Regulation of Terry stops of pedestrians by police requires articulation of the reasonable and individualized bases of suspicion that motivate their actions. Nearly five decades after Terry, courts have found it difficult to articulate the boundaries or parameters of reasonable suspicion. The behavior and appearances of individuals combine with the social and spatial contexts in which police observe them to create an algebra of suspicion. Police can proceed to approach and temporarily detain a person at a threshold of suspicion that courts have been unable and perhaps unwilling to articulate. The result has been sharp tensions within Fourth Amendment doctrine as to what is reasonable, why, and in what circumstances. The jurisprudence of suspicion is no clearer today than it was in the aftermath of Terry. This issue has taken center stage in both litigation and policy debates on the constitutionality of the stop-and-frisk policing regime in New York City. Under this regime, police state the bases of suspicion using a menu of codified stop rationales with supplemental text narratives to record their descriptions of suspicious behaviors or circumstances that produced actionable suspicion.
Evidence from 4.4 million stops provides an empirical basis to assess the revealed preferences of police officers as to the bases for these Terry stops. Analyses of this evidence reveal narratives of suspicion beyond the idiosyncrasies of the individual case that police use to justify their actions. First, we identify patterns of articulated suspicion. Next, we show the individual factors and social conditions that shape how those patterns are applied. We also show how patterns evolve over time and become clearer and more refined across a wide range of police stops. That refinement seems to follow the capacious interpretative room created by four decades of post-Terry Fourth Amendment jurisprudence. Next, we assess the extent of constitutional compliance and examine the neighborhood and individual factors that predict noncompliance. The results suggest that the observed patterns of narratives have evolved into shared narratives or scripts of suspicion, and that these patterns are specific to suspect race and neighborhood factors. We conclude that scripts are expressions of the norms within the everyday organizational exercise of police discretion and that these scripts defeat the requirement of individualization inherent in case law governing Fourth Amendment stops.