Student Spotlight: Racial Bias in the NYPD Stop-and-frisk Policy

Donald Trump recently came out in favor of an old New York Police Department’s (NYPD) “stop-and-frisk” policy that allowed police officers to stop, question and frisk individuals for weapons or illegal items. This policy was under harsh criticism for racial profiling and was declared in violation of the constitution by a federal judge in 2013.

An earlier post by Krit Petrachainan showed a potential racial discrimination against African-Americans within different precincts. Expanding on this topic, we decided to look at data in 2014, one year after the policy had been reformed, but when major official policy changes had not yet taken place.

More specifically, this study examined whether race (Black or White) influenced the chance of being frisked after being stopped in NYC in 2014 after taking physical appearance, population distribution among different race suspect, and suspected crime types into account.

2014 Data From NYPD and Study Results

For this study, we used the 2014 Stop, Question and Frisked dataset retrieved from the city of New York Police Department. After data cleaning, the final dataset has 22054 observations. To address our research question, we built a logistic regression model and ran a drop-in-deviance test to determine the importance of Race variable in our model.

Our results suggest that after the suspect is stopped, race does not significantly influence the chance of being frisked in NYC in 2014. A drop-in-deviance tests after creating a logistic regression model predicting the likelihood of being frisked gave a G-statistic of 8.99, and corresponding p-value of 0.061. This marginally significant p-value shows we do not have enough evidence to conclude that adding terms associated with Race improves the predictive power of the model.

Logistic regression plot predicting probability of being frisked from precinct population Black, compared across race

Figure 1. Logistic regression plot predicting probability of being frisked from precinct population Black, compared across race

To better visualize the relationship in interactions between race and other variables, we created logistic regression plots predicting the probability of being frisked from either Black Pop or Age, and bar charts comparing proportion of suspects frisked across sex and race.

Interestingly, given that the suspects are stopped, as the precinct proportion of Blacks increases, both Black and White suspects are more likely to be frisked. Furthermore, this trend is more profound for Black than White suspects (Figure 1).

Additionally, young Black suspects are much more likely than their White counterparts to be frisked, given that they are stopped. This difference diminishes as suspect age increases (Figure 2).

Logistic regression plot predicting probability of being frisked from age, compared across age

Figure 2. Logistic regression plot predicting probability of being frisked from age, compared across age

Finally, male suspects are much more likely to be frisked than females, given they are stopped (Figure 3). However, the bar charts indicate that the effect of race on the probability of being frisked does not depend on gender.

Proportion frisked by race, compared across sex

Figure 3. Proportion frisked by race, compared across sex

Is stop-and-frisk prone to racial bias?

Our results suggest that given that the suspect is stopped, after taking other external factors into account, race does not significantly influence the chance of being frisked in NYC in 2014. However, after looking at relationships between race and precinct population Black, age, and sex, there is a possibility that the NYPD “stop-and-frisk” practices are prone to racism, posing threat to minority citizens in NYC. It is crucial that the NYPD continue to evaluate its “stop-and-frisk” policy and make appropriate changes to the policy and/or police officer training in order to prevent racial profiling at any level of investigation.

*** This study by Linh Pham, Takahiro Omura, and Han Trinh won 2nd place in the 2016 Undergraduate Class Project Competition (USCLAP).

Check out the 2016 USCLAP Winners here.

Enter your e-mail address to receive notifications of new blog posts.
You can leave the list at any time. Removal instructions are included in each message.

Powered by WPNewsman

Please like & share:

Does Marriage Affect Earning Potential?

Using DASIL’s United States Income Data by Marital Status, Race, and Sex visualization, one can see how the effect of marriage on a person’s earnings is multifaceted in nature: it depends on who we focus on and other factors at play. However, there are general trends that do prevail.


Married people overall have higher earnings, although the difference between divorced people is smaller than that of single people. Married people with a spouse present earned over $33 annually, while single people earned on average well over $10,000 less than married people with a spouse present. While it may appear that being single correlates to lower earnings, inter-related variables may explain some of the earning discrepancies observed.


