The current issue and full text archive of this journal is available at www.emeraldinsight.com/1363-951X.htm PIJPSM 33,2 Crowd psychology, public order police training and the policing of football crowds 218 James Hoggett and Clifford Stott School of Psychology, University of Liverpool, Liverpool, UK Received 12 July 2008 Accepted 28 April 2009 Abstract Purpose – This study seeks to examine what theory of crowd psychology is being applied within public order police training in England and Wales and what accounts of crowds, police strategies and tactics subsequently emerge among officers who undertake this training. Design/methodology/approach – The study uses a multi-method approach including observations of public order training courses, interviews with students and instructors, and the dissemination of questionnaires. Findings – The analysis suggests that a form of crowd theory associated with the work of Gustave Le Bon has become institutionalised within police training. This in turn is leading to a potentially counter-productive reliance on the undifferentiated use of force when policing crowds. Practical implications – The study illustrates that such training outcomes not only are counter to the recent developments in evidence, theory and policy but also undermine the police’s ability to develop more efficient and effective approaches to policing crowds. Originality/value – The study provides a systematic review of public order training which demonstrates how crowd theory is used as a rationale and justification for the use of tactics based on undifferentiated force. It makes suggestions for improving police training so that updates in policy and theory can be translated into operational practice. Keywords Psychology, Police, Training Paper type Research paper Introduction There can be little contention that Gustave Le Bon’s seminal work The Crowd has had a lasting impact on how crowds and their psychology are understood (Le Bon, 1895). Central to Le Bon’s theory was the idea that the anonymity of the crowd leads to a loss of individual identity and thus the capability of individuals to resist any passing ideas or emotions. People can thus become “submerged” within an atavistic collective unconscious or “group mind”, characterised by primitive barbarism, contagion and power. In contrast Floyd Allport (1924) argued that crowds generate a spiralling atmosphere of emotionality; this mutual-excitation eventually leads to over-stimulation, the suppression of individuality and the emergence of a biological survival instinct. Thus, despite their apparent differences both these “classic” theories put forward the view that being in crowds leads ordinary people to lose the meaningful control of their own behaviour. Policing: An International Journal of Police Strategies & Management Vol. 33 No. 2, 2010 The authors would like to thank the ACPO football portfolio holder Stephen Thomas for his help pp. 218-235 and support. The work was funded by the Economic and Social Research Council (ESRC) Case q Emerald Group Publishing Limited 1363-951X Award PTA/033/2005/00044. Thanks also to all the police officers who helped to facilitate the DOI 10.1108/13639511011044858 work and support the authors. Thanks to John Drury for his help in editing the work.  These theories of the crowd are still relevant today because they can underpin the Policing football psychology of those who are charged with controlling crowds (Hoggett and Stott, n.d.). crowds Among the police in the UK, for example, these two classical theories combine in what have been termed the “agitator” model of crowd disorder (Stott and Reicher, 1998a,b). This model is based on the premise that crowds attract groups and individuals who seek to “stir-up” or agitate violence. They are then able to “hijack” crowds because of the inherent gullibility of the mass. Put differently a crowd is a situation whereby the 219 bad can easily lead the mad. These “classic” theoretical models of crowd psychology have now been largely superseded in the scientific literature by theories that accept the rational and normative nature of crowd action (e.g. McPhail, 1991; Turner and Killian, 1987). Chief among these in psychology is the Elaborated Social Identity Model (ESIM) of crowd behaviour (Drury and Reicher, 2000; Reicher, 1996; Stott and Reicher, 1998a,b). Drawing from social identity (Tajfel and Turner, 1979) and self categorisation (Turner et al., 1987) theories, ESIM proposes that collective behaviour during a crowd event is made possible through the shared salience among crowd participants of a common and socially determined identity (Reicher, 1982, 1984, 1987). Both the form and content of a crowd’s identity is viewed as context dependent and therefore can change as a function of the inter-group interactions that occur during a crowd event. Such processes were examined in detail by Stott and Drury (2000). Their study suggested that during the early stages of a demonstration against a new form of taxation participants’ collective identity was defined in terms of non-violence and differentiation from those understood to be seeking disorder. However, the police held a view of the crowd based upon “classic” theory. Therefore, a small sit-down protest during the demonstration led the police to see the crowd as being “whipped up” by a small band of extremists and therefore beginning an inevitable slide toward “disorder”. In order to try to restore control, the police then drove into the crowd using relatively undifferentiated force (e.g. baton charges). Such indiscriminately coercive tactics changed the inter-group context for the crowd. Consequently, demonstrators redefined their collective identity in terms of the illegitimacy of their inter-group relationships with the police. This in turn led to a widely shared understanding among demonstrators of the legitimacy of conflict with the police and shared psychological group membership with those prepared to confront them. The redefined psychological unity of demonstrators also led directly to a sense of collective empowerment which meant that conflict with the police was not just seen as legitimate but also possible social action. What is also illustrated by this study is that police public order tactics based on “classical” theoretical models of crowd behaviour can set in motion a cycle of escalating collective violence. According to the ESIM this is because police decision-making and tactics based on classic theory inadvertently created the social conditions for the escalation of crowd conflict to take place as a kind of self-fulfilling prophecy (Stott and Reicher, 1998a,b; see also Jefferson, 1990). Practical application More recently ESIM research and theory has been focused on exploring the implications of the ESIM for the policing of football crowds. Reicher et al. (2004) proposed a set of principles for the effective management of crowds. Within the UK, the Association of Chief Police Officers (ACPO) produces a Manual of Guidance for PIJPSM Keeping the Peace (Association of Chief Police Officers, 2003). This manual sets the 33,2 national guidelines for the policing of public order in England and Wales. The Manual was updated in 2003 to include these ESIM-derived principles (see also Reicher et al., 2007). For example, the Manual notes the importance in inter-group interaction by stating that “policing crowds is a dynamic and interactive process” and that “a trigger incident may be a result of actions by crowd participants, by the police, or as a result of 220 altered perceptions arising from the interactions between them”. Furthermore it rejects “classic” theory by recognising the meaningful nature of crowd action and asserting that “members of a crowd do not necessarily get carried away by ‘crowd hysteria’” (all p. 27). The relevant section of the ACPO manual concludes by recommending four principles for successful crowd management closely reflecting those set out by Reicher et al. (2004, p. 28). Subsequent research suggests that where these principles are reflected in policing policy and practice results have been productive. Perhaps the best example of this was the 2004 European Football Championships in Portugal (Euro2004) where the ESIM was used to inform the security policy for the tournament of the Polı´cia de Seguranc¸a Pu´blica (PSP), one of Portugal’s two main police forces. This process began with the development of a model of good practice in the policing of football matches with an international dimension (Adang and Stott, 2004; Stott and Adang, 2009). This model of good practice emphasised the importance of maintaining fans perceptions of the legitimacy of their collective relationship with the police. Moreover, where such legitimacy was perceived it was predicted that fans would tend to enforce the maintenance of non-violent norms (i.e. they would “self-police”). Thus, instead of confronting fans, officers in normal police uniform would patrol areas where crowds were gathered and interact with fans in a friendly manner. The intention would be to facilitate legitimate behaviours but also to monitor for and gather information on potential threats to public order posed by individuals or small groups. Such tactics would allow for the early identification of emerging tensions, which, if necessary, could be dealt with by larger squads of police. It was proposed by the theoretical model that this graded tactical approach would increase the likelihood that if police use of force was necessary, it would be differentiated and proportionate. The success of the tournament in terms of the absence of collective disorder among fans is now widely acknowledged in policy circles throughout Europe. Moreover, research on the policing of and crowd psychology among fans attending the tournament largely confirmed the ESIM derived pre-tournament predictions (Stott et al., 2007, Home Office, 2005). Subsequently, updates have been made to important areas of policy within Europe. The Police Cooperation Working Party of the Council of the European Union developed its recommendations[1] for policing football matches with an international dimension in line with the Euro 2004 model of good practice (Stott and Pearson, 2007). But at present there is not yet any systematic research to determine whether these principles, policies and theories are being incorporated into police training and practice in England and Wales. Current paper On average the police in the UK spend approximately £30m annually policing football. Yet incidents of “disorder” still occur at around 15 per cent of football matches within England and Wales, resulting in approximately 3,000 arrests per annum (NCIS, 2005). Moreover, the Greater Manchester police, one of England’s largest police forces, Policing football estimate that around three quarters of its entire public order policing budget is spent crowds on policing football. If measured in these terms, then the management of football crowds is by far the largest and financially relevant public order issue confronting the police within the UK. Stott et al. (2008) have argued that during football crowd events the autonomy of police forward commanders to deal with emergent risks is central to the effective 221 policing of crowds. Yet they also argue that this “tactical autonomy” must be set against the extent to which those commanders are adequately trained to understand and deal effectively with the dynamics of crowds. However, what little research on police public order training that does exist focuses exclusively upon the command decisions of senior officers and their concerns about the political accountability of their decisions (Cronin and Reicher, 2006; cf. Waddington, 1993, 1994). This paucity in the literature coincides both with UK Home Office recognition of a need for enhanced training at all levels of the police service (Home Office, 2001) and a recent international agreement to establish a pan-European training course for police commanders who take responsibility for the policing of football matches in their host nation[2]. However, there is currently insufficient data available in the literature to allow a clear understanding of the current state of public order training. Therefore questions remain about what police training will need to deliver in order to promote an adequate understanding of crowd dynamics and their relationship to public order policing. The current study therefore begins address these limitations and questions by examining the nature of police training for the policing of football crowds within and across England and Wales. More specifically, the current study focuses upon a series of training sessions involving hypothetical exercises of football crowd policing operations. Using these “real time” events we will: . examine in situ what form of crowd theory is being applied within football-related public order training; . explore what relationship, if any, this has to the perception of football crowd dynamics and crowd psychology held by officers on the courses; . analyse the understanding of strategies and tactics that flow from such theoretical models of the crowd; and . explore the implications of this theoretical understanding for police training in England and Wales as this relates to the effective policing of football crowds in practice. Method Background There is currently no training within England and Wales specifically oriented to the policing of public order at football matches. Instead, training for policing football is provided within a framework of courses that deal with the policing of public order more generally. These public order training courses are provided across the 43 separate police forces in the England and Wales. These courses span all police ranks from command to constable and also include non-rank specific specialised roles such as public order tactical advisors[3]. PIJPSM All courses follow a curriculum created by the National Policing Improvement 33,2 Agency (NPIA), which is a British non-departmental public body established to support police by providing expertise in a variety of different areas, including public order. Course materials are then provided by the NPIA to regional accredited police trainers, who in turn then deliver the courses that conform to this national framework. Given this national framework courses were sampled from different regional centres. 