Abstract: Crowd violence has interested researchers in social psychology for many years and is an important issue for sports psychology (STOTT, ADANG, LIVINGSTONE & SCHREIBER, 2007; STOTT, HUTCHINSON & DRURY, 2001; RUSSELL, 2004; MUSTONEN, ARMS & RUSSELL, 1996). Riots in crowds have been explained from different theoretical perspectives (HYLANDER, 2008), such as individual differences, de-individuation (PRENTICE-DUNN & ROGERS, 1989), group interaction (DRURY & REICHER, 2000), history (GUTTMAN, 1986, 1998) and cultural perspectives (CRABBE, 2003). In this study, a social psychology model focusing on group interaction, the Aggravation and Mitigation (AM) model (GUVÅ & HYLANDER, 2008; GRANSTRÖM, 2008; GRANSTRÖM & ROSANDER, 2008), is used as a means of analysis. This article applies the AM model to a sporting event to identify if and how peacemaking processes can be detected. Furthermore, the intention is to discern and illuminate organizational strategies that maybe linked to peacemaking processes. The main results indicate that when arrangements are based on (a) "festival-making," (b) arrangements for basic needs and recognizable order and (c) the creation of a superordinate identity, then the outcome of mass events may turn out peaceful, which is also in line with the AM model.

Key words: mass event; social psychology; large groups; riots; grounded theory; football; Aggravation and Mitigation (AM) model

The Aggravation and Mitigation (AM) model for describing and explaining dynamics in crowds is a substantive theory that emerged from classic grounded theory analyses (GLASER & STRAUSS, 1967; GLASER, 1978), based on a series of qualitative studies of crowd events in Sweden between 2001 and 2006 (HYLANDER & GUVÅ, 2008; GUVÅ & HYLANDER, 2008; GRANSTRÖM & ROSANDER, 2008; GUVÅ, 2008). The database for the AM model consists of observational data—interviews with police officers and protesters before, during and after a series of major events—that is unusual in riot research. The types of events included political protests, "Reclaim the Streets" activities, and neo-Nazi marches with counter-protests (Table 1).

Grounded theory (GT) is sometimes criticized for being restricted to one substantive area and not linked to important theories in the field. This article attempts to use a theoretical model developed by GT in a wider area than where it was developed and to link it to theories in the field. The AM model is based on results from political demonstrations. The present study is the first time this theoretical model is applied to a sporting event. In this article, the model is first explained and linked to current research literature; it is then applied to a high-risk football match and elaborated in relation to this substantive field. [2]

Typically, the focus of research on crowd events has been on the crowd and its members, but in recent years, the police and police strategies in connection with such crowd events as political protests and sports events have also been studied (DRURY, STOTT & FARSIDES, 2003). Less emphasis has been placed on the organization of the event itself; therefore, this is a field in need of research. Additionally, most crowd research has studied the occurrence and escalation of violence, while only a few studies have focused on the prevention of violence (STOTT et al., 2007) or on peacemakers instead of troublemakers (RUSSELL & MUSTONEN, 1998). This study sets out to explore the presence and impact of peacemaking processes. [3]

The Aggravation and Peace Process model (APP, later labeled the AM model) (GUVÅ & HYLANDER, 2008; GRANSTRÖM, 2008; HYLANDER & GRANSTRÖM, 2010) supports and complements the Elaborated Social Identity model (ESIM) (REICHER, 1996, 1997; DRURY & REICHER, 1999, 2000; HOPKINS & REICHER, 1997; STOTT & REICHER, 1998; STOTT & DRURY, 2000). According to the ESIM, the out-group (the police) may understand the identity and actions of crowd members in ways that differ from crowd members' views (DRURY et al., 2003). For example, the police may define a protest as a threat to public order when the protesters see it as legitimate. The police thereby regard all the members of the crowd as potential threats, while the members may regard themselves as peaceful demonstrators. When there is an asymmetry of categorical representations between the in-group and the out-group, and the out-group (the police) has the power to instigate interventions—such as dispersion of a crowd—on the basis of their categorization, a social identity change may occur for the crowd members. Finding themselves in opposition to the police, they may even condone violent behavior, which they would not have otherwise done (DRURY & REICHER, 2000). According to the ESIM, the main explanations of crowd events are the normative rather than anti-normative structure of a crowd, social identity rather than de-individuation in the crowd, generated historically rather than generically, and intergroup rather than individual (DRURY et al., 2003; DRURY & REICHER, 2000; STOTT & REICHER, 1998). That implies that a crowd is not seen as a number of individual savages without norms, but rather as a social constellation governed by some kind of social norms and that crowd events are caused by interaction between groups. [4]

The AM model has developed independently of the ESIM but supports the ESIM's main conclusion. As the AM model is a grounded theory, it is a substantive theory that applies to the substantive area from where it was derived, i.e. Swedish political demonstrations and protests, and is not yet a formal theory (GLASER, 1978, 2006, 2007). However, it is a practical tool for structuring observations of crowd events, and has proved to be a convenient way of describing dynamic group interactions with the police (GRANSTRÖM, 2008; GUVÅ & HYLANDER, 2008). The AM model focuses specifically on mitigation processes, suggesting that peacemaking strategies are essential and are not merely the absence of aggravating strategies. [5]

The AM model illustrates the variation and dynamics in the escalation and de-escalation of violence in crowd events. In the events studied, two types of interactive crowd processes have occurred: promoting and escalating violence, or aggravating, and supporting peaceful activities and de-escalating violence, or mitigating. When aggravation dominates, conflicts escalate and members of the groups involved lose trust in each other, i.e. they do not believe the other party's peaceful intentions. Trust is seen as an intermediate factor that influences aggravation and peacemaking. If mitigation dominates, conflicts de-escalate and mutual trust increases. Furthermore, when members of one group perceive the other group as being dominated by distress and aggression, they are likely to use aggravation in their contact with that group. When members of one group perceive the other group as being peaceful, they are likely to respond in a mitigating way. Consequently, in events starting with mutual distrust, aggravation is more likely than mitigation whereas in a situation starting with mutual trust, mitigation is more likely to develop. Mass events without official permits from the police (a mutual distrust situation) are much more likely to turn into riots compared with mass events with legal permits and preceded by dialogues between the police and the crowd (a mutual trust situation) that tend to remain peaceful (GUVÅ & HYLANDER, 2008). A challenge for the police and for event participants are trust-distrust situations, where only one party trusts the intentions of the other group. These situations may escalate into violence or they may remain peaceful depending on the use of mitigation or aggravation strategies. Table 2 illustrates the main differences between the two processes.

Table 2: The AM model illustrating the main differences between aggravation and mitigation [6]

Variations in how events are organized, how groups treat members of other groups and how members of different groups categorize each other are crucial for the course of events. This is illustrated by the model and illuminates how some events turn into violence and others remain peaceful.

