Harnessing metacognition to assess the accuracy of memory reports from children in the Criminal Justice System

每年，数以百万计的儿童证人，最小的只有两岁，根据他们对世界各地刑事司法系统中事件的记忆提供证据。包括受害者在内的儿童证人越来越多地被警察要求回忆他们在调查期间的经历。这些回忆可以在法庭上作为关键证据重述。儿童记忆证据的准确性需要在法律程序的多个阶段得到正确评估，以确保其可靠性：警察根据证词决定他们应该优先考虑哪些调查线索；检察官决定证据是否足以指控嫌疑人犯罪；法官决定证据是否足够可靠，是否允许在法庭上使用；陪审团将决定是否有证据证明嫌疑人有罪。目前，尚无政策或实践指导来评估儿童记忆证据的准确性。PI的工作清楚地表明，法律从业者认识到缺乏经验指导。有一些指导可以用来判断孩子是否在撒谎。但很少有孩子故意编造证词，大多数诚实的证词既包含准确的信息，也包含不准确的信息。在没有指导的情况下，法律决策者主要依靠证人的年龄来决定是否相信记忆证据。这是有问题的。年龄通常不能预测准确性，也不能区分个人证词中准确和不准确的信息。对年龄的依赖导致了司法不公：儿童在他们的证词被错误地认为不如成年人准确后受到严重伤害或被谋杀，无辜者被错误地定罪并在监狱中度过数年。我们的项目将确定法律决策者（如警察、检察官、法官、陪审员）如何估计儿童诚实记忆报告中信息的准确性。它将通过调查和利用记忆的一个以前被忽视的、但至关重要的信息方面：元认知来做到这一点。元认知是一种监控我们自己记忆准确与否的能力。儿童的元认知线索可以用来确定他们证词中信息的准确性，但现有的研究使用有限的措施，并没有为告知实践提供强有力的基础。新的进展和PI的试点数据表明，这种方法的潜力现在可以实现。元认知线索可以是信心等级（高信心表示高准确性），也可以是非语言的，如耸耸肩（表示低准确性）。我们将测试4-5岁、7-8岁和10-11岁的儿童，以促进对元认知发展的理解，并确定哪些线索可以用来评估信息的可能准确性。我们将首次研究可能对幼儿准确性提供信息的新颖元认知线索，并采用统计技术，使我们能够分别测量表现的不同元素：他们的记忆准确性和元认知能力。此外，元认知研究与其在现实世界中的应用之间存在差距。我们将通过创建和测试警官培训来解决这个问题，通过使用儿童的元认知线索来提高他们评估儿童信息准确性的能力。我们还将与法律从业人员和其他专家合作，制定建议（针对实践、政策和未来研究），以评估全球儿童记忆证据。我们的目标是影响一系列需要评估儿童记忆准确性的法律决策者，以及负责儿童保护和受害者辩护的政府机构、政策制定者和组织。为了确保结果对这些社区具有实际用途，它们将贯穿整个项目。

Each year, millions of child witnesses, from as young as 2 years old, provide evidence based on their memory of events in Criminal Justice Systems around the world. Child witnesses, including victims, are increasingly being asked to remember what they experienced by police officers during an investigation. These recollections can go on to be recounted to a courtroom, as key evidence during a trial.The accuracy of child memory evidence needs to be correctly assessed at multiple stages of the legal process to ensure its reliability: police officers determine which investigatory leads they should prioritise based on testimony; prosecutors decide if the evidence is strong enough to charge the suspect with an offence; judges decide if the evidence is reliable enough to be allowed in the courtroom; and jurors decide whether the evidence warrants the suspect receiving a guilty verdict.Currently, no policy or practice guidance exists to assess the accuracy of children's memory evidence. The PI's work has made it clear that legal practitioners recognise the lack of empirically-informed guidance. Some guidance exists to try to determine if a child is lying. But few children intentionally fabricate testimony, and most honest accounts contain both accurate and inaccurate information. With no guidance, legal decision-makers largely rely on witness age to decide whether to trust memory evidence. This is problematic. Age often does not predict accuracy, and cannot discriminate between accurate and inaccurate information within an individual's testimony. Relying on age has resulted in miscarriages of justice: children have been seriously injured or murdered after their testimony was erroneously deemed to be less accurate than an adult's, and innocent people have been wrongfully convicted and spent years in prison.Our project will determine how legal decision-makers (e.g. police officers, prosecutors, judges, jurors) could estimate the accuracy of information in honest memory reports from children. It will do that by investigating and harnessing a previously overlooked, but crucially informative, aspect of memory: metacognition. Metacognition is the ability to monitor when our own memories are accurate or inaccurate.Children's metacognitive cues can be used to determine the accuracy of the information in their testimony, but existing research using limited measures has not provided a strong basis for informing practice. New advances and the PI's pilot data suggest the potential of this approach can now be realised. A metacognitive cue could be a confidence rating (with high confidence indicating high accuracy), or non-verbal, such as shrugs (indicating low accuracy). We will test children ages 4-5, 7-8, and 10-11, to advance understanding about metacognitive development and determine which cues could be used to assess the likely accuracy of information. For the first time, we will investigate novel metacognitive cues that are likely to be informative of accuracy in younger children, and employ statistical techniques that allow us separately to measure different elements of performance: their memory accuracy and metacognitive ability.Moreover, there is a gap between metacognitive research and its use in the real-world. We will address this by creating and testing police officer training to improve their ability to assess the accuracy of information from children, by using children's metacognitive cues. We will also work with legal practitioners and other experts to develop recommendations (for practice, policy, and future research), for assessing child memory evidence worldwide.Our goal is to impact a range of legal decision-makers who are required to assess child memory accuracy, as well as government agencies, policy-makers, and organisations responsible for child protection and victim advocacy. To ensure outcomes are of practical use to these communities, they will guide the project throughout.