Memorandum
City of Lawrence
Police Department
TO: Tom Markus, City Manager
Gregory Burns, Chief of Police
FROM: Captain Adam Heffley
Jayme Wehmeyer, Crime Analyst
CC: Diane Stoddard, Assistant City Manager
Casey Toomay, Assistant City Manager
Brandon McGuire, Assistant to the City Manager
DATE: April 30, 2018
RE: Police Department Patrol Division Staffing Study
Background
The Lawrence Police Department Patrol Division has not undertaken a data driven patrol personnel needs assessment in recent history. Obtaining accurate workload information as well as determining what data to measure have been key factors hindering such a study. Upgrades to the Department’s Computer Aided Dispatch (CAD) and Record Management System (RMS) has made data collection more accurate in recent years. The current theoretical maximum number of officers assigned to the Lawrence Police Department’s Patrol Division is 92. This number includes 9 police recruits that are still in academy and are not available to be deployed as Patrol resources, other officers unavailable for duty, and unfilled positions. This maximum is theoretical and almost never realized. Examining the past four years of data, the average actual number assigned has been 82.9 with a low of 74.5 during the fall of 2016 and a high of 86 in 2014 when the department utilized an “over-hire” process.
In 2012, researchers at Michigan State University received a grant through the Community Oriented Policing Services (COPS) program at the United States Department of Justice to provide a methodology for what data to examine. The Michigan State Study, “A Performance-Based Approach to Police Staffing and Allocation” (Weiss & Wilson, 2012) attempted to answer the question, “how many police officers does a particular agency need?” Researchers considered “per capita”, “minimum staffing”, “authorized level” and “workload-based” patrol officer staffing models. The authors concluded that the “workload-based approach” was the most appropriate in terms of measuring patrol staffing needs. In the study, the definition of “work” to be accomplished was to be determined by individual law enforcement agencies due to the vast differences in department service levels as influenced by community needs and expectations.
One additional factor affecting staffing levels is the structure of the Patrol Division and how resources are deployed. The Lawrence Police Department Patrol Division has been operating with four 10-hour overlapping shifts consisting of two teams each since 2012. The 10-hour shift structure does require additional employees to maintain as compared to the 8-hour structure, but offers several benefits to the organizations that utilize them. There are three purposes for the 10-hour shift deployment strategy at the Lawrence Police Department. First, the daily overlap of officers on duty allows for a higher concentration of officers during times of increased activity. Second, the weekly overlap of teams allows for more opportunity to train special units and skills in several different areas within the department across teams. Scheduling training with groups across the department was very difficult with the previously model. Third, several studies have shown employee wellness and stress reduction, especially in first responders, is increased with additional recovery time which is achieved by a 10-hour shift schedule.
Alternate shift designs could be considered but each have a cost. A 2011 study, The Shift Length Experiment: What we Know About 8-, 10-, and 12-Hour Shifts in Policing (Amendola, Weisburd, Hamilton, Jones, & Slipka, 2011) by the Washington, D.C., Police Foundation in collaboration with the U.S. Department of Justice influenced the department in adopting the 10-hour shift model. In addition to the aforementioned internally identified organizational benefits, advantages presented in the study concerning 10-hour shifts for policing include:
· Significantly reduced overtime as compared to 8-hour shifts.
· Significantly more sleep for officers as compared to 8-hour shifts.
· Increased quality of work life as compared to 8-hour shifts.
· Those on 12-hr shifts worked less overtime than 8-hour shifts, but still more than 10-hour shifts.
· Those on 12-hr shifts demonstrated significantly more sleepiness than those on 8-hour shifts.
· Those working 12-hour shifts were significantly less alert than those on 8-hour shifts.
· For medium to large agencies, there are likely cost savings with 10-hour shifts.
· Increased safety and wellness with 10-hour shifts while maximizing alertness on the job.
The study concluded that “Ten- hour shifts appear to offer some advantages over 8-hour shifts, both individually and organizationally, with no noted disadvantages.” (Amendola, Weisburd, Hamilton, Jones, & Slipka, 2011, p. 5). A review of the research by the Police Foundation found “that 10-hour shifts were optimal in that they were associated with the greatest benefits to the agencies and officers, while offering no apparent limitations in terms of sleepiness or lack of alertness as found among the 12-hour group” (The Police Foundation, 2013).
