New York Police Diversity Management
finishit
ContentServer.asp-13.pdf
RESEARCH ARTICLE
Assessing the long-term effects of officer race on police attitudes towards the community: a case for representative bureaucracy theory
James R. Lasleya*, James Larsonb, Chandrika Kelsob and Gregory Chris Browna
aDepartment of Criminal Justice, California State University at Fullerton, Fullerton, CA, United States; bDepartment of Criminal Justice, National University, La Jolla, CA, United States
This study examines longitudinal changes in police officer attitudes towards the com- munity within the theoretical context of representative bureaucracy theory. The sample consists of 405 male Caucasian, African-American and Latino/Hispanic patrol officers who began employment with the LAPD from 1985 to 1991 under affirmative action hiring. The same officers comprising the panel study were surveyed at two points in time, 1992 (Wave 1) and 2007 (Wave 2). Results indicated that, over the study’s 15 year time frame, African-American and Latino officers significantly increased their desire to engage in active representation or ‘partnerships’ with the community. Also discovered was a similar time effect among Caucasian officers, who at Wave 1 of the study held significantly lower desires for community interaction than the minority officer sample. Overall, the study findings lend support to representative bureaucracy theory’s general assumption that establishing racial parity between police and citizens may increase the willingness of officers to represent the interests of others with similar demographic backgrounds.
Keywords: LAPD; policing; police; officers; minorities; diversity
Introduction
This study examines, and attempts to explain from a theoretical perspective, the long- term effects of police officer attitudes towards the community they serve as a function of officer race and ethnicity. The study setting is the LAPD (Los Angeles Police Depart- ment), which imposed mandatory minority officer recruitment, hiring and promotion poli- cies beginning in the early 1980s as a means of creating racial balance between the Department and the community it served for purposes of improving relations between police and citizens. The study design is unique in that it involves longitudinal data gath- ered from the same police officers in 1992 and in 2007. This ‘panel’ study design enables the assessment of both immediate and long-term attitude change between minority and non-minority officers. It also enables the statistical evaluation of differences in attitudes between Caucasian, African-American and Latino officers’ attitudes towards their percep- tion of the community they serve while in transition from LAPD’s traditional policing culture in 1992 to a new community-based policing culture which began in mid-1990. From a theoretical standpoint, this study provides a longitudinal test of the basic underly- ing assumption of representative bureaucracy theory that, by increasing racial parity between an organizational workforce and population it serves (i.e. ‘passive representa- tion’), representation of the interests (i.e. ‘active representation’) of clients with a similar
*Corresponding author. Email: [email protected]
Police Practice and ResearchAquatic Insects Vol. 12, No. 6, December 2011, 474–491
ISSN 1561-4263 print/ISSN 1477-271X online � 2011 Taylor & Francis http://dx.doi.org/10.1080/15614263.2011.589567 http://www.tandfonline.com
demographic background will be increased as well. Thus, by extending this theory to the question of police organizational change and community relations, it can be argued that the infusion of minority officers into the ranks of a predominantly Caucasian, traditional policing culture (such as the LAPD) can indeed facilitate attitudinal change favoring community-based policing ideologies and methods. This line of inquiry bears particular importance for policing agencies both inside and outside of the United States that are cur- rently undergoing, or at least contemplating, hiring policies to increase racial and ethnic diversification of its officer workforce in an effort to improve police–community relations.
Representative bureaucracy theory and policing
In general, representative bureaucracy theory (hereafter referred to as ‘RBT’) states that, the greater the racial parity between an organization and the community it serves, the more likely the organization’s workforce will be willing to represent the interests of per- sons with whom they share similar demographic backgrounds (Seiden, 1997). This is based on the assumption that persons with shared racial demographics also share com- mon values and beliefs which, in turn, make them more sensitive to each other’s cultural and social needs (Meier & Stewart, 1992). Furthermore, RBT contends that there are two levels on which organizational change can take place as the result of racially diversifying the organization’s workforce. The first, known as ‘passive representation’, refers to the act of balancing the racial composition of the organization to the racial proportions of the population it serves. This is the necessary initial step to the second level, or ‘active repre- sentation’. Active representation results when the organization’s common demography with the community becomes the catalyst for workers to become advocates of policies and increase their use of discretion in favor of persons of their own race (Hindera, 1993). However, both the community and workplace environment must foster the type of organi- zational and operational change that enables this second level, or ‘active representation’, to be achieved (Meier et al., 1999).
