The URL can be used to link to this page

Home My WebLink AboutBozeman Police Department Staffing StudyBozeman Police Department Patrol Stafﬁng and Deployment Study The city of Bozeman is located in Gallatin County Montana, about 290 miles from Boise, ID and 330 miles from Spokane, WA. Surrounded by four Rocky Mountain ranges, the city is considered the hub of southwest- ern Montana’s Gallatin River Valley. Skiing, snowboarding, kayaking, raft- ing, ﬂy-ﬁshing, biking, rock-climbing, and more awaits the adventurous soul just outside the city’s borders. The city had an estimated population in July 2005 of 33,535 people. Bozeman’s population has grown by 21.9% since July of 2000 and is continuing to expand.1 Assuming continuos growth since July of 2005, the population of Bozeman could be estimated at 35,945 at the time of this report. In addition to this population, there is an estimated daytime population change due to commuting of +7,453.1 The city encompasses 18.3 square miles with a population density of 1,964 people per square mile. Bozeman is home to Montana State University, a Carnegie Foundation top-tier research uni- versity. With roughly 13,000 students, MSU is considered a mid-sized public university with 826 instructional faculty.2 Educational services is the most common industry in Bozeman for both males and females with overall percentages of 16% and 19% respectfully. Bozeman’s racial composition is predominately white non-hispanic (93.8%) with small popu- lations of American Indian, Asian, Hispanic, and mixed races.1 The median resident age for Bozeman is 25.4 years old compared to the median resident age of 37.5 years old for the state. The university may have a large impact on this statistic. Based on the most recent census data, 53% of housing in Bozeman is renter-occupied. This could also be due to the presence of the University. Higher rates of renter-occupied housing could indicate a more transient population and a more ﬂuid demographic composite of the city. Between 2000 and 2001, the average violent crime for the city was below the state average. Murder, robberies, and aggravated assaults in 2004 were lower while sexual assaults were above the average. The average for property crimes during that same time period was above the state average. While motor vehicle thefts and burglaries were lower, they were overshadowed by a large number of larceny/theft cases. A citizen survey was conducted in 2005 to estimate the community’s perception of public safety. Respondents indicated they felt safe in their neighborhoods, in the downtown area, and in the city parks. The average respondent felt safe from violent crime, property crimes, and ﬁre. However, satisfaction of police service was at 58% and trafﬁc enforcement was only 44%. In September of 2006, Etico Solutions, Inc., was requested to assist the Bozeman PD in de- ter mining their optimal staff size for their patrol function. Etico worked in conjunction with Bozeman PD staff to develop and implement a stafﬁng and deployment model for the agency. This report now summarizes the methodology used and the results that were obtained. BOZEMAN, MONTANA POLICE DEPARTMENT - PATROL STAFFING AND DEPLOYMENT STUDY SPRING 2007 COPYRIGHT 2007 - ETICO SOLUTIONS INC. PAGE 1 OF 29 www.eticosolutions.com “A D E Q U AT E P O L I C E P R OT E C T I O N , L I K E B E AU T Y, L I E S I N T H E E Y E O F T H E B E- H O L D E R . T H E O P T I M A L O R A P P R O P R I AT E R AT I O O F O F F I C E R S TO P O P U- L AT I O N , T R A F F I C VO L- U M E S , R E P O RT E D C R I M E S O R AC C I- D E N T S , E T C ., I S N OT A M AT T E R O F M AT H E- M AT I C S O R S TAT I S T I C S . I T I S A M AT T E R O F H U M A N J U D G E M E N T A N D C O M M U N I T Y R E- S O U R C E S .” -J O H N S C H U I T E M A N , T H E P O L I C E C H I E F, J U LY 1 9 8 5 ..............................................I. Summary of Patrol Stafﬁng Methodologies 4 .......................................................................................................................................................................Ofﬁcer to Population Ratios 4 .................................................................................................................................................................................................Benchmarking 4 .....................................................................................................................................................................Empirical Qualitative Analysis 5 ...................................................................................................................................................................................Chosen Methodology 5 .............................................................................II. Police Resource Analysis 6 .....................................................................................................................................................Total Reactive Time 6 ..........................................................................................................................................................................................Reactive Activities 6 ........................................................................................................................................................................................Proactive Activities 6 .................................................................................................................................................................Average Call-For-Ser vice Times 6 ............................................................................................The Need for Proactive Time In Patrol Operations 7 ................................................................................................................................................................................Cross Beat Dispatching 8 .................................................................................................................................................................................................Patrol Interval 8 ..............................................................................................................................................................................Probability of Saturation 9 .....................................................................................................................................................Performance Factor 9 .......................................................................Performance Variables as a function of the selected MR value 10 .......................................................................................................................................................................Unit Mix 11 ......................................................................................................................................................Shift Relief Factor 11 ..................................................................................................................................................................Regularly Scheduled Days Off 11 ..........................................................................................................................................................................................Beneﬁt Days Off 11 ..........................................................................................................................................................................................Non-Patrol Days 11 .................................................................III. Patrol Resource Deployment 12 ...........................................................................................................Characteristics of Police Work Schedules 12 ..........................................................................................................................................................................Fixed Days Off Schedules 12 .......................................................................................................................................................................Locked Rotating Schedules 12 ...................................................................................................................................................................Unlocked Rotating Schedules 13 ............................................................................................................................................................................Unstructured Schedules 13 ...............................................................................................................................................................................................Team Integrity 13 .....................................................................................................................................................................................Unity of Command 13 ............................................................................................................................................................................................Schedule Equity 14 BOZEMAN, MONTANA POLICE DEPARTMENT - PATROL STAFFING AND DEPLOYMENT STUDY SPRING 2007 COPYRIGHT 2007 - ETICO SOLUTIONS INC. PAGE 2 OF 29 www.eticosolutions.com .........................................................................................................Evaluation of The Current Patrol Schedule 14 ..................................................................................................Recommendations for Schedule Modiﬁcations 17 ...........................................................................................................................................................Modiﬁcation of Existing Schedule 17 ............................................................................................................................................................Recommended Schedule Change 18 ............................................................................................................................................................................Unstructured Schedules 19 ..............................................................................................................................................................................................12-Hour Shifts 19 ................................................................................................IV. Beat Design 21 ...............................................................................................................................................................Methodology 21 .........................................................................................................................................................Three-Beat Plan 22 ............................................................................................................................................................Four-Beat plan 23 ...........................................V. Final Recommendations and Observations 24 ......................................................................................................................................................Resource Analysis 24 ..............................................................................................................................................Resource Deployment 24 .................................................................................................................................................................Beat Design 24 ............................................................................................................................................................Future Studies 24 .............................................................................................VI. Addendum A 25 ................................................................................................VII. References 29 BOZEMAN, MONTANA POLICE DEPARTMENT - PATROL STAFFING AND DEPLOYMENT STUDY SPRING 2007 COPYRIGHT 2007 - ETICO SOLUTIONS INC. PAGE 3 OF 29 www.eticosolutions.com I. Summary of Patrol Stafﬁng Methodologies The law enforcement profession presents a unique challenge to those re- sponsible for stafﬁng and scheduling their patrol staff. Not only must they schedule a 24 hour operation every day of the year, they must also attempt to staff pro- portionally to a workload that is con- stantly changing. Patrol workload can be best de- scribed as “non-uniform but predictable”. Calls-for-service do not reach the dis- patch center unifor mly; one at a time and at consistent