Voting before discussing: Electronic voting as social interaction

Voting before discussing: Electronic voting as social interaction

ABSTRACT

A field study is presented which used voting before discussing (VBD) as a means of social influence and communication in a computer supported group interaction, rather than using voting as the final stage in a rational decision making process. The approach is based on a cognitive, three-process model of group interaction, which proposes that group cohesion and agreement arise primarily from normative rather than informational or personal influence. This initial investigation found that the VBD technique can result in higher agreement of group members with the decisions of the group, higher satisfaction with the computer-mediated interaction, higher satisfaction with group performance, and higher group awareness. The voting before discussion method may be useful in situations where agreement is an important group output, or where interpersonal conflict is creating problems in meetings.

KEYWORDS

Agreement, cognitive, CSCW, conflict, group interaction, GDSS, normative influence, CMC, social influence, voting

INTRODUCTION

Voting before discussing (VBD) is a proposed method of computer support for group interaction where electronic voting is used as a pre-discussion social interaction tool, rather than a post-discussion decision making tool. The VBD method uses computer-supported normative influence to establish a cohesive social environment before embarking on the risks of face-to-face discussion. The two major differences from traditional group support voting use are when voting occurs (before discussion rather than after), and why voting occurs (to support agreement rather than support decision making). The aim of this research was not to compare computer supported with non-computer supported meetings, but to explore whether voting before discussing was feasible for a realistic organizational problem, and explore the potential advantages and problems of the approach. This approach is novel, as media-lean computers are traditionally seen as deficient in social functions like building agreement and consensus (Adrianson & Hjelmquist, 1991; Fjermestad & Hiltz, 1999; McGrath & Hollingshead, 1991; McLeod,1992).

Socially generated agreement has significant group benefits. Making a wrong task decision may be the most temporary problem people face in interacting groups. Since many real life problems cannot be solved by reason alone (Daft et al, 1987), even the most rational person can expect "mistakes" every day. Most of these are repairable. But to lose a friend (or make an enemy) is more serious, as relationships usually continue over many tasks. A single enemy can affect everything a person does. Even worse is the disintegration of an important social group, like a religious, cultural or work group, as such groups usually persist as members come and go. Relationships and groups are contextual to task activity. Agreement seems more fundamental than task correctness, as groups must first agree on a position, before that position can succeed or fail. A group that cannot agree cannot even act, so task success is irrelevant. A group that makes task errors is in many ways a normal group, but a democratic group that cannot agree can fall apart, so that it is not a group at all.

The theoretical basis of the VBD method is a cognitive three process (OP) model, which suggests that for people interacting in groups, task information exchange usually occurs within a context of interpersonal relationships, and both in turn usually occur within a context of a unified group identity (Whitworth, et al, 2000). The cognitive processes used to resolve task information, build interpersonal relationships, and develop group unity are fundamentally different, because each presents a different type of problem. Any group discussion, while on the surface involving only the risk of task information errors of analysis, also risks negative social context outcomes. For example, personality clashes between individuals may lead to open conflict between individuals, which may then cause the task analysis to fail. Or if the group fails to agree it may split into two opposing factions. The OP model proposes that computer support for all three underlying processes is important, however supports for one process can hinder another. For example while increasing task information exchange improves decision quality (by increasing the group's domain of information), it also reduces member confidence and group unity (Sniezek et al, 1990).

For groups, successful online interaction involves more than factual information exchange and the rational analysis of task information. Contextual relational and group unity issues must also be supported. Rather than using computers to exclude social influence, the VBD method seeks to use computers to support the normal social function of normative influence. This approach can reduce interpersonal conflict and enhance unity, and is indicated where agreement is an ongoing problem with group interaction.