Collaborative Research on Team versus Individual Play

我们提出了一系列的实验，比较团队和个人在决策中的行为。 许多重要的决策环境涉及群体或团队，但许多相关的理论和实验工作并没有区分团队和个人的决策。 最近的研究表明，在战略互动（因此预测他人选择的能力）很重要的环境中，团队和个人行为之间存在巨大而重要的差异。 在这样的环境中，团队会产生积极的协同效应，表现得比各部分的总和还要好。 此外，与个人不同，团队表现出积极的学习迁移，将他们在一个游戏中学到的东西（例如，具有战略互动的设置）正确地应用于另一个相关游戏。我们的新实验跟进并扩展了这些结果。 (1)在早期的工作中，我们使用了团队成员之间的对话，以深入了解团队如何学习战略行动。 将采用可比较的程序来研究个人的学习过程，并提供与团队进行比较的基础。 (2)先前的工作表明，有意义的（而不是抽象的）背景下的主体的能力，他们在一个游戏中学到的其他相关的游戏中的重要作用。 提出了新的实验，以更好地理解这些过程。 (3)在真实的世界环境中，团队成员经常变化。 第三组实验是为了研究这如何影响学习策略性行动的速度。 (4)迄今为止的研究主要集中在密切相关的游戏之间的学习迁移。 我们计划研究更远距离的游戏之间的转移。 (5)心理学文献中广泛研究了隐性轮廓问题（Hidden Profile problems），即团队成员在不同的信息集合中明显无法汇集信息以实现更优化的决策。 我们计划研究提供明确的激励措施和更好的记忆私人信息的方法是否会缓解这些问题。

We propose a series of experiments comparing the behavior of teams and individuals in decision-making. Many important decision-making environments involve groups or teams, yet much of the relevant theoretical and experimental work does not distinguish between team and individual decision making. Recent research shows large and important differences between team and individual behavior in settings where strategic interactions (and hence the ability to anticipate others' choices) are important. In such environments teams generate positive synergies, performing better than the sum of their parts. In addition, teams, unlike individuals, exhibit positive learning transfer, taking what they have learned in one game (e.g. a setting with strategic interactions) and applying it correctly in another related game. Our new experiments follow up on and extend these results. (1) In earlier work we have used dialogues between teammates to gain insight into how teams learn to act strategically. Comparable procedures will be implemented to study the learning processes of individuals and provide a basis of comparison with teams. (2) Previous work shows an important role for meaningful (as opposed to abstract) context in subjects' ability to apply what they have learned in one game to other related games. New experiments are proposed to better understand these processes. (3) In real world settings, team memberships change frequently. A third set of experiments is proposed to study how this affects the speed of learning to act strategically. (4) Research to date has focused on learning transfer between closely related games. We plan to study transfer between more distantly games. (5) Hidden profile problems - the apparent inability of team members with different information sets to pool their information to achieve more optimal decisions - have been studied extensively in the psychology literature. We plan to study whether providing clear incentives and better devices for remembering private information will alleviate these problems.