222 The courses sampled covered the range of training available nationally and therefore are representative of the whole range of public order training. Furthermore, the participants chosen for the study were all those officers who were enrolled on the courses sampled. Agreement for their participation in the study was granted prior to the commencement of the courses by the course instructors and formed part of the requirement for the successful completion of the training. Access to the regional centres was established through the ACPO portfolio holder for policing football[4], from existing contacts whilst the first author was Football Liaison Officer (FLO) and through networks established while attending the courses. The seven courses attended were as follows: . Two Initial Public Order Commander (IPOC) courses: the IPOC course is a one-week course for forward commanders (known as Bronze[5] commanders within the UK public order policing model). . One Major Sporting Events (MSE) course: a three-day course for police commanders who will be responsible for policing inside sports stadia for all major sporting events. . One Tactical Advisors course: this is a week-long course for all officers seeking to become public order Tactical Advisors in England and Wales. . Three public order Common Minimum Standard (CMS) courses: these are public order training courses that all police officers must complete in order to be able to police public order events (such as football) or provide mutual aid to other police forces as part of a nationally standardised Police Support Unit (PSU[6]). Qualitative data collection and analysis The courses took two forms; firstly the IPOC, MSE and TAC Ad courses were classroom based. Students started the courses by discussing and familiarising themselves with the background material provided by the NPIA, after which knowledge tests would be conducted by the trainers. There were then “table top exercises”, which are hypothetical policing operations for various different types of public order situation. Each exercise could last up to a day. Students would be allocated to various roles (i.e. Senior Commander, Tactical Advisor, etc.) and then begin by developing strategy and tactics for each operation based on the information received from the trainers. Different scenarios would then be “fed into” the event by the trainers during the exercise via “paper feed” or radio (e.g. a group of hooligans had just arrived at the main railway station). Students then had to deal with the developing scenario using their strategic and tactical plan. Finally, at the end of each exercise and the course as a whole debriefs were held where students and trainers drew their conclusions. In contrast, the CMS courses had no classroom-based component as they took place in an outdoor centre or public order training “village”. Here the students had to take part in physical activities such as a test for physical fitness, working together as a PSU, baton charging, shield formations, and working under fire from petrol bombs Policing football and “bricks”. Here students were evaluated on their speed, efficiency and competency crowds in performing the tactics. The first author obtained access to all course materials from the course trainers, prior to or during the respective training courses. Each training course was attended by the first author and observations made on what was taught and how. During each observation and with the consent of participants the comments of instructors and 223 general discussions with other students during each course were recorded on a digital recorder or by written field notes. All of the discussions and table-top exercises that were recorded were transcribed at the earliest opportunity. Observations cumulatively totalled 25 days, during which time approximately 50 hours of discussion were digitally recorded and 15 A4 pages of field notes made. The analysis began by transcribing all recordings from the courses and combining them with all the field notes so that a comprehensive data corpus was created. All material that explicitly related to policing football crowds was then identified. Initially this data was organised into two broad categories, students’ references to crowd theory and their references to strategy and tactics. The material in each category was then re-read and the data analysed using a constructionist revision of grounded theory (Pidgeon, 1996; Glaser and Strauss, 1967). From this rereading sub-themes were developed which then form the overall structure of the analysis presented here. The data included in the analysis section of this paper was selected for its representativeness in terms of indicating the wider body of data within the thematic category. Quantitative data collection and analysis A survey questionnaire developed by Drury et al. (2003) was distributed to officers after each training course had been completed. A total of 65 questionnaires were returned. For the purposes of this study, measures were taken of police understanding of: crowd composition; the effects of crowds on participants; explanations of crowd violence; and tactical response to crowd violence. For all of these themes students were given a series of statements to which they responded on a six-point scale ranging from 1 (strongly disagree) to 6 (strongly agree). The rationale for the questionnaire was to test the representativeness and reliability of the qualitative analysis. Rather than being overly concerned with the scalar representation of overall attitudes the study therefore conducted a simple analysis on the questionnaire variables relevant to officers’ understanding of football crowds and the tactics best suited to policing them. In line with Stott (2003) we therefore report on the significance levels of the difference of the means from the midpoint (3.5) on these specific items using a one sample t-test. Analysis Training crowd theory The first issue arising from the analysis was the almost total absence of any explicit reference to crowd theory in any aspect of any of the courses. In fact the only such reference was contained within a chapter entitled “Crowd psychology” within the National Centre for Policing Excellence (NCPE)[7] (IPOC) Distance Learning Handbook. The opening to the chapter states: PIJPSM Although many psychologists have written numerous theories concerning disruptive crowd behaviour, this information sheet is an attempt to provide some useful background 33,2 information and some ideas that are pertinent to operational considerations. Everett D Martin; a crowd psychologist of the 1920s wrote in his book “Behaviour of Crowds”: – “A crowd is a device for indulging ourselves in a kind of temporary insanity by all going crazy together”. All psychologists seem to agree, that membership of a crowd results in a lessening 224 of an individuals ability to think rationally, whilst at the same time his/her more primitive impulses are elicited in a harmonious fashion with the emerging impulses of all other crowd members. The result being the establishment of a collective mind (NCPE Handbook, p. 24). The extract is interesting in three respects. First is its clear and unequivocal reference to the theoretical idea that crowds are an arena in which ordinary people lose their individual rationality such that their behaviour is governed by emotional impulses. Second is its assertion that such a theoretical position is undisputed within the scientific literature. Third, its affirmation that such a theoretical model has direct relevance to police operational practice. The idea of the inherent irrationality of the crowd was also invariantly reflected in discussions between trainers and students. For example, in the following extract taken from a debate between a training Inspector and his students the Inspector asserts (through implicit and inaccurate reference to Solomon Asch’s famous conformity paradigm; Asch, 1951) that crowds invariantly lead to a loss of individuality and rationality, which in turn leaves them open to casual social influence: Remember that once a crowd assembles they lose their individuality. How many times have you heard a person at a football match say “I don’t know what came over me”. Apparently 75 per cent of the population can be influenced by the behaviour of others and can be made to conform (IPOC Course A, Trainer N). Students themselves then recapitulated this theoretical model of the crowd in their own discussions, even when trainers were