Error Analysis of Sampling Frame in Sample Survey
Li, Zhengdong, Studies in Sociology of Science
Sample survey is the process of observing, describing, and analyzing a sample of units drawn randomly from the survey population to draw an inference. In sample surveys, the sampling frame is a major component in the entire sampling design, and it plays an important role in sample surveys (2). Ideal sampling frame means that every element in the population is recorded and only recorded once, excluding other elements which are not in the population. In practice, this ideal condition is very rare. As for probability sampling, essentially it is to select a subset of some units from a finite set of all units, with the probability of a subset being selected known. Sampling frame is a unit of finite population, however, the population of sampling frame units do not necessarily equivalent to the population of data collection3 (Li Jinchang, 1996:30-32). As a way to identify the elements in the population, the sampling frame usually includes a large number of additional information useful for sampling design in its composition, which often has a strong influence on sampling design. The structure of sampling frame, together with its additional information and the quality of this information will determine the type of survey sampling design and the estimation procedure and its deviation.
1. SAMPLING FRAME AND ITS TYPE AND STRUCTURE
Simply put, sampling frame is the sampling range or the list of sampling units, and it is the list of all sampling units in the survey population. For the sampling frame, different researchers have different cognitive styles. In order to facilitate the implementation of probability sampling, sometimes the population will be divided into several parts that are non-overlapping and exhaustive, each part is called sampling unit, sampling unit is not necessarily the individual. In this sense, the sampling frame is actually the roll or list of all the sampling units. In preparing the sampling frame, if the sampling unit divided too finely, a lot of manpower resources and financial resources will be spent. For example, it is difficult to prepare the sampling frame of village committees and neighborhood committees in a province. At this time sampling units can be graded. For instance, to obtain the sampling frame of all village committees and neighborhood committees in a province, county can serve as the primary sampling units, so that it is easy to obtain the sampling frame of all county-level units in a province. Similarly the sampling frame of all townships in a county and the sampling frame of all villages in a township can be obtained easily.
American scholar Judith and William defined the sampling frame systematically based on the past concept. In their view, the sampling frame consists of the materials, procedures and methods to identify, differentiate, and approach the elements of a target population. Sampling frame is composed of a set of finite units, from which probability sampling can be conducted. The mechanisms or rules linking the sampling frame units and the population elements are an integral part of the sampling frame. The sampling frame also includes auxiliary information (such as size measurement, demographic information), which is used for special sampling techniques (such as stratification sampling and probability proportional to size sampling) and special estimation techniques (such as ratio estimate or regression estimate). the concepts involved in the sampling frame are: (1) The target population is a finite population with its elements identifiable; (2) The sampling is a set of selected units, but this set may not belong to the target population; (3) The linkage mechanism must exist between the target population and the set of selected units; (4) In order to obtain information from the elements, they must be able to distinguish and find; (5) There may be a variety of linkage forms between target elements and sample sets. The linkage procedure determines the type of sampling design and the estimation procedure in a survey; (6) For some sampling designs and estimation procedure, the auxiliary information about the population elements is required. …