What are the characteristics of traditional inspection and analysis methods in China?

1.t test is suitable for small sample comparison between two groups with homogeneous, normal distribution and variance measurement data. It includes three types: paired data, sample and mean, and comparison between two sample means. The calculation formulas of the three types cannot be confused.

2. The application conditions of T ′ test are almost the same as T ′ test, but when T ′ test is used in the case of uneven variance between two groups, the calculation formula of T ′ test is actually a modified formula of T ′ test in the case of uneven variance.

3. The application conditions of U- test are basically the same as that of T- test, except that U- test is used for large samples, T- test is used for small samples, and T- test can replace U-test.

4. Analysis of variance is used to compare the normal distribution and homogeneity of variance between groups. The commonly used methods include the comparison of multiple sample mean values in single factor grouping and the comparison of multiple sample mean values in two factor grouping. Analysis of variance first compares the total differences between groups, and then compares them in pairs if the total differences are significant. Q test or LST test was used for comparison between groups.

5.X2 test is the main significance test method of counting data. Used to compare two or more percentages (ratios). Common situations are as follows: four-grid table data, paired data, data with more than 2 rows *2 columns, and intra-group grouping X2 test.

The zero response test is used to calculate the data. When the occurrence probability of the experimental group or the control group is 0 or 100%, it is a special form of X2 test. It belongs to the direct probability calculation method.

6. Rank sum test, sign test and Ridit test are all nonparametric statistical methods, which are as simple, fast and practical as * * *. It can be used for the analysis of various non-normal distribution data, unknown distribution data and semi-quantitative data (grade data). Its main disadvantage is that it is easy to lose the information contained in the data. So those who have normal distribution or can be transformed into normal distribution by data, try not to use these methods.