In practice, several competing liquefaction evaluation procedures (LEPs) are used to compute factors of safety against soil liquefaction, often for use within a liquefaction potential index (LPI) framework to assess liquefaction hazard. At present, the influence of the selected LEP on the accuracy of LPI hazard assessment is unknown, and the need for LEP-specific calibrations of the LPI hazard scale has never been thoroughly investigated. Therefore, the aim of this study is to assess the efficacy of three CPT-based LEPs from the literature, operating within the LPI framework, for predicting the severity of liquefaction manifestation. Utilising more than 7000 liquefaction case studies from the 2010–2011 Canterbury (NZ) earthquake sequence, this study found that: (a) the relationship between liquefaction manifestation severity and computed LPI values is LEP-specific; (b) using a calibrated, LEP-specific hazard scale, the performance of the LPI models is essentially equivalent; and (c) the existing LPI framework has inherent limitations, resulting in inconsistent severity predictions against field observations for certain soil profiles, regardless of which LEP is used. It is unlikely that revisions of the LEPs will completely resolve these erroneous assessments. Rather, a revised index which more adequately accounts for the mechanics of liquefaction manifestation is needed.