1, massive observation method
This is the basic method of collecting data in the process of statistical activities (that is, the stage of statistical investigation): that is, observing and studying enough individuals in the phenomenon population under study to understand the regular quantitative characteristics of the population. The mathematical basis of a large number of observations is the law of large numbers, which means that although each individual is influenced by different accidental factors and has differences in quantity, they can cancel each other out and present a stable regularity on the whole. Therefore, only when a sufficient number of individuals are observed, the comprehensive results of observation will tend to be stable, and general conclusions will be given based on a large number of observed data. All kinds of statistical investigation methods belong to a large number of observation methods.
2. Statistical grouping method
Because of the complexity, difference and multi-level of the phenomena studied, we need to study the phenomena in groups or classifications in order to explore the differences between different groups or classifications on the basis of homogeneity. Statistical grouping plays an important role in the whole process of statistical activities. In the stage of statistical investigation, different kinds of data can be collected by statistical grouping method to improve the sample representativeness of sampling investigation (that is, stratified sampling method). In the stage of statistical arrangement, various data can be processed and stored by statistical grouping, which provides the basis for compiling distribution series; In the stage of statistical analysis, we can divide the phenomenon types, study the overall internal structure, compare the differences between different classes or groups (significance test), and analyze the correlation between different variables by statistical grouping method. The statistical grouping methods in statistics include traditional grouping method, discriminant analysis method and cluster analysis method.
3. Comprehensive index method
The quantitative characteristics of statistical research phenomena are reflected by statistical comprehensive indicators. The so-called comprehensive index refers to the categories and numerical values used to reflect the quantitative characteristics and quantitative relations of the studied phenomena as a whole, and the common ones are total index, relative index, average index and symbolic variation index. The comprehensive index method plays a very important role in statistics, especially socio-economic statistics, and is the core content of descriptive statistics. How to record, describe and reflect the quantitative characteristics and quantitative relations of the studied phenomena most truly and objectively is a major topic in the theoretical research of statistical indicators.