This study quantifies the impact of two common ground motion (GM) selection methods, included in U.S. standards, on the seismic performance evaluation of steel special moment frames. The methods investigated are a “traditional” approach, herein referred to as the target maximum considered earthquake (TMCE) method, and a newer approach known as the conditional mean spectrum (CMS) method. The TMCE method selects GMs using the risk-based maximum considered earthquake (MCER) spectrum as the target spectrum, while the CMS method uses the CMS that anchors the MCER at multiple conditioning periods. Three special steel moment frames of 4, 8, and 16 stories are designed in accordance with ASCE/SEI 7-10, and their seismic performance is assessed with the nonlinear dynamic procedure prescribed in ASCE/SEI 41-13 using GMs selected and scaled in accordance with the aforementioned methods. A comparison of statistical parameters for the reduced beam sections and column hinges is conducted using the normalized demand-to-capacity ratio ( DCR), as the output parameter. The buildings are evaluated at the collapse prevention performance level for a far-field site located in Los Angeles, CA. In general, the CMS method results in lower DCRs of the frame components and smaller output parameter dispersion. In addition to the spectral shape, the demands are largely influenced by the spectral accelerations prescribed for each evaluated method. The consideration of collapse realizations is also documented as well as the existing and proposed statistical methods to account for these realizations. The study shows that the GM selection process can cause significant differences in structural response that may lead to different retrofitting decisions.