9781405124003_4_012.

qxd 10/31/07 3:11 PM Page 244

The Psychology

12 of Groups: Basic
Principles
Bernard A. Nijstad and Daan van Knippenberg

K E Y CO N C E P TS

cohesion
commitment
entitativity
expectation states theory
group socialization
initiation
interpersonal cohesion
need to belong
role
role transition
socio-emotional behaviour
speaking hierarchy
staffing level
status
task behaviour
task cohesion
transactive memory
9781405124003_4_012.qxd 10/31/07 3:11 PM Page 245

CHAPTER OUTLINE

Groups are pervasive in social life. In this chapter, we discuss why people form and join groups, and
what types of groups can be distinguished. We further discuss three levels of analysis. At the indi-
vidual level, we discuss the (changing) relations between the group and its members. At the group
level, we discuss group development, group structure (status and roles) and group norms. At the
intergroup level, we discuss how the (intergroup) context shapes the behaviour of group members
and the structure of groups.

Introduction
Imagine you’re spending a weekend in Amsterdam. You enter a subway station, which is quite
crowded. From the way people are dressed – many are wearing red and white Ajax shirts – you
infer that they must be Ajax fans going to support their football team. These fans show remarkable
behaviour: they sing and shout in ways they would not normally behave in public. Yet, most of
them are adults (and not all are drunk), and they only show this behaviour when there is an Ajax
match. The most striking aspect of their behaviour is that the fans behave so similarly. However,

Plate 12.1 These fans share membership of a social group: they are all Ajax football supporters.
9781405124003_4_012.qxd 10/31/07 3:11 PM Page 246

246 CHAPTER 12 THE PSYCHOLOGY OF GROUPS: BASIC PRINCIPLES

many of them do not even know each other, and their behaviour their behaviour. At the broader level, one could argue that these
is quite out of the ordinary: normally, people would not sing and fans only show this behaviour because of the context: there is
shout in a subway. going to be a football match in which Ajax will play against
The only reason these fans behave so similarly is that they share another team. Indeed, if there had been no such match, the fans
membership of a social group: they are all Ajax fans. In this chap- would behave quite differently.
ter, we argue that in order to understand their behaviour and beha- In this chapter, we use this three-level framework to discuss
viour in other groups, we need to consider three levels of analysis: some basic characteristics of groups and some basic processes
the individual level, the group level and the wider context in which in groups. We first examine the issues of what a group is, why
groups are situated. At the individual level, all Ajax fans in the sub- people form or join groups, and what types of groups can be dis-
way are individually aware of their group membership (being an tinguished. We then move on to the individual level and discuss
Ajax fan) and of the fact that the other people in the subway are how individuals join groups and how their group membership
Ajax fans as well. At the group level, the fact that their behaviour develops over time. We then consider the group level, as we discuss
is so similar indicates that it cannot be caused by idiosyncratic ten- group development and group structure. Finally, we discuss the
dencies of individual Ajax fans, such as their individual personali- (intergroup) context in which groups exist and how this context
ties. Rather, there is something ‘groupy’ going on which guides affects processes that occur in groups.

THE PHENOMENOLOGY that one can only talk about groups to the extent that there are
people who do not belong to the group, although they belong to
other groups.
OF GROUPS
Why groups?
What is a group?
Why do people form, join and distinguish groups? Why do humans form, join and distinguish groups? Several theo-
What kinds of groups can be distinguished, and what are their retical perspectives can be applied to answer that question. We
characteristics? will discuss three: a sociobiological, a cognitive and a utilitarian
What is group entitativity, and what contributes to perceptions of perspective (also see Baron & Kerr, 2003). These three perspect-
entitativity? ives are complementary rather than mutually exclusive.
Following Darwin’s evolution theory, the sociobiological
perspective (e.g., Bowlby, 1958) emphasizes the adaptive value of
Defining groups forming groups. Forming groups enables humans (and other social
animals) to deal more effectively with enemies or predators, and
Groups are everywhere: we see groups of friends in a bar, groups allows cooperation in such areas as raising children, farming
of colleagues in an organization, groups of fans in a stadium. But or hunting. Especially earlier in our evolutionary history, when
what exactly do we mean by the word ‘group?’ Many authors have food was often scarce and enemies and predators were dangerous,
suggested different ingredients towards a definition of groups. forming groups had a significant advantage. A predisposition to
Lewin (1948) suggested that common fate is critical: people are form groups increased the chances of survival of the individual
a group to the extent that they experience similar outcomes. and, through the evolutionary principle of natural selection, this
Sherif and Sherif (1969) proposed that some form of social struc- predisposition was selected and passed on to later generations.
ture (status or role differentiation, e.g., a leadership role) is essen- This human predisposition
tial, because otherwise the ‘group’ would just be a loose collection to form and maintain stable,
need to belong the fundamental and
of individuals. Bales (1950) stressed the importance of face-to-face strong and positive relation- innate human motivation to form positive,
interaction. We suggest a broader definition of groups: following ships with others is called the strong and stable bonds with others
Tajfel (1981), we argue that a group exists when two or more indi- need to belong (Baumeister &
viduals define themselves as members of a group. Leary, 1995). Baumeister and
A few things should be noted. First, many different groups Leary argued that this human need is innate and universal. Indeed,
would fit this definition, including religious groups (Christians), evidence indicates that the tendency to form groups is found across
national groups (the British), organizational groups (the psycho- all cultures and situations, suggesting that this tendency is evolu-
logy department) and friendship groups (a student society). Second, tionarily ‘built in’.
it is subjective and does not include any ‘objective’ characteristics According to the cognitive perspective, groups help us to under-
of groups, such as common fate or face-to-face interaction. Rather, stand our world. Social comparison theory (Festinger, 1954; see
it emphasizes common identity: sharing the view with others that Chapters 5 and 10, this volume) argues that people want to hold
you belong to the same group. Third, it is important to recognize accurate views of the world. They can do this by validating their
9781405124003_4_012.qxd 10/31/07 3:11 PM Page 247

THE PHENOMENOLOGY OF GROUPS 247

beliefs either against ‘physical reality’ (e.g., ‘Will this glass crack Social exchange theory argues that social relations involve costs
if I hit it with a hammer?’) or against ‘social reality’ (e.g., as well as benefits, and as long as the benefits exceed the costs the
‘I like this new music; I wonder what my friends think about it?’). relation will yield a ‘profit’. There is much evidence that people
People turn to others especially for beliefs for which there is no are unhappy about relations if they feel that they invest more in
physical reality (e.g., preferences). Building on these ideas, social them (e.g., time) than they get back (e.g., approval) (e.g., Le &
identity theory (e.g., Tajfel & Turner, 1986; see Chapter 5, this Agnew, 2003; see also Chapter 10). Furthermore, satisfaction with
volume) and self-categorization theory (Turner, Hogg, Oakes, an exchange relationship depends on the degree to which alter-
Reicher & Wetherell, 1987; see Chapters 5 and 11, this volume) native relationships exist that yield more profit. Thus, people join
argue that people define themselves and others partly in terms groups because they derive benefits from their group membership.
of group membership. The theory argues that seeing oneself People may leave groups (if possible) when they are unhappy
and others as members of groups helps to reduce uncertainty and about the benefits relative to the costs of group membership, or
make sense of our world (e.g., Hogg & Abrams, 1993). Being a when alternative groups exist that have a better cost–benefit ratio
member of a group often provides guidelines for the way we (also see Rusbult & Farrell, 1983). In general, people will leave
should behave and think. If you think about the Ajax football groups when better alternatives are available, including the option
fans we started this chapter with, their behaviour is clearly guided of being alone.
by their group membership and the behaviours thought to be
appropriate for that group (see our later discussion of group
norms). Further, seeing other people as members of certain groups
helps to interpret their behaviour: knowing that the people in the
Types of groups and group
subway are Ajax fans makes it much easier to understand what entitativity
is going on.
A utilitarian perspective argues that people derive benefits from As we noted earlier, our definition of groups is relatively broad
groups. Social exchange theory (e.g., Thibaut & Kelley, 1959; see and many types of groups may be included. However, there are
Chapter 10, this volume) argues that social relations (including different types of groups with different characteristics. Further,
those within groups) help to fulfil the individual’s needs and often some groups seem more ‘groupy’ than other groups, a pheno-
take the form of exchange processes. These exchanges might menon often referred to as the
involve material goods (e.g., borrowing a tool, selling your car) or entitativity of groups: the de-
interpersonal helping (helping a friend move house), but also psy- gree to which a collection of entitativity the degree to which a
collection of persons is perceived as being
chological ‘goods’ such as love, friendship or approval. Enduring persons is perceived as being bonded together in a coherent unit
exchange relations between two or more people are more effect- bonded together in a coher-
ively organized when people form a (more or less stable) group. ent unit (Campbell, 1958).
Thus, groups exist because they facilitate mutually beneficial social So, what different types of groups can we distinguish? Lickel
exchange. et al. (2000) wondered whether people spontaneously distinguish
between different types of groups. They provided their participants
(American and Polish students) with a sample of 40 different
groups, such as ‘members of a family’, ‘blacks’, ‘members of a jury’
and ‘people in line at a bank’. Participants had to rate these differ-
ent groups on eight dimensions: importance of group members
to each other, common goals and common outcomes for group
members, degree of interaction among members, size, duration,
PIONEER permeability (how easy it is to join or leave the group) and
similarity among group members. The groups were also rated
John Walter Thibaut (1917–1986) was born in Marion, Ohio. on the degree to which the group really was a group (group enti-
He studied philosophy at the University of North Carolina. tativity). After they had done the ratings, participants were asked
During World War II, he came into contact with psychology to sort the 40 groups into different categories using their own
when he was assigned to the Aviation Psychology Program. individual criteria, including as many or as few categories as they
In 1946, he moved to the Massachusetts Institute of Techno- wanted.
logy to study with Kurt Lewin. After Lewin’s death in 1947, Lickel et al. (2000) found that some of their 40 groups were con-
Thibaut moved to the University of Michigan where he sistently sorted into one common category, whereas other groups
received his PhD. His subsequent career took him to Boston were consistently sorted into other categories. Further, groups that
University, Harvard University and back to the University of were sorted into the same category were also rated similarly on
North Carolina. Thibaut is best known for his 1959 book (co- the eight dimensions. Lickel et al. identified four types of groups:
authored with Harold Kelley) The Social Psychology of Groups. intimacy groups, task groups, social categories and loose associ-
In that book, Thibaut and Kelley laid out the foundations of ations. In Table 12.1 we give a summary of their findings and some
social exchange theory, arguing that social relations take the examples of the different types of groups. As can be seen in the
form of social exchange processes. table, the types of groups differed along the different dimen-
sions. For example, intimacy groups (e.g., a family) were seen as
9781405124003_4_012.qxd 10/31/07 3:11 PM Page 248

