Witnesses who discuss an event with others often incorporate misinformation encountered during the discussion into their memory of the event. Two experiments were conducted to establish whether this memory conformity also occurs in the context of an interview and whether it is possible to reduce the effect. Participants viewed a crime-video which they then discussed with a co-witness. Some participants were warned they may have been exposed to misinformation during the discussion before all were interviewed individually. In Experiment 1, participants made remember/know judgments about each component of their free recall, and in Experiment 2 they were asked to indicate the source of their memories. Co-witness information was incorporated into participants' testimony, and this effect could not be significantly reduced using warnings and source-monitoring instructions. Remember/know judgments may be useful in distinguishing ‘real’ memories from false memories. We make some recommendations regarding the interviewing of witnesses. Copyright © 2009 John Wiley & Sons, Ltd.