an article co-authored by john drury (2007)

Knowledge-Based Public Order Policing: Principles and Practice

 I‘ve put this here on the site, because there’s no way to access it anywhere else on the internet, and it has some uses for anyone looking at the Aufhebengate affair and new developments in soft cop policing.

It’s co-authored by Stephen Reicher, Clifford Stott, John Drury, Otto Adang, Patrick Cronin and Andrew Livingstone.

See also this :“Chaos Theory”

I haven’t bothered to organise the justification and spacing on this as it involves too much time and isn’t worthwhile


Knowledge-Based Public Order Policing: Principles and Practice

Much public order policing is still based on the assumption that crowds are inherently irrational
and dangerous. We argue that this approach is both misinformed and counter-productive because it
can lead to policing interventions that increase the influence of those advocating violence in the crowd.
We challenge traditional assumptions about crowd psychology and demonstrate how widespread conflict
derives from the interactions between police and crowds. From this, we develop general guidelines as
to how policing can reduce crowd violence and lead crowd members themselves to self-police violent
groupings in their midst. We then use examples from anti-globalisation protests and the Euro 2004 football
championships to show how these guidelines can be applied in practice and how effective they can be. We
conclude by arguing that such knowledge-based crowd policing can turn crowd events into opportunities
to overcome seemingly intractable conflicts between the police and groups within our society.
Most policing is in some way related to the
maintenance of public order. So how come
public order policing refers specifically to
crowd events? Part of the answer lies in
the assumption that crowds pose an inherent
threat to order. We use the term ‘public
order policing’ precisely because we associate
crowds with public disorder.
This assumption has profound implications
for the ways in which crowds are policed.
Most crucially, because it locates the cause
of violence as lying entirely with the crowd
as opposed to arising out of the interaction
between crowds and the police, it neglects the
possibility that police actions may contribute
to the production of conflict and hence provides
no basis for developing strategies, tactics
and technologies that might minimise such
a possibility. Rather, the focus lies on how
to contain crowds and prevent them from
expressing their natural belligerence. In short,
the crowd is seen as a problem and treated as
There are two major problems with this
approach. The first is that it runs the risk
of creating a self-fulfilling prophecy: if you
define people as hostile and if you then act
towards them in ways that make it obvious,
you think of them as being hostile, then they
are very likely to become hostile in response.
The first time the chairman of Manchester
United described his fans as ‘animals’, the fans
responded by chanting ‘we hate humans’ at
the next game (Robins, 1984).
The second problem is that this approach
misses out on a major opportunity. If the
police can interact with crowd members in
ways that lead to a deteriorating relationship
and increase conflict, they can equally
interact in other ways that lead to improving
relationships and reduce conflict. Moreover,
to the extent that police-crowd relationships
are emblematic of relationships with the wider
groups from which crowdmembers are drawn
(for instance, events like Brixton and Toxteth
were seen to crystallise negative relations
between the police and black people
in Britain), then crowd policing can have a
profoundly positive effect upon policingmore
Our aim, in this paper, is to show how
an informed understanding of crowd psychology
can be used in order to develop
forms of policing that promote reconciliation
rather than conflict. The paper is divided
into twomain sections. In the first, we explain
the principles that underlie our proposals for
knowledge-based public order policing. That
is, we provide a critical review of classic crowd
psychology, outline a more contemporary
approach to the subject, and then draw out the
general implications for effective policing. In
the second section, we provide an example of
how these principles can be applied in practice
and of how they are effective in transforming
negative relations between police and crowd
into positive relations.
The principles of knowledge-based
public order policing
The classic psychologies of crowd
The July 2006 distance learning handbook for
the Initial Public Order Commanders Course
of the National Centre for Policing Excellence
(NCPE—now incorporated into the National
Policing Improvement Agency) contained the
following claims:
EverettDMartin; a psychologist of
the 1920’s, wrote in his book ‘The
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 the lessening of an individual’s
ability to think rationally, whilst
at the same time his/her more
primitive impulses are elicited in
a harmonious fashion with the
emerging primitive impulses of all
the other crowdmembers (Martin,
Downloaded from by guest on January 20, 2011
Knowledge-Based Public Order Policing Article Policing 405
It is certainly true that Martin’s words
express what we have termed the ‘classic
view’ of crowd psychology (Reicher et al.,
2004)—not surprisingly, since his book
‘The Behaviour of Crowds’ (1920) starts by
acknowledging that the theoretical approach
is taken from Gustave Le Bon, the founding
figure of the field (see Le Bon, 1895/1947).
Like Le Bon, Martin characterises crowd psychology
in terms of irrationality, primitivity
and aggressivity. If this last element is implicit
in the quote included in the NCPE document,
it is quite explicit elsewhere in Martin’s text.
