Typically the analysis is done by digging trenches across the fault and dating pieces of charcoal found in the sediment with radiocarbon. This study is rare but not unique in using cored trees, or slices of stumps if possible, and dating the event with tree ring dating (dendrochronology), which is vastly superior when it's possible (i.e. when it is demonstrable that an earthquake kills trees). In both cases you want as much as you can get--ideally 5-10 ages at a minimum I think?