Using a literature review, the study found that predictive methods are generally in four categories: crime predictors that identify locations of crime concentrations ("hotspots"); offender predictors that identify individuals at high risk of committing crime; perpetrator identity predictors based on crime-scene information; and crime victim predictors, which blend predictions of crime location and offender risk to predict risk of victimization. The research then identified a four-step process in which predictive policing leads to preventing and countering crime. The four steps are data collection, data analysis, the design of police intervention, and the resulting impact on crime. The following predictive-policing myths are discussed: "the computer knows the future;" "the computer will do everything for you;" "each department needs a high-powered, costly model;" and "accurate predictions equal major decreases in crime." Based on this research, three conclusions are offered. First, predictive policing methods and tools can benefit agencies of all sizes in all regions, regardless of proximity to urban areas. Second, small agencies with low crime rates and routine crime distribution will likely need only basic statistical and display capabilities that are often free and interface with existing software. Third, large agencies that have considerable incident and intelligence data for analysis would benefit from more complex and sophisticated predictive-policing tools. Recommendations pertain to having accurate and complete data, choosing the best predictive model and intervention techniques, and referencing the full report for a complete list of considerations at http://www.rand.org/pubs/research_reports/RR233.html.
- A Novel Two-Step Method for the Detection of Organic Gunshot Residue for Forensic Purposes: Fast Fluorescence Imaging Followed by Raman Microspectroscopic Identification
- Simultaneous Imaging of Latent Fingermarks and Detection of Analytes of Forensic Relevance by Laser Ablation Direct Analysis in Real Time Imaging-Mass Spectrometry (LADI-MS)
- Random Forest Processing of Direct Analysis in Real-Time Mass Spectrometric Data Enables Species Identification of Psychoactive Plants From Their Headspace Chemical Signatures