"Grooming Gangs" Are Not One Criminological Phenomenon
Updated: Sep 1
A causal typology of group-based child sexual exploitation
Abstract
The expression "grooming gang" compresses several distinct empirical questions into a single politically charged label. It can refer to a technique used to gain access to and control a child, an incident involving several perpetrators, a friendship group, a status economy in which access to victims is circulated, a locally embedded exploitation network, an organised crime group (OCG), or a national or transnational trafficking system. These formations may overlap, but they are not criminologically interchangeable. This article develops six ideal types: (A) the lone sexual exploiter or trafficker; (B) the co-offending sexual group; (C) the victim-sharing or status network; (D) local networked child sexual exploitation; (E) the OCG involved in child sexual exploitation; and (F) the national or transnational exploitation network. It moves beyond description by examining the different generative motives, group processes, opportunity structures, exploited dependencies, institutional failures and scaling mechanisms associated with each type. The article then uses the Police Foundation's Bristol study as a methodological case study. That research documented serious organised features, but its presumption that almost all adult multi-perpetrator CSE was connected to organised crime illustrates the category error that the proposed framework is intended to prevent. Finally, the article treats trafficking and National Referral Mechanism relevance as a separate, victim specific axis. The central argument is that "grooming gang" should begin an inquiry, not conclude one: grooming is a method, offender plurality is a fact about participation, and gang, network, OCG and trafficking classifications must each be empirically established.
Keywords: child sexual exploitation; grooming; co-offending; organised crime; trafficking; National Referral Mechanism; criminal networks
1. Introduction: a label that has become an explanation
When several adults sexually exploit the same child, the plurality of offenders can appear to supply its own explanation. The offenders become a "gang"; their conduct becomes "grooming"; and repeated abuse becomes evidence of "organised crime". Each step, however, introduces a distinct empirical proposition. Evidence that two people abused the same child does not by itself establish that they formed a durable group, possessed a shared criminal identity, divided tasks, controlled territory, pursued a collective benefit or participated in a trafficking enterprise.
This insistence on classification is not semantic fastidiousness, and it does not diminish the gravity of the abuse. On the contrary, it directs attention to how abuse was made possible. If a friendship group produces offending through imitation and peer reinforcement, its causal structure differs from a local network in which children, vehicles, premises and drugs circulate among loosely connected actors. Both differ again from an OCG that coordinates recruitment and exploitation as a collective criminal purpose, and from a distributed trafficking system whose resilience depends upon movement across police-force or national borders. A generic response to these configurations is likely to miss the mechanisms that sustain them.
Official definitions have often been intentionally broad for safeguarding and investigative reasons. The Home Office described group-based CSE as sexual exploitation involving multiple connected offenders, while also finding no single organisational structure: identified networks ranged from close groups to loose associations rooted in friendship, kinship, employment and locality (Home Office, 2020, paras 84- 108). The Independent Inquiry into Child Sexual Abuse (IICSA) similarly adopted a working definition of an organised network encompassing two or more associated people involved in or facilitating CSE. It expressly included cases regardless of how organised the association proved to be, because a wide investigative frame helped prevent loosely connected facilitators from disappearing from institutional view (IICSA, 2022, paras 20-28).
That breadth is defensible as a safeguarding gateway. It is not proof that every case admitted through the gateway is the same criminological object. IICSA itself distinguished people who actively worked together and passed children between them from individuals connected only through looser local or online associations (IICSA, 2022, paras 20-23). More recent national scrutiny has reached a related conclusion. The Casey audit found that "grooming gangs" were not consistently captured in official datasets and that broad group-based categories could combine exploitation, intra-familial abuse, institutional offending and child-on-child abuse, obscuring rather than clarifying patterns (Casey, 2025). HMICFRS (2023) likewise found no common operational understanding of group-based CSE across policing.
Co-offending research supplies a further warning. Criminal collaboration is often fluid and temporary rather than organisationally stable (Reiss, 1988). Groups that sexually abuse children can range from transient pairings to associations with recognised membership, terminology and collective identity. A minimal definition of a co-offending group therefore establishes connected offending, not hierarchy, durability or an overarching enterprise (Cockbain, Brayley and Sullivan, 2014). Interview research with people convicted alongside others also found heterogeneous and often informal connections through family, friendship, work and locality, although the self-presentational character and limited generalisability of convicted-offender accounts require caution (Senker, Scott and Wainwright, 2020).
The article develops a causal typology capable of preserving safeguarding breadth while restoring criminological discrimination. It asks five questions of every formation: why did the offenders come together; what did different participants seek; which opportunities and resources made the abuse practicable; what allowed it to persist; and what enabled it to expand to additional victims, perpetrators or places? These questions reveal that "grooming gangs" are not one phenomenon with variable degrees of sophistication. They are a family of sometimes overlapping phenomena generated and sustained by different mechanisms.
Grooming is a method. Multiple perpetration is a pattern of participation. A gang, network, OCG or trafficking system is an organisational claim that must be proved.
2. Separating conduct, connection, organisation and reach
Four analytical levels are commonly collapsed. The first concerns conduct towards the child: grooming, coercion, rape, exchange, transportation, harbouring and exploitation. The second concerns the relationship among offenders: strangers, friends, relatives, colleagues, neighbours, drug-market associates or members of a pre-existing peer group. The third concerns organisation: whether offending was spontaneous, recurrent, reciprocal, role-differentiated, profit-seeking, territorially embedded or coordinated through an enduring criminal purpose. The fourth concerns reach: whether activity remained local, crossed several towns or police areas, or distributed functions across national borders.
