You are on page 1of 260
PATROL STAFFING INSAN DIEGO ONE-ORTWO> OFFICER UNITS ihoet SYSTEM DEVEIOPMENT CORPORATION ‘The Police Foundation is a privately funded, independent, nonprofit organization estab- lished by the Ford Foundation in 1970 and dedicated to supporting: innovation and Improvement in policing: This book is published by the Foundation as a serviee. The ‘opinions and recommendations are those of the authors, and not necessarily those of the Police Foundation, System Development Corporation, or the San Diego Palice Department Copyright © 1977, Police Foundation, 1900 K Street, NW., Washington, D.C. 20006 Library of Congress Card No. 77-80188 FOREWORD ‘This research report represents the third joint effort of the San Diego Police Department, the System Development Corporation, and the Police Foundatign to develop ‘empirical data through repeatable measures, and to bring the resulting hard facts to bear fon a matter of major importance to policing. Barier we dealt with issues of Meld Interrogation and of testing a community oriented style of policing. "Now ote have examined an issue which to date has been characterized by emationally charged claims and counterclaims, but has generated litte, if anything, in the way of testable proof that would advance rational policy-making. The issue is whether to staff police patrol cara with one or two officers. The importance of this iaaue can hardly be exaggerated, given the increasing budgetary problems in many of the nation's cites. Putting two offcers in patrol unit can cost almost twice as much as using one officer in patrol unit. 0. W. Wilson, in Police Administration, asserted his convietion that an officer patrolling alone is more efficient, effective, and safe than two officers ina patrol ear. But Until now, Wilsor’s view has remained an assertion. Using a wide array of rigorously employed measures over a period of a year, the study documented in this report shows, atleast in San Diego, that clearly and unequivocally itis ‘more efficient, safer, and at least se effective for the police to staf patrol care with one officer. OF great importance to police administrators and municipal managers nationally is that every paice agency which has cllected or ean collet data can use the formulation land methods from this study to determine, by beat or watch If they lke, how one- and two- oflcer patrol unite compare in ther cities. ‘Most police departments can expect to have the same results as San Diego. And they can find out for themselves what conditions would have to be, oF how much conditions ‘would have to change, to justify staffing patrol units with two officers. ‘This study ie one of the most complex and difficult—operationally, technically, and sanalyticallythat the Police Foundation has yet sponsored. Chit William B, Kolender and the San Diego Police Department deserve great credit for holding experimental conditions {or a whole year, despite manpower stringencies thatthe department encountered during that year. Ie it only with the leadership of chiefs of police ike William Kolender that policing will be able to undertake research necessary for is advancement. Patrick V. Murphy Prosident Police Foundation PREFACE Over the years, the San Diego Police Department has faced the same dilemma as ‘evory other major police department in the United States: How to allocate sesourees and provide for an efficent and effective patrol force while maintaining the safest working Conditions possible for its officers. During the energy eriss of 1974 our research and analysis unit was attempting to {develop a patrol system that would lend itself to energy conservation, One suggestion as to eanvert from a basically one-offcer system to an entirely twolficer system, signif. cantly reducing the number of vehicles on the streets while keeping the number of personnel constant. The hoped-for result was fuel conservation. Some administrators Supported the plan, but others believed it tobe highly inefficent and a waste of personnel ‘At the same time, the staff ofthe Police Foundation was examining the issue of one. and twopperson patrol units. Tt initial study indicated that no reliable data existed that ‘would aesist a police administrator in making decision on the merits of one- oF two. person patrol units. Therefore the Foundation and the San Diego Police Department agreed that a research experiment would be the best way to resolve the issues. ‘The innues involved are extremely complex: not only the techincal anpects, but the ‘emotional impact upon the officers involved, Beats that had beon staffed by twootfcer lune for more than ten years were converted to one-oficer units for purposes of the sty. ‘To maintain the integrity of the study areas, personnel shortages required a shift of officers from other areas of the ety to staf the units Because tight control wes maintained, we believe that the project was a success and ‘thatthe reaulting study and report are a significant contribution to law enforcement. This report should prove to be » valuable instrument not only to the San Diego Police Department, but to departments throughout the nation. Te is only through this type of careful research that many of the questions that have troubled police administrators for years can begin to be answered. The San Diego Police Department is proud to have undertaken this study, and we are particularly pleased to hhave cosponsored the research with the Police Foundation, ‘Our thanks are expressed to Joseph H. Lewis and Richard A. Staufenberger of the ‘Foundation for their support, and tothe representatives of System Development Corpora ton, John B. Boydstun and Michael B. Sherry. 1 also want to extend my personal appreciation to the project staff, Captain Jim Kennedy, Ruben Rumbaut, Norman Borgen, and Donna Weaver. Without their tireless efforts, the support of ranking officers throughout the department, other erimina justice fagoncies in the county, and most important, the patrol officers who were the sgmifieant Ingredient in the project, this study would net have been possible. My thanks to each of you. W. B. Kolender Chief af Police ‘San Diego, California A NOTE ON EVALUATION Evaluation researchers and experimenters will find this study methodologically interesting. A frst in a diffiutt Held, It required considerable ingenuity on the part ofthe authors as wel as of thove in the San Diogo Police Department who participated in the esign, data collection and processing, and analyses that were done. The challenge was to surmount the measurement and ascription problems that the research staff encountered, "Two points are particulary interesting. The frst i that it was operationally impose. le to freeze officer assignments in the study units for the 12-month peried ofthe study. ‘Normal rotation and reassignment, which in San Diego traditionally has been relatively rapid, continued unchanged throughout the study. Although this presented the researc ers with difficult measurement problems for certain aspects of the study, the fact that a large number of ofieers partiipsted in each stafing mode, and that more than half of ‘them participated in both, rules out the possibility that effects could be the result of hhaving different kinds of officers in one-officer cars from those in twocficer ears. This ctor greatly strengthens the overall findings and conclusions ofthe study. "The second point is that the research staff applied two different methods af analysis to ‘the data of this study. Skilled researchers sometimes disagree about the independence of ‘observations such as some of those included in the data base for the research. Such disagreement leads, of cours, to questions as to whether one analytical method or anther {is more appropriate from the standpoint of yielding sound rerults and accurate insights into the meaning of the data. In this instance both incremental and MANOVA analyses ‘were used, and they yielded the same fundamental results, making such arguments moot. Bach method yielded some additional richness of detailed understanding ofthe data. It Is relatively rare in evaluative research that time and resources allow full Gocumentation of two analytical treatments of the same lange data set. Teachers of ‘methodology may find this report a useful addition to ther store of cates, Like any first attempt, this one undoubtedly ean be improved