Hello,

I am conducting a distinguishability test using the Actor-Partner Interdependence Model with Mediator (APIMeM) in AMOS. Initially, I thought this process was similar to measurement invariance testing, but if it differs, I would like to confirm whether the following approach is correct:

  • In the first step, I created a saturated model where all regression paths were freely estimated for both partners (male/female).
  • In the second step, I applied equality constraints to the corresponding paths across partners (e.g., the path from "investment size" to "commitment" was constrained to be equal for both male and female partners).
  • I compared the two models using the chi-square difference test to evaluate distinguishability.
  • Is this approach correct for testing distinguishability in AMOS? Also, how does it differ from measurement invariance testing? I would appreciate any additional recommendations or insights.

    Thank you!

    More Selin Onaylı's questions See All
    Similar questions and discussions