Proof at one level does not settle another. Grooming is a process of access, conditioning, dependency and control. It may be used by a lone offender, transmitted through a friendship network, or incorporated into an OCG's operating methods. Conversely, multiple-perpetrator rape can occur without an identifiable grooming phase. The government's practitioner definition of CSE identifies an exploitative power imbalance, coercion, manipulation or deception, and some form of exchange or advantage; it defines conduct towards a child, not the organisation of offenders (Department for Education, 2017). Horvath and Kelly's (2009) work on naming multiple-perpetrator rape is instructive here: an event involving two or more perpetrators is not necessarily evidence of a durable gang. Research on group sexual offending likewise points to a combination of individual propensities, sociocultural beliefs, immediate situations and group processes rather than a single group effect (Harkins and Dixon, 2010).
"Gang" and "organised crime" are also not synonyms. The Eurogang framework, designed principally for youth and street-group research, defines a street gang through durability, street orientation, youthfulness and a group identity that includes illegal activity (Weerman et al., 2009). Its direct application to adult CSE networks would be inappropriate, but it demonstrates why shared ethnicity, neighbourhood or appearance in the same trial is insufficient. Research on London street gangs similarly shows that organisation and criminal commitment vary within and between nominal gangs (Densley, 2012).
The domestic statutory OCG benchmark is different. Under section 45 of the Serious Crime Act 2015, an OCG comprises three or more people who act, or agree to act, together to further a purpose of carrying on criminal activities. The relevant activities must be planned, coordinated or continuing, capable of constituting serious offences, and undertaken with a view to obtaining a gain or benefit (Serious Crime Act 2015, s.45). The legislation does not demand a mafia-like hierarchy. Its explanatory notes state that benefit can include a tangible non-monetary advantage, but whether sexual access, status or another claimed benefit satisfies the statutory test is a case-specific legal question, not something that follows automatically from the abuse. The United Nations Convention against Transnational Organized Crime applies a distinct international formulation, requiring a structured group of three or more people, existing for a period and acting together to commit serious crime for financial or other material benefit (United Nations, 2000, art. 2). The difference cautions against treating "OCG" as a universally self defining term.
The expression "victim-sharing" also requires ethical and analytical care. It is used here to describe offenders' circulation or exchange of access to a child, not to endorse the proprietary logic implicit in the offenders' conduct. Its evidential advantage is that it describes an observable relation: one actor introduces, supplies, delivers, refers or makes a child accessible to another. "Status", by contrast, is a possible motivation. It must be demonstrated through communications, testimony, boasting, reciprocity or other evidence rather than inferred from the fact of multiple abuse.
3. From a descriptive typology to a causal framework
A typology that only renames offender arrangements has limited explanatory value. The analytical task is to identify the mechanisms that produce each arrangement and allow it to reproduce itself. Popular discussions of the "root causes" of trafficking often bring welcome attention to poverty, gender inequality, weak protection, discrimination, migration insecurity and failures of justice. Yet they may place offender motives, market incentives, victim exposure and state failure in one undifferentiated list (for an illustrative educational formulation, see IPJC, 2014). This can accidentally make a child's poverty, care experience, missing episodes or drug dependency appear to cause exploitation.
A more precise account separates six domains.
1. Generative drivers are the incentives and orientations that produce reasons to offend: sexual entitlement, domination, peer approval, reciprocal access, financial return, criminal diversification and demand from other abusers or clients.
2. Group and network mechanisms explain why association matters: social learning, imitation, diffusion of responsibility, network trust, mutual validation, task allocation, secrecy, reputation and enforcement capacity.
3. Opportunity and facilitation concern the practical convergence of offenders, children and settings: employment, neighbourhoods, social media, cars, hotels, takeaways, flats, party houses, alcohol, drugs, legitimate businesses and existing criminal-market infrastructure.
4. Exploited exposure and dependency concern conditions offenders identify and manipulate: material need, family violence, unstable care, homelessness, school exclusion, loneliness, disability, mental distress, addiction, insecure immigration status and earlier victimisation. These conditions shape exposure and constrain exit; they do not transfer causal responsibility to the child.
5. Institutional sustaining conditions explain continuity: disbelief, victim-blaming, criminalisation, fragmented intelligence, failures to link missing episodes with exploitation, weak regulation of premises, racialised assumptions, poor witness protection and jurisdictional gaps.
6. Scaling mechanisms explain expansion: circulation of access, coerced recruitment of peers, differentiated roles, pooled resources, digital communication, links between local markets and exploitation of force or border fragmentation.
This framework draws on opportunity theory without reducing exploitation to opportunity. In six British internal-trafficking investigations, Cockbain and Wortley (2015) found that routine associations and everyday settings helped facilitate, sustain and spread offending; the data did not support a stereotype of uniformly sophisticated criminal organisations. Better-connected offenders were implicated in more offences, and abuse occurred in public and semi-public as well as private spaces. Contextual safeguarding makes the complementary point that extra-familial harm is produced within peer, school, neighbourhood and online contexts, so protective work must engage those environments rather than concentrating only on the child's family (Firmin, 2018; Firmin and Lloyd, 2022).
Social embeddedness is equally important. Organised offenders commonly recruit and collaborate through existing social relations because trust is scarce and betrayal costly (Kleemans and van de Bunt, 1999). This helps explain why offender networks often reflect pre-existing patterns of friendship, family, work and neighbourhood. It does not make ethnicity a causal variable. Ethnic composition may indicate the social settings through which people met, but neither ethnicity nor culture should substitute for evidence of motivation, criminal agreement or organisation. Specific norms, rationalisations or community protections can be examined when evidence demonstrates their relevance; they cannot be inferred from group composition.