Thus he states that ‘every crowd is potentially
if not actually homicidal in its tendencies’
(1920, p. 105). That is, even where the crowd
seems peaceable enough, violence can always
erupt at any time. In practice, thismeans that
crowd members must always be treated with
suspicion. They certainly cannot be recruited
as partners in themaintenance of public order.
By the 1920s, when Martin was writing,
the Le Bonian approach was already highly
controversial. Allport (1924) in particular dismissed
the notion of reversion to a primitive
groupmind. Hiswork initiated a tradition that
explains crowd action in terms of the character
of those individuals drawn to the crowd.
According to this perspective, if crowds are
violent it is not because peaceable individuals
are transformed in themass but rather because
violent individuals are drawn to crowd events
where they can express their true nature. If
the classic view, in popular parlance,might be
called the ‘mad mob’ approach, then this latter
view is the ‘hooligan’ approach. But despite the
differences, the two approaches share the view
that crowds are inherently mindless. Second,
both approaches see crowd action as inherently
meaningless; third they see crowds as
inherently aggressive and fourth they therefore
locate the responsibility for collective conflict
as lying exclusively within the crowd. So for all
the theoretical argument about the explanation
of crowd behaviour, there is no argument
about the primitive and pathological nature of
that behaviour. Were we still in the 1920s, the
claim that all psychologists agree with Martin’s
description of crowds would be eminently reasonable.
But psychology has moved on over
the last 80 years.
The new psychology of crowd behaviour
The core conceptual premise which underlies
both Le Bonian crowd psychology and its
Allportian critics, is that the standards which
control our behaviour are associated with individual
identity. If either individual identity is
stripped away in the crowd (Le Bon) or else
individual crowd members have flawed identities
(Allport), then the crowd action will be
uncontrolled and the normal restraints against
aggression will be removed.
Over the last 30 years, social identity
research (Tajfel and Turner, 1979)—by now
the dominant approach ingroup psychology—
has systematically dismantled the particular
notion of identity which underlies the
classic crowdpsychologies. Indeed, as its name
suggests, the social identity tradition rejects the
idea that people only have a single personal
identity. Rather, it argues, identity should be
seen as a system inwhich different parts govern
our behaviour (i.e. are psychologically salient)
in different contexts. Certainly there are times
when we do think of ourselves in terms of our
personal identities: what makes us unique as
individuals and different from other individuals.
But at other times, we think of ourselves
in terms of our group memberships (I am
British; I am a police officer; I am a Catholic,
or whatever) and of what makes our group
Downloaded from by guest on January 20, 2011
406 Policing Article S. Reicher et al.
unique compared to other groups. That is, we
think of ourselves in terms of our social identities.
Psychologically, the shift from personal
identity to social identity is what makes group
behaviour possible (Turner, 1982).
To illustrate the argument, consider the following
situation. You are on a crowded train
full of commuters, each of whom is trying to
ignore the presence of the others, seeking to
avoid eye contact and to minimise any physical
contact. Then the train stops. After some time,
there is an announcement which gives some
excuse for the delay. People start to think of
each other as fellow passengers (in opposition
to the train company) rather than as separate
individuals. They begin to turn towards each
other, to talk, and to smile.
Three points can be drawn from this illustration.
First, we need to distinguish between
a physical group of people (which we will call
an aggregate) and a psychological group. The
former simply refers to a set of people who are
co-present, while the latter refers to a set of
people who, subjectively, think of themselves
as belonging to a common social category.
The same aggregate may contain no psychological
groups (as in our example of the train
before the delay), one psychological group (as
in the train after the delay) or indeed multiple
different psychological groups (say in the
case of rival football fans crowded into the
same carriages). What is more, the psychological
groupings contained in the self-same
aggregate can shift as a function of unfolding
Second, when people shift from seeing
others as individuals to seeing others as
groupmembers, their relationships with them
undergo a fundamental transformation. Common
ingroup members are treated with
warmth and respect; they are trusted, supported
and they receive cooperation (see also
Tyler and Blader, 2000; Levine et al., 2005).
This does not usually extend to outgroup
members. The reason for this fundamental
transformation is that, once people define
themselves in terms of a group membership
the fate of the group as a whole (and hence
of others in the group), the well-being of
the group, the prestige and reputation of the
group becomes their fate, their well-being,
their prestige and their reputation.
Third, while peoplemay do things in groups
that theymay not probably do in other circumstances,
this need not necessarily be anti-social
(in the example that we have given, people are
more sociable and generous than on their dayto-
day commutes) and it certainly does not
indicate that they have lost control over their
actions. Indeed people in groups conform to
the beliefs and standards that are associated
with the relevant identity. In contrast to the
classic view that people lose identity and hence
lose control in collective settings, what is suggested
here is that people shift identity and
hence shift the bases of behavioural control
in groups. What they will then do depends
upon the particular group under consideration
(Reicher, 1987, 2001).
