A technique to separate the phase-induced penalty of a cascade of optical filters into dispersion, dispersion slope, and higher-order terms is introduced and its impact on the proper design and engineering of high-speed Dense Wavelength Division Multiplexed (WDM) optical systems and networks is demonstrated. As the currently deployed fiber optic systems and networks strive for higher speeds to respond to the growing global needs for more bandwidth, the impact of physical layer impairments (such as optical filter dispersion slope) which were not significant at lower speeds are now becoming increasingly important and worth looking at. In this article we demonstrate that at speeds of 40 Gbps and beyond, where the next generation systems will be operating, optical filter dispersion slope is at least as important as filter dispersion. As a result, separating the above contributions and accounting for each using the described modeling technique proves to be an effective way for designing and engineering such systems.