Relevant Scientific Research

The purpose of this chapter is to review relevant bodies of scientific research to provide input into the development of a methodological framework for multi-stakeholder processes. It is intended to provide some theoretical and empirical basis for the suggested framework and its individual steps and mechanisms. The goal is to develop an overview of issues addressed in social and organisational psychological research, as related to the specific aspects of multi-stakeholder processes, particularly various aspects of effective decision-making processes in groups of high diversity.

We will start by looking at some basic findings in the field of social and organisational psychology. Most research in this area is conducted in laboratory settings as a means of controlling the multiple conditions of real life social processes. Controlling these complex conditions is a means of enabling research to look at a single phenomenon separately without all the other conditions of "real" social processes interfering with the analysis.

It is particularly noteworthy that, while there is an extensive body of research in the area of social psychology [2] into group processes, group dynamics, communication and decision-making within groups, research into the specifics of multi-stakeholder processes has hardly been undertaken. Inter-group cooperation and conflict in realistic settings has been addressed by organisational psychology, however, with a clear focus on team-based, often hierarchical structures within corporations. These function under conditions which are in many ways different from those in multi-stakeholder processes, where representatives from different sectors of society aim to discuss or collaborate on a certain issue for a certain period in time. Therefore, some of the research findings reported here can be transferred only to a certain extent.

Clarifying the impact of diversity on communication patterns and decision-making processes will lead us to examining the impact of various methods for achieving consensus. We will explore different forms of diversity, such as gender and ethnicity in more detail and look at the consequences of these and other differences, such as status and power, on effective decision-making and implementation. The chapter will conclude by looking at the role of leadership, mediation and interactive conflict resolution as a means of assisting diverse groups in achieving their full potential.

The intention is to make the information obtained in existing research accessible, relevant and applicable to multi-stakeholder processes. The suggested analytical framework for multi-stakeholder processes has been checked against these findings.

The increasing popularity of group-based decision-making reflects a widely shared belief that group decision-making offers the potential to achieve outcomes that could not be achieved by individuals working in isolation. Diverse perspectives allegedly are beneficial to decision-making processes. Members with diverse perspectives are supposed to:

· provide the group with a comprehensive view of possible issues on the agenda, including both opportunities and threats;

· alternative interpretations of the information gathered and creative courses of action and solutions that integrate the diverse perspectives (Triandis, Hall & Ewen, 1965).

Diverse groups offer immense potential for increased quality of group performance and innovative decision-making (Jackson, 1996; Seibold, 1999; Phillips & Wood, 1984; Pavitt, 1993). The direct involvement in the decision-making process is likely to lead to a change of norms and to individual commitment. However, benefits from decision-making groups are not automatic.

When analysing the potential problems that can emerge through diversity in decision-making groups from a social psychological perspective, stereotypes are of particular importance. A social stereotype is "a set of beliefs about the personal attributes of a group of people" (Ashmore & DelBoca 1981: 16). Such sets of beliefs are being 'activated' (i.e. start influencing perception in a given situation) through identifying the group membership of a person. In other words, once we identify a person as a woman, for example, our stereotypical beliefs about women will be influencing our perception and judgment towards this person.

Once stereotypical beliefs come into play in the cognitive process, they affect people's perception, attitude, and behaviour. The impact of stereotypes can increase in difficult decision-making processes when strong emotions like anxiety, irritation or anger arise and overshadow our judgment (Mackie & Hamilton, 1993). However, contact with members of the stereotyped group might be the first step in overcoming prejudice if the information given is given repeatedly and comes from more than one - typical - group member (Pettigrew 1989). In many cases, the best strategy in order to overcome prejudice has proven to be to get both groups engaged in a common activity – working together, particularly if the activity is successful [3] can significantly contribute to reducing prejudice and improve relations between different groups (Sherif & Sherif, 1953; Smith & Mackie 2000).

Stereotyping does not necessarily imply negative evaluation but often it does – then it implies social discrimination: A person is being judged negatively merely because they belong to a certain social group. It is important to note that being discriminated against can elicit "counter-discrimination", hence increase distance between social groups (Hemmati et al. 1999).

Overcoming stereotyping and prejudice is therefore an important component of successful processes with groups of high diversity.

In practice, diverse decision-making teams have often not achieved their potential. The interaction problems associated with diversity often lead to lower performance than if the group had fewer resources. The need for integration of diversity is great (Maznevski, 1994).

Diverse groups are designed to differ with regard to various characteristics, such as the demographic composition of the group, e.g. gender, age and ethnicity; educational and occupational background; knowledge and area of expertise; attitudes and values; as well as status and power – or, in the case of multi-stakeholder processes, they differ with regard to a mix of those characteristics. An additional facet to diversity in groups is specified by Belbin (1993). Based on training experience with management teams, he distinguishes nine functional team roles that contribute to effective performance of decision-making teams: plant, resource investigator, co-ordinator, shaper, monitor evaluator, team worker, implementer, specialist, completer and perfectionist. Optimal group composition is given when all roles are represented, leading to a high degree of compatibility within the team (for a further discussion see Beck et al., 1999).

In the context of groups consisting of representatives from various stakeholder groups, Belbin's approach cannot easily provide us with pragmatic recommendations. However, his categories of team roles make a strong point about the significance of diversity in appreciating personal and functional differences. Differences provide a space to build on each other's strengths and can be a means to reduce competition and enable co-operation.

