I am using Mplus to test multi-group measurement invariance on binary items with the help of the Mplus user guide. I fitted a free-baseline configural model and a fully contrained scalar model. I did model difference test and it's significant. So I am interested to find out where the partial invariance is. I got kind of confused now. My professor told me to constrain an item at a time and compare this model to the configural model to test whether that item is invariant across groups. It makes sense to me but I am not sure how to constrain an item when doing multi-group analysis in Mplus. Also according to Mplus user guide, it feels like I actually have to free an item at a time to find out whether the item is invariant across groups. If so, then when I compare this model to the free-baseline configural model, it sounds really weird. Can you give me some details how to test partial invariance and some resources? Thanks Angela!