How to decentralize intermediary effect
Publish: 2021-04-12 07:49:13
1. 1. The mediating effect analysis does not need data centralization and standardization
2. Forced centralization or centralization, only the non standardization coefficient is different, the standardization system is the same
(provided by Nanxin)
2. Forced centralization or centralization, only the non standardization coefficient is different, the standardization system is the same
(provided by Nanxin)
2. This can be operated according to the path relationship. First, make a single path, and then make other indirect paths and direct paths.
3. Mediating effect is not needed, but moderating effect is needed.
4. The mediating and moderating effects can be realized by hierarchical regression in SPSS, that is, in which dialog box of multiple linear regression analysis, there is a
Block dialog box, you can move the independent variables and moderating variables to which dialog box one by one, and the regression results will show the changes of moderating effects
Block dialog box, you can move the independent variables and moderating variables to which dialog box one by one, and the regression results will show the changes of moderating effects
5.
If it is very convenient to use spssau as mediating effect, the operation methods are as follows:
(1) select [questionnaire research] & gt Mediating role]
6. First, we should construct the moderator, then do the mediating effect analysis and stepwise regression analysis. But it's better to use structural equation model Mediating effect and structural equation model analysis of the southern heart network
7. It can't be used. The result is too bad.
8. It's better for the two dimensions to act as an independent variable to mediate separately. Generally speaking, since they can be divided into two dimensions, it shows that the two dimensions are relatively independent to a certain extent, and the correlation is not so high, so the two dimensions can no longer form a high-order factor, so they can act as two independent variables to mediate separately. The following process can be used for any dimension:
assume that the independent variable is x, the dependent variable is y, and the intermediary variable is m. do the regression of y to x alone to get the coefficient C, which represents the total effect (indirect effect + direct effect). Do the regression of y to X and m to get the regression coefficient C & # 39; Then according to the definition of mediating effect, mediating effect is a * B. some special structural equations can directly test a * B, such as LISREL, Mplus, Amos, etc. if SPSS is used, then it is generally used to test a coefficient and B coefficient in turn. If they are significant, SIG & lt; If the mediating effect is significant at the same time, C & # 39; If it is not significant, it is completely mediated, that is to say, the effect of independent variable on dependent variable is completely mediated by intermediary variable m, and if it is significant, it is partially mediated. The above is the significance test of mediating effect. To evaluate the role of mediating effect in the whole effect, we usually divide the mediating effect a * B by the total effect C, and take this percentage as the effect value
if your two dimensions are highly correlated, for example, more than 0.6, then we can consider establishing a higher-order factor, that is, the score of the two dimensions can be added, At that time, you can add a total score to act as an intermediary
as SPSS is not specialized in structural equation, it is generally used as an intermediary to test in turn, and the total score is the factor score. The error estimation of SPSS is not as good as structural equation method, but the sequential test method is more strict, and the significant results are more convincing
it is said that your equation here does not have the proct term of x1x2. If you use the proct term, it is to analyze the regulatory effect. This and mediation are two different things. If you do the regulation of these two dimensions, you can directly multiply X1 and X2 as a variable, and then do the regression of y to x1, X2, X1 * x2. If the regression coefficient of X1 * X2 is significant, the regulatory effect is significant
hope to help you
assume that the independent variable is x, the dependent variable is y, and the intermediary variable is m. do the regression of y to x alone to get the coefficient C, which represents the total effect (indirect effect + direct effect). Do the regression of y to X and m to get the regression coefficient C & # 39; Then according to the definition of mediating effect, mediating effect is a * B. some special structural equations can directly test a * B, such as LISREL, Mplus, Amos, etc. if SPSS is used, then it is generally used to test a coefficient and B coefficient in turn. If they are significant, SIG & lt; If the mediating effect is significant at the same time, C & # 39; If it is not significant, it is completely mediated, that is to say, the effect of independent variable on dependent variable is completely mediated by intermediary variable m, and if it is significant, it is partially mediated. The above is the significance test of mediating effect. To evaluate the role of mediating effect in the whole effect, we usually divide the mediating effect a * B by the total effect C, and take this percentage as the effect value
if your two dimensions are highly correlated, for example, more than 0.6, then we can consider establishing a higher-order factor, that is, the score of the two dimensions can be added, At that time, you can add a total score to act as an intermediary