There is, it should be said, some ambiguity about the importance of group composition. Group composition can be seen as an important determinant of the performance of a group. However, group composition is also merely a determinant of the resources available to a group.

Studies on task- or expertise-based status have received little empirical attention. An interesting phenomenon observed within groups composed of experts and relative novices is the "assembly bonus effect", which occurs when both expert and non-expert members of the group perform better within the team context than they would alone (Shaw, 1981). One explanation for this effect is that experts learn during interactions with non-experts because of a need to clarify assumptions they automatically make when dealing with issues in their domain of expertise. Findings such as these suggest that performance is enhanced when both experts and novices are represented in one problem-solving group (Jackson, 1996).

The implications of diversity are far-reaching in the way that members of a group process information, make decisions and implement them. No single theory explains the complex relationship between the different dimensions of diversity and its possible consequences on effective performance of the group, e.g. on communication patterns within a group, communication across group boundaries, cohesiveness and so on. A variety of perspectives have guided the studies, including Social Identity Theory (Tajfel, 1979; Turner et al., 1987 [4]) and research on management composition (Hambrick, 1994).

The following section describes some of the consequences of diversity in more detail.

Communication is an essential process in the development of group culture. The type of communication structure determines leadership, roles and the status hierarchy within the group; group morale and cohesiveness; and it limits or enhances productivity (Hare, 1992).

The balance between task-focused and socio-emotional communication is crucial if a group is to be effective. Different types of communication are needed for different tasks. If a group's task is relatively simple, a centralized communication network, in which interaction between members is limited, tends to increase effectiveness. Complex problem solving is facilitated by decentralized communication networks (Shaw, 1981). As recommended by Wheelan (1994: 33), the choice of a communication network might be more effective if strategies of decision-making were outlined in advance and if urges to stabilize the structure too early were resisted, as there is considerable resistance to change once these structures are established. Awareness of these issues is usually low and it is one of the tasks of the group leader or facilitator to bring them to the group's attention. It is notable that a decentralized communication network does not exclude the existence of a group leader (see discussion below).

Communication standards, and thus performance, are raised if the group has clear, performance-oriented goals; an appropriate task strategy and a clear set of rules; fairly high tolerance for inter-member conflict and explicit communication feedback to ensure that information is understood (Maznewski, 1994: 532).

Decision-making is not simply rational information gathering (Jackson, 1996). For example, information held by only one member of the group is often ignored. Research on social influence and conformity indicates the value of having on a team at least two people who agree on an answer. The well-known social influence studies are the classic experiments by Salomon Asch, who asked people in a group to judge line length after hearing the erroneous judgments of several other people. This research revealed that when a person's private judgment was unlike the judgments expressed by others, the person soon abandoned his or her own judgment, even when their answer was verifiably correct. However, in the presence of just one other person who agreed with the subject, subjects persevered in the face of opposition (Asch 1951, 1956).

Also, just as an individual is likely to lack confidence, the team may lack confidence that, in an ambiguous situation, a deviant opinion could be correct. This is particularly true if the individual with the correct answer is of relatively low status. Such evidence suggests that for diverse groups to fulfil their potential, group members should have overlapping areas of expertise, instead of a sole expert for each relevant knowledge domain (Jackson, 1996).

As demonstrated by a substantial body of research (Poole, 1991; Seibold, 1999), applying formal procedures might control the potential problems of “free” group discussions. Formal procedures offer various models to lessen social influence which can undermine the value of contributions from low status members, as described above, and facilitate effective group discussions (see discussion of various procedures under 8.).

If individual members of a group initially have opposing views on an issue and the number of supporters on both sides are (more or less) evenly split, the communication process usually results in compromise (Wetherall, 1987). Through processes of social influence, the position reflected in the final decision becomes more moderate, an effect called ‘depolarisation’. Divergence between a final decision and member views is generated. This process can reduce the motivation of individual members to participate up to their capacity in group decision-making, thus reducing the chances that decisions will reflect their views (Latane et al., 1979).

A consensus cannot be trusted if it arises from reliance on others' positions without careful consideration of contamination by shared biases [5], based on the belief that we can better trust a consensus because multiple individuals have reached the same conclusion, particularly if these individuals differ significantly in a relevant variable. Public conformity, defined as people behaving consistently with norms they do not privately accept as correct, can potentially undermine true consensus. Such a consensus only offers the illusion of unanimity.

A series of experiments claiming the exact opposite to research findings on conformity had a big impact on the field of group dynamics:

When a majority of the group initially leans toward one position, their consensus tends to influence others in the group that hold a more moderate position. Both their positions and arguments make a polarization of group positions more likely, leading to a more extreme position. The consensus makes the majority arguments more persuasive: they are more numerous, receive more space for discussion and are usually presented in a more compelling fashion, as members of the majority use a less cautious style of advocacy. Thus, majority viewpoints are reinforced and advocates of the minority viewpoint are won over. Group interaction moves the group's average position in the direction favoured by the majority initially or to an even more extreme position. Group polarization toward a more extreme pole can be the consequence (Moscovici & Zavalloni, 1969).

An additional explanation is based on Festinger's social comparison theory (1954) that proposes that polarization is caused by group members competing with one another to endorse the socially most desirable viewpoints. Agreeing with a consensus (or going even beyond that) fulfils people's desire for holding the 'correct' views.

