences, adjectives or adverbs. For each attribute, please rate how closely related the attribute is to the concept of organized crime.’

Next, participants were presented with the list of features generated by a sample of the same nationality in Studies 1a-b. Participants rated each category on a scale from 1 ( =  not at all related) to 7 ( =  extremely related; the middle point of the scale was labeled as 4 = moderately related). The presentation of features was randomized. Embedded within the feature list, we included two attention checks (i.e. items asking participants to respond using a specific number on the scale).

After completing the centrality ratings, participants completed measures of attention towards organized-crime-related content in three media. The items included news (‘How much attention do you pay to news involving organized crime on TV, radio, printed newspaper or the internet?’; 1 = not at all, I pay no attention, 7 = Extremely high, I pay huge attention), TV shows and films (‘How interested are you in television shows or feature films involving organized crime?’; 1 = not at all, I have no interest, 7 = Extremely high, I have a huge interest), and videogames (‘How interested are you in videogames involving organized crime?’; 1 = not at all, I have no interest, 7 = Extremely high, I have a huge interest).

Finally, participants were asked to provide demographic information (gender, age). We measured participants’ subjective social status (SSS) to control for its effects in the multilevel model. The measure of SSS presented participants with a brief set of instructions, ‘Think of the ladder above as representing where people stand in society. At the top of the ladder are the people who are best off – those who have the most money, most education and the best jobs. At the bottom are the people who are worst off – who have the least money, least education and the worst jobs or no job. The higher up you are on this ladder, the closer you are to people at the very top and the lower you are, the closer you are to the bottom. Where would you put yourself on the ladder?’ (Adler, Epel, Castellazzo, & Ickovics, 2000). Participants were then asked to select a number between 1 (the top of the ladder) and 8 (the bottom). Finally, participants were thanked, compensated for their time and debriefed. The studies were approved by the ethics review board of the University of Kent.

Results

Centrality Ratings and Exploratory Graph Analysis. Six participants in the Italian sample and five participants in the American sample did not complete the attention checks correctly and were excluded from further analyses. Mean centrality ratings for both samples are summarized in Table 2a and b. The centrality ratings were only moderately correlated with the features’ frequency in the Italian sample, r = .38, p = .001, and the American sample, r = .35, p = .001. The size of these correlations is consistent with previous research employing the prototype approach in other contexts (e.g. Cross et al., 2014). The size of the correlations is consistent with the idea that frequency and centrality provide distinct types of information about the features. Whereas frequencies are an index of ease of retrieval, centrality ratings are underpinned by a more systematic processing of the information, revealing the relevance of each feature to the lay concept of organized crime.

Following previous research (e.g. Gregg et al., 2008; Kinsella et al., 2015), we used a median split across samples to identify features considered as either central or peripheral. Central features in the Italian sample included the mafia groups (e.g. Mafia, Camorra, ‘Ndrangheta), activities such as ‘drug trafficking’, ‘murder’, and ‘control of the territory’, and characteristics such as ‘fear’, ‘violence’ and ‘power’. Peripheral categories were characteristics such as the ‘ethnic connotation’ of the groups, ‘pride’, ‘loyalty’ and ‘honor’, and activities such as ‘robbery’ and ‘kidnapping’. In the American sample, features rated as central were specific groups (e.g. The Mob, Gangs, Cartels), activities such as ‘drug trafficking’ and ‘fraud’, and characteristics such as ‘secrecy’ and ‘hierarchical structures’. More peripheral features included ‘protection racket’, ‘respect’ and ‘prostitution’.

To investigate the latent structure of individuals’ lay definitions of organized crime, we carried out a separate EGA in each country. Results are displayed in Figures 1a and 1b. In the figures, each node (the circles) represents a feature. The edges (the links among nodes) represent the strength of the partial correlations among features, conditioned on all other features in the dataset. Features belonging to the same latent category share the same color. The existence of a latent category indicates that participants’ centrality ratings of the features included in that category tend to covary. We inferred a name for each latent category by considering the features included.

Figure 1. a. Exploratory Graph Analysis of the features of organized crime in Study 2a (Italy). b. Exploratory Graph Analysis of the features of organized crime in Study 2b (US). Note. Thicker lines indicate larger partial correlations between variables conditioned on all the other variables in the dataset. See the online article for the color version of these figures.