Thus far, we have referred to groups and
crowds as if the two were interchangeable
and, at one level, all we have argued about
groups in general does indeed apply to the
crowds in particular. Notably, there is by now
an extensive historical literature which shows
that crowd action is not random and uncontrolled
but rather is a faithful reflection of
the social beliefs of the groups involved (see,
for instance, Thompson, 1971; Reddy, 1977;
Davis, 1978). Even where crowds are violent,
the nature of that violence—both the targets
Downloaded from by guest on January 20, 2011
Knowledge-Based Public Order Policing Article Policing 407
that are chosen and the manner of the attacks
upon them—reflects belief systems that are
current in the relevant community.
However, there is one critical difference
between crowds and other types of group.
Most groups have formal means of discussing
and agreeing on group norms—and how to
apply these norms to novel situations. They
have meetings and conferences to discusswhat
they believe in and formal hierarchies to enact
decisions on a day-to-day basis. Crowds, characteristically
have neither. It is difficult to sit
down in the midst of a riot and discuss what
to do. Equally, there is rarely a formal authority
structure in the crowd. Yet crowd events
typically throw up unprecedented situations
such that it is all but impossible simply to
apply preformed responses. This combination
means that the process of defining appropriate
group action becomes more volatile andmore
fraught than elsewhere (Reicher, 1982, 1984).
There are liable to be multiple voices within
the group arguing for more or less confrontational
action. What, then, determines which
voice will be heard?
In answering this key question, it is necessary
to highlight another difference between
the crowd and other groups. Whereas the
psychological salience of group identity is
always associated with the (implicit) presence
of another group, that other group is not necessarily
a physical presence. By contrast, crowd
events generally involve face to face contact
between the different parties—either one
crowd versus another (say left wing and right
wing political groups) or else—very often and
of immediate interest here—between crowd
members and the police. We suggest that the
relationship and the balance between groupings
within the crowd is critically dependent
upon the interaction between the crowd and
The interactive dynamics of crowd conflict
There are three elements to our argument.
First, the relationship between outgroup
action, common fate and group formation
is much more immediate in crowd settings
due to the fact that the fate of crowd members
is directly determined by what outsiders
(notably the police) allow them to do. That is,
where the police have both the inclination and
the power to treat allmembers in a crowd event
as if they were the same, then this will create
a common experience amongst crowd members
which is then likely to make them cohere
as a unified group (Drury and Reicher, 1999,
2000, 2005; Drury et al., 2003). Thus, in one of
our studies (Reicher, 1996) the decision of the
police to throw a cordon across Westminster
Bridge—thus preventing all those attending a
student anti-loans protest from accessing the
House of Commons—transformed a multitude
of small grouplets into a unified crowd
who felt attached to all fellow demonstrators,
even complete strangers, in opposition to the
Second, where the common fate imposed
by the outgroup violates ingroup conceptions
of legitimacy—either imposing experiences
which are considered illegitimate or impeding
acts which are seen as legitimate—then the
crowd will unify around hostility and anger
towards that outgroup. To continue with the
Westminster Bridge example, students had
come to London so as to lobby parliament and
they felt that this was a fundamental democratic
right. When, from their perspective, this
right was denied, even the most peaceable of
participants supported attempts to breach the
police line. (Reicher, 1996, for other examples
Downloaded from by guest on January 20, 2011
408 Policing Article S. Reicher et al.
of this process, see Adang, 1998; Stott and
Reicher, 1998; Stott and Drury, 2000).
Our point is that one cannot specify in general
terms a set of actions which will arouse
crowd hostility or produce collective violence
(as is suggested by the common metaphor
that once there is tinder any spark will lead
to a riot). Rather, the events which precipitate
such violence will be different from group to
group as a function of their different conceptions
of legitimacy and their different histories
with outgroups (see Reicher, 1984; Adang and
Standaar, 1993).
Third, we are not arguing that the shared
perception of outgroup illegitimacy will lead
all crowd members into direct conflict with
outgroup members. Rather than assuming
that there is a simple dichotomy between violence
and non-violence, we need to consider a
spectrum of behaviour which goes from active
participation in violent acts to active suppression
of violence (see Adang, 1991). Outgroup
actions can impact at all points along this
spectrum and what may be critical is the way
they disrupt the willingness of crowd members
to contain the violence of those in their
midst—what we term self-policing. A powerful
illustration of this can be found in our
study of the behaviour of England and Scotland
fans during the 1998 world cup in France
(Stott et al., 2001).
Domestically, there is as much (if notmore)
violence in Scottish as in English football. So
why, at an international level, are the English
involved in more conflict than the Scottish?
The answer lies in their differential reputations
and hence in the ways in which they
have been treated historically. The English
were seen as dangerous, so England fans in
general were heavily policed and subject to
hostility—even violence—from local people
and rival fans. Under these conditions fans
who initially condemned conflict began to
sympathise with those who expressed hostility.
Far from impeding such people, the mass
of fans increasingly condoned them as both
expressing and protecting the group interest.