Table 1 sets out the six ideal types. They are not mutually exclusive boxes or a ladder from less to more serious. A Type B assault may involve catastrophic violence; a Type C network may generate extensive trafficking risk without satisfying an OCG test; and a highly organised Type E group may remain geographically local. Cases may also change over time as recurrent co-offending becomes routinised, victim access is circulated and roles differentiate.
Table 1. Six-type causal typology
Type | Defining relation | Dominant causal question |
A. Lone sexual exploiter or trafficker | One offender acts independently, although third parties may be unwittingly used | How does an individual acquire access, establish dependency and avoid detection? |
B. Co-offending sexual group | Two or more offenders participate together without an established victim-circulation system or collective criminal enterprise | How do immediate group processes, pre-existing peer relations and settings generate or reinforce offending? |
C. Victim-sharing or status network | Access to children is introduced, referred, exchanged or circulated among connected actors | How does access become social capital, reciprocity, payment or a means of network reproduction? |
D. Local networked CSE | Recurrent exploitation is sustained by overlapping local people, places and resources without a proven network-wide criminal purpose | How does a local opportunity ecology allow exploitation to persist and spread? |
E. OCG involved in CSE | Three or more actors pursue a demonstrable collective criminal purpose through planned, coordinated or continuing activity | How do shared infrastructure, roles, benefit and enforcement capacity turn abuse into organised criminal activity? |
F. National or transnational exploitation network | Recruitment, movement, harbouring, control or exploitation is functionally coordinated across areas or borders | What advantage is obtained from geographic distribution and jurisdictional fragmentation? |
4. Type A: the lone sexual exploiter or trafficker
Type A is essential even though it falls outside any ordinary meaning of "gang". It demonstrates that grooming, sexual exploitation and trafficking do not depend on offender plurality. A lone actor may establish a relationship online or offline, identify unmet emotional or material needs, provide affection, money, accommodation, alcohol or drugs, and progressively normalise sexual access. The generative drivers may include sexual gratification, entitlement, domination, attachment needs, financial gain or a combination of motives. What distinguishes the type is not motive but independent operation.
Opportunity is commonly produced by legitimate social roles and routine contact. Employment, intimate relationships, family connections, housing, transport, leisure settings and digital platforms can provide access and credibility. Grooming reduces the offender's perceived risk by making disclosure harder, producing ambivalence, isolating the child or creating a record that the offender later misrepresents as consent. Institutional disbelief, especially where professionals interpret exploitation as a child's lifestyle or "risky behaviour", can become a sustaining condition.
"Lone exploiter" and "lone trafficker" are not synonymous. Under section 2 of the Modern Slavery Act 2015, trafficking involves arranging or facilitating another person's travel with a view to exploitation; the statutory action can include recruitment, transportation, transfer, harbouring or receipt (Modern Slavery Act 2015, s.2). Domestic and very short-distance movement can qualify, and British children can be internally trafficked (Brayley and Cockbain, 2014). But grooming or repeated sexual exploitation does not automatically prove a trafficking action. Calling every lone groomer a trafficker would reproduce the same analytical inflation that this article criticises in collective cases.
Type A can develop into other types. A lone offender may introduce the child to friends, exchange access, recruit facilitators, or sell access to others. The transition is empirically significant: it changes the risk of repeated exploitation and the likely disruption strategy. Investigators should therefore ask not only whether the offender acted alone at a particular incident, but whether communications, travel, locations or referrals reveal a wider system.
5. Type B: the co-offending sexual group
Type B describes connected participation without assuming an established system for circulating victims or a continuing collective criminal purpose. Three subtypes are analytically useful. Type B1 is episodic collective rape: offenders converge for a single incident or immediate sequence. Type B2 is recurrent friendship-based co-offending: friends, relatives, colleagues or associates repeatedly facilitate or participate in abuse, but do not distribute access through a wider network. Type B3 is offending by a pre-existing peer or territorial group whose identity exists independently of CSE. Sexual violence may be central to that group, but it may instead be episodic, ancillary to drugs or violence, or used in status competition and punishment.
The formation pathway is usually social rather than entrepreneurial. Offenders are already together. One person may initiate, others may imitate or comply, and the group can intensify physical control over the victim. Harkins and Dixon's (2010) review indicates that explanations must combine individual, sociocultural, situational and group-level factors. Qualitative research with convicted multiple perpetrator rapists has identified perceived pressure, acceptance, camaraderie and "adventure" among the accounts offenders give for participation, although its Portuguese sample and convicted-offender design limit transferability (da Silva, Woodhams and Harkins, 2018).
Peer influence should not be converted into an excuse. Group participation can disinhibit conduct, provide an audience and validate misogynistic beliefs, but each participant remains an agent. Nor should diffusion of responsibility be treated as a universal law. In a study of 71 victim accounts involving 189 suspects, Woodhams, Taylor and Cooke (2020) found that the behaviour of an apparent leader was associated with group violence, while simple group size did not operate as a straightforward escalator of hostility. The finding supports a relational analysis of influence and leadership rather than an assumption that more offenders always produce the same group effect.
Opportunity often lies in party settings, alcohol and drug use, vehicles, private houses, hotels or spaces controlled by the group. Repetition may be sustained by shared secrecy, retrospective normalisation, victim intimidation and the knowledge that disclosure implicates several people. In a territorial group, reputation and local control may make complaints appear dangerous or futile. Yet these features still do not establish an OCG. The empirical questions are whether the group has durability and identity beyond the incident, whether sexual offending is part of that identity, and whether members agreed to pursue a continuing criminal purpose.
The boundary with Type C turns on circulation. If the same friends repeatedly offend together, Type B2 may be sufficient. If an initial abuser begins introducing the child to separate associates, if access is reciprocated, or if participation becomes a resource that connects people who do not otherwise co offend, the case is better understood as a victim-sharing network. This distinction matters for trafficking assessment because an isolated collective assault may present fewer trafficking indicators than a system of transfer, receipt and repeated exploitation.