The extremes of these social processes—mutual treatment, organizing, and categorizing—are featured as aggravating, provoking, creating chaos, and negative stereotyping, or as mitigating, disarming, peaceful organizing, and peaceful categorizing (differentiation and positive stereotyping). [8]

Provoking actions are behaviors by the out-group that the in-group interprets as threatening or confronting. Provoking actions may be intentional or unintentional. For example, riot uniforms, seen as normal working clothes for the police, may provoke crowd members and signal distrust. Similarly, the police may regard black hooded jackets as potential masks for political protesters, for example, while they are just ordinary clothes to the protesters. Each group sees these signals as provocative depending on the degree of trust or distrust between the groups. [9]

The above assumption is in line with the ESIM conclusions that violence is an interactive phenomenon in crowd events (REICHER, 1996, 1997; DRURY & REICHER, 1999, 2000; HOPKINS & REICHER, 1997; STOTT & REICHER, 1998; STOTT & DRURY, 2000). Explanations from other traditions have pinpointed de-individuation in the crowd instead of group interaction (ZIMBARDO, 1969) or individual differences as the cause of violence and provoking behavior (RUSSELL & ARMS, 1998; MUSTONEN et al., 1996). VAN HIEL, HAUTMAN, CORNELIS and DE CLERCQ (2007) investigated provoking behavior such as the use of violence, and proposed that self-reported attitudes towards violence were the most marked predictor variables of both physical and verbal aggression. However, BARON and RICHARDSON (1994) assumed that situational factors may mask the effect of personality, which is in accordance with the analyses of the AM model. In very similar events, provocative behavior by protesters could be present or absent depending on the strategies of the police. There are several examples of the police recognizing an event without official permission as being legitimate and the event remained peaceful. Afterwards, the police explained that the peaceful outcome was possible because the provocateurs were absent from the scene. In contrary, the scientific analyses suggested that it was the positive interaction between the police (as a group) and the demonstrators (as a group) that prevented riots (and not the absence of provocateurs) (HYLANDER & GUVÅ, 2008). VAN HIEL, HAUTMAN, CORNELIS and CLERCQ (2007) concluded that social identity theory could better predict crowd violence than de-individuation theory. Thus, current research seems to support the importance of out-group provocations for crowd violence to occur. [10]

Generating chaos means intentional or unintentional acts of organization or disruption of the in-group that the out-group perceives as chaotic and disorganized and that creates a disrupting setting for the out-group. For example, a crowd event may be organized with performances, food stands, music, and dancing, but as long as the police see no formal organization or leader to negotiate with, they may regard the organization as chaotic. In addition, a very strict organization of the police with horses in line, barriers, and troop formations cause a chaotic context for marching protesters. Chaos is supported through the spreading of rumors, giving no or faulty information and not answering questions. The ESIM proposes that a change in social identity from a peaceful identity to opposition to the police is enhanced by an ambiguous situation (REICHER, 1997). [11]

Negative stereotyping has two main features: first, the out-group members are seen as prototypes for the group, all given the same attribution; second, these attributions are negatively loaded. The police said, for example, in connection with some of the protests in Sweden, that all demonstrators were troublemakers while protesters regarded the police as a troop of "stonefaces" (HYLANDER & GUVÅ, 2008). For sports events, it would be like the police regarding all fans as hooligans. [12]

The concept of stereotyping is used in several different ways in social psychology. The ESIM basically relies on social identity theory (SIT) and social categorization theory (SCT) (TURNER, 1999) where stereotyping (self-stereotyping or stereotyping of the out-group) is seen as a normal cognitive process in order to distinguish perceptions. It is sufficient to be categorized as part of a group to have an influence on the individuals to be categorized in the same way (TAJFEL, BILLIG, BUNDY & FLAMENT, 1971). The self-categorization theory can be described as social identity theory for groups. Self-categorization based on visible social identity leads to self-stereotyping and the de-personification of self-perception. Similarity on the basis of relevant dimensions within the group and differences compared with other groups is emphasized. [13]

According to the AM model, processes of provoking, creating chaos, and negative stereotyping interact and form a negative spiral of mutual violence between groups. In a chaotic situation, negative stereotyping is enhanced and forms the basis for provoking actions, which, in this situation, may be experienced by the out-group as much more provoking than intended. One important result from earlier studies is that in order for a conflict to escalate to violence and riots, all three aggravation processes have to be present (GRANSTRÖM, GUVÅ, HYLANDER & ROSANDER, 2009). It is not enough to have a chaotic situation with only a single provocative incident or negative stereotyping. It is when the processes are combined and interact that conflicts escalate between groups. [14]

Disarming is expressed through mutual understanding, kindness, or a positive surprise. Laying down weapons (literally and/or metaphorically) makes it easier for the other group to do so as well. When the police were dressed in plain uniforms and caps, worked in pairs instead of in troops, mingling, talking in a friendly way and even joking with protesters, the protesters responded in a friendly way (GRANSTRÖM & ROSANDER, 2008). Mutual disarming between groups is typically carried out by means of a continuous dialogue. [15]

RUSSELL (2004) reported several studies on good interaction between police and citizens; where humor was a salient aspect, it defused potential conflicts in crowd events. RUSSELL (2004), RUSSELL and MUSTONEN (1988), and RUSSELL, ARMS and MUSTONEN (1999) found that a quarter of the spectators at ice hockey matches were ready to intervene as peacemakers in a fight in the stands, 5% would applaud, and only 2% would join in. Similar figures are found in other studies. Evidence of the effectiveness of "low-profile policing" in connection with high-risk football tournaments was also provided by STOTT et al. (2007). Self-policing, which refers to the internal control exercised against anyone who the crowd or sectors of the crowd see as acting illegitimately, has been discussed as resulting from respectful treatment by the police (REICHER, STOTT, CRONIN & ADANG, 2004). Thus, disarming is a phenomenon reported in several studies. [16]

Peace organizing refers to any acts that hamper chaos within the in-group and, as a result, between groups. The police regarded protesters who followed their rules as being peacefully organized (GUVÅ & HYLANDER, 2008). They asked for permission to demonstrate, which meant planning, leaders, goals, and structure—a type of organization that the police recognize. This kind of organization may be labeled a "modern" organization, with an organizer who negotiates with police and takes responsibility. The group emphasizes factors of peaceful organizing similar to those emphasized by the police, such as permits, planned routes, specific organizers, demonstration guards and the banning of masks (GUVÅ & HYLANDER, 2008). In contrast, a "post-modern" organization is common at events without official permits. Singing, dancing, different kinds of music, slogans, and other spectacular performances may attract people and draw attention to peaceful actions and is a kind of "post-modern" organization of peaceful activities. [17]

Positive categorizing can have two different meanings. It may imply a positive stereotyping of the out-group: for example, the leader of a group of demonstrators will say through a megaphone: "The police are here to protect us," thus telling the demonstrators that all police officers are trustworthy. Or it can be a differentiation of the out-group in such a way that differences between individuals or subgroups in the out-group are expressed or acted on, as happens when the police only intervene directly with single individuals when they break the rules, and not with the group as a whole. Negative traits are also attributed to the in-group, which is a reason for peacemakers to become involved in self-policing (REICHER et al., 2004) and to take care of deviant members who may instigate a conflict with other groups. [18]