The Scots, by contrast, were seen as colourful,
boisterous and entertaining. They were generally
welcomed, liked and trusted by outsiders.
As a result, any Scottish fan who was violent
was perceived to be threatening the ingroup
reputation. So, while there were plenty of incidents
where individuals began to be violent,
others would characteristically intervene to
stop them. Overall, generalised outgroup hostility
increased the influence of violent fans
and the support given to them. Conversely,
generalised outgroup friendliness decreased
the influence of violent fans and the active
intervention against them.
Putting all three elements together, our
argument explains why policing practices
based on the classic crowd psychologies may
not only bemisinformed but actively counterproductive.
If one believes that all crowd
members are potentially if not actively dangerous,
then one will (1) treat all crowd members
alike and hence create unity amongst
them, (2) react to the violence of some crowd
members by imposing restraint on all, thus
increasing the likelihood of violating ingroup
conceptions of legitimacy and uniting the
crowd in hostility and opposition to the police,
and (3) increase the influence of those advocating
conflict in the crowd and undermine
self-policing amongst crowd members.
This is whatwemeanby our earlier reference
to a self-fulfilling prophecy. By responding to
the acts of the few by clamping down on the
many, a limited problem may be transformed
into a general conflagration. What we are
Downloaded from by guest on January 20, 2011
Knowledge-Based Public Order Policing Article Policing 409
proposing is that, instead of focusing exclusively
on the (potential) violence of crowd
members, we focus on the processes through
which violence escalates and de-escalates. By
understanding these processes, and their contribution
to them, the police will be better able
to deal with the crowds. We do not promise
that such an understanding will always allow
conflict to be avoided. However,we do suggest
that this understanding can guide the police
to act in ways that minimise conflict and
maximise the opportunities to engage crowd
members themselves in achieving this end.
The practice of knowledge-based
public order policing
General guidelines for policing crowds
The implications of contemporary crowd psychology
for public order policing can be
summarised in terms of four general guidelines.
These build on guidelines that we have
described previously (Adang, 1998; Reicher
et al., 2004) and which were incorporated in
the ACPO Manual on Public Order Policing
The first guideline relates to issues of
information and intelligence. Currently, the
emphasis is on criminal intelligence relating
to the presence of individuals with a known
history of violence (Cronin, 2001). While
this work has proved very effective, taken
in isolation, it ignores the process whereby
the violence of the few does (or does not)
become collective. If the police are to understand
this, they need to understand the social
identities of crowd members.We have shown
how the behaviour of crowd members, their
notions of right and wrong, and the ways
in which they will react to police interventions
are all a function of the beliefs and
values associated with these social identities.
Knowledge-based public order policing therefore
starts with information about these social
Understanding of social identities is a necessary
condition for following the second of our
guidelines. That is, the primary focus of police
strategies during crowd events should be to
maximise the facilitation of crowd aims.While
some groups within the crowd may intend to
act in ways that the police cannot permit, and
members of these groups may be prepared to
confront the police in order to achieve their
aims, the majority of participants generally
will identify with groups that have entirely
legal aims and intentions. By facilitating these,
the police will not only avoid violence from
these participants, they will also gain their
cooperation in dealing with the minority of
others. But this only becomes possible where
there is information which allows the police to
understand the priorities of these groups and
to devise practices which will allow legal aims
to be met.
There will be times when such an approach
will not be possible for practical reasons; that
is, one cannot facilitate the legal aims of some
groups without thereby facilitating the illegitimate
aims of others. For instance, to use
an example that we will elaborate on later,
it may be that the majority of demonstrators
in an anti-globalisation event have the aim
of voicing opposition to the actions of financial
institutions. However, to allow them to
march past their headquarters could make it
easier for those factions that want to disrupt,
damage or even destroy the building. In such
cases, however, the police can still work with
the organisers and participants in order (1) to
clarify that the police are seeking to facilitate
legitimate crowd action; (2) that the reasons
Downloaded from by guest on January 20, 2011
410 Policing Article S. Reicher et al.
for policing constraints lie in the illegitimate
aims of certain participants; (3) to explore
ways in which the police can facilitate alternative
ways in which legitimate aims can be
Overall, the aim of these various procedures
is to reach a position where themajority
of crowd members do not react to police
presence as something which impedes them
but rather as something which enables them.
More technically, the aim is to shape interactions
between police and crowd in such a
way as to lead peaceful crowd members categorise
themselves along with the police and in
opposition to violent factions rather than categorising
themselves along with violent factions
against the police.