In the Italian sample, we obtained an eight-category structure. The first category captured the Characteristics of Criminal Groups, including ‘power’, ‘hierarchical structures’, ‘pervasive’, ‘secrecy’, and ‘revenge’. The second category included Contextual Features such as ‘poverty’, ‘ignorance’ and ‘environmental crimes’, reflecting both organized crime’s damage to society and the social environment where these groups operate. The third category encompassed characteristics linked to a perception of Danger, such as ‘death’, ‘blood’, but also ‘collective undertaking’ and ‘ethnic connotation’. The fourth category included Dishonestly Taking Crimes, including ‘theft’ and ‘robbery’. The fifth category included Illegal Financial Crimes related to ‘corruption’, ‘bribes’, ‘money laundering’, and a reference to the investigation ‘mafia capitale’ which involved the local government of the city of Rome. The sixth category referred to Organized Crime Groups (e.g. the Mafia, the ‘Ndrangheta, the Yakuza) together with some key features strongly associated with them (e.g. ‘omertà’ [Law of Silence], ‘connivance’). The seventh dimension described Relationships among Affiliates, i.e. ‘pride’, ‘loyalty’, and ‘honor’ (the latter two constructs were also rather strongly linked to the feature ‘family’). Finally, the eighth dimension referred to Violent Criminal Activities such as ‘intimidation’, ‘drug trafficking’, ‘arms trafficking’, ‘murder’, and ‘exploitation’.

In the American sample, we obtained a 7-category structure. The first category referred to the Aesthetic of Organized Crime, with terms such as ‘New York’, ‘white people’, ‘sunglasses’ and ‘suits’. These terms seem to capture features related to the visual representations of organized crime in the media, involving both people and places (Larke, 2003). The second category included Characteristics of Criminal Groups such as ‘hierarchical structures’, ‘planned and premeditated activities’, and ‘brotherhood’. The third category captured Criminal Activities such as ‘money laundering’, and ‘extortion’, describing also organized crime’s interactions with authorities (‘collusion’ and ‘corruption’, ‘opposing authority’, ‘escaping authorities’, ‘prison’, and ‘tax evasion’). The fourth dimension referred to La Cosa Nostra and included ‘Cosa nostra’ together with the Yakuza and RICO. The fifth category included Large Scale Social Harm such as ‘arms trafficking’, ‘drug trafficking’, ‘human trafficking’, ‘cybercrime’ and ‘prostitution’. The sixth category referred to various concepts that could be linked to the Romanticization of Organized Crime, involving terms such as ‘movies and TV shows’ and ‘old school mafia’ covarying with terms such as ‘loan sharks’, ‘hitmen’, and ‘cartels’. The seventh category could be named Violent Activities and included features such as ‘guns’, ‘evil’, ‘murders’, ‘violence’, and ‘threat’.

Media and Centrality Ratings. To examine whether attention towards organized-crime media content was associated with the features’ centrality ratings, we estimated a multilevel model in each country. Predictors in the model were centered. Results from the two models are summarized in Table 3a and b. Controlling for gender, age, SSS and frequency, in the Italian sample, only attention towards news was associated with centrality ratings, b = .10, SE = .037, p = .008. In the American sample, only interest in TV-shows and feature movies was associated with centrality ratings, b = .12, SE = .041, p = .006.

Table 3. a. Summary of the Results of the Multilevel model in Study 2a (Italy). b. Summary of the Results of the Multilevel model in Study 2a (US).

Effect	Parameter	b	95 CI	SE	t ratio	p
	(Intercept)	5.65	5.49–5.80	.080	70.113	< .001*
Fixed	Frequency	0.02	0.01–0.03	.016	3.27	= .002*
	TV and Feature Films	0.04	−0.01–0.09	.030	1.41	=.160
	Video Games	0.01	−0.05–0.06	.030	0.11	=.920
	News	0.10	0.03–0.17	.040	2.67	=.008*
	SSS	0.01	−0.05–0.05	.026	0.08	=.933
	Age	−0.01	−0.01–0.01	.001	−0.25	=.807
	Gender	−0.03	−0.20–0.15	.090	−0.33	=.740
Random	Var(participants)	0.34
	Var(feature type)	0.44
	Nakagawa's R^2marginal / R^2conditional	0.03 / 0.36
Note. SSS = Subjective Social Status. Gender: 1 = man, 2 = woman. * indicates a significant association at the specified p-value.
Effect	Parameter	b	95 CI	SE	t ratio	p
	(Intercept)	5.22	5.05 - 5.38	.085	61.32	<.001*
Fixed	Frequency	0.02	0.01 - 0.03	.005	3.56	< .001*
	TV and Feature Films	0.12	0.03 - 0.20	.041	2.79	=.006*
	Video Games	−0.01	−0.07 - 0.07	.035	−0.57	=.954
	News	0.01	−0.08 - 0.10	.047	0.19	=.854
	SSS	−0.04	−0.11 - 0.01	0.03	−1.61	=.109
	Age	0.02	0.01 - 0.03	.001	3.68	<.001*
	Gender	0.24	0.02 - 0.46	0.11	2.110	= 0.04*
Random	Var(participants)	0.64
	Var(feature type)	0.41
	Nakagawa's R^2marginal / R^2conditional	0.05 / 0.42

