scieee AI-readable full text Open interactive document viewer

Police View vs. Forensic View: The network you think exists vs. The network that actually exists

Harvey, Dayle

Abstract

Police intelligence links cases only when people or addresses overlap, producing fragmented networks that collapse when a “key actor” is removed. Forensic intelligence (fingerprints/DNA) adds missing offenders and connections, transforming isolated cases into a highly connected structure. Police collapse the network they think exists. Forensics reveal the network that actually does.

Full text

VS Police intelligence links cases if suspects or addresses are shared. This produces a partial network, missing many relationships. Forensic intelligence (fingerprints, DNA) adds: new nodes (offenders previously unknown), new edges (connections between cases). Forensic networks are structurally stronger Forensics converts chaos into structure, increasing connectedness LCC 90%+ → Cases become one network, not fragments. Density 3–5× → More relationships, fewer blind spots. Clustering 4–6× → Reveals real criminal cells within networks. Police disrupt the wrong nodes The network you think exists When police remove the highest-degree actor, their network collapses the Average Shortest Path Length doubles (+100%). But in the fingerprint-enhanced network, removing the same actor barely matters (+10%). Police collapse the network they think exists. Forensics reveals the one that actually does. How we modeled disruption 1.We built 1000 iterations of simulated police only and forensic enhanced networks. 2.We measured the baseline differences in all social network analysis metrics and then applied a variety of node based and edge based disruptions. 3.Finally, we measured the after effects of the disruptinos including sequential disruption. Forensic View The network that actually exists What police think the network looks like “Criminal networks aren’t small. Our visibility is.” Police View By: Dayle Harvey and Christophe Vandeviver