In business survey methodology, increasing attention is given to the “unit problem”. The unit problem is the set of issues associated with the application of the statistical concept of the unit in business statistics. Various types of units may be defined and applied by statisticians, administrative registers and businesses themselves, which greatly complicates the production of statistics. Issues include the delineation of units, for instance if an enterprise is composed of several legal units. They include relating collection and observation units to the target units of business statistics, for instance if local units are used for collecting data on businesses. They also include identifying the units to be listed in business registers, especially if different types of units are used in business statistics, as is the case in the European Statistical System (ESS). Next to the enterprise, which is defined as an independent business actor in the economy, the ESS makes use of, for instance, the kind-of-activity unit, which is roughly defined as part of an enterprise carrying out a single industrial activity and administered as an entity in its own right. There are associated issues of sampling methodology, of coordination and integration of business statistics, of costs, efficiency and data availability, and more. Furthermore, all issues can be seen from the perspective of the errors they may cause in the statistics produced. Given the high relevance of the unit problem, the book The Unit Problem and Other Current Topics in Business Survey Methodology by Boris Lorenc et al. (eds) is a very welcome contribution to understanding such issues. However, the book, which comprises an edited and enriched selection of papers presented at the 2017 European Establishment Statistics Workshop, also covers other business survey topics, related or less related. Examples include sampling coordination, managing response burden, questionnaire design, using new data sources for price statistics, and data visualization. The book will not only benefit survey methodologists and producers of business statistics in general, but also the users of such statistics, widening their understanding of business statistics and their intricacies. Roughly half the book is concerned with the unit problem. The unit problem itself is explained in a chapter by Smith, Lorenc and van Delden, in which they consider unit errors in the context of the five dimensions of output quality as applied in the ESS.

中文翻译：

---

书评

在业务调查方法中，人们越来越关注“单位问题”。单位问题是与单位统计概念在商业统计中的应用相关的一组问题。统计人员，行政登记册和企业本身可以定义和应用各种类型的单位，这极大地增加了统计数据的生成。问题包括单位的划定，例如，如果一个企业由几个法律单位组成。它们包括将收集和观察单位与业务统计的目标单位相关联，例如，如果使用本地单位来收集有关业务的数据。它们还包括识别要在业务注册中列出的单位，尤其是在业务统计中使用不同类型的单位时，就像欧洲统计系统（ESS）一样。在被定义为经济中独立业务参与者的企业之后，ESS会利用例如活动种类单位，该单位被粗略地定义为从事单一工业活动的企业的一部分，并且本身作为一个实体进行管理。存在抽样方法，业务统计数据的协调和集成，成本，效率和数据可用性等相关问题。此外，可以从产生的统计信息中可能引起的错误的角度看待所有问题。考虑到单元问题的高度相关性，鲍里斯·洛伦茨（Boris Lorenc）等人撰写的《商业调查方法论中的单元问题和其他当前主题》一书。（eds）是对理解此类问题的非常欢迎的贡献。然而，该书包括在2017年欧洲企业统计研讨会上发表的精选论文的编辑和丰富选集，还涵盖了其他相关或较不相关的商业调查主题。示例包括采样协调，管理响应负担，问卷设计，使用新的数据源进行价格统计以及数据可视化。这本书不仅将使调查方法学家和一般的商业统计产生者受益，而且还将使此类统计的用户受益，从而拓宽了对商业统计及其复杂性的理解。大约有一半的书涉及单位问题。Smith，Lorenc和van Delden在一章中解释了单位问题本身，他们在ESS中应用的输出质量的五个维度中考虑了单位误差。其中包括在2017年欧洲企业统计研讨会上发表的精选论文的编辑和丰富选择，还涵盖了其他相关或较不相关的业务调查主题。示例包括采样协调，管理响应负担，问卷设计，使用新的数据源进行价格统计以及数据可视化。这本书不仅将使调查方法学家和一般的商业统计产生者受益，而且还将使此类统计的用户受益，从而拓宽了对商业统计及其复杂性的理解。大约有一半的书涉及单位问题。Smith，Lorenc和van Delden在一章中解释了单位问题本身，他们在ESS中应用的输出质量的五个维度中考虑了单位误差。其中包括在2017年欧洲企业统计研讨会上发表的精选论文的编辑和丰富选择，还涵盖了其他相关或较不相关的业务调查主题。示例包括采样协调，管理响应负担，问卷设计，使用新的数据源进行价格统计以及数据可视化。这本书不仅将使调查方法学家和一般的商业统计产生者受益，而且还将使此类统计的用户受益，从而拓宽了对商业统计及其复杂性的理解。大约有一半的书涉及单位问题。Smith，Lorenc和van Delden在一章中解释了单位问题本身，他们在ESS中应用的输出质量的五个维度中考虑了单位误差。