Response amplitude operator (RAO) curves are commonly employed to assess the dynamic behavior of floating offshore structures in the frequency domain. There are multiple methods used to obtain RAOs for numerical models, scaled physical models, and full-scale tests. While for numerical modeling many studies detail the precise methods used, the literature around experimental RAO curves often do not detail them or leave methodological information incomplete. There exists inadequate experimental evidence in assessing the differences in results obtained by following different RAO generation methods from scaled physical testing. This paper addresses this gap by comparing two most popular RAO generation methods: the energy spectra (ES) and the cross spectral auto spectra (CSAS) method. These are experimentally compared on scaled semisubmersible and spar-buoy platforms in an ocean wave basin. Differences of heave and pitch RAOs generated by different methods are investigated. A method for reasonably collating multiple tests to create a representative RAO is also presented. RAO amplitudes vary significantly and how they decay off beyond certain frequencies is dependent on the method adopted to create them. This variation can be a source of significant uncertainty for floating structures for further analysis, design, control, or repair. Some RAOs (e.g., pitch) are sensitive to scaling and should be considered when converting scaled tests to full-scale equivalent. Detailing methods of RAO generation and comparing approaches of developing them can be important for crucial decisions from scaled physical testing of floating structures at design/development stages.