The article first reviews the history of how the science of psychology has sought to assess and characterize the reliability and accuracy of eyewitness evidence under various circumstances. This is followed by a detailed discussion of how assessments of and procedures for obtaining eyewitness evidence are based on pseudo-science, quasi-science, and intuition rather than on scientific research regarding the variable accuracy of eyewitness evidence, depending on a variety of factors, including the methods investigators use to obtain eyewitness evidence. The discussion addresses some basic points about how memory works as a reconstructive process, inherently prone to errors of commission and omission. This contrasts with competing perspectives that claim some memories in some people’s brains are especially stable and resistant to interference, distortion, forgetting, or other limitations. The discussion then turns to eyewitness evidence procedures, with attention to lineup construction and lineup administration. The article advises that best practices for eyewitness evidence procedures should be based on logic and research, not intuition and past practice, and they should be consistent with the legal, moral, and ethical values expected of the legal system. The best way to reduce unreliable eyewitness evidence is to take the frailties of memory into account. 3 notes and 40 references