Almost all the studies in which polarization has been found were conducted in laboratory settings with ad hoc groups in which the outcome was almost always hypothetical. In naturalistic settings the polarization effect is less consistent. An explanation for these discrepancies might be that more permanent bodies establish norms about the communication structure, which might inhibit polarization (Brown, 2000).

Making a decision by establishing consensus rather than some voting procedure typically increases the effectiveness of decisions, i.e. decisions that have a high potential to be implemented in due course. Effective groups have a sound commitment to a clear goal and a combination of members' personalities, skills and roles; morale; and appropriate experience (McGrath, 1984).

Productivity of the group is increased if group members have a communication network that allows for maximum communication. Leavitt (1972) (see Hare 1982: 33) emphasized that "if the group's problem require that every member carry out of the group a desire to act positively on the group's decision, then it is imperative that every one accept, both consciously and unconsciously, the decision reached by the group".

Dialogue practitioner Hare (1982) has produced a set of guidelines for the consensus method, based on Quaker and Gandhian principles and results from laboratory experiments that have demonstrated the advantages of consensus over majority votes.

2. Participants must argue on a logical basis, giving their own opinion while seeking out differences.

3. Participants are asked to address the group as a whole, while showing concern for each point of view, rather than confronting and criticizing individuals.

5. It is essential not to press for agreement, but to hold more meetings if necessary and to share responsibility in the group for the implementation of the consensus (Hare, 1982).

Effective leadership (see 11.) can also be crucial for achieving consensus. Maier (1970) suggests a list of nine principles for the discussion leader to take into account:

1. Success in problem solving requires that effort be directed toward overcoming surmountable obstacles.

5. Disagreement can either lead to hard feelings or to innovation, depending on the discussion leadership.

6. The "idea-getting" process should be separated from the "idea-evaluation" process because the latter inhibits the former.

7. Choice-situations should be turned into problem-situations (a choice between two alternatives directs the energy toward making a choice and thus detracts from the search for additional / innovative alternatives)

8. Problem-situations should be turned into choice-situations (problem situations tend to block behaviour - the discovery of the first possibility tends to terminate the search for alternative and often better and innovative solutions. Decision-making requires both choice behaviour and problem-solving behaviour. It is desirable to capitalise on the differences and thereby upgrade each).

9. Solutions suggested by the leader are improperly evaluated and tend either to be accepted or rejected.

Minorities can influence the consensus reached by a majority in a group if they turn the processes of social influence to their own advantage. According to Moscovici (1976, 1980), a minority can undermine confidence in the majority consensus if they agree among themselves, remain consistent over time, and offer a positive social identity, i.e. being a member of a group that is highly regarded in the respective society with implications on individual self-esteem and behaviour. However, the minority's consistency may be interpreted as rigidity if taken too far and thus be ineffective. Moscovici suggests that minority dissent promotes systematic processing of information as the minority's suggested alternatives create uncertainty about reality as interpreted by the majority which in turn stimulates deeper reflection among majority members. More systematic processing can lead to private acceptance of attitude change but not necessarily to overt agreement with the minority.

Faced with the complex consequences of group diversity, groups should adopt the mode of "learning organisations", that is action should be based on available knowledge and take into account new knowledge generated in the process (Dodgson, 1993; Starbuck, 1983). For an effective decision-making process it is essential to construct a view of the negotiation process that is shared by all participants (Maznevski, 1994: 539)[6].

Group members should be made aware by the facilitating body of the communication process and of the role of communication in group performance. In addition, group members should be provided with specific information on the effects of the types of diversity relevant to their group. Understanding differences is the first step to managing them synergistically - acknowledging that the result of a cooperative effort between different parties can produce a stronger outcome than parties working in separation.

These findings refer to a need for meta-communication, i.e. space for communicating about the way the group communicates. Members of decision-making groups can improve their effectiveness by satisfying the preconditions for communication. Therefore, they should be provided with specific information on the effects of diversity on communication to understand effective and ineffective communication behaviours (Maznevski, 1994).

An inherent feature of decision-making processes in diverse groups is the expression and discussion of alternative or conflicting opinions and perspectives. Exposure to alternative views allegedly improves the learning process of the group and the quality of argumentation. However, dissent and disagreement often arouse negative emotional reactions, impeding the problem-solving process (Nemeth & Staw, 1989).

For decision-making groups, studies of how positive feelings influence negotiations are of particular interest. Group dynamics [7] stresses the role of cohesiveness – the result of the feeling of mutual regard and the commitment to the group and its activities. Without cohesiveness, the group will fall apart. It may translate into greater motivation to contribute and perform well as a means of gaining approval and recognition and thus lead to greater productivity of the group as a whole (Festinger, et al., 1950). Emotions are likely to be particularly beneficial for improving performance where flexible and creative thinking can lead to more effective resolutions than compromise (Jackson, 1996). A very high degree of cohesiveness, on the other hand, can have harmful effects.

When loyalty as a correlate of cohesiveness becomes the paramount aim, i.e. groups become more concerned with reaching consensus than with making the right decision, "groupthink" can be the result. Irving Janis (1972, 1982) applied the term groupthink to situations in which the drive to reach consensus at any cost outweighs the desire to adequately assess alternative courses of action and thus interferes with effective decision-making. Implementation of decisions is threatened.