248 CHAPTER 12 THE PSYCHOLOGY OF GROUPS: BASIC PRINCIPLES

et al. (2000) also considered which of their eight group character-

istics best predicted group entitativity. They found that the single
most important predictor was interaction among group members:
higher levels of interaction were associated with higher entitativ-
ity. The other characteristics also contributed to entitativity:
importance, common goals and outcomes, group member sim-
ilarity and duration showed a positive relation (the higher the
importance, common goals, etc., the higher the perceived entita-
tivity), whereas group size and permeability showed a weak
negative relation (larger groups and highly permeable groups
were rated lower in entitativity). Note that some of the possible
components of a definition of groups that we described earlier
(common fate, face-to-face interaction) were positively associated
with perceived group entitativity: they indeed make groups more
‘groupy’.
Plate 12.2 Intimacy groups, e.g. a family, are seen as high in
entitativity.

SUMMARY

important, with high levels of interaction, common goals and out- Forming, joining and distinguishing groups has a number
comes, a high degree of similarity, fairly small, of long duration of advantages: groups help us to make sense of our world
and low permeability. Social categories (e.g., women), in contrast, and to coordinate more effectively mutually beneficial
were rated low on importance of members to each other, with low social exchange. The tendency to form groups probably is
levels of interaction, common goals and outcomes, and member evolutionarily built in, as groups are found everywhere.
similarity, and were rated to be large, of long duration and low However, not every type of group is equally important
in permeability. or ‘groupy’: especially intimacy groups and task groups are
With regard to group entitativity, intimacy groups and task seen to be important and high in entitativity, while social
groups were seen as high in entitativity, loose associations as low, categories and loose associations are less so.
and social categories occupied an intermediate position. Lickel

Table 12.1 Characteristics of different types of groups (based on Lickel et al., 2000)

Characteristic Type of group (examples)

Intimacy group (family Task group (jury Social category Loose association (people
members, friends, members, cast of (women, blacks, at a bus stop, at the cinema,
romantic partners) a play, sports team) Americans) living in same area)

Entitativity High High Moderate Low

Interaction High Moderate/High Low Low

Importance High Moderate/High Low Low

Common goals High Moderate/High Low Low

Common outcomes High Moderate/High Low Low

Similarity High Moderate Low Low

Duration Long Moderate Long Short

Permeability Low Moderate Low High

Size Small Small Large Moderate

9781405124003_4_012.qxd 10/31/07 3:11 PM Page 249

INDIVIDUALS IN GROUPS: THE INDIVIDUAL LEVEL OF ANALYSIS 249

INDIVIDUALS IN GROUPS: one stage to the next involves

a role transition. Thus, mov-
role transition a change in the relation
between a group member and a group
ing from prospective member
THE INDIVIDUAL LEVEL (the stage of investigation) to commitment the degree to which a group
member identifies with the group and its
new member (the stage of
OF ANALYSIS socialization) involves the role
transition of entry. Further
goals and wishes to maintain group
membership

role transitions are accept-

What stages of group socialization can be distinguished? ance (from new member to full member), divergence (from full
What are role transitions and what determines their occurrence? member to marginal member) and exit (from marginal member to
How does dissonance theory explain severity of initiation? ex-member). As can be seen in Figure 12.1, the five different stages
differ in the degree of commitment of the individual to the group,
In this section we consider the individual within the group: in other words, the degree to which a group member identifies
that is, we focus on the individual level of analysis. In particular, with the group and its goals and wishes to maintain group mem-
we discuss Moreland and bership. Commitment increases gradually as people become full
Levine’s (1982) model of members, after which it decreases towards the point that indi-
group socialization the efforts of the
group socialization, which is viduals wish to leave the group.
group to assimilate new members to
existing group norms and practices depicted in Figure 12.1. The Role transitions occur as a result of evaluation processes in
model is applicable to groups which the group and the individual evaluate one another’s ‘re-
that exist for comparatively wardingness’, or the extent to which the group is rewarding for
long periods of time and have direct interaction between mem- the member and the member is valued by the group. When the
bers, but that experience changes in membership. Examples would group is rewarding for members, they will try to enter the group
include a sports team, a team within an organization or a student or maintain group membership (i.e., feel commitment). Similarly,
society (i.e., many intimacy groups and task groups). when a group values a (prospective) member, the group will en-
Moreland and Levine’s model distinguishes five stages of group courage the person to become or stay a member of the group (i.e.,
membership: investigation, socialization, maintenance, resocial- the group is committed to the member). This is related to our ear-
ization and remembrance. According to the model, moving from lier discussion of social exchange processes and the benefits people

Time
Prospective New Full Marginal
member member member member Ex-member
ACCEPTANCE
Acceptance
criterion

DIVERGENCE Divergence
criterion
Commitment

ENTRY Entry criterion

EXIT Exit criterion

INVESTIGATION SOCIALIZATION MAINTENANCE RESOCIALIZATION REMEMBRANCE

Figure 12.1 The Moreland and Levine (1982) model of group socialization.
9781405124003_4_012.qxd 10/31/07 3:11 PM Page 250