The third of our guidelines concerns the
centrality of communication with crowd
members. Take the case of the Notting Hill
Carnival, where the police routinely cordon
off certain roads as overspill areas in cases of
emergency. Yet participants had no knowledge
of this reasoning and, for them, the
cordons reflected police control on a day when
the streets belonged to ‘the people’. Hence
the cordons became potential flashpoints. The
general lesson to be taken from this is that
actions taken by the police for the interests of
the crowd will be ineffective, or even counterproductive
unless they are perceived as such
by the participants themselves. This can only
be achieved through a comprehensive communication
Prior to events, it is important to plan strategies
along with the event organisers which
clarify collective aims and address how they
can best be facilitated. Before, or at the start
of events, these agreements need to be communicated
to all participants through use of
the net, through leaflets or else through visual
and aural communications systems. Here, the
police can play a part by providing facilities
to organisers that allow them to communicate
betterwith their own constituency. During the
events, it is particularly critical to develop a
means of addressing crowd members so that,
if unexpected events arise, organisers and the
police can explain what they are doing and
how it is tied to the agreed aims of the event.
Uncertainty always provides a space in which
those drawing on historical distrust of the
police can gain influence.
Our fourth and final guideline overarches
all that we have said so far. That is, in every
aspect of public order policing, it is critical to
maintain a differentiated approach with the
crowd. As we have stressed repeatedly, crowds
do not typically start off as homogenous entities
and they are only likely to act as such if
one treats them that way. The key to policing
is to treat the participants with respect and,
where some of theminitiate conflict, to ensure
that the response does not drag the others in.
Amongst other things, this has implications
for the development of new tactics sincemany
existing techniques treat all crowd members
equally. It has implications for technology,
since certain types of equipment are too inaccurate
to separate the perpetrator from the
bystander. It even has implications for clothing
since, seen through a scuffed Perspex visor,
the faces of the crowd merge into one and
make targeted reactions impossible. In short,
the guidelines we have outlined here—and
differentiation in particular—are not mere
‘add-ons’ which can be tacked onto existing
practices. Rather, they represent an alternative
perspective which needs to be taken into
account in each and every decision that is
made about policing crowds.
Downloaded from by guest on January 20, 2011
Knowledge-Based Public Order Policing Article Policing 411
Let us give some concrete examples of
how our approach is beginning to be applied
in practice and how, we believe, the benefits
(in terms of limiting disorder, reducing
officer deployment and improving policecommunity
relations) are sufficiently great to
exceed any costs.
Applying the guidelines—examples of
knowledge-based policing in practice
We provide two examples. One relates to a
very specific issue on tactics. The other is
muchmore extended and shows the approach
being employed to cover all aspects of a series
of high profile crowd events.
Example 1:
Developing differentiated tactics of
containment. In recent years, one of the
fastest growing forms of public protest in the
United Kingdom and elsewhere has been associated
with the anti-globalisation movement.
Specifically, since the late-1990s there have
been major annual demonstrations in London.
These protests pose a number of novel
challenges to the police. Unlike traditional
marches which tended to have clear leadership
and organisation and a specified aim,
the anti-globalisation protests bring together
a loose coalition of multiple groups with multiple
aims, no clear leadership and multiple
targets. In the middle of London, virtually
any building, could be seen as a target to at
least some in the crowd. One aspect of the
police response was the use of corralling tactics,
whereby protestors would be surrounded
by the police and would not be allowed
to leave. However the tactic became highly
controversial, especially after the May Day
anti-globalisation protests of 2001, both for
legal reasons and because of the anger that it
Anger stemmed precisely from the undifferentiated
nature of corralling. People would be
corralled simply for being part of the crowd
irrespective of who they were or what they
had done. This led to considerable concern at
the palpably innocent being treated as guilty.
This, in turn, gave rise not only to a shared
experience amongst crowd members but also a
shared sense of police illegitimacy—precisely
the factors that we have identified above as
leading to conflict escalation. Itmay have been
that, in the event itself, participants may have
felt sufficiently disempowered as to refrain
from expressing their anger. However, the
likelihood that they would conflict with the
police in future events would be increased.
Accordingly, we were asked by the Metropolitan
Police to consider how to develop the
corralling tactic (Cronin, 2002; Cronin and
Reicher, 2002).
We stressed, first, the need for officers to
understand the meaning of their tactic from
the perspective of the participants. In particular,
the anger of participants should not be
dismissed simply as reflecting a prior hostility
to the police. Rather, officers need to consider
how they might be producing hostility
in those who started off being sympathetic
towards them. Next, we stressed that, if crowd
members had to be contained out of fear that
some amongst them might be violent, it was
critical tocommunicate to the people as towhy
they were being contained and how this was
necessitated by minority actions. Part of this
may involve the development of new communications
technologies such as high-powered
mobile loudspeaker systems and giant LCD
screens. Third, procedures of selective filtering
should be developed for enabling those
Downloaded from by guest on January 20, 2011
412 Policing Article S. Reicher et al.
with specific needs to exit the containment
area—and this should also be communicated
to the crowd. Moreover, it should also
be stressed that conflict within the containment
area would disrupt the selective filtering
process and hence act against the interests
of crowd members. Fourth, once those in
need had been allowed to leave, it should
be stressed to the remaining crowd that the
police also wish to let them proceed as well,
but that this could only occur under conditions
that will prevent some amongst them
from causing violence. These conditionsmight
include the removal of clothing that obscures
individual identity, abandoning placards, bottles
and other objects that could be used as
weapons. This advice has been taken on board
by the Metropolitan police and we are told
through personal communication that it has
been applied on a number of occasions to
considerable effect.
