A complex thing often has many factors that restrict and depend on each other. The purpose of variance analysis is to find out the factors that have a significant impact on this matter, the interaction between factors and the optimal level of significant factors through data analysis.
In practical application, it is often necessary to judge whether there are significant differences between several groups of observation data or processing results. For example, you want to know whether there are differences in the average monthly consumption level of credit card users in different regions, such as whether there are differences in multiple sets of data, and whether there are differences in the results of different treatments.
For example, whether there are differences between several drugs used to relieve postoperative pain, that is, the average duration of drug action, is actually to investigate whether there are differences in the results of different treatment methods (drugs applied to patients).
Extended data
There are several types of explanatory variables in ANOVA, such as research variables, control variables, adjustment variables and intermediary variables:
1. Research variable: it only appears in the explanatory model and is the most critical variable in the model. For example, the variable of sales volume in the marketing scene is the research variable;
2. Control variables: Except the research variable, any variable that has an influence on Y is a control variable, and the control variable here has no regulating effect on the research variable, but only plays the role of bearing the variance component. For example, education level and age have an impact on income, and age and education level may be related, but the change of age has no impact on education level and income;
3. Regulating variables: for example, the improvement of employee loyalty by the company's welfare fund investment is affected by the employee's wage income level, so the employee's wage income is a regulating variable;
4. Intermediary variables: If one variable influences Y through another variable, then the other variable assumes the role of intermediary variable. For example, the improvement of restaurant service level can bring customer satisfaction, and customer satisfaction can bring dining loyalty, so customer satisfaction is an intermediary variable.
Baidu encyclopedia-analysis of variance