Collaborative Research: Experiments with Real World Social Networks

社交网络让我们既能了解世界，又能了解他人。因此，网络一方面可以促进技术进步和消费者的选择，另一方面也是社区信任的来源。我们的目标是分解这两种现象运行的渠道。PIs在哈佛大学对大约2500名学生进行了两次大型实地实验。在第一项研究中，他们将向随机选择的一组学生提供新的和复杂的产品样本，如PDA。他们将使用包括印刷和在线广告在内的各种处理方法来获得评估互动渠道的有效工具。通过测量初始和最终估值以及受试者对每种产品功能的了解，研究设计可以将社交学习从直接改变偏好的社会影响中分离出来，无论是从信息中学习(强社交学习)还是从观察朋友的评价中学习(弱社交学习)。在第二个学习项目中，PI将生成关于四家假想公司质量的私人信号，然后观察受试者如何在社交网络中汇总这些信息。通过使用相同的主题库，我们可以进一步完善我们的社会学习模型。这个项目的最后一部分是对信任的研究，将使用秘鲁的实地实验技术。我们通过一个新的小额信贷计划来衡量信任度，借款人可以通过从社区中愿意共同签署贷款的预定群体中找到一个赞助人来获得低息贷款。这项研究将帮助我们更好地理解人们是如何从他人的经验中学习的。研究设计将允许PI通过测量结果和信息传递将这种“社会学习”分为两个不同的部分。更好地了解社交学习的渠道可以帮助决策者设计更有效的干预措施，比如鼓励为退休储蓄计划缴费。

Social networks allow us to both learn about the world and learn about others. Networks can therefore promote technological progress and consumer choice on the one hand, and are sources of trust in communities on the other hand. Our goal is to decompose the channels through which both of these phenomena operate. The PIs conduct two large field experiments with about 2,500 students at Harvard University. In the first study, they will provide samples of new and complex products such as PDA's to a randomly selected group of students. They will use various treatments including print and online advertising to obtain valid instruments for estimating interaction channels. By measuring both initial and final valuations as well as subjects' knowledge about the functionality of each product the research design can separate social learning, whether it be learning from information (strong social learning) or learning from observing friends' valuations (weak social learning), from social influence which changes preferences directly. In the second learning project the PIs will generate private signals about the quality of four imaginary companies and then observe how subjects aggregate this information in the social network. By using the same subject pool we can further refine our model of social learning.The last part of this project is a study of trust that will use field experiment techniques in Peru. We measure trust through a new microfinance program where borrowers can obtain low-interest rate loans by finding a sponsor from a predetermined set group of people in the community who are willing to cosign the loan. This research will help us to better understand how people learn from the experiences of others. The research design will allow the PIs to break this kind of "social learning" into two different pieces by measuring both outcomes and information transmission. A better understanding of the channels of social learning can help policy makers design more effective interventions such as encouraging contributions to retirement savings plans.