Webinar presenter Peter Bellmio answered a number of your questions after his presentation, "Measuring Patrol Workload." Here are a few of his responses.
Audience Question: Before everybody joins us again in December, is there homework they should do? Is there anything they should prepare something to get the most from the next webinar?
Peter Bellmio: One of the things you can do is to try to get a month of unit records for patrol that has everything they logged in CAD. That information is usually in your unit history file in the CAD system. You can start to look at and look for outliers. It is usually easier to find that kind of data and just look for somebody that can do the programming and analysis work to generate the line graphs on time in the presentation Another thing they could do is look at the wording of the definitions for call priorities and do they really have a system that’s going to track true emergency and priority ones and intermediate and the threes. I think that would be a good thing to do. Start to get a picture of where you are. Those two things, starting off with some data would really be a good start and also are the calls that patrol is dispatched to and do you really think we shouldn't be going out them? Look at your priorities when sending units out to tell somebody something we could tell them over the phone in communications. Look if there are opportunities for workload reduction.
Audience Question: How much data do you really need to do this analysis? Is a month sufficient, are we talking more to a year, five years, how much data do we really need?
Peter Bellmio: Well, pretty much a year. The problem with a month is you have four Tuesdays. If you want to see anything by day of the week and you have one Tuesday that has a thunderstorm, or you have a parade or some event, like July 4th, it will mess up the data. I'm trying to say a year of pretty good workload data will give you seasonality. But then you have to calculate the glide slope on call rates. I'm working on one community right now in the Hampton Roads area of Virginia that has 3 percent rate of call growth which is significant. We're going to have a snapshot of time data for the past year. Then we have to correct the call rates moving forward. At least a year. I think for cleaning up your data you can work with a couple of months to get started. You’re going to see the same data quality problems whether it’s a month or two or twelve months. The smaller data set is just easier to analyze.
Audience Question: How do you handle the missing times from CAD like the time of unit arrival? How do you handle the data when there is missing data?
Peter Bellmio: I just don't count those averages in the data and hope that you don't have so many that it causes the representativeness of the data change. That's the key. Never throw out a call for service from the public. You never throw out that we sent x number of patrol units to a call. The averages for the times come from the time data that passes your business rules for data quality.
Audience Question: For budgeting purposes, when asking for additional personnel, is it still safe to use the old standard, officer’s time split three ways, admin, patrol, non-patrol, 33% each. It makes for easy math but really how accurate is this model?
Peter Bellmio: It's really just the old Greek triad. I remember when I started out as a consultant in Public Administrations services in 1980, I looked into the literature, called everybody I could at PERF and IACP but there's nothing. It is really is about what is the strategic direction for a police agency and how much time do so the work we tell the public what we have to do to reach those goals such as community-based policing.. You may need time to do Directed Patrols (DPs) to deal with crime problems to include gangs. DPs usually also should include some public contact. If you can show that you’re going to add value to the safety of the community based on your strategic goals you need to be able to describe those results. This is like asking a community if you want to get the street swept once a year, once a month, or once a year? What level of service do you have? That really drives the pie chart and staffing costs.
Audience Question: You mentioned a training at the beginning of the webinar, can you say what that was again?
Peter Bellmio: This is a five-day basic police planner course. You can get the outline on IALEP website. We spent half a day just on patrol staffing analysis because it is the biggest part of the police budget. The course also covers strategic planning starting from general to the specific. We deliver sessions on Crime Prevention Through Environmental Design, applied research methods, IT planning, RFPs, and surveys methods. The course is one of the requirements for ALEP’s basic police planner certification. There's also an advanced certification. IALEP certification shows your agency knows that you have the knowledge, skills, and abilities that IALEP, as an association, believes are essential. The courses are scheduled in the spring and fall each year and move around North America at locations of host agencies.
Audience Question: Have you seen situations where these reports that you described, that these reports have been used against the agency by a police union?
Peter Bellmio: No. I really haven't. One of the issues we've had raised by Unions is the level of specialization that takes police officers away from patrol duty. It’s been the issue at Baltimore, and it still is, The BPD had 1900 police officers when I started working with them in 2012 but they had only 800 of them in patrol staffing radio cars. The union actually used the data we generated to say to the city we need more people on patrol. I guess we use this as part of public policy debate. In Hartford, Connecticut, the Mayor wanted to hire extra people but the union said we don't need them. All you're going to do is reduce our benefits because you're hiring police because you think we need them and we don’t think it’s necessary. I completed a staffing study and the proposal died when it showed the officers really were not needed. Unions can use the data for the public policy issues and it helps if we all use the same numbers.