in the repeating. In research, it is @ truism to'claim that in eompetent reanalysis and repliation there is Strength. The Foundation continues to hope for both, Joseph H. Lewis Director of Evaluation Police Foundation ACKNOWLEDGMENTS Dr. Richard A. Staufenberger ofthe Police Foundation initiated the chain of events ‘tha led to thie study of patrol unit staffing. His personal efforts and professional interest in addressing the problems of police patrol operations were instrumental in making this study possible. ‘We wish to express special appreciation to Chief William B. Kolender ofthe San Diego Police Department, whose cooperation and support tustained the San Diego study of patrol unit stating rues, His department has ones agsin been responsive in providing an ‘operational seting suitable for the controlled study of complex issues facing the law ‘enforeement community. It is our experience that, too often, attempts at serious field ‘esearch of policing istes have not met expectations because of organizational infexiil- ity or unwillingness, when operational difficulties arise, to maintain the experimental ‘controls essential to rigorous research. Such has not been the case with the San Diego Police Department. ‘The department. project staff under the direction of Captsin James L. Kennedy, project supervisor, was the backbone of the day-to-day data collection effort during both fhe planning and the sady phase Their effort also were instrumental mn the formula tion of enter for selection of the study units. The department's staff, compoaed of two civilian research analysts, Ruben Rumbaut and Norman Borgen, made an essential contribution during the study phase by expediting, collecting, and aggreyating: massive Amounts of data. A fourth membor of the department's staf, Donna Weaver, provided Ctficient secretarial support in addition to maintaining comprehensive information fle ‘on each ofthe patrol units participating in the study. ‘We wish also to acknowledge the fine work of Oficer Pred England during the closing, ‘weeks ofthe study phase, He collected and sorted information and prepared 28 profiles of Critieal incidents involving study unit officers. ‘Recognition should also be piven to Professor J. Ward Keesling, Graduate School of Education, University of California at Los Angeles, who conducted the multivariate analysis of patrol anit staffing, and to Dr. Kathie S. Teilmann, University of Southem California Social Seience Research Center, who consulted on the analysis of the Officer Autitude Survey responses. "The members of the Police Foundation’s Evalustion Advisory Group EAG) provided ‘valuable assistance in their constructive eiteism and suggestions fr the design and later interpretation of the evaluation findings. The group includes Dr. Francine Rabinovitz, Dago FutdIsteroptirn G48) a Son Dhan Commanty Brahe iva we ne Deru: Som ‘School of Urban and Regional Planning, University of Southern Californ Alhert Relss, Jr, Department of Sociology, Yale University; Professor Lee Sechrest, Department of Peychology, Florida State University; and Professor Hans Zeisel, Univer: sity of Chicago Law School Finally, we wish to thank Joseph H. Lewis, Director of Evaluation for the Police Foundation, for providing us the opportunity to perform this study, and for his eantinued ‘encouragement and steadfast support of our efforts Jobo E. Boydston Boutuation Director Michael B. Shorny Principal Insestigntor Nicholas P. Molter Research Analyst " m v CONTENTS FOREWORD PREFACE, NOTE ON EVALUATION ACKNOWLEDGMENTS EXECUTIVE SUMMARY RESEARCH BACKGROUND, ISSUES, AND ENVIRONMENT EVALUATION OBJECTIVES, DESIGN, AND METHODOLOGY PATROL UNIT PERFORMANCE PATROL UNIT EFFICIENCY PATROL OFFICER SAFETY ATTITUDES ON PATROL UNIT STAFFING APPENDIXES LIST OF TABLES Chapter 2 chapter 3 4 6 16 w 8 1» 20 a 2 25 ” ‘Experimental Sample Groups Demographie Profile of Experimental Areas Cal for Service Data Profile of Experimental Aress Reported Crime Profle of Experimental Areas ‘Experimental Units by Watch Assignment Experimental Units by Beat Assignments and Ares of the City Experimental Unite by Previous Patrl Stating Comparisons of Actual and Expected Study Unit Staffing Comparison of Mean Watches per Officer ‘Synopsis of Study Data Resolution of Differences in CallforServiee Data from Dispatch Records ‘snd from Officer Journals Cell Classifiations for MANOVA Analysis Moan Total Calor Service Assignments by Unit Type and Wateh Comparison of Study Unit Assignments and CallsforService by Type of “Assignments Comparison of Primary Ascignments by Number of Officers Recommended ‘and Type of Study Unit CallsforService: Comparative Analysis of Assignments Based on Recom- ‘mendations Comparison of Officer and Unit Assignments With Complaint Operator ‘Recommendations Comparison of Differences in Proportion of Call for Service Assienments to ‘Study Units by Classification of Ineidents and Significance (Comparison of Differences in Proportion of Cal-for Service Assignment to ‘Single Study Units by Classification of Incidents and Significance CallforService Primary Assignments to Stady Unite by Incident Classif- ‘ations and Dispositions Calbfor Service Assignments to Single Study Unite by Incident Classifin- ‘tions and Dispositions Comparison of Primary Assignmenta by Type of Unit and Dispatch Priory (Comperisan of Calls Originating on Study Beats and Proportions Serviced ‘by Assigned Study Units ‘Comparison of Mean Primary Assignments On-Beat and Off Best (Comparison of Offcer-Initiated Patrol Activity Mean Rates for Study Units ‘Comparison of Disposition of Subjects Arrested by Study Units ‘Comparison of Adult Felony Arrest Changes by Type of Unit and Charge Disposition 10 1 2 2B 2 2 18 1% tv 28 Comparison of Adult Misdemeanor Arrest Charges by Type of Unit and ‘Charge Disposition st 29 Comparison of Juvenile Felony Arrest Charges by Type of Unit and Charge Disposition ” 30 Comparison of Juvenile Misdemeanor Arrest Charges by Type of Unit and ‘Charge Disposition 30 51 Mean Frequencies of SelfDispatched Cover Provided and Received per Unit by Watch and Study Group « 82 Comparison of Distribution of Complaints Against Officers by Watch, Area of City, and Provieus Staffing a 33 Comparizon of Citizen Complaint Tneidente Tnvalving Officers Working ‘Study Units by Type of Unit and Type of Complaint a 54 Comparison of Citizen Complaint Tnedents Involving Officers Working ‘Study Units by Type of Unit and Unit Activity a 85. Comparison of Citizen Complaints Against Officers Working Study Units by ‘Type of Unit and Disposition of Complaints 2 86 Explanation of Disposition of Citizen Complaints Against Police Personnel 2 37 Summary of the Incremental Analysis Results for Patrol Performance “3 BE Summary ofthe Multivariate Anslysis of Staffing Effects 4245 Chapter 39. Comparison of Primary Assignments to Study Units by Use of Backup a 4 Comparison of Total Requests for Backup Support 10-8ts) hy Type of Unit ‘and Wateh “ 41 Comparison of Requests for Backup Support (10-87s) by Type of Unit and ‘Preceding Activity 4 42 Comparison of Noed for Assigned Backup Units ” 45 Comparison of Overall Mean Response Times by Mix of Units Assigned to Calls for Service 80 44 Compariaon of Mean Insidont Reeponse Time to Calleforarvice by Mix ‘of Unite Assigned and by Incident Classification 5 45 Comparison of Mean OutotService Times on Calls-for Service by Mix of Units Assigned and by Tncident Classfiation se 48 Comparative Analysis of Unit Efficiency by Type of Unit i 47 Comparison of Sick Leave Hours Taken by Study Unit Officers 5 48 Comparison of Injury Leave Hours Taken by Study Unit Officers a 49 Study Beat Workload Ansiysis by Pereentage of CalleforService Recom- ‘mended for Assignment to One Officer 56 50. Study Beat Workload Analysis by Mean Calls and Service Times 58 51 Summary of the Analysis Results for Patrol Efficiency 59 Chapter 5 12 Comparison of Officer Safety Measures 2 59 Comparative Frequencies and Frequency Ratios of Study Unit Involve ‘ment in Critical Incidents a 54 Comparison of Critical Incidents by Unit Activity, Type of Critical Incident, ‘and Study Unit Staffing 6 55 Comparison of Projected Mesn Unit Exposure to Potentially Hazardous Patrol Activites 6 56 Comparison of Annual Resisting Arrests with Total Arrests Made by Study Unite co 51 Second Watch Comparison of Resisting Arrest Situations with Total Ar. Tests 65 58 Comparison of Critical Incident Injuries Involving Study Officers by Sever ity of Injury 6 59 Mean Critieal Incidents Involvement of Single Patrol Units a 60 Measures of Critical Incidents Involving Single Patrol Units by Type of Unit and Significance of Differences 61 Comparison of Single Unit Critical Incident: Injuries Involving: Study ‘Officers by Severity of Injury. 