Hello everyone!
This is the Yinaoyun scientific research circle, I am sister Miaojun~
In this tweet, we will introduce to you a software for dealing with structural equation modeling - Amos.
01
Summary
What is Amos?
Amos was developed by James L. Arbuckle, whose full name is Analysis of Moment Structures.
Amos has been a member of the SPSS family since version 6.0. Can work independently before version 5.0. The name Amos (Amos) is really creative, because it was originally the name of the Hebrew prophet in the 8th century BC, and it also means the book of Amos in the Old Testament.
Amos is suitable for Analysis of Covariance Structures (Analysis of Covariance Structures), a software that deals with structural equation modeling (SEM).
SEM is suitable for exploring and analyzing complex multivariate data. Amos can analyze many variables at the same time and is a powerful statistical analysis tool.
Amos builds models (path diagrams) by means of visualization and mouse dragging to represent the relationship between variables, without having to write program instructions from the beginning to the end, and it is done in one go, which really improves the efficiency of data analysis.
At the same time, the SEM established by Amos will be more accurate than the standard multivariate statistical analysis.
In addition, Amos allows us to check whether the data fits the established model, and to perform model exploration (gradually build the most appropriate model). -Baidu search
Download of Amos: You can download it from the official website Version: 21.0 22.0 23.0
Official website URL:
https://www.ibm.com/products/structural-equation-modeling-sem?from=singlemessage
02
Main page and function introduction
Main page main interface
If you feel that the position of the middle view is too small, you can click the view-interface properties-paper size in the upper left corner to set it.
components
This piece is the essence of amos. All the components are here, and all of them are dragged to the main interface in the middle by the mouse to build the structural equation model.
If you want to know what function each component represents, just place the mouse on the component, no need to click, lightly place it on it, and the function of the component will be displayed. Next, we will introduce the functions of each component in turn in the intermediary model introduced.
03
Mediator effect
Mediator plays an important role in researches on psychology, education, society and management.
If the independent variable X has a certain influence on the dependent variable Y through a certain variable M, then M is called an intermediary variable between X and Y or M plays an intermediary role between X and Y (as shown in Figure 1, where the circle represents the variable can be explicit or latent variables).
The significance of mediation research is to help us explain the mechanism of the relationship between independent variables and dependent variables, and also integrate the relationship between existing variables.
According to whether there is interaction among multiple mediator variables, multiple mediator models can be divided into single-step multiple mediator model and multiple-step multiple mediator model.
The single-step multiple mediation model means that there is no mutual influence between multiple mediator variables (removing the M1→M2 path in Figure 2 is a single-step multiple mediation model), also known as the parallel multiple mediation model.
The multi-step multiple intermediary model refers to the mutual influence among multiple intermediary variables, and multiple intermediary variables show sequential characteristics, forming an intermediary chain (X→M1→M2→Y path in Figure 2), also known as chain multiple mediation model.
04
Intermediary inspection steps
Building a simple mediation effect model with amos
Amos can be used for confirmatory factor analysis, model establishment of mediation effect test & moderation effect test. Today we will first introduce how to use amos to do simple mediation effect test.
A: The square represents the variable, we need to click with the mouse, and then draw the components we need in the central white paper area. Then use arrows to connect the relationship between them, that is, independent variable-dependent variable; independent variable-intermediary variable-dependent variable.
B: Name the variable: Double-click the component, and the property box will pop up, just enter the name of the variable directly.
C: After the input is complete, we need to add errors to all other variables except the independent variable. Click the component selected below, and drag it to other variables except the independent variable. Only one can be dragged at a time.
Similar to the picture below, that is, the addition is successful
After the addition is successful, you need to name the error. There is a quick way to name all errors, as shown in the figure below. Plugins-name-unobserved-variables is enough.
In addition, if you want to move the variables that have been added, you need to click the second of the hand first, select the variable or path that needs to be moved, and then click the truck to move the added variables. After the transportation is completed, click the third of the hand , release the selection, and the handling is completed.
Subsequently, the complete simple mediation model of the obvious variables was established successfully.
However, if we want to make a simple mediation model of obvious variables, we can do it in spss, and we can use a plug-in in the regression analysis of spss-process to complete it, and the method is simple and easy to operate.
Compared with amos, process is more suitable as an intermediary model of obvious variables. The advantage of amos is that it can be used to test the mediation effect of latent variables.
In amos, circles represent latent variables, and we build a simple mediation model of latent variables.
C: Import data and run model results. Validate the fitted mediation effect.
In addition, Amos can also do moderation effect analysis, as well as mediation moderation and moderation mediation effect. However, complex mediation effect analysis amos cannot bear, amos can only do simple mediation effect test, and complex mediation model has to pass M-Plus.
Finally, we also recorded a small video explaining the content of this issue. We hope that everyone can become more familiar with the operation of amos to test the mediation effect through the video explanation.
If you are interested in related issues, welcome to discuss with us!
references:
[1] Fang Jie, Wen Zhonglin, Zhang Minqiang, & Sun Peizhen. (2014). Analysis of Multiple Mediating Effects Based on Structural Equation Modeling. Psychological Science (03), 225-231. [2] Wen Zhonglin, & Ye Baojuan. (2014
) . Mediation Effect Analysis: Method and Model Development. Advances in Psychological Science, 022(005), 731-745.
[3] AMOS official website: https://www.ibm.com/products/structural-equation-modeling-sem?from =singlemessage
Author｜Benben
Typesetting｜Qiuqiu
Proofreading｜Sister Blue Eucalyptus Miaojun