Testing the Pristine Conditions Hypothesis of the Confidence-Accuracy Relationship in Eyewitness Identification

据估计，目击者身份识别每年在大约80 000起刑事案件中发挥作用。 目击者可以在刑事案件中提供关键证据，但目击者的指认并不总是准确的。 目击者可能会犯两种错误。 他们可能无法从列队中识别出有罪的人，也可能从列队中错误地识别出无辜的人。 这两类错误都对公共安全构成威胁，因为它们有助于犯罪者自由地犯下更多的罪行。 此外，错误的指认会危及无辜嫌疑人的自由。由于这些原因，更好地了解与目击者准确指认相关的情况和程序是很重要的。可以用来预测目击者准确性的一个重要因素是目击者在指认时的信心。 许多研究发现，当证人以高度的信心指认嫌疑人时，证人很可能是正确的。当警察使用理想程序来获得身份时，情况尤其如此。 然而，最近的一些研究发现，即使使用了理想的程序，高度自信的证人也不总是高度准确的。研究人员提出了一系列使用虚拟现实再现犯罪的研究，以帮助了解高度自信的证人何时可能高度准确，何时可能不太准确。 使用虚拟现实可以创造出比传统的目击者研究更逼真、更吸引人的事件。 这项研究的重点是测试两个想法。 第一个想法是，一些证人可能比其他证人更好地理解他们自己的记忆和感知是如何工作的。 那些对自己的记忆和感知更熟悉的目击者可能只会在正确的概率很高的时候才说他们有信心。 如果目击者对自己的记忆和感知如何运作的理解不好，他们可能会错误地说自己很自信，即使正确的几率很低。 第二个想法是，可能有见证条件是如此之差（例如，即使是最好的证人也不可能非常准确，证人在作出信心判断时可能没有充分考虑到作证条件有多差。这些假设将通过一系列测试儿童和成人证人的实验进行测试，并将提供一个更好的理解目击者的信心和准确性是如何相关的。这个奖项反映了NSF的法定使命，并已被认为是值得通过使用基金会的智力价值和更广泛的影响审查标准进行评估的支持。

It has been estimated that eyewitness identification plays a role in approximately 80,000 criminal cases per year. Eyewitnesses can provide crucial evidence in criminal cases, but eyewitness identifications are not always accurate. Eyewitnesses can make two kinds of mistakes. They can fail to identify guilty individuals from lineups and they can falsely identify innocent people from lineups. Both types of errors present a threat to public safety because they contribute to perpetrators being free to commit additional crimes. In addition, mistaken identifications risk the liberty of innocent suspects. For these reasons, it is important to better understand the situations and procedures that are associated with accurate eyewitness identifications.One important factor that may be used to predict eyewitness accuracy is the witness’s confidence at the time of the identification. A number of studies have found that when a witness identifies a suspect with high confidence that the witness is quite likely to be correct. This is particularly true when ideal procedures are used by police to obtain the identification. However, more recently, some studies have found that there are situations where, even when ideal procedures are used, highly confident witnesses are not always highly accurate. The researchers have proposed a series of studies using virtual reality recreations of crimes to help understand when highly confident witnesses are likely to be highly accurate and when they are likely to be less accurate. Use of virtual reality can create more lifelike and emotionally engaging events than have traditionally been used by eyewitness researchers. The research focuses on testing two ideas. The first idea is that some witnesses might have a better understanding of how their own memory and perception works than other witnesses. Witnesses who are more in tune with how their memory and perception works are likely to only say they are confident when the probability of being correct is high. Witnesses who have a poor understanding of how their memory and perception works, may mistakenly say they are confident, even when the odds of being correct are lower. The second idea is that there may be witnessing conditions that are so poor (e.g., poor lighting, very long distance) that even the best witnesses are unlikely to be very accurate and witnesses may not fully take into account how poor the witnessing conditions are when making their confidence judgments. These hypotheses will be tested through a series of experiments testing both child and adult witnesses and will provide a better understanding of how eyewitness confidence and accuracy are related.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.