Exploring Conditional Indirect Effects (Methodolory for AMOS)
Description
Many researchers struggle with testing conditional indirect effect models (aka moderated mediation). They often don’t know how to properly set them up or analyze them using AMOS. This repository hosts supplementary material for the explanatory study titled "Exploring Conditional Indirect Effects: A Structural Equation Modeling Approach," which breaks down—in a simple and practical way—how to run these analyses step by step using AMOS (a structural equation modeling [SEM] software). The original study: 1 Clarifies the key differences between mediation, moderation, and moderated mediation (where "if... then..." conditional effects come into play). 2 - Provides a hands-on guide for setting up your model in AMOS and interpreting results, complete with clear examples. The material shared here does not replace the original paper but serves as a visual aid, featuring AMOS screenshots to enhance understanding.
Files
Steps to reproduce
Steps for conducting a CIE analysis using CB-SEM. 1-Verify the existence of a mediating effect 2- Create the syntax to compute the set of CIE statistic values 3-Evaluate and interpret data 4 -Estimate the CIE statistic value (interacton effect) A prerequisite for conducting moderate mediation studies is to first determine whether there is a mediation effect. The calculation of all parameters of interest to test for mediation is virtually straightforward and automatic with CB-SEM techniques. However, extant literature recommends calculating the parameters using bootstrap procedures with at least 1,000 resamples, which is a reasonable number to ensure the robustness of the model parameters and their respective p-values when using CB-SEM techniques. The subsequent step involves generating a series of statistical estimates to determine the existence of a conditional indirect effect. To accomplish this, an inferential statistical procedure must be employed to confirm the presence of a CIE across various values of the moderator variable. Additionally, it is also crucial to generate and compute the value of the interaction value (IMM) to determine whether this index is statistically significant. However, it should be noted that the computation of the IMM and the CIE across different values of the moderator and the MMI cannot be obtained automatically in CB-SEM. Therefore, some syntaxes will have to be created and typed manually into the software to generate them. We explain how and where to manually configure the syntax. Finally, to verify that the MMI score is different from zero, an inferential test or confidence interval was employed. P-value minor to 0.05 of MMI indicate that the interaction effect is statistically significant.
Institutions
- Universidad de Costa Rica