This memorandum is presented in
two parts. Part A is a presentation of the Michigan State Study utilized “IACP
One-Third” (IACP) model and the researcher developed
“Michigan State Study One-Fourth Model” (MSS). Both models utilize
recommended parameters to arrive at a patrol staffing suggested minimums. As
police departments can differ in structure, training needs, public
expectations, and the real impact of staffing challenges, we conducted an additional
analysis which we believe more accurately represents the Lawrence Police
Department’s patrol staffing need. This is presented in part B of this
memorandum. Each are offered for comparison and discussion purposes.
General Application(s) of the Study
The Michigan State University study identifies citizen generated “calls for service” as the prime measurable in regards to patrol staffing needs. A Call for Service (CFS) is defined as any instance in which a citizen requests an officer respond to a dispatched call until the time task is completed. The researchers applied the CFS data collected from multiple cities using the two different models as a basis for determining what number of officers that city should utilize. In the first model (IACP) used, “follow-up” – which is a by-product of the CFS – is included in the total CFS number. In the second model (MSS), which emphasizes a patrol level investigative expectation, follow-up time spent on investigations of reported crimes is made a separate measurable. Follow-up accounts for any time an officer would spend investigating or working on a “case” after the officer had gone back in service or cleared to take another call from the CFS.
1. The International Association of Chiefs of Police (IACP) model divides the officer’s time into thirds:
· 33% Calls for Service (follow-up included)
· 33% Proactive (patrol) Time
· 33% Administrative Activity
2. The authors of the Michigan State study created a model that set performance measures based on dividing the officers’ time into quarters:
· 25% Calls for Service (follow-up excluded)
· 25% Patrol and Self-Initiated
· 25% Administrative
· 25% Criminal Investigation Follow-up
The Michigan State study model most closely aligns with the current practice at the Lawrence Police Department which is heavily dependent on patrol officers conducting follow-up investigations. This is due to the Department’s culture of viewing all officers as investigators to better serve citizens, and is a consequence of an Investigative Division that is usually at or above capacity in terms of the number of cases that can be assigned to it.
Although, the department is most like the Michigan State quarters model, the IACP model will also be employed to provide contrast as presented in the study.
Data Collection
In August 2014, the Lawrence Police Department implemented a new Computer Aided Dispatch (CAD) system as well as a new and Records Management System (RMS). These new systems are better able to produce accurate data in regards to Calls for Service (CFS) and officer time in relation to the CFSs.
When the new systems were brought on-line in August 2014, they replaced older systems which did not lend themselves well for this specific type of data collection. The 2014 data exists in two different systems which makes any 2014 data examination problematic. Generally good data is available beginning in 2015 and the Department’s Crime Analysis Unit (CAU) was able to produce complete data for the entire 2015, 2016, and 2017 years. The CAD system tracks each officer’s time from the initial “dispatch” to the time the officer went back “in service”. The system will also track how the incident was generated, such as a citizen CFS or an officer initiated incident. Additionally, if multiple officers were involved in a call, each officer’s time is tracked and can be calculated.
Measurable data for the Lawrence Police Department is:
1. Calls for Service (CFS), IACP One-Third Model. Using the best statistical information from the new CAD and RMS the following table represents the Patrol Division’s work in hours on citizen generated Calls for Service (CFS) in 2015, 2016, and 2017. This time also represents follow-up that resulted from these calls.
|
Year |
Calls for Service Total in Hours |
|
2015 |
42,236 |
|
2016 |
36,006 |
|
2017 |
36,638 |
2. Calls for Service (CFS), Michigan State One-fourth Model. Using the best statistical information from the new CAD and RMS; the following table represents the Patrol Division’s work in hours on citizen generated Calls for Service (CFS) in 2015, 2016, and 2017 alone. These totals do not count time spent later on follow-up work.
|
Year |
Calls for Service Total in Hours |
|
2015 |
35,794 |
|
2016 |
33,498 |
|
2017 |
34,624 |
Part A: Michigan State Study Offers Two Variations
Both models presented in the study identify a component that must be accounted for which they have termed “Shift Relief Factor” (SRF). If no consideration was given for compensated time off (vacation, sick time, or regular days off), one officer would be able to work one shift a day for 365 days. The Patrol Division currently utilizes a 10-hour shift model to provide for overlapping coverage during peak CFS times, to facilitate ongoing officer training requirements, to be able to address special initiatives and events, and to positively influence officer health and welfare. The study researchers compensate for this time away from patrol by calculating the Shift Relief Factor. The study advocates for a calculation of this factor utilizing several of the regularly defined types of compensated time off and the constant SRF is then applied to both models.