There are several ways that might prevent groupthink without losing the benefit of cohesiveness (Janis, 1982; Smith, 1996). Firstly, to ensure adequate consideration of alternatives, open inquiry and dissent should actively encouraged. "Devil's advocates" could be appointed to ensure that weaknesses in the group's favoured decision are pointed out and that the opposing views are heard. Secondly, outsiders can be brought in to validate the group's decision and look out for shared biases. Different groups with different perspectives could work simultaneously on the same problem or the group could break into subgroups that take different points of view. Thirdly, to reduce conformity pressure, public votes should be the exception rather than the rule. The role of the leader should be minimised, and the expression of objections and doubts should be encouraged.

However, some of these solutions might prove impractical due to a lack of resources. Also, some of these "solutions" might again have undesirable and corrosive side effects such as prolonged debates, damaged feelings caused by too open criticisms or a lack of loyalty to the final decision due to break-up groups. This discussion shows again some of the complexities of group dynamics. As discussed above, a learning approach should be adopted to account for idiosyncrasies of each situation and specific group composition. Meta-communication should be encouraged to make group members aware of underlying group processes and possible implications.

When inter-group contact is established, pre-existing categories of, say, ethnicity or gender, are likely to be overlaid by other dimensions of categorization - for example, the new emerging category of a working group. Co-operation provides repeated opportunity to challenge certain stereotypes. Doise (1978) has argued that discrimination in regard to the original category will be reduced. A common identity becomes salient, i.e. more prominent, subsuming the - often problematic - division. This form of recategorization might be a crucial step in achieving general attitude change (Brown, 2000: 344).

Gender, age and ethnic group membership are the most salient characteristics of a person. Therefore, these characteristics have a relatively great impact on how we perceive and explain peoples' behaviour. The same behaviour can be perceived differently if shown by a man or a woman, a young or an old person, a white or a black person. Hence, categories of ethnicity, age and gender in decision-making groups are high-impact categories. This impact is enhanced even further as members of minorities are usually under-represented in decision-making groups (e.g. women; Indigenous Peoples; Black Americans in the US; youth; etc). This effect may reinforce the impact of stereotypes on people's perception and judgement so that their behaviour and opinions is even more perceived as due to them being a woman, black, young, etc. (e.g. Ashmore & DelBoca, 1986).

The goal of understanding multicultural and gender-specific group processes is both to maximise advantages such as multiple perspectives and creativity, and to minimise weaknesses such as mistrust and miscommunication.

Despite the fact that women are said to perform a more integrating style of communication and leadership than men and often act as "informal peacemakers" in cases of organisational conflict (Kolb, 1992), in many cases they are not in a position to fulfil their potential in decision-making groups. We have to consider a multitude of factors to understand this seeming contradiction.

Prejudice against women is one of the main factors. According to stereotypic beliefs women are less competent in management qualities such as initiative, strategic thinking, tactical skills, assertiveness and authority. In addition, the stereotypic belief about women being emotional possibly causes men to mistrust women and expect them not stay calm and rational in critical situations. Stereotypic beliefs that women are unpredictable and therefore less trustworthy are equally harmful (Kuepper, 1994; Hemmati 2000b).

Numerous psychological studies have demonstrated differences between women and men with in regard to their communicative behaviour and their styles of collaboration in groups (Dion, 1985). However, differences in overt behaviour of women and men are less prevalent than stereotypes might suggest. This is also true for men's and women's leadership behaviour (Friedel-Howe, 1990; Rustemeyer, 1988).

In gender-mixed meetings, women speak less often and more briefly, interrupt others less, and are interrupted more often. Women do more often express their feelings than men, who show a rather factual, technical and unemotional style of communication (Dion, 1985). This does not mean that women are in fact more emotional, but it can be a reason why they are less likely to be perceived as skilful and self-controlled strategists. The tendency of women to behave less competitively in groups often makes their contributions seem less important. Men are often more visible in teams because they tend spontaneously to take a leading role.

It has also been shown that the same kind of behaviour will be judged as perfectly ‘normal’ and ‘acceptable’ when shown by a man but will be judged differently when exhibited by a woman – not only as non-feminine but also as more extreme in its aggressiveness or assertiveness (Friedel-Howe, 1990; Ashmore & DelBoca, 1986, Wintermantel 1993).

Women and men also differ in regard to resources of power and ways of exercising power. Female strategies of exercising power are usually indirect. In organisational or micro-politics women often employ "soft" strategies such as showing friendliness, empathy, sympathy and loyalty or demonstrating devotion (Dick, 1993). These behavioural responses, however, are not very functional in order to succeed and having success attributed to oneself (Hemmati 2000b).

The dynamics in the communication process of diverse groups described above may lead to experiences of exclusion, having effects on the quality of instrumental exchanges, self-censorship, and withdrawal (Elsass & Graves, 1997). Vital information might get lost in the process, as judgments may not be expressed.

The salience of social categories, for example demographic characteristics such as gender, is dependent on the context and on the proportion of representatives in a group. One way to achieve that the category of gender loses some of its salience so that women would be less associated with gender stereotypes is by raising the percentage of women involved above the “critical level” of about 15 - 20%; (Friedel-Howe, 1990; Wintermantel, 1993). This rule of thumb also applies to other social categories / minorities.

However, group members do not belong just to one but to multiple relevant categories. Social categories overlap. Depending on the context, a different category apart from gender might come to the fore and thus influence the perception and judgment of other group members.

As stated above, group membership does not exist in a vacuum, but depends on the cultural context. Hofstede (1980, 1991) defines culture as "the collective programming of the mind, which distinguishes the members of one human group from another" (Hofstede, 1980: 1). Until recently, social psychological research into small groups has mostly been conducted in North America (and other Western societies) with the assumption, usually implicit, that findings are representative of other cultures. A focus on cultural diversity within North American society as well as experiences of an increasing number of international work groups in the organizational context have revealed this cultural bias more clearly.