6. Type C: the victim-sharing or status network
Type C begins where access to children circulates. One offender may act as a gateway, introducing a child to friends or associates; participants may exchange access reciprocally; money, drugs or favours may mediate introductions; or an exploited child may be deceived or coerced into recruiting peers. The network is reproduced through referral rather than simply through simultaneous participation.
The principal generative driver is the conversion of access into a network resource. For some actors it may produce status, approval or "kudos". For others it may repay a favour, secure reciprocal access, generate money, consolidate friendship or signal trust. A gatekeeper may acquire influence by controlling introductions. Different participants in the same network need not share a motive: one may groom, another provide drugs, a third supply a location, and a fourth seek prestige. What holds the configuration together is functional complementarity.
Network trust lowers the perceived costs of offending. Approaching a child independently carries risks of rejection, exposure and police attention. A trusted introduction transfers information about the child's circumstances, the groomer's methods and the locations in which access is possible. Repeated circulation also normalises abuse: when associates behave similarly, individual offenders receive social confirmation that their conduct is accepted within the network. Senker, Scott and Wainwright (2020) found that convicted participants described access spreading through casual connections among friends, relatives, partners and colleagues, often without a formal hierarchy. Their accounts cannot establish the true absence of organisation, but they demonstrate why horizontal social relations deserve investigation.
Victim peer networks can become an additional scaling mechanism. A child may introduce friends or siblings while seeking companionship, acting under pressure, attempting to reduce danger, or because grooming has distorted their understanding of the situation. These actions should not be redescribed as voluntary facilitation without examining coercion and control. The Police Foundation's Bristol data found that 70 per cent of identified young people were linked to another child considered at risk, illustrating how exploitation could propagate through peer relationships as well as offender ties (Crocker et al., 2017, p. 52).
Type C can be extensive without possessing clear boundaries, leadership or a single network-wide purpose. It can therefore carry high trafficking and safeguarding relevance while falling short of Type E. The decisive evidence includes introductions, contact sharing, communications about access, transport or delivery to other offenders, reciprocal arrangements, payment and coerced recruitment. Mere serial abuse by several people is not enough to infer a circulation system.
7. Type D: local networked child sexual exploitation
Type D describes recurrent exploitation emerging from an embedded local ecology of people, places and resources. The network may include abusers, facilitators, victims, friends, relatives, workers, landlords, drivers and drug suppliers; not every actor knows every other actor, and some participants may not share a single criminal purpose. Its boundaries are porous because ordinary social and commercial relationships coexist with criminal ones.
Formation is driven by proximity and repeated convergence. Offender friendship and kinship networks intersect with vulnerable children's peer groups, care placements, schools, transport routes, night-time economies and drug markets. Cars, taxis, takeaways, hotels, flats, parks, private parties and businesses can become enabling infrastructure. Alcohol and drugs may be gifts, tools of dependency, means of incapacitation or features of the setting. An older "boyfriend" relationship can supply an initial bridge from a child to the offender's associates.
The dominant explanation is an opportunity ecology rather than a central plan. Cockbain and Wortley (2015) found that internal child sex trafficking was facilitated by routine activities and everyday associations, often in visible settings, and that offenders were not uniformly specialised or sophisticated. Firmin's contextual safeguarding framework similarly directs attention to how neighbourhood, peer and institutional contexts allocate risk and protection (Firmin, 2018). Deprivation does not cause sexual violence. It can, however, concentrate conditions that offenders exploit, including material need, unstable housing, reduced safe leisure space, open drug markets and over-stretched services.
Persistence is commonly institutional. Missing episodes may be processed separately from drug use, school exclusion and sexual harm; police intelligence may not connect people known under different names or across different cases; social care, health and education may hold fragments of the same pattern. Familiar local offenders can appear ordinary, while children are labelled troublesome or complicit. Racialised assumptions can operate in both directions, producing excessive suspicion toward some communities and reluctance to investigate others for fear of controversy. A defensible analysis examines decisions, resources and evidence, not a single motive attributed retrospectively to all institutional failures.
The distinction between Type D and Type E is not hierarchy. A decentralised network can qualify as organised crime, while a locally notorious group may not. The question is whether evidence demonstrates a collective criminal purpose and coordinated capacity across the relevant actors. In Type D, exploitation can proliferate through overlapping opportunities even when no one directs the whole network. In Type E, planned, coordinated or continuing criminal activity is a purpose of the group, supported by shared resources, roles or enforcement.
8. Type E: the organised crime group involved in CSE
Type E applies when the OCG classification is independently established rather than inferred from multiplicity. Three or more people must act or agree to act together in furtherance of criminal activity, and the evidence should show planning, coordination or continuity alongside a contemplated gain or benefit (Serious Crime Act 2015, s.45). The group need not have a name, rigid hierarchy or permanent membership. Networked and adaptive forms of organised crime are well documented, and local embeddedness can be a source of resilience rather than evidence of disorganisation (Kleemans and van de Bunt, 1999; Hobbs, 2013).
Two causal variants should be distinguished. In Type E1, child sexual exploitation is a principal collective activity. Recruitment, grooming, transport, accommodation, control and abuse may be divided among actors. The enterprise can be sustained by financial return, sexual access or both, but the legal benefit element must be evidenced rather than presumed. In Type E2, a polycriminal OCG primarily engaged in drug supply, violence, fraud or another market also becomes involved in CSE. Sexual exploitation may arise through access to drug-dependent children, use of properties already controlled for criminal purposes, debt and punishment, or members treating sexual access as a privilege of criminal standing.