Using the study’s definitions of time off which includes use of all accrued benefit time, the average compensated time off and average training time for a Lawrence police officer is listed below in hours:
|
Personal Time |
24 |
|
Vacation Time |
157 |
|
Holiday Time |
72 |
|
Sick Leave |
96 |
|
Training |
137 |
|
Regular Days off (3) |
1560 |
|
Total hours off per year |
2046 |
Using the example equation provided by the researchers, the Shift Relief Factor would be calculated as follows:
365 x shift length / (365 x shift length – time off)
or
365 x 10 / (3650 – 2046)
3650 / 1604 = 2.28 SRF
The amount of time officers spend on Calls for Service listed previously 42,236 (2015), 36,005 (2016), and 36,638 (2017) with follow-up or Calls for Service 35,794 (2015), 33,498 (2016), and 34,624 (2017) without follow-up identified, and the SRF (2.28) calculated, the specific model (IACP one-third or MSS one-fourth) can be applied to arrive at minimum recommend patrol staffing level:
1. International Association of Chiefs of Police One-Third Model. The IACP model is applied below to determine the total amount of “work needed to be done”:
Year 2015
42,236 (CFS hours) = 1/3 X
X = 126,708 hours of work
Applying the study’s formula:
(Hours of “work” to be done) / (potential shifts worked) = (Officers needed)
126,708 / 3650 = Y
Y = 34.7 (officers)
The equation identifies the minimum number of officers that would be required to meet the CFS performance goal. This figure assumes the officers work a 10-hour shift every day of the year, which is obviously not sustainable. Applying the constant SRF (2.28) to the one-third IACP model produces the following results:
(Hours of work) / (Shift hours worked by officer)
126,708 / 3650 = Y
Y = 34.7 (officers)
(Officers required to cover work) x (Shift Relief Factor) = Total Minimum Recommended
34.7 x 2.28 = 79.1 (79 officers)
The resulting recommendation for 2015 is for a minimum of 79 officers assigned to the Patrol Division to handle the CFS and meet minimum performance goals using the one-third IACP model. The table below illustrates the totals when the formula is applied to years 2015, 2016, and 2017.
|
Year |
“Y” Value |
Total Recommended |
|
2015 |
34.7 |
79 Officers |
|
2016 |
29.6 |
67 Officers |
|
2017 |
30.1 |
69 Officers |
*Calculations are attached
The resulting 3 year average recommendation from the IACP model is a minimum of 72 officers assigned to the Patrol Division.
2. The Michigan State Study Model (Delaware, Ohio) One-Fourth Model. The Michigan State model is applied below to determine the total amount of “work needed to be done”. Applying the model corresponding Call for Service (CFS) number and Shift Relief Factor (SRF) to the second model produces the following data:
Year 2015
35,794 (CFS hours) = 1/4 X
X = 143,176 hours of work
143,176 / 3650 = Y
Y = 39.23 (officers)
(Officers required to cover work) x (Shift Relief Factor) = Total Minimum Recommended
39.23 x 2.28 = 89.44 (89 officers)
The resulting recommendation for 2015 is for a minimum of 89 officers assigned to the Patrol Division to handle the CFS and meet minimum performance goals. The table below illustrates the totals when the 1/4 formula is applied to years 2015, 2016, and 2017.
|
Year |
“Y” Value |
Total Recommended |
|
2015 |
39.23 |
89 Officers |
|
2016 |
36.71 |
84 Officers |
|
2017 |
37.94 |
87 Officers |
*Calculations are attached
The resulting recommendation based on the 3 year average for the Patrol Division would be 87 officers per the Michigan State Model.