250 CHAPTER 12 THE PSYCHOLOGY OF GROUPS: BASIC PRINCIPLES

can derive from them (e.g., to gain social approval or receive help that the group is not as attractive as initially believed. This would
or material goods). Indeed, according to Moreland and Levine lead to cognitive dissonance: members can no longer maintain that
(1982), commitment is a function of the past, present and expected they had good reasons to undergo the harsh treatment when they
future rewardingness of the group as compared to the rewarding- admit that the group is not so attractive after all. Thus, the mem-
ness of alternative groups. In the remainder of this section, we con- ber will deny that the group is unattractive and will maintain a
sider the five stages shown in Figure 12.1. high level of commitment to the group.
Aronson and Mills (1959) performed an experiment to test this
reasoning. They offered female students the opportunity to join a
Joining a group and group discussion group about sexuality. However, some of the prospect-
ive members first had to undergo the embarrassing experience of
socialization: Becoming a full reading aloud sexually explicit passages, while other prospective
member members did not have to do this. Next, the participants listened
to an actual group discussion that was recorded on tape. This
Investigation In the stage of investigation, groups look for peo- discussion was in fact quite boring and was about the secondary
ple who might make a contribution to the attainment of group sexual behaviour of lower animals. Participants were next asked
goals. Task groups will often search for people who have the re- to rate the attractiveness of the group. In line with the dissonance
quired skills and abilities, whereas intimacy groups will tend to explanation, the women who had to read the embarrassing pas-
emphasize compatibility (e.g., similarity) with the existing mem- sages rated the group more attractive than those who did not.
bership. Prospective members, on the other hand, will look for Lodewijkx and Syroit (1997), however, did not find a positive
groups that may potentially fulfil their needs. For example, when relation between severity of initiation and group liking. They
you have just moved to a new city to start college, you will prob- conducted a field study among the prospective members of the
ably try to identify certain groups that may help to fulfil your sorority mentioned above and found, in fact, that severe initiations
social needs. Thus, you may join a student society, hoping to find decreased the liking for the group. Thus, prospective members of
people with whom you can start a new, positive and stable relation the sorority who rated the initiation as more severe liked the group
(i.e., fulfil your need to belong). less. The reason was that severe initiations led to loneliness and
frustration, and this in turn reduced the liking for the group. What
Entry and initiation When the level of mutual commitment Lodewijkx and Syriot did find was that, during the initiation, pos-
between group and prospective member reaches an entry cri- itive relations developed among prospective members and these
terion, a role transition will occur: entry. Entry is often marked by increased liking for the group.
some ritual or ceremony that makes it clear that the relation Thus, severe initiations do not always increase liking for the
between the group and the (prospective) member has changed. group, as they may lead to loneliness and frustration. In the
In an organization, this may take the form of a welcome speech, Aronson and Mills study, in which the initiation was very brief,
and in social groups it may this probably did not happen. Severe initiations may also have
initiation the role transition of entry into a
be a party. At other times the other functions: they deter potential members who are not eager
group, often accompanied by some ritual entry or initiation ritual can enough to join the group, and prospective members can show
be quite unpleasant and painful their interest in the group by undergoing these harsh treatments
for the prospective member. (Moreland & Levine, 1982).
Lodewijkx and Syroit (1997) studied initiation into a Dutch
sorority (a student society for female students). The novices first Socialization After entry, the stage of socialization begins.
stay in a campsite for a week. Everyone wears a shapeless, sack-like In this stage, new members learn the norms of the group: the
uniform, they are not called by their real names and they have to (unwritten) rules that prescribe the attitudes and behaviours that
undergo physical hardship (hard work and lack of sleep and food). are (or are not) appropriate in the context of the group. In addi-
After a week they return to the city and participate in ‘evening tion, new members may acquire the necessary knowledge and
gatherings’ for a further one and a half weeks. During these gath- skills to function effectively
erings, which are regarded as threatening by the novices, they as a group member (i.e., learn
role the behaviours expected of a person
are often bullied and embarrassed. Then, finally, the inauguration their role in the group: the set with a specific position in the group
ceremony takes place, after which they have a meal with the senior of behaviours associated with
members – the so-called ‘integration party’ – and the initiation is a certain position in the group).
ended. Thus, the group tries to assimilate the member to fit the expecta-
As these severe initiations take place in many different groups tions of the group. However, socialization is a two-way street,
(e.g., the military, some sports teams, student societies), the ques- and the new member may also try to influence the group in such
tion arises as to why groups perform these harsh rituals. Aronson a way that the member’s needs are best met. For example, a new
and Mills (1959) suggested a classic argument. They maintained member may try to change the group’s norms or customs (e.g.,
that severe initiations increase the liking for and commitment to ‘I think that we should meet more often’). Research close-up 12.1
the group. Their argument is based upon cognitive dissonance theory describes a study of newcomer influence.
(Festinger, 1957; see Chapter 7, this volume). Suppose a prospect- During socialization, the commitment of the member towards
ive member has undergone harsh treatment but it later appears the group and the commitment of the group towards the member
9781405124003_4_012.qxd 10/31/07 3:11 PM Page 251

INDIVIDUALS IN GROUPS: THE INDIVIDUAL LEVEL OF ANALYSIS 251

RESEARCH CLOSE-UP 12.1

Conditions under which newcomers can influence monitoring plane characteristics. Some groups could choose
a group their preferred strategy, while others were given no choice. Then
the first trial, lasting 15 minutes, was performed. The groups
Choi, H.S. & Levine, J.M. (2004). Minority influence in work teams:
received either positive or negative feedback about how well
The impact of newcomers. Journal of Experimental Social
they had done in that first trial (the performance manipulation).
Psychology, 40, 273–280.
One of the specialists was then replaced by a newcomer, who
in fact was a confederate of the experimenter. To get acquainted,
the two real participants were allowed to have an electronic chat
Introduction with the newcomer. During this chat, the newcomer proposed
Choi and Levine studied the impact of newcomers on groups. using the other strategy (i.e., the one the group had not used in
In particular, they were interested in the degree to which the the first trial). A second 15 minute trial followed, in which the
other group members would accept a newcomer’s suggestion groups made an assessment whether to stay with their old strat-
to change the way in which the group works. Choi and Levine egy or adopt the newcomer’s suggestion to change.
argued that this would be dependent on task success: when the
group had been successful before the newcomer’s arrival, they Results
would be unmotivated to change their strategy, whereas they
would more likely consider it after failure. Second, they argued Results are shown in Figure 12.2. As predicted, both group
that when the group had chosen their own way of working, they choice and group performance affected the adoption of the
would be less likely to give it up because they would feel more newcomer’s suggestion. After failure and when the initial strat-
committed to it. On the other hand, groups that had been egy had been assigned, groups were more likely to change
assigned a specific way of working would more easily accept the strategies than after success or when they had chosen their
newcomer’s suggestion. initial strategy themselves.

Method Discussion
Participants Choi and Levine conclude that newcomers are not merely
Participants were 141 male undergraduates who took part in passive recipients of influence. Under some conditions, such as
47 three-person groups. failure on the group task, newcomers can have a substantial
influence on the practices of the group. As such, newcomers can
Design and procedure bring about changes and introduce innovations to the group.
Participants performed a task twice. After the first task trial,
one of the group members was replaced by a newcomer, who 100
Failure
suggested changing the way of working. The experimental Success
90
design was a 2 (group performance: failure/success) × 2 (group
Percentage of groups that changed

choice: no choice/choice) factorial. Group performance was 80

manipulated by giving false feedback after the first task trial. Group
70
choice was manipulated by having the groups choose their own
way of working before the first trial or not giving them that choice. 60
The task the groups had to perform was an air-surveillance
50
task. The three group members were seated at different com-
puters. One of them was randomly appointed commander, the 40
other two specialists. The two specialists had to monitor eight
30
characteristics of planes flying through a simulated airspace,
such as airspeed, direction and weapons. They had to pass the 20
information on to the commander, who had to use a formula to
10
integrate the information and assign a threat value to each plane.
Based on the accuracy of that value, the group could earn points. 0
After task training, the first independent variable was intro- No choice Choice
duced. Groups were given a description of two strategies of how
to divide the workload between the two specialists – one Figure 12.2 Effects of performance feedback and group choice
according to the importance and one based on the difficulty of on acceptance of newcomer suggestions (after Choi & Levine, 2004).
9781405124003_4_012.qxd 10/31/07 3:11 PM Page 252