The distinction is causally important. An exploitation-centred OCG recruits resources for CSE; a polycriminal OCG may repurpose existing resources. Vehicles, premises, cash, weapons, contacts and corrupt or coerced facilitators reduce the marginal cost of abuse. Children criminally exploited in drug supply may also be sexually exploited, while sexual violence may function instrumentally to intimidate, punish, blackmail or bind them to the group. The overlap between child criminal and sexual exploitation is therefore not merely the coincidence of two offence categories; it may reflect a shared coercive system.
Organisation can increase persistence through role differentiation and replaceability. Recruiters, drivers, premises providers, enforcers and abusers need not all participate in the same sexual offence. Arresting a visible perpetrator may leave the infrastructure intact. Reputation, violence and access to wider criminal networks can intimidate victims, families and witnesses. Financial resources may fund legal defence, relocation or replacement of compromised premises.
At the same time, drug links should not be treated as a shortcut to OCG classification. Drug supply can actively enable exploitation; drug-involved offenders and vulnerable children can simply occupy the same local environment; and police may know more about drug-linked suspects because those people are already surveilled. These competing hypotheses were raised in the Police Foundation study and require case-specific testing (Crocker et al., 2017, pp. 49-50). An association with drugs is an investigative lead, not a conclusion about the structure or cause of CSE.
9. Type F: national or transnational exploitation networks
Type F is defined by functional coordination across areas or borders, not by size alone. Recruitment may occur in one town or country, transportation through another, accommodation elsewhere and exploitation by a further set of actors. Digital communications, payments, family or business links and enforcement arrangements can connect the nodes. A network can operate nationally or transnationally without rigid hierarchy; conversely, a sophisticated OCG can remain entirely local.
The distinctive driver is advantage gained from geographic distribution. Moving a child can isolate them from family, peers and trusted professionals, create dependence on the people controlling transport or documents, provide access to new offenders or clients, and exploit differences in law enforcement and data systems. Offenders may use hotels, private rentals, transport businesses and online platforms to coordinate at distance. Immigration insecurity, threats concerning authorities or document control can deepen dependency in cross-border cases, while domestic victims can be similarly isolated by movement between towns.
Scale must not be inferred from movement alone. Taking one child to another city may constitute trafficking, but it does not prove a national network. Similar offending in several towns may reflect parallel local patterns rather than coordination. Type F requires communications, referrals, financial flows, shared personnel or differentiated functions showing interdependence across locations. The same evidential discipline applies to transnationality: nationality of offenders or victims is not evidence of a transnational criminal system.
Institutional fragmentation is a sustaining condition. Separate police forces may each see an apparently local fragment; national borders create differences in legal definitions, intelligence standards and victim support; and victims may be recorded under different identities. The appropriate response therefore depends upon joint intelligence, financial and communications analysis, cross-force ownership and, where necessary, international cooperation. Yet the use of those capabilities is a response to evidential complexity, not retrospective proof that every case was an OCG.
10. The Police Foundation study as a worked methodological test
The Police Foundation's study of organised crime in local communities is a central source because it displays both the value and the danger of an organised-crime lens. Its Bristol analysis used police data from 2013-2014 to identify 371 alleged, suspected, charged or convicted CSE perpetrators. Of these, 192 were linked to at least one other offender and were allocated to 43 groups with an average size of five; the great majority of those people had not already been mapped by police as OCG members (Crocker et al., 2017, pp. 47-49). The underlying material revealed drug supply, cross-area movement, premises, firearms, intimidation, financial gain, older-boyfriend recruitment and children being induced to recruit other children.
Those findings show why serious-crime capabilities may be necessary. They also require methodological qualification. The study classified CSE cases involving two or more offenders working together as groups. Peer-on-peer cases were excluded because they were considered less likely to involve organised crime, and the remaining adult multiple-perpetrator cases were presumed to have an organised-crime connection for the purpose of analysis (Crocker et al., 2017, p. 49; pp. 96-97). The resulting 43 groups were therefore not 43 independently verified OCGs. The initial classification built the organised-crime connection into the sample.
The problem is not cured by the seriousness of what researchers later found. Rather, the evidence should be reclassified. An older boyfriend who grooms a child and then introduces them to associates may represent a transition from Type A to Type C. A sex party may be Type B1 if episodic, Type B2 if repeated among the same friends, or Type C/D if access and locations are recurrently circulated. A drug dealing group may be Type D where exploitation emerges through a local opportunity ecology, or Type E2 where a proven OCG repurposes premises, transport, drugs and enforcement capacity. Kidnapping, confinement and firearms are control methods, not organisational types. Financial gain is a motive or benefit, while a child being compelled to recruit peers is a scaling mechanism with potentially strong trafficking implications.
The report itself supplies evidence for this more discriminating analysis. It found an older-boyfriend pattern in 16 groups; links with drug supply in 11; financial gain in nine; and extreme threats, kidnapping, violence or firearms in eight (Crocker et al., 2017, pp. 49-50). These characteristics overlapped and operated at different analytical levels. Treating them as equivalent "OCG features" would conflate recruitment, setting, market connection, control and motive.
Its treatment of drug markets is especially valuable. The researchers considered whether drug activity directly facilitated CSE, whether drug offenders and vulnerable children were simply exposed to the same local conditions, or whether enforcement against drugs made certain offenders more visible to police (Crocker et al., 2017, pp. 49-50). This is causal reasoning rather than labelling. It asks which mechanism links two observed phenomena and recognises that police data are generated by enforcement priorities as well as offending.