Discussion
Both models identify the 2015 suggested minimum staffing levels for the Patrol Division to be somewhere between 79 (IACP One-Third model) officers to 89 officers (Michigan State One-Fourth model). The three year averages for suggested minimum staffing levels for the Patrol Division is between 72 (IACP One-Third model) officers to 87 officers (Michigan State One-Fourth model). As stated before, the second (MSS) model was developed for the Delaware, Ohio Police Department which tasks patrol officers with a large amount of follow-up investigations which is similar to the Lawrence Police Department.
At 92 officers, or 83 if the recruits are excluded, the current theoretical maximum number of patrol officers assigned to the Lawrence Police Department’s Patrol Division has not supported maintaining the recommended minimum patrol staffing of the model that most closely resembles the function of the Lawrence Police Department. As previously mentioned, the theoretical maximum is almost never realized and the average has been 82.9 for the last four years. The issue of constantly being below the theoretical maximum is affected by two factors. First, the amount of time it takes to recruit, hire and train a new police officer to replace one that has left the department and second, the allocation of sworn Police Officers to tasks other than the Patrol Division. The hiring and training process issue is not unique to the Lawrence Police Department. There is a large investment of time and resources in hiring and training an officer. Once authorization is received to hire an officer, it is commonly over a year before they are able to independently respond to a call for service.
Currently, the Chief of Police is working to civilianize several positions in the department. This is a long term project and will need to be weighed with needs such as Crime Scene Investigators, officers involved in community engagement, the new Mental Health unit, and the impending implementation of a Body-worn camera system. To achieve the identified minimum staffing levels on Patrol the department should increase the authorized sworn staffing level by 5 officers to meet the minimum recommended staffing levels (87 Officers) of the model that closely resembles the community expectations for Patrol.
Part B: Lawrence Police Department
Scheduling Data and Practice
To address what staff feels is a more accurate representation of the actual Shift Relief Factor (SRF) that may be used, 2015, 2016, and 2017 department schedules were examined. In doing so, Lawrence Police Department scheduling, training, and community expectations could be incorporated into the SRF for the department. This information is present in the following section.
The study’s example for calculating the Shift Relief Factor (SRF) does not take into account several additional factors that affect the Patrol Division for the Lawrence Police Department. To accurately demonstrate the SRF for the Department, the process should adjust the SRF variable to capture other “time away from Patrol” that continually occurs. Since the Department is not heavily staffed in the support functions, it relies on the expertise of officers, many of whom are assigned to Patrol, to perform various special assignments, take part in initiatives, and provide training and instruction for inside and outside of the department. For example, during recruit training, several patrol officers are tasked with training recruits in specific job functions such as car stops, building searches, crisis intervention, and defensive tactics. An impending shortage in crime scene investigators is currently causing a personnel draw from Patrol so that officers can learn and receive the required experience to develop that skill. These are just a few examples of how Patrol personnel are utilized to meet the specialized needs of the community and maintain a high level capability and service.
The Department could choose to permanently assign these trainers to the training division but this would be a less efficient use of the remainder of the officer’s time. Additionally, a loss of experienced and technically proficient officers from patrol would hinder many investigations and learning opportunities that occur daily on Patrol. Another factor that impacts staffing levels is injury. During the course of a year it is very common to lose several officers to on-duty injuries to hands, fingers, and knees. A minor on-duty injury can commonly remove an officer from the shift for two to five business days and a moderate injury could reduce an officer’s availability to staff a patrol position for a month or more. This loss of time occurs in other departments as well but was not specifically mentioned during the study. Examining actual data, we were also able to account for officer attrition which has been consistently 8-12 officers a year before a spike in 2016 due to many retirements of a large group of employee cohorts.