Note. SSS = Subjective Social Status. Gender: 1 = men, 2 = woman. * indicates a significant association at the specified p-value.

Discussion

Studies 2a-b provide valuable insights into how organized crime is viewed in the public sphere in both Italy and the US. There was some overlap between the features considered central in both countries (see Table 2a-b) but also some notable differences. For instance, ‘terrorism’ was categorized as peripheral (below the median of centrality ratings) in both samples. However, ‘gangs’ were categorized as peripheral in the Italian sample but as central in the American sample. Moreover, ‘secrecy’ was categorized as peripheral in the Italian sample, but as central in the American sample.

EGA revealed the existence of eight latent dimensions in Italy and seven in the US. The dimensions highlighted clusters of features rather similar to those identified by academics (see Table 1), for instance pertaining to organized crime’s activities and characteristics. However, in the American sample, there were two clusters of features consistent with the existence of a ‘mafia-myth’ in the population. Specifically, the two dimensions of Romanticization and Aesthetic of Organized Crime (which were absent in the Italian sample) encompassed a set of characteristics and activities potentially derived from feature films and representations of organized crime in Hollywoodian movies (Larke, 2003).

Findings from the multilevel models only partially supported our hypotheses. Centrality ratings in the American sample were predicted by individuals’ attention to movies. Individuals’ interest in news content predicted the ratings in the Italian sample. Although news content may be sensationalistic or partial (cf. Di Ronco & Lavorgna, 2018), we assume it represents organized crime less romantically than TV Shows and Films. Overall, this evidence is also consistent with the presence of a stronger ‘mafia myth’ in the US.

Studies 3a-b: to what extent are central features diagnostic

In the previous studies, we identified the features that individuals associate with organized crime in Italy and the US and investigated the features’ relative centrality. The objective of Studies 3a-b was to examine whether central features are more diagnostic (namely, descriptive) of the ‘organized crime’ concept. Compared to peripheral features, central features should carry more information about a category. Therefore, an entity described using central features should be perceived as more representative of that category. To test this idea in the context of organized crime, we presented participants with two descriptions of criminal entities constructed using either central or peripheral features. We hypothesized that individuals should perceive the criminal entity described using central (vs peripheral) features as more representative of the category of organized crime.

Moreover, we also explored some potential implications of perceiving a criminal entity as more representative of organized crime. Specifically, we examined whether entities more representative of organized crime were also seen as more dangerous and whether individuals were less likely to express the intention to report them to the police. Organized crime is notoriously violent. Criminal groups may elicit stronger social concern than other forms of crime (Woodiwiss & Hobbs, 2009). Thus, we expected a criminal entity described using more central features to be perceived as more dangerous and less likely to be reported to the police.

Methods

Participants

One-hundred participants from Italy (49 women, 51 men, 1 unreported; M_age = 25.50 SD_age = 6.79) took part in Study 3a. One-hundred-one participants from the United States took part in Study 3b (49 women, 49 men, 3 other; M_age= 30.02 SD_age = 10.33). Sample size was predetermined at one-hundreds in each Country. The remaining participants were oversampled by the platform. A power analysis using the software G*Power showed that this sample size enabled us to detect a small-to-medium effect (d_z= .28) 80% of the time, assuming α = .05 and a two-tailed test. Participants took part in the study via the Qualtrics survey software and the website ‘Prolific’.

