What can big data teach us? Assistant Professor Joscha Legewie discusses how he used statistical analysis to find out the times and neighborhoods where New Yorkers complained the most.
Assistant Professor Joscha Legewie is a sociologist whose research explores questions related to education, social inequality, and ethnic relations. Recently, to better understand the conditions under which conflict among New Yorkers emerges, Legewie studied data from 7.7 million time and geo-coded 311 service requests to determine the times and neighborhoods where people complain about their neighbors.
You use what you call ‘quasi-experimental research designs’ to understand sociological behavior. Can you explain that method to us?
Quasi-experimental research designs try to imitate traditional experimental designs or randomized control trials usually when experimental data is not available. There are many situations in which it is invisible to conduct an experiments. Some of my work, for example, focuses on the effect of extreme violence against police officers on the subsequent police treatment of residents and particularly minority groups in pedestrian stops.
Designing an experiment to address this question is hard and faces clear ethical challenges. As an alternative, I use an quasi-experimental design or a natural experiment. This design compares police stops right after with similar stops right before certain events such as the shooting of a police officer. This design relies on the exogenous nature of many events and allows me to examine whether the pattern of stops changed after incidents of extreme violence against police officers and whether there is a race specific pattern to this response. Another advantage is that quasi or natural experiments often use data from real-world settings.
Your current research looks at New Yorker’s complaints. How do you go about figuring out who complains the most and in what neighborhoods the complaints are files?
My work relies on data from 7.7 million time and geo-coded 311 service requests to examine when and where New Yorkers complain about their neighbors making noise, blocking the driveway, or drinking in public. The study tries to understand the conditions under which conflict between neighbors emerges. 311 is a centralized non-emergency telephone number, Internet platform, and smart phone application that allows city residents to file a request for or complain about issues as diverse as birth certificate services, fallen tree removal, or broken heating.
These service requests provide a unique opportunity to get fine-grained information about everyday life tensions between neighbors. It is an interesting example of how social scientist can use “big data” to address new questions or work with information on an entirely new scale. But these new data sources also have limitations and raise privacy concerns. They are generally not designed for social science research and lack important information. The 311 data, for example, does not include information on the identity of the caller so that my work can only study patterns across neighborhoods and over time.
New Yorkers filed 42,896 noise complaints about neighbors in 2011, about 117 per day. Manhattan is clearly ahead of the other boroughs with 16,082 noise complaints in 2011 (37%) followed by Brooklyn with 12,162 complaints (28%). The difference is even more pronounced when you consider that Manhattan is substantially smaller: In Manhatten there are 10 complaints per 1,000 residents compared to 4.8 in Brooklyn.
The map gives a more detailed view showing the number of complaints per 1,000 residents across census tracts in NYC. Some areas around Midtown Manhattan stick out with a very high rate of complaints. But it’s also interesting to see that the rate is higher in some of the gentrified parts of Brooklyn such as Williamsburg.
Some of these patterns reflect relations to neighborhood characteristics that have played an important role in research on urban communities. We find, for example, less complaints in neighborhoods with concentrated disadvantage—i.e. a higher poverty and unemployment rate, lower educational attainment, and more households that receive public assistance.
This pattern might reflect that citizens in disadvantaged communities are less likely to contact a city agency. But residential mobility also plays an important role. We observe more complaints in neighborhoods with fewer homeowners and many residents who recently moved into the neighborhood.
Presumably, this residential instability undermines friendly relations between neighbors so that residents are more likely to call 311 instead of knocking on someones door.
*** Note on references: The two working papers “Racial Profiling in Stop-and-Frisk Operations: How Local Events Trigger Periods of Increased Discrimination” and “Contested Boundaries: Explaining Where Ethno-Racial Diversity Provokes Neighborhood Conflict” (together with Merlin Schaeffer) are available from Joscha Legewie (email@example.com).