The substantive ideas of RBT can easily be extended to policing organizations. First, it can be argued that, by improving the passive representation of the police organization’s workforce (i.e. improving its racial parity with the community it serves), increased positive ‘active representation’ will be attained. In other words, improving demographic parity between police and the community may perhaps result in an increased likelihood that offi- cers will use their discretion to positively increase the level of service they provide to citi- zens of their own race. The latter assumption is based on the RBT contention that officers and citizens also share beliefs and values that are the unique product of similar socialization experiences. However, in order for a police organization’s racial diversification efforts to improve police–community relations, RBT supposes that the police organization itself must advocate such actions through the passage and support of policies that support active repre- sentation efforts by their officers. In addition, the need for improved levels of police service must be a salient issue for both the police organization and its civilian constituents.
This study examines RBT within the substantive context of police officer attitudes towards the community. That is, traditional aspects of active participation that would involve the analysis of specific changes in actual officer-level behavior towards citizens as a result of racial and ethnic commonalities are not addressed. Rather, it is here assumed that attitudinal measures serve as an indicator of active representation among officers. In other words, a critical assumption of this analysis is that that increased police–community demographic balancing (i.e. passive representation) will produce
Police Practice and Research: An International Journal 475
attitude change among officers that favors policing and practices that improve service to minority communities and the community at large. Thus, it is similarly assumed that atti- tude change among officers is favoring police–community partnerships is a necessary ‘mental’ first step in the active representation process.
The representative bureaucracy and police attitudes
Very little research evidence is available that addresses relationships between police atti- tudes towards the community and the various RBT hypotheses raised regarding demo- graphically produced organizational change. For the most part, the handful of existing police studies focusing on this subject can be categorized as either supporting or oppos- ing the hypothetical claims of RBT. Research in support of RBT suggests that efforts to improve demographic representation in policing agencies will result in positive officer attitude change and perhaps lead to improved and more equitable community service by police to the minority community. Proponents of this research contend that the current lack of racial and ethnic diversity within police organizations contributes to the formation of an organizational culture wherein officers learn to believe that it is acceptable to treat minority citizens differently than their nonminority counterparts (Dowler, 2005; Eterno, 2003; Wilson, 1968). Furthermore, some theorists and policy-makers claim that this insti- tutionalized form of racially disparate treatment by police can be evidenced by many neg- ative varieties of police behavior (Wilson, 1968), including brutality, selective enforcement of laws and a general insensitivity towards specific cultural needs of the minority community (see Christopher Christopher Commission, 1991).
Also in accord with the RBT paradigm is the notion that police insensitivity directed towards minority citizens is due, in part, to a long history of recruiting and hiring Caucasian working-class male recruits that hold pre-employment attitudes and demographics that are similar to those held by existing officers and the police organizational culture at large. This trend, according to Chappell (2009), has perpetuated a policing culture that has remained racially unbalanced for generations. Many researchers argue that the only sure way to break the cycle of ‘all cops thinking alike’ is to change police hiring practices so that they result in the proactive employment of officers who have attitudinal perspectives that are divergent from those held by the dominant traditional policing culture (Haarr & Morash, 1999; Crank, 2004). Some studies have suggested that ethnocentric enforcement attitudes instilled in offi- cers by a traditional policing culture can be lessened, or perhaps eventually be eliminated, by increasing the numbers of minority officers within the ranks of predominantly Caucasian departments (Sun & Payne, 2004). In his pioneering study of minority police officers, Alex (1969) suggested that increased hiring of African-Americans would introduce ‘different’, more racially sensitive attitudes to traditional police culture because of these officers’ socially divergent backgrounds and experiences.