The study therefore supports two propositions at once. Some group-based CSE plainly overlaps with trafficking, organised drug markets, profit, coercive premises and cross-area offending, and should not be reduced to a collection of isolated sexual offences. But multiple adult perpetrators are not sufficient evidence of an OCG. The proposed typology can be understood as a reclassification framework for the heterogeneous formations that the study placed under one initial organised-crime heading.
11. Trafficking and the National Referral Mechanism: a cross-cutting, victim-specific axis
Trafficking is not the highest rung of the typology. It is a legal and safeguarding classification of what was done to a particular person. Under the Modern Slavery Act 2015, arranging or facilitating travel with a view to exploitation can include recruitment, transportation, transfer, harbouring or receipt. Current statutory guidance also treats the transfer or exchange of control as relevant action. In child cases, only the action and exploitative purpose must be established; the coercive "means" required in adult trafficking definitions need not be shown (Home Office, 2026a, paras 2.3-2.12).
The Home Office's 2026 statutory guidance provides a useful evaluative framework for sexual exploitation. Trafficking is more strongly indicated where offending is repeated or prolonged, connected offenders create a risk of continued exploitation, the conduct forms part of a wider pattern, an offender obtains monetary, social or political benefit, or grooming occurs. An isolated rape or sexual assault with no wider features is less indicative, although every case remains fact-sensitive (Home Office, 2026a, paras 2.26-2.30).
This supports a critical distinction: collective rape describes participation in an event; circulation of access describes transfer, receipt or exchange within an exploitative process. The latter is often more relevant to the NRM. A flat friendship network may therefore present strong trafficking indicators without being an OCG, while an established gang or OCG does not automatically make every sexual offence committed by its members an instance of trafficking.
Table 2. Indicative NRM relevance across the six types
Type | Likely NRM relevance | Reason for the assessment |
A. Lone sexual exploiter or trafficker | Variable for a lone exploiter; very high once trafficking action and purpose are established | A lone person can recruit, move, harbour or receive a child for exploitation, but grooming alone is not synonymous with trafficking |
B. Co-offending sexual group | Variable, often lower for a genuinely isolated assault | Relevance rises where the child is brought to the group, moved, controlled, repeatedly exploited or placed at risk of further exploitation |
C. Victim-sharing or status network | Often high | Introductions, delivery, receipt, exchange of control and coerced recruitment can closely track trafficking actions |
D. Local networked CSE | Often high, but not automatic | Local distance is no bar; transport, harbouring, premises, dependency and repeated receipt may establish internal trafficking |
E. OCG involved in CSE | Frequently high, but not inherent in OCG status | Organised infrastructure may facilitate trafficking, while an isolated offence by OCG members may not qualify |
F. National or transnational exploitation network | Very high where the network is truly organised around trafficking | Cross-area coordination, movement and distributed control are powerful indicators, but geography alone is insufficient |
The NRM consequence is practical. Where a First Responder considers that a person who is, or may be, under 18 is a potential modern-slavery victim, a referral must be made and the child's consent is not required. Adults generally must give informed consent, including adults referred for exploitation experienced in childhood; the substantive assessment of historical childhood exploitation applies the child trafficking criteria (Home Office, 2026a, paras 2.3-2.12; Home Office, 2026b). Trafficking assessment should therefore be conducted for each victim across all six types rather than reserved for cases already labelled organised crime.
12. An empirical classification protocol
A defensible classification should be reconstructed from relationships and conduct rather than adopted from a press label, prosecution title or intelligence category. At minimum, investigators and researchers should examine the following.
7. Duration and stability: How long did offender relationships persist, and did the association survive individual incidents or arrests?
8. Co-offending pattern: Who offended with whom, how often, and in what sequence?
9. Recruitment and circulation: Who first contacted each child; who introduced, referred, transported or delivered them; and were contact details, images or access exchanged?
10. Collective identity and territory: Did a durable peer or street group exist independently of CSE, and what place did illegality and sexual violence occupy within its identity?
11. Resources and settings: Who controlled vehicles, hotels, businesses, flats, party houses, drugs, alcohol, money and communications?
12. Role differentiation: Were there identifiable recruiters, gatekeepers, drivers, premises providers, enforcers, abusers, clients or financial beneficiaries?
13. Purpose and benefit: What collective purpose, if any, can be proved, and what financial, material, reputational, sexual or other benefit was sought by particular actors?
14. Enforcement capacity: Were threats, weapons, criminal reputation, debt, blackmail or violence used to control victims, families, witnesses or other offenders?
15. Polycriminal connections: Did drug supply, criminal exploitation, fraud or violence merely coexist with CSE, facilitate it, or form part of one collective system?
16. Geographic coordination: What communications, payments, referrals or shared personnel connect apparently separate localities or countries?
17. Victim-specific trafficking indicators: For each child, what recruitment, movement, transfer, harbouring, receipt or exchange of control occurred, and for what exploitative purpose?
18. Institutional production of the record: Which patterns may be missing because of non-reporting, misclassification, intelligence fragmentation, enforcement priorities or failures to identify victims?
Ethnicity, occupation and neighbourhood may explain how people encountered one another, but none proves criminal organisation. A joint indictment does not necessarily map a network's true boundary: prosecutors may divide a large network across trials or combine defendants with unequal and peripheral relationships. Likewise, an OCG map can reflect intelligence priorities rather than an exhaustive social structure. Social network analysis can assist, but links must be substantively coded; a shared telephone contact, co-presence at an address and proven victim referral are not equivalent ties. Recent network research has reinforced the need to separate structure from activity, finding that some CSE networks lack the cohesion or centralised power often imagined while still displaying specialised roles (Dixon, 2025).