not directly present. What was evident is that implicit within such discussion was the idea of the crowd’s irrationality. As such, it was universally and explicitly accepted that “hooligans” (i.e. violent individuals or groups) could always become the primary factor governing the behaviour of the crowd as a whole. For example, during one exercise a student acting as a Commander for the hypothetical policing operation specifically briefed the other students about the impact such minorities can have: If protagonists or ringleaders immerse themselves into the middle of a crowd, they can be the sort of brain or engine of that crowd driving them on (MSE Course A, Student C). Furthermore, where hooligans were understood to be present it was always the case that the crowd as a whole was described as a threat to public order and that there was therefore a necessary requirement to police all fans that were present. Thus discussions of crowd dynamics sat neatly alongside judgements concerning how to react to large crowds containing suspected “hooligans”: Well we worked on the basis that within that group of 300 [ordinary fans] you probably had your hardcore nominals[8] and crowd dynamics say that the behaviour of these nominals can affect the behaviour of the others there (MSE Course A, Student D). The relationship between theory and practice Policing football As the above extract implies the important feature of conversation about crowd crowds dynamics is that they invariantly served as a rationale and justification for students’ proposed strategy and tactics. In particular there was a focus on arresting and containing hooligans as the primary preventative measure. For example, in discussion during another table-top exercise a student, also acting as the Senior Commander, defined his views to his subordinates on the strategic focus of their hypothetical 225 football policing operation: The arrest policy for today is to identify ringleaders and protagonists at an early stage and arrest them. This will then impact on crowd dynamics and assist in early resolution and affects the ability of the groups to organise disorder (IPOC Course A, Student G). In other words their theory of crowd dynamics led directly to a strategic focus on arresting “ringleaders” because students understood that this removes the overall threat to public order posed by the crowd. This of course raised important questions for the students about how this targeting of hooligans could best be achieved. When discussing solutions students emphasised the importance of police spotters[9] to identify hooligans and feed this “intelligence” into their operation: You have intelligence about the types of clothing and uniform that the people we are interested in are going to be wearing so the spotters are automatically going to be looking for those persons (TAC AD Course A, Student T). Having dealt with the identification of hooligans, discussion turned toward how these individuals and groups could be contained, disrupted or otherwise removed from the crowd through the use of force. A student tactical advisor discussed with the Commander how they could use spotters to coordinate PSUs, allowing them to separate hooligans from the crowd at the end of their hypothetical match: Outside we have a PSU and a spotter and they are going to pull out the risk element as they see them coming out and corral them with the PSU (TAC AD Course A, Student R). Despite this focus, during all exercises students acknowledged and accepted that for various reasons it is actually often not possible to even identify let alone separate hooligans from crowds. Moreover, in such circumstances it was described as perfectly appropriate to take action against the crowd. During one of the table-top exercises students were discussing the fact that a large crowd of away fans had arrived simultaneously at the railway station. Noting the perceived impossibility of isolating any hooligans, they chose instead to forcibly corral the entire contingent of fans and remove them from the area: If we don’t know who the goodies or baddies are we will have to clear them all (IPOC Course B, Student F). Another tactical response in such circumstances included highly visible displays by the police of their capability to use force, described as a means of deterring any hooligans thought to be present in the crowd: If they don’t have a lot of police officers there they will think well what’s going on, and they will go and play. Whereas if we show them that we are there we can take control of them (MSE A, Student C). PIJPSM Thus far from isolating those suspected of holding violent intent the police instead choose 33,2 tactics that actually involved the relatively undifferentiated use of force. Moreover, while attention was ubiquitously paid to the use and display of force, there was at the same time a relative absence of discussion about alternative approaches. Where such discussion did occur these alternatives were treated with some scepticism. For example, in discussion about a recent operation to deal with a crowd of Millwall fans, a trainer and student 226 debated the efficacy of an approach that focused upon facilitating the crowd: Student A: We had Millwall last year and obviously Millwall come with a reputation but we decided that perhaps if we treat them normally then they might respond to that. And they thought it was great, even the Millwall thugs thought how good it was that the police were looking after us and dealing with us as human beings rather than being rounded up with batons like they get everywhere else and it worked for us, it worked very well. Trainer B: Sometimes these novel approaches work very well. But what was your contingency if they had turned nasty? Student A: Batons [laughter]. Well we always have that to fall back on to don’t we! (MSE course, Student A, Instructor B.) Thus, despite an initial focus upon targeting hooligans, the less discriminating use of force was central to the tactical plans that students developed. At the same time there was very little discussion about the potentially negative impact that this undifferentiated use of force may have upon crowd dynamics. This ultimately culminated in the tactical position whereby the use of force would be the defining and central characteristic of tactical plans related to policing football crowds. Indeed, as one instructor asserted, should that not be the case then students could legitimately begin to question the credibility and competence of their Senior Command: Arrest, containment and dispersal are the three key tactical considerations and if you don’t see that bleeding through Silver’s[10] tactical plan as it drips down to Bronze level then you should be asking questions because if you haven’t got elements of the aforementioned then it could be that the plan isn’t quite up to it (IPOC A, Instructor B). The above extract also demonstrates how discussions about tactics were not just framed in terms of a shared theory of the crowd. The analysis also identified the importance of accountability in their decision-making. At times students expressed concerns that being unable or unwilling to exercise the use of force would leave them professionally vulnerable. For example, during the planning phase of one table-top exercise a group of commanders sought to justify their request to senior colleagues to utilise large numbers of officers where the intelligence suggested that they may not be needed: We thought, let’s be realistic, we looked at the intelligence and really it shows that realistically the risk groups from both teams actually get on quite well and