Demand generation and network loading models under evacuation conditions are crucial yet challenging components of evacuation studies. This study aims to assess network-wide impacts of behavioral response curves, which are the most popular evacuation demand generation models in the literature. System optimal dynamic traffic assignment (SO-DTA) formulation originally proposed by Ziliaskopoulos (2000) is used to model a simplified version of the Cape May County network in New Jersey. This multiple origin-single destination SO-DTA case study model is then studied under different demand generation conditions simulated through the use of various behavioral response curves (S-curves). The results show that the choice of the demand curve changes some important evacuation planning parameters such as clearance times and average delays. Thus it is shown that choice of response curve cannot be stated as just an assumption, but needs to be well calibrated to ensure its validity in terms of representing demand conditions in the study region