1. Pay attention to the corresponding relationship of events in contingency table and the determination of related values, so as not to be confused. The calculation formula of the observed value K of K2 is very complicated, and it is easy to confuse the meaning of some data when solving the problem. When substituting into the formula, there is an error, which leads to errors in the whole calculation result.

2. When describing the judgment results, we should pay attention to the accuracy of object selection, which should be a probabilistic judgment of the hypothetical conclusion, not something else.

Second, the basic idea of independence test

1. The basic idea of independence test is similar to the reduction to absurdity in mathematics, and its purpose is to confirm the credibility of the conclusion that "two classified variables are related".

2. First of all, it assumes that the conclusion is not valid, that is, "there is no relationship between the two classified variables". Under this assumption, the value of the constructed random variable 2X should be very small. If the observed value of 2X calculated from the observed data is very large, it shows that the hypothesis is unreasonable to some extent and can be used according to random variables.

3.2X to determine the irrationality of this assumption. If 2x >: 6.635, the irrationality of this hypothesis is 99%, then the reliability of the conclusion that "two classified variables are related" can be considered as 99%.

Third, the concept of independence test-2× 2 contingency table

Generally speaking, if there are two classified variables X and Y, their range is 12 12{}xxyy, and their sample frequency series contingency table is called 2x2 contingency table.