Work

An Analysis On Stop-And-Frisk In New York City In 2014

Year: 2019
Type: dissertation
Abstract: Stop-and-frisk is a policing tactic in which police officers may stop an individual whom they suspect of being on the verge of committing a crime. Its goal is to prevent a crime before it happens by stopping and sometimes searching an individual, which might lead to arrest or summons issued. The top priority is to seize illegal guns. But how do we differentiate such individuals from others in order to make it a successful stop? The program has been plagued by controversies that police abuse their power and stop innocent people for no reason other than their skin color. In this thesis, we analyse stops in 2014 by New York Police Department. We use Exploratory Data and Exploratory Spatial Data Analysis to assess stops as a whole based on race, stop outcomes, and across New York City geographically. We fit logistic models to predict the probability that a person might get arrested to address the difference in stopping different ethnic groups. We also use Naive Bayes classification to classify arrest given the conditional probability of demographics, weapons found, time observed, and time spent with the suspects. The results show that although blacks only constitute about 24.5% of NYC population, they were stopped as much as the other three ethnic groups combine. In terms of getting arrested, only 14.7% of total stops lead to arrests in which black were arrested the most at 7.6% followed closely by Hispanics at 5.2% and at the distance 1.4% for whites. Compared to blacks, Hispanics are more likely to get arrested and being white decreases the odds. In the context of seized contraband, white stops are more efficient as whites are more likely to get arrested compared to blacks while black and Hispanic stops are disproportionately higher, which implies lesser standards were applied to stopping minority groups.
Author: Nguyen, Tam
Language: English
Open Access status: green
OpenAlex ID: W6979922092
Cites: 1
Cited by:
Related to: 10