This chapter presents a structured organization of the various types of seismic-hazard estimates, approaches, scaling parameters, techniques, and data used in the estimation of potential earthquake sizes. This organization is designed to facilitate the use of multiple techniques, so that a greater amount of data can be incorporated into the size estimate. An earthquake size analysis begins with the determination of the type of seismic hazard to be estimated, such as characteristic, maximum, maximum credible, or floating earthquakes. The characteristic earthquake is defined as one that is characteristic of a particular fault or area. The maximum earthquake generally is defined as the largest to occur during a given time period, whereas the maximum credible earthquake is the largest that is reasonably physically possible, irrespective of the frequency of occurrence. Floating earthquakes occur along unidentified sources, e.g., along smaller faults adjacent to the major fault zones. The various types of estimates can be made at different levels of conservatism and probabilities of occurrence. Five approaches to earthquake size analysis are: historical earthquake, paleoseismic, source characterization, regional, and relative comparison approaches. For each of these, various parameters are used to scale earthquake size, including fault rupture length, fault rupture area, fault displacement, seismic moment calculations, and strain rates. Techniques include the specific correlations and equations used in the analysis. Data include the information collected and its associated uncertainties. The combined use of several approaches, scaling parameters, techniques, and data may reduce the overall uncertainty (or increase the confidence) in the analysis. Logic trees offer a useful format for presenting multiple estimates and uncertainties in an explicit manner. It is important to understand the type of earthquake size estimate made (e.g., local magnitude, surface-wave magnitude, or seismic moment), and to be internally consistent about the type used throughout the analysis, most importantly in the techniques and data. Current research in seismic-hazard analysis includes studies in intraplate regions (e.g., central and eastern United States) and on fold-related, subduction-zone, and volcanic earthquake sources.