CALLS FOR SERVICE AND OFFICER-INITIATED ACTIVITY

This section analyzes recorded police activity and Calls for Service to understand community requests for public safety services and identify areas where police activity may not be aligned with Calls for Service.

Insights on the type and volume of Calls for Service initiated by community members help identify the scope of community needs and the extent to which police are being asked to perform activities that should or could be handled by non-police responders. Mismatches between Calls for Service and deployment can shed light on potential inefficiencies that may be contributing to inequitable policing.

We include data on all community requests for service (“Calls for Service”) as well as all other recorded police activity, including activities that are officer-initiated. We request data only for events involving police, and not those solely addressed by fire or emergency medical responders. Police activity may also be influenced by community requests for public safety services that are not captured in the Calls for Service data, such as public safety concerns communicated to officers in community forums or council meetings. Calls for Service data also do not capture public safety concerns that community members did not call 911 or seek police services to address, making them an incomplete measure of community members’ public safety priorities. For more details on the types of events included in these analyses, see “More information” under the first chart in this section.

We were unable to conduct some analyses that compare Officer-Initiated Activity to Calls for Service because WPD data did not distinguish officer-initiated events from events initiated by a Call for Service. Mapping Calls for Service against Officer-Initiated Activity can reveal areas where officers are disproportionately initiating police activity, as well as what types of enforcement police are initiating that the community has not asked for.

CALLS FOR SERVICE AND OFFICER ACTIVITY BY EVENT TYPE

Police recorded 86,965 total events from January 1, 2014 – July 1, 2020, including Officer-Initiated Activities and police responses to Calls for Service.

For more detail on how events were categorized, see the More Information drop down.

What does this show?

The pie chart above shows two categories of events: those involving Bodily Harm, Property Harm, or Threats, and those involving All Other Event types. The percentages are based on the number of events police recorded in each category, rather than the amount of time officers spent on each event type.

“Bodily Harm, Property Harm and Threats” includes all events involving either Bodily Harm (such as assault, robbery, or kidnapping), Property Harm (such as theft, burglary, arson, or vandalism), or Threats (such as harassment, shots heard, or extortion), regardless of whether the police engagement was Officer-Initiated or requested by the community via a Call for Service.

“All Other Events” includes all events officers engaged in that did not involve Bodily Harm, Property Harm, or Threats, regardless of whether it was Officer-Initiated or requested by the community via a Call for Service.

Calls for Service are reported by community members via 911 calls, non-emergency calls to the department, direct contact with officers, or 311 calls that were routed to the police for response.

Officer-Initiated Activity includes events that officers initiated based on their own observations or assignment, rather than in response to Calls for Service. This includes officers being assigned to particular activities, such as a patrol or school resource assignment, and any requests for assistance from other officers or from outside agencies.

How did we calculate this?

First, we categorized all recorded events into one of the two categories: those involving Bodily Harm, Property Harm, or Threats, and those involving All Other Event types. We then calculated the percentage of all events officers recorded for each category by dividing the number of events in each category by the total number of events in both categories combined.

Data required for this analysis:

COMMUNITY CALLS FOR SERVICE

92% of recorded Calls for Service that officers responded to did not involve reports of Bodily Harm, Property Harm, or Threats. These include requests related to public assistance, nuisances, and medical/fire assistance. Some of these calls may not require an armed response or may be better addressed via community centered response models.

What does this show?

This figure only includes events that were requested by community members via 911 calls, non-emergency calls to the department, direct contact with officers, or 311 calls that were routed to the police for response. It does not include any events categorized as Officer-Initiated Activity.

This figure is meant to highlight the reasons for which community members are most often calling the police. This can help evaluate how often a police response is necessary, and whether there are sufficient Calls for Service that may not require a police response to warrant considering the use of alternative response models.

How did we calculate this?

First, we excluded all events categorized as Officer-Initiated Activity. If the data provided by the police do not indicate which events were Officer-Initiated, we categorized event type descriptions involving warrants, alarms or non-dispatched calls as Officer-Initiated, and then grouped all other event types as Calls for Service. Non-dispatched calls include requests for assistance from other officers or outside law enforcement agencies, activities that appear to be officer assignments (such as “directed patrol” or “walking the beat”), and other activities police typically have some discretion over (such as “traffic stop,” “subject stop,” “area check,” or “follow up”).

Then we categorized the Calls for Service based on the caller’s complaint type and calculated the percentage of calls in each category.

Data required for this analysis:

OFFICER ENGAGEMENT IN MENTAL HEALTH CRISES

Officers engaged in 580 events that were recorded as involving individuals experiencing mental health crises.

Center for Policing Equity | Data provided by Watertown PD

Officers and 911 dispatchers are not often equipped to recognize or assist people experiencing mental health crises, so analyses of recorded “mental health” calls may underestimate the full extent of opportunities to reduce police involvement in behavioral health incidents. Relying on police to respond to calls related to mental health issues can unnecessarily escalate a mental health emergency to an incident ending in incarceration, use of force or death. Community-based systems of crisis response are preferred for these call types because they help connect people with appropriate care that may prevent future crises.

These 580 recorded events represent some incidents that may be better addressed using appropriate non-police responders or co-response models.

What does this show?

This figure shows the number of times officers recorded being involved in or responding to reports of people who were perceived as experiencing mental health crises. This figure only represents incidents that police or dispatchers coded with labels that clearly indicate a mental health-related event. Therefore, these findings likely underestimate the full extent of police involvement in managing the types of health and social issues that are routinely handled by existing unarmed community-based response programs around the country.