Classification should be longitudinal and multi-level. One case may begin with Type A grooming, move into Type C circulation, acquire Type D local infrastructure and eventually satisfy Type E. Different subsets of the same network may occupy different types. The unit of analysis must therefore be explicit: incident, dyad, offender cluster, victim pathway, local network or criminal enterprise.
13. Prevention, investigation and disruption
Different causal formations require different interventions. Type A places emphasis on access controls, online and occupational safeguarding, relationship-based work with children, and rapid recognition of coercive dependency. Type B requires attention to peer norms, group leadership, high-risk settings, bystander intervention and evidence that captures encouragement and facilitation as well as principal offending. Type C demands referral-chain and communications analysis, identification of gatekeepers, protection for children coerced to recruit peers, and disruption of the social rewards attached to circulating access.
Type D requires place-based and contextual safeguarding. Agencies should connect missing episodes, education exclusion, drug harm, hotels, transport, housing and local premises rather than process each as a separate problem. Mapping both victim and offender social relations can identify bridges, hotspots and overlooked facilitators. Disruption may involve licensing, civil orders, tenancy and hotel safeguards, but such measures must not displace criminal investigation or simply move the problem elsewhere.
Type E justifies the full repertoire used against serious and organised crime: financial investigation, communications and network analysis, asset denial, covert tactics where lawful and necessary, witness protection and targeting of infrastructure rather than only visible abusers. In Type E2, joint CSE and child-criminal-exploitation strategies are essential because the same debt, properties, drug supply and violence may control both. Type F adds cross-force ownership, national tasking, international cooperation, compatible data and attention to payments, travel, documents and distributed roles.
The Casey audit recommended that group-based CSE investigations receive the seriousness and capabilities associated with major and organised crime (Casey, 2025). That is an operational proposition, not a classificatory shortcut. Police can deploy sophisticated investigative methods against a loose but dangerous network without declaring that network an OCG. Conversely, accurate OCG classification can reveal facilitators, assets and replacement capacity that conventional sexual-offence case building may overlook.
The typology also changes prevention. A universal campaign against an imagined offender profile will not address the different engines of abuse and can reinforce racialised stereotypes. Prevention should instead target mechanisms: access, peer reinforcement, victim circulation, local infrastructure, criminal coercion and jurisdictional gaps. Cultural and community-specific work may be appropriate where locally evidenced, but it should follow analysis rather than replace it.
14. Limits and research priorities
The six types are ideal types, not natural kinds. Boundaries will sometimes be uncertain, and organisational features can be hidden by offenders or distorted by case selection. Conviction samples exclude undetected and unprosecuted cases; police records reflect reporting and enforcement; offender interviews involve denial and self-justification; victim accounts may be fragmented by trauma and institutional responses. Inconsistent definitions make numerical comparison particularly hazardous (Home Office, 2020; HMICFRS, 2023; Casey, 2025).
Future research should follow victim pathways and offender relations over time rather than treating prosecution groups as fixed organisations. Mixed methods are needed: social network analysis to map ties; qualitative work to establish what those ties mean; place analysis to identify opportunity structures; financial and communications evidence to test coordination; and survivor-centred research to understand movement, dependency and institutional contact. Comparisons should examine not only ethnicity but gender, age, class, occupation, criminal-market involvement, online/offline interaction and the distribution of police attention.
Most importantly, research should test causal alternatives. Did a drug market supply resources for exploitation, create contact, or merely make suspects more visible? Did a child introduce peers because of coercion, distorted attachment, companionship or an attempt to manage danger? Did movement serve exploitation, or was it incidental? Did a group's pre-existing identity generate sexual violence, or did shared abuse create a new group? These questions produce evidence capable of supporting differentiated prevention.
15. Conclusion
"Grooming gang" is best treated as a claim requiring disaggregation. It can obscure the difference between a lone trafficker, an episode of collective rape, recurrent friendship-based co-offending, a status economy of victim access, a local exploitation ecology, an OCG and a distributed trafficking network. Those forms can overlap and evolve, but they are driven by different combinations of motivation, group influence, opportunity, dependency, institutional failure and scale.
The analytical sequence should therefore be disciplined. Establish what happened to each child; reconstruct what connected the offenders; determine what, if anything, they organised together; identify the resources, benefits and control mechanisms involved; and assess trafficking separately for every victim. Only then should labels such as gang, network, OCG or transnational trafficking system be applied.
The gain is more than conceptual clarity. Precise classification improves safeguarding, directs investigative resources, reduces guilt by association and identifies the mechanisms that prevention must change. Grooming describes a method of exploitation. Whether the offenders constitute a gang - and, if so, what kind - remains an empirical question.
References
Brayley, H. and Cockbain, E. (2014) 'British children can be trafficked too: towards an inclusive definition of internal child sex trafficking', Child Abuse Review, 23(3), pp. 171-182. https://doi.org/10.1002/car.2307
Casey, L. (2025) National Audit on Group-Based Child Sexual Exploitation and Abuse. London: Home Office. Published 16 June 2025. https://www.gov.uk/government/publications/national-audit-on-group-based-child-sexual-exploitation-and-abuse
Cockbain, E., Brayley, H. and Sullivan, J. (2014) 'Towards a common framework for assessing the activity and associations of groups who sexually abuse children', Journal of Sexual Aggression, 20(2), pp. 156-171.
Cockbain, E. and Wortley, R. (2015) 'Everyday atrocities: does internal (domestic) sex trafficking of British children satisfy the expectations of opportunity theories of crime?', Crime Science, 4, article 35. https://doi.org/10.1186/s40163-015-0047-0
Crocker, R., Webb, S., Garner, S. and Skidmore, M. (2017) The Impact of Organised Crime in Local Communities. London: Police Foundation and Perpetuity Research. June 2017. https://www.police-foundation.org.uk/wp-content/uploads/2017/10/oc_in_local_communities_final.pdf
da Silva, T., Woodhams, J. and Harkins, L. (2018) '"An adventure that went wrong": reasons given by convicted perpetrators of multiple perpetrator sexual offending for their involvement in the offense', Archives of Sexual Behavior, 47(2), pp. 443- 456. https://doi.org/10.1007/s10508-017-1011-8
Densley, J. (2012) 'The organisation of London's street gangs', Global Crime, 13(1), pp. 42-64.