Congruent with RBT assumptions, and irrespective of the need for internal organiza- tional change within police organizations, various arguments have been raised for the hiring of minority officers on the general basis that their unique cultural background will cause them to be more perceptive than Caucasian officers to the law enforcement needs within neighborhoods of color (Kuykendall & Burns, 1980). Other studies have raised similar claims that minority officers have unique perspectives on their roles as police officers, their modes of interaction with the public, their status in police departments and their relation- ships with co-workers (Campbell, 1980). Hawkins (2001) discovered that officers of color were less likely than Caucasian officers to feel depressed or negative about their work as the result of enduring negative treatment by community members and co-workers.
476 J.R. Lasley et al.
Researchers in opposition to the hypothetical claims of RBT assume that, regardless of efforts to increase the ranks of minority officers within a policing agency, the domi- nant traditional policing culture will continue to prevail over officers’ perceptions of their role as police officers in the community. Thus, despite efforts to increase ethnic diversity within the ranks, no changes in police officer attitudes will result. This counter claim to the RBT model is based on several theoretical assumptions. First, with regard to hiring practices, agencies may attempt to only hire minority candidates who reflect the status quo thinking of the police organization. Second, through socialization processes within the organization, minority officers who do hold different perceptions of their policing role prior to being employed as a police officer may eventually abandon these attitudes in favor of more traditional outlooks as the result of ‘conforming’ or being socialized into the traditional policing culture. Third, although minority officers may resist being social- ized into the traditional beliefs of the non-minority officer value system (Paoline, Myers, & Worden, 2000; Paoline, 2004), they may be marginalized or the victim of what Wertsch (1998) refers to as police officer ‘tokenism’.
In general, opponents to the RBT perspective claim that, even if minority officers are hired for purposes of fulfilling a quota or presenting an ethnically diverse agency image to the public, these officers will be given few opportunities to express their divergent beliefs within the policing organization or at the community level. Thus, the increased representation of minority officers within status quo policing agencies will have little, if any, impact on ‘unfreezing’ traditional attitudes held by officers towards their role as law enforcers or their outlook towards the citizens they serve. For the most part, these researchers predict that the application of RBT as a model for increasing officers’ attitudi- nal sensitivity to the beliefs and values of the minority community within established police organizations will inevitably fail because producing demographic change in the organization’s workforce will not be sufficient to overcome the institutionalized and dom- inant non-minority policing culture (Wilkins & Williams, 2008).
Study objective and setting
The primary objective of this study is to examine the basic assumptions of RBT within a major policing agency that has undergone significant and prolonged racial diversification (i.e. passive representation) of its officer workforce and to measure any effect of this organization change (i.e. active representation) in the form of attitudinal change among both minority and non-minority officers. To carry out this objective, The Los Angeles Police Department (LAPD) was selected as the study setting. For many reasons, the LAPD organization provides an ideal research environment for the examination of attitu- dinal change within a traditional American policing culture produced by the infusion of minority officers. Not only is the LAPD America’s second largest policing agency, but it is also considered by police historians to be the birth place of the ‘professional model of policing’. This model, which is commonly known as the ‘Dragnet Model’, was instituted in the Department in 1950 and operated until 1993. For numerous decades the LAPD ‘Dragnet Model’ of operation called for police officers to ‘detach’ themselves from involvement with community members, and the Department was reputed to have the nation’s strongest traditional organizational policing culture. In addition, for decades, the majority of officers (80%) were male Caucasians who were hired from working-class or military backgrounds.
As the result of a 1981 lawsuit filed against LAPD, Fanchon Blake v. City of Los Angeles, the ethnic composition of LAPD patrol officers was changed radically as a
Police Practice and Research: An International Journal 477
result of court-mandated hiring of minority and women recruits (Felkenes & Unsinger, 1992). Reforms in LAPD hiring practices following the Blake decision were imple- mented as a result of recommendations from the Christopher Commission, which pro- vided oversight of the Department’s policies following the 1992 Los Angeles riots. Taken together, these policy changes resulted in court-imposed hiring quotas for the LAPD that required the proactive hiring of minority officers at a rate nearly five times that of Caucasian male officers. The predominant theory behind these radical shifts in LAPD hiring practices was that the infusion of minority officers into the Department’s nearly all non-minority ranks would not only ‘break the back’ of the nation’s oldest and largest traditional policing culture, but also bring much needed diversity in thinking to the LAPD and promote stronger police–community relations (Felkenes & Unsinger, 1992).