Looking at historical data, the realized SRF for the Lawrence Police Department Patrol Division can be calculated using the 2015 Patrol assignments and schedules. The collected data shows the number of officers initially assigned to each shift and the average number of officers that were on-duty. The total averages on each shift in 2015 are:
|
Shift |
Average Officers Assigned |
Average Working |
Shift Relief Factor |
|
Shift One |
25.57 |
10.04 |
2.55 |
|
Shift Two |
16.50 |
6.52 |
2.54 |
|
Shift Three |
20.43 |
8.77 |
2.33 |
|
Shift Four |
25.57 |
9.47 |
2.71 |
|
Total Avg. |
|
|
2.53 |
*calculations are attached
In this case, the SRF is derived from actual 2015 Department personnel deployed to Patrol. It is generally reflective of what is normal on an annual basis and will more accurately demonstrate the relief factor that effects patrol staffing. Subsequently, the SRF experienced by Patrol was calculated for 2016 (2.52) and 2017 (2.36). The three year average SRF experienced by the Lawrence Police Department Patrol division was 2.47.
Applying this derived SRF to the two staffing models yields:
1. International Association of Chiefs of Police One-Third Model. Applying the 2015 Lawrence Police Department calculated Shift Relief Factor (SRF) to the one-third IACP model produces the following results:
(Hours of work) / (Shift hours worked by officer)
126,708 / 3650 = Y
Y = 34.7 officers
(Officers required to cover work) x (Shift Relief Factor) = Total Minimum Recommended
34.7 x 2.47 = 85.7 (86 officers)
The resulting recommendation for 2015 is for 86 officers assigned to the Patrol Division to handle the CFS and meet minimum performance goals.
|
Year |
“Y” Value |
Total Recommended |
|
2015 |
34.7 |
86 Officers |
|
2016 |
29.6 |
73 Officers |
|
2017 |
30.1 |
74 Officers |
|
Average |
|
78 Officers |
*calculations are attached
The resulting recommendation based on the three year average would be for a minimum of 78 officers assigned to the Patrol Division for this model.
2. The Michigan State Study Model (Delaware, Ohio) One-Fourth Model. Applying the 2015 Lawrence Police Department calculated Shift Relief Factor (SRF) to the one-fourth Michigan State Study model produces the following results:
35,794 = 1/4 X
X = 143,176 hours of work
143,176 / 3650 = Y
Y = 39.23 officers (officers)
(Officers required to cover work) x (Shift Relief Factor) = Total Minimum Recommended
39.23 x 2.47 = 96.89 (97 officers)
The resulting recommendation is for 97 officers assigned to the Patrol Division to handle the CFS and meet minimum performance goals.
|
Year |
“Y” Value |
Total Recommended |
|
2015 |
39.2 |
97 Officers |
|
2016 |
36.7 |
91 Officers |
|
2017 |
37.94 |
94 Officers |
|
Average |
|
94 Officers |
The resulting recommendation based on the three year average would be for a minimum of 94 officers assigned to the Patrol Division for this model.
Discussion
After applying the 3 year averaged Lawrence Police Department specific Shift Relief Factor (SRF) of 2.47 to both models, the resulting suggested staffing levels for the Patrol Division fall between 78 officers (IACP one-third model) and 94 officers (Michigan State one-fourth model). Since actual schedules and their resulting manpower levels were used to calculate this SRF, staff believes it is a more accurate representation of the actual need. As seen in the study’s calculations the department Patrol staffing that is realized (82.9) falls below the minimum suggested levels (94) in the model that most closely represents the department’s form and function.
References
Amendola, K. L., Weisburd, D., Hamilton, E. E., Jones, G., & Slipka, M. (2011). The Shift Length Experiment: What We Know about 8-, 10-, and 12-Hour Shifts in Policing. Washington, DC: Police Foundation.
The Police Foundation. (2013). Police Foundation: Advancing Policing Through Innovation and Science. Retrieved from policefoundation.org: https://www.policefoundation.org/projects/the-shift-length-experiment/
Weiss, W. &., & Wilson, J. M. (2012). A Performance-Based Approach to Police Staffing and Allocation. Officer of Community Oriented Policing Services, U.S. Department of Justice.
