Abstract: Although the term ``pebble count'' is in widespread use, there is no standardized methodology used for the field application of this procedure. Each pebble count analysis is the product of several methodological choices, any of which are capable of influencing the final result. Because there are virtually countless variations on pebble count protocols, the question of how their results differ when applied to the same study reach is becoming increasingly important. This study compared three pebble count protocols: the reach-averaged Environmental Monitoring and Assessment Program (EMAP) protocol named after the EMAP developed by the Environmental Protection Agency, the habitat-unit specific U.S. Forest Service’s PACFISH/INFISH Biological Opinion (PIBO) Effectiveness Monitoring Program protocol, and a data-intensive method developed by the authors named Sampling Frame and Template (SFT). When applied to the same study reaches, particle-size distributions varied among the three pebble count protocols because of differences in sample locations within a stream reach and along a transect, in particle selection, and particle-size determination. The EMAP protocol yielded considerably finer, and the PIBO protocol considerably coarser distributions than the SFT protocol in the pool-riffle study streams, suggesting that the data cannot be used interchangeably. Approximately half of the difference was due to sampling at different areas within the study reach (i.e., wetted width, riffles, and bankfull width) and at different locations within a transect. The other half was attributed to using different methods for particle selection from the bed, particle-size determination, and the use of wide, nonstandard size classes. Most of the differences in sampling outcomes could be eliminated by using simple field tools, by collecting a larger sample size, and by systematically sampling the entire bankfull channel and all geomorphic units within the reach.