Journal of Information Security Research Martin Lopez Nores 17 2 2026 https://doi.org/10.6025/jisr/2026/17/2/57-75 https://www.dline.info/jisr/fulltext/v17n2/jisrv17n2_1.pdf The rapid evolution of cyber threats has exposed the limitations of traditional, reactive cybersecurity frameworks, necessitating data driven and adaptive defense mechanisms. This study investigates the dynamics between threat characteristics and system response latency through a comprehensive statistical and predictive analysis of 1,000 cybersecurity incident records. Employing chi-square tests, correlation analyses, effect size measurements, and Ridge regression modeling, the research evaluates how variables such as threat severity, access level, and detection status influence breach occurrence and response time. Results reveal that threat severity and privilege levels exhibit negligible associations with breach likelihood and response latency, challenging conventional risk prioritization assumptions. Instead, threat detection status emerges as the dominant predictor of response time (n²  0.76), indicating that current systems operate primarily on reactive, event triggered paradigms rather than proactive, severity aware mechanisms. Furthermore, the analysis identifies critical data quality challenges, including zero inflated response time distributions and timestamp induced feature leakage, which artificially inflate predictive metrics and compromise model generalizability. These findings underscore a significant gap between theoretical cybersecurity frameworks and practical system behavior. The study concludes that enhancing threatdetection accuracy, implementing rigorous data-preprocessing protocols, and integrating meaningful temporal and contextual features are essential for developing robust, proactive cybersecurity architectures. Future research should prioritize real world dataset validation and advanced machine learning techniques to overcome current analytical constraints.