Part A: Formula Application
Year 2016 1/3 Model
36,006 (CFS hours) = 1/3 X
X = 108,018 hours of work
Applying the Michigan Study formula:
(Hours of “work” to be done) / (potential shifts worked) = (Officers needed)
108,018 / 3650 = Y
Y = 29.6 (officers)
(Officers required to cover work) x (Shift Relief Factor) = Total Minimum Recommended
29.6 x 2.28 = 67.49 (67 officers)
Year 2017 1/3 Model
36,638 (CFS hours) = 1/3 X
X = 109,914 hours of work
Applying the Michigan Study formula:
(Hours of “work” to be done) / (potential shifts worked) = (Officers needed)
109,914 / 3650 = Y
Y = 30.1 (officers)
(Officers required to cover work) x (Shift Relief Factor) = Total Minimum Recommended
30.1 x 2.28 = 68.63 (69 officers)
Year 2016 1/4 Model
33,498 (CFS hours) = 1/4 X
X = 133,992 hours of work
133,992 / 3650 = Y
Y = 36.71 officers (37 officers)
(Officers required to cover work) x (Shift Relief Factor) = Total Minimum Recommended
36.71 x 2.28 = 83.69 (84 officers)
Year 2017 1/4 Model
34,624 (CFS hours) = 1/4 X
X = 138,496 hours of work
138,496 / 3650 = Y
Y = 37.94 (officers)
(Officers required to cover work) x (Shift Relief Factor) = Total Minimum Recommended
37.94 x 2.28 = 86.50 (87 officers)
Part B: Formula Application with LPD Calculated Shift Relief Factor
(SRF was calculated for each shift and averaged over 2015, 2016, and 2017)
|
|
2015 |
2016 |
2017 |
|
Shift 1 |
2.55 |
2.73 |
2.57 |
|
Shift 2 |
2.54 |
2.58 |
2.31 |
|
Shift 3 |
2.33 |
2.24 |
2.22 |
|
Shift 4 |
2.71 |
2.51 |
2.35 |
|
Average |
2.5325 |
2.515 |
2.3625 |
|
3 Yr Avg |
|
|
2.47 |
Year 2016 1/3 Model
36,006 (CFS hours) = 1/3 X
X = 108,018 hours of work
Applying the Michigan Study formula:
(Hours of “work” to be done) / (potential shifts worked) = (Officers needed)
108,018 / 3650 = Y
Y = 29.6 (officers)
(Officers required to cover work) x (3 year average Shift Relief Factor) = Total Minimum Recommended
29.6 x 2.47 = 73.11 (73 officers)
Year 2017 1/3 Model
36,638 (CFS hours) = 1/3 X
X = 109,914 hours of work
Applying the Michigan Study formula:
(Hours of “work” to be done) / (potential shifts worked) = (Officers needed)
109,914 / 3650 = Y
Y = 30.1 (officers)
(Officers required to cover work) x (3 year average Shift Relief Factor) = Total Minimum Recommended
30.1 x 2.47 = 74.34 (74 officers)
Year 2016 1/4 Model
33,498 (CFS hours) = 1/4 X
X = 133,992 hours of work
133,992 / 3650 = Y
Y = 36.71 officers (officers)
(Officers required to cover work) x (Shift Relief Factor) = Total Minimum Recommended
36.71 x 2.47 = 90.67 (91 officers)
Year 2017 1/4 Model
34,624 (CFS hours) = 1/4 X
X = 138,496 hours of work
138,496 / 3650 = Y
Y = 37.94 officers (officers)
(Officers required to cover work) x (Shift Relief Factor) = Total Minimum Recommended
37.94 x 2.47 = 93.71 (94 officers)
|
Patrol Officer Time on Calls for Service (1/4 Model) |
|
|
|
(Calculated from when the Patrol Officer/ Teleserve Officer is Dispatched to Complete) |
|
|
|
|
CFS Time |
Average Time |
|
2015 |
35794:14:41 |
0:28:44 |
|
2016 |
33498:14:12 |
0:27:20 |
|
2017 |
34624:02:04 |
0:26:36 |
|
Patrol Officer Time on Calls for Service (1/3 Model; Follow-up included) |
|
|
|
(Calculated from when the Patrol Officer/ Teleserve Officer is Dispatched to Completed) |
|
|
|
|
CFS Time |
Average Time |
|
2015 |
42236:07:21 |
0:30:00 |
|
2016 |
36005:50:55 |
0:28:03 |
|
2017 |
36638:18:16 |
0:26:57 |
