Meta-Analysis of Facial Identification Research: A Reappraisal

面部识别研究的荟萃分析：重新评估

• 批准号: 0010140
• 负责人: Brian Bornstein
• 金额: $ 14.06万
• 依托单位: University of Nebraska-Lincoln
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2001
• 资助国家: 美国
• 起止时间: 2001-05-15 至 2004-04-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0010140&HistoricalAwards=false
• 关键词: Meta; Analysis; Facial; Identification; Research

This project will examine mistaken eyewitness identifications as a principal source of erroneous convictions. Although a growing body of empirical research by psychologists provides insights into the factors contributing to mistaken identifications, and with increasing frequency, psychologists are called upon to provide courtroom testimony about those factors. Such expert testimony, however, is often criticized as unreliable. This project will examine the empirical questions of reliability and generalization of findings using a meta-analysis of existing studies of face recognition, in both the eyewitness and laboratory research paradigms. In particular, the project will examine the strength and reliability of research findings concerning a large number of factors thought to influence identification accuracy and the relationships among research methods and the magnitude of effects produced by manipulations of substantive variables. Factors to be addressed include stable (e.g., sex, race) and malleable characteristics (e.g., stress, disguise) of both participants and targets, as well as situational (e.g., exposure duration) and procedural (e.g., lineup presentation) variables. Careful attention will be given in the meta-analysis to the different types of study characteristics, independent variables, and their operationalization employed in the research. The project, as a result, will identify research and theoretical domains where findings are on the strongest and weakest footings and, most importantly, will test a series of hypotheses about methodological and substantive factors that enhance or limit the generalizability and theoretical implications of research results.

本项目将审查错误目击者身份作为误判的主要来源。尽管越来越多的心理学家的实证研究提供了对导致错误识别的因素的洞察，而且频率越来越高，但心理学家被要求就这些因素提供法庭证词。然而，这种专家证词经常被批评为不可靠。本项目将通过对现有人脸识别研究的荟萃分析，在目击者和实验室两种研究范式下，审查调查结果的可靠性和概括性的实证问题。特别是，该项目将审查关于被认为影响识别准确性的大量因素的研究结果的强度和可靠性，以及研究方法和操纵实质性变量产生的影响的大小之间的关系。需要解决的因素包括参与者和目标的稳定性(例如，性别、种族)和可塑性特征(例如，压力、伪装)，以及情景(例如，暴露时间)和程序性(例如，阵容呈现)变量。不同类型的研究特点、自变量及其在研究中的可操作性将在元分析中得到认真的关注。因此，该项目将确定研究和理论领域的研究结果是最坚实和最薄弱的基础，最重要的是，将检验关于加强或限制研究成果的概括性和理论影响的方法学和实质性因素的一系列假设。