Oakland Police Stop Data

Authors: Daniel Ochoa Morales and Beatriz Santos

Advisor: Andrea Arauza Rivera

Introduction

The city of Oakland, California is home to over 425,000 people and has one of the most diverse populations in the country. In an effort to better serve residents, the Oakland Police Department (OPD) has been the focus of federal monitoring and has participated in studies to identify policies and practices that have caused harm and distrust.

As a part of these efforts the city makes data on police stops publicly available. The city also releases quarterly reports on the data. In this work we dig deeper into the data on police stops and provide visualizations of various components of the data. We focused on two components of police stops:

The origins and outcomes of stops made in the 2021 July-September Quarter data. This includes examining various aspects of the origin of the stop, the demographic breakdown of those being stopped, and the outcomes of the stops.
Additionally, we used data available for 2016-2019 and 2021 to see the impact of policy change on the frequency of stops. We also examine how policy change can mitigate the inequities in stops.

Origins and Outcomes

In order to understand the origins of stops by OPD, we first get a sense of why the stop was initiated. We focused on 3 factors that pertain to the initiation of a stop:

Was the officer dispatched to the scene to make the stop, or was the stop made by the officer while on patrol? In either the dispatch or non-dispatch setting, was the stop supported by “Intel-led factors”, meaning OPD requires officers to have knowledge of possible criminal activity? The knowledge can be specific such as having the information and or description of a suspect or can be vague such as general idea of a criminal trend in certain areas.
The above yields 4 types of stop origins. For each type we find the demographic breakdown of those being stopped.
Finally we examine both the reasons given to justify the stop as well as the outcomes of the stops.

Based on data provided by OPD in the OPD Quarterly Stop Data Report, we can see that 57.9% of stops are initiated because an officer was dispatched to the scene and 42.1% of stops were made by an officer on patrol. We then look for the presence of intel-led factors in each category, and find that dispatched stops are more likely to be based on previous intel (52.9% of dispatch calls are intel-led) while non-dispatch calls are more likely to be done without previous intel (41.6% of non-dispatch calls are intel-led).

We next looked at the demographics of people in each type of stop: dispatch stops with and without intel led factors, and non-dispatch stops with and without intel led factors.

In terms of gender we observe that on average 26.65% of those stopped are female and 72.7% are males. These are the averages across the 4 types of stops. Specific breakdowns for each of the 4 types of stops are below.

Dispatch with no intel led factor present: 29.6% Females, 70.3% Males, 0.1% Unknown sex
Dispatch with intel led factors present: 27.4% Females, 72.5% Males. 0.1% Unknown sex.
Non-dispatch with no intel led factors present: 27.8% are Females, 72.2% Males, and 0.0% Unknown sex
Non-dispatch with intel led factors present: 21.8% females, 77.8% males, 0.3% Unknown sex

We next looked at the racial demographics of those being stopped. Across the 4 types of stops we see African Americans are stopped at higher rates than those in other racial groups. Comparing the proportions in the charts below to the chart displaying the demographics of the city of Oakland, we can see that African Americans are being stopped at disproportionately higher rates.

Reasons for Stop

Having looked at the demographics of those being stopped, we next look into what reasons are stated for the stops. We note some observations based on the charts provided below:

We can see that there is a significantly larger proportion of stops made for traffic violations during non-dispatch stops with no intel led factors, 80.6%, compared to dispatch stops with no intel led factors which was, 2.7%.
We can also see that when intel led factors are present we have a greater percentage of probable cause stops, 43.8% for dispatch stops and 47.8% for non-dispatch stops. These percentages drop when intel led factors are not present: 39.3% for dispatch stops and 11.4% for non-dispatch stops.

Result of Stop

Next we turn to the results of the stops. Here we have 6 possible outcomes: Arrest, 5150, Citation, Warning, No Action, Other. The outcome for 5150 is code for ‘Psychiatric hold’. The label ‘ Other’ can be a combination of different outcomes such as referral to school administration or transportation by ambulance (at times this results in rows summing to greater than 100%). Below we have a table indicating race, stop type, and result. The table entries indicate the percentage of stops yilding each result.

Effects of Policy Change

Following Allen v. City of Oakland in 2003, the Oakland Police Department was placed under federal monitoring and was required to force reforms. The monitoring was only supposed to last five years but is still in place to this day. It has spanned over nine-teen years, the longest of any federal monitoring program. Several policy changes have been implemented over the past two decades but the one that has impacted stop and searches during these past five years is cutting back on low-level traffic violations starting in 2017-2018. “Capt. Christopher Bolton of OPD said that they might not even stop motorists for rolling through a stop sign, if no one is crossing the street and the car doesn’t pose an imminent threat to public safety” [Swan]. OPD stops for low-level traffic violations has been criticized as a method to search ethnic minorities for a reason to arrest them [Swan]. Now OPD has dramatically cut back on enforcing these stops.

From OPD yearly stop data, we can see that traffic stops accounted for most of the total stops from 2016-2019. The total number of stops per year continued to decrease and the total stops for traffic violations were cut in half by 2018. By 2021, fewer than 5,000 were stopped for traffic violations from the original 25,000 in 2016. Also by 2021, probable cause had taken over as the leading reason for stops but

probable cause has also been decreasing since 2017. Probation/Parole stops practically ceased to exist. We believe that it is correlated with the decrease in stops for traffic violations because the data provided by OPD shows that traffic violations were the main reason for probation searches. We believe that this indicates that officers were at times filing in the reason for stop as probation/parole instead of traffic violation.

Prior to the policy change, African-Americans accounted for almost 70% of arrests in 2016 with Hispanics in a distant second of 17.5%. Men also accounted for 75% of all arrests in 2017. Starting from 2017 when the policy change occured, arrests for African-Americans decreased significantly. By 2021, arrests for all ethnicities were lower than in 2016. Additionally, by 2021 African-Americans made up 56% of total arrests, a 14% decrease from 2016. Arrests by gender only showed minor changes with the total percentage of men arrested slightly increasing.

Conclusion

We believe that increased transparency and citizen involvement is a good first step in building trust between the citizens of Oakland and OPD. While OPD generates quarterly reports on its stops, we believe that resorts which include a narrative of the analysis and visualizations of the data can facilitate meaningful engagement with the community. Additionally, we have shown how policy change can have a significant effect on greater equity when it comes to police stops. For additional work studying the data released by the Oakland Police Department see [Voigt].

References

[OPD] Oakland Police Stop Data. https://www.oaklandca.gov/resources/stop-data

[Swan] Rachel Swan.To curb racial bias, Oakland police are pulling fewer people over. Will it work? San Francisco Chronicle, 2019. https://www.sfchronicle.com/bayarea/article/To-curb-racial-bias-Oakland-police-are-pulling-14839567.php

[Voigt] Voigt, Rob, et al. "Language from police body camera footage shows racial disparities in officer respect." Proceedings of the National Academy of Sciences 114.25 (2017): 6521-6526.

Page updated

Report abuse