the chances are that we are totally over-reacting and that actually we could have a police-free game, but we just don’t know and nobody is going to be brave enough to make that decision so let’s be really firm about it and do it properly and then we can back it all the way back down (MSE Course, Student E). Qualitative analysis The qualitative analysis identified a basic pattern in the data whereby public order training emphasised “classic” theory of the crowd. At the same time instructors and students discussions reflected an understanding of the crowd as irrational and Policing football therefore easily influenced by hooligans. This in turn was associated with a strategic crowds emphasis on the removal, containment or disruption of hooligans through the use of force lest they manifest their ability to “hijack” the crowd. When it was not possible to achieve this, then a strategic and tactical shift toward the use of force against crowds as a whole was evident. The pattern of responses by students to questionnaire items is representative of this 227 analysis. For example, students agreed that the majority of people in football crowds have peaceful intentions (M ¼ 5:385 ^ 0:722, tð65Þ ¼ 21:028, p , 0:001, d ¼ 2:60) but that even the most respectable and ordinary people can become irrational and violent when in a football crowd (M ¼ 4:185 ^ 1:401, tð65Þ ¼ 3:967, p , :001, d ¼ 0:48). They agreed that football crowds always contain a significant minority of people with violent aims (M ¼ 4:415 ^ 1:647, tð65Þ ¼ 4:479, p , 0:001, d ¼ 0:55) who are skilled at inciting crowd violence (M ¼ 4:123 ^ 1:352, tð65Þ ¼ 3:715, p , :001, d ¼ 0:46) for fun and excitement (M ¼ 4:738 ^ 1:107, tð65Þ ¼ 9:013, p , 0:001, d ¼ 1:11). Students agreed that once violence occurs in a football crowd, otherwise law-abiding decent people get affected by crowd dynamics such that they are drawn into the emerging disorder (M ¼ 3:861 ^ 1:285, tð65Þ ¼ 2:268, p , 0:027, d ¼ 0:28). Students also agreed that all football crowds must be strictly controlled in order to prevent widespread disorder (M ¼ 3:969 ^ 1:550, tð65Þ ¼ 2:439, p , 0:017, d ¼ 0:30). Further, they agreed that if a few members of a football crowd become violent it is important for the police to intervene against the crowd quickly and with force to ensure the violence does not escalate (M ¼ 4:107 ^ 1:469, tð65Þ ¼ 3:334, p , 0:001, d ¼ 0:41). Finally, given such agreement the students therefore disagreed with the statement that the police are often responsible for the eruption of violence because of their inflexible and indiscriminate response to violence by small numbers of crowd members (M ¼ 2:492 ^ 1:336, tð65Þ ¼ 6:081, p , 0:001, d ¼ 20:75). Discussion This paper had four central objectives. The first was to examine what model of crowd theory is being applied within public order police training in England and Wales. The second was to explore the potential impact such theory has upon police students’ views of the crowd. Third, to analyse the strategy and tactics that flow from this model of crowd dynamics. Finally, the fourth was to explore the implications of this analysis for theory and practice. Turning to the first objective it is evident that there is a relative absence of explicit theory in public order police training in England and Wales. In all of the written police training materials we collected there was only one explicit reference to a theory of the crowd. In line with classic theory this single reference points unequivocally to the idea that in the crowd individuals will collectively undergo a loss of rationality and be subject to contagion and impulsivity. This theoretical position is given added weight and credibility through the inaccurate assertion that this theoretical model of crowd dynamics and psychology is undisputed within the scientific literature. In terms of our second objective the analysis illustrates how this theoretical model then coincides with similar descriptions of the crowd by trainers and students. In particular, students and trainers asserted that the football crowd was an arena within which ordinary fans could lose rational control of their behaviour. Moreover, that PIJPSM football crowds invariantly attract violent minorities that are capable of hijacking the 33,2 crowd directing it toward violence. Thus, students agreed that football crowds were inherently dangerous since they could be exploited by “hooligans”. In other words, the very presence of those defined as hooligans would lead officers to perceive a threat to public order posed by the crowd as a whole. Third, this corresponds with specific strategic and tactical orientations. Strategically 228 the orientation was toward the use of force at two levels. Initially the focus would be on identifying and removing hooligans from the crowd. However, it was acknowledged that identifying and isolating these minorities was at the very least problematic. In such situations the strategic orientation would shift to using force against the crowd as a whole. Tactically, this was either in terms of corralling and containing crowds of fans arriving in particular locations or through obvious and indiscriminate displays of police capability to use force. Where strategic and tactical alternatives were discussed these were either dismissed or their success viewed as contingent on the use of force. Implications for theory Stott and Reicher (1998) and Drury et al. (2003) suggest that the police in England and Wales hold an “agitator” model of crowd dynamics. They also assert that such theoretical understanding leads to an increased likelihood that during emergent disorder police will use force in an undifferentiated manner and thus could inadvertently initiate the dynamics of widespread disorder as a kind of self-fulfilling prophesy. The current data and analysis is consistent with this analysis. The current study is able to move beyond these previous studies in a number of ways. First, data has previously been gathered post hoc. In this sense previous studies were always open to the criticism that the data reflected post hoc rationalisation where officers were seeking rhetorically to deflect blame and create justifications (Wetherell and Potter, 1989). In this study we have been able to examine police articulations of crowd theory in training situations and during hypothetical scenarios. This provided a unique opportunity to examine how police theoretical modelling of crowds related to their real time decision making about how to deal with them. Previous studies have argued that the agitator model leads the police to view the crowd as heterogeneous in composition but also as homogonously dangerous. This ideological position then combines with tactical constraints to lead police to treat the crowd uniformly in situations where disorder is expected (Stott and Reicher, 1998a). The current study supports and develops this contention. It highlights how the entire focus of police training is on the control and disruption of suspected ringleaders. Our analysis suggests that this is not simply because they are perceived as an inherent threat to public order, in and of themselves, but also because of their perceived ability to influence the irrational crowd. However, this study also suggests that this places the police in a problematic position, as for various reasons, it is often difficult to identify and isolate hooligans (Stott et al., 2008). In this situation then the police tend to fall back upon the threat of or use of force against the crowd as a whole. This study suggests that they do so because the agitator model quite reasonably leads them to assert that the very presence of the minority renders the crowd dangerous and in need of strict control. Such a view of the crowd also leads to a relative lack of reflexivity about the potentially negative impact that such indiscriminate tactics can and do have upon ordinary fans or crowd dynamics (e.g. Stott et al., 2001, 2006).  