The maximum % variation with changing FRD was calculated for two case: one with a boxcar only including the two outermost spatial pixels of the in-focus PSF and one with a boxcar extraction using all 6 spatial pixels of the  in-focus PSF. The expectation was that the boxcar extraction only including the edges/wings of the PSF would be more sensitive to FRD variations, so % variations would be larger toward the edge of the detector and at lower vertical (bluer) wavelength pixels. ExtractionMethods.png shows the results of this extraction as it would appear across the detector, where the downward triangles correspond to the wing-only boxcar and the circles correspond to the full boxcar. The global trend between these extractions seems to be quite similar, though it is possible that a better metric (i.e. SNR difference between images) may be more sensitive to PSF variations with FRD even using the different extraction methods; this is being investigated.