Example 2:
Graded policing at the Euro 2004.
By far the most systematic application of
our approach occurred at the 2004 European
Football Championships in Portugal (Euro
2004). This work was based on a long-standing
programme of research on police-crowd interactions
by two of the authors,Adang and Stott,
which had already had an impact on policing
practices (Adang, 1991, 1998; Adang and
Cuvelier, 2001; Stott and Adang, 2003). These
researchersworkedwith the Portuguese Public
Security Police (Pol´ıcia de Seguranc¸a P´ ublica
(PSP)) to produce a model of good practice
in policing football matches. The model was
implemented in all the areas under the PSP
control (which covers all the major cities in
Portugal and seven of the ten tournament
venues—the other three, in the South, were
under the control of the Guarda Nacional
Republicana (GNR). It was evaluated by a
team of sixteen observerswho made over eighteen
hundred structured observations during
themajor gatherings of fans in the cities where
the matches were held. In addition, English
fans were surveyed for their perceptions, feelings
and behaviours (for a full report of the
research, see Stott et al., 2006; see also Stott
et al., 2007a,b; Stott and Pearson, in press).
The core feature of the model lay in the
notion of graded policing. That is, four levels
of policing intervention were developed
with the aim of creating a positive and close
relationship with crowd members, but also
of monitoring incipient signs of disorder.
This would then allow early, appropriate and
targeted interventions before conflict could
escalate to a level where only draconian measures
would suffice.
Level 1 was characterised by officers in normal
uniform, working in pairs spread evenly
throughout the crowd within the relevant geographical
location—not merely remaining at
the edges.Their primary functionwas to establish
an enabling police presence. Officers were
specifically trained to be friendly, open and
approachable. They would interact with the
crowd members and generally support the
aim of Euro 2004 as a ‘carnival of football’.
At the same time, the presence (and acceptance)
of these officers in the crowd allowed
them to spot signs of tension and incipient
conflict (such as verbal abuse against rival
fans). They could therefore respond quickly
to minor incidents of emergent disorder and
ensure that they targeted only those individuals
who were actually being disorderly without
having impact on others in the crowd. This
tactic therefore gave the police the option to
intervene in a low-keymanner during the early
Downloaded from by guest on January 20, 2011
Knowledge-Based Public Order Policing Article Policing 413
stages of the emergent disorder often without
others in the crowd even noticing that an arrest
had been made.
Where disorder endured or escalated, policing
shifted to level 2. This involved larger
groups of officers moving in, still wearing
standard uniforms. Their remit was to communicate
with fans in a non-confrontational
manner, to reassert shared norms concerning
the limits of acceptable behaviour, and to
highlight breaches of those norms and the consequences
that would flow from them. Should
this fail, the intervention would shift up to
level 3. Officers would don protective equipment
and draw batons, but always seeking to
target their actions as precisely as possible.
If this was still insufficient, then the PSP’s
riot squads, the Corpo de Intervenc¸˜ao, in full
protective equipment and with water cannon
were always ready at the fourth tactical level.
The approach proved to be highly successful.
Overall, the police maintained low
visibility and although there were large numbers
of riot police around but out of sight,
interventions by these units were very rare.
This reflected a very low level of conflict. In
only 0.2% of our observations was there any
sign of violent behaviour, and even where
there was violence it involved very few people.
By contrast there were a number of notable
instances of ‘self-policing’ amongst fans at critical
moments. For example, just prior to one
of England’smatches, as one groupof England
fans attempted to assault the police, other England
fans confronted the group and actively
prevented the attack from taking place.
This positive picture is corroborated by
the official figures. There were no recorded
instances of major collective disorder during
Euro 2004 in the areas under PSP jurisdiction
and there was only one arrest of an England fan
for a violence-related offence. This figure is in
contrast to the 965 arrests of England fans during
the earlier European championships held
in 2000. However, perhaps the most telling
comparison would have been with those areas
under the control of the other Portuguese
police force, the GNR, during Euro 2004. The
GNR employed a more traditional ‘high profile’
approach to crowd policing. They either
avoided close contact with the crowd or else
intervened in force with full riot gear. This
resulted in two major ‘riots’ during which 52
England fans were arrested.
Taken together these findings suggest, first,
that a positive attitude and facilitative relationship
to the crowd is effective in reducing levels
of disorder; second, that this relationship is
effective in promoting self-policing amongst
fans; third, that the capacity to mount early,
low impact and targeted interventions allows
the police tomanage the low-level dynamics of
emergent disorder; fourth, that this capability
in turn decreases the likelihood of having to
use more forceful and inherently indiscriminate
tactical options (e.g. riot squads/PSUs in
protective equipment).