Department for Education (2017) Child Sexual Exploitation: Definition and a Guide for Practitioners, Local Leaders and Decision Makers Working to Protect Children from Child Sexual Exploitation. London: DfE. February 2017. https://www.gov.uk/government/publications/child-sexual-exploitation-definition-and-guide-for-practitioners
Dixon, S. (2025) 'Child sexual exploitation (CSE) networks: reassembling structure and activity', Journal of Sexual Aggression, 31(3), pp. 396-412. First published online 9 July 2024. https://doi.org/10.1080/13552600.2024.2374076
Firmin, C. (2018) 'Contextualising case reviews: a methodology for developing systemic safeguarding practices', Child and Family Social Work, 23(1), pp. 45-52. https://doi.org/10.1111/cfs.12382
Firmin, C. and Lloyd, J. (2022) 'Green lights and red flags: the (im)possibilities of contextual safeguarding responses to extra familial harm in the UK', Social Sciences, 11(7), article 303. https://doi.org/10.3390/socsci11070303
Harkins, L. and Dixon, L. (2010) 'Sexual offending in groups: an evaluation', Aggression and Violent Behavior, 15(2), pp. 87-99. https://doi.org/10.1016/j.avb.2009.08.006
HMICFRS (2023) An Inspection of the Effectiveness of the Police and Law Enforcement Response to Group-Based Child Sexual Exploitation in England and Wales. London: His Majesty's Inspectorate of Constabulary and Fire & Rescue Services. https://hmicfrs.justiceinspectorates.gov.uk/publication-html/effectiveness-of-police-and-law-enforcement-response-to group-based-child-sexual-exploitation/
Hobbs, D. (2013) Lush Life: Constructing Organised Crime in the UK. Oxford: Oxford University Press.
Home Office (2020) Group-Based Child Sexual Exploitation: Characteristics of Offending. London: Home Office. Published 15 December 2020. https://www.gov.uk/government/publications/group-based-child-sexual-exploitation-characteristics-of offending
Home Office (2026a) Modern Slavery: Statutory Guidance for England and Wales (under s49 of the Modern Slavery Act 2015) and Non-Statutory Guidance for Scotland and Northern Ireland, version 4.6. London: Home Office. Published 11 May 2026. https://www.gov.uk/government/publications/modern-slavery-how-to-identify-and-support-victims
Home Office (2026b) Guidance on the National Referral Mechanism for Potential Adult Victims of Modern Slavery: England and Wales. London: Home Office. Updated 7 April 2026. https://www.gov.uk/government/publications/human-trafficking victims-referral-and-assessment-forms/guidance-on-the-national-referral-mechanism-for-potential-adult-victims-of modern-slavery-england-and-wales
Horvath, M.A.H. and Kelly, L. (2009) 'Multiple perpetrator rape: naming an offence and initial research findings', Journal of Sexual Aggression, 15(1), pp. 83-96. https://doi.org/10.1080/13552600802653818
Independent Inquiry into Child Sexual Abuse (IICSA) (2022) Child Sexual Exploitation by Organised Networks Investigation Report. London: IICSA. February 2022. https://www.iicsa.org.uk/reports-recommendations/publications/investigation/cs organised-networks.html
Intercommunity Peace and Justice Center (IPJC) (2014) Root Causes of Human Trafficking. Seattle, WA: IPJC. https://ipjc.org/wp-content/uploads/2016/09/USCSAHT%20-%20HT%20and%20Root%20Causes%20module.pdf
Kleemans, E.R. and van de Bunt, H.G. (1999) 'The social embeddedness of organized crime', Transnational Organized Crime, 5(1), pp. 19-36.
Modern Slavery Act 2015, c.30. https://www.legislation.gov.uk/ukpga/2015/30/section/2
Reiss, A.J. Jr (1988) 'Co-offending and criminal careers', Crime and Justice, 10, pp. 117-170. https://doi.org/10.1086/449145
Senker, S., Scott, M. and Wainwright, L. (2020) An Explorative Study on Perpetrators of Child Sexual Exploitation Convicted Alongside Others. London: IICSA. https://www.iicsa.org.uk/reports
recommendations/publications/research/perpetrators-child-sexual-exploitation-convicted-alongside-others.html Serious Crime Act 2015, c.9. https://www.legislation.gov.uk/ukpga/2015/9/section/45
United Nations (2000) United Nations Convention against Transnational Organized Crime, General Assembly resolution 55/25, 15 November 2000. https://www.unodc.org/unodc/en/organized-crime/intro/UNTOC.html
Weerman, F.M., Maxson, C.L., Esbensen, F.-A., Aldridge, J., Medina, J. and van Gemert, F. (2009) Eurogang Program Manual: Background, Development, and Use of the Eurogang Instruments in Multi-Site, Multi-Method Comparative Research. University of Missouri at St Louis. https://pure.manchester.ac.uk/ws/files/45767040/FULL_TEXT.PDF
Woodhams, J., Taylor, P.J. and Cooke, C. (2020) 'Multiple perpetrator rape: is perpetrator violence the result of victim resistance, deindividuation, or leader-follower dynamics?', Psychology of Violence, 10(1), pp. 120-129. https://doi.org/10.1037/vio0000255
Comments