Procedure

Participants in both countries were invited to take part in a study examining ‘how people perceive criminal entities operating in society’. After reading the consent form, participants were provided with brief descriptions of two criminal entities (named X and Y). Each entity was described using eight features derived from Studies 1a-b. ‘Entity X’ was described using eight of the most central features as rated in Studies 2a-b. ‘Entity Y’ was described using eight of the most peripheral features. In this study, we did not use names of specific mafia groups (e.g. La Cosa Nostra or the Mafia) because they are immediately recognizable as organized crime. The presentation order of the two entities was randomized for each participant. Participants only read descriptions containing features generated from a sample of the same nationality.

The features and descriptions employed in the studies are summarized in Table 4. After reading each of the descriptions, participants were asked: ‘Please rate how representative Entity [X/Y] is of organized crime’ (1 = Not at all representative, 7 = Extremely representative), ‘To what extent do you think it is dangerous to oppose Entity [X/Y]’ (1 = Not at all dangerous, 7 = Extremely dangerous), and ‘If you became aware that Entity [X/Y] committed a crime, how likely is that you would report Entity [X/Y] to the police’ (1 = Not at all likely, 7 = Extremely likely). Questions were generated in English, translated to Italian and back-translated to ensure consistency across studies. Finally, participants were asked to provide demographic information (gender and age), debriefed, thanked, and compensated for their time. The studies were approved by the ethics review board of the University of Kent.

Table 4. Features and descriptions used in Studies 3a and 3b.

Study	Criminal Entity	Description
Study 3a (Italy)	X (Central)	Entity X is headed by a Boss and is engaged in illegal actions and activities that include money laundering, drug trafficking and extortion. Entity X is known for episodes of corruption of politicians, magistrates and law enforcement agencies. Entity X uses intimidation to achieve its goals and represents a damage to society.
	Features	Boss; Money Laundering; Illegal Activities; Use di intimidation; Drug Trafficking; Corruption of police, politicians and/or other institutions; Extorsion; Damage to Society
	Y (Peripheral)	Entity Y has enduring structures, is made up of thieves, and is engaged in activities that generate poverty. Entity Y is known for its ability to achieve its objectives and, for this reason, is in the sights of the police. Entity Y is characterized by a strong ethnic connotation and its members consider loyalty and pride important.
	Features	Police; Thieve; Poverty; Pride; Durable Structures; Loyalty; Ability to achieve objectives; Ethnic connotation
Study 3b (US)	X (Central)	Entity X includes cartels and many famous gangsters. Entity X is engaged in illegal activities, including money laundering and drug trafficking. Entity X is known for its violence and is considered very dangerous. Entity X plans and premeditates activities carefully.
	Features	Illegal activities; Famous gangsters (e.g. Al Capone, Baby Face, Dilinger); Money laundering; Dangerous Cartels; Drug trafficking; Violence; Planned and premeditated activities
	Y (Peripheral)	Entity Y includes families of white people, and many Italians and/or Russians. Entity Y’s members often wear suits and sunglasses. Entity Y is known for the presence of white collars. Entity Y’s members consider respect important.
	Features	White collar; Family; Italians; White people; Respect; Russians; Suits; Sunglasses

Note. Descriptions in Italian were translated to English (descriptions in their original language are available at https://osf.io/sxt7u/).

Results

Across studies, data were analyzed using paired sample t-tests. We adjusted for multiple comparisons and divided the p-values obtained in the analyses by the number of comparisons conducted in each sample (p/3). The results below report the adjusted p-value.

Study 3a: Italy. Results indicated that individuals rated the criminal entity as more representative of the concept of organized crime when it was described using central (M = 6.47, SD = .90) rather than peripheral (M = 5.36, SD = 1.37) features, t(99) = 8.38, 95% CI = [1.37–0.85], p < .001, d = 0.84. In addition, the criminal entity was perceived as more dangerous when it was described using central features (M = 6.24, SD = .88) compared to the peripheral features (M = 5.24, SD = 1.16), t(99) = 8.18, 95% CI = [1.24–0.76], p < .001, d = 0.82. Finally, individuals showed lower intentions to report to the authorities a crime committed by the criminal entity described with central features (M = 4.42, SD = 1.52) compared to peripheral ones (M = 5.18, SD = 1.45), t(99) = −6.63, 95% CI = [−0.53 – −0.99], p < .001, d = 0.67.

