When investigating crimes police frequently use eyewitness identification tests such as lineups in an attempt to establish the culprit's identity. It is now well documented that witnesses commonly make mistakes, sometimes identifying innocent suspects or failing to identify the culprit. Lineup administrators also make mistakes in the sense that they use procedures that increase the likelihood of eyewitness error. Eyewitness memory research has provided a basis for many useful guidelines regarding the conduct of identification tests. Here, we review the relevant research underpinning effective procedures for lineup construction, lineup presentation, recording the witness's decision, and providing feedback to the witness. The systematic implementation of these guidelines is likely to reduce significantly the extent of eyewitness errors at the identification test and improve the probative value of courtroom testimony about identity. However, further significant advances may well require the development of novel procedures for accessing witnesses' memories.