2 Comparison of Single Unit Critical Incidents by Unit Activity, Type of Critical Incident, and Study Unit Staffing 3 Comparison of Differences in Proportion of CalLfor-Service Assignments to ‘Single Study Units by Classification of Ineidents and Simifieance 64 Comparison of Police Vehile Accidents 65 Summary of the Incremental Analysis Results for Safety Chapter & (45 Mean Scores on Attitudinal Variables by Respondent Group and Test Period 61 Response Percentages on Atttudinal Variables by Indicated Preference ‘and Test Period for All Respondents (68 Consistently Significant Pearson Correlation Coofficents 69 Comparison of Patrol Staffing Preferences by Respondent’ Group and Survey Period TO Summary of the Analysis Results for Officer Attitudes Apvendis A ‘Al Popation of Areas Served and Ratio of Patol Officers to 1,000 Population ‘A2 Percentage of Patrol Deployed as One Officer Units Appondic B B-l Disposition Codes for Califor Service BE Distribation of Calls for Service Dispositions by Classifiation and Types of Tneidente Aprende © C1 Priorities for Calls-for-Service Appendie D Calls for Service Activities Callsfor Service Activities: Means for Significant Stafing Effect Variables (Criminal Classification Da D2 D3 D-& Criminal Classification: Means of Significant Stafing Variables DS Ds a (Criminal Clasaifistion: Means of Significant Watch Variables ‘Tralfe Classiseation: Multivariate Analy ‘D1 Traffle Classification: Means of Significant Stafing Variables D-8 Peacekeeping Ciassifeation D9 Poseskeeping Classification: Means of Significent Sttfing Variables D-10 Medical Clasifienton D-11 Medical Clasiieation: Means of Significant Stafing Variables DAZ Medical Classifeaton: Within-Cell Correlations D-12 Medieal Clasifieation: Univariate and Step-Down Tests for Staffing Effect Dd Security Checks DAS Security Checks: Means for Significant Staffing Variables DIG Operational Classification DAI? Operational Cassfieation: Mesns for Significant Watch Variables DAR Other Cassifietion D-19 Report Writing Classifieation D.20 Report Writing Cinssifietion: Means for Significant Staffing ect ‘Variables a 66 a 8 m0 3 " 1% a BA Bopd ec D2 be D4 Da Dé D6 Da Da bs Dao io pil Dit Da Daa Das Dis bat bat Das Das Dae Das Report Writing Classfestion: Means for Significant Watch Efect Variables [Report Writing Classification: Means for Significant Area Etfect Variables Ofer Initiated Activities, Officer Safety Ofer Safety: Means and Correlations Onicer Safety: Univariate and Step-Down Tests LIST OF FIGURES List of Figures 1 2 Sen Diego Police Department Patrol Unit Staffing Experimental Beats ‘Comparison of Offer Assignments to Stady Units pas p20 p20 Da Daa Das 18 u PATROL STAFFING IN SAN DIEGO OFFICER UNITS, EXECUTIVE SUMMARY PATROL MANAGEMENT ISSUBS ADDRESSED BY THE STUDY ‘The implications of the decision to use one officer instead of two in patrol units are nat trivial. For example, the San Diego Police Depart- rent spends $112,045 or 88 pereent more money each year to Feld 8 two-fficar unit on a 2¢hour basis than to fled a oneaficer unit. For single eight-hour watch period, the difference in cost ‘exeeeds $100 per unit. The San Diogo department tan field 18 one-officer units for lesa than the coat of 10 two officer units {In deciding how to staff and use patrol units, managers need to be able to predict how many. officers will be needed for any given situation, Dased on acceptable levels of unit performance, efficiency, and officer safety. Patrol units are ex: pected to provide a variety of law enforcement and community services, in dissimilar areas, and ‘under highly variable circumstances. This, man agers must address the question of patrol unit staffing in terms of specifi loeal expectations, areas, and circumstances—including specifically the time of day or wateh period. Each unit assizn- ment should meet acceptable levels of perform lance, efficiency, and safety, and requirements vary. If two afflers actually are needed to per form the majority of tasks required of a particular patrol unit, itis clearly more economical to put ‘them in the same car, rather than in two cars But if one officer can acceptably handle most of the expected tasks, it is cheaper fo put one officer Jn each ear and send two or more cars when backup assistance is required. "Because iti diffieultto predict the number of petrol officers required for en assignment, dopart- ment administrators most allow themselves a ‘mangin for error. A second patel officer in each car provides a built-in safeguard against underes: fimating the number of officers required, but it an extremely expensive one. Hoaving enough units to ingure a reterve and ‘operating an effective dispatch system for sending belt tnits from the reserve aso are saferuards against error. The auccost of this approsch de pends on careful evaluation of each eal for-aervice fo that the necessary number of unite can be ‘assigned initially, and on making & possible for assigned units to request assistance when they ned it. Because the frequency of cals for service land the number of officers needed per call are variable the numbers of beth assigned and avall able units algo vary. Systems for assigning prior- ity to the ealleforservice are used to help dis ppatchers allocate the units according to both emand and availabilty. These di Unit performance. Department administrators ‘ust consider them in making unit staffing dec sions, and in comparing unit performances on calle for service, ‘When units are not assgmed to calle-fr-serv. ice, the officers are expected to initiate patrol activities, The kinds of potentially valuable activ ties possible are mang, and available opportuni ties for each type vary. Patrol officers generally have a great deal of diseretion over the kinds of petrol duties they perform, and how much they fo, Inasmuch ae patrol unit staffing ean be ex pected to influence unit performance on officer. Initiated activities ae well at on ealLforservice assignments, patrol managers must consider the Importance they wish to give to these activities, as well as to any differences in performance by one And twootieer unite. For each kind of patrol activity. unit per forms, whether sasigned or seiFnitiated, there should ideally, be some measure of performance quality as well es quantity. Few such measures Ihave been developed. For this study, cal-forserv- fo results in torms of errests mado, formal crime) reports filed, medical emergency transportation provided, and other (no formal report) dispositions Served as measures of the quality of primary call Torservice assignments. Arrest subjects held to answer and arrest charges filed as requested were Used to assess the quality of arrests. Citizen com. plaints were used at an overall measure of the Diblie’s satisfaction with services provided, EMeieney of patrol units is commonly meas- tured by the use of backup units. The efficiency janes involved in patrol unit staffing also require ‘determination of the time and resouroes expended for the types and qualities of patrol unit activities employed. The mast direct approsch, and the one ‘used in this study, is to ealeulate unit cost for a standard interval of time and then measure the faversge amount of time unite are spending, on ‘each activity. Subeategories of duties (such a call. Torservice assignments to criminal calls resulting in arrests) can be used to include mensures of unit time and money expended for a specified level of quality. Tn addition to the type, quantity, quality, and cost of services performed by patrol units, patrol ‘management must consider officer safety in mal Ing unit staffing decisions. Because of the vara tions in patrol unit activity itis important to try ‘to measure unit exposure to potential safety haz fds as well as to tabulate the occurrences of assaults, Injuries, and accidents. Both exposure Itgelt and the results of expowure may be inf. ‘enced by unit staffing decisions. The hazards of fan assignment on a given beat and watch change fver