Disarming, peaceful organizing, and positive categorizing interact in such a way that when an event is peacefully organized, peaceful categorizing is more likely and then disarming behavior is much easier to accomplish. One major conclusion from the research is that the peaceful intentions of one group, publicly expressed, have an impact only if the same group also links them to peaceful strategies and actions (GUVÅ & HYLANDER, 2008). The absence of aggravation processes is not enough to produce a peaceful event: active mitigation strategies have to be applied. In conjunction with the EU summit meeting in 2001, large peaceful demonstrations took place simultaneously with riots (GUVÅ & HYLANDER, 2008). The peaceful intentions of the organizers were expressed very clearly. The city of Gothenburg welcomed the demonstrators and expressed a strong conviction that the organizers and the police could control events and assure a peaceful outcome of the mass event. The peaceful intentions of large groups of demonstrators were also clearly and publicly expressed. However, there were also small groups of demonstrators who expressed violent intentions. The results showed that in spite of the peaceful intentions of the organizers, there was little preparation for peaceful strategies. The preparations for the event were governed by the "strategies for violence" which then dominated the course of events. When the situation became chaotic, the strategies that had been prepared for war-like situations came into play and there were no mitigating strategies that could compete with the strategies for violence (GUVÅ & HYLANDER, 2008). [19]

The application of the AM model to different political demonstrations supports the following assumptions:

The AM model has been developed based on political demonstrations. It has not been applied to major high-risk sports events. Football and ice hockey account for a majority of violent incidents (RUSSELL, 2004 causing significant disturbances and destruction. Therefore, it is interesting to apply the AM model to a sports event and focus on the organizers' strategies for peace, i.e., to investigate what measures could be taken to promote peace. There is already clear evidence that police tactics influence the course of events in a crowd situation (REICHER, STOTT & ADANG, 2004). However, it remains to be explored how the organizers' arrangements influence mitigation in a crowd. [21]

The assumptions mentioned above are based on a number of observations of mass events (GRANSTRÖM, 2008). So far the AM model has not been applied to large football events thus the aim of this article is to continue to elaborate on the AM model by applying it to a different type of event—a sports event—in order to identify if mitigation (peacemaking processes) can be discovered in this context and if so how it may be conceptualized. A second aim is to illuminate what organizational strategies are linked to mitigation processes in the crowd. The focus is on arrangements, as they appear in observations and interviews with football supporters, but also includes the organizers' perspectives from interview data. [22]

The situation described in this article is based on activities in conjunction with a football match in Dortmund between Germany and Poland in the 2006 World Championship (14 June 2006). In view of historical conflicts, the police (HAU, 2006 and the media in Germany considered this match to be one of the high-risk events during the Championship. [23]

Sports riot researchers have typically used archival data, interviews with and self-reports from supporters, and participant observation (STOTT & REICHER, 1998). This study takes an ethnographic approach (HAMMERSLEY & ATKINSON, 1983): data are collected through participant observation and field interviews in the setting where the event occurs. This approach was recently proposed in order to extend and enhance our understanding of sports psychology (KRANE & BAIRD, 2005). But given that the focus is on a single event at a specific time, it is classified as a field study (HAMMERSLEY, 2006) and is part of a larger project involving a wide range of different methodological techniques and theoretical and practical issues (GRANSTRÖM, 2008). [24]

Several hours before and also during the match a number of "events" took place, such as supporters arriving at the railway station and meeting other supporter groups. There were crowds moving around in the city, police officers patrolling. Furthermore, during the match a large number of supporters watched the play on large video screens, others strolling around. Three pairs of observers were following the events at different places. These places were selected after consulting with police management. Two teams observed two places where large video screens were situated. The third team moved around in the city where the remaining fans were staying. [25]

The AM model, as presented above, is a substantive grounded theory that emerged by employing the constant comparison method (GLASER & STRAUSS, 1967): similarities and differences in observational and interview data on violent as well as peaceful political crowd events have been elaborated into concepts, patterns, and, finally, a substantive theory (GLASER, 1978). Theoretical sampling was used in order to secure a wide variation of different types of political protests and to let the emerging theory guide the samples, sampling methods, and focus of observation and inquiry. The database of the AM model is presented in Table 1. To elaborate this substantive theory into a formal theory with a broader range of applicability, it needs to be grounded in several sets of new and varied data (GLASER, 2007). The result of the present study is not yet a formal theory but is one step further to widening the implications of the theoretical model, particularly the mitigation process. In the present study observational data and interview data from one sports event have been coded, compared, and conceptualized. In this way several of the categories generated in the AM model have been elaborated. New subcategories have emerged altering or adding to the meaning of the concepts. When anomalies are found in the present set of data or when new data are not in line with prior hypotheses (GLASER, 1978, 2007), the theoretical model is elaborated to take into account and explain this new and incongruous information. Thus, the theoretical model is continuously being elaborated. New concepts are formed and relations between concepts are transformed. Existing and dominating theories in the field are compared to the substantive theory (AM). If other theories confirm the findings, this validates the grounded theory. However, suffice it to say, a grounded theory is only a set of hypotheses that earns its credibility by fitting to data, comprehension (work), and usefulness in explaining and governing practice for the actors in the field that is being researched (GLASER, 1978). [26]

To achieve a wide variation of data, participant observations and interviews were used. [27]

The main method for data collection was participant observations. Three pairs of senior researchers covered different strategic locations before, during, and after the football match. The observations covered ten hours and were semi-structured (see Appendix). The course of events, from the observers' point of view, was recorded as closely as possible in the form of narratives. Observations were either recorded directly by an audio recorder or as notes transcribed immediately afterward. In both cases, time tags were noted every five minutes, which made it possible to compare different observations from the same event. Observation locations were chosen according to information from the police authorities about local arrangements, the location of large video screens (Friedensplatz¹⁾ and Westfalenhalle²⁾), times for supporters' trains arriving at the railway station etc. In this way, the observers were able to be physically and temporally present in different strategic locations throughout the day. The observations were in the form of multi-point observations, which means that the six observers observed and recorded the development of incidents from various locations. Recording the time of all observations enabled descriptions of incidents from a variety of starting points to be made. All tapes and field notes were transcribed verbatim and totaled 60,000 words. The observations reported in this article were the result of the observers' consensual reports. Data were also checked against photos. [28]

Short interviews were carried out with various groups of fans, police officers, and other service agents during the course of events before and after the game and at halftime. Interviews took place in the streets and at the two locations where large video screens were placed. Attempts were made to conduct interviews with a varied selection of spectators (Polish and German, varying by age and gender). As the situation was rather unpredictable and noisy, the interviews were carried out whenever there was an opportunity to do so. A short interview guide was set up covering questions such as: "How do you like being here?" "What has it been like so far?" "Have you been in contact with the Polish/German fans?" "What is it like?" "What do you think of the way the police are acting?" "Have you seen or experienced any serious incidents?" "Do you expect it to stay calm or do you expect turmoil between the fans?" 33 interviews were taped and five interviews were documented only by notes. The interviews lasted from a couple of minutes to ten minutes. All interviewees were informed about the research project and gave their informed consent. [29]

An interview with organizers was conducted afterward and used as a means of extending the variation of data and saturating categories. The interview, a three-hour, face-to-face interview with the police commissioner, was conducted after the tournament, when the outcome had been analyzed. The whole interview has been published in a specific report (HAU, 2006). Below, the observations are linked to statements by the commissioner. [30]

