New study examines the impacts of the pandemic and George Floyd’s death on crime and policing
After the killing of George Floyd in 2020, protests erupted throughout many cities and calls for defunding the police became the new norm. This also occurred at the same time as the pandemic, which police responded to in myriad ways. A new study by researchers at Arizona State University examined the independent impacts of George Floyd’s death and the pandemic on crime and police work. Overall, the study found that both events had independent impacts on some measures of crime and police activity. The pandemic seemed to have a greater impact on crime while George Floyd’s death seemed to have a greater impact on police proactivity. The impacts on police work were massive and are consistent with the de-policing hypothesis.
Two theories guided this research: routine activities and legal cynicism.
Routine activity theory posits that crime is more likely to occur when a motivated offender and suitable target converge at the right time and place, when no “capable guardian” (e.g., a police officer) is present. Changes in population mobility and routine activities, such as those brought on by the pandemic, could theoretically increase or decrease the chances of crime occurring. Alternatively, changes in these factors could impact police activity, which could then subsequently affect crime patterns. For example, some argue that decreases in certain types of crimes might be due to changes in policing strategy that reduce enforcement of certain crimes.
Legal cynicism refers to the public’s perception that the legal system (e.g., law enforcement) are “illegitimate, unresponsive, and ill equipped.” Legal cynicism of police increases when citizens perceive that police abuse their authority. Thus, it makes sense that legal cynicism increases around times of “viral” police encounters, such as in the case of George Floyd’s death. Research has found that legal cynicism subsequently increases criminal conduct. Further, many believe that increased legal cynicism can affect police morale and reduce their proactive policing efforts, a phenomenon dubbed the “de-policing effect.”
The pandemic dramatically altered the routines of many people across the globe, including police. Police departments had to adapt policy and practice following the pandemic. This included things like eliminating in-person roll calls, suspending academy classes and in-service training, cancelling community-oriented policing efforts, and re-allocating officers from specialized units to patrol. Many departments changed shift schedules to limit officers’ exposure to each other, and other new policies limited in-person contact between officers and citizens. For example, some agencies required officers to stay in their patrol vehicles whenever possible during citizen encounters, while others pressured officers to reduce their arrests for low-level crimes. All of these changes altered the routine activities of nearly everyone, which likely played a role in terms of the pandemic’s impact on crime.
In recent years, the public has become fixated on instances involving police brutality and use of force. This intensified following the death of George Floyd, to the point where there is now a widespread movement committed to de-funding the police. Not surprisingly, these factors likely contribute to increased legal cynicism, which can negatively impact police morale and their proactivity in responding to crime.
Domestic violence has been a key focus of pandemic-related research. Many people hypothesized that domestic violence incidents would increase due to the nature and duration of stay-at-home orders, which extended periods of contact between family members and domestic partners. This phenomenon might be explained by routine activity theory, because all three conditions specified by the theory would converge: a motivated offender, a suitable target, and lack of a guardian. Sure enough, an increase in domestic violence cases was shown in some research, including studies in Dallas, Los Angeles, Indianapolis, and Chicago. Similarly, a multi-site study documented increases in domestic violence cases in five jurisdictions, including New Orleans, Seattle, Salt Lake City, Montgomery County (MD), and Phoenix (they did not find the same pattern in Cincinnati, though). The impacts on domestic violence seemed somewhat short-term though, with increases being most pronounced during earlier stages of stay-at-home orders and social distancing protocols.
Early evidence also suggested small increases in overall crime throughout the United States during the pandemic, which was seen in Los Angeles, San Francisco, Detroit, and Chicago, among other places. One multi-site study examined aggregate-level data from 10 cities, including Baltimore, Cincinnati, Los Angeles, New Orleans, Phoenix, San Diego, Seattle, Sonoma County (CA), and St. Petersburg. They found that overall crime across the cities trended slightly downward, but impacts varied by crime type. Notably, there were large increases in certain types of crime, such as murder. The overall downward trend was likely driven by substantial reductions in certain types of calls (such as traffic-related offenses), or possibly by reductions in crime reporting for certain types of crime. Another multi-site study of 25 cities found similar results following the onset of the pandemic, documenting immediate decreases in drug crimes, theft, and residential burglary, alongside increases in car theft and murder. Further, FBI statistics released in September 2021 officially confirmed that the United States experienced an alarming increase in violence from 2019 to 2020, including nearly a 30% increase in homicide.