252 CHAPTER 12 THE PSYCHOLOGY OF GROUPS: BASIC PRINCIPLES

will generally increase (except when the new member or the group roles to members in such a way that the group’s goals can be best
is dissatisfied). At a certain point in time (when the acceptance cri- achieved. One of the more important roles within the group is that
terion is reached), the member will no longer be treated as some- of group leader (see Chapter 13, this volume). However, there are
body who needs special attention, the socialization stage is ended often other roles that need to be fulfilled within groups, such as
and the new member is accepted as a full member. The new mem- those of ‘recruiter’ (who identifies and evaluates prospective
bers may gain access to information that was previously hidden, members) and ‘trainer’ (who has a role during socialization of new
join certain informal cliques, and their behaviour is monitored less members). According to the model, the relation between the
strictly. As with entry, there may be some ritual to mark the tran- group and the member will be rewarding and commitment will
sition of acceptance as a full member. A well-known example is remain high to the degree that role negotiations are successful.
the bar mitzvah ceremony for Jewish boys at the age of 13, after Being in a group is more extensively examined in the next section,
which the boy is accepted as a full member of Jewish society in- where we discuss norms, roles and status.
stead of being considered a child.
Being accepted as a full member is easier in some groups than
in others. In part, it depends on the staffing level of the group: Leaving a group: Divergence and exit
the degree to which the actual
number of group members Divergence After a time, group members may lose interest in
staffing level the degree to which the
actual number of group members is similar
is similar to the ideal number the group, for example because they are dissatisfied with their role
to the ideal number of group members of group members. Groups in the group or because they have identified other groups that are
can be overstaffed (have too more rewarding. On the other hand, the commitment of the group
many members) as well as to its members may decline when members fail to live up to group
understaffed (have too few members). One might expect that expectations. For example, members may not perform well in their
understaffed groups will be less demanding of new members (it is role or may violate important group norms. This will lead the
easier to become a full member) than overstaffed groups. group to relabel these members as marginal members or deviates.
Cini, Moreland and Levine (1993) conducted a study among The group might, for instance, no longer give marginal members
93 student groups, including fine arts clubs, social groups and full information, or other group members may exclude marginal
political groups. They held interviews with the president of each members from informal cliques (e.g., they are no longer asked to
group in which they gathered information about the staffing level come along for a drink after work). Often, considerable pressure is
of the groups and about recruitment and socialization practices. exerted on deviates to realign or even to leave the group (espe-
It appeared that both understaffing and overstaffing caused prob- cially if the group is overstaffed).
lems. Understaffing led to a loss of resources (e.g., too few mem- Schachter (1951) experimentally demonstrated the pressure
bers contributing membership fees), poorer group performance that is exerted on deviates. He had groups discuss a delinquency
and fatigue among group members. Overstaffing led to apathy and case. In each of the experimental groups there were confederates
boredom, alienation (i.e., group members felt ‘lost in the crowd’), playing different roles: the ‘mode’ who accommodated to the
and confusion and disorganization. The solution to understaffing, group’s average judgement, the ‘slider’ who initially took an
not surprisingly, was to recruit new members. Consequently, the extreme position but then moved towards the group norm, and
groups that were understaffed were more open: they were less the ‘deviate’ who also took an extreme position but maintained
selective (it was easier to become a new member), and also less it throughout the discussion. Initially, the group discussion was
demanding for new members (it was easier to become a full mem- primarily aimed at the two deviating members (the slider and the
ber). For example, new members were evaluated and expected to deviate) in each group, trying to change their minds. When it
perform special duties less often in understaffed as compared became apparent that the deviates would not change, the groups
to overstaffed groups. Solutions to overstaffing, in contrast, were eventually excluded them, refusing to talk to them and ignoring
to restrict membership, but also to punish deviance from group their contributions (see Figure 12.3).
norms more harshly, in the hope that deviant members would
leave the group.
to mode to slider to deviant
2
communications from

Being in a group: Maintenance and

Average number of

group per minute

1.5
role negotiation 1

After acceptance, the stage of maintenance begins. This stage is 0.5

characterized by high levels of commitment, and for both the
0
member and the group the relation is seen as rewarding (see Figure 5 to 15 15 to 25 25 to 35 35 to 45
12.1). The major way in which groups and members try to increase Time interval (minutes)
the rewardingness of their relationship is through role negotiation.
Thus, the member tries to occupy the role within the group that Figure 12.3 Communications directed towards the mode, slider
best satisfies his or her need, whereas the group tries to appoint and deviant over time (based on Schachter, 1951).
9781405124003_4_012.qxd 10/31/07 3:11 PM Page 253

GROUP DEVELOPMENT AND STRUCTURE: THE GROUP LEVEL OF ANALYSIS 253

Resocialization and exit Divergence might be followed by of revenge. Workplace shootings (e.g., in Kansas City, USA, in
a period of resocialization. In this period, the group might try to 2004), in which employees who had been dismissed shot their boss
persuade marginal members not to leave, or they might try to or former colleagues, are extreme examples. Fortunately, these
accommodate to the wishes of marginal members (e.g., give them incidents are rare.
a different role). Similarly, group members may try to convince the
group not to expel them, and might try to assimilate to the group’s
expectations again. This might result in re-entry to the group when
successful. However, when resocialization fails, group members SUMMARY
may reach an exit criterion and leave the group. As with other role
transitions, this may involve some ritual, such as a goodbye speech Individuals move through different phases of group mem-
or a party. Alternatively, the group may expel the member, which bership (prospective member, new member, full member,
can be quite a painful experience. For example, an employee might marginal member and ex-member). These stages of group
be fired or a church member might be excommunicated. membership differ in the degree of commitment of the
Research has shown that social exclusion from groups has group and the member to each other. Moving from one
enormous negative effects on excluded members. Consider the stage to the next involves a role transition, and role transi-
following situation. You are invited to come to the psychology tions can both be extreme (e.g., severe initiation rituals) and
lab to participate in an experiment and are asked to wait in a wait- have a large impact on members (e.g., after exit).
ing room until the experiment starts. In that room two other
participants are also waiting (they are, in fact, confederates of the
experimenter). One of them has brought a tennis ball and playfully
throws it to the other participant. That participant joins in and
throws the ball to you. For a while, the three of you play this ball-
tossing game. After some time, however, the other participants no
GROUP DEVELOPMENT
longer throw the ball to you, but only to each other, and this goes
on for several minutes. How would you feel? AND STRUCTURE: THE
Williams (2001) reports extensive evidence concerning the
power of social exclusion. Using the ball-tossing game (and other GROUP LEVEL OF
situations), he found that social exclusion produces severe negat-
ive moods and anger, and leads to lower ratings on belongingness ANALYSIS
and self-esteem. Further, Eisenberger, Lieberman and Williams
(2003) found that exclusion quite literally is a form of ‘social pain’.
These researchers had participants play a computerized version of What are the five stages of group development?
the ball-tossing game while lying in an fMRI (functional magnetic What is interaction process analysis, and how is it helpful when
resonance imaging) brain scanner. Using the fMRI scanner, the studying group development and group structure?
researchers could identify which brain areas were active during What are the functions of group norms?
social exclusion. Participants were led to believe that, by pushing How do status and role differences come about?
a button, they could throw a (virtual) ball to another participant,
who could then throw the ball back to them or to a third partici- In the previous section we discussed the (changing) relation of the
pant. In fact, there was only one real participant, and the computer group member with the group. In this section we explore the
was programmed in such a way that this participant received the group level of analysis. First, we discuss how groups themselves
ball nine times, after which the ball was no longer thrown to him can also change over time. Second, groups have certain charac-
or her. While being excluded from the game, an fMRI brain scan teristics, such as norms to govern their behaviour and a group
was made. Eisenberger et al. (2003) found that social exclusion structure, in which certain
activates an area in the brain (the anterior cingulate cortex) that is members have more status status evaluation of a role by the group in
normally activated when a person is in physical pain. Furthermore, than others or in which dif- which a role is contained or defined
the level of activation of that brain area was correlated with par- ferent members occupy dif-
ticipants’ reports of distress. ferent roles in the group. These issues are examined below. It
should be noted that this section is mainly relevant for groups with
Remembrance The last stage of the Moreland and Levine direct (usually face-to-face) interaction.
model is remembrance (see again Figure 12.1). In this stage, the
ex-member and the group retrospectively evaluate each other.
Thus, remaining group members will evaluate the ex-member’s Group development
contributions to the group and will maintain some degree of com-
mitment to the ex-member if these contributions are seen as Some groups are formed for a special reason and end after a cer-
positive. Similarly, ex-members look back on their time with the tain time. Examples include therapy groups, project teams and the
group with either fond or bitter memories. In extreme cases, group of students in a psychology seminar. These groups will gen-
ex-members may even try to destroy their former group in an act erally develop: the interaction patterns among group members
9781405124003_4_012.qxd 10/31/07 3:11 PM Page 254

254 CHAPTER 12 THE PSYCHOLOGY OF GROUPS: BASIC PRINCIPLES

Forming Storming Norming Performing Adjourning

Group members Group members Group members Group members Group members
get to know each resist influence: share a common work together leave the group:
other: high disagreement purpose: high towards their feelings of
uncertainty and high conflict friendship and goal: accomplishment
cohesion performance- or failure,
oriented sometimes grief
relations or relief

Figure 12.4 The five stages of group development (after Tuckman & Jensen, 1977).