time and need to he reviewed perioaeally, particularly with regard to decisions made on the bass of hazardous encounters that ooeurred dur ing past, particularly vicient times. The reputa- tion of a’partieuar unit assignment often outlives the actaal situation, Finally, management must consider the pret ferences and opinions of patrol officers in making patrol unit staffing decisions. In some eities,con= fract nogetitions already have established spe- cific requirements for patrol unit staffing, but in ‘most eases these decisions have not been based on. 4 thorough, joint analysis of the performance, efficiency, and offcer safety issues involved nor of the many’ contract options available for the same or lower cont DESIGN OF THE STUDY In order to address these management issues, the study design included « number of compart sons between groups of one- and twovofier units operating in similar areas under similar eondi- tlons. The fundamental assumption underlying the comparative design was that the substantial extra cost of twooicer unite would not be just fied for a given situation unless there were (avor able and important differences between one- and twocficer units in one or more of the following Unit Performance ‘The type, quantity, and quality of serv ews performed Unit Bficieney ‘The unit time and cost associated with comparable levels of performance. Osicer Safety ‘The rates of assaults on officers, resist. ing arrest situations, vehicle accidents, and officer injuries Teauling from com parable levels of unit exposure, Officer Atiudes "The preferences and opinions of as- sgned patrol officers ‘The research staff used a stratified sampling design considering patrol area, watch, and former staffing to select 44 patrol unite, Using the same Stratification criteria, the project staff assigned half ofthe seleeted units as oneofficer study units and the other half ac two-officer study units. Unit staffing (one or two officers) was controlled ‘hroughout the year ofthe study. and during that period the staif made a number of matched obser ‘vations (both sampled and continuous) ofthe two types of study units. The selected design permit- ted analysis ofthe overall difference between one- fand two.olicer units ab well as more specific ‘comparisons of differences by area, watch, former staffing, and assigned workloads. ‘The primary analysis was ineremental and ‘treated each ofthe many separate measures indi- idly. The anslysis proceeded from initial over. all comparisons of the two types of patrol units to progressively detailed comparisons under partieu- lar conditions and eiteamstances, A secondary form of analysis examined various combinations of information describing the units and their workloads. It attempted to identify and isolate the Influence of factors other than cnit staffing on observed differences in unit performance, eff tency, and officer safety. FINDINGS ‘The overall performance of patrol units in terms of the type and frequency of eall-for-service Activities and officer initiated activities proved to be about equal for units having one or two of cers. Twoolficer units produced more traffic cite tons, and there were minor variations in the types of calsforservice handled. Most of these latter variations were explained by the presence of petrol ambulances in the group of two.afieer unite Tn terme of the quality of service provided, calls serviced by primary one-officer units were ‘more likely to result in arrests and formal erime reports than wore those serviced by two-officer Units. Persons arrested by both types of units ‘were equally lkolyto be held to answer, and there ‘were signifiantly fewer citizen complaints result- {ng fom incidents handled by oneofficer tinits than from those handled by twooffcer units. BY these measures, the quslity of service of one- officer units was at least as good as and perhaps better than that provided by twoofficer unit, ‘The overall efficiency of one-officer patrot unite clearly exceeded that of unite staffed wi ‘bwo officers even though one-aficer units required ‘more backup support. The biggest difference in the use of backup between one: and twolficer tants was that two-ofheer units received signifi cantly more single unit assignments than did ane- officer units. The main reason for this situation, however, was that dispatchers had assigned two. officr units to 63.5 percent ofall alls for which a ‘tingle officer unit had been recommended, and hand assigned oneolfonr tnt with hacia to BR peroent of all alls recommended for neo officers. ‘As an indication that the level of backup provided in San Diego may have been excessive, ‘study unite reported that the backup support they received from other units was not needed in fully 50 percent of the eases. In contrast, only 28 percent of all assigned calls were underdis. Patched, that is, resulting in a unitinitiated re ‘quest for backup support. Even if only dispatches made as primary assignments to oneofficer units fre considered, the comparisons do not change ‘much—58.5 pereent of the backup received was ‘ot needed, and 6.0 percent of the total assigned calls resulted in uninitiated requests for backup. ‘Aithough the absolute differences were smal ig: nifeantly more of the total requests for backup support on allforservice assignments eame from ‘one-offcer units than from two-ticer units. ‘Tho study found that twooffieer units re quired less time to service ealls than did one- tfficer units, either singly oF in pairs, but the relative time saving in minutes per cal was not great enough overall to offset the eost-per-minute Aisadvantage of two-offier unite, The time and cost comparisons were even more unfavorable for throcfficer nits with regard to officer initiated activities completed during the time when units ‘were not involved in servicing alls. With the exception of traf citations, one officer units pre- duced equivalent amounts of officer initiated ac- tivity, and did go within a shorter period of avail able time and at substantilly lower cost than did tbrocfficer units One-ticer units were shown to have overall safety advantages over twoffcer units. With an ‘equivalent’ amount of estimated exporure, one- officer unite had less involvement in resisting arrest, and only equal invalvement in assaults on officers and in olfieer injuries, One. and two officer units had equivalent safety records with regard to vehicle accidents and exposure (mies driven) to vehicle accidents. Single oneoffcer tunts, which had a lower estimated exposure to potential hazards, also had significantly lees ite volvement than single twoofficer unite in assaults on officers and in resisting arvest situations, O8F- cer injuries were equivalent for both types of ‘ingle unite ‘Officer attitudes overall indicated slight proference for two-fficer units althoogh opinions Appeared to be neutral on which staffing mode was more advantageous with regard to perform nce, effcleney, and safety. However, those patrol officers who were assigned to study units were stronger in their preference for twoofTeer units fand saw them as providing advantages in per- formance efficieney, and safety. Patrol experience in a twoolficer unit and in Southenst San Diego (he patrol area considered to be most hazardous) ‘was shown to be correlated with preferences for ‘bwoofficer units and with opinions that they were more advantageous than one-fficer units. Age and rank oro also correlated with survey re sponses: Younger and lower-ranked officers Showed more support for twoofficer units, CONCLUSIONS AND RECOMMENDATIONS. On the bass ofthis study, two-ofcer regular patrol units do not appear to be justified in San Diego. Separate comparisons of unit performance, cfcency, and officer safety under current condi. tions all suggest that one-ofcer nit ate at least equal to and often more advantageous than two- officer units. Given the cost differential, which ‘enables 18 one-officer unite to be fielded for Is then the cost of 10 twoofficer units, it seems clear that one-officer nite should be the normal pat- tern in the ety ‘Police ambulance unite are an exception to this general conelusion bees the sestimption is that frequently two officers are needed to handle and transport medical emergeney patients. Spe- ial situations tactieal assignments, the need for field training, and other temporary conditions nay justify temporary