Justice at a Crossroads in New York City. Studying Crimes in New York City Using the National Crime Victimization Survey (NCVS)

By Min Xie, Preeti Chauhan, Michael Rempel, and Jeremy Travis

This study relies on data from the National Crime Victimization Survey (NCVS) and presents trends from 1996 to 2022 in crime victimization, rates at which victims report crimes to the police, confidence in the police, and victims’ use of services in New York City. This is one of two studies falling under the umbrella of the Crossroads Project. Its goal is to trace New York City’s trends in crime victimization, enforcement, incarceration, and racial disparities from the 1990s to the early 2020s in the hopes that empirical data over a long timeframe might provide a much-needed perspective capable of informing future policy. Both of the two resulting reports and a joint executive summary may be found at the project landing page. How was the NCVS used in the present study? Although the NCVS is typically used to assess victimization patterns for the nation as a whole, the data can be used to study New York City victimization trends. This is because the New York City’s population base is greater than 5 million and there are fewer concerns regarding identification of victims within the data and unstable estimates (too small of a population may provide volatile estimates). A small body of research has used the NCVS data for local crime analysis including for Chicago, Los Angeles, Dallas, and Houston (Rezey & Lauritsen, 2023; Xie, Ortiz Solis, & Chauhan, 2024). Other research has examined New York City data (Langan & Durose, 2009; Xie, Ortiz Solis, & Chauhan, 2024) but the analyses was more limited in the scope and did not extend to 2022. Our study examines the NCVS data in New York City from 1996 to 2022 by nonfatal violence and property crimes. We also assess whether these crimes were reported to the police and, if not, what proportion were not reported because the victims believed the police would not help (a proxy for public confidence in the police). Next, we examine what proportion of victims of violence used victim services, regardless of whether the crime was reported to the police or not. Lastly, we compare New York City to the nation as a whole and to other cities with a population over one million. The analyses are also conducted by race/ethnicity with the caution that the sample size may be too small for reliable estimates. This is the first study to unpack New York City victimization data in this way. Table 1 presents the average number of interviews conducted in New York City annually from 1996 to 2022, compared to the number of interviews for the United States and other cities with populations exceeding one million. The sample size decreases when the data is disaggregated by race/ethnicity, which is a limitation of the NCVS as smaller sample sizes result in larger standard errors of estimates, reducing the results’ precision. To further assist with this issue, a data point in this report for a specific year is a three-year moving average. Three-year moving averages produce more reliable estimates because the annual count of violent incidents in a single city tends to be small. The three-year average violence rate in 1997 is the average violence rate from 1996 to 1998, and so on.

New York: Data Collaborative for Justice, at John Jay College, 2025. 18p.