Several photos were taken to complement the transcriptions of taped reports and interviews. These supported and illustrated six different narratives based on the transcripts written immediately after the events. Thus, recorded observations and interviews from six researchers, covering about ten hours of observations, constituted the data. [31]

All the narratives were read several times, and incidents and excerpts were coded according to the grounded theory method. New codes and concepts emerged. However, the main categories of the AM model still seemed relevant, although they were elaborated by new subcategories forming new aspects of the concepts. The authors independently conducted open coding on different parts of the data, while researchers independently did theoretical coding; discrepancies were then discussed and analyzed. Critical incidences, which might have turned violent according to hypotheses generated by the AM model, were specifically targeted and analyzed. New categories emerged that might be specific to this substantive area while other concepts of the AM model had a good fit with data. Finally, the findings were discussed in relation to earlier research and the Elaborated Social Identity model (ESIM) (REICHER, 1996, 1997). [32]

The city of Dortmund had prepared for an invasion of football supporters. The local authorities estimated that the number of visitors would exceed 100,000 fans, although the football stadium only holds 80,000. Certainly, this was a huge challenge for the local community and was considered to be a high-risk match by the local authorities (HAU, 2006. The risk of confrontations, given the historical conflicts between the two nations, was discussed in the media. The media also reported that Polish hooligans were known to be violent; true or not, this was the atmosphere the newspapers promoted. Aggression and violence would not be far away if nationalist supporters from the two countries perceived each other as real enemies and not just as combatants on the football pitch. According to the AM model, if no trust existed between the groups of supporters, aggravation would be likely, and if there was no trust on the part of the police, aggravation could dominate the interaction between the police and fans. The AM model also proposes that experiencing the other group as peacefully organized—disarmed and not easily stereotyped—enhances trust. Consequently, all arrangements facilitating such processes will be recognized as mitigating. [33]

The overall picture of the Dortmund football match was that it was peaceful. There were no fights between the Polish fans and the German fans. Interaction between the police and the fans was mostly peaceful except for two major incidents. At one point, Polish fans singing and shouting at the police were arrested, and there was turmoil outside the two areas where the police separated groups of German fans. These incidents did not, however, spread to other crowds. Since these incidents are reported and analyzed elsewhere by the research team (ROSANDER & GUVÅ, 2009), they are not included in this article. However, critical incidents that could have turned into aggravation according to the AM model are reported and analyzed. Below, a number of arrangements are analyzed in accordance with the AM model. [34]

The publicly expressed intention of the organizers was that it was "time to make friends," which signaled to the participants that they were all taking part in a joyful festival—a high-quality sporting event—and diminished the risk of confrontation between the fans. When football supporters (mainly Polish) arrived at the railway station, a number of police officers created a "human avenue," helping the fans find their way out of the railway building. The police officers were uniformed but without helmets or visors, i.e., they were disarmed. The supporters then entered a square where the houses were covered with enormous paintings giving an impression of a football stadium, with large stands filled with supporters around the arena. Supporters' yelling and cheering mixed with music sounded from loudspeakers. A large streamer with text in block letters said that Dortmund welcomed the guests. Thus, the first impression given to the visitors was that they were all welcome to the same magnificent party regardless of the team they supported. This can be seen as disarming in a metaphorical sense and was clearly expressed in the interviews.

Supporter: "The atmosphere is hot and up to this point I haven't seen anything, i.e., no hostility between Poland and Germany, as regards anyone being aggressive. It is a quite normal and wonderful atmosphere. Have had lots of fun and will have still more." (Male, around 25, Polish) [35]

Different types of police officers moved about among the supporters. Special "school police" officers, who usually teach and give advice to young children and pupils, chatted with the supporters, initiating contact by means of a large teddy bear in a football shirt and football boots. Loud rhythmic music sounded from some of the police cars. The supporters from both nations blew whistles and horns, they were singing and chanting, not to provoke but to celebrate. [36]

"We try to keep it nice and calm. One way is to have very firm police. On the contrary, we try to be very friendly, welcome them at the train station—we play music and we give information ... friendly. There won't be any incidents. We have done everything; we have police from many countries. We try to be friendly." (Male police officer, middle-aged, German) [37]

An interesting incident was a make-believe fight at an old market square. A game ("the cock of the walk") where supporters from both sides tried to climb onto a statue took place. When a young man succeeded, his supporters cheered enthusiastically. The competition continued and new winners climbed onto the statue. Police officers watched the "fight" without any reaction other than smiling. [38]

A similar event was a table-football match between "Poland" and "Germany" outside a cafe. With supporters shouting behind the two players, every goal resulted in applause and cheering in a joyful, friendly, and open atmosphere. [39]

Conclusions: Two factors had a disarming function, supporting a sense of festivity and preventing provocations and forerunners of violence. First, there were physical arrangements giving the impression of a festival, rather than a fight between antagonists. These arrangements met the visitors as soon as they arrived in Dortmund. The motto for the championship was "A time to make friends" and this ambition was supported by the arrangements. Furthermore, the arrangements aimed at mixing the supporters and the police, that is, the strategy of not separating them, proved to be disarming in the sense that national groups were not provoked by fences or police. Such arrangements also signaled trust that no violence was expected. The second important factor was the role taken by the police. As the official preservers of law and order, their promotion of a festival spirit served as a symbolic disarming. Equipment used by the police was not provocative, so supporters relaxed and had no reason to suspect the police of antagonism. The police commissioner confirmed that the decided strategy was that police officers in normal police uniforms should appear in small groups and mingle with fans and that heavily equipped police should be kept out of sight. [40]

Not all the fans arriving in Dortmund had tickets for the actual match. The danger of disorder, chaos, and provocation would not be far away with thousands of fans, under the influence of liquor, from two different nations going astray in the city. In order to give all the visitors an opportunity to follow the football match, two areas with huge video screens were made available. One was located in a square (Friedensplatz) in the city; the other was set up in an exhibition hall (Westfalenhalle) on the way to the football stadium. The shortest route to these two places was marked with a red carpet, fastened on the ground with bolts, starting at the railway station and leading the supporters to the Friedensplatz. The arrangement with the red carpet had the obvious effect of keeping the supporters together and preventing them from wandering about the city. Those satisfied with watching the screen outside stopped at the Friedensplatz. The carpet went further to Westfalenhalle, which swallowed most of the remaining ticket-less supporters. Entry was free. Those with tickets for the match only had to pass Westfalenhalle and follow the carpet to the stadium. [41]

On both sides of the carpet there were several coffee stalls, beer vans, hamburger and sausage stands, and other refreshment stalls. Garbage pails located in various places, as well as large flowerpots decorated with flags of the participating nations kept the streets and open places tidy and festive. In addition, older people went around picking up beer cans and garbage and pulling mobile garbage pails. As they gave the impression of being grandmothers and grandfathers, they could keep the supporters from littering the streets. [42]

The main arrangements to avoid disorder were the temporary areas with the video screens, a measure that made it possible for the visitors to take part in the great football event without tickets. The supporters were allowed to enter the areas after being searched for dangerous items and had to leave bottles and liquor outside the enclosed area. Civilian security guards in yellow or grey shirts carried out the inspections. Police officers were present, but were not involved in the inspections. According to the police commissioner, it was an intentional strategy to use civilian guards instead of police officers. The procedure was much the same as at airport security gates, a situation most people are used to nowadays. The fans collaborated and no objections could be observed. People with plastic beer glasses in their hands were allowed in, but all bottles had to be thrown away. This seemed to mean that drinking and partying were allowed, but not dangerous objects. Food and beer were also served from several stands in the marketplace. Portable toilets were also available to the supporters. [43]

The role of the police, in connection with the crowd moving from the railway station to different places, was just to be present and visible in and around the large procession of supporters. They were also ready to answer questions from the visitors. The majority of the police officers patrolled in pairs giving no impression of being a military force. In the neighboring streets, police vehicles were visible, but according to the fans interviewed, this was seen as a sign of security. The atmosphere was such that Polish as well as German fans mentioned that they felt secure because police were present and visible. Also the inspections of fans were mentioned as a factor contributing to order and security.