Study background and design
The study design is longitudinal and utilizes a two-wave panel analysis. This methodol- ogy simply compares data for the same officers over two separate time periods (i.e. ‘waves’), one at an early point in their career and the other at a much later point in time. The first wave (Wave 1) represents data initially collected from LAPD officers during January 1992. Throughout the beginning and latter part of 2007, data for the second wave (Wave 2) were collected from the same LAPD officers originally partici- pating in the 1992 data collection wave. The attrition rate between Waves 1 and 2 was estimated at approximately 31.6% (Wave 1, N = 621; Wave 2, N = 425). Such rates of attrition are typical of panel designs with data collection wave separated by large amounts of time (Wooldridge, 2001). For reasons explained in the discussion of study findings, it is here assumed that selection bias was not a serious threat to the general study conclusions.
Data and sample
Wave 1: 1992
Data used here were collected as part of a larger personnel study conducted within the Los Angeles Police Department. These data were collected with the cooperation of the LAPD; however, tabulation and analysis were performed independent of any departmen- tal influences. Sampling was carried out using a multistage cluster design with the main objective of creating a probability sample of all LAPD patrol officers. First-stage selec- tion involved random selection of LAPD patrol districts; selection at the second- and third-stage involved random selection at the patrol unit and shift levels, respectively. In all, approximately 1500 (N = 1483) active patrol officers were sampled. These officers represented slightly less than 50% of the active LAPD patrol force. With respect to eth- nicity and length of service, LAPD patrol officer population parameters were within five percentage points of statistics representing the final sample: 32% were ethnic minorities and the median length of experience was 12 years.
For the immediate study, a subset of 425 male patrol officers was selected who were hired by the LAPD between 1985 and 1991. At the time of the original study, the range of experience levels for these officers was 1–7 years (excluding those officers on first-year probation). Broken down by self-reported officer ethnicity, the sample was approximately 43% (N = 184) Caucasian, 32% were Latino and 25% (105) were
478 J.R. Lasley et al.
African-American. The study ethnicity percentage breakdown compared with the present ethnic composition of LAPD under-represents Caucasians by 3% (LAPD = 46%) and Latinos by 1% (LAPD = 33%), and over-represents African-Americans by 11% (LAPD = 14%).
Wave 2: 2007
Officers in Wave 1 were contacted via telephone, email and/or personal meetings to soli- cit participation in the follow-up phase of the study. Approximately one out of every three of the original participants in the first wave were disqualified (no longer worked a patrol assignment), unavailable or refused to take part in the 2007 data collection effort. This resulted in the creation of a Wave 2 minority officer cohort of the same officers in Wave 1, but now with experience ranging from 16 to 22 years. Although the exact extent to which attrition has led to bias in the study panel data is not known, statistical compari- sons between the race and age of officers included and excluded from the second wave sample failed to reveal significant differences.
Survey instrument
Instrumentation consisted of a self-administered questionnaire. The instrument, during Wave 1, was introduced to sample subjects during routine roll-call via a 10 minute video- tape prepared by then Chief of Police Daryl Gates. Officers’ responses were placed in a response envelope after completion and were not viewed by their supervisors at the roll- call session. Substantive content of the videotape included research goals and objectives, general instructions and the voluntary nature of study participation. Officers electing to participate were given the option of completing the questionnaire during roll-call (which was extended for survey purposes) or at the location of their choice. A modified version of the 1992 instrument was used to collect data for Wave 2. The final version was self- administered, as was the first, and was reduced from 5 pages to 1 page in length. Meth- ods used for Wave 2 survey distribution included both personal interviews and mail-out questionnaires.
Measures
Independent variables
Independent variables examined in this study include six measures of officer attitudes towards the potential benefits and hazards of forming community-based policing partner- ships with citizens. Substantively, these items reflect both the ‘pros’ and ‘cons’ of enhanced police involvement with the community. One dependent variable measuring overall officer attitudes towards initiating a closer relationship with community members was also included. These items are presented below:
In your opinion, closer personal contact between police and the community will . . .
‘Pro’ Items:
� Improve citizen’s opinions of police officers � Make citizens more willing to help police stop crime � Make officers more sensitive to community members and their needs.