change over time. Further, there may be similarities in the way on to the third stage. In the third stage, norming, group members
different groups develop. The basic idea is that every group faces develop close ties. In this stage, the group members come to agree
certain challenges and has certain goals, and these challenges and upon the group’s goals and develop norms that govern group
goals change over time. This, in turn, has consequences for the interaction. Once this has been achieved, the group enters the
way group members interact with each other, as well as for group performing stage. Because group structure and group norms have
performance and the rewardingness of the group to its members. been established, the group’s efforts can be directed towards
Tuckman (1965) and Tuckman and Jensen (1977) introduced a achieving the group’s task. Although it is probably still necessary
classic five-stage model of group development: forming, storming, to engage in behaviours to maintain a positive atmosphere in the
norming, performing and adjourning (see Figure 12.4). In the first group, most activities will be task-related. The final stage of group
stage, when the group is forming, group members feel insecure development is adjourning. When the task has been accomplished
because they do not know each other and do not know what is or is abandoned, the group will end. This might be associated either
expected of them. As a consequence, interactions are usually with feelings of accomplishment or with feelings of disappoint-
polite and inhibited. In this first stage, people get to know each ment (dependent, of course, on task success).
other and develop a shared identity as members of the same group. According to the Tuckman and Jensen (1977) model, the dif-
This might happen at the beginning of a psychology seminar: ferent stages of group life should be characterized by different
students still feel insecure, engage in polite conversation, and the interaction patterns within the group. But how can we establish
atmosphere is quite subdued. Once people have got to know each whether this really is true? To answer that question, it is necessary
other, they enter the second stage (storming). The challenge in the to code group interactions into certain categories and see whether
second stage is to develop a group structure. Here issues of lead- certain types of behaviour are more frequent in the early or the
ership and influence are at stake, and as group members may com- later stages of group life. Probably the best-known coding system
pete about different roles in the group, there may be conflicts and of group interaction is Bales’s (1950) interaction process analysis (IPA;
disagreements. Most groups will overcome this, and when a group see also Bales & Slater, 1955, and Chapter 2, this volume). IPA
structure and group roles have been established, they can move makes the basic and import-
ant distinction between task task behaviour behaviours during group
behaviours (all behaviours dir- interactions that are directed at task
ected at task completion) and completion
socio-emotional behaviours socio-emotional behaviour behaviours
(all behaviours directed at during group interactions that are directed
interpersonal relations within at interpersonal relations
the group). In the socio-
emotional domain it further distinguishes between positive and
negative behaviours. According to Bales, task-related behaviour
is necessary for task completion but can lead to conflicts when
people disagree. In order not to disturb the functioning of the
group, socio-emotional behaviour is necessary to restore group
harmony. The coding system of IPA is shown in Figure 12.5. As
can be seen in the figure, the scheme distinguishes between 12 dif-
ferent categories, divided into socio-emotional behaviours that are
positive, task-related behaviours (which are emotionally neutral)
and negative socio-emotional behaviours.
Plate 12.3 Some groups are formed for a special reason and also Now, according to the Tuckman and Jensen (1977) stage
end after some time, e.g., a group of students in a psychology model, these 12 categories of behaviour should occur to differing
seminar. degrees in the different stages of group life. The forming stage
9781405124003_4_012.qxd 10/31/07 3:11 PM Page 255

GROUP DEVELOPMENT AND STRUCTURE: THE GROUP LEVEL OF ANALYSIS 255

1. Shows solidarity, raises other’s status, gives help, reward.

Socio-emotional behaviour, positive 2. Shows tension release, jokes, laughs, shows satisfaction.

3. Agrees, shows passive acceptance, understands, concurs, complies.

4. Gives suggestions, directions, implying autonomy for other.

5. Gives opinion, evaluates, analyses, expresses feelings and wishes.

6. Gives orientation, information, repeats, clarifies, confirms.

Task behaviour, neutral
7. Asks for orientation, information, repetition, confirmation.

8. Asks for opinion, evaluation, analysis, expression of feeling.

9. Asks for suggestion, direction, possible ways of action.

10. Disagrees, shows passive rejection, formality, withholds help.

Socio-emotional behaviour, negative 11. Shows tension, asks for help, withdraws out of the field.

12. Shows antagonism, deflates other’s status, defends or asserts self.

Figure 12.5 The coding scheme of interaction process analysis (after Bales, 1950).

should be characterized by much positive socio-emotional

PIONEER behaviour, whereas in the storming stage more negative socio-
emotional behaviour should occur. In the norming stage, there
should be both positive socio-emotional behaviour and task-
Robert F. Bales (1916–2004), a pioneer in the development related behaviour, and the performing stage should be dominated
of systematic methods of group observation and measure-
by task-related behaviour. Is this what really happens? At a gen-
ment of interaction processes, received his BA and MS
eral level, the answer seems to be yes. For example, Wheelan,
degrees in Sociology from the University of Oregon. He
Davidson and Tilin (2003) found time together to be related to
entered graduate study in sociology at Harvard in 1940
socio-emotional behaviours (the longer the group was together,
(with Talcott Parsons as his dissertation advisor), received his
PhD in sociology in 1945, and was appointed Professor of
the fewer of these behaviours) as well as to task-related behaviours
Social Relations in 1957, retiring in 1986. During the 1944–45 (the longer the group was together, the more of these behaviours).
academic year, Bales spent a formative year as Research On the other hand, stage models such as Tuckman and Jensen’s
Associate at the Section on Alcohol Studies at Yale University. can easily be criticized as an oversimplification of reality. Some
His research on the interactions in therapeutic group settings groups, for example, may never have a storming stage, whereas
for alcohol addicts formed the basis for his first and classic other groups are in conflict continuously. Further, groups may
book, Interaction process analysis: A method for the study of sometimes return to a previous stage instead of progressing to the
small groups, published in 1950. Bales hoped that by study- next (as the model would assume). Finally, it will often be impos-
ing the interaction of many such groups, he would discover sible to establish which stage the group is in, and the assumption
recurring patterns that might help to understand and to that the different stages are qualitatively different from each other
predict the functioning of problem-solving groups. His is difficult to maintain. Rather, different activities occur in each
interaction process analysis proved an extremely useful tool stage, although they may vary in intensity. Most researchers would
for studying group interaction, group member roles and therefore argue that there are no abrupt changes in the way group
group development. This research reflected his conception members interact with each other, but rather that these changes
of social psychology as the scientific study occur gradually and that one can see this as a gradual development
of social interaction with the group and its of groups over time.
activity, rather than the individuals, as the
primary unit of analysis. With this research
program he sought to integrate the psycho-
logical and sociological sources of social On being similar: Norms, shared
psychology.
cognition and cohesion
Group norms Group norms are (unwritten) rules shared by the
members of a group, which prescribe the attitudes, behaviour and
beliefs that are, and are not, appropriate in the context of the group
9781405124003_4_012.qxd 10/31/07 3:11 PM Page 256

256 CHAPTER 12 THE PSYCHOLOGY OF GROUPS: BASIC PRINCIPLES

(see Everyday Social Psychology 12.1). Because norms are pre- regulate group interaction. Group norms are also an important
scriptive, they serve as guides for attitudes and behaviour and in that source of information about social reality. Often, people rely on
way perform an important regulatory function. Group members what many people see as valid and true as an accurate reflection of
tend to conform to group norms (i.e., think and act in accordance (social) reality. Another important function of norms is that con-
with group norms), either because group norms are internalized, formity to group norms illustrates one’s commitment to the group
that is, become part of the individual’s belief and value system – it shows that one is ‘a good group member’ (cf. Hollander, 1958).
(Turner, 1991), or because group norms are enforced by the (an- This is not to say, however, that all group members always
ticipated) reaction of other group members to normative and anti- conform to group norms. Individual group members may show
normative behaviour (Deutsch & Gerard, 1955). Because of this deviant behaviour. If they do, however, they are likely to run into
adherence to group norms, groups function more smoothly than the negative responses of their fellow group members, even to the
without norms. For instance, if everybody adheres to group norms, extent that they may be excluded from the group (Schachter,
other group members’ behaviour becomes more predictable 1951). Because social exclusion is a highly unpleasant experience
and therefore can be anticipated. In that sense, group norms help (Williams, 2001; see above), such pressures to conform to group