Supporter: "Very good. We came from Castrup [the airport in Copenhagen] and were checked immediately. That is good; people should buy this so that there won't be any riots like in France where they had bashed up Nivel [a comrade]. They were German hooligans. We are all here in order to have fun." (Male, around 20, German) [44]

When the Friedensplatz market was filled up and was closed off, people waited in line outside the gates hoping to get in. This certainly created a dangerous situation—one that could have perhaps been foreseen by the organizers. The restricted number of spectator places represented a danger of "chaoticizing." [45]

The way the toilets were arranged—right beside the gates—was also a danger. The lines were so long that people could hardly get back from the toilets without being pushed into the fence at the gate, which also made the fence fall in front of the people outside the gate. No one, however, was observed taking advantage of this to try to get in. Once again, a blunder in the physical arrangements risked creating chaos. [46]

Conclusions: Certainly, the arrangements contributing to maintain order and organization, as described here, are some of many efforts, but two main features seem to recur. First, there was an attempt to provide for the visitors' basic physical needs: familiar inspection procedures, relaxed and service-minded police officers, food and beverages stands, and hygienic facilities such as garbage pails and toilets. The police commissioner confirmed that the organizers had carefully planned for an open access to food and beverages. [47]

The other factor had to do with a recognizable order in an otherwise crowded and confusing event. As long as the visitors could understand "the infrastructure" of the crowd event, they could relate to the seemingly chaotic situation and behave in accordance with well-known and internalized behaviors, such as throwing garbage in pails instead of littering the street. Following a line (on the red carpet) instead of getting lost looking for foreign street names probably induced a sense of psychological security and comprehensibility and served as a mitigating strategy. This type of arrangement most certainly contributed to the maintenance of comprehensible order among the visitors. [48]

With the absence of provocative organizing activities and expressions of distrust, a high sense of responsibility among the supporters was maintained. Maintaining responsibility seems to be an important prerequisite for order and the avoidance of chaotic situations. [49]

Even the dangerous situation of the crowd not getting in did not result in violence. Mitigation strategies still dominated. The police informed people, they let single individuals in who had lost their friends, and they talked and treated people in a friendly way, i.e. they performed the mitigating strategies of organizing, differentiating, and disarming. [50]

Thus, no obvious signs of distress were present among the visitors. Our interpretation is that the arrangements designed to satisfy the supporters' basic physical needs as well as their need for safety, security, and a recognizable order were successful in the sense that they established and maintained a sort of intelligible organization. [51]

As mentioned above, the motto for the championship was "time to make friends." This could be seen as an unrealistic ambition as the supporters spoke different languages and had different historical memories. In the case of Poland and Germany, there might still be distrust between groups of fans due to what happened during World War II. Thus, this specific football match was a challenge to the organizers. In the Westfalenhalle, there was a planned activity to unify the supporters. The hall was filled with about 12,000 German and Polish supporters about one hour before kick-off, with mostly German fans standing in the middle, and Polish supporters seated along the stands. When the arena was opened three hours before the match, the supporters of different nationalities were mixed when seated. Gradually, however, the supporters re-grouped into more homogenous sections. Most of the Polish supporters remained in the seating areas, while the Germans gathered in the center of the hall. This was an autogenous process neither initiated by the organizers nor by any cheerleaders. It could be seen as a kind of self-regulation, with supporters gradually moving together. Certainly, this "self-categorization" in a physical sense could be interpreted as an assembly to fight. The fight in this case, however, concerned a cheering competition, not physical confrontation. [52]

About one hour before kick-off, a disc jockey came on to the stage. With loud and rhythmic music he tried to involve the crowd in cheering and singing together. As long as he spoke German, he was not very successful, but when he introduced a familiar melody with the simple text of la-la-la-la, he succeeded in getting the whole arena to take part in the singing. When he greeted and welcomed the German supporters, the Poles started booing, and vice versa. Such attention to the different groups seemed to promote the "cheer war." However, the DJ succeeded in calming the 12,000 or so supporters by playing a soft and well-known melody. The supporters sang (in English) and lit their mobiles and waved them. As soon as the song faded away the "cheer war" emerged again. Not until the DJ played well-known English football songs did he succeed in uniting all the supporters in a common activity and was even able to get the whole crowd to wave their arms. The English cheers disarmed the antagonistic groups and united them in singing well-known songs; the fans demonstrated a sense of unity and affinity reducing the risk of negative stereotyping. [53]

Another example of positive categorization was a collaborative task. At Friedensplatz, organizers threw white balloons to the fans as a means of unifying them. The balloons were handed from one to the other and the challenge of keeping them in the air as long as possible was a common task for all the fans. During the game, at times the crowd divided into competitive groups of fans, verbally fighting each other and shouting for their own teams, while at other times the supporters were united in a superordinate common identity as football fans, and it was impossible to distinguish Polish from German fans. [54]

Still another example of affinity and unity was the arrangements in the city before and after the game. The local authorities had arranged a festival with dance, entertainment, refreshments, etc. Police officers mingled in the streets before the game, gluing Polish and German mini-flags on the visitors' cheeks. Many of the supporters appeared with a German flag on one cheek and a Polish flag on the other, thus indicating a festival for both German and Polish supporters. [55]

When the supporters returned to the city after the game, they could join in the same activities regardless of nationality—and they did. The "fight" was over and the hosts had organized a variety of festive activities. Both German and Polish fans expressed their appreciation of the other group. The German fan cited below had only good things to say about the Polish fans.