According to Hofstede's results, the US is the most individualistic nation in a study comparing 53 nations and thus is the most atypical nation (Smith & Noakes, 1996). Individualism - Collectivism and Power Distance have been identified as key dimensions describing differences between cultures in social behaviour patterns.

Power Distance indicates the degree of maintaining a respectful distance from superiors to having more informal and equal relationships with superiors (Smith & Noakes, 1996: 480).

Groups in collectivist cultures are more concerned with long-term commitment, more deferential toward authority, and more concerned with harmony in the group, but are just as competitive with the outside (Triandis, 1989, 1990). Collectivist or interdependent cultures, like most in Asia, South America, and Africa, foster and reinforce views of the self in group terms (Markus et al., 1996). People from these cultures tend to see themselves as members of larger groups. In contrast, people in more individualistic cultures think of themselves in more idiosyncratic terms.

It is important to note that NGOs, business & industry, Indigenous Peoples, trade unions, and the like are also "cultures" (see Hofstede's definition above) which (can) differ in regards to these characteristics. Multi-stakeholder processes are ones of cultural diversity.

Another view of culture that can be helpful in working with MSPs conceives of culture as "the way we do things around here". members of a culture understand those ways and generally honour them, though without necessarily being conscious of doing so. Since MSPs bring people with different cultural orientations into interaction with one another, sensitivity to cultural differences is essential. More precisely, cultural sensitivity involves awareness of norms (standards of behaviour) and beliefs (assumptions about the way things are) and values (standards of importance) on which the cultural norms are based.

In addition to dilemmas facing mono-cultural groups, multi-cultural groups must initially overcome language problems and differing understandings of how to get to know one another. At a later stage, alliances of those who share cultural norms may form, which may impede effective decision-making. Reliance on stereotyped expectations also will be strongest during the early phase of group development. The challenge for the group is to move beyond stereotypical expectations, enabling individuals to become more aware of their own and others' assumptions and to use the information given. This process will be impeded if some team members experience their status (e.g. national, or stakeholder group) as privileged over others, thereby determining whose opinions are sought and acted upon. Feelings of inequalities and mistrust can originate from colonial history, historical antagonisms and economic dependence of some countries on others (Smith & Noakes, 1996: 491, referring to studies by Bartlett & Ghosal, 1987 and Ohmae, 1990). It is vital to take these underlying feelings into account, particularly in multi-stakeholder processes, as they impact more or less directly on the relations of different stakeholder groups.

Adaptation can best be accomplished by appreciating the cultural relativity of conceptions and practices within the group, while creating a sensitivity at the individual and group level, sometimes referred to as "valuing difference". If these problems are dealt with effectively, the group may then capitalize upon its diversity rather than be obstructed by it (Smith & Noakes, 1996: 495). People should become aware that most of what we experience as ‘natural’ is actually culturally specific. Again, meta-communication as part of the discussion and decision-making process can be suggested as a way forward, as well as other joint activities in which sharing about ones cultures is made possible in a more informal environment.

Another important issue is the choice of communication channels: Switching from face-to-face to electronic communication, for example, can provide a good basis for neutralising differences in status and personality, as related to gender, age and ethnicity. Nonverbal stimuli like personal characteristics, such as charisma, mimic and gesticulation can be displayed less effectively in the process of communication und thus be less successful in preventing others from contributing/contradicting (Kiesler et al. 1988; Hiltz & Turoff, 1993). Representatives of groups with less status, such as women or members of ethnic minorities would benefit primarily from this filtering of personal characteristics. Without participants being physically present, more attention can be given to the contents of the communicative act (Turkle, 1995; Geser, 1996).

Research suggests that IT supported communication is more suitable for producing heterogeneity. Thus, the internet could be the ideal tool for collecting suggestions to a given problem in a brainstorming or for getting an overview of the diversity of opinions on a given subject matter. If the goal is to convince others or to generate unanimity, the internet would not be the most useful tool (Geser, 1996; Kerr & Hiltz, 1982; Sproull & Kiesler, 1993).

Behaviour in decision-making teams reflects status and power differentials within the group. Numerous studies have investigated the effects of socially defined status, i.e. status based on age, gender, ethnicity, profession, income level, etc. Status usually correlates with demographic characteristics that are not necessarily relevant to performance in the group (Ridgeway, 1987).

There are wide-ranging behavioural differences between people of different social status: "Compared to those with lower status, higher status persons display more assertive non-verbal behaviours during communication; speak more often; criticize more; state more commands and interrupt others more often; have more opportunity to exert influence, attempt to exert influence more, and actually are more influential" (Jackson, 1996: 62). If lower status is not based on task-relevant attributes, differences in status appear to contribute to process losses because the expertise of lower-status members is not fully used.

In a review of formal group discussion procedures, Pavitt (1993) looks at the impact of formal procedures on small group decision-making. A formal discussion procedure (e.g. reflective thinking; brainstorming as a method of proposal generation; "nominal group technique" or NGT; devil's advocacy; dialectic inquiry) consists of an ordered sequence of steps for decision-making groups to follow in their discussions (see further discussion under 7.9. Implementation). To ensure more equitable discussion, formal group discussion procedures like NGT encourage equal participation for all members of the group regardless of power and status. NGT (Delbecq et al., 1975) is characterized by limitations of group discussions to exchanges between group members and an official group leader. After group members silently generate proposals on paper, the content of subsequent discussion is limited to the presentation and clarification of proposals, discouraging verbal clashes of differing ideas, the criteria for an ideal solution, and the extent to which proposals meet these criteria. Thus, NGT emphasizes individual decision-making over group interaction (see Pavitt, 1993: 219).