intervals. Further more, the time required to handle a call-for-service can vary greatly depending on the nature of the call. However, police agencies can usually predict when their most active times will be and when they will experi- ence their lowest call levels. Call-for-service loads are very im- portant but they are not the only consid- erations that must be made when deter- mining staff sizes and scheduling ofﬁcers. Minimum stafﬁng levels must sometimes be considered when an agency has a large geographic area to cover or a small num- ber of ofﬁcers available for patrol. Even at times when call volume is expected to be low, agencies may need to staff addi- tional ofﬁcers to ensure their ability to answer multiple two-ofﬁcer calls safely and promptly. Patrol divisions operate in an unsta- ble environment, the real world. When the volume of work begins to exceed the available number of ofﬁcers, the agency cannot close its doors or stop answering the phone. They must prioritize the calls, respond without back-up, or hold the calls until a unit becomes available. The patrol division functions in an environment that is void of walls, roofs, or fences. Their working conditions and workload is effected by weather, national events, political activities, natural disas- ters, demographic shifts, and numerous other environmental, economic, and social factors that affect their community. Ofﬁcer to Population Ratios Optimal patrol stafﬁng has been deter mined in a number of ways over the last several decades. One of the more popular gauges of adequate stafﬁng is the ofﬁcer to population ratios published each year by the FBI in their report enti- tled “Crime in the United States”. This report provides a table displaying self- reported statistics on the number of sworn ofﬁcers per 1000 population. The ratios provided in the table are based on two criteria, a population range, and a general location within the United States. Such ratios are not particular to individual police agencies since they do not take local criteria into account. The chart does not consider local demograph- ics, socioeconomic status, crime rates, geographic size, or a host of other impor- tant considerations. It should be noted that the authors of the “Crime in the United States” re- port speciﬁcally state that the statistics provided are not to be used as stafﬁng guidelines. Their report states: “Because of law enforcement's differing service requirements and functions as well as the varied demographic traits and character- istics of jurisdictions, use caution when drawing comparisons between agency stafﬁng levels based upon police employ- ment data from the Uniform Crime Re- porting (UCR) Program. The data merely reﬂect existing stafﬁng levels and are not preferred ofﬁcer strengths rec- ommended by the FBI. In addition, it must be remembered that the totals given for sworn ofﬁcers for any particular agency reﬂect not only the patrol ofﬁcers on the street but also ofﬁcers assigned to various other duties such as those in ad- ministrative and investigative positions as well as those assigned to special teams.”3 Benchmarking A second method that is often used is a comparative analysis based on a number of “similar” agencies. This is often referred to as a form of “bench- marking”. This process is also fraught with inherent assumptions and limita- tions. The ﬁrst obvious assumption in this process is that the comparable agencies are in fact, “similar”. Agencies must be found that share similar populations, agency sizes, and geographic locations. This is just the tip of the iceberg on nec- essary similarities if a true benchmark comparison is to be made. Other consid- erations such as demographics, socioeco- nomic status, geographic size, crime rate, and population density must also be con- sidered. The list of comparable charac- teristics could be endless as an agency seeks to ﬁnd their ultimate comparable agencies. The second assumption that must be made of comparable cities is that they are operating under the same philosophy and mission of patrol as the agency under study. Some communities applaud an agency that uses strict enforcement and zero tolerance to maintain a safe com- munity while others would view such tactics as oppressive and overzealous. One community may be willing to fund more ofﬁcers per population in order to gain greater visibility and ofﬁcer presence while another community merely toler- ates the police department and believes they should only be seen when they are called. BOZEMAN, MONTANA POLICE DEPARTMENT - PATROL STAFFING AND DEPLOYMENT STUDY SPRING 2007 COPYRIGHT 2007 - ETICO SOLUTIONS INC. PAGE 4 OF 29 www.eticosolutions.com Benchmarking: Continuous process of measuring products, services, and practices against the toughest competitors or those companies recognized as industry leaders. (David T. Kearns, CEO, Xerox Corporation) The third and often most erroneous assumption in this process is the assump- tion that the chosen “similar” agencies are staffed appropriately. If an agency chooses four similar agencies that are all understaffed, the entire exercise becomes futile. If inquiries were to be made to the similar agencies about the appropriate- ness of their current stafﬁng, one may receive many different answers, all de- pendent on who is answering the ques- tion. Patrol ofﬁcers, police administra- tors, and city ofﬁcials may have differing perspectives on current stafﬁng levels. Both methods previously mentioned are based on a jurisdiction’s current population. Population is an exter nal workload that does not change based on the initiative and self-motivation of the ofﬁcers on the patrol division. More importantly, population does not adequately measure the amount of work created for the patrol division. Citi- zens residing in upscale gated communi- ties may not require the same police serv- ices as apartment residents or govern- ment subsidized housing residents, yet all are counted the same in the census. Census populations include the peo- ple who reside in the community as resi- dents. Ratios and comparisons based on population do not take into account the additional people that commute into a community for work, tourists that are drawn to a community, or migrant work- ers that do not appear on any US census poll. Even though these additional groups are not reﬂected in the city’s population, they must still be afforded police services and protection. Empirical Qualitative Analysis The best way to avoid the pitfalls of population-based studies is to select an alternative workload that accurately re- ﬂects the demands placed on the agency’s patrol division. By selecting an inter nal workload, derived from the agency’s own historical data, an agency can deter mine an optimal staff size for patrol that re- ﬂects their unique community character- istics regardless of the people who are being policed. For police patrol divisions, a measure of an appropriate internal workload al- ready exists in the for m of historical CAD (Computer Aided Dispatch) data. CAD systems typically capture each ac- tivity that an ofﬁcer is called to along with important dates and times such as dis- patching times, arrival times, and cleared times. By carefully analyzing an agency’s CAD data over past years, a forecast can be made for the total hours of work that a patrol division can expect in the current year and several upcoming years. This workload measure, the total hours of expected work, can be effectively used as the basis of an empirical qualita- tive stafﬁng and allocation study. Once the workload is accurately deter mined, the agency can set their own performance levels for patrol based on the amount of proactive time that is allotted to the aver- age patrol ofﬁcer. After determining the ofﬁcer availability ratios for the agency, an administrator can then make a deter- mination on optimal patrol stafﬁng that is relevant for their own community. This later method is not a one-size- ﬁts-all methodology for stafﬁng. It is unique to the agency under study and is born from their own historical data. The method is replicable in future years and does not contain the assumptions of comparative methods. Most importantly, the process is pliable enough to be modi- ﬁed appropriately based on available data within the agency and special cir- cumstances that may exist within an agency or community. Chosen Methodology This study utilized an empirical methodology based on internal qualita- tive data provided by the Bozeman Police Department. Four years of historical CAD data were analyzed in order to produce a model that would accurately forecast patrol workload into future years. In addition, the agency provided data pertaining to ofﬁcer leave times, training times, and other non-patrol days. This information was used to deter mine ofﬁcer availability so a total staff size could be deter mined for patrol. This report is broken down into three sections, each dealing with a differ- ent aspect of patrol analysis. All sections utilize the same CAD data and deﬁne workload in ter ms of hours of work. The ﬁrst section details the process of police resource analysis. This analysis examines the number of calls-for-service and the average time required to com- plete those calls. Ofﬁcer availability is also calculated and the two results are combined to form recommendations on the optimal total staff size for patrol based on various selected performance levels. The second section of this report uses the same workload data from the CAD to determine appropriate work schedules for the ofﬁcers on the patrol division. This is referred to as patrol de- ployment. By using historical CAD data, the shift schedules can be modiﬁed to provide stafﬁng by hour of the day that best matches the call-for-service load by hour of the day. Scheduling the right amount of people at the right time in- creases efﬁciencies among the patrol divi- sion. The third and ﬁnal section addresses the issue of geographic beat design. Using the same CAD data generated for the ﬁrst two sections, the city can be partitioned into a number of distinct beats based on equal workload distribution among all beats. Efﬁcient beat design can have a beneﬁcial effect on the operation of the agency in a number of ways. By strategi- cally placing ofﬁcers in predetermined beats based on calls for service, the agency has a better chance of reducing cross beat dispatching, increasing ofﬁcer and citizen safety, and lowering response times to calls-for-service. This report will conclude with ﬁnal recommendations for change as a result of the data analysis. In addition, obser- vations made during this study that were not directly related to the scope of this study will also be cited. BOZEMAN, MONTANA POLICE DEPARTMENT - PATROL STAFFING AND DEPLOYMENT STUDY SPRING 2007 COPYRIGHT 2007 - ETICO SOLUTIONS INC. PAGE 5 OF 29 www.eticosolutions.com II. Police Resource Analysis The ﬁrst step in the resource analysis portion of this study was to obtain CAD outputs for ﬁve previous years and pre- pare them for analysis. CAD data records contain a large amount of data related to each activity perfor med by the patrol ofﬁcers. There are many CAD software vendors in the public safety market and each vendor’s product varies slightly from the rest in the way is collects and stores data. The Bozeman Police Department has been using Intergraph CAD software since February of 2001. The software creates a separate record for each ofﬁcer that responds to every dispatched call. Items stored in each record include the time and location of the original call, the type of call, the arrival time of each ofﬁ- cer, the cleared time for each ofﬁcer, the unit ID for each responding ofﬁcer, the priority of the call, the emergency service zone for the caller’s location, and a host of additional information. CAD systems almost always utilize a proprietary database to store the call information. A secondary software appli- cation, known as a records management system (RMS), imports the information from the CAD database to provide managerial data for agency use. RMS software will usually have a preset collec- tion of database queries based on the software designer’s understanding of the most useful infor mation to users. The Bozeman Police Department uses a RMS system called “I/LEADS” from Inter- graph Corporation. I/LEADS (Inter- graph’s Law Enforcement Agency Data System) can provide agency data on traf- ﬁc crashes, citations, incidents, and a host of other activities within the agency. Five years of CAD data, 2002 through 2006, were analyzed by Lt. Rich McLane of the BPD. The total number of calls-for-service in each CAD category (frequency) along with the average times spent on calls in each category were pro- vided to Etico Solutions. These frequen- cies and