The Current Study
The current study, conducted in Tempe (AZ), sought to examine the independent effects of two events–the pandemic and George Floyd’s death–on crime and police work. The researchers examined whether these events: 1) contributed to the spike in violence; 2) altered proactive policing, arrests, and police use of force; and 3) affected officers’ professionalism when interacting with citizens.
Like other departments, the Tempe Police Department (TPD) enacted several policy changes following the pandemic. One of the major changes was the implementation of a ‘tier system’ that allowed for re-allocation of personnel to different assignments as needed, including remote telecommuting options for certain personnel. The department also encouraged officers to use citations in lieu of arrests for low-level offenses, and sworn personnel were instructed to prioritize calls for service that presented a danger to one’s life or property.
The death of George Floyd on May 25, 2020 also greatly impacted the city of Tempe, with citizen protests occurring throughout June and July 2020. Not surprisingly, tension between police and the public increased during this time. On July 27, these protests caused tension between officers and protestors that resulted in seven arrests and an officer injury.
The authors tested four hypotheses, two that related to the pandemic and two that related to George Floyd’s death. Regarding the pandemic, the authors expected to see declines in both crime and police activity following the onset of the pandemic. They also expected to see more negative dynamics during police-citizen interactions (e.g., procedural justice, de-escalation, aggression) following the onset of the pandemic. Regarding George Floyd’s death, the authors expected to see increases in both crime and police activity. They also expected to see more negative dynamics during police-citizen interactions following George Floyd’s death.
The study examined weekly trends in crime and officer activities from January 1, 2017 to January 24th, 2021. This time frame covered approximately 3 years before the pandemic, 10 months after the onset of the pandemic (in the United States), and 8 months after George Floyd’s death. Researchers collected data on calls for service, reported crime, arrests, and use of force. Data were extracted from police department repositories, including the records management system (RMS), the case management system (IAPro), and body-worn camera (BWC) footage. Crime was measured based on reported crime numbers and citizen-initiated calls for service, while police activity was measured in terms of officer-initiated calls for service, arrests, and use of force. All of these outcomes were examined and compared across three time periods. Period 1 is the “pre-pandemic” period (covering about three years preceding the pandemic, from January 1, 2017–March 12, 2020), period 2 is the “pandemic only” period (covering 8-10 months following the onset of the pandemic, with exact dates depending on outcome) and period 3 is the “pandemic and George Floyd” period (covering 3-8 months following George Floyd’s death, with exact dates depending on the outcome).
Weekly offense totals for reported crime were calculated for the study period and were separated into three broad categories (crimes against persons, crimes against property, and crimes against society). The authors also examined assaults, burglary, and domestic violence incidents separately. Homicides were not examined separately because of the low occurrence in Tempe. Weekly counts of calls for service were also examined, which were divided by citizen-initiated calls (i.e., 911 calls) and officer-initiated calls (i.e., proactive contacts). Weekly counts of arrests and use-of-force incidents were also examined. Lastly, the researchers extracted and reviewed randomly-selected BWC footage to observe officer behaviors during 475 interactions and determine whether officers handled citizen encounters in a consistent manner before and after both events. All interactions were coded by multiple people with high inter-rater reliability between coders.
A three-step approach was used to analyze data. First, the authors conducted aggregate-level descriptive examinations to describe overall trends both throughout the study period (January 2017 to January 2021) and in the three smaller time periods (pre-pandemic, pandemic only, and pandemic + George Floyd). Second, the authors used a statistical analysis technique known as interrupted time series analysis (i.e., an ARIMA model), which allowed authors to compare pre-intervention and post-intervention values for each outcome. The authors used this to test whether weekly counts for each measure changed significantly with the onset of the pandemic and George Floyd’s death. The first time series model tested for immediate impacts in the weeks following each event (i.e., the week of March 8 for the pandemic and the week of May 24 for George Floyd). The second time series model tested for gradual impacts of varying durations ranging from one month through the end of the study period.
Third, to examine the data extracted from BWC footage, the researchers used bivariate analysis (i.e., independent samples t-tests) to assess whether the prevalence of each variable changed across the three time periods. There a total of 50 variables that were examined that captured the dynamics of how officers handled encounters (e.g., spoke in a calm manner, applied de-escalation tactics).
Impacts on Crime
Weekly averages of citizen-initiated calls and weekly averages of reported crime both decreased significantly from pre-pandemic to pandemic only periods, by about 22% and 19%, respectively. The authors also identified separate statistically significant declines in both citizen-initiated calls and reported crime beginning the week after George Floyd’s death.
