A Methodological Innovation in Creating Representative Samples of Organizations

由于没有组织数据库，甚至没有通常被认为有效的抽样程序，对组织进行概括的努力受到严重阻碍，通过抽样程序可以选择具有代表性的组织样本。解决这个抽样问题的一个潜在的有希望的方法是选择一个有代表性的个体样本，并以某种其他方式询问他们。这种方法是有希望的，因为人们对如何对个体人口进行抽样调查有很大的了解。然而，尚不能确定一个有代表性的个人样本是否会产生一个有代表性的组织样本。需要的是一些独立产生的组织分布，据信代表组织人口，可用作评估个人产生的分布的有效性的标准。这样的标准可以在国税局的文件中找到。它们可以提供一个全面的--可能是详尽的--组织清单，对照这些清单可以检查单独生产的组织分布。这项研究将考察一个大致与洛杉矶大都市相对应的地区。去年春天，通过电话对近1000人进行了抽样调查，得出了1000多个志愿组织和工作组织的信息。不同类型的组织出现在国税局与税务有关的记录中的频率将与从大洛杉矶调查中获得的分布进行比较。任何观察到的差异的重要性将使用Chow测试进行评估。

Attempts to generalize about organizations have been seriously hampered by the absense of an organization data base, or even a sampling procedure, generally recognized as valid, by which a representative sample of organizations could be selected. One potentially promising approach to this sampling problem entails selecting a representative sample of individuals, and asking them in some other manner. This approach is promising because a great deal is known about how to conduct sample surveys on populations of individuals. However, there is no known certainty that a representative sample of individuals will yield a representative sample of organizations. What is needed is some independently generated distribution of organizations, believed to be representative of the population of organizations, which can be used as a criterion to assess the validity of the individual - produced distribution. Such a criterion may be found in the files of the Internal Revenue Service. They may provide a comprehensive--and potentially exhaustive -- listing of organizations against which an individually - produced distribution of organiztions can be examined. This study will examine an area that corresponds roughly with metropolitan Los Angeles. A sample of nearly 1,000 individuals surveyed by telephone last Spring, produced information on over 1,000 voluntary and work organizations. The frequency with which different types of organizations appear in the tax-related records of the Internal Revenue Service will be compared to the distribution obtained from the greater Los Angeles survey. The significance of any observed differences will be assessed using Chow tests.