EVERYDAY SOCIAL PSYCHOLOGY 12.1

Jonestown of all their current social ties. It also implied that they were only
in contact with like-minded people. As this chapter shows, one
On 18 November 1978, more than 900 men, women and chil- function of groups is to provide us with knowledge of our social
dren died in a mass suicide/murder in Jonestown, a jungle en- and physical world. In isolation, people may even begin to
campment in Guyana, South America. Most of them drank, or believe bizarre things, such as the concept of ‘revolutionary
were forced to drink, a fruit punch that had been laced with suicide’. The members of the cult, for example, believed that an
cyanide and tranquillizers. Parents first gave it to their children, end to Jonestown would mean not only an end to the promised
then they drank it themselves. How could this have happened? land they had believed in, but also torture and imprisonment by
Why did a whole group of people resort to this desperate the US government. When faced with the grim prospect of
measure? losing all social ties, all hopes, everything they believed in, and
The people of Jonestown were members of a religious cult, torture and imprisonment, they saw no reason to live.
called the People’s Temple. The cult was founded in the USA by Although the Jonestown case is clearly extreme, and fortu-
James Warren Jones and had moved to the jungle encampment nately very rare, it does illustrate the power of the social group
in Guyana in the mid-1970s. There, the members of the cult had (and of an autocratic leader). It is one of social psychology’s
to work hard on the fields and lived in isolation from the outside goals to understand these tragedies and hopefully prevent them
world. Immediately before the tragedy, US Congressman Ryan in the future.
had visited Jonestown with some journalists, investigating
accusations that people were being held there against their will. Plate 12.4 Members of the People’s Temple at Jonestown,
Eighteen people indeed wanted to leave with Ryan. However, Guyana, committed mass suicide in 1978. They were socialized
cult members attacked them at the airstrip as they were leav- to accept the cult’s norms, especially group loyalty.
ing, killing the congressman, three journalists and one defector,
and wounding 12 others. Back in Jonestown, Jones proclaimed
that the end had come, and that in this extreme situation ‘revo-
lutionary suicide’ was their only option. The members of the
People’s Temple obeyed, committing mass suicide and killing
those who were unable or unwilling to kill themselves (including
children and the elderly).
To begin to understand why they chose death, one must
firstly realize that the members of the People’s Temple were
socialized to accept the norms of the cult. One of the more
important norms was loyalty to the group, a norm that was quite
strictly enforced. Second, the members of the People’s Temple
lived in isolation from the outside world and had no contacts
with relatives or others outside Jonestown. One implication of
their isolation was that an end to Jonestown would imply a loss
9781405124003_4_012.qxd 10/31/07 3:11 PM Page 257

GROUP DEVELOPMENT AND STRUCTURE: THE GROUP LEVEL OF ANALYSIS 257

norms tend to be quite effective in many situations. Thus, groups Groups may share not only cognition but also emotions
may enforce and maintain their group norms. (George, 1990). Research in group emotions is still in its infancy,
As already noted in the discussion of group development, but there is emerging evidence that groups may come to share
groups develop group norms relatively early in their existence emotions, and that these shared emotions affect group functioning.
(Tuckman, 1965). This is not to say that group norms do not Barsade (2002), for instance, found that affect introduced by a con-
change. Norms may change over time. This change may occur federate in an experiment spread within the whole group and
because the environment of the group changes. It may also occur affected group members’ ratings of group functioning. In a similar
because the membership of the group changes. New members tend vein, Sy, Coté and Saavedra (2005) showed that the affect dis-
to be socialized into the group and its norms (Moreland & Levine, played by a confederate leader of a group transferred to the
1982), but they may also introduce changes to the group. Indeed, members of the group and affected group performance: groups
as research on minority influence shows (see Chapter 11, this vol- performed better when the leader displayed positive affect than
ume), if the conditions are right, a deviant minority may convert when the leader displayed negative affect.
a whole group towards a different way of thinking. Group norms
should therefore be seen, on the one hand, as enforcing their own Group cohesion Group cohesion (or ‘cohesiveness’) is the force
maintenance and, on the other hand, as subject to change over that binds members to the group and induces them to stay with the
time and situations. Group norms are thus both an influence on group (Festinger, 1950). Group cohesion is assumed to be import-
group process and an outcome of group process. ant to group functioning, be-
cause it helps keep the group cohesion the force that binds members to
Socially shared cognition and affect An aspect of groups together and motivates group the group
that is receiving more attention in recent years is shared cognition members to exert themselves
task cohesion cohesion based on
(Thompson, Levine & Messick, 1999; Tindale & Kameda, 2000). on behalf of the group. Evid- attraction of group members to the group
Over time, groups may develop a shared understanding of differ- ence for this proposition is task
ent aspects of group life, such as the tasks the group performs, the mixed, however, and research
interpersonal cohesion cohesion based
role of each member in the group, and each member’s particular suggests that it is useful to on liking of the group and its members
knowledge, skills and abilities. For each individual group member, distinguish between types of
such understanding is important, but when it is shared within the cohesion. Task cohesion refers
group it has the added advantage of setting the stage for smooth to the shared commitment to the group’s tasks, while interpersonal
coordination, communication and cooperation, because all group cohesion refers to the attraction to the group. As a meta-analysis by
members have a similar understanding of what they are supposed Mullen and Copper (1994) shows, only task cohesion is (positively)
to do and who does what. Socially shared cognitions, when accur- related to group performance. Further, cohesion may not always
ately reflecting the demands faced by the group, may therefore improve performance, as can be seen in Research close-up 12.2.
improve group functioning and performance (Mohammed &
Dumville, 2001).
A nice illustration of the
influence of shared cognition
On being different: Status
transactive memory a system of
knowledge available to group members is found in work on trans- and roles
with shared awareness of each other’s active memory. Transactive
expertise, strengths and weaknesses
memory refers to shared Whereas norms make group members’ behaviour more alike,
knowledge about how know- there are also clear differences between group members in the way
ledge is distributed in the group. Rather than having all the infor- they behave and the position they have in the group. Take, for
mation themselves, group members know who knows what and instance, a football team. Clearly, different players have different
whom to ask for information about specific things (Wegner, 1986). roles defined by their position in the field (goalkeeper, defender,
Transactive memory makes it possible for groups to operate forward). Besides these formal roles, there will also be informal
efficiently and adequately because it helps locate information and roles. For example, a more experienced team member (even
‘the right person for the job’. though not formally the team captain) may have more influence
Liang, Moreland and Argote (1995) experimentally studied on the other players than a newcomer, and another team mem-
groups that had to assemble a radio. Before they assembled the ber may always take the initiative to reconcile people after an
radio as a group, participants received training to prepare them argument.
for the task. The critical manipulation was whether individuals Earlier, we discussed Bales’s (1950) interaction process analysis
received this training as a group or individually (after which they (IPA). It appears that IPA is a useful tool for looking at status and
performed the task in newly formed groups). As predicted, roles inside a group: it is possible to keep track of the 12 different
groups that were trained together performed better than those types of behaviour (see Figure 12.5) for each group member, to
who were trained alone. This effect could be explained because see whether there are differences among group members.
groups that were trained together had more accurate knowledge Research using IPA (or other coding systems) to code behaviour
about who was good at which part of the task: they had thus in freely interacting groups has revealed a number of important
formed a better transactive memory system (see also Chapter 13, insights (see McGrath, 1984, for a summary of findings), two of
this volume). which we will discuss now.
9781405124003_4_012.qxd 10/31/07 3:11 PM Page 258

258 CHAPTER 12 THE PSYCHOLOGY OF GROUPS: BASIC PRINCIPLES

RESEARCH CLOSE-UP 12.2

Group cohesiveness leads to better performance

Low goal acceptance High goal acceptance
when the group accepts performance goals

Performance (% of maximum)
100
Podsakoff, P.M., MacKenzie, S.B. & Ahearne, M. (1997). Moderating
effects of goal acceptance on the relation between group cohesive- 90
ness and productivity. Journal of Applied Psychology, 82, 974–983.
80

70
Introduction
60
Podsakoff and colleagues argued that higher cohesion does
not always lead to better performance. They argued that the 50
Low High
relationship between cohesiveness and performance should
Cohesion
be contingent on the group’s acceptance of performance goals.
If the group accepts the performance goals of the organization, Figure 12.6 The relation between cohesion and performance
cohesiveness should be positively related to group perfor- for crews high and low in goal acceptance (after Podsakoff
mance. If, however, the group does not accept performance et al., 1997).
goals, then cohesiveness is expected to be unrelated (or even
negatively related) to performance. Results
Figure 12.6 shows the results. As predicted, group cohesion
Method and group goal acceptance interacted in predicting task per-
Participants formance. When groups were relatively accepting of perfor-
The study participants were 218 members of 40 work crews at mance goals, the relationship between group cohesion and
a paper mill in the USA. Crews consisted of 5.25 members on group performance was positive. However, when groups were
average, most participants were male (96 per cent), and their not accepting of performance goals, the relationship between
average age was 39 years old. group cohesion and performance tended to be negative.

Measures and procedure

Discussion
Two measures were obtained through a questionnaire distributed
among the crew members: group cohesiveness and acceptance This study illustrates that group cohesion does not necessarily
of the performance goals of the company. Thus, all group mem- motivate performance. Rather, it motivates group members to
bers individually rated their perception of group cohesiveness and exert themselves for causes that are seen as important to the
their acceptance of performance goals. Performance of each crew group (see van Knippenberg & Ellemers, 2003). When group
was obtained from company records. It consisted of the amount members do not accept the company’s performance goals,
of paper produced as a percentage of total machine capacity. higher cohesion will generally not improve performance.

