Despite achievements in dolphin conservation for the tuna purse-seine fishery of the eastern Pacific Ocean, debate continues about the magnitude and importance of dolphin mortality caused by small (unobserved) vessels. In-port sampling of tuna catch size composition is a potentially cost-effective means of identifying unobserved vessels that may be catching tunas associated with dolphins because yellowfin tuna caught in association with dolphins are larger, on average, than those caught in other types of purse-seine sets. A classification algorithm to predict purse-seine set type (“dolphin” vs. “nondolphin”) was built from port-sampling data on yellowfin tuna length-frequencies and the date and location of fishing of large (observed) vessels. This classification algorithm was used to screen the port-sampling data of small vessels collected during 2006-2009, assuming the fishing practices of the two groups resulted in similar catch characteristics. From these results, hypothetical time series of dolphin mortality for small vessels were constructed and incorporated into a population dynamics model, along with mortalities of large vessels. Results suggest that any dolphin mortality of small vessels is unlikely to be substantially affecting trends in dolphin abundance. These results underscore the importance of in-port sampling, in combination with at-sea observation and fishery-independent surveys, to effective management.