09-May-2022

Validity During Cfa Made Easy

Alright, let's say, you have a nasty model like this, a huge confirmatory factor analysis measurement model. And you'd like to produce some tables that look like this the validity and reliability table with the composite reliability, the AV e, MSB and ASV and a factor correlation matrix with the square root of the AV. E on the diagonal in order to establish discriminant and convergent validity and reliability. Well, to do this takes quite a bit of time, even with some tools already developed and. Let's see you get to this point you create these tables, and then you realize, oh, no it's, not good enough. I need to change the model.

Well, then I mean, you're, just in a rut. You've got to go and do it all over again. It takes another hour or two real pain in the rear, and it's subject to a lot of human error. So I've developed a new tool to do this in a flash let's. Do it let's see?

So you just run this as usual go to the output, put estimates scalars go to the correlations table, you're going to copy the. Correlations table copy and then go to my new stats tool paste, the correlations' table in a one of the validity master tab. Then go get the standard regression weights table copy that over paste that in f1 and then hit this button, and it's done. It builds these tables for you here's that convergent and screen validity table with C R as well and here's. The cofactor correlation matrix with the square root of AV, e on the diagonal and everything that is a problem is highlighted in red and described. Below so, for example, this AV e shows right here in K, 18 it's, a convertible edit. E issue where the AV e for C LED is less than 0.5.

So we know that it's got to be greater than 0.5. So that is a can any issue. Anyway, then if you want to change the model, you just go back and do that names run it again and then click to reset and before I, you can recreate those tables in a split.

Second, I hope, that's helpful, it's been very useful for me.