Compare original model to model with deleted loadings

Once you have a revised version with that weak loading deleted, it's time to compare it to the original theory-based model. If you were the developer of this measure, you would probably hope that the original version would fit better, since you would have carefully constructed items to measure your theoretical constructs. In practice, though, it can be tricky to know whether or not items are actually performing well, and sometimes you might want to drop items to create a stronger measure. For this exercise, we'll compare the CFIs and RMSEAs.