Police Practice and Research: An International Journal 479
‘Con’ Items:
� Give citizens too much control over officers � Result in officers knowing more about citizens’ private lives than is really necessary � Be less safe for officers than current patrol methods
Dependent variable
One five-point Likert-style dependent variable was included in the present study. This was as follows:
All things considered, do you think it is a good idea or bad idea that the police develop per- sonal contacts with the community?
Responses to each independent variable were rated on a five-point Likert scale, rang- ing from ‘1’ (not important at all) to ‘5’ (extremely important). Responses to the depen- dent criterion measure were also Likert-style ranging from ‘1’ (Very bad idea) to ‘5’ (Very good idea).
Findings
Descriptive analyses
Officer ethnicity group analyses: Wave 1
Means and between-group t-tests for all study variables representing Wave 1 (1992), controlling for officer ethnicity, are presented in Table 1. These results, on a descrip- tive level, support the general assumption underlying RBT theory that minority offi- cers possess attitudes more favorable to methods of policing involving increased community contact and interaction than do nonminority officers. Mean scores for Latino and African-American officers for all three ‘Pro’ police–community contact measures were higher than those obtained for non-minority officers, indicating that minority officers held a more favorable outlook towards increased officer involvement with citizens. Results of independent group t-tests revealed that these observed mean differences were highly statistically significant (p < 0.05 or greater). Interestingly, sim- ilar t-test comparisons between both minority officers groups on the ‘Pro’ variable set revealed that the two minority officer groups had statistically equivalent means for one of the variables (‘Improve citizens’ opinions of officers’), yet had significantly higher mean scores (p < 0.05) on the other two ‘Pro’ measures (‘Citizens more will- ing to help police’; ‘Officers more sensitive to citizens’). This finding may in fact be a statistical artifact owing to the time when these data were collected, which was less than 1 year after the LAPD controversy over the Rodney G. King beating occurred in Los Angeles. Political fallout over this event may have served to increase African- American officers’ sensitivity towards improving their relations with the Los Angeles community, and the minority community at large.
Mean comparisons for the ‘Con’ police–community contact measures revealed a simi- lar pattern to that discovered for the ‘Pro’ variable data set. For all three measures, minor- ity officers were less likely than non-minority officers to have a negative perception towards the prospect of increasing their contact with community members. Again, t-tests comparing means obtained for each ‘Con’ variable between the minority and non-minority
480 J.R. Lasley et al.
T ab le
1 .
M ea n s an d b et w ee n g ro u p s t- te st re su lt s fo r al l o ffi ce rs , W av e 1 (1 9 9 2 ).
O ffi ce r et h n ic it y
V ar ia b le
L at in o
A fr ic an -A
m er ic an
C au ca si an
B et w ee n g ro u p s t- sc o re s
(1 )
(2 )
(3 )
1 v s. 3
2 v s. 3
1 v s. 2
P ro
it em
s Im
p ro v e ci ti ze n s’
o p in io n s o f o ffi ce rs
3 .8 8
(1 .1 3 )
4 .1 4
(0 .9 3 )
3 .4 2
(1 .3 9 )
4 .7 1 * * *
3 .1 8 * *
�1 .8 9
C it iz en s w il l h el p p o li ce
w it h cr im
e 3 .6 3
(1 .1 2 )
4 .0 6
(0 .7 3 )
3 .3 6
(1 .2 2 )
5 .3 0 * * *
1 .9 7 *
�3 .6 2 *
O ffi ce rs
m o re
se n si ti v e to
ci ti ze n s
3 .2 5
(1 .1 1 )
3 .6 0
(1 .0 5 )
3 .0 4
(1 .1 0 )
4 .1 9 * * *
1 .6 5 *
�2 .4 8 *
C o n it em
s G iv e ci ti ze n s to o m u ch
co n tr o l
3 .0 7
(1 .2 6 )
3 .1 4
(1 .2 5 )
3 .6 1
(1 .3 2 )
�2 .9 7 * *
�3 .6 7 * * *
�0 .4 2
In v ad e ci ti ze n s’
p ri v ac y
3 .0 4
(1 .3 1 )
3 .1 4
(1 .1 3 )
3 .5 3
(1 .4 0 )
�2 .4 2 * *
�3 .1 6 * *
�0 .6 1
L es s sa fe
fo r o ffi ce rs
3 .1 5
(1 .0 3 )
3 .2 0
(1 .0 9 )
3 .6 7
(1 .1 1 )
�3 .5 2 * * *
�4 .2 6 * * *
�0 .3 3
C ri te ri o n m ea su re
O ffi ce rs
cl o se
to co m m u n it y
3 .9 1
(1 .0 8 )
3 .9 1
(1 .2 4 )
3 .2 3
(0 .8 7 )
4 .1 8 * * *
4 .8 0 * * *
0 .0 3
N o te : S ta n d ar d er ro rs
p re se n te d in
p ar en th es es
b el o w
m ea n s. * * * p < 0 .0 0 1 ; * * p < 0 .0 1 ; * p < 0 .0 5 .