"Yes, well, very hospitable, very hospitable, very generous. Well, for example, I ordered another beer 'No, no don't pay. We pay, we pay' and so on. Very OK; very good. And then also taking photos all the time just like the Japanese. That is really, that's really great." (Male, around 25, German) [56]

"Yes, but the German fans up to now have been very friendly. We have twice greeted a bit by shouting 'Germany' and we were greeted with 'Poland' but actually they are very peaceful and they were a bit drunk but all this is okay." (Male around 25, Polish) [57]

"Well, I think it's very good between the fans, Polish and German. Sometimes they even walk with their arms hooked (arm in arm) or sing their songs to each other. Well, at least it's still peaceful." (Female, around 18, German) [58]

Most comments by the supporters, irrespective of nationality, were of a similar nature, indicating positive instead of negative stereotyping. [59]

Conclusions: Three processes or arrangements seemed to be important for reducing negative stereotyping and creating a sense of unity and affinity. These were arrangements for festive activities, participation in common activities and no separation of groups of fans. [60]

The whole situation with two antagonistic nations and an important fight for advancement in the Championship could be seen as a basis for aggravation and violence. The idea behind a tournament is combat between the contending parties. Certainly, the supporters also ought to contend with each other. The situation, as such, presents a danger of negative stereotyping and riots. In the case of the Dortmund events, this danger was reduced by arrangements for festive activities: both Polish and German supporters were invited to the same party. The participants, as well as the police officers, seemed to have no problem in identifying the supporters as basically peaceful. [61]

The second important factor was the common activities, for example, being physically involved in the same activities. When the DJ introduced songs and cheers that were familiar to both groups and also linguistically neutral, he succeeded in breaking the stereotype, shouting "Polska-Polska" met with "Deutschland-Deutschland." Thus, finding common activities seems to be an important means of reducing stereotypical manifestations and making possible a superordinate social identity as World Cup fans. [62]

Third, the different teams' supporters were mixed and not separated, as is usual in national league matches. German and Polish supporters were not separated in different lines or areas, but mingled together in the square before and after the game, which created an obvious impression of being part of the same festival. Even if the fans themselves took part in self-categorization, gathering together in national groups, the arrangements were such that differentiating came to dominate negative stereotyping. Symbols were exchanged (mini flags), and the mix of supporters encouraged arranged photos, small talk, and loud greetings. [63]

The outcome of this event was peaceful and joyful. Signs of aggravation were few and were nipped in the bud. A summary of the results is presented in the table below.

It is obvious, from Table 3, that the intentional efforts to create a festival atmosphere seem to have a great impact on the progress of events. That impact is elaborated further in the discussion section below. [65]

The event described in this study proved to be a large and peaceful "happening" entirely without riots or violence between fans. The reasons for this have been preliminarily interpreted by using the AM model, developed from studies of political demonstrations (GUVÅ & HYLANDER, 2008; HYLANDER & GUVÅ, 2008; GRANSTRÖM & ROSANDER, 2008; GUVÅ, 2008). The AM model distinguishes between two types of concurrent processes in crowd events, aggravation and mitigation. [66]

According to the AM model, it is assumed that one group's perception of another group's intention has an obvious impact on the first group's behavior. When one group does not believe in or distrusts another group's peaceful intentions, aggravation processes are likely, which in turn may lead to mutual distrust. Both groups blame defensive and destructive behavior on the other group. If, however, mutual trust is publicly expressed and is also linked to mitigation strategies, the most likely outcome is peaceful, according to the AM model. In the account above, as illustrated with excerpts, the fan groups' trust each other and the organizers (discussed below). Additionally, the organizers' trust in the supporters' behavior could be mediated by the arrangements. Several such examples were found in Dortmund. The police officers, mingling without helmets, signaled trust rather than distrust. The easy access to beer may have been seen as a sign of trusting the supporters' sound judgment regarding alcohol. [67]

The response from the fans also signaled trust in the organizers. In the marketplace there were families with children, and as time passed more girls entered through the gates. Teenage girls, who were interviewed, said that they were there to have fun and join the party. Another group of very young teenagers was observed getting meal tickets from an adult and then being left on their own in the marketplace, showing that the adults trusted that the organizers could keep the marketplace calm. Also, several fans remarked that they felt secure because the police were there. [68]

A significant signal of trust was the limited number of civilian security guards in the Westfalenhalle. Inside the Westfalenhalle as well as at Friedensplatz there were no police officers, and only civilian guards dressed in colored T-shirts carried out the personal inspections. The arrangement gave an impression of a security check rather than a police action. Inside the Westfalenhalle, it was difficult to discern the guards, even though they most certainly were present. The supporters had to find their seats themselves—all seats were free of charge and the guards seemed to trust the visitors' ability to fill up the seats in an orderly way. [69]

The great majority of Polish fans interviewed expressed trust in the German fans, and vice versa. The fans referred to the organization as a guarantee for the trust between the fans, saying that here (at Friedensplatz) nothing will happen, or that as long as the police and security are here nothing will happen. [70]

Earlier research (REICHER, 1996) has pointed to the fact that it is not the intention of one group that creates a springboard for the other group's actions; instead, it is the idea of the other group's intention that matters, i.e., the imagination or expectations of the other group's actions. The conception of the other group's intentions comes from the acts of the other group, not from the perception by that group itself. Recurrent in the interviews was a belief in the other group's peaceful intentions, labeled "trust." The word "friendly" appeared frequently in the narratives. What is noticeable, however, is that trust between the groups of fans was supported by the arrangements. The interviewees said that nothing would happen, not where there is a party. This outcome supports the interpretation that being part of a common, peaceful, and celebrating community increases trust between groups. The few supporters believing that there would be fights between fans were the only ones interviewed who had not talked and intermingled with the other group. The conclusion is that when the arrangements support a higher order social identity, trust increases between groups and those few not taking part will not make a difference. BARON and RICHARDSON (1994) assumed that situational factors can mask the effect of personality, as shown by these results from analyses using the AM model. [71]

The model discerns three mitigating processes: disarming, peaceful organization, and positive categorization, and three aggravation processes: negative stereotyping, creating chaos, and provocative actions. Theoretically, crowd events ought to be organized in such a way that the risk of aggravation is reduced. However, based on earlier research (GUVÅ & HYLANDER, 2008; GRANSTRÖM & ROSANDER, 2008) there is good reason to assume that it is not enough for organizers and the police to plan for violent situations. It is much more important to plan for peace, which means working out plans and strategies for peace interventions, i.e. strategies for disarming, peaceful organizing, and positive categorizing. Apparently, this was the case in Dortmund at the World Championship in 2006. [72]

As previously mentioned, previous mutual disarming between groups of, for example, fans and guards typically involved a functional dialogue. In a few cases, some supporters with flags climbed up to the stage. These rebels were "talked off" by a single guard. No massive or provocative actions were carried out in these cases; a guard just asked the supporter to leave the stage. The procedure was the same when some supporters had climbed up onto a small roof structure above the entrances. The guard succeeded in "talking" the climbers down, except in one case. Then, the guard asked a fellow supporter to take over the persuading, which eventually succeeded. An expression of trust typically precedes the dialogue but the dialogue itself increases mutual trust between groups. Certainly, this was a strategy for disarming. The friendly police behavior probably disarmed any tendency towards antagonism between police and supporters, but also between different supporter groups. Thus, the covert police strategy was to behave in a friendly way and to treat the fans as festival participants rather than rival supporter groups. This echoes several studies reported by RUSSELL (2004) where there was good interaction between police and citizens, with humor as a salient aspect that defused potential conflicts. Low-profile policing in conjunction with high-risk football tournaments has also been discussed by STOTT et al. (2007). The most successful disarming, however seemed to be all the party arrangements. [73]

All the arrangements for maintaining a recognizable order contributed to the success of the Dortmund approach. The red carpet and large video screens in combination with easy access to beverages and food and well-organized security procedures were all arrangements that reduced the danger of chaos. The organizing did not, however, take responsibility away from the participants. They still had to decide by themselves where to stay, what to eat and drink, and how to interact with other fans and with the police. [74]

