Introduction
In this article we take a look at data collected on police involved shootings from 2015- May 2020. Using the dashboard below you will be able to examine for yourself the data and view our key findings. The original data set was generated and distributed on GitHub by the Washington Post. The data provided by police seems somewhat incomplete. Many fields are undetermined or not collected at all. Even so, it is one of the more complete data sets available at this time. We collected our version here and combined it with generalized data on racial demographics by state from the year 2018, provided by the Kaiser Family Foundation.
Let’s take a moment to clarify the possible limitations in the data before we begin. Here are some important things to keep in mind before diving in:
- This goes through only mid-2020.
- The data tells us what happened and not what may happen or what is likely to happen. This is a snapshot that provides a starting place to identify important questions and additional data to be collected.
- The states of Rhode Island, Vermont, North Dakota, and Wyoming have relatively small sample sizes (fewer than 25 people killed by police) and therefore the percentage of unarmed appears heavily skewed. This is true of other states with marginally higher numbers and the question of statistical significance is an important consideration.
- We are considering people to be unarmed if they were recorded as unarmed, had a toy weapon, or it was undetermined whether they had a weapon or not.
A Look at the Data
The charts below can be adjusted by making changes to the filters shown at the top. The filters themselves reveal insights into how much and what type of data was collected. For instance, the vast majority of shootings across the United States during the period do not have body camera footage.
As you examine the data you will see differences across states and racial demographics that raise important questions. The first set looks at racial differences and provides a generalized reference for population distributions. The second set takes a deeper dive into the data and focuses on state differences. After you have a chance to examine the charts a bit you can view some of our key findings below. There are arrows to navigate between charts in the bottom left corner of the dashboard.
Key Findings
- The vast majority of people shot by police were armed with either a gun or knife (approximately 75% overall).
- Overall 15% were unarmed. This skews higher for black and hispanic people. In many states with significant black and latinx populations we see more were killed by police proportionate to their population and more of those killed were unarmed when compared to their white counterparts.
- Those killed with an undetermined/”other” threat level (when excluding “attack”) were unarmed 20-40% of the time across most states.
Final Thoughts
The findings here support a great deal of superb investigation that has been done by the Washington Post and other publications and foundations. It is important to consider that police face people that are armed and dangerous in the vast majority of cases where they fatally shoot someone. That said, there are a significant number of people being killed by police across the country in situations that likely could be avoided.
There is a clear racial bias at play with police involved shootings in most states. Some states may have differences that appear to diminish this. Regardless, there are an alarming number of people that have been killed by police in the past five years who were black, hispanic, unarmed, not fleeing, not recorded by body cameras, or no identifiable threat level.
Do you have thoughts or insights to share from the data here? Use the form below to get in touch and share your thoughts. We look forward to your input!