Police Practice and Research: An International Journal 481
officer groups revealed significantly lower scores (p < 0.05 or greater) for minority officers for all three variables. This finding indicated that minority officers were less skeptical about a negative outcome from their increased contact with citizens than were their non-minority officer counterparts. No statistically significant differences were discovered for t-tests performed on ‘Con’ variable means between African-American and Latino offi- cers for the ‘Con’ variables. This finding suggests that both groups shared similar ‘less skeptical’ attitudes towards increased police–community involvement.
The final Wave 1 comparison for the study criterion measure, ‘Increased community closeness,’ revealed no statistically significant differences between the means representing the African-American and Latino/Hispanic officers. Means for both minority officer groups, however, were significantly higher than that discovered for the Caucasian offi- cers. Again, these findings provide tacit support for the ‘change hypothesis’ assertion that ethnic officers may possess unique personal backgrounds that increase their professional belief that police should take a more active role in working with community members to help solve crime and social problems.
Officer ethnicity group analyses: Wave 2
Table 2 presents means and t-test comparisons for all officer ethnicity groups during Wave 2 (2007) of the study. Overall, these results once again provide support for the RBT hypothesis that ethnic diversification of a traditional policing agency is a potential facilitating source for police officer attitude change. This argument is buttressed statisti- cally by the overall data presentation in Table 2, indicating that nearly all mean scores between officers of all ethnicities during Wave 2 (2007) failed to exhibit statistically significant differences. (The only exception to this overall trend was the statistically sig- nificant mean differences between Caucasian and Latino officers between African-Ameri- can officers on the ‘Pro’ measure of ‘increased sensitivity towards citizens – again perhaps a carry-over effect of the Rodney King incident.). This finding of attitudinal par- ity between all officer races during Wave 2, compared with the finding of attitudinal dif- ference between minority and nonminority officers in Wave 1, suggests that Caucasian officers may have changed their attitudes towards police–community interaction as a result of prolonged exposure to the influences of ethnic diversity within the LAPD orga- nization. This finding, however, must be considered speculative at best considering the limitations of the present study with regard to testing for direct causal effect of diversity on organizational attitude change.
Percentage change analysis: Waves 1 and 2
Descriptive analyses presenting mean differences and between Waves 1 and 2 tested by matched pairs within-group t-tests, and the percentage of increase or decrease in officer attitude scores between these time periods, are presented in Table 3. These data indicate that all officers, regardless of their ethnicity, expressed more favorable attitudes towards police–community interaction during Wave 2 compared with Wave 1. All of these dif- ferences were discovered to be highly statistically significant at or beyond the 0.001 level. However, the shifts in attitudes for Caucasian officers between the two study time periods were, on average, nearly five times those evidenced for minority officers. Comparatively speaking, the greatest percentage change in attitude measures among non-minority officers was revealed for the ‘Con’ variable set – indicating a marked decrease in ‘skepticism’ from Wave 1 to Wave 2 regarding the notion of increased
482 J.R. Lasley et al.
T ab le
2 .
M ea n s an d b et w ee n g ro u p s t- te st re su lt s fo r al l o ffi ce rs , W av e 2 (2 0 0 7 ).
