A synthesis is presented of published statistics on the lengths, scarp heights, and spacing of normal faults on the flanks of the East Pacific Rise (EPR) between 19¿S and 21¿N. Nine data sets are included which were collected at different times employing one of four different sonar mapping systems (the Scripps deep-tow instrument, SeaMARC I, SeaMARC II, and GLORIA) with widely ranging data collection capabilities. The EPR in this region is morphologically quite similar, and in some cases the data sets were obtained from the same area of the seafloor. It is assumed, therefore, that data collection procedure and system resolution have had a greater influence on the fault statistics observed than varying tectonic processes. We try to account for these sampling effects such that quantitative comparisons of fault development at midocean ridges may then be made. We discuss the problem of identifying faults in sonar images and the errors associated with measurements of the relevant parameters. We also investigate the statistical properties of the fault populations and derive expressions of correcting observed values of the population means at finite resolution to obtain the corresponding values at infinite resolution. The theoretical predictions compare well with the results of a decimation procedure applied to real data sets to simulate diminishing resolution. We also discuss the effects that surveying geometry can have on the interpretation of fault lengths. The data synthesis is augmented by an analysis of the correlation between scarp height and fault length for areas near 3.5¿S and 12¿N on the EPR. The correlation, which we find to be linear is used to estimate (1) the minimum operational fault target size for each of the four mapping systems, and (2) the topography due to faulting in an area where only fault lengths are known. ¿ American Geophysical Union 1994