Second, what this study suggests is that while there is little explicit reference to Policing football crowd theory, implicitly the “classic” model of crowds permeates almost every aspect crowds of public order training throughout England and Wales. Moreover, the course material provided by the NPIA rather than developing a critical assessment of classic theory reinforces the model by presenting it as unproblematic fact. It is no surprise then that those police officers who deliver the training courses therefore consistently assert the model and students consensually articulate an understanding of their role and the 229 effectiveness of their tactics in its terms. At no point was the factual accuracy of this theoretical model discussed or challenged. In effect the “classic” model of crowds is institutionalised within public order police training within England and Wales. This research also supports contentions that policing tactics are not merely a matter of police understanding of the crowd. Cronin and Reicher (2006) and Waddington (1993, 1994) have argued that both the setting of strategy, the use of tactics and decision making during public order policing operations are affected by the expectations and demands of those who scrutinise them. In other words, accountability to both internal and external audiences introduces a dynamic to policing operations that must be understood and prepared for. Yet our study demonstrates that training does little if anything to formally address this issue. Implications for public order policing Martin (1920) acknowledges that Le Bon provided a basis for the development of this theory (Reicher et al., 2007). This stands in stark and obvious contrast to the ACPO Manual of Guidance for Keeping the Peace (Association of Chief Police Officers, 2003) which makes explicit reference to the fact that people in crowds do not necessarily get carried away by “crowd hysteria” (p. 27). The Manual also asserts that policing of crowds is a dynamic and interactive process and that therefore conflict can emerge not just from crowd participants but also from the dynamics of interaction between the crowd and the police. It is evident from our research that significant work is required in order to integrate ACPO policy more formally and explicitly into the NIPA curriculum. The Manual recommends four key crowd management principles. It states that these principles should be considered during the planning, briefing and deployment stages of any policing operation involving the management of crowds. For example, ACPO guidelines state that “intelligence” gathering should be used to assist understanding of different group’s intentions and cultures (p. 28). The rationale for this is that understanding the cultures and intentions of those who make up a crowd is vitally important so that perceptions of the legitimacy of police action among them can be predicted, established and maintained. However, the current analysis suggests that in training for football emphasis is exclusively upon intelligence as the location and identification of suspected hooligans. The identity and cultural perspectives of ordinary fans is largely ignored. These ACPO guidelines further suggest that officers should not view everybody in the crowd as the same, particularly when disorder begins to occur. However, our analysis suggests that when police suspect that hooligans are present they would do precisely the opposite and shift strategically and tactically to the indiscriminate use of force. While the ACPO guidance recognises that such indiscriminate tactics may sometimes be unavoidable, it recognises the need to communicate with the crowd in order to ensure the maintenance of perceived legitimacy. Yet while the Manual stipulates the importance of communication as a PIJPSM principle of crowd management, this is something that seems largely ignored in 33,2 training. Finally, the ACPO Manual identifies facilitation as a key principle, highlighting the importance of allowing crowds to pursue lawful aims whilst at the same time dealing with groups acting unlawfully. Indeed, research has demonstrated that facilitation can coincide with a “self-policing” culture in crowds, whereby those perceived as hooligans by ordinary fans are marginalised physically and 230 psychologically (Stott and Pearson, 2007). Yet this analysis suggests that in training the focus is exclusively upon the forceful control and disruption of groups suspected of intending to act unlawfully while little time is spent discussing means through which the legitimate behaviour of all fans can be facilitated or undermined by police action. Limitations of current study The current analysis is based on a data set gained from observations conducted at a relatively small number of police public order training events. Moreover, two courses do currently exist that are not included within our observations – the Advanced Public Order Commanders Course (APOC) and the Football Intelligence Officers Course (FIO). However, both of these are recently new developments and were either unavailable to us or simply not being run during the five-month period within which the observations took place. None the less, the courses that were observed cover the majority of the public order training courses available to police officers in England and Wales, and it is therefore unlikely that the two courses absent from the study would dramatically alter the themes identified in the analysis. There is also the possibility that courses may be delivered differently or the content altered slightly in different regional training centres. However, all the courses are nationally accredited and designed to support the development of national minimum standards and mutual aid whereby officers from different police forces are trained to a common national standard so they can work together during a public order operation. As such there is a need for consistency in both content and delivery. Indeed the courses have to be broadly representative of the NPIA curriculum that must be delivered by the regional training centres. It is therefore unlikely that there will be large scale variations in the content and delivery of these courses in other regional centres and it is therefore likely that our sample will be representative of public order training nationally. Another potential criticism is our data draws from discussions of largely hypothetical scenarios and as such does not reflect what actually goes on in the operational context. However, the scenarios used during training were actually quite realistic since they were often based on actual events. Exercises were also conducted in real time and under stressful conditions. Indeed, these scenarios and exercises form the basis of training precisely because they are understood to be a close approximation of “real life”. None the less such criticism is well founded and it remains to be demonstrated how police understandings relate in real time to actual policing operations. This study provides further support for the contention that there is a systematic relationship between theoretical models of the crowd and subsequent police responses to them. However, this relationship between theory and practice has been tested here only in the context of the UK in