
S1E12: Measurement (Non)Invariance -- Can We Ever Fail to Not Incorrectly Reject It?
Quantitude
00:00
The Omnibus Model Fit
The omnibus test is, have we done harm in our ability to reproduce the data that we observed? Is there a significant decriment a model fit? So when i do this myself, i do look at that omnibus test. And what those give you in this case is, if you were to free these two loadings, what would the improvement and fit be? Alltha says is, at least one loading is significantly different across the two groups. Well, here's the problem, well, which one? How do you know? Right? Because you get this wonderful thing called partialn ras, right? The simple question is, all loadings are equal, or all
Transcript
Play full episode