I am estimating peer effect of students on their GPA score using Friendship network model. The model suffers from reflection problem (Manski,1993),

In short, I have a model after local transformation like the following

(I-G)Y= (I-G)GY +(I-G)X +(I-G)GX+ (I-G)e

Where "I" is block diagonal identity matrix and 'G" is block diagonal of the normalised adjacency matrix representing friendship network at the classroom level. Y is GPA and X is some exogenous variable. So 'GY' is the average GPA of all the friends of an individual and GX is the average contextual effect of the friends. GY by definition is endogenous therefore theory suggest G_Square_X which represent friends' friend characteristic can be used as IV for GY.

I am confused to use exploratory data analysis on which variables such as a scatter plot. I mean Shal I use Y and GY or shall I use transformed model (I-G)y and (I-G)GY.