I picked a "reference" energy spectrum which was far enough away from the Galactic Center so that we wouldn't expect to see WIMP annihilation signals there. In this case, I used galactic longitude (absolute) between 7-10 degrees, latitude (absolute) less than 3-degrees. The binning for the energy spectrum was the "standard" 10 Egret energy bins from 30-MeV up to 10-GeV. Then I took the same definition for energy spectra for angular bins at the galactic center and close to the galactic center. These spectra were divided, bin-by-bin, to obtain a spectral ratio as a function of energy. Thus, for the spectral ratio for energy bin 1-2GeV, the flux from the galactic center between 1-2GeV was divided by the flux between 1-2GeV from the "reference" region.

Finally, to obtain the double ratio, I divided each spectral ratio, bin-by-bin, by the diffuse ratio. So, if the Egret data is in perfect agreement with diffuse modeling, then we should have a flat line at 1 as a function of energy. Instead, we have some disagreement at the 10-20% level in angular bins near the Galactic center, and within 1-degree of the galactic center a disagreement which increases with energy to the level of 60% above 1-GeV. I think this indicates that we have 10-20% systematics associated with diffuse modeling. Presumably with better estimates of the hydrogen column density we can beat down this systematic error, and the spectral double ratios should become more consistent with unity (away from the Galactic Center).