The police strategy of patrolling in pairs without helmets and shields made it possible for the fans to differentiate and not negatively stereotype the police. This behavior signaled that police forces could act in different ways, that is, peacefully and supportively, instead of in a military and repressive manner, which many supporters may have expected. The strategy followed by the police helped the supporters to differentiate between a friendly and a repressive attitude among the police. This prevented a negative stereotyping by supporters, as well as by the police, from taking place. [75]

The results are also in line with the social identity perspective (TAJFEL, 1978; TAJFEL & TURNER, 1979) and the ESIM (REICHER, 1996). The social identity of the crowd is to a high degree created in relation to its surroundings and interaction with other groups. It is closely related to the context that, in turn, is made up of the other group's actions. In Dortmund, the trusting and festive atmosphere constituted the context for the supporters from both groups and enhanced the trust between the two groups by means of distinct unifying symbols, like the German and Polish flags on the same items and people. [76]

The results of the arrangements in Dortmund contributed to a superordinate social identity of peaceful fans or party participants. As a result, the arrangements and the course of action promoted a social identity among the supporters as peaceful and responsible participants in a festival rather than as combating rivals in a warlike fight. The social identity among the police officers was also perceived as friendly. Earlier, we assumed that avoiding violent strategies is not enough and that active peacemaking strategies, mitigation, must also be employed. We now assume that "festival-making" is a specific mitigation strategy that supports a superordinate common social identity and thus prevents negative stereotyping between groups. It proved to be a means of securing orderly organization, thereby preventing chaos as well as being a means of securing disarming contacts between groups. [77]

An objection to this interpretation could be that irrespective of the organizers' efforts, the attitude and behavior of the spectators in Westfalenhalle and at the Alter Markt were such that no matter what happened they would have stayed calm and that the course of events depended on the category of people who were there. On the other hand, this match was one of the high-risk matches in the tournament. From the interviews, we can conclude that people did not expect trouble and that they felt that this was due to the atmosphere of festivity and friendliness promoted by the festival arrangements and the presence of the low-profile police. Thus, the way the mitigating strategies of the organizers were perceived by the football supporters seemed to be of great significance. [78]

Another objection to our interpretation could be that the crowd in the streets stayed calm because the police, in an event that occurred before the game, had taken individuals into custody who they thought might start a riot. On the other hand, that event may also be used as an illustration of a police intervention that did not escalate into an extensive riot. As the huge crowd of fans adhered to the common and peaceful activities and developed a peaceful social identity, promoted by the arrangements, isolated police intervention did not change that situation. [79]

The objections described above should be considered and tested in future research. Thus far, however, our conclusion is that the AM model could be used to describe how the arrangements in Dortmund served to exclude aggravation processes and to promote mitigation processes such as disarming, peaceful organizing, and positive categorizing. In particular, the organizers' "festival-making" combined all these mitigation strategies and helped to create a superordinate social identity of the crowd, which could best be described as supporters taking part in a joint and peaceful football party. The very important lesson to be learned from this study is that active planning for peace in large sports events seems to be as important as planning to avoid war. [80]

We proposed applying the AM model to sports events. Our conclusion is that it is feasible to use in this substantive area and that mitigation arrangements in sports events may be understood by making use of AM model analyses. Trust is present when mitigation processes are identified. The new set of data has made it possible to elaborate the AM model in two ways. First, new features have been added to the category of disarming. Creating a festival or festive atmosphere seems to be one of the most powerful ways to make opposing groups disarm. Second, trying to create a superordinate identity for different groups has a peace-making effect. The feasibility of creating such a superordinate identity—"we are all football fans"—seems to necessitate the strategy of not separating opposing fans. This feature may, however, be specific to the substantive area of sports events. In the case of political protests, it is difficult to imagine how such a superordinate identity of, for example, anti-fascists and neo-Nazis, could be created. Earlier results point instead to different places and times as necessary prerequisites for avoiding violence between protesters and anti-protesters. However, creating a superordinate identity should be added to the features of peaceful categorizing in the AM model, and should be understood as differentiation, positive stereotyping of the other group, and creating a common superordinate identity. Third, the organizer of an event may promote mitigation by maintaining a recognizable order of equipment and infrastructures (e.g. food stalls and the red carpet) and thus preventing chaos. Earlier research has only focused on the way the different groups have been organized. This study clearly points to the vital importance of the organizers' arrangements. [81]

The research presented here was supported by The Swedish Emergency Management Agency.

Appendix: Observation Schedule, Semi-Structured Observation, Dortmund, June 14, 2006

2) A congress hall with galleries and an open space in the middle, planned for 120,000 visitors. <back>

Baron, Robert A. & Richardson, Deborah R. (1994). Human aggression. New York: Plenum.

Crabbe, Tim (2003). "The public gets what the public wants". England's football fans, "truth" claims and mediated realities. International Review for the Sociology of Sports, 38(4), 13-425.

Drury, John & Reicher, Stephen (1999). The intergroup dynamics of collective empowerment: Substantiating the social identity model of crowd behavior. Group Processes & Intergroup Relations, 2(4), 381-402.

Drury, John & Reicher, Stephen (2000). Collective action and psychological change: The emergence of new social identities. British Journal of Social Psychology, 39(3), 579-604.

Drury, John; Stott, Clifford & Farsides, Tom (2003). The role of police perceptions and practices in the development of "public disorder". Journal of Applied Social Psychology, 33(1), 1480-1500.

Glaser, Barney G. (1978). Theoretical sensitivity. San Francisco: The Sociology Press.

Glaser, Barney G. (2006). Doing formal theory: A proposal. San Francisco: The Sociology Press.

Glaser, Barney G. (2007). Doing formal theory. In Antony Bryant & Kathy Charmaz (Eds.), The Sage handbook of grounded theory (pp.98-113). Los Angeles: Sage.

Glaser, Barney G. & Anselm Strauss (1967). The discovery of grounded theory. New York: Aldine de Gruyter.

Granström, Kjell (2008). Soziale Identität und Massenereignisse. Freie Assoziation, 11(3), 7-14.

Granström, Kjell & Rosander, Michael (2008). Friedensbewahrende Intergruppenprozesse zwischen Polizei und Demonstranten bei den "Reclaim the street"-Protesten. Freie Assoziation, 11(3), 51-64.

Granström, Kjell; Guvå, Gunilla; Hylander, Ingrid & Rosander, Michael (2009). Riots and Disturbances: How riots start and how order is secured. FOG-report 64, Linköping University, Department of Behavioural Sciences and Learning.

Guvå, Gunilla (2008). Hilfe oder Selbstkontrolle—über Intragruppenprozesse bei der Polizei bei Massenereignissen. Freie Assoziation, 11(3), 65-84.

Guvå, Gunilla & Hylander, Ingrid (2008). Aggravation und Friedensprozesse. Freie Assoziation, 11(3), 33-50.

Guttman, Allen (1998). The appeal of violent sports. In Jeffrey H. Goldstein (Ed.), Why we watch (pp.7-26). New York: Oxford University Press.

Hammersley, Martin (2006). Ethnography: Problems and prospects. Ethnography and Education, 1(1) 3-14.

