The exploratory research of the criminal investigation statement analysis

• Main Author: 李泱輯
• Other Authors: 廖有祿
• Format: Others
• Language: zh-TW
• Published: 1040
• Online Access: http://ndltd.ncl.edu.tw/handle/60905035051309260744

碩士 === 中央警察大學 === 刑事警察研究所 === 103 === In criminal investigations, besides instincts, the main tool for criminal investigators to detect a deception in a statement within a short period of time is biological lie detection technology. However, while biological lie detection is not efficient enough and instructs are not scientifically reliable, are there any scientific methods that can help investigators detect deceptions? The purpose of this research was to investigate whether Scientific Content Analysis can be applied to detect deceptions and offer timely feedback for investigators, and to build a Scientific Content Analysis model to facilitate front-line investigators. The instrument for the statement analysis in this study was Linguistic Inquiry and Word Count (LIWC) and one lab experiment and one field study were made. In the lab experiment, a total of 132 students from Central Police University were randomly assigned to 4 groups: (A) offenders with true statements, (B) offenders with false statements, (C) non-offenders with true statements and (D) non-offenders with false statements. In the experiments, each of the participants was required to act and offer a statement based on his/her corresponding group instructions and then the statements were analyzed and compared by C-LIWC criteria. In the filed study, a total of 120 statements of real victims, witnesses and offenders in 2014 offered by the Crime Investigation Bureau were collected and then they were analyzed by C-LIWC criteria and compared with the statistics obtained in the lab experiment. The result of this study concluded that the language forms can be potentially applied to detect deceptions in a statement. In the lab experiment, it can be observed different states of mind based on different variables (true statements vs. false statements) significantly affected the language forms. In the filed study, similar findings indicated that offenders who attempted to make a false statement tended to use exclusive and filler words. To observe the language transition by a particular participant in different timing can help investigate whether an offender is offering a statement honestly or deceptively. In other words, despite individual differences, Scientific Content Analysis can detect the language forms which reflect the state of mind of the speaker. If more valid samples can be collected and analyzed in future studies, this system will have higher liability and validity and can be applied by frontline criminal investigators to detect deceptions.