Collaborative Research: Group Decision-Making in Strategic Market Games

许多重大的经济、政治、科学、文化和军事决策都是由群体作出的。决策小组有多种形式，包括家庭、委员会、董事会和立法机构。许多这样的群体决策，特别是经济、军事和政治决策，都是在高风险的战略环境中做出的，在这种环境中，群体决策的收益或成本取决于对手群体的决策。但为什么决策责任往往分配给群体，而不是个人？重要的信息是否由组成一个组的不同个体成员拥有？群体天生就是比个人更理性的决策者吗？在某些或所有环境中，群体比个人做出更好的决策吗？如果是这样的话，这是否反映了拥有更多信息的优势？这些都是有趣的经验性问题。由这笔拨款支持的研究比较了在财务风险和战略环境中进行的受控实验室实验中，竞争小组所做的决定与竞争个人所做的决定。这种环境是普通价值拍卖的环境，在这种拍卖中，投标人不知道他们竞标的物品的价值，但对该价值有一个或多个估计。许多自然产生的拍卖市场的特点是，对价值不确定的物品进行投标，由公司提交，这些公司已经确定了出价金额。相反，理论假设投标人是一个理性的个体，以前的大多数共同价值拍卖实验都是由个体决策者参与竞价，在这些实验中，群体决策者比个体决策者(在不同的会议中)拥有不同程度的更多、相同或更少的信息。将群体和个人决策与其偏离最低理性决策的量化度量进行比较，这些决策产生的预期利润为零，而不是负的。除了将群体决策与个人决策进行比较外，研究还将包括旨在洞察竞争群体在高风险战略环境中做出决策的过程的实验。在一种类型的实验中，信息量或市场规模将在实验期间改变，小组讨论将被录像和录音，以便随后转录和分析小组对这些治疗做出反应的推理。之前对个别决策者进行的实验发现，他们会通过提高出价来应对竞标者数量的增加。这与理性竞价理论形成鲜明对比，理性竞价理论暗示，竞价与竞标者的数量成反比。群体对这些治疗会有什么反应？在另一类实验中，计算机化的群体支持系统将被用来控制群体成员可以交换的信息内容，并实施替代决策规则，如简单多数投票和一致同意。这种类型的实验旨在揭示战略博弈中的群体决策是否可以通过使用这些系统来合理化，以及通过群体而不是面对面的会议进行互联网交流是否可以支持理性竞价。

Many important economic, political, scientific, cultural, and military decisions are made by groups. Decision-making groups have many forms including families, committees, boards of directors, and legislatures. Many of these group decisions, especially the economic, military and political ones, are made in risky strategic environments in which the benefits or costs that result from a group's decisions depend on the decisions made by opponent groups.But why is decision-making responsibility often assigned to groups, rather than individuals? Are important pieces of information possessed by different individual members who combine to form a group? Are groups inherently more rational decision-makers than individuals? Do groups make better decisions than individuals in some or all environments? And if so, does this reflect an advantage from having more information available? These are interesting empirical questions.Research supported by this grant compares decisions made by competing groups with decisions made by competing individuals in controlled laboratory experiments conducted in a financially risky, strategic environment. The environment is that of a common value auction in which bidders do not know the value of the items they are bidding on but have one or more estimates of that value. Many naturally-occurring auction markets are characterized by bids for items of uncertain value, submitted by firms in which groups have determined the amounts to bid. In contrast, theory has assumed that the bidder is a single rational individual and most previous common value auction experiments have involved bids by individual decision-makers.In these experiments, group decision-makers variously have more, the same, or less information than individual decision-makers (in different sessions). Group and individual decisions are compared with a quantitative measure of their deviation from minimally-rational decisions that yield the decision-maker zero, rather than negative, expected profit.Beyond comparing group decisions with individual decisions, the research will include experiments designed to provide insight into the processes that competing groups use to arrive at their decisions in risky strategic environments. In one type of experiment, the amount of information or size of the market will be changed during experiments, group discussions will be video- and audio-taped for subsequent transcription and analysis of the reasoning used by groups to respond to these treatments. Previous experiments with individual decision-makers have found that they respond to an increase in the number of competing bidders by increasing their bids. This is in sharp contrast to rational bidding theory, which implies that bids vary inversely with the number of bidders. How will groups respond to these treatments? In another category of experiment, computerized group support systems will be used to control the content of information that members of groups can exchange and to impose alternative decision rules such as simple majority voting and unanimous consent. The experiments of this type are designed to reveal whether group decisions in strategic games can be rationalized by the use of these systems and whether internet communication by groups, rather than face-to-face meetings, could support rational bidding.