Study 3b: The US. In this study, results also indicated that individuals rated the criminal entity as more representative of the concept of organized crime when it was described using central (M = 6.15, SD = 1.29) rather than peripheral (M = 3.93, SD = 1.79) features, t(100) = 9.53, 95% CI = [1.76–2.68], p < .001, d = 0.95. Moreover, individuals perceived as more dangerous the criminal entity when it was described using central features (M = 6.32, SD = 1.03) compared to peripheral features (M = 3.86, SD = 1.75), t(100) = 14.32, 95% CI = [2.12–2.80], p < .001, d = 1.42. However, individuals were not less likely to report the criminal entity to authorities when it was described using central features (M = 3.90, SD = 2.03) compared to peripheral features (M = 4.29, SD = 1.92), t(100) = −1.93, 95% CI = [−0.78–0.01], p = .169, d = 0.19.

Discussion

The findings of Studies 3a-b supported the idea that, compared to peripheral features, central features are more diagnostic of the concept of organized crime. Specifically, criminal entities described using central features were rated as more representative of the ‘organized crime’ category in both the Italian and American samples. Interestingly, entities described using more central features were also rated as more dangerous, consistent with the idea that organized crime elicits stronger public concern than other forms of criminality. Only in the Italian sample the likelihood of reporting entities more representative of organized crime to the police was lower than less representative entities. In the US sample, the difference was not statistically significant. A plausible explanation for these results is that, in the US, individuals’ intentions to report a crime to the legal authority may be driven by different considerations, for instance, a higher level of trust in the institutions or a perceived greater competency of the authorities (Catterberg, 2006).

General discussion

Organized crime is an extremely impactful phenomenon affecting societies worldwide (Travaglino, Burgmer, et al., 2023). Public discourses around organized crime have coalesced in culturemes, namely country-specific and pervasive representations of organized crime (Caliendo & Scotto Di Carlo, 2019; Sergi, 2022; Whithorn, 2014). While previous research has addressed institutional and media representations (e.g. Di Ronco & Lavorgna, 2018; Sergi, 2017), the nature of such culturemes among lay people has remained empirically unexplored and an object of speculation among social scientists. Culturemes can be likened to shared prototypes, fuzzy categories defined by features varying in typicality and frequency among people (Mervis & Rosch, 1981). In the present research, we employed the prototype approach to investigate the ‘black box’ of laypeople's representations of organized crime. We examined lay prototypes of organized crime in Italy and the US, two countries characterized by the presence of notorious criminal organizations and an influential media industry revolving around the phenomenon.

In Studies 1a-b, we asked participants to indicate features they associate with organized crime. In both countries, participants generated a substantial number of features, of which only approximately half overlapped. Several of the unique features generated by the US sample could be traced back to the portrayal of organized crime groups in movies and TV shows.

In studies 2a-b, we investigated what features were perceived as more central to the definition of organized crime. Centrality ratings uncover the perceived relevance of the features to the concept of organized crime. We used centrality ratings to explore the latent categories underlying the features and to summarize the complex network of concepts associated with organized crime.

Exploratory graph analyses yielded eight categories in Italy and seven in the US (see Figures 1a and 1b). Generally, the latent categories that emerged across the two samples referred to organized crime’s activities, characteristics, and different forms of harm. However, in the US also emerged two categories related to the romanticization and aesthetic of organized crime. Interestingly, whereas in the Italian sample, the features’ perceived centrality was predicted by participants’ interest in organized-crime-related news, in the US the centrality ratings were predicted by participants’ interest in TV shows and feature films. It should be highlighted that the labeling of latent macro-categories involves a high degree of subjectivity (e.g. Reio & Shuck, 2015). However, EGA provides a transparent visual representation of how features are interconnected. This representation might be employed by other researchers to address more specific questions about how features covary and suggest alternative ways of labeling the macro-categories.

Finally, Studies 3a-b demonstrated that central features carry more information about the concept of organized crime than peripheral features. Individuals were more likely to perceive a criminal entity as belonging to the ‘organized crime’ category when the entity was described using more central features. Studies 3a-b also addressed some of the implications of perceiving an entity as more closely related to the concept of organized crime. Specifically, participants perceived it as more dangerous and – in Italy – expressed significantly lower intentions to report a criminal entity to the police when the entity was described using central features.

Taken as a whole, this set of studies showed the existence of sophisticated and – to a certain extent – country-specific prototypes of organized crime. The dimensions emerging in Studies 2a-b parallel many of the dimensions identified by experts (cf. Table 1). Participants also listed a broad and heterogeneous number of features, involving different activities, consequences and actors. Notably, in line with the idea that prototypes have fuzzy boundaries, features in both countries included ‘gangs’ and ‘terrorism’, reflecting the links, also acknowledged by experts, between organized crime and these distinct actors (Decker & Pyrooz, 2014).