The method of reflective thinking (Dewey, 1910) which has been the starting point for the development of many other procedures for discussion and problem-solving, gives equal opportunity to all proposals. It proceeds through a sequence of decision-making steps:

4. evaluation of the extent to which each proposal meets the criteria for an ideal solution; and

According to Pavitt, "formal discussion procedures can be a force for democracy in decision-making, and this fact alone may warrant their employment in institutions in which democracy is valued" (Pavitt, 1993: 232). However, during the described stages of a formal procedure, appropriate chairing and facilitation needs to ensure equal and universal participation.

The fundamental task facing decision-makers is how to go about developing a prescription for action and get it implemented. Most of the studies on diversity in groups are conducted in laboratory settings in which teams have to come to solutions and agree on courses of action. In this regard the actual implementation of decisions receives little attention. If diversity of perspectives makes consensus reaching difficult, groups might try to accommodate opposing perspectives through compromise and majority rule instead of persisting to reach a creative solution by consensus. Reliance on compromise or majority rule may decrease group members' acceptance of the team's resolution and thus be an obstacle to effective implementation (Jackson, 1996).

Pavitt (1993) states that the impact of formal group procedures on the quality of the decision-making process is unclear. There is no firm basis for recommendations to practitioners. However, it is possible that formal procedures improve individual and, in turn, group performance. Referring to White et al. (1980), he concludes that groups using formal procedures tend do be more satisfied with their decision, and are therefore more committed to its implementation (Pavitt, 1993).

Further research suggests that agreements produced through mediation are characterized by very high rates of implementation (see below: 12. Mediation). According to Bingham (1987), the most significant factor in determining the likelihood of implementing a mediated agreement appears to be direct participation in the negotiation process of those with authority to implement the decision. For a mediated agreement to stay in effect over time, a monitoring group should be established to ensure implementation (see Baughman, 1995).

To conclude, implementation of a consensual agreement depends on a sense of ownership by all participants, be it those with high or low degrees of authority and power. To achieve and strengthen that sense of ownership, representatives participating in negotiations should also have opportunities to report back to their constituencies to ensure their backing and support.

In contrast to personal decision-making, a commitment to a particular course of action within a group does not necessarily provide decision-makers with the possibility of complete control over the consequences of their decision.

Vary et al. (unpublished manuscript) emphasize that the assumption of a shared understanding of the problem by all the stakeholders cannot be taken for granted. To ensure a jointly shaped and shared representation of the problem, the objectives and the goals of a group process, the following points should be kept in mind: first, the similarities and differences of the problem representation of the stakeholders should be acknowledged; and second, divergences should be discussed; and, third, the group should, if possible, aim to develop a common definition and view. Tools that help develop a common problem definition include addressing still open and possibly already decided questions and the background knowledge underlying the problem (Vary et al.). Hence a group needs to be open to explicitly revisit first stages of problem definitions if differences become apparent.

Humphreys (1998) points out that different representations also arise at different hierarchical levels. He describes the discourses employed in decision-making by identifying these different levels of representations of the issue under discussion. These levels set constraints on what can be talked about at the next level, thus establishing a common representation about the situation, by means of which prescription for action may be legitimated. The decision-making group, in order to act, must limit the number of problem representations until a common course of action is prescribed and can actually be embarked upon.

Therefore, working toward a shared definition of the problem needs to be the first step of all problem-solving procedures. Cultural differences and constraints due to hierarchical levels between representatives of the same stakeholder groups need to be kept in mind when working towards shared representation.

Leadership can be defined as a process in which a group member is granted the power to influence and motivate others to help attain the group goals (Forsyth, 1999; Smith & Mackie 1995). It is important to note that leadership is defined as a process (not a fact or stable position) and that the definition now prevalent in the social sciences explicitly acknowledges that leadership is being granted by the ones being led – without that, there is no leadership (but control, dictatorship, etc.) (see Neuberger, 1990).

Leaders have a disproportionate influence on team dynamics. Through their attitudes and behaviour, leaders may amplify, nullify, or moderate some of the natural consequences of diversity. They can shape informal norms and structure the process used for decision-making (Jackson, 1996: 70).

Effective leadership ideally involves both enhancing group performance and maintaining cohesion (task- vs. relationship-focused style). A high degree of quality and acceptance of the decision is needed for effective decisions. Therefore, the effective leader must recognize and distinguish between facts/ideas and feelings/biases - a distinction not easily made, as feelings/biases are often veiled behind made-up reasons or rationalisations. Hence, diagnostic skill is important another leadership requirement (Maier, 1970).

There has been substantial disagreement over the years in regards to the most effective leadership style with some studies favouring democratic over autocratic leadership and vice versa and others finding no significant effects of different styles at all (see review in Wheelan, 1994: 111). A learning approach, as described above, may be most likely to succeed if the group has a leader with strong leadership skills, i.e. being supportive and participatory but not too directive. Fiedler's "contingency model of leadership" (Fiedler, 1946) was a starting point for a lot of research, including new kinds of analysis, e.g. factor and cluster analysis, which have become possible through development of computer-based statistics. Based on that large body of research examining Fiedler's model, it seems that effective leaders vary their styles to meet the demands of the situation. The essence of good leadership may therefore be the flexibility to adapt to the needs of the group and the respective problem.

