Decentralization of continuous variables as interaction terms

1. According to Hou Jietai: the so-called centralization refers to subtracting the mean value of a variable from its expected value. For sample data, each observation value of a variable is subtracted from the sample average value of the variable, and the transformed variable is centralized
for your question, subtract the mean from each measurement.

2. Not necessarily, centralization is only for the convenience of explanation, and does not affect the regression coefficients Central treatment of regulatory effect of South Heart Network

3. Indeed, it is meaningless to analyze the interaction between continuous variables for the following reasons:
first of all, you need to understand the definition and significance of interaction. For this, you can refer to my other answer: http://..com/question/271101881.html
then, you can imagine that for continuous variables, there should be many different values, Each different value is considered as a category, so a continuous variable is defined by the statistical program as having many categories. It is not difficult to imagine that when two continuous variables are analyzed interactively, a lot of classification combinations will be formed. Among so many classification combinations, there will always be some classification combinations showing significance e to accidental reasons (this is the problem of multiple comparison). As long as there is one classification combination with significance, it will lead the statistical program to think that there is interaction between the two continuous variables, but in fact it is not. In addition, because there are many classification combinations between two continuous variables, it is difficult to explain the meaning of interaction, so there is no practical significance.

4. There can be interaction items, robot sister

5. The causal step method was proposed by Baron and Kenny (1986), and its test steps are divided into three steps. First, the regression of X to y tests the significance of regression coefficient C; second, the regression of X to m tests the significance of regression coefficient a; Third, the regression of X and m to y, testing the regression coefficients B and C & # 39; There is no significant difference between the two groups. If the coefficients C, a and B are significant, there is a mediating effect. In this case, if the coefficient C & # 39; If it is not significant, the mediating effect is called full mediation; If the regression coefficient C & # 39; Significant, but C & # 39& lt; c. This mediating effect is called partial mediation. The effect size of mediating effect is usually AB / C or AB / C & # 39; To measure. Although the causal step method is widely used, it has been controversial since it came into being.

6. Generally speaking, if there are interaction terms, you need to explain them first. If the interaction terms are not significant, compare the changes before adding the interaction terms.
you need to run at least three regression equations:
assume that the independent variable is x and the dependent variable is y, The adjusting variable is Z
the first regression equation is y = a1 + B1 (x)
the second regression equation is y = A2 + B1 (x) + B2 (z)
the third regression equation is y = A3 + B1 (x) + B2 (z) + B3 (XZ)
before that, I still need to ask, are these continuous variables or mmy variables?

7. This is used to analyze the regulatory effect.
after centralizing the independent variable and the regulatory variable, the interaction term can be obtained by multiplying them.

8. The interaction coefficient measures the influence of one variable on "the ability of another variable to influence the dependent variable". Usually, X1 is mmy. For example, X1 is male and female, X is IQ, and the dependent variable is income. Its coefficient measures the gender difference of the effect of IQ on improving income

9. Through the block function, you first move the main research independent variable B to the independent variable dialog box, and then click next of the block to switch to the next level, and then move variable a to the independent variable dialog box. In this way, under the control of a, you can study the influence of B on C separately

10. Logistic regression can be divided into three categories: one is binary logistic regression with dependent variable, which is called binomial logistic regression; the other is disordered multiple logistic regression with dependent variable, which proct is preferred, which is called multiple logistic regression. There is another kind of logistic regression in which the dependent variable is ordinal and multi classification, such as the degree of illness is high, medium, low, etc. this kind of regression is also called cumulative logistic regression, or ordinal logistic regression.