A Study on the Integrated Statement Comparative Analysis Model
Korea Safety Culture Society, Jung Sik Gong · Forum of Public Safety and Culture · 2025
This study proposes the Integrated Comparative Statement Analysis Model (I-CSAM), a novel framework designed to address the limitations of traditional statement analysis techniques that primarily rely on single-statement evaluations, often focusing solely on the victim's account. I-CSAM enables a structured, segment-by-segment comparison of both the victim’s and suspect’s statements within the temporal and contextual flow of the event. To empirically evaluate its validity and practical applicability, the model was applied to 43 real-world criminal cases in South Korea involving sexual assault, child abuse, dating violence, and other interpersonal offenses. I-CSAM was independently applied alongside three established methods—Criteria-Based Content Analysis (CBCA), Reality Monitoring (RM), and Scientific Content Analysis (SCAN)—with the results compared against the final court verdicts. I-CSAM achieved an accuracy rate of 86.0%, with a sensitivity of 0.87 and specificity of 0.84, outperforming the other methods across all metrics. It also demonstrated high inter-rater reliability (Cohen’s Kappa = 0.78), indicating strong consistency among expert analysts. Notably, I-CSAM effectively identified structural inconsistencies, shifts in emotional and sensory information across repeated statements, and gaps in narrative coherence—elements often overlooked by conventional tools. By integrating empirically validated criteria from existing models with a structured coding system that allows for both quantitative scoring and qualitative interpretation, I-CSAM enhances analytical objectivity, replicability, and explanatory clarity. The findings support the potential of I-CSAM as a highly reliable tool for assessing the credibility of conflicting statements in investigative and judicial settings. Furthermore, its structured design offers a scalable foundation for future integration with artificial intelligence and natural language processing technologies, paving the way for more efficient and standardized automated statement analysis systems.