We examined the effects of research disturbance on the behavior and abundance of Steller sea lions (Eumetopias jubatus) at rookeries on Marmot and Ugamak Islands in Alaska. During 3 of 6 yr, researchers intentionally drove all adult and juvenile sea lions off at least part of the beach in order to permanently mark and measure sea lion pups. The research disturbance occurred after the majority of females had bred and when most pups were 1 mo old. We used generalized linear models to determine the relationship between research disturbance and sea lion behavior or abundance. Research disturbance was related to changes in the proportion of sea lions exhibiting two to three of nine behavior metrics: agonistic and resting females and active males at Marmot, and active and resting males and females at Ugamak. Model results indicated that changes lasted between 3 and 20 d depending on the sex, behavior, and rookery. Inclusion of research disturbance into Marmot abundance models did not improve the fit to the data, if variability between years was permitted. Optimally timed, low-frequency research disturbance did not appear to have long-term effects on sea lion behavior or abundance and was largely associated with changes that were similar to natural variation.