The conformity studies show group influence by peers, but Milgram (1974, cited by Brown 1996) distinguished this situation from the influence of persons of authority in hierarchical situations. Milgram investigated the strength of pressure to obey with a now-famous methodology. Milgram’s subjects were led to believe that they were giving electric shocks to learners (actor confederates), and they were prompted to do so a clearly marked ‘dangerous levels’ despite the pre-recorded cries of the learners. The subjects exhibited signs of nervous tension, but in the situation where they were alone twenty-six of forty subjects administered the highest shock level.
In a variation of this Milgram demonstrated the positive influence of peer pressure in a group situation. The subject was joined by two confederates who were instructed to resist the experimenter’s authority. In this case the subject joined the peer rebellion and defied the experimenter in thirty-six out of forty trials. This can be interpreted in different ways. If the subject is seen as a pliant stooge, then this is a demonstration of group influences in a more complex, real way.
There are competing influences in most situations which must be incorporated by the individual. It could be that in group situations the individual is affected to a greater extent by a peer than somebody in authority. Equally, it could be a simple aggregated effect with two confederates outnumbering the single experimenter. More positively, if we assume that the subject really does not want to shock the learner, then it may show the enabling effect of the support of others.
Taken together these experiments begin to suggest that the process of influence in groups is not clear-cut or homogenous. As Brown (1996, p19) puts it “[T]here is no universal way in which individuals respond to group pressure…participants will be affected by the meaning a situation has for them which itself may be influenced by cultural variables”. This is a theme that was taken up in experiments into minority influence.
Mucovici et al. (1969, cited in Brown 1996) showed the effects of different kinds of peer behaviour on the strength of influence. Groups of six including two confederates were required to name the colour of a slide. The confederates called blue slides as green either consistently or consistently. Results showed that the minority influence was only significant where a consistent behavioural style was evident. Mugny (1975, cited in Brown 1996) held group discussions on topics of contemporary concern and found that the influence of a minority required that an appropriate argument style be employed. Where differences in opinion were large a flexible negotiation style was more effective, but with smaller differences a more rigid style was more influential.
Such experiments show a tendency towards reductionism, groups are reduced to peers, authorities, majorities and minorities. Group pressures become types of influence and influence becomes behavioural style. The more specific that group experiments become the less that they seem to be talking about groups and the more they refer to individuals and circumstances. However, some experiments may have more universal implications. Wetherell (1996, p203) refers to two series of experiments that may indicate that “[G]roup membership in itself has profound effects upon the psychology of the individual regardless of personality and individual differences”.
The ‘Summer Camp Experiments’ of Sherif and Sherif (1969, cited in Wetherell 1996). used a process of in-group formation and inter-group competition which led to positive identification with the in-group and to overt out-group hostility. They found that specific psychological tendencies were not necessary for this to occur, and concluded that “[T]he objective relationship between two groups causes the various subjective or psychological states characteristic of intergroup relations”.
Tajfel and Turner’s (1979, cited in Wetherell 1996) ‘Social Identity Theory (S.I.T)’ based on the findings of the ‘Minimal Group Experiments’ goes further in identifying the particular processes that occur in individuals when they are in groups. The minimal group idea removed circumstantial factors by taking the extreme case of a group defined by the mere recognition of it. Using schoolchildren, groups were created ostensibly from arbitrary categorisation according to stated preference for the work of either Klee or Kandinsky.
Tasks showed that individuals consistently gave preferential treatment to their own group even where there was no-contact between members and there was no overt self-interest involved. They concluded that competition for goals is not required for group conflict, but that conflict occurs due to the self-esteem of the individual being tied to the status of the group. This happens through a process of ‘social categorization’, ‘social identification’ where individuals self-categorizes themselves as group members and attach value to that membership, and social comparison with out-groups.
Some of the problems of S.I.T. reflect the wider limitations of experimental work on groups. The ‘sense of belongingness’ required for defining a group seems to be activated too easily by the individual in isolation. This presents two problems. First, such sensitive activation makes it difficult to determine which single or combination of groups an individual is being influenced by at any one time. Second, if I read the Financial Times and feel a sense of belonging and esteem by doing so, it is not clear whether this says anything at all about the group called ‘Financial Times Readers’. Whilst group conflict experiments show that no particular personality traits or circumstances are needed for individuals to be affected by group membership, the analysis still fails to enter the ‘domain of the group’ because the processes identified are not essentially at the system level.
Sapsford’s (1996, p70) ‘Domains of Analysis’ tool might usefully be employed to examine the experimental approach to groups. Group experiments based on participants having effects on each other are within the ‘interpersonal/personal domain’ which “[T]reats the person as a whole – living interaction and relationship with other people, but analytically separate from the….[which] presupposes the idea of the individual or person as something distinct from the social world”.
The experimental methodology therefore prevents work on groups from being located, as might be expected, in the ‘domain of the group’. This domain “[I]s concerned with what people create between them – though not primarily with what participating individuals do and think; the focus is either on the system of which the individuals are a part or the meanings they create between them”. Social constructionist and psychodynamic perspectives on groups fit more neatly into this domain, since the group itself is seen as the fundamental unit of analysis.
The ‘group psychodynamic paradigm’ offers a more inclusive vision of group processes. Aschbach and Schermer (1994, adapted in Morgan and Thomas, 1996, p77) describe the paradigm across three systems. First, the internal/intrapsychic systems of each individual in the group including unconscious motivations; Second, systems of communications, both conscious and unconscious between two or more individuals; Finally, The group as group.
Some psychodynamic work on groups differs in assumptions from the experimental approach whilst still retaining a focus on what happens to individuals in groups. Instead of measurable behaviour there is an emphasis on subjective experience, a dynamic unconscious and the influence of primitive motivations within the group context. However, it is in the notion of the ‘group mind’ that the approaches can be most clearly distinguished. Turner (1984, cited in Morgan and Thomas 1996, p68) forthrightly rejects the concept: “It is a basic assumption of modern psychology that psychological processes reside only in individuals – in the most literal sense, at least, there is no such thing as a ‘group mind'”.
Thus it would seem that there are no surprises in what experimental evidence can tell us about what goes on in groups. Scientific positivism is scientific positivism. Whilst it is shown that individuals are complex and do not react in just one way, patterns of influence are described in cause-and-effect terms without recourse to emergent properties. The value that we place on this evidence depends on how we define a group. If it is simply a collection of people with affiliation then we must accept that group experiments are some of the most interesting and telling in psychology. But, if we incorporate the possibility of a ‘groupishness’ (Bion’s word) beyond that collection, then experiments can contribute little to our knowledge of the group so defined.