When looking at crimes against persons, there were no significant changes resulting from the pandemic nor from George Floyd’s death. Overall crimes against persons decreased slightly following the onset of the pandemic and increased slightly after George Floyd’s death, but neither of these findings were statistically significant. Similarly, assaults also decreased slightly following the onset of the pandemic and increased slightly after George Floyd’s death, but these findings were also non-significant. Robberies remained even throughout the pre-pandemic, pandemic only, and George Floyd periods, with no significant changes over time. Domestic violence cases experienced small increases following both the onset of the pandemic and following George Floyd’s death, but neither increase was significant.
Crimes against property and crimes against society appeared affected by the pandemic but not affected by George Floyd’s death. Overall crimes against property declined significantly by 16% from pre-pandemic to pandemic only periods. After George Floyd’s death, the weekly average began increasing again, but this was not statistically significant. Burglaries followed a similar pattern, with a statistically significant decline of 19.3% from pre-pandemic to pandemic only periods. After George Floyd’s death, this increased again, but the increase was not statistically significant. Overall crimes against society also declined significantly by nearly 28% from pre-pandemic to pandemic only periods. This started increasing again after George Floyd’s death, but the increase was not statistically significant.
Impacts on Police Activity
Weekly averages of officer-initiated calls for service (i.e., police proactivity) decreased by about 11% after the onset of the pandemic. There was another massive decline of 42.6% following George Floyd’s death. Both declines were statistically significant, though officer-initiated calls did begin to increase again throughout the remainder of the study period.
Weekly arrests declined significantly by 45.5% after the onset of the pandemic, but this only lasted a few weeks. By April 2020, arrests were increasing again, though this increase was not statistically significant, and arrests remained lower than pre-pandemic levels during the remainder of the study period.
Police use of force remained steady throughout the pre-pandemic period, but increased 22% during the pandemic only period. This increased again by 7.5% after George Floyd’s death. Both increases were statistically significant.
Several important findings emerged from the BWC review as well. First, the average length of encounters declined by 22.5% (3.5 minutes) over time. Authors surmise that officers may have resolved encounters more quickly to reduce infection risk to themselves and others, as well as in response to increased scrutiny following George Floyd’s death.
Second, officers displayed high levels of procedural justice with citizens throughout all three time periods. Across the three time periods, officers continued to handle encounters in ways that minimized the potential for a violent outcome. For example, more than 90% of officers consistently spoke in a calm manner, maintained sufficient personal space, controlled their emotions, and and used time to their advantage. In addition, fewer than 6% of officers used charged body language, ignored what the citizen was saying, or lost their patience.
However, there were a handful of significant differences reflecting both negative and positive change. After George Floyd’s death, officers were less likely to use friendly language (39.8% compared to 53.6% pre-pandemic), less likely to compromise (20.6% compared to 53.2% pre-pandemic), less likely to encourage citizens that a positive outcome could be reached (19.5% compared to 60.7% pre-pandemic), and less likely to attempt to build rapport (19.5% compared to 50.7% pre-pandemic). Alternatively, several dimensions trended positively after George Floyd’s death. For example, officers were slightly more likely to speak in a respectful manner (98.8% compared to 92.9% pandemic only) and were less likely to act impersonally (2.4% compared to 11.4% pre-pandemic).
What Does it Mean?
The impacts of the pandemic and George Floyd’s death on crime appeared mixed. George Floyd’s death had an independent impact on some measures of crime and police activity, though the pandemic appeared to have stronger impacts for nearly all outcomes. One exception is officer proactivity, which was more impacted by George Floyd’s death than the pandemic.
All reported crime dropped significantly after the pandemic began (-19%), as did citizen-initiated calls for service (-22%). Most of this was driven by declines in property offenses such as burglaries, which likely were affected by underreporting and stay-at-home orders. The decline lasted five months, continuing after Floyd’s death. Crimes against property and society, as well as burglary, experienced major declines associated with the pandemic, but those effects lasted less than a month (ending before George Floyd’s death). Crimes against persons, as well as assaults, robbery, and domestic violence were not affected at all by the pandemic.
Crime did not spike after George Floyd’s death as expected; in fact, there were actually declines in overall reported crime (about -25%) and citizen-initiated calls for service (about -20%). This might be impacted by the fact that police made fewer arrests, or it could also reflect reduced reporting rates. Another possible explanation may have been revealed in the BWC footage: prior to George Floyd’s death, officers exhibited high levels of procedural justice when interacting with citizens, and this did not change following his death. This may have partially mitigated the rising tensions between police and the public that occurred in many other cities.