as SPSS is not specialized in structural equation, it is generally used as an intermediary to test in turn, and the total score is the factor score. The error estimation of SPSS is not as good as structural equation method, but the sequential test method is more strict, and the significant results are more convincing
it is said that your equation here does not have the proct term of x1x2. If you use the proct term, it is to analyze the regulatory effect. This and mediation are two different things. If you do the regulation of these two dimensions, you can directly multiply X1 and X2 as a variable, and then do the regression of y to x1, X2, X1 * x2. If the regression coefficient of X1 * X2 is significant, the regulatory effect is significant
hope to help you
9. It's better for the two dimensions to act as an independent variable to mediate separately. Generally speaking, since they can be divided into two dimensions, it shows that the two dimensions are relatively independent to a certain extent, and the correlation is not so high, so the two dimensions can no longer form a high-order factor, so they can act as two independent variables to mediate separately. The following process can be used for any dimension:
assume that the independent variable is x, the dependent variable is y, and the intermediary variable is m. do the regression of y to x alone to get the coefficient C, which represents the total effect (indirect effect + direct effect). Do the regression of y to X and m to get the regression coefficient C & # 39; Then according to the definition of mediating effect, mediating effect is a * B. some special software of structural equation can directly test a * B, such as LISREL, Mplus, Amos, etc. if SPSS is used, then it is generally used to test a coefficient and B coefficient in turn. If they are significant, SIG & lt; If the mediating effect is significant at the same time, C & # 39; If it is not significant, it is completely mediated, that is to say, the effect of independent variable on dependent variable is completely mediated by intermediary variable m, and if it is significant, it is partially mediated. The above is the significance test of mediating effect. To evaluate the role of mediating effect in the whole effect, we usually divide the mediating effect a * B by the total effect C, and take this percentage as the effect value
if your two dimensions are highly correlated, for example, more than 0.6, then we can consider establishing a higher-order factor, that is, the score of the two dimensions can be added, At that time, you can add a total score to act as an intermediary
because SPSS is not a special software for structural equation, it is generally used as an intermediary test method, and the total score is the factor score. The error estimation of SPSS is not as good as structural equation method, but the sequential test method is more strict, and the significant results are more convincing
it is said that your equation here does not have the proct term of x1x2. If you use the proct term, it is to analyze the regulatory effect. This and mediation are two different things. If you do the regulation of these two dimensions, you can directly multiply X1 and X2 as a variable, and then do the regression of y to x1, X2, X1 * x2. If the regression coefficient of X1 * X2 is significant, the regulatory effect is significant
hope to help you
assume that the independent variable is x, the dependent variable is y, and the intermediary variable is m. do the regression of y to x alone to get the coefficient C, which represents the total effect (indirect effect + direct effect). Do the regression of y to X and m to get the regression coefficient C & # 39; Then according to the definition of mediating effect, mediating effect is a * B. some special software of structural equation can directly test a * B, such as LISREL, Mplus, Amos, etc. if SPSS is used, then it is generally used to test a coefficient and B coefficient in turn. If they are significant, SIG & lt; If the mediating effect is significant at the same time, C & # 39; If it is not significant, it is completely mediated, that is to say, the effect of independent variable on dependent variable is completely mediated by intermediary variable m, and if it is significant, it is partially mediated. The above is the significance test of mediating effect. To evaluate the role of mediating effect in the whole effect, we usually divide the mediating effect a * B by the total effect C, and take this percentage as the effect value
if your two dimensions are highly correlated, for example, more than 0.6, then we can consider establishing a higher-order factor, that is, the score of the two dimensions can be added, At that time, you can add a total score to act as an intermediary
because SPSS is not a special software for structural equation, it is generally used as an intermediary test method, and the total score is the factor score. The error estimation of SPSS is not as good as structural equation method, but the sequential test method is more strict, and the significant results are more convincing
it is said that your equation here does not have the proct term of x1x2. If you use the proct term, it is to analyze the regulatory effect. This and mediation are two different things. If you do the regulation of these two dimensions, you can directly multiply X1 and X2 as a variable, and then do the regression of y to x1, X2, X1 * x2. If the regression coefficient of X1 * X2 is significant, the regulatory effect is significant
hope to help you
10.
Spssau can be used to analyze the regulatory role / mediating role:
click start analysis, and the system will automatically process the data according to the set data type
the suggestion analysis should be combined with the analysis suggestions and help manual instructions
Hot content