Comment by dadrian

As a reminder, peer review is not designed adversarially. It is not supposed to catch people who are fraudulent before they publish---that's the point of replication (which happens rarily). Peer review is designed to ensure a the publication as a packaged work of science describes a valid experiment. But you're necessarily assuming that the authors did what they said they did. As a peer reviewer, you're ensuring that what the authors said they did constitutes valid science.