The generalizability crisis so ably articulated by Yarkoni can lead to particularly unfortunate consequences in applied fields such as the law. Some of the most egregious examples of this occur in the fields of memory and eyewitness testimony. General statements are made asserting that lay opinions about memory, including those of jurors and lawyers, are frequently in error (Berkowitz & Frenda, Reference Berkowitz and Frenda2018; Clifasefi, Garry, & Loftus, Reference Clifasefi, Garry, Loftus and Della Sala2007; Lynn, Evans, Laurence, & Lilienfeld, Reference Lynn, Evans, Laurence and Lilienfeld2015). Memory is claimed to be error-prone or unreliable without the qualification that it may be accurate under other conditions (Lynn & Payne, Reference Lynn and Payne1997; Zajac, Garry, London, Goodyear-Smith, & Hayne, Reference Zajac, Garry, London, Goodyear-Smith and Hayne2013). Some psychologists refer to “laws” of memory (Howe, Reference Howe2013), even though memory phenomena are known to be highly dependent on such factors as who the participants are, the conditions present at encoding, what is being recalled, the encoding conditions, and how memory is assessed (Roediger, Reference Roediger2008). Such overgeneralizations can then become incorporated in legal and judicial documents.

The general public is supposed to think that memory involves playing back events exactly as they happened, literally “like a video camera” (Lacy & Stark, Reference Lacy and Stark2013; Lilienfeld, Lynn, Ruscio, & Beyerstein, Reference Lilienfeld, Lynn, Ruscio and Beyerstein2010). This often repeated, but false, assertion involves generalization from responses to single survey questions such as “Human memory works like a video camera, accurately recording the events we see and hear so that we can review and inspect them later” (Simons & Chabris, Reference Simons and Chabris2011). In fact most people do experience some memories, particularly those involving important events, as a connected series of scenes rather like a videotape. Simply asking additional questions reveals that memory beliefs are much more nuanced than this and people are well aware that their recollection is not always reliable (Brewin, Li, Ntarantana, Unsworth, & McNeilis, Reference Brewin, Li, Ntarantana, Unsworth and McNeilis2019). However, the idea that the public had mistaken ideas about memory led to the state of New Jersey instructing jurors that “Research has revealed that human memory is not like a video recording that a witness need only replay to remember what happened. Human memory is far more complex” (New Jersey Courts, 2020). The danger here is that testimony could be discounted purely on the basis that the witness described their memory as “like a video recording,” when this statement did not at all imply a naïve or mistaken view of memory.

For many years eyewitness confidence was thought to be only weakly related to accuracy, but this conclusion was overturned when techniques for analyzing the question were improved (Wixted & Wells, Reference Wixted and Wells2017). Eyewitness confidence is not a guarantee of accuracy, but is very highly related to accuracy when memory is uncontaminated and suitable interviewing procedures are used (Wixted, Mickes, & Fisher, Reference Wixted, Mickes and Fisher2018). However, the inappropriate generalization had already been incorporated into compulsory jury guidance issued by the Supreme Court of New Jersey (http://www.judiciary.state.nj.us/pressrel/2012/pr120719a.htm) for cases involving eyewitness identification, which stated “Although some research has found that highly confident witnesses are more likely to make accurate identifications, eyewitness confidence is generally an unreliable indicator of accuracy” (p. 4).

Witnesses have often been impugned by claims that memory is unreliable, for example, invoking research on “false memory” (Brewin, Andrews, & Mickes, Reference Brewin, Andrews and Mickes2020; Wade, Nash, & Lindsay, Reference Wade, Nash and Lindsay2018; Wixted et al., Reference Wixted, Mickes and Fisher2018). This term is applied to several different experimental paradigms such as associative illusions (the Deese-Roediger-McDermott paradigm), manipulating memory for detail by providing misleading post-event information, artificially inflating the perceived likelihood that a non-remembered event occurred, and implanting memories for childhood events that never happened. Implications for the legal system are then discussed without distinguishing between these very different paradigms, and without recognition that there is no overall proclivity to experience all types of false memory (Lacy & Stark, Reference Lacy and Stark2013). The fact that memory can be manipulated in the laboratory is important but does not by itself allow of any conclusion about how reliable memory is under normal, real life circumstances. Despite this, some organizations have endorsed statements such as “Science shows that the memory of an honest witness who gives evidence in international arbitration proceedings can easily become distorted” (International Chamber of Commerce, 2020).

Similarly, it is sometimes suggested on the basis of memory implantation research that it is relatively simple to create false memories (Conway, Reference Conway, Nadel and Sinnott-Armstrong2012). This overgeneralization overlooks the special procedures that are used, including at times a high degree of deception, and the fact that only a small minority of participants may succumb to them. Little is known about the durability of the effects or whether they are associated with the degree of conviction necessary to sustain legal procedures such as cross-examination (Brewin & Andrews, Reference Brewin and Andrews2017).

Overgeneralizations also occur when defence experts comment on contemporary witnesses reporting historic crimes. The simple idea that “memory decays with time” is sometimes put forward, without acknowledgment that most studies involved meaningless experimental materials and that there are numerous counter-examples (Roediger, Reference Roediger2008). Moreover, the claim does not make clear that with significant personal experiences initial decay in the total amount recalled typically plateaus, resulting in stable and largely accurate long-term recall of the remainder (Diamond, Armson, & Levine, Reference Diamond, Armson and Levine2020; Hirst et al., Reference Hirst, Phelps, Meksin, Vaidya, Johnson, Mitchell and Olsson2015). Despite this, the compulsory jury guidance referenced above contained the generalization: “Memories fade with time…. In other words, the more time that passes, the greater the possibility that a witness's memory of a perpetrator will weaken” (p. 5).

In 2013, a UK judge, Justice Leggatt, commented “Psychological research has demonstrated that memories are fluid and malleable, being constantly rewritten whenever they are retrieved (EWHC 3560 (Comm); Case No. 2011 Folio 1267). A person would be in error to suppose: that the stronger and more vivid is our feeling or experience of recollection, the more likely the recollection is to be accurate.” These statements do not accurately reflect a complex literature, and fail to take into account the number of times an event happens, its importance, and memory rehearsal, whether deliberate or spontaneous. Legal professionals are not in a position to know that these are overgeneralizations. Psychologists, as Yarkoni has demonstrated, are the ones who are responsible and this can have serious consequences when their statements are repeated in the real world.