staffing regular patrol ‘nits with two officers However, in San Diego, urrent patrol conditions and unit performances ‘would have to change markedly to justify contina- Ing to assign two officers tothe units sted, as matter of routine ‘Although one-officer staffing provides equsl unit performance with fewer officer complaints, freater efficiency, and greater officer safety than toes twoottcer staffing, San Diego police officers in general have a alight personal preference for ‘trooier stalling. Among thove officers assigned to study Units, this personal preference for bwo- offer staffing Is combined with a strong opinion that twoolficer units are more advantageous in terms of performance, efficiency, and safety. "Ths it appears the findings from this study may not match the preferences of San Diego patrol oficersparticulariy those. with exper- fences in twooicer unite and in Southeast San Diego. When patrol officers were asked to identify changes that might be justified by a shift of current twoofficer nits to one-officer units, ‘changes that presumably would make sucha shift Imre acceptable o them, the responses mos re ently suguested that additional personal protec: on in the form of thotgune or police dogs be ‘made available, Either or both of these options ‘ould te Implemented at lower cost than that of placing second officer in a patrol unit. Improved backup was the second most frequent responce Backup requests (10-879) averaged 1.2 minutes of dispatch delay and 60 minutes of unit response time—a total of 72 minutes to respond to a unit's request for backup support. The additional units made available by converting current twe-ficer Units to singleofficer units should help In etforts to reduce these delays "The study also found indications that exces- sive numbers of backup unite are dispatched with: fut request, chus reducing the number of Unite fvallable to respond to unit requests. One way to improve unit availability is to examine the cur rent practice of dispatchers of assigning, more aficers to cals than the complaint operator rec: fommends. Similarly, seltdispatching practices also should be reviewed. "To justify continuing to assign two officers to patrol units reguies that performance, efiieney, Band safety be made more competitive with that of ‘oneofficer units: There appears to be ample room for improvement in the productivity of two-lfeer tunts. On eallforservice assignments, dispatch ‘operators should make more efficient use of two- officer units by assigning them more of the total of ealls coming from outside their beat that re- ‘ire two offers, and fewer of the total of these calls that require only a single officer. Regarding offcerinitiated activities, would seem veacona. ble for patrol management to expect mone fom twootfcer unit than from 8 one-offiee® unit This ‘expectation should be independent of the specie nature of the offcernitiated activities. Th this study, the available measures of oficer initiated activity levels were those traditionally associated ‘with law enforcement: arrests, trafic itations, field intorrogations, citizen contacts, and so fort San Diego Police Department's Community Ori. ‘ontad Policing programs sugrests that other oF st least additional, measures of officer performance sre needed to Telect the degree to which patrol officers focus their attention and efforts on the specific problems of their assigned bests. Whit ever unit performance measures are eventually chocen, the unit staffing issue will remain. The justification Tor twootficer units will always de pend in part on the departments ability to demon Strate that they are a least as cost effective as oneoficer units Tn order to reduce the involvement of two- olfcer patrol unite in resisting arret situations may be desirable to conduct unit team training of Dalrs of officers in the handling of srrest subjects ‘reduction in resisting arrest occurrences wold rot only reduce potential hazards to police oficers but would increase the safety of arrest subjects, For other police departments, the San Diego study should provide a new point of reference for considering patrol unit staffing. The commonly wecepted assumption that twootficer Unite are Safer, more efficient, and more productive than ‘oneficer units ean no longer go unchallenged particulary in view of escalating labor costs. In San Diego these assumptions proved to be un founded. Although ne two cites nor police depart ‘ments are exactly alike, most are sufficiently simi Jar thet the patrol unit comparison methods used in San Diego ean bo readily applied. Tt seems likely that other cities may find ae Sen Diego did, ‘that one-ffier patrol units offer equal perform ance and increased officer eafety as well as su Stantial eos saving aver twosfficer units 1 RESEARCH BACKGROUND, ISSUES, AND ENVIRONMENT BACKGROUND Controversy and debate among police practi- toners regarding the overall subject of police patrol methods have persisted over the past 25 years, The issue of staffing a police ear with one tficer patrolling alone or with two officers work Jng'as 8 team has been a source of the eantro- versy, but empirical research on these two meth. ods of police patrol staffing has been lacking. When police departments first began using automobiles for patrol, they departed from the traditional practice of & single officer walking & beat and afsigned two officers to each ear. The rationale for this assignment pattern was that the petrol area was freatly enlanged by the use of Automobiles, and that the potential for danger to the single officor was thus also inereased; in addition, the large area made rapid cover astict- ance from other officers very difficult. The rising crime rates and a generally greater emphasis on ssurvellance and detection of criminal behavior also were considered to produce eanditions too demanding for singleofficer motor patrol! Over- al officer safety factors were suid to be the main reasons for lack of serious consideration of one- officer patrol units, But in 1958 the Kansas City, Missouri, Police Department. changed from the traultional two-fficer patrol car to 8 one-officer system, becoming the frst major metropolitan area police department to break that tradition ‘Since that time, police departments generally hhave relied on experiential judgments rather than on operations research oF experimental testing to fgude their deployment of patrol nits with one officer, two officers, ora combination THE ISSUES? ‘The ramifications of the patrol unit staffing debate are complex. Ironically, the arguments of patrol efficiency, effectiveness, and officer safety hhave been put forth by proponents of both one and twooffcer patrol A nationwide survey of 65 ‘major city police department patrol staffing pali- ‘es (conducted during the planning phase of this project by the San Diego Police Department, 1975) as well a2 formal interviews conducted by the project staff with SDPD personnel, and a review of the literature indicate a wide range of opinion on these matter. One Officer Patrol: Pro Proponents of one-officer patrol claim that this method is more efficient, permits grester area, coverage, Increases police vsiilty, and provides fuleker response time and a higher volume of police activity. In addition, advocates say that onesofieer patrol makes performance evalvation ‘and individual recognition easier for supervisors, ‘and that individual patrelling encourages respom sibility, alertness, motivation, and self-reliance One Officer Patrol: Con Critics of one officer patrol sugxest that single officers are more likely to avoud involvement in suspicious eases and are more easily bored or fatigued during a tour of duty. Both these factors, critics say, have an adverse effect on an officers ‘morale and effectiveness Two Officer Patrol: Pro Advocates of twoofficer patrol say that it provides built-in cover, eliminating the need to atl, Ree Beiction nd pntapea aeebay cree ‘pyeplan Parl Unita!” Poze Revel Becebar Bernard 6B Re Ay Srnec ep ng crt SEE Sarr ction mow ra Skelter deta Game ees SAvEuiary ofthe maton ervey rule ined ‘nove? . Alispateh cover units; inerenses officer safety and fauteness of observation; and leads to a more tffective, teamwork kind of polieng. In addition, ‘bro-lfcer patrol provides opporsunitis for train: ing, improves coordination and decisionmaking when two officers rust work together