One important variable to consider is the effect of age. As we discuss in another blogpost, workers ages 15-24 earn less than those of other age brackets. Studies suggest that those belonging to the 15-24 age bracket are less likely to be married, so some of the earning trends shown may not be strictly due to marriage. In addition, as illustrated in the aforementioned blogpost, 25-34 year-olds and 65+ year-olds make about the same and the next least age demographic (about $25000 more in 2013 dollars), and 35-64 make about $20,000 more on average. The 35-64 year-olds are more likely to be established in their careers, earning their highest-paying years within this age bracket. So, some earnings trends may be attributed to the pace of a career’s trajectory.

Breaking down by gender, the general trend persists: married men make a lot more than divorced and single men of all races, $44k, $33K, and $20k respectively. Married women have been making more than single men in recent years, averaging about $2K more in 2006 and persisting into 2010. While single women made more than married women in the 80s, the trend has reversed in recent years.



Breaking down by race, both Asian single men and women make more than any other singles demographically, at both averaging about $21K in 2010. Hispanic single women make the least of all demographics of men and women, at $15.1K, although Black single men are a close second. Earnings of Black single men peaked in 1998, only separated from white men by about a $200 difference. Studies attribute this peak to the economic boom of the 1990s and the transition of Black men into higher-skilled service-industry jobs.



Married Hispanic women still make less in comparison to all other married women, at $19.1K, but still substantially more than if they are single. Black females top the earnings compared to women of other races, at $26.6K, with the trend moving more or less in the same way as Asian married women.

Enter your e-mail address to receive notifications of new blog posts.
You can leave the list at any time. Removal instructions are included in each message.

Powered by WPNewsman

Please like & share:

Investigating Police Brutality in Los Angeles

Excessive use of force by law enforcement is by no means a novel phenomenon in the United States. However, with high-profile cases like Michael Brown, Eric Garner, and most recently Greg Gunn, fueling national movements such as #BlackLivesMatter, race-related incidences of police brutality are receiving worldwide media attention.

I investigated geographic trends in reported police brutality, using Los Angeles County at the census tract level and data from The Guardian’s project “The Counted,” a comprehensive dataset that records all people killed by police and other law enforcement agencies in the US, for the year 2015.

To measure the effect of location on incidences of police brutality, I conducted a hot spot analysis, which identifies statistically significant spatial clusters of high (hot spots) and low police brutality (cold spots). Essentially, the hot spots/cold spots indicate whether observed spatial clustering of police brutality events is more pronounced than if the values were randomly distributed. We specified the spatial relationship for the analysis as Contiguity Edges, meaning that census tracts that share a boundary or overlap with a census tract that contains a police brutality event will be weighted more that those that don’t in the analysis.

Below is a map depicting the results of the hot spot analysis.


The hot spots depicted in the map reveal the relationship between location and the occurrence of police brutality. The neighborhoods enveloped in hot spots are those with an abnormally high number of police brutality events, indicating that these areas may be disproportionately affected by excessive use of force by law enforcement.

Looking demographically at both the incidences themselves and these hot spot neighborhoods can shed some light on why these areas have abnormally high police brutality. Right off the bat, the number of blue and green dots (Hispanic/Latino and black victims, respectively), dominates the map. Breaking down by race, there were 30 victims of Hispanic/Latino descent, 11 black, 4 Asian/Pacific Islander, 7 white, and 1 Arab-American. In addition, most of the incidences with blacks as victims happen in LA neighborhoods that have a large population of blacks, such as Willowbrook and Westmont. The same trend also appears when focusing on Hispanic/Latino victims: most Hispanic/Latino victims died in neighborhoods with large populations of Hispanics/Latinos, such as Los Angeles proper and Eastern LA County (Baldwin Park, Irwindale, West Covina).

Enter your e-mail address to receive notifications of new blog posts.
You can leave the list at any time. Removal instructions are included in each message.

Powered by WPNewsman


Please like & share:

The Mass Shooting Epidemic in the United States

An examination of Stanford University’s Mass Shootings of America (MSA) dataset shows why shootings have been making the headlines in the U.S. and gun violence has become a big issue addressed in the campaigns of presidential hopefuls. Stanford MSA defines a mass shooting as “3 or more shooting victims (not necessarily fatalities), not including the shooter. The shooting must not be identifiably gang or drug related” (Stanford Mass Shootings in America, courtesy of the Stanford Geospatial Center and Stanford Libraries).