Audience Question: Do you have any suggestions on how to convince senior management to sustain investment in monitoring workloads?
Peter Bellmio: That is the $64,000 question. If they are intimidated by it, they fall back on their crime fighter role they know well. Some police managers just want to make all their troops happy. The data is going to raise issues like the 12-hour shift doesn’t match staffing to workload and you have a chief that’s interested in making your troops happy, you're going to have a problem. I think you can make your troops happy and have a good schedule. It’s really a question of leadership. At the LAPD, we went through the staffing analysis process and eventually the Union and the City agreed that 12-hour shifts alone don’t fit staffing to workload. They implemented a schedule with, 12-hour shifts, 10-hour shifts and in some cases some 8-hour shifts as well based on employee choice. It has been in place since the mid-1990’s and it’s good for management, good for crime fighting, good for the union, good for the staff. It took leadership in the LAPD and the city to make that happen.
Audience Question: How would you recommend we track time for serving orders of protection or other civil documents?
Peter Bellmio: If patrol officers doing that work, you would have to create a code for that and use it either create a call type or out of service. Creating a call and coding it as proactive, not citizen-generated, would give you an address and x-y coordinate for where the work was completed. If you just use it as a change of unit status, the data you get is the time from when the unit logged out on that unit status code and then came back in service. Call records are easier to analyze. It is. It is very difficult to use data from a unit history file to calculate the interval of time from when units log-on log-off an out of service code. It can be done, however. If the work is being done in an agency with a large patrol area, you need to decide how to count the travel time for getting to the location for serving those orders. Service time is travel plus time on the scene. That’s probably why creating a call type for that work makes the most sense. To be sure it is not counted as a citizen-generated call, you can use a priority level that is for this work. Some CAD systems can have up to nine priorities available.
Audience Question: If an agency does not have codes, are there certain ones you absolutely start with?
Peter Bellmio: I have a model set that evolved from work with a couple of agencies. I can send them out if anybody's interested. My email is [email protected] You can use that example to group activities and work in code ranges that make some sense for your agency.
Audience Question: If an agency isn’t using their codes, how do you train to get the staff to use them regularly and correctly?
Peter Bellmio: I think it goes back to having that officer activity report and having the sergeant’s use it for evaluations. You would think that a sergeant may not want to deal with that but many of them like the CAD data because they don’t get accused of favoritism. You are what your records say you are. How you look in CAD is something that you should be able to account for as an officer. They say I did more, I didn’t put it in the CAD. It may sound like George Orwell in 1984 but if it's not in the machine, it did not happen. The machine is your friend, log your time in CAD. You can have some excessive compliance on that too. Those people add so much activity their data jump outs immediately as abnormal. How can they have possibly done that? It’s really getting that ownership CAD data as a tool for supervisors and officers.
Audience Question: Is there an open sharing data project among agencies where people can compare data across agencies, types of agency, size of agency, locale, that type of thing?
Peter Bellmio: None that I know of. The Ontario government in Canada has passed a New Police Act which was going to include more of this operational data in call for service and how they compare their response time and calculate all that stuff (indiscernible 1:06:11). I don't think it's out there. The fundamental problem agencies to not have standard definitions for creating calls for service and account for time differently. That makes it hard to compare data from agency to agency.
Audience Question: Which of these reports do you recommend that we share to the public on our website? How often should we update these publicly available reports?
Peter Bellmio: What we want to do with the reports is to translate them into more of the graphics. The example of the graphics that you saw. You can probably come up with a graph on the emergency response line, the pie chart, percentage of time, units answering, peak calls and use that. I think the graphics speak volumes for analysis and data.
Audience Question: Can the same type of analysis be done for probation, for courts, jails, etc.?
Peter Bellmio: Yes. The question is how you log your time. I‘ve done some corrections work. In fact, we’ve done a lot in the entire Ontario Corrections system. I could refer you to Sean Tout at the Ministry of Community Safety and Corrections. It’s really how you collect time and how you document it and make it in the employee's interest to book their time properly and have a quality patrol system. Self-report, the activity sheets we used to use in policing were not very accurate. Instead, it’s got to be something the people do when they go to work. In a study I worked on for Los Angeles Transit, we recommended a phone app that would allow staff to document their activity on their phone because they are on foot in the transit system and could not use CAD.
Click Here to Watch a Recording of "Measuring Patrol Workload."