Prevenindo injustiças : intervenções baseadas em evidências para o reconhecimento de pessoas

Abstract

Eyewitness identification is an evidence that can result in unfair convictions due to false identifications (i.e., when an innocent person is identified as the perpetrator of the crime). Researchers have studied variables that may decrease the risk of false identification resulting reforms of identification procedures in countries such as the United States and England. In Brazil, identification procedures are still based on the penal procedure code that dates from 1941, and research on this subject in Brazilian-portuguese literature is still incipient. This dissertation approaches evidence-based procedures to prevent false identification through five chapters. The first chapter presents a review of the variables that reduce the risk of false identifications, such as the use of fair lineup (e.g, presenting the suspect among similar nonsuspects), and the use of appropriate instructions (e.g, mentioning that the perpetrator of the crime may not be among the faces presented to the witness). The second study experimentally compares the performance of witnesses in a lineup with moderately or highly similar lineup members, demonstrating that regardless of similarity, the lineup is effective in reducing the risk of false identification. Even in countries that have developed evidence-based procedures for eyewitness identification, there is a lack of research on effects of training police officers to perform such procedures, thus chapter three, four, and five aimed to develop a training program for police officers. Chapter three presents a review of the importance of training police officers in investigative interview techniques, whose training serve as model for the development of the eyewitness identification training program. Chapter four presents the steps taken for the development of the FAIR (Find a suspect, Avoiding bias, Instructing the witness, Recording the procedure) training program, and presents police perceptions prom a pilot training. Finally, chapter five tests the effectiveness of the FAIR) training program with Brazilian civil police officers by comparing their performance before and after training. In a mock identification task performed before training police officers generally built lineups with multiple suspects and did not provide inappropriate pre-lineup instructions or feedback. In the posttest results of the mock identification task police officers performed fair lineups, and used appropriate instructions and feedback more frequently. The FAIR training program presents a novel approach to be replicated in other states of Brazil and adapted to other countries. In short, this dissertation presents a literature review, the use of experiments, and the development of training to approaches fairer and more effective eyewitness identification procedures. The results of this dissertation point towards the need to reframe identification procedures in Brazil, and propose procedures to be followed and training program for police officers in order to preserve witness memories and the human rights of those involved in a criminal investigation.