The American National Standards Institute/American Nuclear Society national standards 8.1 and 8.24 provide guidance on the requirements and recommendations for establishing confidence in the results of the computerized models used to support operation with fissionable materials. By design, the guidance is not prescriptive, leaving freedom to the analysts to determine how the various sources of uncertainties are to be statistically aggregated. Due to the involved use of statistics entangled with heuristic recipes, the resulting safety margins are often difficult to interpret. Also, these technical margins are augmented by additional administrative margins, which are required to ensure compliance with safety standards or regulations, eliminating the incentive to understand their differences. With the new resurgent wave of advanced nuclear systems, e.g., advanced reactors, fuel cycles, and fuel concepts, focused on economizing operation, there is a strong need to develop a clear understanding of the uncertainties and their consolidation methods to reduce them in manners that can be scientifically defended. In response, the current studies compare the analyses behind four notable methodologies for upper subcriticality limit estimation that have been documented in the nuclear criticality safety literature: the parametric, nonparametric, Whisper, and TSURFER methodologies. Specifically, the work offers a deep dive into the various assumptions of the noted methodologies, their adequacies, and their limitations to provide guidance on developing confidence for the emergent nuclear systems that are expected to be challenged by the scarcity of experimental data. To limit the scope, the current work focuses on the application of these methodologies to criticality safety experiments, where the goal is to calculate a bias, a bias uncertainty, and a tolerance limit for keff in support of determining an upper subcriticality limit for nuclear criticality safety.