Hammersley, Martin & Atkinson, Paul (1983). Ethnography: Principles in practice. London: Routledge.

Hau, Stephan (2006). Communication as the most important police strategy at the Football World Cup Final 2006. FOG-report No 61, Department of Behavioural Sciences and Learning, Linköping University, Linköping.

Hopkins, Nick & Reicher, Stephen (1997). The construction of social categories and processes of social change. In Glynis Marie Breakwell & Evanthia Lyons (Eds.), Changing European identities (pp.69-93). London: Butterworth.

Hylander, Ingrid (2008). Massenbewegungen, Massenproteste und Soziale Identität. Freie Assoziation, 11(3), 15-32.

Hylander, Ingrid & Granström, Kjell (2010/in print). Police use of active and passive mitigation strategies. In Johannes Knutsson & Tamara Madensen (Eds.), Preventing crowd violence. Boulder, CO: Lynne Rienner Publishers.

Krane, Vikki & Baird, Shannon M. (2005). Using ethnography in applied sport psychology. Journal of Applied Sport Psychology, 17(1), 87-107.

Mustonen, Anu; Arms, Robert L. & Russell, Gordon W. (1996). Predictors of sport spectators' proclivity for riotous behaviour in Finland and Canada. Personality and Individual Differences, 21(4), 519-525.

Prentice-Dunn, Steven & Rogers, Ronald W. (1989). Deindividuation and the self-regulation of behavior. In Pauluse Hillsdale (Ed.), Psychology of group influence (pp.87-109). New Jersey: Lawrence Erlbaum.

Reicher, Stephen (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, 26(1), 115-134.

Reicher, Stephen (1997). Collective psychology and the psychology of the self. BPS Social Section Newsletter, 36(1), 3-15.

Reicher, Stephen; Stott, Clifford; Cronin, Patrick & Adang, Otto (2004). An integrated approach to crowd psychology and public order policing. Policing, 27(4), 558-572.

Rosander, Michael & Guvå, Gunilla (2009). Keeping the peace—police behaviour at a mass event. (Submitted)

Russell, Gordon W. (2004). Sport riots: A social-psychological review. Aggression and Violent Behavior, 9(4) 353-378.

Russell, Gordon, W. & Arms, Robert L. (1998). Towards a social psychological profile of would-be rioters. Aggressive Behavior, 24(3), 219-226.

Russell, Gordon W.; Arms, Robert L. & Mustonen, Anu (1999). When cooler heads prevail: Peacemakers in a sport riot. Scandinavian Journal of Psychology, 40(3), 153-155.

Russell, Gordon W. & Mustonen, Anu (1998). Peacemakers, those who would intervene to quell a sports riot. Personality and Individual Differences. 24(3), 335-339.

Stott, Clifford & Drury, John (2000). Crowds, context and identity: Dynamic categorization processes in the "poll tax riot". Human Relations, 53(2), 247-273.

Stott, Clifford & Reicher, Stephen (1998). How conflict escalates: The inter-group dynamics of collective football crowd "violence". Sociology, 32(2), 353-377.

Stott, Clifford; Hutchinson, Paul & Drury, John (2001). "Hooligans" abroad? Inter-group dynamics, social identity and participation in collective "disorder" at the 1998 World Cup Final. British Journal of Social Psychology, 40(3), 359-384.

Stott, Clifford; Adang, Otto; Livingstone, Andrew & Schreiber, Martina (2007), Variability in the collective behaviour of England fans at Euro2004. "Hooliganism", public order policing and social change. European Journal of Social Psychology, 37(1), 75-100.

Tajfel, Henri (1978). Differentiation between social groups: Studies in the social psychology of intergroup relations. London: Academic Press.

Tajfel, Henri & Turner, John C. (1979). An integrative theory of intergroup conflict. In William Austin & Stephen Worchel (Eds.), The social psychology of intergroup relations (pp.33-47). Monterey: Brooks/Cole.

Tajfel, Henri; Billig, Michael G.; Bundy, Robert P. & Flament, Claude (1971). Social categorization and intergroup behaviour. European Journal of Social Psychology, 1(1), 149-178.

Turner, John (1999). Some current issues in research on social identity and self-categorization theories. In Naomi Ellemers, Russell Spears & Bertjan Doojse (Eds.), Social identity: Context, commitment, content (pp.6-34). Oxford: Blackwell.

van Hiel, Alain; Hautman, Lobke; Cornelis, Ilse & De Clercq, Barbara (2007). Football hooliganism: Comparing self-awareness and social identity theory explanations. Journal of Community & Applied Social Psychology, 17(2), 169-186.

Zimbardo, Phil G. (1969). The human choice: Individuation, reason, and order vs. deindividuation, impulse, and chaos. In William J. Arnold & David Levine (Eds.), Nebraska symposium on motivation (pp.237-307) Lincoln: University of Nebraska Press.

Ingrid HYLANDER is Associate Professor in Educational Psychology at Linköping University, Senior Researcher at Karolinska Institutet, and licensed psychologist. Her research interest is in group psychology, team work, and professional consultation and collaboration.

Center for Family Medicine
Karolinska Institutet
Alfred Nobels Allé 12, S-141 83 Huddinge
Sweden

Kjell GRANSTRÖM, Professor Emeritus in Educational Psychology, is a reader in communication and licensed psychologist. His main research interests concern communication and dynamics within and between groups, such as school classes, staff teams, football supporters, and groups in demonstrations.

Department of Behavioural Sciences and Learning, Linköping University
S-581 83 Linköping
Sweden

Hylander, Ingrid & Granström, Kjell (2010). Organizing for a Peaceful Crowd: An Example of a Football Match [81 paragraphs]. Forum Qualitative Sozialforschung / Forum: Qualitative Social Research, 11(2), Art. 8, http://nbn-resolving.de/urn:nbn:de:0114-fqs100283.

Forum Qualitative Sozialforschung / Forum: Qualitative Social Research (FQS)

ISSN 1438-5627

Creative Commons Attribution 4.0 International License

Event	Type of study	Data base	Word count
Protests against EU-summit 2001	Focus groups	Demonstrators, four groups Police officers, two groups	41,000 18,000
Reclaiming activities on 1 May, 2004, City 1	Field study Observation, interviews two observers	Demonstrators at the event Police officers at the event Police planning sessions	5,800
Reclaiming activities on 1 May, 2004, City 2	Field study Observation, interviews two observers	Demonstrators at the event Police officers at the event	4,900
Neo-Nazi demonstration Counter-protests 2004	Field study Observation, interviews six observers Interviews before and after	Demonstrators at the event Police officers at the event Police planning sessions Organizers of protests	20,000 3,800

Relational conditions	Aggravation	Mitigation
Mutual treatment	Provoking	"Disarming"
Organizing	Creating chaos	Peaceful organizing
Categorizing	Negative stereotyping	Positive categorization

Mitigation processes	Observed features
"Disarming" instead of provoking	Festival-making: physical arrangements for a festival, police officers promoting a festival spirit
Peaceful organizing instead of chaos	Arrangements for basic needs Creating a recognizable order
Positive categorizing instead of stereotyping	Creating a superordinate identity: opportunities for participating in common activities, not separating groups of fans