There were massive drops in officer proactivity following the onset of the pandemic and following George Floyd’s death. The pandemic altered every aspect of police activity, resulting in large significant reductions in officer-initiated calls (-11%) and arrests (-46%), while George Floyd’s death was associated with a 43% reduction in officer-initiated calls (but no change in arrests). The decrease in police proactivity initially following the pandemic is likely explained policy changes to reduce infection risk (e.g., shift and unit adjustments, reduced informal contacts). The decline after George Floyd’s death is not explained by such a policy change, though. The more plausible explanation is the de-policing theory: officer-initiated calls declined in response to increased scrutiny and outrage after George Floyd’s death that led to an informal pullback in police proactivity. Importantly though, police proactivity climbed back up to pre-pandemic levels before the end of 2020. Nonetheless, these findings raise particular concerns about how police can maintain their effectiveness and proactivity despite external forces (i.e., policy changes associated with the pandemic) and internal fears of becoming the next “viral” police incident.
Police use of force actually increased with the onset of the pandemic (+22%), and continued to increase after George Floyd’s death (+7.5%). In effect, there were far fewer police-citizen encounters, but they were more likely to involve force. There are a few possible explanations for this. First, as the pandemic began, officers focused their response to the most serious calls, indicating they were responding to higher-risk incidents more often. Secondly, citizen frustration with the pandemic and anti-police sentiments increasing after George Floyd’s death may have led to greater aggression between the police and the public. Relatedly, it is possible that police were dealing with more individuals who expressed animosity and behaved in threatening manners while engaging with police.
Changes over time associated with the pandemic might be explained by routine activities theory. Stay-at-home orders, closures of schools and businesses, and fear of infection limited the availability of victims and restricted the movement of offenders, greatly reducing opportunities for crime. At the same time, capable guardians (e.g., police) were still present, though they too were affected by the pandemic. Many of the effects on crime were short-term, while other crimes (i.e., assaults, robbery, domestic violence) were not impacted. One possible explanation is that the stay-at-home orders were only effective for a few weeks and people may have quickly decided to not abide by the order, which was not strictly enforced by police (only encouraged). After Arizona’s stay-at-home order expired on May 15, 2020, crime rates started to climb back up. Another possibility is that significant increases in alcohol use in 2020 contributed to rising crime rates.
Of course, the study did have limitations. First, this study tells the story of only one city, and the extent to which Tempe’s story aligns with the experiences of other jurisdictions remains unknown. Local variation in the response to the pandemic likely contribute to differing impacts on crime and police work. The same can be said for George Floyd’s death. In essence, the impact of these events likely differs from one city to the next. Tempe also appeared to have higher levels of procedural justice during their interactions with citizens, which is not necessarily true in all cities. However, Tempe has characteristics common to many other jurisdictions in both Arizona and across the U.S. — it is a medium-sized city that borders a large metropolitan area and is home to a large university. Tempe tends to experience few homicides but has higher-than-average property crime rates. Even though local context can certainly impact results, these findings still have implications that extend beyond Tempe.
Second, each data source examined here also has limitations. The issues of administrative police data are well-known, and BWC footage also suffers from many shortcomings, such as failure to record encounters, obstructed views, and the like. Further, the 474 videos reviewed in this study were randomly selected and only represent a very small proportion of the total encounters. Lastly, the authors treated the global pandemic and George Floyd’s death as events with exact start dates, implying that each is a one-time occurrence, which they are not.
The current study is one of the most comprehensive efforts to date to capture the impact of COVID-19 and George Floyd’s death on police work. Overall, the pandemic had a much larger impact on crime and other features of police work than did George Floyd’s death. The one exception was the impact on officer-initiated calls for service, which were more impacted by George Floyd’s death than the pandemic.
The research found that independent impacts of the pandemic and George Floyd’s death on crime were mixed and varied notably by crime type, while both events were significantly associated with decreases in proactive policing. After the onset of the pandemic, Tempe saw decreases in some types of crime, citizen-initiated and officer-initiated calls for service, and arrests. Interestingly, police use of force spiked following the start of the pandemic and continued to rise after Floyd’s death. In the face of these changes, BWC footage revealed consistency in how officers effectively handled citizen encounters. In some cases, they improved their interactions with citizens following George Floyd’s death.
This research supports the idea of the de-policing effect. One major concern going forward is how to encourage proactive policing despite increased scrutiny of the police and changing social factors.