First, some group mem-

speaking hierarchy hierarchy within a Five Six Seven Eight
group based on who talks most bers talk more than others,
and the discrepancy increases 30
Percentage of time talking

with the size of the group. 25

Thus, groups develop a speaking hierarchy (Bales, 1953) in which
20
members higher in that hierarchy talk more than those lower in
the hierarchy (see Figure 12.7). Further, people who talk more are 15
usually seen as more influential. Later research has shown that 10
group members do not distribute their participation evenly
5
throughout the discussion, but rather that contributions are con-
centrated in periods of high activity (Dabbs & Ruback, 1987). Thus, 0
1 2 3 4 5 6 7 8
if a person has recently spoken, he or she is more likely to speak
Member rank (leader excluded)
again. Often this takes the form of a dyadic exchange, in which
two group members alternate speaking turns. When this happens, Figure 12.7 Speaking hierarchy for groups of five, six, seven and
we say that the group is in a floor position (i.e., two group members eight members (taken from Stephan & Mischler, 1952).
9781405124003_4_012.qxd 10/31/07 3:11 PM Page 259

GROUP DEVELOPMENT AND STRUCTURE: THE GROUP LEVEL OF ANALYSIS 259

Task success in the past

High seniority
Specific status High
characteristics (high expectations of
perceived ability) High status More influence
successful task
Diffuse status contribution
characteristics
Personality

Figure 12.8 Expectation states theory (Berger et al., 1980).

‘hold the floor’; Parker, 1988). Parker found that four-person things happen. The important question, then, is: what determines
groups were in a floor position no less than 61 per cent of their these performance expectations?
time, much more than would be expected if all group members The theory assumes that performance expectations are
contributed equally. influenced by so-called status characteristics. The theory dis-
Second, research using IPA has found that some people are tinguishes between diffuse status characteristics (not necessarily
consistently more task-oriented (i.e., they engage mostly in task- related to the group task), including, for example, gender, age and
related behaviours, categories 4–9 in Figure 12.5), whereas others race, and specific status characteristics, such as skills and abilities
are more relationship-oriented (i.e., they engage more in socio- (i.e., characteristics that are necessary for the group task, previous
emotional behaviours) (Slater, 1955). The former person has been task success). These characteristics carry certain cultural expecta-
labelled the task specialist, and the latter the socio-emotional tions about competencies. For example, women are generally
specialist: clearly a case of (informal) role differentiation. It further seen as less competent than men (especially on tasks that are more
appeared that these two group members interacted with each ‘masculine’; e.g., Pugh & Wahrman, 1983), and more senior
other quite frequently, and much more than would be expected people may be seen as more competent (up to a certain age) than
according to chance (i.e., they were often in a floor position). younger people (Freese & Cohen, 1973). Similarly, higher expec-
Finally, the task specialist was seen as most influential, but he or tations are formed for people who are more experienced, have a
she was liked less than the socio-emotional specialist. higher status in society more generally, or have a relevant area of
Who talks most in the group and who takes which role is expertise. Obviously, these expectations may sometimes be false
dependent on personality and individual abilities. For example, an (i.e., a woman may in fact be more competent than a man), but
extroverted person will probably talk more than an introverted they nevertheless affect people’s status in the group and the
person. However, this is not the whole story. There are other amount of influence they have. The reason is that expectations
factors that determine who is more and who is less influential. need to be explicitly falsified before they lose their influence, and
The most comprehensive as long as they are not, they continue to have their effect in a self-
theory about status in groups fulfilling way. There is extensive evidence supporting the theory.
expectation states theory argues that is expectation states theory For example, Driskell and Mullen (1990) found that characteristics
status differences within a group result from
different expectations that group members
(Berger, Rosenholtz & of group members affected their status and power through the
have about each other Zelditch, 1980). It deals with expectations of other group members (for more evidence, see
the issue of how status struc- Ridgeway, 2001).
tures emerge in groups, and
how they are shaped by the outside status of group members (see
Ridgeway, 2001, for an overview of the theory and the evidence
for it). A simplified graphical depiction of the theory is presented
in Figure 12.8. SUMMARY
Expectation states theory is applicable to groups in which mem-
bers strive for a common goal or perform a common task. It as- Groups develop over time, in the sense that their interac-
sumes that several inequalities within a group, such as inequalities tion patterns change. Further, some processes cause group
in participation and influence, are highly correlated because they members to become more similar to each other, both in
are all derived from performance expectations. That is, because of terms of their behaviour (as prescribed by group norms) and
certain characteristics of group members, other group members in terms of their cognitions and emotions. Finally, differ-
form expectations about the usefulness of each group member’s ences between group members may also emerge, for which
contributions. These expectations then serve as a self-fulfilling expectation states theory offers a theoretical account. We
prophecy: the greater the expectations, the more likely a person is now turn to the last level of analysis: the contextual or
to speak up, offer suggestions and be evaluated positively by the intergroup level.
others. The lower the expectations, the less likely it is that these
9781405124003_4_012.qxd 10/31/07 3:11 PM Page 260

260 CHAPTER 12 THE PSYCHOLOGY OF GROUPS: BASIC PRINCIPLES

GROUPS IN THEIR These processes are well illustrated in a study by James and
Greenberg (1989). They conducted two experiments in which they
had students from their university work on a task solving ana-
ENVIRONMENT: THE grams. The task objective was to solve as many anagrams as pos-
sible and participants’ performance on the task (i.e., the number of
INTERGROUP LEVEL anagrams solved) was the main variable of interest. James and
Greenberg argued that students would be more motivated, and
OF ANALYSIS therefore perform better, when their university membership was
made salient in the context of a comparison between students from
their university and students from another university.
In what ways does the (intergroup) context affect intragroup James and Greenberg experimentally manipulated the extent
behaviour? to which students’ affiliation to their university was salient. In their
How does behaviour in groups change when group membership is first experiment, they manipulated group membership salience
made salient? by letting participants work in a room that was painted either
white (low salience condition) or red and blue (the colours of the
Going back to our opening example of the football fans, it is clear university: high salience condition). All participants were led to
that these people do not always behave in this way. They are also believe that the experiment was part of a larger study compar-
supporters of their team when the team is not playing, but it is the ing the performance of students from their university with that
context of the match that draws them together and that brings out of students from a ‘rival’ university. As expected, participants in
their behaviour in the subway station. Playing against another the high group membership salience condition solved more ana-
team renders these supporters’ affiliation with their favourite team grams than did participants in the low group membership salience
salient and evokes the quite uniform behaviour that clearly iden- condition.
tifies them as a group. In their second experiment, James and Greenberg aimed to
What holds for these supporters holds for all groups. Groups do show that this effect would only be found in the presence of in-
not live in isolation. Other groups are part of the environment in tergroup comparison and not in the absence of this intergroup
which groups function. Understanding the psychology of groups comparison. In order to demonstrate this, they manipulated not
therefore requires studying the influence of the intergroup context only group membership salience but also the presence or absence
on the thoughts, feelings and behaviour of group members. Part of of the comparison with the other university. Intergroup compari-
this involves the study of intergroup relations – the way group son was manipulated by telling half of the participants that their
members think, feel and act towards members of other groups (see performance would be compared with that of the rival university,
Chapter 14, this volume). The intergroup context may, however, also whereas the other half did not receive this instruction. This time,
affect intragroup processes, and that is an issue we deal with here. salience was manipulated by giving participants a practice anagram
that solved either as wildcats, which referred to their university
mascot (high salience condition), or as beavers, which had no relev-
ance for university membership (low salience condition). Results
The intergroup context and the indicated that group salience had no effect when the intergroup
salience of group membership comparison was absent, but that group salience led to higher (and
the highest) performance when intergroup comparison was pre-
The fact that individuals are members of a certain group does not sent (see Figure 12.9).
mean that this group membership is always at the forefront of their
minds. Self-categorization as a group member needs to be cognit-
ively activated, or rendered salient, for the group membership to Low group salience High group salience
exert its influence on people’s self-definition (see Chapter 5, this
volume). Group membership then influences group members’ 90
attitudes and behaviour via this self-definition (i.e., social identity; 80
Percentage correct

see Turner et al., 1987; see also Chapters 11 and 14, this volume). An 70
60
important influence of the intergroup context is that of rendering
50
group membership salient. Exposure to other groups in a sense 40
‘reminds’ us of our own group memberships. Especially in the con- 30
text of an intergroup confrontation of some kind, this may work 20
to render group membership a salient influence on group members’ 10
thoughts, feelings and behaviour. Such confrontations may involve 0
Absent Present
explicit competition, as in sports or in the political arena, or com-
Group comparison
petition for scarce goods (e.g., customers, funding), but may also
involve more implicit forms of competition, such as competition Figure 12.9 Percentage of anagrams solved correctly as a
for social status (e.g., which is the most important department within function of ingroup salience and comparison condition (after
an organization? which street gang has the toughest reputation?). James & Greenberg, 1989).
9781405124003_4_012.qxd 10/31/07 3:11 PM Page 261