A number of studies on mediation has emerged over the last two decades. For example, the University of Washington's Institute for Environmental Mediation describes mediation as "a voluntary process in which those involved in a dispute jointly explore and reconcile their differences. The mediator has no authority to impose a settlement. His or her strength lies in the ability to assist in settling their own differences. The mediated dispute is settled when the parties themselves reach what they consider to be a workable solution" (Cormick, 1987 in Baughman, 1995: 254).

To attain consensual agreements, the focus in mediation lies upon collective rather than individual interests. Mediators often work with the different stakeholders individually to determine both the differences in values parties place upon the issue under discussion and the range within which each party is able or willing to negotiate. This enables the mediator to create alternative solutions and group discussions may then focus on commonalities rather than differences. It is therefore one of the tasks of the mediator to limit discussions to the extent that it appears to serve the achievement of consensus (Baughman, 1995).

A study of local governmental mediation in municipal boundary disputes in Virginia, US by Richman et al. (1986) describes dispute resolution processes and the role of mediation in settling these, providing valuable information for other negotiation processes. One aspect is an analysis of the non-explicit and non-rational dynamics involved in negotiations. Contrary to the dominant impression, negotiations are not necessarily a purely cognitive and emotionally 'cool' process in which people focus on their immediate stakes on the matter under discussion. Reality seldom fits this rational image.

In the negotiation process, the bottom line of each party is usually determined by a sense of vital interests or wants which it seeks to satisfy. Bottom lines are usually more ambiguous and vague and, though stakes are felt to be immutable, almost always difficult to be translated into concrete negotiable positions. According to Richman, one reason for this is that Western culture - capitalistic and competitive as it is - teaches not to identify and seek what one wants but to get as much as one possibly can. Attention is turned toward the external situation, and the focus lies on how to outdo the other side (Richman, 1986: 129). Here the mediator can manage the process by improving communication, and increasing comfort with the other: "As comfort with the negotiating relationship grows, so does trust. The bottom line payoff of mediation is that it nurtures the trust required as a foundation for the parties' moving to dialogue at the level of vital interests and wants." (Richman, 1986: 140).

Ronald J. Fisher (1997) analyses how interactive methods can influence decision-making processes and policy formation at the inter-communal and international levels. The method of interactive conflict resolution is defined as involving small-group, problem-solving discussions between unofficial representatives of parties (groups, communities, states) engaged in protracted social conflict, mediated by a third party (Fisher, 1997). It takes a social-psychological approach by asserting that relational issues (of misperceptions, miscommunication, distrust) must be addressed and that satisfactory solutions will only be attained through joint interaction. It is therefore seen as a complex process that allows for new mechanisms to develop to achieve constructive dialogue. The overall goal of the intervention is to transform a mutually hostile "win-lose" orientation into a collaborative "win-win" scenario. Numerous interventions described as dialogue can be considered applications of interactive conflict resolution (Fisher, 1997).

The methodology of dialogue can be regarded as a prerequisite to other processes, such as negotiation or problem solving. It puts emphasis on simply understanding the other party and acknowledging the conflict as a mutual problem. The goal is to discover new ways out of complex problems in which integrative solutions emerge that were not at first perceived by anyone, leading to consensus (see above; Fisher, 1997).

Seibold (1999) describes a range of procedures that help groups agree. He lists six rules for a non-competitive method of reaching a group decision, in which all members eventually agree to agree, notwithstanding individual preferences:

Based on the review of findings in this paper on effective communication and decision-making processes in groups of high diversity, considering and utilizing these rules to achieve consensus might have the potential to equalize participation and integrate many of the benefits of diverse decision-making groups.

Multi-stakeholder processes of any type are a novel approach to public participation, either in informing, monitoring, implementation or full decision-making processes. Creative and innovative solutions have to be found for a process that has not occurred in that same form before. It is important to appreciate the setting in which such processes should take place and optimum decisions would be reached. The setting is a competitive, knowledge-rich and complex world, in which our decisions affect the world in some momentary way to those of global and lasting proportions and in which it is often difficult to determine the consequences of our actions due to an increasing interconnectedness of people, organisations, corporations and states. Each decision, therefore, requires the use of knowledge of all kinds and maximum amounts. Even dialogues 'only' aiming at informing decision-makers deliver more than information given by each of stakeholder group separately. Through following the discussions taking place among stakeholders, rather than by asking each group individually, participants gain additional insights and more clarity about differences and commonalities between stakeholder groups. Multi-stakeholder processes have a great potential to assemble, transform, multiply and spread necessary knowledge and to reach implementable solutions.

Several points come out strongly from reviewing social and organisational psychological research:

There are strong arguments that any MSP should take a learning approach – towards its procedures and, in some cases, issues developing over time. Pre-set agendas, time-tables, definitions of issues, group composition, goals, procedures of communication and decision-making won't work. Participants also need to take a learning approach – being prepared to learn from and about others (new knowledge; overcoming stereotypes) and to 'teach' others about their views (assembly bonus effect). The same applies to chairpersons who need to be able to respond flexibly to a groups needs and developments.

