Public Good Experiments with large Groups in a virtual laboratory

公共产品的特点往往是，它们需要由大集团提供，但每个代理人为提供公共产品所作的贡献所获得的回报(以MPCR边际替代率衡量)很小。在以前的公共物品实验室实验中，不可能研究这种组合(大群体，小MPCR)，因为MPCR从下到下的界限是1/N，N(受试者数量)的大小由于几个原因而受到限制。WE 2027/10-1项目的目的是首次在实验室条件下对具有较小MPCR的大团体进行公共物品实验。这是通过将四个独立的实验室合并为一个大实验室而实现的。到目前为止，该项目已经使我们能够成功地对8、60和100名受试者进行实验，MPCR值分别为0.25、0.04和0.02。到目前为止，这项工作的结果证明是非常有趣的，并在第一份工作文件中作了部分介绍。我们特别观察到，当群体规模没有显著影响时，强烈的MPCR效应。一个令人惊讶的结果是，MPCR值为0.04的100名受试者组成的大小组与MPCR值为0.25的小小组(N=8)表现出非常相似的合作水平。然而，这两种组合的直接比较相当困难，因为群体规模和MPCR参数都是不同的。然而，可以说，群体规模和MPCR效应之间肯定存在交互作用，这一点到目前为止还没有被认识到。继续项目的目的是通过更多的实验和新的设计来赢得对这两种效应相互作用的更多见解。在先前进行的关于购买排放权的实验中，使用了一组60名受试者。有一个相对稳定和恒定的平均贡献率为初始捐赠的20%-25%。然而，这些贡献遵循了与公共产品实验不同的模式。

Public goods are often characterized by the fact that they need to be provided by large groups but that the return each agent receives for their contribution for the provision of the public good (as measured by the MPCR-marginal rate of substitution) is very small. It has not been possible to investigate this combination (large groups, small MPCR) in previous laboratory experiments on public goods because the MPCR is bounded from below by 1/N and the size of N (the number of subjects) is limited for several reasons. The aim of project WE 2027/10-1 is to, for the first time, carry out public goods experiments with large groups with small MPCRs under laboratory conditions. This is made possible by combining four separate laboratories into one large laboratory. Up until now the project has enabled us to successfully carry out experiments with groups of 8, 60, and 100 subjects and MPCRs of 0.25, 0.04 and 0.02. The results of this work have proved to be very interesting thus far and are partially presented in a first working paper. We observed, in particularly, a strong MPCR effect whilst group size has no noteworthy influence. A surprising result was that large groups of 100 subjects with an MPCR of 0.04 showed a very similar level of cooperation to small groups (N=8) with a large MPCR (0.25). A direct comparison of these two combinations is rather difficult, however, since both the group size and the MPCR parameters were varied. However, it is possible to say that there must be an interaction between group size and MPCR effect which has so far not been recognized. The aim of the projects continuation is to win additional insights into the interplay of these two effects through more experiments and a new design. With the experiments previously carried out regarding the purchasing of emissions rights a group of 60 subjects was used. There was a relatively stable and constant average contribution of 20-25% of the initial endowment. However, the contributions follow a different pattern than in the public goods experiments.