These community-based first response models, such as CAHOOTS (Eugene, Oregon), STAR (Denver, Colorado), BHEARD (New York City), and Portland Street Response, have generated early evidence showing that unarmed crisis services can effectively divert people experiencing mental health emergencies away from arrest and hospitalization, decrease the repeated use of crisis services, and improve the health of people who need emergency mental health care.

How did we calculate this?

To calculate this, we counted the number of events – both Officer-Initiated Activities and Calls for Service – in which the event reason included terms such as “mental health,” “emotionally disturbed,” “suicide,” or “behavioral health.” This figure may undercount the number of behavior health-related calls that officers responded to, as we did not include incidents police or dispatchers coded with labels such as “citizen assist,” “welfare check,” or “disturbance.” Such labels sometimes do – and sometimes do not – indicate a mental health-related event.

Cited work

Data required for this analysis:

PUBLIC ORDER CONCERNS, NUISANCES, AND SUSPICIOUS BEHAVIOR

The number of recorded events involving Public Order Concerns, Nuisances, and Suspicious Behavior per capita was relatively evenly distributed throughout the jurisdiction.

These event types generally include minor violations such as truancy, homelessness, drug use, missing persons, suspicious persons, and noise complaints, many of which would benefit from interventions by trained social workers or mental health professionals rather than police.

For further details, see the More Information drop-down below.

What does this show?

This figure shows where in the jurisdiction events involving Public Order Concerns, Nuisances, and Suspicious Behavior– both Officer-Initiated Activity and community-requested via Calls for Service – are most highly concentrated per capita. Neighborhoods shaded in the darkest colors had the highest concentrations of these events.

Areas shaded in the lightest or second-lightest colors represent Census tracts with relatively low or average concentrations of these events, respectively. Areas shaded in the darker third, fourth and fifth colors have high enough concentrations that they are respectively considered low-level, moderate or high statistical outliers relative to other tracts in the jurisdiction. Not all jurisdictions will include tracts shaded in each of these five colors. For example, a city with no high outliers would not have tracts in the fifth (darkest) color; a city with no moderate outliers would not have tracts in the fourth color, and so on.

How did we calculate this?

To represent neighborhoods, we use Census tracts, which are small geographic areas of approximately 4,000 residents each, defined by the Census Bureau.

We calculated the number of events involving Public Order Concerns, Nuisances, and Suspicious Behavior recorded in each neighborhood per 1,000 residents. We then sort these into 5 categories of neighborhoods so that the 20% of neighborhoods with the fewest events are in the first category, the 20% of neighborhoods with the most events are in the last category, etc. Finally, we map these neighborhoods, with darker colors indicating neighborhoods with more of these events. Both Officer-Initiated Activity and Calls for Service are included in this calculation.

Data required for this analysis:

DENSITY OF OFFICER INITIATED ACTIVITIES RELATIVE TO CALLS FOR SERVICE

This analysis examines which neighborhoods had higher levels of Officer-Initiated Activity relative to the number of Calls for Service.
  • The neighbourhood with the highest ratio had 551 recorded Officer-Initiated Activities for every 1,000 Calls for Service.
  • All other neighbourhoods combined had an average of 446 Officer-Initiated Activities for every 1,000 Calls for Service.

What does this show?

This figure shows where in the jurisdiction there might be a mismatch between how much the community is asking police to do and how much the police are actually doing. The neighborhoods with the highest ratio have more Officer-Initiated Activity – which is not requested by the community – relative to the number of Calls for Service.

Areas shaded in the lightest or second-lightest colors represent Census tracts with relatively low or average ratios of Officer-Initiated Activities to Calls for Service, respectively. Areas shaded in the darker third, fourth and fifth colors have high enough ratios that they are respectively considered low-level, moderate or high statistical outliers relative to other tracts in the jurisdiction. Not all jurisdictions will include tracts shaded in each of these five colors. For example, a city with no high outliers would not have tracts in the fifth (darkest) color; a city with no moderate outliers would not have tracts in the fourth color, and so on.

How did we calculate this?

To represent neighborhoods, we use Census tracts, which are small geographic areas of approximately 4,000 residents each, defined by the Census Bureau.

First, we calculated the number of recorded events that were Officer-Initiated Activities and the number that were police responses to Calls for Service. If the data provided by the police had an indicator of which events were Officer-Initiated, we used that. If the data do not indicate which events were Officer-Initiated, we categorized event type descriptions involving warrants, alarms or non-dispatched calls as Officer-Initiated, and then grouped all other event types as Calls for Service. Non-dispatched calls include requests for assistance from other officers or outside law enforcement agencies, activities that appear to be officer assignments (such as “directed patrol” or “walking the beat”), and other activities police typically have some discretion over (such as “traffic stop,” “subject stop,” “area check,” or “follow up”).

Next, we calculate the number of Officer-Initiated Activities per 1,000 Calls for Service by dividing the number of Officer-Initiated Activities by the number of calls and multiplying by 100. We sort these into 5 categories of neighborhoods so that the 20% of neighborhoods with the lowest ratio of Officer-Initiated Activities to Calls for Service are in the first category, the 20% of neighborhoods with the highest ratio are in the last category, etc. Finally, we map these neighborhoods, with darker colors indicating neighborhoods with higher levels of Officer-Initiated Activity relative to the number of Calls for Service.

Data required for this analysis:

Scroll to Top