Eyewitness Memory

Abstract Eyewitness testimony during a criminal trial, even when made in good faith, is widely considered to be unreliable because (a) basic-science research has shown how malleable eyewitness memory can be and (b) many real-world wrongful convictions involve eyewitness misidentifications of innocent defendants. However, like other forms of forensic evidence, there are conditions under which declarations based on eyewitness memory are reliable and conditions under which they are unreliable. Precisely because memory is so malleable, declarations based on eyewitness memory are the most reliable on the first test conducted early in a police investigation. Indeed, the very act of testing memory changes (i.e., contaminates) memory, and this is true whether memory is tested by a police interview (recall of details) or by a police photo lineup (face recognition). For example, because of the contaminating effect of the first test, a witness who initially recalls a detail with low confidence is at risk of later recalling that same detail with higher confidence. Similarly, a witness who initially identifies a suspect from a photo lineup with low confidence is at risk of later identifying that same suspect with higher confidence. In addition, memory can be contaminated by extraneous factors that occur after the first test (e.g., conversations with other people), leading to higher confidence or even to an altogether different memory decision (e.g., initially claiming that a person well known to the witness was not the perpetrator but later remembering that he was). Unbeknownst to police investigators, any change in a witness’s good-faith memory of events after the first test is far more likely to reflect memory contamination than it is to reflect a more successful search of memory. Moreover, and critically, this is true no matter how the witness rationalizes the inconsistency (e.g., “I was nervous on the first test but then I calmed down and searched my memory more carefully, and now I am positive he is the perpetrator”). A witness has no awareness of the insidious effects of memory contamination and certainly has no scientific expertise in the underlying memory mechanisms involved. Therefore, asking why a witness’s memory-based declaration changed from one test to the next is a question for a memory expert, not an eyewitness. A memory expert can explain that the main problem with relying on eyewitness memory is not that it is unreliable. Instead, the main problem is that the criminal justice system ignores the results of the reliable first test of uncontaminated memory and instead relies on the results of subsequent unreliable tests of contaminated memory to win a conviction. Assuming good faith on the part of the eyewitness, this practice should be reversed by relying on the results of the first test of uncontaminated memory and ignoring all later tests, especially the maximally contaminated test of memory that occurs under oath at trial.