Several points emerge as the guiding lines aim for group composition: 1. Aim for sufficient diversity of views. 2. Aim for equitable distribution of views, aiming to create a symmetry of power (at least of some sort). 3. Include at least two representatives of each stakeholder groups (gender balanced). 4. Do not invite people to represent more than one stakeholder group. 5. Avoid groupthink by checking that not a significant number of participants are dependent on another member (who would easily assume leadership and dominate the process).

The human being is not a 'homo oeconomicus'. Communication and decision-making are not merely rational processes and should not be to approached as such. People's feelings, attitudes, irrationalities in information processing, etc. need to be taken into account and respected. Discussions need to be based on factual knowledge – however, cultural values, ideals, fears and stereotyping (positive and negative) are human, too, and not to be ridiculed.

Trust building is an essential pre-requisite of successful groups of high diversity. There are a number of fundaments for building trust, above all honesty and integrity of participants, fairness, transparency and equity of the process. Necessary processes of overcoming prejudice and stereotyping need to take place before people are able to truly "dialogue"; they take time and concrete experience.

Employing formal procedures of communication and decision-making within groups of high diversity is certainly beneficial, for several reasons: There raise communication standards (being clear; speaking equitably; listening to others; taking each other seriously; etc.). They ensure that everybody gets the same amount of speaking time and help create more equitable discussion. They can thus help keep in check factors of social influence such as power and status, charisma, eloquence, etc. For MSPs, such procedures need to be agreed on by participants. One option is for the group initiating the process (ideally a stakeholder mixed group) to present possible procedures to potential participants and include discussion and decision-making on procedures in the (common) MSP design process. Formal procedures can make a group process more task-oriented, but successful groups need socio-emotional components, too. The challenge will be not to over-formalize a process and keep it flexible and at the same time reap the benefits of formal procedures.

Several reasons make it advisable to create space for meta-communication in MSPs. Groups increase their effectiveness if they work on the basis of an agreed set of rules – hence they need to communicate about the way they communicate. Meta-communication also allows space for dealing with problems which arise when members feel others aren't playing by the rules. In culturally mixed contexts (such as many MSPs), it enables participants to find out what are indeed cultural differences, perhaps more than we tend to believe.

[1] Thanks and acknowledgments go to
Dr. Dieter Beck, Deutsche Hochschule für Verwaltungswissenschaften Speyer:
www.dhv-speyer.de, Forschungsinstitut fuer oeffentliche Verwaltung bei der
Deutschen Hochschule fuer Verwaltungswissenschaften Speyer (FÖV)
www.foev.dhv-speyer.de
Dr. Humphreys, London School of Economics and Political Sciences, UK
Prof. Dr. Klaus Jonas, Technical University Chemnitz, Germany
Prof. Dr. Oswald Neuberger, University of Augsburg, Germany
Dr. Kai Sassenberg, Friedrich Schiller University Jena, Germany
Prof. Dr. Erich Witte, University of Hamburg, Germany
I also thank Minu Hemmati and Chris Church for their critical comments on a draft version of this article.

[2] "Social psychology: The scientific study of the effects of social and cognitive processes on the way individuals perceive, influence, and relate to others" (Smith & Mackie 2000. Glossary). Organisational psychology can be defined as a specialised and applied field of social psychology, focusing on the specific social context of organisations such as private sector companies, public administration bodies and voluntary sector organisations.

[3] Based on the review of research undertaken since Sherif & Sherif's (1953) summer camp experiments, Smith & Mackie (2000) summarize the conditions for successful cooperation between groups:

· a valued common goal, which eliminates competition for material and social resources

· repeated opportunities to expose and disconfirm out-group stereotypes

· successful results

· equal partners, at least for the task at hand

· shared social norms

[4] Social Identity Theory (Turner et al., 1987; Tajfel, 1979): Social Identity Theory (SIT) is one of the few social psychological theories dealing with individual (psychological) reactions to social / societal realities. SIT describes society as composed of social groups or social categories rather than individuals. SIT asserts that individuals partly perceive themselves as members of social groups or categories, and are being perceived as members of social groups by others. Different social groups differ with regard to their resources or status. Individuals are aware of their group membership and its social consequences. Low social status can lead to a threat of a positive social identity which individuals desire. Therefore, low status groups (minorities in power or number) strive for increased social identity. SIT describes various ways for individuals and for groups to achieve positive social identity, for example, through discriminating against groups of higher status. In other words: Group membership will dominate the individual perception of oneself and others and be the main source of identity. Attention will be focused on the conditions that sustain or modify the boundaries between groups (status and power of different groups, the legitimacy of these variables, boundaries between groups). Social context is of utmost importance (groups are in dynamic states of alliance or conflict). The context determines whether, at a given moment, people consider themselves as members of a specific group. For example, a certain social categorisation such as ethnicity can be more or less salient (i.e. obvious) and therefore more or less important for perception and behaviour in different social situations (Hemmati et al. 1999).

[5] A bias can be defined as a predisposition; an inclination or prepossession toward an object or view.

[6] This recommendation is consistent with symbolic interactionism (Blumer, 1969; Charon, 1989) and social constructionism approaches (Gergen & Gergen, 1988) , which assume that perspectives on reality are negotiated.

[7] Group dynamics as an area of research was born in the late 1930s - it endorsed beliefs in the collective strength of people and the value of cooperative interaction (Phillips & Wood, 1984).

Effective Communication and Decision-making, particularly processes in groups of high diversity.

A review of scientific research.

Table of Contents / Chapter 7 of the report