O ffi ce r et h n ic it y
V ar ia b le
L at in o
A fr ic an -A
m er ic an
C au ca si an
B et w ee n g ro u p s t- sc o re s
(1 )
(2 )
(3 )
1 v s. 3
2 v s. 3
1 v s. 2
P ro
it em
s Im
p ro v e ci ti ze n s’
o p in io n s o f o ffi ce rs
4 .2 9
(1 .0 4 )
4 .3 9
(0 .9 3 )
4 .1 7
(1 .3 9 )
1 .7 5
0 .9 7
�0 .7 5
C it iz en s w il l h el p p o li ce
w it h cr im
e 4 .1 6
(1 .0 )
4 .3 5
(0 .7 0 )
4 .0 2
(1 .2 2 )
1 .0 5
1 .1 4
�1 .6 5
O ffi ce rs
m o re
se n si ti v e to
ci ti ze n s
3 .8 3
(1 .1 4 )
4 .1 6
(1 .0 3 )
3 .5 8
(1 .1 0 )
1 .1 7
2 .9 2 *
�2 .2 7 *
C o n it em
s G iv e ci ti ze n s to o m u ch
co n tr o l
1 .3 3
(0 .6 9 )
1 .3 5
(0 .7 3 )
1 .3 4
(1 .3 2 )
0 .0 5
�0 .1 2
�0 .1 5
In v ad e ci ti ze n s’
p ri v ac y
1 .6 8
(0 .7 6 )
1 .7 7
(0 .8 1 )
1 .6 7
(1 .4 0 )
1 .0 3
0 .1 2
�0 .8 6
L es s sa fe
fo r o ffi ce rs
1 .4 7
(0 .7 0 )
1 .4 0
(0 .6 8 )
1 .5 4
(1 .1 1 )
0 .6 2
�0 .8 8
0 .7 8
C ri te ri o n m ea su re
O ffi ce rs
cl o se
to co m m u n it y
4 .6 2
(0 .7 9 )
4 .6 1
(0 .7 6 )
4 .5 8
(0 .8 7 )
0 .3 1
0 .4 0
0 .0 6
N o te : S ta n d ar d er ro rs
p re se n te d in
p ar en th es es
b el o w
m ea n s. * * * p < 0 .0 0 1 ; * * p < 0 .0 1 ; * p < 0 .0 5 .
Police Practice and Research: An International Journal 483
T ab le
3 .
W it h in -g ro u p m ea n co m p ar is o n s (W
av e 1 , 1 9 9 2 to
W av e 2 , 2 0 0 7 ) an d co rr es p o n d in g p er ce n ta g e ch an g e b y o ffi ce r ra ce .
V ar ia b le
L at in o
A fr ic an -A
m er ic an
C au ca si an
W 1 – W 2 %
W 1 – W 2 %
W 1 – W 2 %
Im p ro v e ci ti ze n s’
o p in io n s o f o ffi ce rs
0 .4 1
1 0 .6
0 .2 5
6 .0
0 .7 5
2 1 .9
C it iz en s w il l h el p p o li ce
w it h cr im
e 0 .5 4
1 4 .9
0 .2 9
7 .1
0 .6 6
1 9 .6
O ffi ce rs
m o re
se n si ti v e to
ci ti ze n s
0 .5 8
1 7 .8
0 .5 6
1 5 .5
0 .5 4
1 7 .8
G iv e ci ti ze n s to o m u ch
co n tr o l
�1 .7 4
�5 6 .6
�1 .7 9
�2 .2 7
�5 7 .0
�6 2 .9
In v ad e ci ti ze n s’
p ri v ac y
�1 .3 6
�4 4 .7
�1 .3 7
�4 3 .6
�1 .8 6
�5 2 .7
L es s sa fe
fo r o ffi ce rs
�1 .6 8
�5 3 .3
�1 .8 0
�5 6 .2
�2 .1 3
�5 8 .0
O ffi ce rs
cl o se
to co m m u n it y
0 .7 1
1 8 .1
0 .7 0
1 7 .9
1 .3 5
4 1 .8
N o te : P ai re d t- te st re su lt s fo r al l m ea n d if fe re n ce
co m p ar is o n s sh o w n , w it h in
ra ce
ca te g o ri es , ac h ie v ed
st at is ti ca l si g n ifi ca n ce
at o r b ey o n d th e 0 .0 0 1 le v el .
