11.4.3.1 Group cohesion

Janis considered group cohesion to be the most important antecedent of groupthink (Esser 1998: 127). Cohesion is the ‘resultant of all forces that hold group members together’ (McCauley 1989: 251). Janis regards group cohesiveness as ‘members’ positive valuation of the group and their motivation to continue to belong to it’ (Janis 1982: 4). He adds that high levels of group cohesiveness lead to the expression of ‘mutual liking, and positive feelings about attending meetings and carrying out the routine tasks of the group’. However, this is only one way group cohesiveness is created. McCauley identifies two different forces that lead to group cohesion (McCauley 1989: 251). The first is – as described above – the attractiveness of the group and its members and the prestige it generates. The second force leading to or source of groupthink is based on performance-related rewards. McCauley cites Janis about the difference between those sources of group cohesiveness and the consequences for groupthink: ‘Concurrence seeking tendencies are stronger when high cohesiveness is based primarily on the rewards of being in as pleasant ‘clubby’ atmosphere or of gaining prestige from being a member of an elite group than when it is based primarily on the opportunity to function competently on work tasks with effective co-workers’ (Janis 1982: 247). Therefore, when cohesiveness is based on the attractiveness of the group, the risk of groupthink is greater, as new members want to be accepted by the ones that provide prestige.

Cohesiveness itself is not a bad thing, but too high levels of cohesiveness may lead to low-quality decision-making caused by groupthink. Furthermore, group cohesion is a necessary, but not a sufficient, condition for groupthink (Janis 1982: 245). In addition, group cohesion is necessary for cooperation within a group. However the levels of group cohesion may not rise too far, because that may increase the likelihood of groupthink. Furthermore, low levels of group cohesion are not recommended either, because the quality of decision-making does not increase when people cannot get along with each other. ‘Both the control and service components of the board’s task require extensive communication and deliberation, and board members must have a certain minimum level of interpersonal attraction in order to engage in these things. In addition, board members must trust each other’s judgment and expertise, and such trust will be difficult to sustain on boards with very low levels of interpersonal attraction’ (Forbes and Milliken 1999: 496). Or as Janis puts it: ‘For most groups, optimal functioning in decision-making tasks may prove to be at a moderate level of cohesiveness, avoiding the disadvantages of conformity out of fear of recrimination when cohesiveness is low and the disadvantages of strong concurrence-seeking when cohesiveness is high’ (1982: 248).

O’Connor puts the theory of group cohesiveness in a framework, which she uses to analyse the failures of the board of Enron (2003: 1261-1262). She emphasises that an important source of group cohesion is the willingness to belong to a certain group, which was described above. This is a phenomenon that occurs in board rooms as well. Independent NEDs or members of the supervisory board are selected for their expertise and informal or professional network. The invited NED or supervisory board member might make his decision to join the board dependent on the prestige or reputation of the sitting members. This makes them function as ‘good ole boys clubs’ and ‘elite private clubs with a rubber stamp’ (O’Connor 2003: 1262).

This theory about group cohesion stresses that groups need a moderate level of group cohesion to function well. Too high levels do not contribute to good decision-making, neither do too low levels of group cohesion. This also applies to groups of independent supervisors. If such a group is independent, but suffers from high levels or low levels of group cohesion, the decision-making is likely to be of low quality. In order to prevent this, the independent members had better not receive motives and rewards based on their performance. In addition, the selection procedure should be such that the identity of the other board members is not a reason for accepting a position on the board as NED or a position as supervisory director. By taking these two points of advice into account, the risk of group cohesiveness and possibly accompanying groupthink is decreased.

In a literature review of Esser the results of case studies and laboratory tests on group cohesiveness are described (1998: 126-131). Some analyses of case studies do not find evidence for group cohesion being a predictor for groupthink. This again stresses that not all antecedents need to be present in order to have groups that suffer from groupthink. Laboratory tests reported that non-cohesive groups reported more self-censorship than cohesive groups, which found more alternative ideas. In addition, laboratory tests reported that within cohesive groups dissent was discouraged more than in non-cohesive groups. Therefore, no or weak evidence is found in empirical research for the theory about the relationship between group cohesiveness and groupthink.