GROUPS IN THEIR ENVIRONMENT: THE INTERGROUP LEVEL OF ANALYSIS 261

What this study shows is that group membership needs to In contrast to the other countries of the European Union, the
be salient to affect behaviour, but that the context in which it is largest religious denomination in Turkey is Islam rather than
rendered salient affects whether and how group membership Christianity. This fact seems to have highlighted the shared roots
salience translates into behaviour (for more on this issue, see in Christianity of the current EU countries in the perception of
Haslam, 2001; van Knippenberg, 2000). In the intergroup context many parties partaking in the discussion – an attribute that until
created by James and Greenberg, performing well could help now never really seemed at the forefront of perception within the
establish that one’s own group was superior to the comparison European Union.
group. Because salient group memberships reflect on how we see Changes in the intergroup context may occur because old
ourselves, the relative standing of our group vis-à-vis other groups groups disappear from the scene (e.g., a competitor goes bankrupt)
(i.e., are we ‘better’?) reflects on how good or bad we can feel or new groups emerge, or because an existing group becomes
about ourselves. Obviously, then, we prefer our groups to com- more relevant as a comparison group (as in the example of Turkey
pare favourably to other groups, and are willing to contribute ac- and the European Union) or less relevant as a comparison group
tively to our group achieving such a favourable comparison (Tajfel (e.g., because a competing firm focuses more on other markets than
& Turner, 1986). In the situation created by James and Greenberg, one’s own firm). Such changes may affect which attributes of the
this led individuals for whom group membership was made salient group are salient (i.e., what differentiates the group from relevant
in the context of intergroup comparison to work harder. other groups), but they may also alter our perception of a given
An important influence of the intergroup context on group attribute of the group. Take the example of a group of psychology
members is, thus, that it may render group membership salient, students who think of themselves as intelligent. Within the larger
and may inform the translation of this salient self-categorization context of society, this probably makes a lot of sense. Imagine,
into attitudes and behaviour. The intergroup context may also however, that this group finds itself in a context where comparison
affect group members’ perceptions of their own group, and by doing with a group of the proverbial rocket scientists becomes relevant.
so may affect attitudes and behaviour that are contingent on these Intelligence may not be seen as the most relevant dimension of
perceptions. This is an issue that is addressed next. comparison, but if it were, the attribute intelligent would likely be
ascribed not to one’s own group but to the other group.

The intergroup context, group

perceptions and social influence SUMMARY

Part of what defines a group is the distinction between who is ‘in’ In sum, the intergroup context may both affect the salience
and who is ‘out’. Groups exist by virtue of their members, but also of group membership and inform group members’ beha-
by the fact that there are some people who are not members of the viour within this context (cf. the intergroup comparison
group and may indeed be members of other groups. Accordingly, in the James & Greenberg 1989 study), and influence per-
people’s perceptions of their membership groups are affected by ceptions of group norms that may feed into attitudes and
the comparison between their own group and other groups, and behaviour.
group members’ perceptions of their group are also contingent on
what differentiates their group from other groups (Turner et al.,
1987). Put differently, we ascribe characteristics to ourselves and to
our groups on the basis of our perception that we possess these
characteristics to a greater degree than others. For example, we SUMMARY AND
will only come to the conclusion that the members of our group CONCLUSIONS
are intelligent if we perceive our group to be more intelligent than
certain other groups. Indeed, such social comparison processes
evaluating ourselves permeate social life (see Chapter 14 for fur- Let us return to the example we began with: the Ajax fans in the
ther detail on intergroup social comparison). The important point subway. The individual Ajax fan is probably looking for an enter-
for our present discussion is that if the intergroup context changes, taining and enjoyable game of football. However, there is more
comparison groups may change and as a consequence our per- than that. As you will probably agree, ‘real’ football fans identify
ceptions of our group may change. very much with their teams: they are proud of the team when the
Take, for instance, the case of political parties. Members of a team wins, and feel sad and depressed when the team loses. Being
party that is the most conservative party within a country’s polit- an Ajax fan is thus part of an individual’s identity, and self-esteem
ical spectrum will probably think of their party as conservative. is derived from the team’s success. When an individual Ajax fan
However, when a new party emerges that is perceived to be more now enters the subway, he or she will know what to expect:
conservative, the attribute conservative may become less suited watching a game of football implies singing and shouting. Because
to distinguish the party from other parties, and party members’ most of the other people in the subway are Ajax fans, and because
perceptions of their party may change to emphasize other charac- they have similar expectations, the behaviour becomes normative:
teristics of their party. Or consider, for example, the discussion it is seen as appropriate. However, the only reason why this beha-
about Turkey’s prospective membership of the European Union. viour is seen as appropriate (or at least acceptable) is because of
9781405124003_4_012.qxd 10/31/07 3:11 PM Page 262

262 CHAPTER 12 THE PSYCHOLOGY OF GROUPS: BASIC PRINCIPLES

the context: Ajax is about to play another team, and this both l Cohesion can be based on attractiveness of the group
makes group membership salient and affects the perception of (interpersonal cohesion) or on attractiveness of the group
group norms. Thus, the behaviour in the subway is caused by task (task cohesion). In general, cohesion motivates group
individual expectations (individual level), which are shared among members to exert effort for causes that are important to the
the fans and constitute behavioural norms (group level), and arise group.
in a context that makes group membership salient (context level).
l Groups develop status and role differences. Expectation
l A group exists when two or more people define themselves states theory explains the emergence of a status structure
as members of a group. in a group. It argues that certain status characteristics
lead to performance expectations that subsequently
l The reasons why people form, join and distinguish groups lead to differences in status and influence.
are sociobiological (evolutionarily built in), cognitive
(understanding our world) and utilitarian (gaining benefits). l The presence of other groups can make group
membership salient. As a consequence, group
l Different types of groups, such as task groups, intimacy members will be more strongly influenced by
groups, social categories and loose associations, differ on a their group membership.
number of important dimensions such as group entitativity,
importance and shared objectives.
l Group members move through the different stages of group
membership (prospective member, new member, full Suggestions for further reading
member, marginal member and ex-member) separated from
each other by role transitions, and these different stages are Haslam, S.A. (2001). Psychology in organisations: The social identity
characterized by different levels of commitment. approach. London: Sage. A detailed review of the influence of
l The role transition of entry can be marked by a harsh group norms and intergroup context on attitudes and
transition ritual. A classic explanation for these rituals is given behaviour in groups.
by dissonance theory, which argues that such rituals increase Lickel, B., Hamilton, D.L. & Sherman, S.J. (2001). Elements of a
commitment to the group. lay theory of groups: Types of groups, relational styles, and
the perception of group entitativity. Personality and Social
l An important determinant of group openness is staffing level: Psychology Review, 5, 129–140. An in-depth discussion of types
it is easier to become a full member of an understaffed as of groups and ‘lay theories’ about them (i.e., how lay people
compared to an overstaffed group. look at groups).
Moreland, R.L. & Levine, J.M. (1982). Socialization in small
l Social exclusion from groups can lead to quite severe anger
groups: Temporal changes in individual–group relations. In
and depression.
L. Berkowitz (Ed.), Advances in experimental social psychology
l Groups develop over time, because the challenges they face (Vol. 15, pp. 137–192). New York: Academic Press. An
and the goals they have change. Tuckman’s classic theory extensive discussion of group socialization.
distinguishes five stages: forming, storming, norming, Ridgeway, C.L. (2001). Social status and group structure.
performing and adjourning. In M.A. Hogg & S. Tindale (Eds.), Blackwell handbook of social
psychology: Group processes (pp. 352–375). Oxford: Blackwell.
l Interaction process analysis is a useful coding scheme for
A good summary of the research on expectation states
group interactions and makes a basic distinction between
theory.
socio-emotional and task behaviours.
Wheelan, S.A. (1994). Group process: A developmental perspective.
l Groups develop shared cognitions, such as transactive memory Boston: Allyn & Bacon. A discussion of group development
systems (i.e., knowing who knows what) and shared emotions. that examines different stages of group life.
9781405124003_4_012.qxd 10/31/07 3:11 PM Page 263