response to the policing of football crowds. None the less our contention is that our data and analysis is relevant to understanding police responses to crowds more generally and to police training in other nation states. However, it remains to be shown the extent to which classical theory still forms the Policing football bedrock of police understanding internationally and as such there is a pressing need crowds for further cross national research examining police training; particularly given the pan-European agreement to develop an international police training programme in response to the policing of football crowds with an international dimension. Conclusion 231 This study has demonstrated that there is, certainly within England and Wales, a lack of focus upon modern crowd theory within public order training. The national training framework within England and Wales therefore does not reflect any significant development in scientific thinking about crowd dynamics since the early part of the twentieth century. But it is also evident that there is a lack of emphasis given to theory and empirical research in general. In this sense the study supports White’s (2006) claims that police training is “proceeding down an intellectual cul-de-sac. It has failed to engage in a debate over ‘ends’, having substituted for this a simplistic discussion over ‘means’” (p. 389). As we have pointed out there is now an extensive scientific literature on the effective management and policing of public order. In particular this literature contains an empirically led debate that contrasts the efficacy of two different approaches referred to as “escalated force” versus “negotiated management” (for an overview see Waddington, 2007); two approaches which rely differentially upon the use of force. It seems somewhat ironic that given the centrality of the use of force in public order training, the existence of this debate and its related evidence was never even referred to in the entire context of our observations. Such absence of theory also seems to us as somewhat problematic given the history of football and previous research which suggests the importance of police competencies in the effective management of football crowd dynamics (Stott et al., 2008). There can be little if any dispute that policing football is inherently difficult, often quite stressful and a highly challenging environment within which officers must make effective decisions. Indeed, ineffective decision-making can be catastrophic (e.g. Taylor, 1989). We would contend that one feature of what we have shown here is that the dynamics of such decisions are at the very least influenced by officers’ knowledge of crowd dynamics. It is our contention therefore that public order police training in England and Wales (and indeed internationally) must be updated to include the latest scientific knowledge so that officers can make these highly demanding decisions in the most informed way possible. Indeed, the evidence suggests very strongly that where such knowledge is made available to senior commanders there can be significant reductions on the overall levels of football related disorder (Stott et al., 2007; Stott and Pearson, 2007). Our research also suggests that the current status of public order training in England and Wales does not entirely reflect the policy set out in the ACPO Manual of Guidance, nor to that in the EU Handbook. In particular the updates made to public order policy and crowd psychology are not being supported or incorporated into training with regard to strategy and tactics. We would argue this is important because training is the primary means through which these developments in policy can and will be cascaded down into practice. In other words, it is not just a matter of policy but of root and branch reform in public order policing. We would contend that integral to such reform is a view of police practice, research, policy and training as equal partners. PIJPSM Indeed, it was just such an initiative that has led to important developments in police 33,2 training within Scotland (Scottish Police Training College, 2007) and Portugal (Adang and Stott, 2004; Stott and Pearson, 2007). Indeed, it is our view that this type of partnership is vital if society is to properly address the long term reduction of football-related “public disorder” and the massive and ongoing costs of policing football across the European Union. 232 Notes 1. Measures to Prevent and Control Violence and Disturbances in Connection with Football Matches with an International Dimension (2006), referred to more commonly as the EU Handbook. 2. Agreement was reached to fund such training at a ‘High Level Conference Towards and EU Strategy Against Violence in Sport’, organised collaboratively by the European Commission, the Portuguese Presidency of the Council, the European Parliament and UEFA in Brussels, 28-29 November 2007. 3. Officers whose role is to provide a police commander with strategic and tactical advice both before and during public order operations. 4. ACPO elects individual representatives to take responsibility for policy within specific areas. These individuals are referred to as Portfolio Holders. At the time of the study he football Portfolio Holder was ACC Steven Thomas, then of Greater Manchester Police. 5. During a public order policing operation, Bronze commanders are charged with taking the primary responsibility in implementing appropriate tactical responses within their geographical or functional area of responsibility 6. PSUs are essentially a paramilitary-style police unit generally composed of 18 Police Constables, three Sergeants and an Inspector. They are trained to a National Minimum Standard specifically to deal with incidents of public disorder and to allow for mutual aid across different police forces. 7. Now part of the NPIA. 8. “Nominals” is a term used by the police to refer to anyone fitting the profile of “hooligans” 9. Police officers whose role is to be able to identify known trouble makers. 10. The role of the Silver Commander is to develop and co-ordinate the tactical plan. It is usually Silver who is expected to be the Senior Operational Commander on the day of an event, brief the senior command team and take an overview of the implementation of and changes to the tactical plan throughout the operation. References Adang, O.M.J. and Stott, C.J. (2004), “Preparing for Euro 2004: policing international football matches in Portugal”, report for the Portuguese Public Security Police, available at: www. liv.ac.uk/psychology/staff/CStott/Preparing_for_Euro_2004.doc Allport, F. (1924), Social Psychology, Houghton Mifflin, Boston, MA. Asch, S.E. (1951), “Effects of group pressure on the modification and distortion of judgment”, in Guetzkow, H. (Ed.), Groups, Leadership and Men, Carnegie Press, Pittsburgh, PA. Association of Chief Police Officers (2003), Manual of Guidance for Keeping the Peace, Association of Chief Police Officers, London. Cronin, P. and Reicher, S. (2006), “A study of the factors that influence how senior officers police Policing football crowd events: on SIDE outside the laboratory”, British Journal of Social Psychology, Vol. 45, pp. 175-96. crowds Drury, J. and Reicher, S. (2000), “Collective action and psychological change: the emergence of new social identities”, British Journal of Social Psychology, Vol. 39, pp. 579-604. Drury, J., Stott, C.J. and Farsides, T. (2003), “The role of police perceptions and practices in the development of public disorder”, Journal of Applied Social Psychology, Vol. 33 No. 7, 233 pp. 1480-500. Glaser, B.G. and Strauss, A.L. (1967), The