We also found evidence aligning with a ‘mafia myth’ in the lay understanding of organized crime. Participants from both Italy and the US identified some features potentially influenced by media portrayals of organized crime. Evidence for a ‘mafia myth’ was more pronounced in the US, where participants’ centrality ratings were significantly predicted by their interest in TV shows and feature films. These findings echo previous studies, highlighting the influential role of the US entertainment industry in shaping public perceptions of organized crime (Larke, 2003).

Limitations and future research directions

The present study empirically explored the features constituting the prototype of ‘organized crime’ in Italy and the US. The studies yielded crucial insights into the culturemes (or prototypes) of organized crime, shedding light, for the first time, on how the public understands the phenomenon across the two national contexts. Nevertheless, this research was affected by some limitations that should be addressed in future work.

A limitation is that we did not explicitly include an analysis of institutional (legal and political) perceptions of organized crime. Institutional perceptions directly inform law enforcement strategies and public policies (Caliendo & Scotto Di Carlo, 2019; Sergi, 2017, 2022). Institutional perceptions could be analyzed using a similar approach to the one utilized here. For instance, future studies could systematically address what features of organized crime appear most frequently in official discourses across countries or are considered more central by practitioners and legal experts.

Consideration should also be given to how institutional perceptions may shape the public's prototypes of organized crime. These perceptions and discourses are likely to reach a large portion of the general public via the media (Di Ronco & Lavorgna, 2018; Surette, 2011; Thompson, 1988, 2011; Young & Allum, 2012). In fact, the media serve as one of the primary sources of crime-related information for individuals (Surette, 2011). Media depictions can also influence the understanding of policymakers and practitioners regarding the issue (Di Ronco & Lavorgna, 2018). Consequently, further research is needed to delve into the intricate, dynamic relationship between public perceptions, institutional viewpoints, and media portrayals. Investigating this relationship could lead to a more comprehensive understanding of the prototypes of organized crime across various societies, the factors shaping such prototypes, and their implications for society's approach to criminal groups.

Another significant area for future research is to investigate the prototype of organized crime within different social groups (e.g. different social classes), geographical locations (e.g. a broader range of nations and regions), and contexts (e.g. communities directly impacted by criminal groups; Poppi & Ardila, 2023; Poppi & Sandberg, 2023). As noted earlier, our findings revealed both shared and country-specific features of organized crime. Similarly, examining lay prototypes across diverse contexts could reveal a mixture of shared core attributes and unique characteristics. For instance, prototypes among populations residing in regions with a prolonged history of interactions with organized crime might develop distinct connotations, influenced by the behaviors of mafia groups in those areas (Sergi, 2022). Thus, leveraging such a comparative approach could offer further insights into how various legal, social, and cultural frameworks affect the formation of lay prototypes. Especially crucial is the investigation of public understanding of organized crime in the Global South, where some of the most impactful manifestations of this phenomenon occurr (Pereda, 2022).

Finally, future research might also explore how various descriptions of organized crime correlate with individuals’ acceptance of different policy responses. The public is an important stakeholder in the social construction of crime-related phenomena (Surette, 2011). However, the impact of people's expectations about what constitutes organized crime on political agendas or media decisions to disseminate specific materials remains ambiguous. Future research could utilize the lists of features generated in our research to explore the connection between centrality ratings and the acceptance of policies with varying levels of restrictiveness and impact on personal freedom.

Conclusions

This research investigated the key features that individuals associate with organized crime, their relative centrality and the existing links among features. By unpacking the ‘black box’ of shared lay prototypes of organized crime, we revealed evidence consistent with a ‘mafia myth’ within the public, especially in the US. However, our findings also uncovered complex lay prototypes of organized crime, suggesting a level of sophistication that might exceed what experts typically credit to the public.

The features identified in this research could serve as a tool for improving communication between policymakers and civil society. For example, these features could help reveal activities and actors that the public tends to overlook, thereby rectifying such misconceptions. The lists of features generated by participants, along with the associated frequency and centrality indices, are made available online (https://osf.io/sxt7u/) in the hope of stimulating further research into civil society’s understanding of organized crime.

Data

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.

Additional information

Funding

This work was supported by the UK Research and Innovation under Grant ‘Secret Power’ No. EP/X02170X/1 awarded to G. A Travaglino under the European Commission’s ‘European Research Council - STG’ Scheme.