Paleoliquefaction research in the last 15 years has greatly improved our ability to interpret the paleoseismic record throughout some large geographic areas, especially in regions of infrequent large earthquakes. Paleoliquefaction studies have been used extensively in the central and eastern U.S. to assess seismic hazards, and could be used elsewhere to good purpose because paleoliquefaction studies in some field settings can reveal more than other methods, such as fault studies, about the prehistoric strength of shaking and earthquake magnitude. We present guidelines for the conduct of a paleoliquefaction study in continental deposits, mainly in terms of the geologic/seismologic setting and geotechnical properties, because a successful interpretation requires factors from all these disciplines. No single discipline suffices alone. Their interactions must be appreciated in order to understand why seismically induced liquefaction features are found in some locales and not in others. The guidelines that we present also relate to three primary issues for which liquefaction features are especially useful for interpretations: Where was the tectonic source? What was the strength of shaking? And what was the magnitude? In discussing these issues we focus on the following aspects of level-ground liquefaction: (1) mechanisms that form seismic liquefaction features in the field; (2) field settings where liquefaction features should be present if strong seismic shaking has occurred; (3) field settings where an absence of liquefaction features indicates an absence of strong seismic shaking; (4) how liquefaction features should be used to interpret the tectonic source locale of a paleo-earthquake; and (5) how effects of liquefaction can be used to back-calculate the strength of shaking as well as earthquake magnitude. Several methods are available to back-calculate the strength of shaking and earthquake magnitude, and the most commonly used methods are presented and critiqued. Our critique of these methods points out the uncertainties attending each. Paleoliquefaction/paleoseismic case histories are presented to illustrate potential uncertainties in back-calculations and procedures to overcome these uncertainties. Reasonable confidence in paleoseismic interpretation generally requires using multiple methods of back-analysis, and achieving similar results from each method. An alternate approach can be used for paleo-earthquakes that were large enough to have caused liquefaction in a variety of geologic settings, in which there were differing factors affecting surface ground motions and liquefaction susceptibility. For this situation, a method such as the cyclic-stress method can be used to make back-calculations that can be cross-checked with results from other settings.