Discovery of Grounded Theory: Strategies for Qualitative Research, Aldine Press, New York, NY. Hoggett, J. and Stott, C.J. (n.d.), “The role of crowd theory in determining the use of force in public order policing”, Policing and Society (forthcoming). Home Office (2001), “Policing a new century: a blueprint for reform”, available at: www. policereform.gov.uk\docs\policingsurvey (accessed 12 April 2008). Home Office (2005), “Impact of measures introduced by the Football (Disorder) Act 2000”, Report to Parliament, UK Home Office, available at: www.homeoffice.gov.uk/documents/Football- Disorder-2006.pdf?view¼Binary (accessed 12 April 2008). Jefferson, T. (1990), The Case against Paramilitary Policing, Open University Press, Buckingham. Le Bon, G. (1895), The Crowd: A Study of the Popular Mind, Ernest Benn, London. McPhail, C. (1991), The Myth of the Madding Crowd: Social Institutions and Social Change, Aldine de Gruyter, New York, NY. Martin, E.D. (1920), The Behaviour of Crowds, Ernest Benn, London. Pidgeon, N. (1996), “Grounded theory: theoretical background”, in Richardson, J.T. (Ed.), Handbook of Qualitative Research Methods for Psychology and the Social Sciences, BPS Books, Leicester. Reicher, S. (1982), “The determination of collective behaviour”, in Tajfel, H. (Ed.), Social Identity and Intergroup Relations, Cambridge University Press, Cambridge. Reicher, S. (1984), “St Paul’s: a study on the limits of crowd behaviour”, European Journal of Social Psychology, Vol. 14, pp. 1-21. Reicher, S. (1987), “Crowd behaviour as social action”, in Turner, J.C., Hogg, M.A., Oakes, P.J., Reicher, S.D. and Wetherell, M.A. (Eds), Rediscovering the Social Group: A Self-categorization Theory, Blackwell, Oxford. Reicher, S.D. (1996), “‘The Battle of Westminster’: developing the social identity model of crowd behaviour in order to explain the initiation and development of collective conflict”, European Journal of Social Psychology, Vol. 26, pp. 115-34. Reicher, S., Stott, C.J., Cronin, P. and Adang, O. (2004), “An integrated approach to crowd psychology and public order policing”, Policing: An International Journal of Police Strategies & Management, Vol. 27 No. 4, pp. 558-72. Reicher, S., Stott, C.J., Drury, J., Adang, O., Cronin, P. and Livingstone, A. (2007), “Knowledge-based public order policing: principles and practice”, Policing: A Journal of Policy and Practice, Vol. 3, pp. 403-15. Scottish Police Training College (2007), Football Match Commander Training Course, Tulliallan Castle, Kincardine, 17-19 August. Stott, C. and Adang, O. (2009), Understanding and Managing Risk – Policing Football Matches with an International Dimension in the European Union, Bavnebanke Press, Slagelse. PIJPSM Stott, C. and Drury, J. (2000), “Crowds, context and identity: dynamic categorization processes in the ‘Poll Tax Riot’”, Human Relations, Vol. 53, pp. 247-73. 33,2 Stott, C. and Pearson, G. (2007), Football “Hooliganism”: Policing and the War on the “English Disease”, Pennant Books, London. Stott, C.J. and Reicher, S.D. (1998a), “Crowd action as inter-group process: introducing the police perspective”, European Journal of Social Psychology, Vol. 26, pp. 509-29. 234 Stott, C.J. and Reicher, S.D. (1998b), “How conflict escalates: the inter-group dynamics of collective football crowd ‘violence’”, Sociology, Vol. 32, pp. 353-77. Stott, C., Hutchison, P. and Drury, J. (2001), “‘Hooligans’ abroad? Intergroup dynamics, social identity and participation in collective ‘disorder’ at the 1998 World Cup Finals”, British Journal of Social Psychology, Vol. 40, pp. 359-84. Stott, C., Adang, O., Livingstone, A. and Schreiber, M. (2007), “Variability in the collective behaviour of England fans at Euro 2004: ‘hooliganism’, public order policing and social change”, European Journal of Social Psychology, Vol. 37 No. 1, pp. 75-100. Stott, C., Adang, O., Livingstone, A. and Schreiber, M. (2008), “Policing crowd dynamics and public order at Euro 2004”, Psychology, Public Policy and Law. Tajfel, H. and Turner, J.C. (1979), “An integrative theory of intergroup relations”, in Worchel, S. and Austin, W.G. (Eds), Psychology of Intergroup Relations, Brooks-Cole, Monterey, CA. Taylor, Lord Justice (1989), The Hillsborough Stadium Disaster, Interim Report, HMSO, London. Turner, J.C., Hogg, M.A., Oakes, P.J., Reicher, S.D. and Wetherell, M.S. (1987), Rediscovering the Social Group: A Self-categorization Theory, Blackwell, Oxford. Turner, R.H. and Killian, L.M. (1987), Collective Behavior, 3rd ed., Prentice-Hall, Englewood Cliffs, NJ. Waddington, D. (2007), Policing Public Disorder: Theory and Practice, Willan, Cullompton. Waddington, P.A.J. (1993), “Dying in a ditch: the use of police powers in public order”, International Journal of the Sociology of Law, Vol. 21, pp. 335-53. Waddington, P.A.J. (1994), Liberty and Order, Public Order Policing in a Capital City, UCL Press, London. Wetherell, M. and Potter, J. (1989), “Narrative characters and accounting for violence”, in Shotter, J. and Gergen, K.J. (Eds), Texts of Identity, Sage Publications, London. White, D. (2006), “A conceptual analysis of the hidden curriculum of police training in England and Wales”, Policing and Society, Vol. 14 No. 4, pp. 386-404. Further reading Braun, V. and Clarke, V. (2006), “Using thematic analysis in psychology”, Qualitative Research in Psychology, Vol. 3, pp. 77-101. Independent Football Commission (2008), “Report on Euro 2004”, available at: www.theifc.co.uk (accessed 12 April 2008). Stott, C., Livingstone, A. and Hoggett, J. (2008), “Policing football crowds in England and Wales: a model of good practice?”, Policing and Society, Vol. 18 No. 3, pp. 1-24. About the authors James Hoggett is a PhD student at the University of Liverpool, and is funded by the Economic and Social Research Council. His research interests relate to exploring the relationships between police understandings of crowd dynamics, the manifestation of these understandings in certain types of strategic and tactical responses to crowd policing, and the outcomes of these tactics on crowd behaviour. He holds a BScEcon (Sociology) from the University of Wales Cardiff and a Policing football BSc (Psychology) from the Open University, as well as an MSc (Investigative Psychology) with distinction from the University of Liverpool. In addition, he was employed by the Avon and crowds Somerset Constabulary as a Football Liaison Officer between 2003 and 2005, with primary responsibility for Yeovil Town Football Club. James Hoggett is the corresponding author and can be contacted at: j.hoggett@liverpool.ac.uk Clifford Stott is a Senior Lecturer at the University of Liverpool. He is currently involved in an ongoing series of primarily qualitative studies examining the underlying causes and psychology 235 of “riots”. In particular he is interested in the relationship between the dynamics of inter-group interaction, social identity processes and the development of violent norms during crowd events. He has worked closely with Professor Steve Reicher (St Andrews) and Dr John Drury (Sussex) in developing the Elaborated Social Identity Model of crowd behaviour. He has done extensive work looking at the extent to which this model can account for acts of “hooliganism” involving English fans travelling to international tournaments. He has also been involved in Home Office working groups into football-related disorder and is currently involved in developing a major European study of public order policing. To purchase reprints of this article please e-mail: reprints@emeraldinsight.com Or visit our web site for further details: www.emeraldinsight.com/reprints