average times were then used to calculate the total reactive time required by the patrol division for past years. By utilizing a forecasting algorithm based on linear regression, estimated workloads for future years were obtained. This methodology avoided the use of comparable agencies by calculating the BPD’s workload from the last ﬁve full years of CAD data. Not only does this method avoid the inherent problems with “comparable agencies”, it also avoids the problem of transient populations. The police department is obligated to respond to calls-for-service whether they are placed by a Bozeman resident, a tourist, or a migrant ﬁeld worker. Therefore, by using CAD data generated by the agency, a true account of the BPD’s workload was obtained. The categories of calls-for-services from the CAD database included all pa- trol activities perfor med by the ofﬁcers of the patrol division. The basis of this study rests on the reactive activities per- formed by patrol. The ﬁrst step, there- fore, was to aggregate the time spent on all reactive activities. To t a l R e a c t i v e T i m e Reactive activities are the responses and services that are expected of the police department by the citizens and visitors within the city of Bozeman. This may include a threshold level of trafﬁc enforcement and directed patrols. Reactive Activities Reactive activities would include, but are not limited to, the following types of activities: criminal calls emergency calls non-emergency calls agency walk-ins trafﬁc crash investigations service activities check welfare motorist assists missing persons mentally ill individuals assisting other agencies ﬁre other government agencies administrative functions roll-call attendance report writing assisting investigators The number of minutes spent on reactive activities during the average pa- trol hour will be referred to as “MR” (Minutes of Reactive time per hour). Proactive Activities Proactive activities would include, but are not limited to, the following types of activities: preventative patrol community oriented policing spontaneous training case follow-ups building checks selective trafﬁc enforcement The number of minutes spent on proactive activities during the average patrol hour will be referred to as “MP” (Minutes of Proactive time per hour). Once all calls were sorted by CAD category, a total reactive time was deter- mined by multiplying the frequency of each reactive CAD category by the aver- age time per call in each category. Average Call-For-Service Times A total time-on-call for every activity in the database was calculated based on the primary ofﬁcer’s en-route time to cleared time as well as all supplemental ofﬁcers that responded to each call. This process is not an exact science and the data to get exact times for each call is never precise. Any time an ofﬁcer con- ducts an activity and fails to call in that activity to dispatch, the obligated time for that call will not appear in the database. When ofﬁcers call out at the station on a “busy” code, it is impossible to sort out reactive times from proactive times. Court time can often be lost if an on-duty BOZEMAN, MONTANA POLICE DEPARTMENT - PATROL STAFFING AND DEPLOYMENT STUDY SPRING 2007 COPYRIGHT 2007 - ETICO SOLUTIONS INC. PAGE 6 OF 29 www.eticosolutions.com ofﬁcer does not call out that they are in court or if an ofﬁcer attends court off- duty unbeknownst to dispatch. Report writing is one of the biggest items missed when average call times are calculated. It is common in law enforce- ment agencies for ofﬁcers to hold their reports during busy times to reduce the holding times for calls in the queue. As things slow down later in the shift, ofﬁcers will then call out at the station to write their reports or will complete the reports in their cars while stationed throughout the city. In more cases than not, reports that are written after the ofﬁcer clears the call and takes other calls, are never re- corded or tied back to the original inci- dent that generated the report. Therefore, the majority of time spent writing re- ports, which is signiﬁcant in most agen- cies, is never credited towards the agency’s reactive time. In the case of Bozeman PD, the CAD does not ade- quately capture report writing times. It is impossible to tell, based on the CAD output, how many reports are being held for later in the shift and how many are being completed before the ofﬁcer clears the call which generates the report. The reactive workload used for this report should be considered as a conservative estimate since it is known that report writing times are not being adequately captured. Despite best efforts to clean CAD times and calls, there are incidents that legitimately take either extraordinarily longer times to handle or incidents that resolve themselves immediately. A raw mean time for calls-for-service will take every event in the database that falls into a particular activity category and use its time to compute the overall average. Thus, outliers such as the two cases just mentioned will effect the aver- age to some degree causing greater stan- dard deviations in the average and reduc- ing the conﬁdence interval of the ﬁnal results. As an alternative to the raw mean time, a trimmed mean can be determined using more advanced statistical software. A trimmed mean will sort all events in each particular activity category by time- on-call and discard the top 5% and bot- tom 5% of the call times as assumed outliers. This eliminates the calls that took an abnor mally long time or an ab- normally short time to handle and com- putes a mean from the middle 90% in each category. If this study were based on a manu- facturing situation, or any other situation where the workload is unifor m and con- sistent, this researcher would be more inclined to utilize a trimmed mean. However, in law enforcement, outliers to the nor mal service times do not necessar- ily mean a record keeping mistake. Real- istically, that is the nature of the envi- ronment in which law enforcement agen- cies operate. Calls can vary from the norm at any time, for any number of reasons. Therefore, a raw mean was used for this study based on the ﬁve years of CAD data mentioned earlier. The reactive time, estimated by a raw mean, for 2007 was 29,035 hours. This equates to an average of 79.547 hours of reactive work each day. With each unit working a ten hour day, the BPD would need to ﬁeld a minimum of 7.955 units every 24 hour period. Note that this is the minimum number of units that can be ﬁelded to handle the reactive workload. Any fewer units per day and calls would begin to stack at the dispatch level with no chance of catching up. T h e N e e d f o r P ro a c t i v e T i m e I n Pa t ro l O p e r a t i o n s To the casual observer it might appear that for optimal efﬁciencies, ofﬁcers should be on reactive time every minute of every hour. Such an assumption is far from accurate and is worthy of further explanation at this point in the study. Including an appropriate amount of proactive time into each patrol hour has ramiﬁcations on the agency, the ofﬁcer, and the citizens whom they serve. Some consequences affect single stake-holders, some affect multiple stake-holders, and some affect all stake-holders. One of the more obvious reasons against ofﬁcers running call-to-call is the inevitable burn-out that will slowly manifest itself. Ofﬁcers will eventually require a break for meals, for rest, and for collecting their thoughts. In many criminal situations, the solvability of a case begins to deteriorate from the moment the incident happens. If the initial re- sponding ofﬁcer is rushed to get the information so they can move on to the next call, there is a greater chance that immediate follow-up opportunities will not be taken and solvability will be diminished. In large agencies where shift bids are based on seniority, it is quite possible to have an entire shift of ofﬁcers who are all within their ﬁrst several years of service. When corrective action is required on behalf of the supervisor, there must be some proactive time for that to occur. If the ofﬁcer clears one call and heads directly to the next, throughout their entire shift, the supervisor will miss that “training oppor- tunity” and the ofﬁcer will continue behaving in the same manner until the correction is made. In order to see the managerial efﬁciencies that are enhanced with proactive time, the patrol function must be viewed from an opera- tions management perspective. From that viewpoint, proactive time can be seen effecting the level of cross-beat dispatching within the agency, the duration of the agency’s patrol inter val, and the probability of saturation when the next citizen calls for police services. BOZEMAN, MONTANA POLICE DEPARTMENT - PATROL STAFFING AND DEPLOYMENT STUDY SPRING 2007 COPYRIGHT 2007 - ETICO SOLUTIONS INC. PAGE 7 OF 29 www.eticosolutions.com Cross Beat Dispatching A main tenet to the community oriented policing philosophy is ofﬁcer familiarization with a small area. By assigning ofﬁcers to the same area for long periods of time, they begin to take ownership of that area and build relationships with the local residents. Often over- looked, however, is the frequency and duration of time that an ofﬁcer is pulled from their assigned beat in order to answer a call-for-service in another beat when the assigned beat ofﬁcer is occupied on a call. Pulling an ofﬁcer from their beat to answer a call in someone else’s beat is refer red to as “cross-beat dispatching”. Based on probability statistics, the amount of time that an ofﬁcer spends on cross-beat dispatched calls each hour, referred to as “Mx”, can be calculated if the minutes of reactive time per hour (MR) and the number of beats (N) are known. Assuming one beat ofﬁcer in each beat: As the minutes of reactive time per hour and the number of beats both increase, the for mula collapses into a much simpler form. Notice that in both formulas, MR is subject to an exponential term. Therefore, as the minutes of reactive time each hour increases, the minutes of cross-beat dispatching increases exponentially. The exponential growth of Mx creates a curve as shown in Chart 1. According to historical CAD data, the patrol division is operating with an MR value of 45 minutes/hour at the time of this study. Assuming the department operated with a three-beat deployment, the department’s Mx would be located on the curve at the point desig- nated by the BPD icon. The ofﬁcers would be cross-beat dispatched out of their area an average of 27.42 minutes out of every patrol hour. Patrol Interval A second efﬁciency to the operation of the organization gained by proactive time is the agency’s patrol interval. By deﬁnition, the pa- trol interval (PI) is the average time interval between two consecutive passes by the same location by units while on random patrol on proactive time. Notice in the deﬁnition that this perfor mance measure is based on the amount of proactive time spent by each ofﬁcer per hour. A patrol interval can be calculated using the for mula shown on the right. Since the “MP” variable and the “Units ﬁelded” variable are both in the denominator, another exponential effect in the patrol interval is seen as additional ofﬁcers are added to the patrol division and the average MP of all ofﬁcers is in- creased. Practically speaking, a patrol interval could also be deﬁned as the number of hours that a crime would go unnoticed until an ofﬁcer happens by on random patrol on proactive time. If individuals with criminal intent were to sit and observe a potential target, the patrol interval indicates how long it would be before an ofﬁcer returned to that location on proactive patrol. BOZEMAN, MONTANA POLICE DEPARTMENT - PATROL STAFFING AND DEPLOYMENT STUDY SPRING 2007 COPYRIGHT 2007 - ETICO SOLUTIONS INC. PAGE 8 OF 29 www.eticosolutions.com € Mx =(MR )2 60 € M x = 2 M R 60 1 −1 N    M R 60       (N −1)    As the number of beats increases, Mx can be estimated as: 0 15 30 45 60 60 55 50 45 40 35 30 25 20 15 10 5 0 Actual Mx Estimated Mx MR MX Chart 1. Cross-beat Dispatch Levels as a Function of Reactive Time per Hour € PI =Street miles in the jurisdiction Average patrol speed ∗M P 60       ∗Units fielded Probability of Saturation The probability of saturation (POS) is deﬁned as the probability that when the next call-for-ser vice is received by the dispatch center, there will be no units available to answer the call. As the minutes of reactive time per hour (MR) increases for the patrol ofﬁcers, the chance that all units will be unavailable increases. This leads to a higher probability of saturation for the agency. The POS for an agency is in constant change as the call volume ﬂuctuates, the time-on-call changes, and the number of ofﬁcers currently on the street