市场调查设计中的概念教程文件.ppt

 无回答对调查的影响主要表现在：由于提供答案的单元减少，使有效样本的数量减少，从而会增大抽样误差。 The main problem is that if the non-response rate is high, the bias can be so severe as to invalidate the survey results. 由于无回答的人通常具有不同于受访者的特征，如果无回答率很高的话调查结果就会出现严重的偏误，调查就会失败。 The second problem is that non-response reduces the effective size of the sample, since fewer units than expected provided answers. As a result the sampling variance increases.  本章的重点在于理解普查与抽样调查的不同点以及各自的应用条件。 在实际调查工作中，要根据不同的调查目的、不同的资源条件选择实施普查或抽样调查。 正确区分和定义目标总体和被调查总体也是实际调查中要注意的重要问题，这项工作将关系到能否正确选择或合理构造抽样框。  抽样框分为目录框和地域框，二者的联合体称为多重抽样框。 要掌握两种抽样框各自的优点、缺点和应用条件。 还要注意掌握目录抽样框的构成要素、抽样框的缺陷以及解决办法。 调查误差一般分为抽样误差和非抽样误差两类。 要掌握抽样误差的影响因素及测量方法、非抽样误差的种类和来源、各自对调查结果的影响等。?? Chapter 3 Introduction to Survey Design * 3.1 Two kinds of surveys-sample survey and census 3.2 Two kinds of populations-target and survey population 3.3 Two kinds of survey frame-list frame and area frame 3.4 Two kinds of survey error-sampling error and nonsampling error * 3.1 Two kinds of surveys-sample survey and census There are two types surveys-sample survey and census. The difference lies in the fact that, in a census, information is collected from all units of the population, while in a sample survey, information is only collected from a fraction (typically a very small fraction) The following list covers the most important factors when deciding between a census and a sample survey: Cost Timeliness Size of the population Small area estimation Prevalence of attributes Survey error Specialized needs Other factors * * 3.2 Two kinds of populations-target and survey population Target population: the population for which information is desired. Survey population: the population that is actually covered by the survey. It is the collection of units that the client is interested in studying. * Survey population: the population t