on some ‘tivity, and allows greater flexibility for report: writing. Two.Offcer Patrol: Con Opponents of twooicer patrol say that two oificers are more likely to become overconfident in hhazardous situations, and feel compelied to dem onstrate excessive baldness.' Most routine situa- tons do not require the presence of twe officers, land assigning two officers to them is a waste of personnel. Finaly, two officers may actually dis- tract each other during the eourse of their patrol, and credit or blame is diffieult for supervisors to "Amidst these conflicting interpretations, the need for research into the merits of one- xs. two officer patrol unite emenges as the principal area of agreement and a topic in need of definitive study. PATROL STAFFING IN SAN DIEGO" Before 1954, the San Diego Police Department ‘signed two officers to all patrol cars. Then, in 1964, after tho Kansas City, Missouri, Police Dex ‘partment had instituted a policy of assigning only fone officer to patrol units the chief of police in San Diego decided to follow suit, basing his dec sion on preliminary indications of inereased patrol ‘effcieney and eflectivencss. A transition period followed, during which, first, traffic unite were deployed as oneofficer units; then the patrol day shifts and finally, by 196%, the night patrol shits, ‘This one officer patrol aystem continued until 1964, when police administrators! concern over & ‘maried inerease in civil onrert in the Southeast ‘area of San Diogo led to an increase in police ‘deployment to that area, From that time until the patrol staffing study wae undertaken, all units in Southeast San Dingo, except those assigned to the day wateh, were staffed by two officers. (South feast San Diogo encompasses the highest concen ‘tration of minovty residents in the Central Patrol Division) By contrast, all other areas ofthe Cen tral Division, excluding police ambulances and special details, remained one-oficer unit, despite varying vates of assaults and other erimes in thoee areas. These considerations suggested that the police administration decision to staff an area with twoolficer units was strongly influenced by perception of community hostility toward the po- Tice and the police function. This perception eae to be institutionalized, accepted by the patrol ‘officers. In turn, this reinforeed the perceived ‘need for continuing the policy EVENTS PRECEDING THE STUDY For some time the Police Foundation had been interested in resolving some of the contro: versy surrounding one- and two-fficer patrol car stating. Tm spring of 1974, the Foundation con tacted the San Diego Police Department, whieh expressed an interest in obtaining fands for 8 ‘comprehensive etudy of the isu, ‘As a result of preliminary meetings and dis- cussions involving the San Diogo Police Depart- ‘ment, the Police Foundation, and System Develop- tent Corporation, the Foundation in October 1974 ‘approved and provided funding for a cooperative planning and design effor involving joint parti. pation of the San Diego Police Department staff land the evaluation staff of System Development Corporation. The purpose of this planning phase ‘was to study the current patrol environment trithin the San Diogo Police Department with the object of establishing an evaluation design and ‘methodology. The ultimate goal was to resolve as ‘many’ ns possible ofthe ines bn the controversy regarding unit performance, efficiency, and officer safety associated with using one. vs. twooffier patrol unite Tn the simmer of 1975, after many design revisions and earefilertial review by the Police Foundation and its Evaluation Advisory Group, the evaluation design was approved. In Septem” ber 10975 the Police Foundation's Board of Direc. tors approved the funding of a 12-month research study to determine differences in patrol perform ance, patrol efficiency, and officer safety between oneotfieer and twoofieer patrol units. inn, as Venton Boat P-t0 "OS “hnodare i, Shel, Tragitional Prevention Parl (Washington BC: RILECH Sune tro.” mney Pore ‘Paraphraed fom Sen Dig Pals Department Propoa. I EVALUATION OBJECTIVES, DESIGN, AND METHODOLOGY OBJECTIVES AND HYPOTHESES ‘The primary objective of this study was to evaluate the differential effects of one officer and twootfcer patrol unit staffing on police patrol operations. The study was designed to meastine objective and subjective outcomes (the dependent variables) of these two patrol staffing methods (the single independent variable) The asesement ‘of objective outcomes focused on patrol perform: lance, patrol efficiency, and officer safety. The assessment of subjective outcomes focused on the attitudes and poreeptions of police personnel with regard to one-fieet and two-ofver staffing. pesiGn ‘The goals of this field research project re- aquired careful selection of experiments! units combined with explicit plans for observing unit performance over the course of the std. Seletion of Bzperimental Areas and Unite In order to aavese the differential effects of oneoicer vs. tweoicer patrol units, many fac tors haul to be controlled to assure the differences in outeomes were attributable to the number of ‘officers in the cars, and not to extraneous sonrves of varianco, Among these factors were character. inties ofthe patrol area, police perceptions of these areas, previous patrol staffing poliey, and time of day Gefore the experiment, twooffieer unite oper. ated only at night in San Diego) "Even though the focus of this study was on the patrol uni, staff recognized that the environ: ‘ment in which the unit operates is of fundamental Importance, because it determines to a lange ex tent the number and kinds of problems that patrol ‘units encounter. Thus the type of area had been a primary factor in decisions on whether to field fone: o two-ffier units In onder to control for these determinants, 2 sample of beats was chosen fs the setting for this experiment, and units Assigned to thote beats became the focus of obser vation and analysis. This control was achieved through the specification of four types of expert ‘mental units described in Table 1. TABLE 1 EXPERIMENTAL SAMPLE GROUPS Group One-off: us prevausy ove-ficr Sess the sats qv) Group li Cneoffier unite in. previously two-ofcer 35 [an sxperimental contol Group I: Two-ofcor units in previously one-fficer sae [an experimental cana) Group W: Two-efear unit in previously two-tfice avons he satus qu) ‘This division of beats allowed for several im- portant comparitons. First, the oneiicer patrol Across all areas was compared with two-ffcer patrol aeroae all areas. A comparison of Group 1 And Group TIL units was used to assess differences in performance between one- and two-fficer units in areas of the eity perceived as being low-rise areas, and to provide a good measure of dilfer fences in effectiveness and efficiency between one ‘officer and two-lficer units when the concern for Safety was st a minimum, Group 11 units were compared with Group IV units to determine ‘whether there was any diferenee in performance between one-officer snd two-ffcer units in the areas of the sty perceived as posing a high risk Analysis of the data along these lines thas pro- vided a much broader hase of data for interpreta ton than stietly assigning units at random and comparing solely along the lines of eurrent stat ing modes, ‘Tho issue of sampling is crucial in statitiea! research, because the nature ofthe sample deter mines whether the results can be generally ap- in some eases simple random sam- particularly when the {otal population ia small'and relatively heteroge neous, and a simple random sample would be likely to exclude areas of potential importance fromm the experimental design. For this study, a Simple random selection of beats within the city ‘would present potentially serious research prob- Jems because: (1) assignment of two-tfier units, not oniy in San Diego, bat in many other police Gepartments a wel, is based on the socioeco ‘nomic or ethnie characteristics of a patrol beat: @) beats or areas of the city vary widely on such factors; (all beats in San Diego have been staffed consistently with either one-otficer or two officer unita for some years, making. previous patrol staffing poliey an important dimension (on the part ofboth the community andthe polio) to fontrol; and (@ the total number of beats in the Central Patrol Division ie #0 stall 69) that to ‘sample randomly decreases the effcieney of the ‘sample from a statistical standpoint Based on these conditions the sample of units selected for this study was’ chosen through a stratified sampling procedure designed to contrel for the determinants outlined above. The frst stratifieation is thus the previous staffing policy of the beats. Ths stratification assigns every beatin the Central Patrol Division to one of the four ‘areas previously described in Table 1 "The next problem was to choose, from within these four groups of beats, those beats to be Stalled with ether one- oF twolficer experimen tal units, Beats were carefully evaluated on a ‘ariety of soci eonomie and demographic vari bles using a Sori Ares Analysis of San Diego. ‘This analysis consisted of factor analyses of 128 census indicators, The three factors under this analysis which best predict all other factors are ethnicity (percentage of minority resents, socio ‘cconomie rents (2 combination of housing value, jneome, and eduention), and faraly status (size of family units and peroentage of families with chil den) Using beat rankings on these demographic tlimensionsy as well aa on various measures o SINS haar etal Ca ‘Stat ute ecumsne 15. (eS Wheto chewed TABLE 2 DEMOGRAPHIC PROFILE OF EXPERIMENTAL AREAS 12970 Consus Data) sourweast SaN o1tco CAST SAN DIEGO, PREVIOUSLY 2-0 AREAS [WATCH 2 & 3} | PREVIOUSLY 1-0 AREAS (WATCH 2 & 3) beMocRAPHC care. | EXPERIMENTAL | EXPERIMENTAL | EXPERIMENTAL | EXPERIMENTAL CHARACTERISTICS OC SUNITS 129 UNITS DUNS SO UNS, (re 6 beets (2 beat) | tn > 2 boats) We UNTS suns Suits unt a5 | numuer se | Numbers | Number % PoruLaTion s7ees i000 | saa 1000 | 193% 1000 | 17400 1000, Ge DSTRBUTION O8 sen 155 | seas | tse az | gas ez eas tea) yen te | 200108 | tas 107 35 gan ‘ee | 39a '20 | “aor ‘az | ae 28 ie19 ne 7a | dew 8 | ass bo | mesa 20.38 jam ata | tees 20a | 200 oat | az mas dea bece iar | ates © te | Sar tat | 2190 ta 30 ane older Beet tse | sven 50 | Barr go7 | ae 37 ace ‘we race a4 | ros 6 | tare 902 | tgm8 880 ‘Spanien Ameriean fas me | ‘Sasa dor | ‘Vs “ag | “zoe Bloc wae 3 | sao as tat! 2 (03 Omer 3s “a7 | “aa = “sa | tz 18 | oes TOTAL go 10 | ayer r0 | 5538 1000 | 5.122 tooo Female Head Zoo aa | tyes "208 | are tga | ana Below Poverty tee 73] is a | "50 | a Avorege Saw ‘20 “20 20 27 MEDIAN INCOME sun $928 suds! seot? EDUCATION IMeclan Yeats) ns 10s wes 3 TABLE CALL-FOR-SERVICE DATA PROFIL a OF EXPERIMENTAL AREAS iSmonth totals) January-June 1975 _ ‘SOUTHEAST SaN DIEGO EAST SAN DIEGO PREVIOUSLY 2-0 AREAS PREVIOUSLY 1-0 AREAS. osearcy | WATCH 2 & a) Watch 2 & Dara BY Bepeniventar |” expenMeENTAL | EXPERIMENTAL | EXPERIMENTAL warcis OUNTS "0 UNS, TOUNTS. PEnioos (o> bau i= Bote (ne 2 be "We uNtTs "we UNS suns roa OK ea. | tou OR oe Tou OR sa. fan caus 783 7042 «1a | 7704 240 459] 2107 1758 676] 2508 22 «2a waa seat | 2186 620 08 | 2218 6.6 162) ox so2 94) 705 588 139 ware rirw | 3529 S80 268] 3721 1033 211) 969 909 m72| 1101 o34 sas (tem to7au) | 1040 455 98 | 1765 510 n14| 542 452 208| 708 sao 146 YNETS DisArcHED | rosa 2714 S48 | 1iiz01 satus 744) aa60 2724 1043 | 068 253) bed Fawtoara) | 3267 935 160| sas1 ors 270| asa 711 164| 1059 eee 189 watch"? wirrmrotten | soe r362 225] 5175 1437 26] 1502 1252 «24/1701 19s 18 Giem rz ami | 2999 686 139 | 2655 737 177| 94 762 455/129 1033 229 ELAPSED INCIDENT reurs 3279 e57 736 | ase 927 207| 9s 789 257| 1005 see 200 Taw to arm) so 267 35 | 1002 278 64| 20 247 €1| 24 200 70 watch 2 wiggagrettrm | 629 453 167 | 1642 asa 107] a4 388 113 | 482 002 104 (iivmrovam | 680 192 36 | 009 102 44] 200 175 ea| 20 198 56 = Mean per month per best Fa ~ sundord dewaton reported crimes on the beats, beats within “prev ously one-officer” and “previously two-aficer” Areas were matched to one another, and assigned equally to either the one-fficer staifing condition ‘oF the twooffcer staffing condition. Comparisons ofthe selected beate are in Tables 2,8, and 4 ‘inanclal considerations made it impossible to ‘choose a saturation sample that would include all the beats in the city, thus eliminating sampling problems. Because the number of unite for sidy hhad to be limited, the decision was that, for the purposes ofthis study, particular attention would be focused on units operating in the controversial area of current twofficer staffing, and that the majority would be assigned equally to either one officer or two officers in these areas. For purposes ‘of control, smaller numberof units in the “previ- ‘ously one-officer area” were chosen to be staffed ‘vith ether one officer er two offers. The mateh- ing procedure already described shows that these u control beats were representative of the rest of the city In summary, to est the proposals put forth in the study design, units were assigned equally in the one- and two-offier conditions, acsording to watch, area of ety, and previous stalfing policy, as Indicated in Tables 5, and 7. [Figure 1 shows the selected study beat loee- tions in relation to other patrol areas inthe city of San Diego. Selection of Officers ‘The Central Patrol Division of the San Diego Police Department had a total of $10 patrol off cers operating a daily average (all watches com bined of 141 patrol unit, Inasmuch as the stady design called for controling the staffing of 44 Units, half of which would result in changes frorn previous staffing patterns, the impact on person ‘hel assignments was potentially serious, Further more, the personnel paiey then in effect allowed TABLE 4 REPORTED CRIME PROFILE OF EXPERIMENTAL AREAS “aly 1974-June 1975 (12-month totals) ‘SOUTHEAST SAN DIEGO EAST SAN DIEGO PReviOUSLY 2-0 AREAS previguscy 0 AREAS |___Mwarenes 2 ano 3) WATCHES 2 AND 3) came EXPERIMENTAL | EXPERIMENTAL | EXPERIMENTAL | EXPERIMENTAL “we "20 UNS, ‘oUNTS “ZO ONT COUNTS, (n= 8beat) (02 S008) (eo 2 beat) (n= 2 boats) ‘eunms ‘euxis eure urs root RS aa [rout RS a. | tors OR 9 | tomtOR ad opbery ‘hea se 19 os | ua 20 136] 1 om om | 2 om 04 Stora Arm sor 1 ogo “99 33 Ose] "5 O21 ote] 13 ost oe urgary endantt rao | za 129 928] 252 195 ase] zm 9 229 Commerc tes | ‘aor "3 2a] ‘eo [2s tie] “ely ie betty The isp | sane 12 ina | 68 238 Lie] so ze 78 ‘ra Ts tes | "ies "s7 Gas | te “et oze| “7s “tle ASsoutBecery aha| He ae tos| ‘e226 los] BF 24 Ges Fave Gao | Se os ont| “som 00s) “e O3s ore Tors [sess S| yer 683 | not as = mean por month pa bet Ba standard devon TABLES TABLE 7 EXPERIMENTAL UNITS BY WATCH ASSIGNMENT EXPERIMENTAL UNITS BY PREVIOUS PATROL GCOS PTRINENTAL UNTTS N= aa STAFFING [EXPERIMENTAL UNITS N= 48 Ste patnowwarcies {“orecmneer | Tweicer” pre wous patna. [EXPERIMENTAL UNTTS N= 44 Unis Unis SrASEING "| One-otcer | Two-cer Fiat wach Funte [7 ante —_| Unite Unis sera Previous One Oreer | ttunite | 17 unt ‘Second Watch Tunis eur see ‘ruta Previous Two-Oficer | tunis | 17 une Thre waten Tons [ Fanta ‘San irae? a TOTAL Zune | 22 one TOTAL Bunks | 22 uni ‘iicer senfority and personal preference to deter- thine changes in unit (beat and watch) assign- TABLE 6 iments, To “freeze” assignments for the oneyear EXPERIMENTAL UNITS BY BEAT ASSIGNMENTS "AND AREA OF THE CITY. [EXPERIMENTAL UNITS N= «8 AREA OF THE CITY ‘One-Offear | Two.0¥feor Soutneart, Gees | Shea Sindee séunte_| _1eunte Zeus | Zea Sen Diego inne Sone ToraL 3 beat ‘8 beate 2 ‘experimental period would have been disruptive of staffing practices then in effect, and would have ‘created an artifielal condition during the observa- tion of unit performance. Based on these considerations, and on the fact that the design focused on units rather than fon the officers themselves, the decision was made hot to control individual officer assignments. Tn the absence of specifi experimental con- trols on the assignment of individual officers, the reader may question the extent to which observed Unit differences can logically be attributed to Staffing by one or two officers because of the BRREPSRETER EET Z Ee ease wouso Twnainion 3a ‘ONLAVS NN 40 34AL A TOULVG NO SOOI4 HOLM 40 WABANN SLINN AGNLS OL SLNSNINDISSY ¥OHHO 40 NostuvAAIOD z aunou “ TABLE & COMPARISONS OF ACTUAL AND EXPECTED* STUDY UNIT STAFFING MEAN OFFICERS PER UNIT ssTUDY UNITS _ EXPECTED ACTUAL (ne Offcor Unie 167 vast Swo-Oftes Units 3 28 4 Expected if ssignment hd beon rozen MEAN WATCH PERIODS TOTAL INDIVIDUAL PER OFFICER OFFICERS INVOLVED EXPECTED. ACTUAL EXPECTED." ACTUAL 50.30 561 u 20% 50 sor a 20 ‘Inetudes 26 wine wanted exclcivelym one ofcor Units and 174 who worked ia both typeof study uit “Includes 108 who worked exzusvyn worotieer and 4 who worked In th types of ut uns possibility that any observed differences might reflect only differences in the performance of two {roupe of individual officers. To address this que tion, research staff made a stady of the duty assignments of individual patrol officers during the 12 one-week sample periods, A total of 66 officers were required to staff the study units for exch 2¢hour period (22 for the 22 fone-officer units plus 44 for the 22 two-officer Unita. Using the department's 167 multiplier to accommodate days off, vacation, sick leave, ete, approximately 110 officers could have been ex pected to staff the experimental units during the Study. OF the $70 patrol officers in San Diego's Central Patvol Division, 306 (82.7 percent) had served one or more duty shifts in one of the 4 Selected experimental units “Two hundred officers had operated in one ur more of the 22 eneotficer study units, with the inean length of assignment per officer 88 days, land the median four days; 280 officers had assign” iments to one or more of the 22 twootfcer study Unita. The mean length of assignment per officer was 126 days and the median was six days. Figure 2 compares the number of officers and ‘watches per officer spent in each type of study ‘it Furthermore, 174 officers (689 percent) of those assigned to study units had had assign rents in both one-fficer and two~icer study tunis, A total of 106 officers (846 percent) worked exclusively in twoofficer study units, while 26 officers (8 percent) worked exclusively’ in one- Cffcer study tnits, Stated differently, 87 percent of the officers who worked in one~ficer study Units alao worked in two-officer study units, and {2 percent of those who worked in two-fficer study unite also worked in one-officer stady units ‘during the course af the study. ‘Because this study focused on patrol units, the most important measures of officer mobility trore those at the specie beat and watch period ‘signment level (A “unit” as used throughout 6 this report, i a stafed patrol vehicle assigned toa specific patrl beat during a speciie watch period) ‘The average one-offieer unit during the course of this study had 1491 different patrol officers. The median number of officers per unit ‘was 15. The range was 9 to 22 officers yer unit. ‘The average number of watches these individual offers worked was 5.61, and the median number of watches per afler in single unit was L The ‘ange was 1 to 50 watches per officer “The average twooicer unit had 27.68 difer- cent officers, with a median of 27 officers per unit ‘and a range of 21 to 34. The average number of ‘watches worked by these officers was 617; the Imedian number of watches per officer in a single ik was 2 and the range was 1 to 6 watches per ‘Table 8 compares these measures of actual staffing with those that would have been expected ‘ftunit assignments had been frozen during the experimental period, Also included is a ttest of the difference In mean watches per officer which showod no significant difference between one- and ‘wocflicer unite (Table 9) ‘There was large numberof officers involved, ‘more than half of those involved worked in both types of study units, and the average officer assigned to a one-officer unit accounted for only '59 (6:7 percent) ofthe 84 wateh periods studied for TABLE 9 COMPARISON OF MEAN WATCHES PER OFFICER ‘WATCHES PER OFFICERS stupvuNits “DURING STUDY PERIOD (One-Ofieer Unis Set 142 (ees Two-Otficer Units 607 ose R= Mean cach unit, while the average officer assigned to a ‘peo-offcer unit accounted for only 607 (72 per cent) ofthe Bl watch periods studied foreach unit. ‘Thus, it does not appear that individual offier Jifferences eguld be expected to have had a major influence on unit performance as measured dur Ing this study. Although the instability of patrol officer as- siinments in San Diego may pose some opera- tional and policy questions, it does enhance this study by inereasing the confidence that observed differences between one- and twro-fficer nits sre sttributable to the number of officers assigned t0 ‘8 unit Tn fact one of the very strong points of this ftady is that despite the relatively small number of patrol unite involved, the extreme diveraty in ‘conditions under which the unite were fielded, that isin diferent areas of the city, with all Unree ‘watches staffed by a very lange and diverse group of officers riding in many different patterns, itis ‘extremely improbable that variables other than the stafing pattern could have accounted for the remarkably consistent rerults in this study ‘Essentially all of the Central Patrol Division was involvad at some time during the study; the following race and sex distributions of central patrol officers are provided for general informa- ‘on, but were not considered in any ofthe subse- quent analysis. Male, white: 89.2 percent Male, black: 88 percent ‘Male; Mexican-American: 1 percent [Males, other minorities: 1.6 pereent Females, al: 73 percent Data Sampling Approach Before the study phase began, staff adopted a ‘data sampling approach to ease some of the data collection activities. Sampling was not applied to all data sources but only to the following, high volume types of data: © patrol officer daily activity reports, ‘which required manual data extraction and coding; ‘© patrol dispatch data, which required fextonsive computer editing and aggre- gation. Based on projections made during the planning phase, the data colletions of the above items for feach ofthe 44 study units would have ereated an “unmanageable eolection problem and an unneces sarily complicated analysis problem. ‘The adopted sampling, approach called for selecting 12 sevenday periods distributed uni ormly over the 12-month study phase, The first 16 period was derived by randomly selecting a Sun- Gay start date and the next si consecutive days from a four-week base period beginning October 26, 1975, and ending November 22, 1975. The 1 ‘additional sampling periods were established by Using a four-week skip interval Limitations of the Experimental Desion In addition to the limitations on sample sizes and offeer assignments previously mentioned, it should be noted that no controls were placed on the assigned workloads (type or amount) of study ‘units. For example, sx combination patrol and ambulance unite were included in the selected {roup of 2 two-tier units, These units provided Public emergency ambulance aerviees—an addi- onal role not performed by one-officer unts. ‘Thus the analysis examined performance under measured workload differences rather than com ‘tolling for equal workload exposure. Interprets ton of observed differences in these cireurstan- tes is more difficult. This point is relevant particularly with regard to unit assignments ‘nade by dispatchers in response to call-fr-serv fee. To offaet the lack of experimental controls necessary to compare performance directly, provi- sions were made to collect dispatch records and analyze dispatch decsions with regard tothe type dnd frequency of assignments made to one-fficer ‘and twoofficer units “Another limitation inthe design was the ca ure to examine the relative amount of nit time devoted to individual types of office initiated ae tivities. In some cases (a8 frequently happens in Seld research) the analysis of collected data posed fasues which could not be adequately addressed with the available aggregated data. At a high level of detail, small data cell sizes and skewed distributions prohibited mesningfal testing in ‘some instances. This problem ie pointed out wher- fever occurs in the text Major Variables and Selected Measures ‘The evaluation design called fora eomprehen- sive set of objective and subjective measures of fone and twooificer patrol work from four tent perspectives: unit performance, unit eff- ciency, officer safety, and officer attitudes. Barly {in the planning phase i beeame apparent that no combination of date currently in use by the de- partment would be suficiently comprehensive to Compare the two staffing modes in a thorough manner. Therefore, several projectspecifie data callection plans were developed for use in corjune- tion with existing patrol activity measures. See ‘Table 10 for a summary of study data

You might also like