PHD Discussions Logo

Ask, Learn and Accelerate in your PhD Research

Question Icon Post Your Answer

Question Icon

1 year ago in Statistical Analysis By Kanhu

Within a Structural Equation Modeling framework, what is the established method for constructing a Wald test or using model constraints to compare the magnitude of two specific indirect effects within the same model?

My model contains two parallel mediation pathways. It's theoretically important for me to argue not just that both are significant, but that one is stronger than the other. I know this involves a constraint or a Wald test, but I'm uncertain about the precise method to set this up and test it statistically.

 

All Answers (1 Answers In All)

By Isaac Answered 4 months ago

The established method uses the Wald test, often implemented via parameter constraints. First, label the two indirect effects' path coefficients meaningfully in your software. You then impose a constraint stating that the first indirect effect (the product of paths a1b1) equals the second (a2b2). The software then provides a Wald statistic (or a chi-square) testing this constraint. A significant result indicates the two effects are not equal, providing evidence for their difference in magnitude. I have found this crucial for advancing theoretical arguments about competing mechanisms.

Your Answer