These benefits are substantial, but there
is more. We have argued that police-crowd
interactions do not only affect the immediate
outcomes of collective events. They also determine
themore enduring relationships between
the police and the groups in the crowd. These
claims are supported from the survey data
with England fans. Respondents described PSP
policing in highly positive terms. They saw the
PSP as behaving legitimately towards them
and, in this context, they did not perceive
their own group (‘England fans’) as being in
opposition to either the fans of other countries
or the police. Rather, they saw themselves
Downloaded from by guest on January 20, 2011
414 Policing Article S. Reicher et al.
in opposition to those ‘hooligan’ fans who
supported violence against such targets.
Critically, however, these perceptions were
not something that pre-dated the tournament
and which explain the peaceful nature of the
crowds in PSP areas. Rather they emerged
during the tournament as a function, we suggest,
of the PSP approach. Thus, before Euro
2004, being an England fan meant distancing
yourself from the police (in statistical terms,
there was a strong and significant negative correlation
between ingroup identification and
perceived similarity to the PSP). However,
after the tournament this relationship was
reversed and being an England fan meant
associating with the police (that is, there was
a strong and significant positive correlation
between ingroup identification and perceived
similarity to the PSP).
At the start of this paper, we argued that
crowds should not be seen as an inherent threat
and that crowd events should not be seen just
as a problem but also as an opportunity. We
have shown how violence derives from interactions—
notably between crowd and police.
Using this knowledge, we have outlined, both
in principle and in practice, an approach to
crowd policing that can produce harmony
both within and beyond crowd events. That is,
as illustrated in Euro 2004, we have developed
a knowledge- based public order policing that
can profitably exploit the opportunities inherent
in crowd events. Against a background
where, for decades, English fan violence has
been seen as an intractable problem, we have
shown not only that fans can be peaceful but
that they can be recruited as allies in subduing
The obvious question is that, even if our
claims over Euro 2004 are accepted, how far
can the Portuguese example be extended to
other settings? At one level, we would readily
join those who counsel against simplistic
overgeneralisation. It would be a serious misunderstanding
of our argument to suggest that
the tactics used in Portugal can be applied
mechanically to other crowds. Our whole
position, as encapsulated in the first of our
guidelines above regarding information and
intelligence, is that the way one treats crowds
must be based on knowledge of the specific
social identities of the groups involved and
hence of their aims, their understandings and
their notions of legitimacy. Such questions as
‘should we facilitate crowd goals and if so
how?’ or, ‘what forms of action will be seen
as violating crowd rights and hence increase
conflict?’ can only be answered in relation to
such knowledge.
In sum there is no such thing as ‘one size fits
all’ public order policing. The specifics must
always be tailored to the given event. What
our approach provides is a means of asking
the questions from which these specifics can
be developed. That was what was done in Euro
2004 and if the same process is followed, there
is no reason why crowd policing could not
be just as effective in turning around seemingly
intractable tensions with other alienated
groups in our society.
ACPO. 2003. Manual of Guidance on Keeping the Police,
Bramshill: National Police Training.
Adang, O. M. J. 1991. ‘‘Gedrag van mensen Tijdens Rellen.’’
In Defares, P. B., and van der Ploeg, J. D. (eds), Agressie,
Assen: Van Gorcum.
1998. Hooligans, Autonomen, Agenten. Geweld en
politie-optreden in Relsituaties, Alphen aan den Rijn:
Downloaded from by guest on January 20, 2011
Knowledge-Based Public Order Policing Article Policing 415
Adang, O. M. J., and Cuvelier, C. 2001. Policing Euro
2000. International Police co-operation, Information
Management and PoliceDeployment, Ubbergen: Tandem
Adang, O. M. J., and Standaar, H. 1993. Jongerendemonstratie
DenHaag, 8 mei 1993,Hoogerheide: PIOV.
Allport, F. 1924. Social Psychology, Boston,MA: Houghton
Cronin, P. D. J. 2001. The Psychology of Crowd Policing, PhD
thesis, University of St. Andrews.
Cronin, P. 2002. Review of the May Day Protests in London,
2001. Briefing to the Metropolitan Police Public Order
Committee. Area 1 Police HQ, Cannon Row Police
Cronin, P., and Reicher, S. D. 2002. Review of the May
Day Protests in London, 2001. Unpublished Confidential
Report to the Metropolitan Police.
Davis, N. Z. 1978. ‘‘The Rites of Violence: Religious Riot in
Sixteenth Century France.’’ Past and Present 59: 51–91.
Drury, J., and Reicher, S. D. 1999. ‘‘The Intergroup
Dynamics of Collective Empowerment: Substantiating
the Social Identity Model of Crowd Behaviour.’’ Group
Processes and Intergroup Relations 2: 381–402.