changes. One way to reduce the agencies POS is to ﬁeld additional units, thereby lowering the probability of saturation to an acceptable level. This option obviously comes at an additional cost. A second option is to screen calls in order to reduce the workload on the existing ofﬁcers in the ﬁeld. A third and ﬁnal method is to reduce the amount of time that ofﬁcers spend on each call. The last two options can have nega- tive results in future citizen surveys. At the current time, with patrol ofﬁcers maintaining an average MR value of 45 minutes out of every hour, 3.71 calls per hour, and an average of 54 minutes per call, the agency’s POS averages approximately 66.04%. During the hours of heavy workload, the calls-for- service per hour increase and calls get stacked. As things begin to slow on the street, the ofﬁcers begin to work their way through the back- log of calls and the POS drops below 100% once again. P e r f o r m a n c e F a c t o r Based on the work previously dis- cussed, it is estimated that in 2007, 7.955 units need to be ﬁelded per day in order to meet the total reactive workload. This assumes that all ofﬁcers will be running call-to-call for their entire shift, everyday of the year. Three different performance meas- ures have already been discussed; cross- beat dispatching, patrol intervals, and the probability of saturation. All three per- formance measures are directly depend- ent on the amount of proactive time per average patrol hour that is afforded to the ofﬁcers of the patrol division. In order to reduce cross-beat dis- patching, shorten the patrol intervals, and reduce the probability of saturation, ad- ditional units must be added to the street each day. Additional units that are added to the 7.955 minimum units each day will be referred to as “performance units” since they are added to the ﬁeld to reach a desired or targeted performance level. Up to this point in the study, the work has been of a mathematical or sta- tistical nature. However, when determin- ing how much proactive time is appropri- ate for the ofﬁcers on patrol, a subjective decision must be made by the agency administrators. That decision, as John Shuiteman once eloquently posed, “is not a matter of mathematics or statistics. It is a matter of human judgment and com- munity resources.4” Based on the level of police presence expected by the citizen’s of Bozeman and the ﬁnancial resources available to the Police Department, an agency adminis- trator will have to decide how many min- utes out of each patrol hour should be dedicated to proactive activities such as free patrol and community oriented polic- ing. This is the MP value that was dis- cussed earlier in this report. Since MP and MR must always equal 60 minutes, an agency administrator could also de- termine the number of minutes each patrol ofﬁcer should spend reacting to calls-for-service and other activities in the average patrol hour (MR). Once a value is chosen for MR, the value is used to determine the total amount of proactive time that must be added to the total reactive time from the CAD data. Dividing the desired MR value by 60 will yield the fraction of each patrol hour that the agency administrator wants spent on reactive activities. If we invert this fraction , we obtain a multiplier. This multiplier is referred to as the “perfor m- ance factor” or Fperf. Assume that an MR value of 30 minutes is chosen. The perfor mance factor for the agency would then be: 60Fperf= = 2.0 30 With an MR of 30, the patrol ofﬁcers would be available to conduct proactive law enforcement for 30 minutes out of every patrol hour. The total reactive time from the CAD data is multiplied by the perfor m- ance factor to yield the Total Patrol Time for the agency. The 29,035 hours of reactive time based on the frequency of each activity category and the raw mean times is multiplied by the Fperf of 2.0 in this example. The Total Patrol Time for the agency then becomes: Total Total Patrol = Fperf x Obligated Time Time = 2.0 x 29,035 hours = 58,070 hours Converting the new total patrol time per year into an average daily workload indicates that 159 hours of work must be done on a daily basis. With each unit working 10 hours a day, the agency will have to ﬁeld 15.9 or 16 units per day. When the number of units based on reactive time was calculated earlier, with an MR of 60, it was deter mined that 7.995 minimum units were needed. Now that the MR is reduced in this example to 30 minutes out of the hour, an additional 7.995 performance units were added to the 7.995 minimum units. They were added to reach a targeted performance level of only 30 minutes of reactive time per hour as opposed to 60 minutes. Since the MR value must be subjec- tively chosen by an agency administrator, this report will not contain a recommen- dation of an exact number of ofﬁcers required for the patrol division. Instead, BOZEMAN, MONTANA POLICE DEPARTMENT - PATROL STAFFING AND DEPLOYMENT STUDY SPRING 2007 COPYRIGHT 2007 - ETICO SOLUTIONS INC. PAGE 9 OF 29 www.eticosolutions.com it will detail the results of various MR values as they effect the ability to respond to calls-for-service and conduct other necessary functions of patrol. It is up to the agency to chose the level of perfor m- ance they feel they can justify and afford. P e r f o r m a n c e V a ri a b l e s a s a f u n c t i o n o f t h e s e l e c t e d M R v a l u e As the total minutes of proactive time each hour increases, the probability of saturation (the probability that no units will be available when the next call-for-service is received by the dispatch center) will decrease, the minutes of cross-beat dispatch per hour (Mx) will de- crease, and the number of required units, ofﬁcers, and total staff will increase. Table 1 shows the cost/return of adding additional units and reducing the average MR value of the patrol division assuming a 3-beat deployment, 3.71 calls per hour, and 54 minutes per call. The deter mination of Ofﬁcers/Day and Total Patrol Staff will be explained in a later section of this report. BOZEMAN, MONTANA POLICE DEPARTMENT - PATROL STAFFING AND DEPLOYMENT STUDY SPRING 2007 COPYRIGHT 2007 - ETICO SOLUTIONS INC. PAGE 10 OF 29 www.eticosolutions.com Ta b le 1 C o s t Re tur n o f A d di n g A d d i t i on a l U n it s to t h e Patro l Divi s i on M R PO S Mx U n i t s /D a y Tot a l Pat ro l S ta f f 60 152%52.46 7.95 19.67 Ofﬁcers 59 148%51.65 8.09 20.0 Ofﬁcers 58 144%50.73 8.23 20.34 Ofﬁcers 57 139%49.72 8.37 20.70 Ofﬁcers 56 134%48.62 8.52 21.07 Ofﬁcers 55 130%47.44 8.68 21.45 Ofﬁcers 54 126%46.19 8.84 21.85 Ofﬁcers 53 122%44.89 9.01 22.26 Ofﬁcers 52 119%43.54 9.18 22.69 Ofﬁcers 51 114%42.16 9.36 23.14 Ofﬁcers 50 111%40.75 9.55 23.60 Ofﬁcers 49 107%39.32 9.74 24.08 Ofﬁcers 48 103%37.88 9.94 24.58 Ofﬁcers 47 99%36.43 10.15 25.11 Ofﬁcers 46 95%34.99 10.38 25.65 Ofﬁcers 45 90%33.55 10.61 26.22 Ofﬁcers 44 85%32.13 10.85 26.82 Ofﬁcers 43 81%30.72 11.10 27.44 Ofﬁcers 42 76%29.34 11.36 28.09 Ofﬁcers 41 71%27.97 11.64 27.97 Ofﬁcers 40 66%26.64 11.93 29.50 Ofﬁcers 39 62%25.33 12.24 30.25 Ofﬁcers 38 58%24.06 12.56 31.05 Ofﬁcers 37 54%22.81 12.90 31.89 Ofﬁcers 36 50%21.60 13.26 32.78 Ofﬁcers 35 46%20.42 13.64 33.71 Ofﬁcers 34 43%19.27 14.04 34.70 Ofﬁcers 33 38%18.15 14.46 35.76 Ofﬁcers 32 34%17.07 14.92 36.87 Ofﬁcers 31 29%16.02 15.40 38.06 Ofﬁcers 30 25%15.00 15.91 39.33 Ofﬁcers U n i t M i x After calculating the total obligated time and selecting a performance level for the agency, a determination was made on the optimal number of “units per day” based on a chosen MR. In that context, a “unit” was deﬁned by what it did and not by how it looks. Therefore, a “unit” could be a one-ofﬁcer, a two-ofﬁcer unit, or any other type of unit the agency chooses to ﬁeld. At the time of this study, the patrol division operates primarily on single- ofﬁcer units. Therefore, the number of ofﬁcers needed per day is equal to the number of units calculated per day. As the agency increases in size, a time may come when two-ofﬁcers units are created. Ta b l e 2 . U n i t M i x Ty p e s U n i t Ty p e /D ay #% One-Ofﬁcer Patrol 26 100 Two-Ofﬁcer Patrol 0 0 Units/Day in Sample 26 100 All calculations in this study are designed to maintain the same percent- ages among the various mix of units as currently represented in the patrol divi- sion. In other words, as alternative MR values are selected and total stafﬁng changes, the number of one-ofﬁcer units will always remain at the current percent- age of total units. The unit mix process determines the number of on-duty ofﬁcers that must be ﬁelded to staff the number of on-duty units required. The last major step was to determine the total number of ofﬁcers that must be assigned to the patrol divi- sion in order to ﬁeld a set number of on- duty ofﬁcers each day. This step deter- mines ofﬁcer availability, referred to in this report as the Shift Relief Factor. S h i f t R e l i e f F a c t o r By deﬁnition, the shift relief factor is “the number of units required to staff one shift position every day of the year.” This is a mul- tiplying factor which will convert the number of on-duty ofﬁcers each day to the total staff sized needed for the patrol division. The shift relief factor is com- prised of three sets of data. •Regular Days Off •Beneﬁt Days Off •Compensatory Leave •Family Leave •Holiday Leave •Maternity Leave •Military Leave •Sick Leave •Vacation Leave •Personal Leave •Other Paid Leave •Non-Patrol Days •Administrative Leave •Court Leave •Light Duty •Special Assignment •Training Regularly Scheduled Days Off The ofﬁcers work four ten-hour shifts each week with three days off. In- tuitively, they would appear to have 156 days off per year (52 weeks multiplied by 3 days per week). However, there are not exactly 52 weeks in a year. For larger agencies, this somewhat minor variation can make a large difference. For this study, we used an exact ﬁgure of 156.4286 days off per year. Beneﬁt Days Off To deter mine ofﬁcer availability, it was essential to determine the amount of beneﬁt leave taken by the average ofﬁcer for each beneﬁt leave category. The cal- culations included only the time used by the ofﬁcers and did not include any vaca- tion, sick, or comp time that was con- verted to pay or carried over into subse- quent years. Compensatory time is treated differ- ently than the other for ms of leave. Since comp time can be earned in various amounts and then either cashed out or taken as leave, a net comp time was ac- quired. Net comp time is the difference between the comp time earned by sup- plementing patrol and comp time taken in place of working a nor mal patrol shift. Once all leave was tallied within the various categories, each category sum was divided by the number of ofﬁcers on patrol to create an average leave time for patrol. Table 3 is a summary of the beneﬁt leave data collected from the City of Bozeman Human Resources Division. Ta b l e 3 . B e n e f i t D ay s O f f B e n e f i t L e av e D ay s Comp Time 9.51 Family Leave 1.56 Holidays 0 Maternity Leave 0 Military Leave 0 Personal Days .34 Sick Leave 6.64 Vacation 13.99 Total 32.04 Non-Patrol Days Non-patrol days are any days in which the ofﬁcer is being paid for their presence without actually being engaged in answering calls-for-service. Training and special assignment are two clear examples of non-patrol days. New ofﬁ- cers will spend 90 to 100 days in initial ﬁeld training and attendance at the Mon- tana Law Enforcement Academy prior to their ﬁrst day as a solo ofﬁcer. This calcu- lation also includes any days when an ofﬁcer participates in specialty training such as K-9 or SWAT training as part of their normal duty shift. Various forms of administrative leave, such as suspensions, would also be considered as non-patrol days. BOZEMAN, MONTANA POLICE DEPARTMENT - PATROL STAFFING AND DEPLOYMENT STUDY SPRING 2007 COPYRIGHT 2007 - ETICO SOLUTIONS INC. PAGE 11 OF 29 www.eticosolutions.com Table 4 is a summary of the non- patrol day data collected from the various places within the agency. Ta b l e 4 . N o n -P a t r o l D ay s N o n -P a t ro l D ay s D ay s Administrative Leave .45 Court Leave .72 Light Duty 5.62 Special Assignment 0 Training 22.11 Total 28.9 Adding the results of the three time off categories provided an average time off patrol for the ofﬁcers within the patrol division. Regularly scheduled days off = 156.4286 Beneﬁt Days off = 32.04 Non-Patrol Days = 28.90 Avg. days off patrol/ofﬁcer = 217.3586 The average days off patrol per year per ofﬁcer is then multiplied by the shift length to convert the days into hours. According to Human Resource data, the ofﬁcers have 2173.586 hours off patrol per year. The shift relief factor is calculated by dividing the entire time that could be worked by the time actually worked. Based on one year of Human Re- source data for beneﬁt days off, and data collected by Bozeman PD personnel for non-patrol days, the shift relief factor for the patrol division was found to be 2.47221. Therefore, for every ofﬁcer that needs to be ﬁelded each day, to staff the required number of units, the agency must have 2.47221 ofﬁcers assigned to the patrol division. On page 9 of this report, an exam- ple was given to demonstrate the use of a perfor mance factor. In that example, we assumed a desired MR of 30 minutes and found that 16 units would have to be ﬁelded each day. Continuing that exam- ple, the unit mix for the patrol division indicates that 16 on-duty ofﬁcers would have to be ﬁelded each day to staff the required 16 units since all units are single- ofﬁcer cars. The 16 on-duty ofﬁcers per day is now multiplied by the shift relief factor of 2.47221 to determine the total staff size for the Bozeman patrol division. In order to daily ﬁeld 16 units, staffed by 16 ofﬁ- cers, the patrol division would have to have a total staff size of 39.33 or 39 ofﬁ- cers. This same process is used repeat- edly to determine the total patrol staff in Table 1. III. Patrol Resource Deployment A properly staffed patrol division is the ﬁrst step in improving efﬁciencies within the patrol function. However, if patrol stafﬁng by day- of-week and hour-of-day is not in harmony with the patrol workload by day-of-week and hour-of-day, efﬁciencies will still be lacking. The second phase of this study addresses the current work schedule in use by the patrol division and it’s relation to the workload that is driving the need for ofﬁcers. Before looking at the current patrol work schedule, it would be prudent to undertake a review of the more popular police work sched- ules in place today. As we review various schedules, it is also important to review a number of characteristics that pertain to various sched- ule types. C h a r a c t e ri s t i c s o f P o l i c e Wo r k S c h e d u l e s Almost all police work schedules can be classiﬁed into one of four schedule types; ﬁxed days off, locked rotating, unlocked rotating, or unstructured. Each schedule type has ramiﬁcations on a number of sched- uling dynamics such as team integrity, unity of command, and schedule equity. Fixed Days Off Schedules On a ﬁxed days off schedule, ofﬁcers will receive an assigned set of consecutive days off that will be the same each week of the bid period. Fixed days off sched- ules will have a duty cycle length (the sum of all days in the duty cycle pattern) of exactly seven days. With few exceptions, ﬁxed days off patterns are usually limited to a 5-on 2-off pattern with an 8 hour shift length, or a 4-on 3-off pattern with a 10 hour shift length. In some cases, a 3- on 4-off pattern may be used with a 12 hour shift length and an occasional pay- back day to get the average workweek back up to 40 hours per week. Fixed day off schedules can provide uniform stafﬁng by day of the week, non- uniform stafﬁng by day of the week, or proportional stafﬁng by day of the week. Locked Rotating Schedules Locked rotating schedules will have a repeating pattern of on and off duty days. The duty cycle length (the sum of all days in the duty cycle pattern) will be an even multiple of 7. Since the pattern is easily divisible by 7 (the number of days in a week), any long on-duty or long off-duty portions of the pattern will al- ways be locked to the same days of the week. BOZEMAN, MONTANA POLICE DEPARTMENT - PATROL STAFFING AND DEPLOYMENT STUDY SPRING 2007 COPYRIGHT 2007 - ETICO SOLUTIONS INC. PAGE 12 OF 29 www.eticosolutions.com For example, a duty cycle pattern of “2on - 2off - 3on - 2off - 2on - 3off ” is a locked rotating schedule. The duty cycle length is 14 days which is divisible by 7. If this pattern is started on a Monday, the three consecutive days off that appear at the end of the pattern will always fall on a Friday-Saturday-Sunday. Note that the days off change over the 14 day cycle, but the unique portions of the pattern remain locked to a particular set of days. Locked rotating schedules can pro- vide uniform stafﬁng by day of the week or variable stafﬁng by day of the week. However, they are not as ﬂexible as ﬁxed days off when workload by day of the week changes signiﬁcantly. Unlocked Rotating Schedules Similar to locked rotating schedules, unlocked rotating schedules will also have a repeating pattern of on-duty and off- duty days. However, the duty cycle length for unlocked schedules (the sum of on-duty and off-duty days in the pattern) will not be a multiple of 7. Unlocked rotating schedules will create a schedule in which consecutive off-duty days will continue to move across the entire seven days of the week until the pattern begins to repeat itself. Unlocked rotating days off can pro- vide uniform stafﬁng by day of the week if the number of patrol ofﬁcers available for deployment is compatible with the schedule (there is a mathematical process to determine compatibility that is beyond the scope of this report.) Unlocked rotating schedules cannot be used to provide proportional stafﬁng by day of the week. Agencies that have vastly different workload curves across the seven days of the week would not be able to match their call for service load effec- tively with such a schedule. Unstructured Schedules Unlike the other three types of schedules previously described, unstruc- tured schedules do not have a repeating pattern of on-duty and off-duty days. Instead, unstructured schedules are cre- ated one month at a time based on the needs of the agency. To create an unstructured schedule, the shift supervisor would begin by post- ing a calendar for the upcoming month which indicated the number of ofﬁcers allowed off each day. Ofﬁcers would then chose their days off using some form of rotating list until all ofﬁcers on the shift have chosen a designated number of days off for the month. The ﬁnal schedule for the month is then reviewed and approved by the supervisor before being posted for the month. Unstructured schedules are rare in law enforcement and are in use by ap- proximately 5% of agencies across the U.S. Team Integrity In a 24 hour /7 day operation such as law enforcement, it can be difﬁcult to build a “team” concept among patrol ofﬁcers. If ofﬁcers are assigned ﬁxed days off each week, the group of ofﬁcers work- ing together on a shift will change throughout the workweek as ofﬁcers take their various days off. Team integrity refers to the consistency of the “team as a unit” from day to day. Fixed days off, as previously men- tioned, provide very little team integrity since the regular days off assigned to the shift ofﬁcers span the seven days of the week. Conversely, rotating schedules will often employ a “squad” methodology. Each squad will rotate through the shift pattern as a “team”. All members of the squad will be assigned the same work days and the same days off. Thus, over time, the squad members begin to pre- dict the actions of their team members through unspoken body language, man- nerisms, and past calls. This unspoken communication that develops over time is believed to increase the efﬁciency of the team as synergy builds amongst the shift. Unstructured schedules typically have the worst team integrity of any schedule since there is no repeating pat- tern of on-duty and off-duty days. Unity of Command Past studies in law enforcement have concluded that a clear chain-of- command is a critical component to a well organized, well prepared, and efﬁ- cient agency. McKinsey & Company authored a report in 2002 entitled “Improving NYPD Emergency Preparedness and Response.”5 Of the 20 improvement opportunities listed in their report, the second was “Better clarity in the chain of command” with a desired outcome of “Clear reporting lines with no gaps or duplication of activities.” In 2006, the United States Govern- ment Accountability Ofﬁce released a report on the “Preliminary Observations Regarding Preparedness and Response to Hurricanes Katrina and Rita.”6 One of their major ﬁndings stated, “...there were multiple chains of command, a myriad of approaches and processes for requesting and providing assistance, and confusion about who should be advised of requests and what resources would be provided within speciﬁc time frames.” Bill Wilkerson, CEO of The Global Business and Economic Roundtable on Addiction and Mental Health, conducted a survey in 2001 on the “Top 10 Causes Of Workplace Stress.”7 The second cause on the list was "lack of communica- tions". Poor communication up and down the chain of command leads to decreased perfor mance and increased stress. When work schedules are reviewed, the issue of communication and chain of command should always be considered. Unity of Command, in the context of police work scheduling, means that each ofﬁcer reports to one designated supervisor. In the “squad” system used by many rotating shifts, the supervisor works and takes their days off with the rest of the squad. Therefore, the same supervisor is present for all days worked by the squad and complete unity of command is ob- tained. Fixed days off violate the concept of unity of command due to the variety of days off across the shift. A patrol ofﬁ- BOZEMAN, MONTANA POLICE DEPARTMENT - PATROL STAFFING AND DEPLOYMENT STUDY SPRING 2007 COPYRIGHT 2007 - ETICO SOLUTIONS INC. PAGE 13 OF 29 www.eticosolutions.com cer may report to one supervisor for the ﬁrst two workdays and a different supervi- sor for the remaining days of the week. Unstructured schedules also violate the concept of unity of command due to their ever-changing nature. Unsupervised “power shifts” (ofﬁcers assigned to peak times that overlap the traditional day/evening/night shifts) are usually required to report to one supervi- sor at the beginning of their shift and then change mid-shift to the oncoming supervisor. This creates unnecessary stress as the ofﬁcer must sometimes change their policing style based on the expectations of the current supervisor. Unity of command is also important to the police administration when it is time for employee evaluations. There is a beneﬁt to having one supervisor oversee an ofﬁcer consistently as opposed to sev- eral supervisors comparing notes and working cooperatively on evaluations based on the ofﬁcers they supervise the most. Finally, unity of command improves efﬁciencies by establishing clear expecta- tions and boundaries for the ofﬁcers on patrol. When an ofﬁcer becomes familiar with the supervision style and the expec- tations of their supervisor, they are more inclined to work to their potential. When the patrol ofﬁcer does not know how their supervisor will respond to various activi- ties, they are more likely to “play it safe’ and provide a mediocre level of perfor m- ance. For illustrative purposes, imagine a dog conﬁned in a yard by a buried elec- tric fence. When the dog crosses the invisible boundary, a shock is received. After the ﬁrst several negative conse- quences (shocks), from a mechanism they cannot see, the dog will not venture near the area where they were shocked. Even- tually, the dog will gravitate towards the center of the yard where they are “safe” from the consequence. Effectively, the dog is limiting it’s own range due to the uncertainty of the exact limits of move- ment. Similarly, when an agency fails to provide very clear limits, or when the limits are interpreted differently by differ- ent supervisors, the ofﬁcers will remain in the center of their potential performance, safe from negative consequences, yet self- limiting their potential. Thus, clear ex- pectations, set by a strong unity of com- mand, can have the ability to increase perfor mance levels among the patrol division. Schedule Equity Schedule equity refers to the overall fair ness of the schedule to all ofﬁcers assigned to the shift. Work schedules based on ﬁxed days off do not provide schedule equity for the ofﬁcers. Typically, the most senior ofﬁcer will receive weekends off while the more junior ofﬁcers are left with mid-week consecutive days. If ofﬁcers are allowed to bid shifts as well as days off, an agency may ﬁnd that their least desirable shift, one which may have the highest work- load, is staffed with their most junior ofﬁcers and least seasoned ofﬁcers. Rotating schedules, by their very name, rotate all ofﬁcers through the duty cycle pattern giving all ofﬁcers the same access to long off duty periods and occa- sional weekends off and providing a high degree of schedule equity. Unstructured schedules can have schedule equity if an equitable process for choosing days off is used. If ofﬁcers chose their days off based on a rotating list, schedule equity may be obtained. If ofﬁcers chose their days off based on seniority each month, schedule equity would not exist. E v a l u a t i o n o f T h e C u r re n t P a t ro l S c h e d u l e The patrol division for the Bozeman Police Department is currently working a 4-on 3-off ﬁxed day off schedule with a 10-hour shift length. At the time this study began, there were 25 ofﬁcers as- signed to the patrol division. Three shifts are ﬁelded each day with 9 ofﬁcers working 7a-5p, 9 ofﬁcers working 5p-3a, and 7 ofﬁcers working 9p- 7a. The 4-on 3-off schedule creates six hours of overlap between shifts each day due to three 10-hour shifts in a 24 hour day. The BPD purposely scheduled the overlap from 9p-3a which coincides with their busiest hours of the day. The exist- ing schedule provides the stafﬁng by shift and day of the week shown in Table 5 . BOZEMAN, MONTANA POLICE DEPARTMENT - PATROL STAFFING AND DEPLOYMENT STUDY SPRING 2007 COPYRIGHT 2007 - ETICO SOLUTIONS INC. PAGE 14 OF 29 www.eticosolutions.com Ta b l e 5 . C u r r e n t S t a f f i n g b y S h i f t a n d D a y o f We e k S u n d a y M o n d a y Tu e s d a y We d n e s d a y T h u rs d a y F ri d a y S a t u rd a y 7 a -5 p 6 5 5 5 4 6 5 5 p -3 a 5 5 5 5 5 6 5 9 p -7 a 3 3 3 5 5 5 4 To t a l 14 13 13 15 14 17 14 In order to gauge the efﬁciency of the current schedule, it must be com- pared to the average workload experi- enced by the agency. One complete year of CAD data was analyzed and sorted by hour of the day and day of the week to deter mine when the agency was experi- encing high and low periods of activity. The analysis was based on the num- ber of hours of work experienced during each hour and each day. This was calcu- lated using the number of calls from each CAD category each hour of the day, and multiplying them by the average time for each CAD category. The product from each category was then summed to get a total amount of work per hour of day and day of week. This method is prefer- able to using just the number of calls for service within each hour. The later method does not take the nature of calls into consideration and may provide an inaccurate picture of the true workload. Workload by day of the week showed Friday and Saturday as the busi- est days. Tuesday through Thursday showed the lightest workloads (Chart 2). An average workload by hour of the day was calculated for the patrol division so a comparison could be made against current scheduling practices. As before, the workload was meas- ured in hours of work per individual hour based on all patrol activities performed by the patrol division. The average, as shown in Chart 3, shows a lull in activity between the hours of 3 am to 8 am and a peak activity time of 10 pm to 2 am. The workload depicted in Chart 3 is consistent with the majority of law en- forcement agencies across the country. The slow times between 3 am and 8 am is widely seen, as is the higher activity be- tween 10 am and 9 pm. The higher spike of activity between 10 pm and 2 am is consistent with a typical college town that sees a large amount of early bar activity and a younger demographic proﬁle within the community. The scale on the right side of Chart 3 represents the percentage of total work- load for the year under study during that particular hour. The axis on the bottom represents the hour of the day starting at 12:01 am and ending with 12:00 mid- night. The stafﬁng by hour of day for the current schedule is depicted in Chart 3 as a green line. During the 3 am to 8 am lull, stafﬁng remains high. When activity increases between 8 am and 8 pm, the stafﬁng falls under the workload curve causing a greater probability of satura- tion for the division. In most cases, it is impossible to ﬁnd a stafﬁng curve that will match the work- load curve exactly. The stafﬁng curve in Chart 3 is based on a weekly average for stafﬁng and for workload. While the stafﬁng curve does not match exactly, it appears to resemble the workload curve in it’s general shape. A greater picture is observed when the stafﬁng for each day of the week is plotted against the workload for that particular day. BOZEMAN, MONTANA POLICE DEPARTMENT - PATROL STAFFING AND DEPLOYMENT STUDY SPRING 2007 COPYRIGHT 2007 - ETICO SOLUTIONS INC. PAGE 15 OF 29 www.eticosolutions.com Sun 14% Mon 14% Tue 13% Wed 13% Thu 13% Fri 16% Sat 17% Chart 2. Workload by Day of Week 0 2 4 6 8 Midnight 3 am 6 am 9 am Noon 3 pm 6 pm 9 pm Stafﬁng Workload Chart 3. Percentage of Stafﬁng and Workload by Hour of Day Pe r ce n t a g e o f t o t a l c a l l s p e r h o u r o f d a y When the curve comparison is broken down by the day of the week, the inefﬁciencies of the current schedule become more ap- parent. The stafﬁng on Friday and Saturday appear appropriate. However, Monday through Thursday show great disparity be- tween the workload curve and the stafﬁng curve. In an ideal schedule, the two lines would appear as one. Less correlation be- tween the lines indicates less efﬁciency in stafﬁng. For instance, the stafﬁng and workload curves on Tuesday and Thursday only cross three times. The curves on Wednesday touch only twice through- out the day as the stafﬁng changes from overstafﬁng to understaff- ing and back. BOZEMAN, MONTANA POLICE DEPARTMENT - PATROL STAFFING AND DEPLOYMENT STUDY SPRING 2007 COPYRIGHT 2007 - ETICO SOLUTIONS INC. PAGE 16 OF 29 www.eticosolutions.com Workload Staffing Workload Staffing 0 1.75 3.50 5.25 7.00 12a 3a 6a 9a 12p 3p 6p 9p Monday Chart 4.b Workload Staffing 0 1.63 3.25 4.88 6.50 12a 3a 6a 9a 12p 3p 6p 9p Tuesday Chart 4.c Workload Staffing 0 1.75 3.50 5.25 7.00 12a 3a 6a 9a 12p 3p 6p 9p Wednesday Chart 4.d Workload Staffing 0 2 4 6 8 12a 3a 6a 9a 12p 3p 6p 9p Thursday Chart 4.e Workload Staffing 0 3.75 7.50 11.25 15.00 12a 3a 6a 9a 12p 3p 6p 9p FridayChart 4.f 0 3.75 7.50 11.25 15.00 12a 3a 6a 9a 12p 3p 6p 9p SaturdayChart 4.g Workload Staffing 0 2.5 5.0 7.5 10.0 12a 3a 6a 9a 12p 3p 6p 9p Sunday Chart 4.a R e c o m m e n d a t i o n s f o r S c h e d u l e M o d i f i c a t i o n s There are several alternatives for the current schedule. Some modiﬁcations may improve the efﬁciency of the patrol division deployment and some may be detrimental. A few characteristics ob- served from chart 2 and charts 4.a through 4.g provide an indication of the various types of schedules that may be used. Chart 2 shows that activity levels on Friday and Saturday are at 16% and 17% of the overall workload respectively. Tuesday, Wednesday, and Thursday rep- resent 13% of the overall workload each. As shown in charts 4.a - 4.g, the workload distribution by hour of the day is vastly different midweek than it is on the weekends. Wednesday’s activity is consistently high from 8 am to midnight with a short spike around 11 am. Fri- day and Saturday’s activities spike from 9 pm to 1 am and remain ﬂat through- out the day. Modiﬁcation of Existing Schedule The most promising option for schedule modiﬁcations is to remain with a ﬁxed day off schedule and realign the shift totals and days off. This modiﬁcation should put the stafﬁng curve closer to the workload curve since both will be based on the num- ber of hours of work per hour of the day. An analysis was done to deter- mine how much work is completed by each shift over a year’s time. The time worked during the overlap was split evenly among the 5p-3a shift and the 9p-7a shift. Once a percentage of total workload was obtained for each shift, the same percentage was used to divide the 25 available patrol ofﬁcers. This led to a realignment from 9 ofﬁ- cers to 10 ofﬁcers on day shift, and from 7 ofﬁcers on nights to 6 ofﬁcers on nights. Afternoon remained the same with 9 ofﬁcers assigned. After the initial shift assignments, the database was analyzed once more to deter mine how much work was created during each hour of the day for each individual shift. These percentages were then used to deter mine how many ofﬁcers should be scheduled per day to match the workload on a daily basis at a shift level. Since each ofﬁcer works four days per week, the number of ofﬁcers assigned to each shift was multiplied by four to de- termine the number of on-duty tours available for allocation across the seven days of the week. The data in Table 7 indicates how many ofﬁcers should be working each day on each shift to match the workload. One ﬁnal modiﬁcation was to re- evaluate the starting and stopping times for the three existing shifts. After modify- ing the days off and the shift realign- ments, the day shift stafﬁng levels were starting to rise too early and shifted the stafﬁng curve too far to the left. The starting and stopping times were changed for each shift as follows: Days: from 7a-5p to 8a-6p Afternoons: from 5p-3a to 6p-4a Nights: from 9p-7a to 10p-8a These ﬁnal modiﬁcations resulted in the stafﬁng curves shown in Charts 6.a through 6.g. BOZEMAN, MONTANA POLICE DEPARTMENT - PATROL STAFFING AND DEPLOYMENT STUDY SPRING 2007 COPYRIGHT 2007 - ETICO SOLUTIONS INC. PAGE 17 OF 29 www.eticosolutions.com Ta b l e 6 . R e c o m m e n d e d O ff i c e r s S c h e d u l e d P e r D a y S u n M o n Tu e We d T h u Fr i S at D a y s 5 6 5 5 6 7 6 A f t e r n o o n s 5 4 5 5 5 6 6 N i g h t s 4 3 3 3 3 3 5 Ta b l e 7 . M o d i f i e d S h i f t S c h e d u l e s f o r C u r r e n t 4 -o n 3 -o ff S c h e d u l e D a y S h i f t 8 a - 6 p A f t e r n o o n S h i f t 6 p - 4 a N i g h t S h i f t 1 0 p - 8 a Office r Na m e D a ys Off Office r Na m e D ay s O ff O ff ic e r N a me Da y s O ff Ofﬁcer 1 Sun, Mon, Tues Ofﬁcer 1 Sun, Mon, Tues Ofﬁcer 1 Sun, Mon, Tues Ofﬁcer 2 Mon, Tues, Wed Ofﬁcer 2 Sun, Mon, Tues Ofﬁcer 2 Sun, Mon, Tues Ofﬁcer 3 Tues, Wed, Thu Ofﬁcer 3 Mon, Tues, Wed Ofﬁcer 3 Mon, Tues, Wed Ofﬁcer 4 Tues, Wed, Thu Ofﬁcer 4 Tues, Wed, Thu Ofﬁcer 4 Thu, Fri, Sat Ofﬁcer 5 Tues, Wed, Thu Ofﬁcer 5 Wed, Thu, Fri Ofﬁcer 5 Thu, Fri, Sat Ofﬁcer 6 Wed, Thu, Fri Ofﬁcer 6 Wed, Thu, Fri Ofﬁcer 6 Thu, Fri, Sat Ofﬁcer 7 Fri, Sat, Sun Ofﬁcer 7 Thu, Fri, Sat Ofﬁcer 8 Fri, Sat, Sun Ofﬁcer 8 Sat, Sun, Mon Ofﬁcer 9 Sat, Sun, Mon Ofﬁcer 9 Sat, Sun, Mon Ofﬁcer 10 Sat, Sun, Mon Recommended Schedule Change Based on the data provided from the Bozeman PD CAD system and the analysis completed during this study, it is recom- mended that the schedule changes listed in Table 7 be initiated at the next shift bid opportunity. The correlation between stafﬁng and workload presented in Charts 5.a through 5.g creates the greatest stafﬁng efﬁciencies of any schedules reviewed. One additional beneﬁt of this ﬁnal option is the minimal effect it will have on the patrol ofﬁcers. Modifying the existing schedule by changing days off, shift assignments, and starting times is far less intrusive than switching shift lengths or shift types. BOZEMAN, MONTANA POLICE DEPARTMENT - PATROL STAFFING AND DEPLOYMENT STUDY SPRING 2007 COPYRIGHT 2007 - ETICO SOLUTIONS INC. PAGE 18 OF 29 www.eticosolutions.com Workload Staffing 0 2.5 5.0 7.5 10.0 12a 3a 6a 9a 12p 3p 6p 9p Sunday Chart 5.a Workload Staffing 0 1.75 3.50 5.25 7.00 12a 3a 6a 9a 12p 3p 6p 9p Monday Chart 5.b Workload Staffing 0 1.5 3.0 4.5 6.0 12a 3a 6a 9a 12p 3p 6p 9p Tuesday Chart 5.c Workload Staffing 0 1.75 3.50 5.25 7.00 12a 3a 6a 9a 12p 3p 6p 9p Wednesday Chart 5.d Workload Staffing 0 3.75 7.50 11.25 15.00 12a 3a 6a 9a 12p 3p 6p 9p FridayChart 5.f Workload Staffing 0 2 4 6 8 12a 3a 6a 9a 12p 3p 6p 9p Thursday Chart 5.e Workload Staffing 0 3.75 7.50 11.25 15.00 12a 3a 6a 9a 12p 3p 6p 9p Chart 5.g Saturday Unstructured Schedules Unstructured schedules, as described earlier in this section, provide great ﬂexi- bility for agencies. The administration would be able to staff proportional to workload and for planned special events. However, it would represent a major change from the current scheduling prac- tices. In the midst of a potential building project, the introduction of a designated beat plan, and a schedule change, it is not recommended that a new type of sched- ule, such unstructured, be introduced. 12-Hour Shifts The variety of workload from day to day indicates that rotating schedules with uniform stafﬁng by day of the week would likely be detrimental. Unlocked rotating schedules provide either uniform stafﬁng by day of the week or uncontrol- lable stafﬁng by day of the week where heavy or light stafﬁng levels ﬂoat across all days of the week. Some locked rotating schedules will provide proportional stafﬁng by day of the week but the ﬂexibility is usually lim- ited to only one heavy or light staffed day per week. One locked rotating schedule was examined and plotted out against the workload curve by day of the week. The locked rotating schedule that was examined was a 2-on 2-off 3-on 2-off 2-on 3-off duty cycle pattern utilizing a 12 hour shift length. The schedule is shown in Table 8. This schedule was designed for a full authorized staff of 33 ofﬁcers divided into 8 squads with one ﬂoating relief ofﬁcers. •Squad A, 4 ofﬁcers, 6a-6p •Squad B, 4 ofﬁcers, 6a-6p •Squad C, 4 ofﬁcers, 6p-6a •Squad D, 4 ofﬁcers, 6p-6a •Squad E, 4 ofﬁcers, 8a-8p •Squad F, 4 ofﬁcers, 8a-8p •Squad G, 4 ofﬁcers, 2p-2a •Squad H, 4 ofﬁcers, 2p-2a Squads A,C, E, and G all start in week 1 while the remaining squads start in week 2. At the end of each week, the ofﬁcers go to the alternate week and then start over again. Squads A-D provide basic coverage 24 hours a day, 7 days a week. Squads E-H provide supplemental stafﬁng during peak times. Squads A-D are self-relieving and work each other’s days off. Among squads A-D, only one of the four squads would be on-duty at a time. The popularity of 12-hour shifts in law enforcement is increasing across the country with the most prevalent use cur- rently in the northwestern states. 