The dramatic change in the number of mass shootings in the past two years is readily apparent. There were 121 mass shooting events from 1966 to 2009, but 116 just in the past 5 years. 2015 alone had 65 separate instances of mass shootings. In terms of total number of fatalities, the past 7 years are noticeably thicker than earlier years. Even in years with low numbers of mass shootings, such as 1991 which only had 5 incidences, there were a large number of fatalities (47).



The Southern states had the largest numbers of mass shootings in 2015. Florida led with 6. Even though Texas had fewer mass shootings (4), the state sustained the most fatalities, 20. North Dakota and New Hampshire are the only 2 states that have not experienced any mass shootings in the 49 year time period covered by the data (not shown). In 2015 39 mass shootings occurred in residential homes & neighborhoods, while 21 happened in public places. Back in the late 90s, schools were the primary target of mass shooters, with 3 incidents in 1997 and 1999, each.


Most of the mass shootings in the past 8 years have stemmed from a variant of an altercation, be it domestic, legal, financial, or school-related. Of course it can always be argued that all mass shooters have mental health issues, but contrary to popular belief, according to these classifications,  shooters’ mental health issues as a direct motive for shootings  has not increased in recent years, with only one incident in 2015 attributed to mental health issues. Perhaps what’s most troubling is the high number of cases where a motive can’t be identified, 23 in 2015, suggesting the need for further, more comprehensive study into the underlying causes of these mass shootings.

Many pundits largely attribute the US-specific phenomenon to things to lax gun policy. However, any progress to change gun laws, even to fund research into the causes of gun violence, has been (and continues to be) stymied by the gun lobby, led by the National Rifle Association (NRA). Re-examining the nation’s access to guns is imperative, and those in Congress who are funded by the gun lobby need to be open to that re-examination. While the data available is informative, unfettered research is integral to truly understanding the nature of gun violence and to finding effective policy solutions.

Enter your e-mail address to receive notifications of new blog posts.
You can leave the list at any time. Removal instructions are included in each message.

Powered by WPNewsman

Please like & share:

Data Across the Curriculum: The Explanatory Power of Data in Global Development & Geography

The field of geography is split into two camps: critical scholars, who are skeptical of data because they believe it silences certain voices within society and fails to explain process and context, and empirical scholars, who incorporate data to create empirical models that explain geographic concepts and trends.  Leif  Brottem, Assistant Professor of Political Science with a PhD in Geography, is a firm believer in the importance of both critical and empirical approaches.  Data analysis can compensate for and expand upon the limits of text and qualitative evidence. His focus on data analysis as a tool that illustrates narrative is evident in the work of each of his three classes, Introduction to Global Development Studies (GDS), Introduction to Geographical Analysis and Cartography, and Climate Change, Development, and Environment.

In his Introduction to Global Development Studies, for instance, Brottem utilizes infographics & charts to explain basic concepts, and utilizes data tools such as GapMinder to illustrate change over time and regional differences pertaining to a variety of development indicators. His students also complete two data analysis exercises as a part of the class: one exercise asks students to study the relationship between economic development and social development indicators, and the second has students explore different aspects of population dynamics such as carrying capacity, limits to growth and the determinants of population growth.

In Brottem’s Introduction to Geographical Analysis and Cartography course, students learn both the basic critical perspectives on how to evaluate maps and understand their overt and covert messages and practical techniques for making maps using Geographical Information Systems software.  Students complete in-class exercises and take-home labs that require creating data and using data to solve problems.


Finally, in Climate Change, Development, and Environment, Brottem utilizes data analysis in the form of topic-modeling: students investigate textual trends in various sources, from tweets to scholarly articles, using the MALLET topic model package. In addition, his students also work with Nvivo to conduct further qualitative analysis, and GIS to visualize spatial trends.

Working with data builds data literacy, a marketable and necessary skill in the real world that Brottem says isn’t typically developed in a liberal arts settings. Building data literacy is especially important in his introductory classes, because he has students who wouldn’t otherwise be exposed to data, and aims to get them comfortable with using data and reduce their fears of data, numbers, and data analysis. Brottem strongly believes that data is a powerful explanatory tool that helps students think of different ways to look at the world and their studies, beyond theory.

Enter your e-mail address to receive notifications of new blog posts.
You can leave the list at any time. Removal instructions are included in each message.

Powered by WPNewsman

Please like & share: