JOURNALOF CRIMINALJUSTICE

PergamonJournal of Criminal Justice 30 (2002) Sl-64

Robin Shepard Engel *

Crime. Law and .lILftice Program. Department of Sociology, The Pennsylvania Stale Ulti~ity. /0/ J Oswald Towel;U,.;~ity Park. PA /6802. USA

Abstract

Scholars have speculated that in an era of community policing, the role of first-line patrol supervisors haschanged. Given this change, can patrol supervisors effectively influence their officers' behavior? This question isaddressed by examining data collected for the Project on Policing Neighborhoods (POPN), a systematicobservational study of patrol officers and first-line supervisors in two metropolitan police departments in 1996-1997. Using four distinct supervisory styles created through factor analysis of attitudinal constructs identified inthe management and policing literatures, the influence of different supervisory styles over the time officers spendper shift conducting particular types of activities and encounters is assessed. Findings from multivariate analysesshow that officers with "active" supervisors spent significantly more time per shift engaging in self-initiated andcommunity-policing/problem-solving activities, while officers with "innovative" supervisors spent significantlymore time engaging in administrative tasks. Implications for policy and future research are explored. 0 2002Elsevier Science Ltd. All rights reserved

Introduction

Many scholars have described the changing philo-sophies, goals, and strategies of modem Americanpolice departments. Although considerable debateexists regarding the definition, implementation, andeffectiveness of community-oriented policing, it isclear that many deparbnents are implementing innov-ative strategies to achieve their goals of buildingcommunity relations and attending to quality of lifeissues, minor crimes, and public disoroen, (Bayley,1994; Goldstein, 1990; Kelling & Moore, 1988;Masb'ofski, Worden, & Snipes, 1995; Skolnick &Bayley, 1986). It has been suggested that thesechanges in police organizations must be accompaniedby changes in the roles of both patrol officers andfirst-line supervisors. Although changes in the roles

and activities of patrol officers have received con-siderable research attention, little is known aboutchanges in first-line supervision. Despite scholars'speculations regarding the changes in supervisors'roles, few empirical studies have explored thesechanges. Furthennore, no research has examinedthe influence of supervisors' role changes on sub-ordinate behavior.

The following research seeks to address some ofthe remaining questions sun"Ounding police super-vision in an era of community policing. Using foursupervisory styles identified in earlier research (Engel.200 I), the varying influences of supervisory styles onsubordinates' behavior are examined. Specifically, theinfluence of supervisors described as "innovative" iscompared to the influence of other supervisory styles.The findings show that while one supervisory styledoes have an important influence over the amount oftime officers spend on police-initiated activities andcommunity-policing and problem-solving activities, itis not the "innovative" supervisor as hypothesized.

. Tel.: +1-814-863-6399; fax: +1-814-863-7216

E-mail address: rscngel@psu.cdu (R.S. Engel).

0047-23S~ - see fnMtt I118tter 0 2002 EJsevier Science Ltd. All rights reserved.

PU:SOO47-23S2(Ol)OO122-2

R.S. Engel I Journal of Criminal Justice 30 (2002) 51-6452

The implications of these and other findings for policy

and future research are discussed.

The evolving role of first-line patrol supervisors

Prior to the community era in policing, patrol fieldsupervisors were described as performing traditionalroles of supervision. They were expected by theirsuperiors to monitor subordinate officers' activitiesand enforce departmental rules and regulations (Man-ni~g, 1977; Van Maanen, 1983). Furthermore, super-visors were expected to disseminate information bothup and down the chain of command in organizationsthat were traditionally para-militaristic, hierarchical,and bureaucratic (Trojanowicz. 1980). Subordinateofficers expected their supervisors to provide a bufferfrom higher levels of management; supervisors whoprotected their subordinates from criticism or punish-ment were especially respected by lower rankingofficers (Rubinstein, 1973; Van Maanen, 1983).These traditional models of supervision were basedon the exchange of favors between subordinate offi-cers and their supervisors through a system of con-tingent rewards and punishments (Allen & Maxfield.1983; Rubinstein, 1973; Van Maanen. 1983). Theultimate goal for supervisors during this era ofpolicing was to control subordinate behavior.

This supervisory role of controlling subordinatebehavior is similar to "transactional" leadership.Transactional leadership is based on the premise thatan exchange or quid pro quo takes place betweenleaders and followers. In this exchange, the subordin-ate performs what is expected of him or her, while theleader specifies the conditions under which thoseexpectations must be met and the rewards that thesubordinate will receive when the requirements arefulfilled (Bass, 1985; Burns, 1978; Downton. 1973).

As police organizations have evolved and greaterstress has been placed on the goals of community-oriented policing and problem solving, scholars havespeculated about the changing role for patrol super-visors (Goldstein. 1990; Van Maanen, 1983; Weis-burd, McElroy, & Hardymann. 1988; Witte, Travis. &Langworthy, 1990; Wycoff & Skogan, 1994). Schol-ars now Stress the role of the field supervisor as acoach and mentor for subordinate officers (Goldstein,1990). Under the principles of community-orientedpolicing, patrol officers are afforded more discretionand are encouraged to develop creative problem-solving approaches. Supervisors can facilitate thistransition in objectives by encouraging team building,raising morale, and emphasizing their role as a coachand mentor. Consequently, supervisors are expectedto rely less on their formal authority. The ultimategoal for supervisors in this environment is to develop

subordinates by emphasizing problem-solving skills,sound judgement, and creativity.

This role is similar to that of the "transfonna-tional" leader articulated by Bass and Avolio (1994).Leaders with a transformational style "stimulateinterest among colleagues and followers to view theirwork from new perspectives, generate awareness ofthe mission or vision of the team and organization.develop colleagues and followers to higher levels ofability and potential, and motivate colleagues andfollowers to look beyond their own interests towardthose that will benefit the group" (Bass & Avolio.1994, p. 2). Transformational leadership is believedto do more than establish exchanges between super-visors and subordinates; it is based on idealizedinfluence (leader as a role model), inspirationalmotivation, intellectual stimulation, and individual-ized consideration (Bass & Avolio, 1994).

Several scholars have noted variation in super-visors' willingness to embrace changes in theirroles and encourage changed within their organiza-tions. Cohen (1980) depicted the transition of theNew York Police Department in the late 1970sfrom a primarily traditional department to areform-oriented department that emphasized profes-sionalism. Based on interviews with commandingofficers, he reported two dominant styles of lead-ership: traditional-oriented and reform-oriented.These two types of leaders di~ significantlyin their willingness to support reforms proposed bythe administration.

Other scholars have noted variations in resistancetoward more recent reform efforts among supervisors.Van Maanen (1983) PTt'f'I',,('/1 that sergeants mayresist changes, ;IJ\ cnt ofcommunity policing, where patrol otficers are givenmore autonomy and control while sergeants have lessauthority and responsibility. Goldstein (1990) alsonoted the hesitancy of some supervisors to embrace anew role under problem-oriented policing, wherethey are less able to rely on their ability to controlsubordinates with departmental restrictions and for-mal authority. Recent research examining differentsupervisory styles of patrol supervisors further con-firms that supervisors' attitudes toward policychanges differ significantly (Engel, 2000).

Tbe acceptance of policy and philosophy changesby patrol supervisors is crucial for the successfulimplementation of these changes-particularly forcommunity-oriented policing and problem solving(Goldstein, 1990; Weisburd et al.. 1988; Witte etal.. 1990; Wycoff & Skogan. 1994). Several scholarshave reported that past failures of particular strategiesand structural changes was due in part to the lack ofsupport among patrol supervisors (Mark. 1976; Sher-man, Milton, & Kelly, 1973; Walker, 1993). Further-

R.s. Engd / .1OIImaJ of Criminal JU3tiC~ 30 (2002) 51-64

more, others have argued that if supervisors supportchanges within a department, their officers are morelikely to implement these changes at the street level(Bayley, 1994; Goldstein, 1990; Spanuw, Moorc, &Kennedy, 1990). This proposition, however, has notbeen empirically tested. That is, although researehhas shown that supervisors' reluctance to supportorganizational changes is rclated to the failure ofthese changes, it is not known if supervisors' supportof innovative policies and strategies influences offi-cers' implementation at the street level. If supervisorsdo not encourage officers to change their behavior,organizational change is likely to fail- but it isunclear if supervisors' encouragement is enough tochange officers' behavior.

Supervisors' influence over patrol officer behavior

an an'est. This effect, however, was contingent on dietype of department. He concludes, "previous researchhas slighted the potential influence of supervisors onpolice officers' arrest decisions... on the basis ofsome ethnographic evidence and, in light of currentfindings, this appears to be a serious omission"

(Smith, 1984, p. 31).2Other studies using officer and supervisor survey

data, however, have found that supervisors' attitudesdo not have an effect on subordinate behavior.Analyses of survey responses from patrol officersand supervisors in the Louisville Police Departmentfound that subordinates' reported behavior was notinfluenced by supervisors' reported preferences norpatrol officers' pen:eptions of supervisors' prefewm-ces (Allen & Maxfield, 1983). Likewise, Brown's(1988) survey of patrol officers and supervisorsshowed that the majority agreed that "the routineactions of field supervisors have but a marginalimpact on the way diey [patrol officers] use theirdiscretion," (p. 121). Finally, Mastrofski et at. (1994)found diat neither officers' pen:eption of their imme-diate supervisor's priority for Driving Under dieInfluence (DUI) enforcement nor officers' perceptionof higher administrators' priorities for DUI enforce-ment had a statistically significant effect on officers'enfon:ernent activities.

Collectively, studies examining patrol officersupervision have produced inconsistent findingsregarding the effect that supervision has over pab'Olofficer behavior. Some of this research lacks rigorousmethodological designs and advanced statistical tech-niques. In addition, supervision is often inadequatelyoperationalized as the quantity of supervision. ratherthan the quality or style of supervision (Engel, 200 1).Although scholars have speculated about die changingrole of patrol supervisors in die community policingera, no research to date has examined the varyinginfluence that supervisors have over the cornmunity-policing and problem-solving activities of their sub-ordinates. Using systematic observation and surveydata from two metropolitan police departments, thepresent research seeks to address these issues.

Although most scholars and practitioners agreethat one role of police field supervisors is to controlthe behavior of their officers. the degree of controlthat supervisors actually have continues to be amatter of debate. Several researchers have hypothe-sized about this relationship (Brown. 1988; Muir,1977; Van Maanen, 1983. 1984; Wilson, 1968) andothers have tested it empirically. Most of the empir-ical research exploring the influence of supervisionover patrol behavior has focused on three generaltypes of behavior: the frequency and duration ofencounters with citizens (Allen. 1980. 1982; Tim,1971). patrol officer discretionary decision makingtoward citizens. including decisions to ~t or issuetickets (Allen & Maxfield, 1983; Mastrofski, Ritti,& Snipes, 1994; Engel, 2000; Smith. 1984), andofficer misbehavior, including worf( shirf(ing anddepartmental violations (Bittner. 1983; Brehm &Gates, 1993; Brown, 1988; Gates & Worden,1989; Reiss. 1971).

Tifft (1971) reported that differences in supervis-ory styles within or between districts had little to noeffect on patrol officer effectiveness. I In contrast,

Allen's (1980.1982) research using systematic obser-vational data of twenty-four police departments inthree metropolitan areas (PSS data) reported thatsupervisor presence had a moderate effect on thetime subordinates spent at encounters with citizens.Allen also noted, however. that supervisors werepresent at only seven percent of all police-citizenencounters observed. Using the same data, Smith(1984) classified police departments based on theirlevel of professionalism and bureaucratization.Employing a multivariate probit analysis. he foundthat the presence of a supervisor at the scene of anencounter with a citizen significantly increased theprobability that the subordinate officer would make

Method

Data

This empirical examination of supervision usesdata collected for the Project on Policing Neighbor-hoods (POPN), a large-scale study of police behav-ior funded by the National Institute of Justice.Systematic observation of patrol officers and fieldsupervisors (sergeants and lieutenants) was con-ducted during the summer of 1996 in the Indian-

R.S. Enge/ / Journal a/Criminal Justice 30 (1002) 5/-6454

apolis, Indiana Police Department (IPD) and duringthe summer of 1997 in the St. Petersburg, FloridaPolice Department (SPPD).

Systematic observation was conducted in IPDwith fifty-eight patrol supervisors (sergeants andlieutenants) during eighty-seven rides, totaling over600 hours - 78 pen:ent of IPD patrol sergeants and

lieutenants were observed at least once. In SPPD,systematic observation was conducted with twenty-six patrol sergeants, four patrol lieutenants, andeight patrol offICers wooong as the acting sergeantduring seventy-two shifts - 96 percent of patrol

sergeants in SPPD were observed at least once. Inaddition, field observation of patrol officers wasconducted for over S,700 hours in twenty-fourneighborhoods across the two sites. In IPD, 194patrol officers were observed during 336 shifts,while in SPPD, 128 officers were observed during360 shifts. Combined, observers recorded infonna-tion regarding encounters between officers andapproximately 12,000 citizens on standardized cod-ing forms (Mastrofski, Snipes, Parks, & Maxwell,2000; Parks, Mastrofski, Dejong, & Gmy, 1999).

Data were also obtained through structured inter-views with sixty-nine of seventy-four patrol super-visors in IPD and all twenty-seven patrol sergeants inSPPD. The interview captured demographic and back-ground information along with supervisors' viewsregarding a number of topics, including managementpriorities and innovative policies such as problemsolving and community-oriented policing. More than9S percent of patrol officers in IPD (398 officers) and97 percent of patrol officers in SPPD (240 officers) alsocompleted a structured interview of similar design.

Research ,viles

The IPO serves the city of Indianapolis, with anestimated population of 377,723 in 1995. At thattime, the city consisted of 39 percent minorities, 8percent unemployed, 9 percent below the povertylevel, and 17 percent female-headed households withchildren. The VCR Index crime rate in 1996 was 100per 1,000 residents. During that year, the departmentemployed 1013 sworn officers, 17 percent of whomwere female, 21 percent minority, and 36 percent witha four-year college degree (Mastrofski et al., 2000).

IPO's patrol division was divided geographicallyinto four districts. Each district differed slightly incharacter, mission, and emphasis placed on commun-ity policing. Twenty-four officers within the depart-ment were designated as community-policing officers(CPOs) working in teams within three of the fourdivisions their approach to problem solving differedslightly across districts (see Dejong, Mastrofski, &Parks, 200 I). Officers were expected to participate in

IMPACT, a program that emphasized innovativeproblem-solving techniques. In addition, top admin-istrators emphasized a more aggressive order main-tenance approach to community policing, whichincluded traditional law enforcement activities toimprove quality of life within neighborhoods. Thechief of IPD and other top administrators stressed thatthe philosophy of community policing and problemsolving should guide the practices of all patrolofficers (Parks et al., 1999; Dejong et al., 200 I).

Supervisory structure and form also differed acrossIPD districts and shifts. It was unclear if a directsupervisor-subordinate relationship existed becausethis department did not have a fonnal evaluationprocess in place. Patrol officers and sergeants wereassigned particular work schedules that provided thatsergeants worked the same schedule as the group ofoffICers they were responsible for. Officers working inspecialized community-policing units were directlysupervised by one sergeant whose sole responsibilitywas to monitor and direct the activities of the officersin that unit

St Petersburg, Florida. (population of 240,318 in1995) is located at the southern tip of PinellasCounty. In 1995, the city consisted of 24 percentminorities, 5 percent unemployed, 6 percent belowthe poverty level, and 10 percent female-headedhouseholds with children (Parks et al., 1999). TheVCR Index crime rate (per 1,000 residents) was 99 in1996. During that year, the SPPD had 505 swornofficers, 13 percent of whom were women, 22percent minority, and 26 percent with a four-yearcollege degree (for details, see Mastrofski et al.,2000; Parks et al., 1999).

SPPD a focused on problem solving and geo-graphic deployment of officers, and had "developedan international reputation as a leader" in theseprograms (Parks et al., 1999). The department wasdivided into three districts and forty-eight commun-ity-policing areas (CPA) with at least one CPOassigned to each area whose full time responsibilitywas "resource management and facilitation." CPOswere allowed to flex their schedules according totheir personal needs and the needs of the community.These officers were freed from dispatched assign-ments and were encouraged to utilize innovativeproblem-solving tactics within their areas. In addi-tion, CPOs were expected to become the mainliaisons with neighborhood organizations. As inIPD, administrators of SPPD stressed that the philo-sophy of community policing and problem solvingshould guide practices of all patrol officers (Parkset al., 1999; Dejong et al., 200 I ).

The structure of supervision within SPPD had

undergone tremendous change in recent years priorto the study.3 During the period of observation,

R.S. Engel / Jountal of Criminal Jus/ice 30 (2002) 51-64

supervision reflected a compromise between geo-graphic deployment and a squad system. Sergeantswere directly responsible for a "team" of officers ona particular shift working in a specific sector orgeographic area (three sectors in each district), alongwith a handful of CPOs assigned to a CPA. Eachsergeant had responsibility for one or more CPAs(and the CPOs working in those areas) within theirsector. Sergeants were scheduled to work three"temporal" shifts (twenty-tour hours) and then"flex" their schedule for the remaining sixteenhours each week. During a "temporal" shift, ser-geants were responsible for the direct supervision ofall patrol officers working that shift for the entiredistrict. not just their sector. During "flex" shifts,sergeants were expected to work on problem solvingin their specific CPA. supervise their CPOs, andcomplete administrative paperwork.

The differences between the community-policingand problem-solving approaches for IPO and SPPOhave been thoroughly described by Dejong et al.(200 I). As they indicate, compan:(i to SPPO, IPDcommitted a much smaller percentage of their patrolofficers as community-policing specialists (5% inIndianapolis compared to 17% in St. Petersburg).Furthermore, specialists in IPO worked within asingle unit, while SPPO specialists were moreintegrated into teams working with patrol general-ists. In regards to management, IPO districtcommanders were more involved in setting priorit-ies, while SPPO managers played a more supportiveand facilitative role. Finally, as described by Dejongand her colleagues, problem-solving efforts in (POfocused on a narrower range of problems than thosein SPPO.

Mea.~ures of supervision

management and policing literatures (Engel, 200 I ).Combining both sites, eighty-one supervisors whowere both interviewed and included in the analyses-seventeen lieutenants and thirty-nine sergeants fromIPD and twenty-five sergeants from SPPD.4 Nineattitudinal dimensions were identified and measuredfrom supervisors' survey responses (i.e., power dis-tribution, decision making, activity level, relations-orientation, task-orientation, inspirational motivation,expect community-relations, expect aggressiveenforcement, and view of subordinates). These atti-tudinal dimensions were combined through factoranalysis to create four different factors, interpretedas distinct supervisory styles and labeled as tra-ditional, innovative, supportive, and active. For adetailed description of dte selection of underlyingattitudinal constructs and creation of supervisorystyles, see Engel (200 I). A brief description of eachsupervisory style is given below.

"Traditional" supervisors expect aggressiveenforcement from subordinates and do not expectcommunity-oriented activity. These supervisors arehighly task-oriented (expect subordinates to producemeasurable outcomes and also emphasize paperworkand documentation) and are decision makers. Thesesupervisors give more instruction to subordinates andare more likely to take over a situation or handle itthemselves. Finally, these supervisors are less likelyto reward and more likely to punish subordinates. Theultimate concern for traditional supervisors is tocontrol subordinate behavior.

"Innovative" supervisors expect community-ori-ented activity from their subordinates. They arestrongly relation-oriented and are not task-oriented(that is, they do not expect production of meas-urable output). Innovative supervisors have moreperceived power in the organization and have pos-itive views of subordinates. They are more receptiveto innovative policies and changes in the organiza-tion. The ultimate goal for these supervisors is todevelop the problem-solving skills of subordinates.This is typically done through mentoring and coach-ing subordinates.

.. Supportive" supervisors are also highly relation-

oriented, however, they show concern for subordinatesby protecting them from unfair criticism and punish-ment. These supervisors provide a buffer betweensubordinates and management. They have high levelsof inspirational motivation, emphasize teamwork, andencourage creativity by supporting subordinates. Theyare also more likely than other types of supervisors toreward subordinates. Supportive supervisors are nottask-oriented; that is, they are not concerned with workoutput measures. The ultimate goal for supportivesupervisors is to develop subordinates, however, thisis accomplished through providing protection and

Using systematic observational data, "super-vision" has been operationalized in past researchas a measure of the number of field contacts betweena supervisor and a subordinate (Allen, 1980, 1982;Gates & Worden, 1989), or as the presence of asupervisor at the scene of a police/citizen encounter(Brehm & Gates, 1993; Smith, 1984; Smith & Klein,1983). Only one study utilizing observational dataattempted to measure styles of supervision. Tifft(1971) systematically observed patrol officers basedon a sampling of supervi~rs who were self-iden-tified as baving a particular style. Note, however,that the validity of these self-categorizations wasnever established.

In the following analyses, the substantive natureof supervision was measured rather than the quantityof supervision, by using the underlying attitudinaldimensions of supervisory styles identified &om the

56 R.s. Engel/Journal of Criminal Justice 30 (2002) 51-64

Hypothesesbeing an advocate. Unlike innovative supervisors,supportive supervisors see management as somethingofficers need to be protected from.

"Active" supervisors are characterized by theirhigh levels of activity in the field. This activityincludes both patrol work and supervisory functions.These supervisors tend to be decision-makers, takingover and handling situations themselves. Unlike tra-ditional supervisors, these supervisors feel they havepower within the organization and have positiveviews of subordinates. Active supervisors also havelower levels of inspirational motivation. The ultimategoal for active supervisors is to be active field super-visors, performing the dual function of street officerand supervisor.

The distribution of supervisory styles for this sam-ple of sergeants and lieutenants is reported in Table I.There is a roughly equal distribution of each style,however, when the styles are examined for eachdepartment separately, significant differences emerge.Traditional supervisors are significantly overrepre-sented in SPPD, while active supervisors are under-represented. IPD sergeants are evenly distributedacross all styles. however. IPD lieutenants are slightlymore likely to be classified as innovative and less likelyto be classified as traditional (Engel. 200 I ).

Differences between male and female supervisorsare also reported in Table I. Female supervisors aredisproportionately represented as traditional super-visors (50 percent of female supervisors are classifiedas traditional supervisors). Few other differences inclassification are apparent. There are no statisticallysignificant differences in classification with regard tothe supervisors' race, rank, age, years of experience,or education. The four types of supervisors also donot differ from one another in their reported views ofthe importance of promotion or moving to a speci-alized unit. or their amount of b'aining and generalknowledge of the principles of community policing.Innovative supervisors, however, report receivingsignificantly more training in supervision, manage-ment. and leadership, while the other three supervisorstyles do not differ in their reported amount of train-

ing (Engel, 2001).

To examine the influence of supervisory stylesover officer behavior in departments emphasizingcommunity policing, the amount of time officersspend on four types of behaviors-proactivity, prob-lem-solving/community-policing activities, admin-istrative tasks, and personal business-were examined.Based on the findings from the literature, two mainhypotheses were tested in the following analyses. It isexpected that supervisors' influence over officerbehavior varied based on: (1) the specific type ofofficer behavior examined and (2) the specific super-visory style examined. That is. it was hypodlesizedthat the strength of the relationship between super-vision and officer behavior varied based on the typeof task the officer was engaged in, and the style ofsupervision experienced by the subordinate.

Wilson (1968) was one of the first researchers tosuggest that police administrators' influence oversubordinates' use of discretion varied with the char-acteristics of the situation. He hypothesized thatadministrators have greater control over patrol offi-cers' behavior in both police-invoked and citizen-invoked "law enforcement" situations compared to"order maintenance" situations. Wilson suggestedthat administrators could control officer discretion inpolice-invoked law enforcement situations "byobserving substantive outcomes or by measuring theoutput of individual officers," and could controlofficer discretion in citizen-invoked law enforcementsituations "by setting guidelines on how such caseswill be handled and by devoting, or failing to devote,specialized resources. . . to these matters" (p. 88). Itwas only in order-maintenance situations that theability of administrators to effectively control policeofficer discretion was questioned. Wilson's hypothesisthat administrators have less control over policeofficer discretion in order-maintenance situations isespecially salient today, as police departments evolveand patrol officer roles emphasize handling order-maintenance situations and quality of life conditions.

In contrast, both Jermier and Berkes (1979) andVan Maanen (1983) have proposed that control over

Tlblel

Supervisory styles

Overall Males Fea.Ies Whi8 N.--WhiIe IPD IieutCDauIs IPD sersents SPPD IeIIeI*

(n=!!L (n=69)~-12) (11=69) (11-12) (n-l7) (n=39) (n=2S)

Supervisory sty/e.fTraditional 0.26Innovative 0.27

Supportive 0.23Active 0.23

0.220.300.250.23

0.500.080./70.2S

0.260.260.250.23

0.250.330.J10.25

0;120.350.240.29

0.180.28

0.260.28

0.480.200.200.12

Entries ~ means representing the percent of supervisors with that predominant style.

R.S. Engel / JOImtQl of Criminal JlLflice 30 (2002) 1/-6457

Likewise, although supervisors may have limitedinfluence over the amount of time officers spend onpersonal business during their shifts, particular typesof supervisors may have more influence than others.It is hypothesized that the amount of time subordi-nates spend conducting personal business willdecrease as the amount of supervision increases.Active supervisors have been described as moreactive on the streets, engaging directly in patrol workand interacting with subordinates. Likewise, tra-ditional supervisors emphasize adherence to rulesand regulations. Innovative supervisors are less likelyto hold officers directly accountable for their time(e.g., they are not "task-oriented"), while Supportivesupervisors have stronger relationships with subordi-nates and are more likely to protect them. Therefore,it is hypothesized that officers with active and tra-ditional supervisors will spend the least amount oftime conducting personal business during their shifts,while officers with supportive and innovative super-visors will spend the most.

Analyses

Sample

Analyses of the effects of supervision on patrolofficer behavior were contingent upon linking offi-cers with their individual supervisors. Officers werenot asked for the name of their direct supervisor toencourage forthright responses on the officer sur-vey, which included sensitive questions about thequality of supervision. Therefore, officers werematched with their direct sergeants through otherinfonnation collected from the officer and super-visor surveys. In IPD, both officers and supervisorswere assigned to districts. shifts. and work sched-ules. Of the 194 officers observed. 82.9 percentwere matched with an individual sergeant fromtheir shift and work schedule.s In SPPD, officersand supervisors Were assigned to districts. shifts,and CPAs. Of the 128 officers observed in thisdepartment, 92.5 peKent were matched to theirindividual sergeant.6 Only the 239 officers matchedwith a sergeant were included in the followinganalyses. Analyses were conducted at the shift (orride) level--officers with an identified sergeantWere observed during 518 shifts.

subordinate behavior may be effected by task dif-ferentiation. Specifically, Van Maanen hypothesizedthat "the less certain the task, the less visible itsperformance, the less direction provided from above,and the lower the official wort load, the moreopportunity a sergeant has to provide his men witha definition for their duties" (pp. 297-298). Admin-istrato~ and superviso~, however, often fail toclearly define the goals of community policing andproblem solving. Therefore, it is believed that super-visors have little control over community-policingand problem-solving tasks because these tasks areoften ambiguous and unclear.

Applying Wilson's (1968) propositions to fieldsuperviso~, one might suppose that street-level super-visors also have varying influences over officer dis-cretion. In situations where work output is easilymeasured, it is hypothesized that supervisors havegreater control over subordinate decision making. Forexample, reports, arrest, and citations create statisticsthat serve as a means for supervisors to measure theirsubordinates' work output. As suggested by VanMaanen (1983), supervisors use these "stats" toinfluence subordinates' behavior because they areconcrete measures that are interpretable by both super-visors and subordinates. Applying these propositions,it is hypod1esized that supervisors will have a strongerinfluence over officers' proactivity and adminisb'ativeactivities and less influence over problem-solving!community-policing and personal business activities.Given traditional and active supervisors' emphasistoward task-orientation (e.g., producing measurableoutcome), it is speculated that officers with tra-ditional and active superviso~ will spend more timeengaging in proactivity and conducting adminisb'at-ive tasks than those office~ with innovative or

supportive supervisors.Although the overall influence of supervision on

community-policing activities undertaken by subor-dinates is likely to be small. the effect is expected tovary by the type of supervisory style. It has beenspeculated that supervisors who support innovativestrategies and philosophies can encourage subordin-ate office~ to do the same (Goldstein, 1990).Therefore, one would expect officers with super-visors who are more "transfonnational" (i.e., super-viso~ who are more likely to coach, mentor, andfacilitate officers) will spend more time engaging incommunity-policing and problem-solving activities.In, the present research, sergeants categorized asinnovative are more likely to coach or mentorofficers and expect office~ to perform problem-solving and community-policing activities. There-ore, it is hypothesized that officers with innovative

superviso~ will spend more time per shift conduct-ing problem-solving activities.

Dependent variables

To explore the effects of supervision on policeofficer behavior, the time officers spend per shiftconducting four general types of behaviors wereexamined: proactivity (i.e., police-initiated activities),

58 R.S. Engel/Journal of Criminal Justice 30 (1001) 5 J-64

activities, while officers with innovative sergeantsspent the least amount of time (8.0 percent) enga-ging in these activities.

One might speculate that increases in problem-solving and community-policing activities will beaccompanied by increases in time spent on admin-istrative duties documenting these activities. Toaccount for the time officers spend on problemsolving, many departments (including IPD andSPPD) require that officers complete additionalpaperwork documenting problems and the strategiesused to solve those problems (Parks et al., 1999). Inaddition, the percentage of time spent per shiftengaging in any administrative activities (e.g., reportwriting, processing evidence, court time, etc.) is alsomeasured. On average, 17.3 percent of an officer'sshift is spent engaging in these types of activitiesand encounters. Officers with active supervisorsspent the least amount of time engaged in admin-istrative activities (13.3 percent), compared to 19.2percent of shifts observed for officers with tra-ditional sergeants.

Finally, personal business was measured as theamount of time spent at nonwork-related encountersand activities. Although the time spent on personalbusiness may have been somewhat attenuated due tothe presence of observers, one may still expectvariation among officers due in part to supervisoryinfluences. Obviously, some time per shift wasexpected to be spent on personal business (e.g.. mcalbreaks, restroom breaks, etc.). Unfortunately, therewas no measure that controls for the legitimacy of thetime spent on personal activities. It was assumed thatthe amount of time spent per shift on personalactivities and encounters increased proportionallywith the amount of time spent "shirking" workresponsibilities (Brehm & Gates, I 993). As shownin Table 3, officers were observed to spend onaverage, 15.5 percent of their shifts engaging inpersonal business. Officers with innovative super-visors spent the most amount of time per shift onpersonal business (18.3 percent), while officers withactive supervisors spent the least amount of time(12.9 percent).

community-po I ic ing/prob lem -so I v ing activities,administrative activities, and personal business activ-ities. Each of these behaviors was measured as thepercent of time spent engaging in that particularactivity per eight-hour shift observed (see Parks etaI., 1999).

Proactivity was measured as the percentage oftime per shift that officers engage in activities thatare self-initiated (i.e., activities that are not initiatedby a dispatcher, citizen, supervisor, or otherofficer). This measure includes activities andencounters with citizens involving suspicious situa-tions, public disorders, victimless or moral offenses,traffic offenses, service needs, and problem focusedactivity (e.g., surveillance, security checks, propertysearches, ordinance enforcement, backing up otherofficers, etc.). This measure however, excludes timespent on general patrol, traveling en route to alocation, personal business, and administrativeactivities or encounters. As shown in Table 2,officers spent an average of 13.2 percent of dleirshift engaging in self-initiated activities andencounters. Officers with active supervisors spentthe most time per shift (15.4 percent), whileofficers with innovative supervisors spent the leastamount of time (10.7 percent).

The amount of time per shift that officers engagein cornmunity-oriented policing and problem-solv-ing activities was also examined. This measureincludes time spent on all activities and encountersbetween officers and citizens that were considered"part of a long-term plan or project," or if officerswere "trying to determine the nature, extent, orcauses" or "prevent the occurrence or recurrence"of a long-term problem (Dejong et al., 2001; Parkset al., 1999). This measure also includes time spentin meetings/phone calls with representatives ofcitizen organizations or service providers, alongwith time spent gathering information. On average,officers spent 9.8 percent of the their shifts enga-ging in what has been defined as problem-solvingactivities or encounters (see Table 2). Again, offi-cers with active sergeants spent the most time pershift (11.3 percent) engaging in problem-solving

Table 2Dependent variables

13.OS9.37

19.1714.94

10.66

7.96

17.54

18.27

13.6510.7317.4816.59

Entries are means representing the percentage of each shift spent on activities and encounters. Also included are the moans forshifts with each type of supervisor.

R.s. Engel / JOIImal of Criminal Justice 30 (2001) 51-64 59

FindingsTable 3Independent variables

Variables Minimum Maximum Mean S.D.

Officer characteristicsRace (1 : White)Sex (1 :ma1e)Years experienceCP assignment

00

0.72

0.81

8.43

0.20

0.410.396.340.40

311.0

0 0.58 0.49

00

0.43

0.38

0.21

0.500.490.9819 2.17

10 100 75.43 14.60

Otgonizational/ shiftDepartment

(I aSPPD)Day shiftEvening shiftConcentrated

disadvantagePercent

discretionary time

Supervisory stylesTraditionalInnovativeSupportiveActive

0000

0.40

0.20

0.20

0.20

0.490.400.400.40

Independent variables

The bivariate relationships described aboveshow a clear pattern of differences in the timespent by officers based on the supervisory styleof their direct sergeant. To test these relationshipswhile controlling for other possible influences,additional variables were included in the multi-variate models that follow. Officer characteristicsincluding sex, race, and assignment as a communitypolicing or beat officer were measured as dicho-tomous variables, while job experience was meas-ured as the number of years working in thedepartment. Variables for the department, shift,and supervisory style of an officer's sergeant werealso measured as dichotomies.

Other control variables (percent of discretionarytime and concentrated disadvantage of the beatassigned) were measured as continuous variables.The percent of discretionary time referred to theamount of time per shift that officers were notassigned to calls or activities by dispatchers, super-visors, or citizens. Neighborhood context wasmeasured as the level of concentrated disadvantageof the beat that officers were regularly assigned topatrol. This measure was a four-item weightedfactor score, which includes the percent poor,percent unemployed, percent female-headed fam-ilies, and percent African - American population ofthe beat (Mastrofski et al., 2000; Sampson, Rau-denbush, & Earls, 1997). These variables are fur-ther described in Table 3.

Ordinary least square regression was used toexamine the multivariate influence of supervisorystyles on patrol officer behavior. Due to the hierarch-ical nature of the data, hierarchical linear modelingestimates were also derived. Although ignoring dienested structure of multilevel data could lead tobiased standard errors and false tests of statisticalsignificance (Bryk & Raudenbush, 1992), die hier-archicallinear model results indicated that there wasnot significant dependence between groups of obser-vations. Therefore, multilevel modeling was notnecessary (Bryk & Raudenbush, 1992). Furthennore,the estimates generated through HLM were similar tothe logistic regression results and therefore were notreported in the text. Table 4 reports the ordinary leastsquare regression coefficients and standard em>rs foreach of the four dependent variables.

Table 4, Model A reports the unstandardizedregression coefficients and standard errors for ananalysis of time spent on police-initiated activitiesand encounters. Findings show that female officers,White officers, less experienced officers, and offi-cers working shifts with more discretionary timespent significantly more time per shift engaging inproactivity. One supervisory style has an influenceover officer behavior--officers with active super-visors spend significantly more time per shiftengaging in proactivity compared to officers withinnovative supervisors.

The analysis of time officers spend on problem-solving and community-policing activities is alsoreported in Table 4 (Model B). The results show thatfemale officers, White officers, officers with lessexperience, and officers assigned to community-poli-cing assignments spent significantly more time pershift engaging in problem-solving and community-policing behavior. In addition. day shift and depart-ment (St. Petersburg) were significant predictors.Again, one supervisory style - active supervi-sors - bad a significant influence over officer behav-ior. Officers with active sergeants spend significantlymore time per shift engaging in problem-solving andcommunity-policing activities compared to officerswith innovative sergeants.

Model C displays the analysis of the amount oftime spent per shift on administrative activities andencounters. Officers with a community-policingassignment and officers who work in SPPD spentsignificantly more time engaging in administrativeduties. In addition, officers with traditional super-visors, supportive supervisors, and active supervisorsspent significantly less time on administrative dutiesthan officers with innovative supervisors.7 That is,officers with innovative sergeants spend significantly

R.s. E"gel / Joumal of Cri",i"aI Justice 30 (2002) 51-6460

Table 4Ordinary least squares resressioo 8II8lyaes for time spent per shift (11=539) ,; ~Ii,..

. .' .Personal business:

Model D

- 4.64 (.58)... - 14.61(3.82)--- 12.73 (3.70)888Constant

OffICer characteristicsWhiteMaleYears experienceCP assignment

-0.32 (1.29)- 2.12 (1.40)- 0.11 (0.09)

3.84 (1.7.1)8

2.58 (1.15)*- 2.87 (1.24).- 0.30 (0.08)...

2.17 (1.56)

2.89 (1.33).- 7.40 (1.44)...- 0.33 (0.10)...

S.97 (1.80)...

9.99 (1.14)888

-0.61 (1.76)

-0.99 (1.42)

0.16 (0.54)

O.OS (0.04)

- 5.03 (0.99)---

3.11(1.52)-- 1.57 (1.23)- 0.26 (0.47)

0.22 (0.03r--

0.67 (1.02)-0.42(1.57)

0.61 (1.27)0.83 (0.48)0.27 (0.03)888

4.78 (1.18)888

3.91 (I.SI)8

-0.40 (1.47)

0.10 (0.56)- 0.06 (0.04)

Organization/shiftDepaI1mentDay shiftEvening shiftConcentrated disadvantagePen:ent discmionary time

Supervisory stylesTraditi<malInnovativeSupportiveActive

1.51 (1.31)- 0.88 (1.56)

0.52 (1.75)4.81 (1.77)..

- 2.92 (I.~I)*

-2.93(1.70)-3.8S (1.71)*

1.54 (

1.94 (I4.62 (I

2.29 (1.46)-0.94 (1.48)

12.71...0.22

12.11...0.22

11.15...0.20

11.39***0.21

Model F statisticKEntries are unstandardi~ coefficients and, in parendteses, standard ~rs. Innovative supervision was the excluded category.. P< .05... P< .01.

... P< .001.

more time on administntive activities compared toofficers with any other type of supervisor.

The amount of time spent per shift conductingpersonal business is examined in Model D. Althoughthere are several significant variables (officer race,officer experience, officer assignment, shift, depart-ment, and percent discretionary time), supervisorystyle does not have a significant influence on the timeofficers spend conducting personal business.

officers widt any odter type of supervisor. Finally, nosupervisory style has a significant impact on the timeofficers spent conducting personal business duringtheir shifts. This finding supports previous researchdtat found supervisors had little or no influence overparticular types of work shirking (Brehm & Gates.1993; Gates & Worden, 1989).

It was interesting that officers with innovativesupervisors did not spend more time engaging inpolice-initiated activities or problem-solving and/orcommunity-policing activities, but did spend moretime on administrative tasks. At first glance, it mightappear that innovative supervisors (or those with amore "transformational" role) were not having theinfluence over officers' behavior at the street levelthat scholars had speculated they would. Otherresearch, however, has shown that community-poli-cing specialists actually spend less time in face-toface encounters with the public and more timeengaging in administrative tasks and other "behindthe scenes" activities (Parks et al., 1999). Therefore,the fmding that supervisors who have embracedcommunity policing have produced subordinateswho spend more time per shift engaging in admin-

Discussion

Collectively, these findings suggest that supervis-ory styles do influence some forms of subordinatebehavior. Compared to other types of supervisorystyles, supervisors with an active style appear to havethe most influence over subordinate behavior. Activesergeants significantly increase officers' time spenton self-initiated and problem-solving activities pershift, while also significantly decreasing time spenton administrative duties. In addition, controlling forother factors, officers with innovative sergeants spentsignificantly more time on administrative tasks than

Police initiated: Problem solving: AdminIstrative:Variables Model A Model B Model C

-3.60 (1.12)...-0.S4 (1.21)

0.21 (0.08)..-4.91 (1.51)...

.3S)

.SI)

.S3)--

R.S. Engel I JOIImaI of Criminal JlLftice 30 (2002) 51-64 61

istrative tasks is not surprising. If the goal of com-munity-policing and problem-solving strategies is tohave officers engaging in more self-initiated. com-munity-policing, and problem-solving activities onthe street. Supervisors who have embraced thesestrategies do not appear to be delivering the desiredoutcomes. Indeed, this research shows that super-visors with an active style (i.e., supervisors who areactive in the field, making decisions and handlingsituations, controlling subordinates' behavior, andhave positive views of their subordinates) have thestrongest influence over officer behavior.

Prior research has shown that active supervisorsdiffer significantly from other types of supervisors intheir own behavior-they spend significantly moretime engaging in general patrol, making traffic stops,and going to incidents on their own initiative (Engel,200 I). They do not rely on "inspirational motiva-tion" to influence subordinates, but rather "lead byexample" (Engel, 2000/200 I). This type of lead-ership is described by one active supervisor.

While discussing different "styles.. of supervisors.51 [the observed supervisor] said that. in a "tra-ditional.. department. sergeants were able to "leavetheir mark on their officers." For example. if thesergeant wanted drug &nests. officen produced drugarrests. if be wanted drunks. they brought himdrunks, etc. But. in an "innovative.. departmentsuch as this one. there are few supetVisors whocan leave their "imprint.. on an officer. 51 said thatsupetVising subordinates in a tJaditional departmentis much different. where they gain compliancethrough fear and discipline. In an innovative depart-ment. gaining compliance was a totally differentproJK>Sition and that is why supervisors need to lead

by example (POPN).

antly more likely to use force against suspects theyencounter (Engel, 2000). In addition, active super-visors themselves engage in more uses of forceagainst suspects than other types of supervisors -at a rate of 0.14 times per eight-hour shift, comparedto 0.009 times per eight-hour shift for all othersupervisors (Engel, 2001). Similarly, the ChristopherCommission's (1991) investigation of LAPD hasspeculated that aggressive supervisors transmit thesesame aggressive values to officers. Together, thisevidence suggests that supervisors do have an influ-ence over subordinates' behaviors, however thisinfluence can be positive or negative depending uponthe example they set.

Although identifying supervisory styles andexamining their influence on subordinate behaviorhas provided interesting findings, caution should beexercised when interpreting them. The data used inthis study of police supervision was limited inseveral ways. First, the POPN utilized a datacollection design created for systematic observationof encounters between patrol officers and citizens.Although systematic observation has proven to be auseful technique to study pab'ol officers' behavior,the utility of this technique for patrol supervisorswas unknown prior to the POPN data collection. Inpilot tests, the use of systematic observation techni-ques proved to be particularly difficult when used toexamine supervisors' interactions with other offi-cers. As a result, POPN observers were instructed tofocus more on detailed narrative writing of super-visors' activities and encounters with other officersand citizens, while relying less on quantitativecoding on a data collection instrument This semi-structured narrative data was collected based on theprinciples of debriefing (Mastrofski & Parks, 1990).Observers were instructed to ask supervisors todescribe their decision-making processes directlyafter observing situations and to directly probesupervisors during the shift for their views ondepartmental policies, management, their roles, andinteractions with subordinates. Utilizing contentanalysis, a technique used to transform qualitativedata into quantitative form (Holsti, 1968), futureresearch will be able to explore some of the issuessurrounding police supervision that have not beenadequately addressed.

Second, while systematic observation and surveysprovide a descriptive slice of police work, they oftendo not provide detailed information about long-termpatterns of police behavior or the effectiveness oflong-term policies and strategies. The study of patrolsupervision may be better captured by some type ofmodified ethnographic research design where detailedinformation about the actual patterns (especially theunderlying rationales, objectives, etc.) of supervisory

Perhaps, supervision in the community policingera could be as simple as showing officers what isexpected of them. Indeed, "transfonnationaI" lead-ership may not be the definitive answer scholarshave proposed. Rather than instructing supervisorsthat their new role in the community era is toencourage subordinates through coaching and men-toring, they should be told to "get in the game"themselves. That is, perhaps, the best way to influ-ence officer behavior is to have supervisors in thefield engaging in the same activities they want theirsubordinates to engage in.

Take note, however, that there are problems asso-ciated with leading by example and the active super-visory style. Other research has shown that activesupervisors also have a significant influence overtheir subordinates' use of force. Controlling for otherfactors, officers with active supervisors are signific-

R.S. Engel / JolmUli of Criminal Justice 30 (2002) 51-6462

by example is only an effective supervisory tool ifsupervisors are engaging in activities that support thelegitimate goals of their departments.

Ackno"'ledgments

Previous versions of this paper were presented atthe Academy of Criminal Justice Sciences AnnualMeeting, Orlando, Florida, March 1999 and theNational Institute of Justice Annual Conference,Washington, DC. July 18-21. This manuscript isbased on data from the POPN, directed by Stephen D.Mastrofski, Roger B. Parks. Albert J. Reiss, Jr., andRobert E. Worden. The project was supported bygrant no. 95-U-CX-OO71 by the National Institute ofJustice, Office of Justice Programs. U.S. Department.Points of view in this document are those of theauthor and do not necessarily represent the officialposition or policies of the U.S. Department of Justice.The author would like to thank the following scholarsfor their helpful comments and suggestions: RobertE. Worden, Hans Toch, David H. Bayley, DavidMcDowall, Alissa Pollitz Worden. Stephen D.Mastrofski. and Roger B. Parks.

Notes

practices could be collected. While the POPNresearch design did have a partial edmographic com-ponent containing detailed information collected dur-ing each ride, examinations of long-term patterns ofsupervision and the structural, environmental, andpolitical factors affecting these patterns was nonethe-less limited. Future research on police supervisionshould address these issues through a blending ofqualitative and quantitative research techniques direc-ted toward examining long-term patterns and super-

visory practices.Taken with previous research, the fmdings pre-

sented in this article have important policy implica-tions. Experience has shown that field supervisorsmay impede the implementation and eventual suc-cess of innovative policies (Mark, 1976; Sherman etal., 1973; Walker, 1993). It has been generallyrecognized that field supervisors play an importantrole in communicating administrators' philosophiesand policies. Their communication of these policieshas a significant influence on their acceptance bysubordinates (Van Maanen, 1983, 1984). Thisresearch, however, suggests that simply acceptingcommunity-policing and problem-solving strategiesmay not change line-officer behavior on the streetWhile officers with innovative supervisors do spendmore time on administrative tasks, they do notspend more time engaging in community-policingand problem-solving activities on the street. Rather,officers with active supervisors are more likely tospend time engaged in these activities. It is possiblethat supervisors have embraced different forms ofcommunity policing-active supervisors may bemore inclined to use and encourage aggressiveenforcement tactics, while innovative supervisorsmay be more likely to use and encourage commun-ity-building tactics. This interpretation lends supportto other findings that suggest officers with activesupervisors (and active supervisors themselves) aremore likely to use force against citizens. It ispossible that officers and supervisors who use moreaggressive policing styles are more likely to findthemselves in situations where the use of force is

needed (Engel, 2000).In conclusion, these findings suggest that to have

a significant influence over behavior, supervisorsmust lead by example. It is the example that theyset, however, that should most concern police admin-istrators and citizens. Active supervisors have thestrongest impact on officers' behavior, however, thisimpact can be positive (e.g., increased time spentconducting community-policing and problem-solvingactivities) or negative (e.g., increased use of force).To influence officers' behavior, supervisors need tolead by example-but the example that they set has

important and wide-reaching implications. Leading

I. Officer effectiveness was operationalized as thenumber of incidents handled per tour, response to radiocalls, time spent handling encounters, down time after anencounter was handled, and the manner of handling crimes,complaints, offenses, and disturbances.

2. Smith's findings must be interpreted with caution,however, because the seriousness of the offense was notadequately controlled. Offense seriousness is correlated withboth supervisor presence and arrest, therefore, the relation-ship reported between supervisor presence and arrest may be

spurious.3. Previously, supervision was organized a.~ a "squad

system" where one sergeant was directly responsible for afixed group of officers working the same schedule. Aftera change in administrative personnel, SPPD implementeda supervisory structure that focused on geographic deploy-ment. Each sergeant in the department was assigned to aparticular geographic area (CPA) that they were directlyresponsible for. As a result, sergeants were responsible forsupervising patrol officers and CPOs who were assigned totheir particular area across every shift. After about a year.this structure of supervision was reorganizcd because of theunrealistic demands it placed on sergeants.

4. Supervisors were excluded from the analyses ifthcyw~ not both interviewed and observed (sixteen sergeantsand lieutenants from IPD and three sergeants from SPPD).Also, eight patrol officers acting as temporal sergeants inSPPD were excluded, as were all the lieutenanl~ from thisdepartmcnt. For details, see Engel (2000).

R.S. Engel / JoW'nal of Criminal Justice 30 (2002) j /-64 63

5. There was not a complete match between officersand sergeants because not all sergeants from IPD were bothobserved and interviewed (thus, they we~ not categorizedwith a particular supervisory style). Also, in a handful ofcases, a positive match could not be made between theofficer and sergeant because the officer changed shifts duringthe observation period. Although a high percent of patrolofficers from lPD were matched to supervisors with the samework schedule, supervision in this department appeared to bem~ collective in natu~ and concentrated at the shift level.

6. In SPPD, officers were not matched to a specificsergeant because they reported working CPAs that were notidentified by supervisors as being their particular areas ofresponsibility. These particular CPAs were not "studybeats" for the POPN project, therefo~, systematic observa-tion of these officers was limited.

7. Note, however, that the coefficient for the suppor-tive supervisory style is only significant using a one-tailedtest ( P=.O83).

References

Cohen. B. (1980). Leadership styles of commanders in theNew Yortc City Police Department. Journal qf PoliceScit'nce and Administro/ion, 8, 125 - 138.

Dejong, C., Mastrofski, S. D., & Parks, R. B. (2001). Patrolofficers and problem solving: an application of expect-ancy theory. Jus/ice Quarterly, /8 (I), 1-62.

Downton, 1. v. (1973). Rebel leadership: commitmentand charisma in the n'rollllionary process. New York:F fee Press .

Engel, R. S. (200 I ). Supervisory styles of patrol sergeants andlieutenants. Journal of Criminal Jus/ice, 29, 341 - 355.

Engel, R. S. (2000). The effects of supervisory styles on patroloffice behavior. Police Quarterly, 3 (3), 262 - 293.

Gates, S., & Worden. R. E. (1989). Principle-agent modelsof hierarchical control in the public bureaucracies: wortc,shirking, and supervision in police agencies. Confermcepaper delivered at the 1989 Annual Meeting of theAmerican Political Science Association, at the AtlantaHilton and Towers, August 31-September 3.

Goldstein, H. (1990). Proble-oriented policing. New Yortc:McGraw-Hili.

Holsti, O. R. (1968). Content analysis. In: G. Lindzey, &E. Aronson (Eds.), 71Ie handbook of .focial p.fychology

(2nd ed.) (pp. 596-692). Reading, MA: Addison-Wesley Publishing.

Jennier,I. M., & Berkes, L. 1. (1979). Leader behavior in apolice command bureaucracy: a closer look at the quasi-military model. Administrative Science Quarterly, 24,1-23.

Kelling, G. L, & Moore, M. H. (1988). From political to re-fonn to community: the evolving strategy of police. In: 1.R. Greene, & S. D. Mastrofski (Eds.). Commllnity polic-ing: rhetoric or ~ality (pp. 3 -26). New York: Praeger.

Manning. P. K. (1977). Po/ice ~rk. Cambridge, MA:MIT Press.

Mark, 1. (1976). Police organizations: the challenges anddilemmas of change. In: A. Niedertloffer, & A. Blum-berg (Eds.), The ambivalent force: perspe(;'li~s on thepolice (pp. 356~366). Hinsdale, IL: Dryden Press.

Mastrofski, S. D., & Partcs, R. B. (1990). Improving obser-vational studies of police. Criminology, 28, 475-496.

Mastrofski, S. D., Ritti, R. R., & Snipes. 1. B. (1994). Ex-pectancy theory and police productivity in DUJ enforce-ment Law and Society Review, 28, 113-148.

Mastrofski, S. D., Snipes, 1. B., Parks. R. B., & Maxwell C. D.(2000). The helping hand of the law: police control ofcitizens on ~uest Criminology, 38, 307 - 342.

Mastrofski, S D., Worden. R. E., & Snipes, 1. B. (1995).Law enforcement in a time of community policing.Criminology, 33, 539-563.

Muir Ir., W. (1977). Police: street corner politicians.Chicago, IL: University of Chicago Press.

Parks, R. B., Mastrofski, S. D., DeJong, C., & Gray, M. K.(1999). How officers spend their time with the communi-ty. Justice Quarterly, /6, 483 - 5 18.

Reiss Jr., A. 1. (1971). The police and the public. NewHaven. CT: Yale.

Rubinstein, 1. (1973). City police. New York: Ballantine.Sampson, R. I., Raudenbush, S. W., & Earls, F. (1997).

Neighborhoods and violent crime: a multilevel studyofcollcctive efficacy. Science, 277, 918-924.

Allen, D. (1980). S~t-level police supervision: the effectof supervision on police officer activities, agency oul-puts, and neighborhood outcomes. PhD dissertation, In-diana University. Ann Arbor, MI: University MicrofilmsInternational.

Allen, D. (1982). Police supervision on the strect: an anal-ysis of supervisor/officer inleraction during the shift.Journal t?fCriminal Justice, 10,91 -109.

Allen, D., & Maxfield, M. (1983). Judging police perfOn11-ante: views and behavior of patrol officers. In: R. Ben-nett (Ed.), Police at work: policy issues and anal)'si,\"(pp. 65-86). Beverly Hills, CA: Sage Publications.

Bass, B. M. (1985). Leadership and perfiJrmance heyondex~lation.r. New York: Free Press.

Bass, B. M., & Avolio, B. J. (1994). Improving organiza-tional effectiveness through transformalionalleadership.Thousand Oaks, CA: Sage Publication.~.

Bayley, D. H. (1994). Police for the futUI-e. New York:Oxford University ~.

Bittner, E. (1983). Legality and workmanship: introductionto control in the police organization. In: M. Punch (Ed.),Control in the police organization (pp. 1 -II). Cam-bridge, MA: MIT ~.

Brehm, J., & Gates, S. (1993). Donut shops and speed traps:evalualing models of supervision on police behavior.American Journal t?f Occupational and OrganizationalPS}'C'hoJogy, 67,69- 78.

Brown, M. K. (1988). Working the street: police discretionand the dilemmas of reform. New York: Russell SageFoundalion.

Bryk, A. S., & Raudenbush, S. W. (1992). Hierarr.'hicallinear models: applicatiOllS and data analysis methods.Newbury Park, CA: Sage Publications.

Burns, J. M. (1978). Leadership, New York: Harperand Row.

Christopher Commission. (1991). Report of the Independent

Commission on the Los Angeles Po/ice Department. LosAngeles, CA: Commission.

R.s. Engel / JOIInIa/ of Criminal Justice 30 (1001) jJ-6464

American police agency. In: M. Punch (Ed.), Control inthe police organization (pp. 275 - 317). Cambridge,MA: MIT Press.

Van Maanen. J. (1984). Making rank: becoming an Ameri-can police sergeant. Urban Life, 13. 155 - 176.

Walker. S. (1993). Docs anyone remember team policing?Lessons of the team policing experience for community~licing. American Joumal of Police. 11.33-55.

Weisburd, D.. McElroy, J.. &. Hardymann. P. (1988). Chal-lenges to supavision in community ~Iicing: observa-tions on a pilot project. American Journal of Police. 7,29-50.

Wilson. J. Q. (1968). Varieties of police behavior. Cam-bridge. MA: Harvard University Press.

Witte, J. H., Travis Ill.. L. F., &. Langworthy. R. H. (1990).Participatory management in law enfon;ement: ~Iiceofficer, supervisor and administrator perceptions. Amer-ican Journal of Police, 9, 1-23.

Wycoff. M. A., &. Skogan. W. G. (1994). The effect of acommunity policing management style on officers'attitudes. Crime and Delinquency. 40. 371 - 383.

Shennan. L.. Milton, C.. & Kelley. T. (1973). Team policing:seven case studies. Washington, DC: Police Foundation.

Skolnick, J. H.. & Bayley. D. H. (1986). The new blueline: police innovation in six American cities. NewYork: F ~ Press.

Smith, D. (1984). The organizational context of legal con-trol. Criminology. 22. 19-38.

Smith. D. A.. & Klein. J. R. (1983). Police agency character-istics and ~ decisions. In: G. P. Whitaker, & C. D.Phillips (Eds.). Evaluating performance of Criminal Jus-tice Agencies (pp. 63-97). Beverly Hills, CA: Sage.

Sparrow, M. K.. Moore. M. H.. & Kennedy. D. M. (1990).Beyond 91/: a new era for policing. New York:Basic Books.

Tim, L. (1971). Comparative police supervision systems: anorganizational analysis. PhD dissertation, Univeristy ofIllinois at Urbana-Champaign. Ann Arbor, MI: Univer-sity Microfilms International.

Trojanowicz. R. C. (1980). The environment of theFirst-Line Police Supervisor. Englewood Cliffs, NJ:

Prentice Hall.Van Maanen. J. (1983). The boss: fIrSt-line supervision in an