Drury, J., and Reicher, S. 2000. ‘‘Collective Action and
Social Change:The Emergence ofNew Social Identities.’’
British Journal of Social Psychology 39: 579–604.
Drury, J., and Reicher, S. D. 2005. ‘‘Explaining Enduring
Empowerment: A Comparative study of Collective
action and Psychological Outcomes.’’ European Journal
of Social Psychology 35: 35–58.
Drury, J., Reicher, S. D., and Stott, C. 2003. ‘‘Transforming
the Boundaries of Collective Identity: from the ‘Local’
Anti-Road Campaign to ‘Global’ Resistance.’’ Social
Movement Studies 2: 191–212.
Le Bon, G. 1895/1947 The Crowd: A Study of the Popular
Mind, London: Ernest Benn.
Levine, M., Prosser, A., Evans, D., and Reicher, S. D.
2005. ‘‘Identity and Emergency Intervention: How
Social Group Membership and Inclusiveness of Group
Boundaries Shapes Helping Behavior.’’ Personality and
Social Psychology Bulletin 31: 443–453.
Martin, E.D. 1920 The Behaviour of Crowds, London: Ernest
Reddy, W. M. 1977. ‘‘The Textile Trade and the Language
of the Crowd at Rouen 1752–1871.’’ Past and Present
74: 62–89.
Reicher, S. D. 1982. ‘‘The Determination of Collective
Behaviour.’’ In Tajfel, H. (ed.), Social Identity and
Intergroup Relations, Cambridge: Cambridge University
Press and Paris: Maison des Sciences de l’Homme.
1984. ‘‘The St. Pauls Riot: An Explanation of the
Limits of Crowd Behaviour in terms of a Social Identity
Model.’’ European Journal of Social Psychology 14: 1–21.
1987. ‘‘Crowd Behaviour as Social Action.’’ In
Turner, J., Hogg, M., Oakes, P., Reicher, S. and
Wetherell, M. (eds), Rediscovering the Social Group,
Oxford: Blackwell.
1996. ‘‘The Battle ofWestminster’: 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:
2001. ‘‘The Psychology of Crowd Dynamics.’’ In
Hogg, M. and Tindale, S. (eds), Blackwell Handbook of
Social Psychology: Group Processes, Oxford: Blackwell.
Reicher, S. D., Stott, C., Cronin, P., and Adang, O. 2004. ‘‘A
new Approach to Crowd Psychology and Public Order
Policing.’’ Policing 27: 558–572.
Robins, D. 1984. We Hate Humans, Penguin:
Stott, C., Adang, O., Livingstone, A., and Schreiber, M.
2006. Policing, Crowd Dynamics and Public Order at
Euro 2004 Unpublished report to the Home Office.
2007a. ‘‘Variability in the Collective Behaviour of
England fans at Euro 2004: Public order Policing, Social
Identity, Intergroup Dynamics and Social Change.’’
European Journal of Social Psychology 37: 75–100.
2007b. Policing, Crowd Dynamics and Public
Order at Euro 2004, Manuscript submitted to
‘Psychology, Public Policy and Law’.
Stott, C. J., and Adang, O. 2003. Policing Football in the
European Union: understanding and managing risk at
football matches with an international dimension. Report
to the Public Security Police. Portugal.
Stott, C., and Drury, J. 2000. ‘‘Crowds, Context and Identity:
Dynamic Categorization Processes in the ‘Poll Tax
Riot’.’’ Human Relations 53: 247–273.
Stott, C., Hutchison, P., and Drury, J. 2001. ‘‘‘Hooligans’
abroad? Inter-group Dynamics, Social Identity and
Participation in Collective ‘disorder’ at the 1998 World
Cup Finals.’’ British Journal of Social Psychology 40:
Stott, C., and Peason, G. in press ‘‘Football.’’ ‘Hooliganism’
and the War on the ‘English Disease’, London: Pennant
Stott, C., and Reicher, S. D. 1998. ‘‘How Conflict Escalates:
the Inter-Group Dynamics of Collective Football Crowd
‘Violence’.’’ Sociology 32: 353–377.
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, Monterey,
CA: Brooks-Cole.
Thompson, E. P. 1971. ‘‘The Moral Economy of the English
Crowd in the Eighteenth Century.’’ Past and Present 50:
Turner, J. C. 1982. ‘‘Towards a Cognitive Redefinition of
the Social Group.’’ In Tajfel, H. (ed.), Social Identity and
Intergroup Relations, Cambridge: Cambridge University
Tyler, T. R., and Blader, S. L. 2000. Cooperation in
Groups. Procedural Justice, Social Identity, and Behavioral
Engagement, Philadelphia, PA: Psychology Press.
Downloaded from by guest on January 20, 2011

Page counter added 29/3/13


Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.