12- hour shifts provide complete unity of command and complete team integrity. However, with this beneﬁt it is easy for each shift to become their own clique. With four self-relieving shifts, the admini- stration must watch carefully to ensure that all supervisors are in step with each other. The agency cannot allow each squad to take on their own independent mission and slowly take the agency mis- sion off course. From an ofﬁcer perspective, every ofﬁcer on the patrol division would re- ceive every other Friday, Saturday, and Sunday off as a three day weekend. Complete schedule equity is provided and the three-day weekend every other week can be used as a strong recruiting tool. Ofﬁcers will work 182.5 days per year and be off 182.5 days per year. If two vacation days are taken during the second week of the duty cycle schedule, the ofﬁ- cer will get 7 consecutive days off. 12-hour shifts have a native average workweek of 42 hours as opposed to the current 10-hour shift with a 40-hour av- erage workweek. The additional 2 hours per week accumulates to an extra 104.28 hours per year per ofﬁcer. With 25 ofﬁ- cers on the patrol division, this adds an additional 2,607 hours of patrol time per year, the equivalent of an additional 1.19 full-time ofﬁcers. Charts 6.a through 6.g show how a 2-on 2-off 3-on 2-off 2-on 3-off 12-hour shift would match the call for service load by hour of day and day of the week. The stafﬁng to workload correlation appears to be equal or closer than the existing schedule for most days of the week. However, it is not as efﬁcient as Charts 5.a through 5.g where the current sched- ule was modiﬁed with new days off, new starting times, and a realignment of staff- ing per shift. BOZEMAN, MONTANA POLICE DEPARTMENT - PATROL STAFFING AND DEPLOYMENT STUDY SPRING 2007 COPYRIGHT 2007 - ETICO SOLUTIONS INC. PAGE 19 OF 29 www.eticosolutions.com Ta b l e 8 .M o n d a y Tu e s d a y We d n e s d a y T h u rs d a y F ri d a y S a t u rd a y S u n d a y We e k 1 Work Work Off Off Work Work Work We e k 2 Off Off Work Work Off Off Off As seen in Charts 6.a-g, the 12-hour shift provides a gradual increase in stafﬁng throughout the evening and then drops back at 8 pm. Since this is a locked rotating shift with uniform stafﬁng, the shape of the stafﬁng curve is the same across all days of the week. This shape is close to the workload curve in only a few cases and fails to provide the stafﬁng necessary in the late evening/early mornings on Thursdays, Fridays, and Saturdays. Neither this schedule nor any other rotating schedule is rec- ommended at the current time for the patrol division due to the current stafﬁng numbers and the workload patterns currently indicated by the CAD. Speciﬁcally, peak workload times during bar hours and increased early morning activities due to the college student population are not met by rotating schedules. BOZEMAN, MONTANA POLICE DEPARTMENT - PATROL STAFFING AND DEPLOYMENT STUDY SPRING 2007 COPYRIGHT 2007 - ETICO SOLUTIONS INC. PAGE 20 OF 29 www.eticosolutions.com 0 2.5 5.0 7.5 10.0 12a 3a 6a 9a 12p 3p 6p 9p Workload StaffingChart 6.a Sunday Workload Staffing 0 1.75 3.50 5.25 7.00 12a 3a 6a 9a 12p 3p 6p 9p Chart 6.b Monday Workload Staffing 0 1.63 3.25 4.88 6.50 12a 3a 6a 9a 12p 3p 6p 9p TuesdayChart 6.c Workload Staffing 0 1.75 3.50 5.25 7.00 12a 3a 6a 9a 12p 3p 6p 9p Wednesday Chart 6.d Workload Staffing 0 2 4 6 8 12a 3a 6a 9a 12p 3p 6p 9p Thursday Chart 6.e Workload Staffing 0 3.75 7.50 11.25 15.00 12a 3a 6a 9a 12p 3p 6p 9p Chart 6.f Friday Workload Staffing 0 3.75 7.50 11.25 15.00 12a 3a 6a 9a 12p 3p 6p 9p Saturday Chart 6.g IV. Beat Design M e t h o d o l o g y Through the process of patrol resource analysis, the optimal number of ofﬁcers for the patrol division has been determined. The shift deployment section of this report determined the most efﬁcient temporal deployment across three shifts and seven days of the week. The ﬁnal portion of this study will now focus on the geographic deployment of ofﬁcers across the city of Bozeman. Beat design, or the designation of smaller areas within the city to which one or more ofﬁcers will be assigned, is the last step in improving efﬁciencies within the patrol division. This com- pletes the process of “putting the right amount of ofﬁcers, in the right place, at the right time.” There are many methods of beat design ranging from very simple manual methods to sophisticated software programs. For this study, a process of “Equalized Beat Workload” was used. The same CAD data that was used for the analysis and deployment phases was analyzed one more time to deter mine where the patrol workload was being generated and in what amounts. The CAD output contained a data ﬁeld that speciﬁed one of 95 different “map grids” within the city of Bozeman. The map grids were a designation that was already in place by the City of Bozeman Emergency Services Team. A total workload was calculated for each map grid area based on the hours of workload completed in that area over the entire year under study. The total work in each area was divided by the combined workload for the entire jurisdiction. The product from this calculation represented the percentage of overall workload attributable to each individual area. Based on meetings with the BPD administration, a three-beat plan and a four-beat plan were created based on the workload percentages from each area. There were two main objectives as the various areas were being aggregated into the ﬁnal beats. The ﬁrst was to keep the beats as circular and compact as pos- sible. It is undesirable to have “long and nar- row” beats or to have “dog legs” that wrap one part of a beat around another beat. In order to maintain consistent drive times from one area of a beat to the other, the beat would ideally be a perfect circle. The second objective was to keep the total work- load for all beats as equal as possible to equalize response times across all areas of the city. One phenomenon that is often seen in beat plans is the “bur- den of centralized location.” If a three or four beat plan could be created with exactly the same estimated workload in each area, it would still not guarantee that the ofﬁcers in those three or four beats would have an equal workload. In most cases, the ofﬁcers in the central or interior beats (beats surrounded by other beats) will typically experience a higher workload than ofﬁcers in outer-lying beats that are tangent to the jurisdictional borders. This is due to various levels of cross-beat dispatching that takes place during peak call times. One way to defeat this phenomenon, is to recognize its po- tential presence and join the three or four beat plan in the center of the jurisdiction so each beat contains a corner quadrant. If additional ofﬁcers are available, a “rover” unit can be assigned to the center of the jurisdiction where the four main beats meet so that the rover can serve as a back-up unit or be dispatched to areas when the unit cars are unavailable. BOZEMAN, MONTANA POLICE DEPARTMENT - PATROL STAFFING AND DEPLOYMENT STUDY SPRING 2007 COPYRIGHT 2007 - ETICO SOLUTIONS INC. PAGE 21 OF 29 www.eticosolutions.com T h re e -B e a t P l a n The workload percentages were ﬁrst aggregated into a three- beat plan that would span the city. If workloads were equalized properly, each beat would contain one-third of the overall workload for the city. The map grid areas and their associated workload per- centages are shown in Table 9. The map grid aggregate produced ﬁnal beat workloads of 33.87%, 33.22% and 32.92% for the three newly created beats. The three beats meet in the center of the jurisdiction at the corner of Main Street and 11th Avenue. The highest workload areas in beats 1 and 2 lie adjacent to this intersection. The highest workload area in beat 3 is only one area to the east of this intersection. This creates an opportunity to lessen the effects of the burden of central location by assigning a “rover” unit to the center of the jurisdiction to overlap the heaviest areas in the three designated beats. The three designated beats are as compact as possible and the design avoids long narrow beats and “dog legs” that would wrap one beat around another. By assigning an ofﬁcer to each beat and re- questing that ofﬁcers remain in their beats during their proactive time, the units available for calls will be more equally spaced throughout the jurisdiction. This improves response times, improves the equality of patrol intervals throughout the jurisdiction, and pro- vides a more equitable probability of saturation across the commu- nity. BOZEMAN, MONTANA POLICE DEPARTMENT - PATROL STAFFING AND DEPLOYMENT STUDY SPRING 2007 COPYRIGHT 2007 - ETICO SOLUTIONS INC. PAGE 22 OF 29 www.eticosolutions.com Table 9. Workload Percentages for Three-Beat Plan Beat 1 Beat 2 Beat 3 Map Grid Workload Map Grid Workload Map Grid Workload 31 0.04%1 0.01%38 0.29% 32 0.27%2 0.00%39 0.05% 33 1.71%3 0.65%47 8.63% 40 0.14%4 0.03%48 2.92% 41 0.51%5 0.01%49 0.16% 42 1.05%6 0.45%56 3.49% 43 2.83%7 1.77%57 4.07% 44 3.34%8 0.00%58 0.63% 50 0.13%9 0.02%59 1.82% 51 0.72%10 0.05%60 0.42% 52 1.56%11 0.17%61 0.08% 53 3.67%12 0.10%68 1.34% 54 4.69%13 0.18%69 2.52% 55 8.15%14 0.06%70 0.24% 62 0.21%15 0.25%71 1.03% 63 0.19%16 0.07%72 0.00% 64 0.20%17 0.34%73 0.07% 65 0.20%18 0.40%77 0.18% 66 1.39%19 0.84%78 1.55% 67 0.46%20 0.04%79 0.47% 74 0.00%21 0.01%80 0.30% 75 0.83%22 0.00%81 0.00% 76 0.82%23 0.06%82 0.01% lg 0.76%24 0.21%83 0.84% 25 1.47%84 0.10% 26 0.61%85 0.16% 27 4.06%86 1.16% 28 0.99%87 0.21% 29 0.42%88 0.00% 30 0.06%89 0.08% 34 2.71%90 0.11% 35 1.00%91 0.00% 36 3.84%92 0.00% 37 2.04%93 0.00% 45 4.12%94 0.00% 46 6.14% 33.87%33.22%32.92% F o u r-B e a t p l a n The workload percentages were aggregated once again into a four-beat plan that would span the city. If workloads were equalized properly, each beat would contain one-fourth of the overall workload for the city. The map grid areas and their associated workload percentages are shown in Table 10. The map grid aggregate produced ﬁnal beat workloads of 25.63%, 24.96%, 26.44%, and 22.97% for the four newly cre- ated beats. The four beats meet in the center of the jurisdiction at the corner of Main Street and 11th Avenue. Once again, this creates an opportunity to lessen the effects of the burden of central location by assigning a “rover” unit to the center of the jurisdiction to overlap the heaviest areas in the four designated beats. The four designated beats are as compact as possible and the design avoids long narrow beats and “dog legs” that would wrap one beat around another. The geographic size of the four beats are not equal due to heavier activity in areas where the population density is higher than the average. BOZEMAN, MONTANA POLICE DEPARTMENT - PATROL STAFFING AND DEPLOYMENT STUDY SPRING 2007 COPYRIGHT 2007 - ETICO SOLUTIONS INC. PAGE 23 OF 29 www.eticosolutions.com Table 10. Workload Percentages for Four-Beat Plan Beat 1 Beat 2 Beat 3 Beat 4 Map Grid Work- load Map Grid Work- load Map Grid Work- load Map Grid Work- load 40 0.14%1 0.01%19 0.84%56 3.49% 41 0.51%2 0.00%20 0.04%57 4.07% 42 1.05%3 0.65%21 0.01%58 0.63% 43 2.83%4 0.03%28 0.99%59 1.82% 50 0.13%5 0.01%29 0.42%60 0.42% 51 0.72%6 0.45%30 0.06%61 0.08% 52 1.56%7 1.77%36 3.84%67 0.46% 53 3.67%8 0.00%37 2.04%68 1.34% 54 4.69%9 0.02%38 0.29%69 2.52% 55 8.15%10 0.05%39 0.05%70 0.24% 62 0.21%11 0.17%46 6.14%71 1.03% 63 0.19%12 0.10%47 8.63%72 0.00% 64 0.20%13 0.18%48 2.92%73 0.07% 65 0.20%14 0.06%49 0.16%75 0.83% 66 1.39%15 0.25%76 0.82% 74 0.00%16 0.07%77 0.18% 17 0.34%78 1.55% 18 0.40%79 0.47% 22 0.00%80 0.30% 23 0.06%81 0.00% 24 0.21%82 0.01% 25 1.47%83 0.84% 26 0.61%84 0.10% 27 4.06%85 0.16% 31 0.04%86 1.16% 32 0.27%87 0.21% 33 1.71%88 0.00% 34 2.71%89 0.08% 35 1.00%90 0.11% 44 3.34%91 0.00% 45 4.12%92 0.00% lg 0.76%93 0.00% 94 0.00% 25.63%24.96%26.44%22.97% V. Final Recommendations and Observations R e s o u rc e A n a l y s i s 1.Educate the patrol ofﬁcers on the importance of calling out all activities that are performed. If stafﬁng allows, encourage ofﬁcers to complete reports before clearing a call from the CAD. This should help determine average times per call in the next study. 2.Review all available CAD codes to eliminate “catch-all” categories that prevent accurate average times. Make an attempt to cre- ate whatever codes are necessary to properly categorize all activities performed by patrol. 3.Create a court log, if necessary, to track non-patrol time that prevents ofﬁcers from working their assigned patrol shifts. 4.Explore any options to accurately record police reports that are written after a call for service has been closed in the CAD. The method should tie the report back to the original CAD event number if at all possible. 5.Consider the use of a software package to track ofﬁcer leave times, overtime (FLSA vs. non-FLSA), court time, and net-comp time effects at the end of the year. Applications such as TeleStaff, In-Time Watch Commander, or Schedule Soft may be helpful in this endeavor. 6.Update call frequencies and average times from year to year to improve the reliability of the forecasting used in this study. R e s o u rc e D e p l o y m e n t 1.Initiate schedule modiﬁcations from Table 8 at the next shift bid opportunity. 2.Continue to monitor the interest in 12-hour shifts for both current ofﬁcer beneﬁts as well as a recruiting tool. 3.Once the data collection methods have been improved (through the recommendations listed above), recalculate the estimated workload for the three shifts being sure to use only dispatched calls for service. If self-initiated activities are used, shifts with higher levels of self-intrinsic motivation will continue to pull more ofﬁcers to their shift in a self-feeding cycle. B e a t D e s i g n 1.Initiate the 3-beat plan as soon as the CAD administrator can program the beats into the CAD system. The next shift bid may be a good opportunity for implementation. 2.Encourage ofﬁcers to remain in their beats as much as possible and to utilize their proactive time in their own neighborhoods. 3.As total available patrol ofﬁcers increase from existing average of 27, to future number in excess of 36, implement 4-beat plan. 4.Be vary cautious not to set hard barriers for the ofﬁcers concerning their beat boundaries. Allow some cross-over from time to time but encourage the staff to take ownership of their beat F u t u re S t u d i e s 1.As data collection methods improve, the average times for various calls may vary from this report. If the variation is great, it may require a new deter mination of beat designation, shifts assignments, and total stafﬁng. It is recommended that this study be updated annually as long as data collection methods are being actively improved. This study was conducted by Timothy J. Freesmeyer, of Etico Solutions, Inc. Questions or concerns may be brought to his attention via email at tim@eticosolutions.com, by phone at (w) 217-641-3205 or (m) 333-309-4906, or by U.S. mail at 524 E. Washington Street, Macomb, IL 61455. Statistical analysis of the CAD data was perfor med using SPSS software and exporting the results to Microsoft Excel. An electronic copy of this report is available to the Bozeman Police Department along with supporting Excel spreadsheets of all underlying data. BOZEMAN, MONTANA POLICE DEPARTMENT - PATROL STAFFING AND DEPLOYMENT STUDY SPRING 2007 COPYRIGHT 2007 - ETICO SOLUTIONS INC. PAGE 24 OF 29 www.eticosolutions.com VI. Addendum A At the conclusion of this study, the Bozeman Police Department received authorization for two additional ofﬁcers for the patrol divi- sion raising their number of patrol ofﬁcers to 27. This addendum was added as a recommended deployment plan for the existing 10-hour 4-on-3-off shift. The same CAD data that was used throughout this study was used once again to deter mine proportional stafﬁng by shift and day of the week for the new compliment of 27 ofﬁcers. Table A1 shows the recommended number of ofﬁcers to be scheduled on each shift assuming that the two newest ofﬁcers were as- signed to the night shift. Based on this deployment, 10 ofﬁcers would be assigned to dayshift, 9 ofﬁcers to afternoon shift, and 8 ofﬁcers to the night shift. New regularly scheduled days off were determined based on the call for service workload for each shift. Table A2 shows the recom- mended days off for each shift that would provide the closest correlation between stafﬁng by hour and workload by hour. When attempts are made to schedule consecutive days off that match a desired stafﬁng level per day, there will only be one solution. The days off con- tained in Table A2 are the only combination of days off that will provide the on-duty stafﬁng levels shown in Table A1. BOZEMAN, MONTANA POLICE DEPARTMENT - PATROL STAFFING AND DEPLOYMENT STUDY SPRING 2007 COPYRIGHT 2007 - ETICO SOLUTIONS INC. PAGE 25 OF 29 www.eticosolutions.com Ta b l e A 1 . R e c o m m e n d e d O ff i c e r s S c h e d u l e d P e r D a y S u n d a y M o n d a y Tu e s d a y We d n e s d ay T h u r s d a y Fr i d ay S a t u r d a y D ay s 5 6 5 5 6 7 6 A f t e r n o o n s 5 4 5 5 5 6 6 N i g h t s 5 3 4 4 4 5 7 Ta b l e A 2 . M o d i f i e d S h i f t S c h e d u l e s f o r C u r r e n t 4 -o n 3 -o ff S c h e d u l e D a y S h i f t 8 a - 6 p A f t e r n o o n S h i f t 6 p - 4 a N i g h t S h i f t 1 0 p - 8 a O ff i c er N a m e D ay s O ff O ff i c e r N a me D ay s O ff Off ic e r N a m e D ay s O ff Ofﬁcer 1 Sun, Mon, Tues Ofﬁcer 1 Sun, Mon, Tues Ofﬁcer 1 Sun, Mon, Tues Ofﬁcer 2 Mon, Tues, Wed Ofﬁcer 2 Sun, Mon, Tues Ofﬁcer 2 Sun, Mon, Tues Ofﬁcer 3 Tues, Wed, Thu Ofﬁcer 3 Mon, Tues, Wed Ofﬁcer 3 Mon, Tues, Wed Ofﬁcer 4 Tues, Wed, Thu Ofﬁcer 4 Tues, Wed, Thu Ofﬁcer 4 Mon, Tues, Wed Ofﬁcer 5 Tues, Wed, Thu Ofﬁcer 5 Wed, Thu, Fri Ofﬁcer 5 Wed, Thu, Fri Ofﬁcer 6 Wed, Thu, Fri Ofﬁcer 6 Wed, Thu, Fri Ofﬁcer 6 Wed, Thu, Fri Ofﬁcer 7 Fri, Sat, Sun Ofﬁcer 7 Thu, Fri, Sat Ofﬁcer 7 Thu, Fri, Sat Ofﬁcer 8 Fri, Sat, Sun Ofﬁcer 8 Sat, Sun, Mon Ofﬁcer 8 Sat, Sun, Mon Ofﬁcer 9 Sat, Sun, Mon Ofﬁcer 9 Sat, Sun, Mon Ofﬁcer 10 Sat, Sun, Mon Charts A1.a through A1.g represent the correlation between workload per hour and stafﬁng per hour based on the deployment recommended in Table A1. As in earlier charts, the stafﬁng on Friday and Saturday are the closest matches while Tuesday and Wednesday show the greatest deviations between the two curves. The beneﬁt to this deployment method is the increase in stafﬁng numbers for the midnight shift during early morning hours. This deployment leaves only one day (Monday) where stafﬁng drops to 3 ofﬁcers for a brief time from 4 am until 8 am. BOZEMAN, MONTANA POLICE DEPARTMENT - PATROL STAFFING AND DEPLOYMENT STUDY SPRING 2007 COPYRIGHT 2007 - ETICO SOLUTIONS INC. PAGE 26 OF 29 www.eticosolutions.com 0 2.5 5.0 7.5 10.0 12a 3a 6a 9a 12p 3p 6p 9p Workload StaffingChart A1.a Sunday 0 1.75 3.50 5.25 7.00 12a 3a 6a 9a 12p 3p 6p 9p Workload StaffingChart A1.b Monday 0 1.63 3.25 4.88 6.50 12a 3a 6a 9a 12p 3p 6p 9p Workload Staffing TuesdayChart A1.c 0 1.75 3.50 5.25 7.00 12a 3a 6a 9a 12p 3p 6p 9p Workload Staffing Wednesday Chart A1.d 0 2 4 6 8 12a 3a 6a 9a 12p 3p 6p 9p Workload Staffing Thursday Chart A1.e 0 3.75 7.50 11.25 15.00 12a 3a 6a 9a 12p 3p 6p 9p Workload StaffingChart A1.f Friday 0 3.75 7.50 11.25 15.00 12a 3a 6a 9a 12p 3p 6p 9p Workload Staffing Saturday Chart A1.g Table A3 shows the recommended number of ofﬁcers to be scheduled on each shift assuming that one of the new ofﬁcers is assigned to the night shift and the other new ofﬁcer to day shift. Based on this deployment, 11 ofﬁcers would be assigned to dayshift, 9 ofﬁcers to af- ternoon shift, and 7 ofﬁcers to the night shift. This assignment by shift is the closest proportional match to the calls for service. New regularly scheduled days off were determined based on the call for service workload for each shift. Table A4 shows the recom- mended days off for each shift that would provide the closest correlation between stafﬁng by hour and workload by hour. When attempts are made to schedule consecutive days off that match a desired stafﬁng level per day, there will only be one solution. The days off con- tained in Table A4 are the only combination of days off that will provide the on-duty stafﬁng levels shown in Table A3. This deployment recommendation calls for only three ofﬁcers to be schedule between 4 am and 8 am on Monday, Tuesday, and Wednesday. The agency is encouraged to strongly consider the need for a fourth ofﬁcer during these times as opposed to a closer correla- tion with workload throughout the remainder of the week. The agency is encouraged not to “staff for exception” but to staff where the greatest need exists and utilize overtime for the minimum stafﬁng exceptions. BOZEMAN, MONTANA POLICE DEPARTMENT - PATROL STAFFING AND DEPLOYMENT STUDY SPRING 2007 COPYRIGHT 2007 - ETICO SOLUTIONS INC. PAGE 27 OF 29 www.eticosolutions.com Ta b l e A 3 . R e c o m m e n d e d O ff i c e r s S c h e d u l e d P e r D a y S u n d a y M o n d a y Tu e s d a y We d n e s d ay T h u r s d a y Fr i d ay S a t u r d a y D ay s 5 7 6 6 7 7 6 A f t e r n o o n s 5 4 5 5 5 6 6 N i g h t s 5 3 3 3 4 4 6 Ta b l e A 4 . M o d i f i e d S h i f t S c h e d u l e s f o r C u r r e n t 4 -o n 3 -o ff S c h e d u l e D a y S h i f t 8 a - 6 p A f t e r n o o n S h i f t 6 p - 4 a N i g h t S h i f t 1 0 p - 8 a O ff i c er N a m e D ay s O ff O ff i c e r N a me D ay s O ff O ff i c e r N a me D ay s O ff Ofﬁcer 1 Sun, Mon, Tues Ofﬁcer 1 Sun, Mon, Tues Ofﬁcer 1 Sun, Mon, Tues Ofﬁcer 2 Mon, Tues, Wed Ofﬁcer 2 Sun, Mon, Tues Ofﬁcer 2 Sun, Mon, Tues Ofﬁcer 3 Tues, Wed, Thu Ofﬁcer 3 Mon, Tues, Wed Ofﬁcer 3 Mon, Tues, Wed Ofﬁcer 4 Tues, Wed, Thu Ofﬁcer 4 Tues, Wed, Thu Ofﬁcer 4 Mon, Tues, Wed Ofﬁcer 5 Tues, Wed, Thu Ofﬁcer 5 Wed, Thu, Fri Ofﬁcer 5 Wed, Thu, Fri Ofﬁcer 6 Wed, Thu, Fri Ofﬁcer 6 Wed, Thu, Fri Ofﬁcer 6 Wed, Thu, Fri Ofﬁcer 7 Fri, Sat, Sun Ofﬁcer 7 Thu, Fri, Sat Ofﬁcer 7 Thu, Fri, Sat Ofﬁcer 8 Fri, Sat, Sun Ofﬁcer 8 Sat, Sun, Mon Ofﬁcer 9 Fri, Sat, Sun Ofﬁcer 9 Sat, Sun, Mon Ofﬁcer 10 Sat, Sun, Mon Ofﬁcer 11 Sat, Sun, Mon Charts A2.a through A2.g represent the correlation between workload per hour and stafﬁng per hour based on the deployment recommended in Table A3. This deployment shows the closest correlation between calls for service and stafﬁng of any schedule reviewed throughout this report. As the number of ofﬁcers increases, the correlation be- tween the workload curve and the stafﬁng curve will continue to improve as long as shift assignments and days off are based pro- portionally on calls for service. This deployment is recommended over all other deployments previously discussed in this study based on the increased correla- tion in the curves. BOZEMAN, MONTANA POLICE DEPARTMENT - PATROL STAFFING AND DEPLOYMENT STUDY SPRING 2007 COPYRIGHT 2007 - ETICO SOLUTIONS INC. PAGE 28 OF 29 www.eticosolutions.com 0 2.5 5.0 7.5 10.0 12a 3a 6a 9a 12p 3p 6p 9p Workload StaffingChart A1.a Sunday 0 1.75 3.50 5.25 7.00 12a 3a 6a 9a 12p 3p 6p 9p Workload StaffingChart A1.b Monday 0 1.75 3.50 5.25 7.00 12a 3a 6a 9a 12p 3p 6p 9p Workload Staffing Wednesday Chart A1.d 0 2 4 6 8 12a 3a 6a 9a 12p 3p 6p 9p Workload Staffing Thursday Chart A1.e 0 3.75 7.50 11.25 15.00 12a 3a 6a 9a 12p 3p 6p 9p Workload Staffing Saturday Chart A1.g 0 3.75 7.50 11.25 15.00 12a 3a 6a 9a 12p 3p 6p 9p Workload StaffingChart A1.f Friday 0 1.5 3.0 4.5 6.0 12a 3a 6a 9a 12p 3p 6p 9p Workload Staffing TuesdayChart A1.c VII. References BOZEMAN, MONTANA POLICE DEPARTMENT - PATROL STAFFING AND DEPLOYMENT STUDY SPRING 2007 COPYRIGHT 2007 - ETICO SOLUTIONS INC. PAGE 29 OF 29 www.eticosolutions.com 1 Bozeman, Montana (MT) Detailed Proﬁle - relocation, real estate, travel, jobs, hospitals, schools, crime, news, sex offenders: Retrieved June 2, 2007 from the World Wide Web: http://www.city-data.com/city/Bozeman-Montana.html 2 About MSU @ Montana State University. Retrieved June 2, 2007 from the World Wide Web: http://www.montana.edu/misc/aboutmsu.php 3 U.S. Department of Justice Federal Bureau of Investigation. Crime in the United States 2005; Retrieved June 22, 2007 from the World Wide Web: http://www.fbi.gov/ucr/05cius/police/index.html 4 Shuiteman, J. (1985). The Police Chief (July). 5 McKinsey & Company. (2002) Improving NYPD Emergency Preparedness and Response; Retrieved July 31, 2007 from the World Wide Web: http://www.nyc.gov/html/nypd/pdf/nypdemergency.pdf 6 Government Accountability Ofﬁce. (2006) GAO-06-365R Preliminary Observations on Hurricane Response; Retrieved July 31, 2007 from the World Wide Web: http://www.gao.gov/cgi-bin/getrpt?GAO-06-365R 7 Wilkerson, B.. (2001) Global Business and Economic Roundtable On Addiction and Mental Health. Top 10 Sources of Workplace Stress. Retrieved July 31, 2007 from the World Wide Web: http://www.mentalhealthroundtable.ca/aug_round_pdfs/Top Ten Sources of Stress.pdf