The risk–disturbance hypothesis states that animals react to human stressors in the same way as they do to natural predators. Given increasing human–wildlife contact, understanding whether animals perceive anthropogenic sounds as a threat is important for assessing the long-term sustainability of wildlife tourism and proposing appropriate mitigation strategies. A study of pygmy marmoset (Cebuella niveiventris) responses to human speech found marmosets fled, decreased feeding and resting, and increased alert behaviors in response to human speech. Following this study, we investigated pygmy marmoset reactions to playbacks of different acoustic stimuli: controls (no playback, white noise and cicadas), anthropogenic noise (human speech and motorboats), and avian predators. For each playback condition, we recorded the behavior of a marmoset and looked at how the behaviors changed during and after the playback relative to behaviors before. We repeated this on ten different marmoset groups, playing each condition once to each group. The results did not replicate a previous study on the same species, at the same site, demonstrating the importance of replication in primate research, particularly when results are used to inform conservation policy. The results showed increased scanning during playbacks of the cicadas and predators compared with before the playback, and an increase in resting after playbacks of avian predators, but no evidence of behavior change in response to playbacks of human speech. There was no effect of ambient sound levels or distance between the playback source and focal animals on their behavior for all playback conditions. Although we find that noise can change the behavior of pygmy marmosets, we did not find evidence to support the risk–disturbance hypothesis.