484 J.R. Lasley et al.
interaction between police and the public. Perhaps most telling among these findings is the fact that the criterion measure of ‘community closeness’ increased approximately 40% for Caucasian officers from 1992 to 2007. This finding again lends support to the RBT assumption that the minority officers possessed police–community attitudes during 1992 that were nearly at a standard that had to be met by Caucasian officers some 15 years later.
Predictive analyses
Procedure and variable transformations
A series of ordinary least squares (OLS) regression models were used to determine the relative contributions of the police officer race as well as the ‘pro’ and ‘con’ variables for predicting officer attitudes towards engaging in community partnerships. These anal- yses were carried out by regressing the overall ‘Community closeness’ dependent vari- able on three independent variables. The first of these measures consisted of a trichotomous race variable which was dummy coded so that Caucasian officers served as the primary officer ethnic group (or ‘criterion’) against which the other two minority officer groups were statistically compared. The second and third independent variables consisted of two composite measures generated from factor scores reflecting the com- bined effects of all three ‘pro’ and ‘con’ independent measures (hereafter referred to as Pro Factor and Con Factor). This transformation was done in order to alleviate the multicollinearity (ranging from 0.48 to 0.76) discovered between individual items com- prising the ‘pro’ and ‘con’ measures.
Regression results: Wave 1
Results for three separate OLS models fitted to the cross-sectional data of 1992 are presented in Table 5. Model 1 includes the race effect only. The results indicate a highly significant race effect for both the African-American and Latino officer ethnic- ity predictors (p < 0.0001). The standardized regression coefficients associated with these effects indicate that African-American officers were only slightly more positive in their attitudes towards their involvement with the community than were Latino offi- cers (b = 0.25 and b = 0.23, respectively). Model 2 adds in the effects of the Pro Factor variable. This effect was also statistically significant (p < 0.0001), and the associated standardized regression coefficient and improvement in the fit of the model (b = 0.39, R2 = 0.20) indicated that the relative importance of pro-community policing attitudes was greater than that of officer race alone for influencing the dependent mea- sure of overall police attitudes towards community involvement. The final model, Model 3, includes all effects by adding the Con Factor predictor. The b coefficient and significance of this predictor (b = �0.19; p < 0.0001) revealed that this variable also produced a reliable change in the dependent measure of ‘community closeness’, but that this effect was approximately one-half that of the effect produced by the Pro Factor measure. Substantively speaking, results from this regression series provide additional support for the RTB conceptual model by indicating that, after all officers’ attitudes towards the pros and cons of police community interaction are held constant (or presumed to be equal regardless of race), Latino and African-American officers still had more positive attitudes towards closer community involvement than did Caucasian officers.
Police Practice and Research: An International Journal 485
T ab le
4 .
F ac to r an al y si s re su lt s an d fa ct o r lo ad in g s, W av e 1 an d W av e 2 .
V ar ia b le
W av e 1
W av e 2
F ac to r 1
F ac to r 2
F ac to r 2
F ac to r 1
Im p ro v e ci ti ze n s’
o p in io n s o f o ffi ce rs
0 .8 0
�0 .2 4
0 .9 2
�0 .1 2
C it iz en s w il l h el p p o li ce
w it h cr im
e 0 .8 9
0 .0 1
0 .8 4
�0 .1 3
O ffi ce rs
m o re
se n si ti v e to
ci ti ze n s
0 .8 5
�0 .1 2
0 .8 6
�0 .1 0
G iv e ci ti ze n s to o m u ch
co n tr o l
�0 .2 0
0 .7 7
�0 .1 0
0 .8 9
In v ad e ci ti ze n s’
p ri v ac y
�0 .2 1
0 .7 9
�0 .0 6
0 .8 0
L es s sa fe
fo r o ffi ce rs
0 .0 7
0 .8 1
�0 .2 0
0 .8 6
V ar ia n ce
ex p la in ed
2 .2 7
1 .9 7
2 .3 7
2 .2 3
C o m m u n al it y
4 .2 4
4 .6 1
N o te : It al ic iz ed
lo ad in g s u se d to
id en ti fy
v ar ia b le s, w h ic h co n si d er ed
to g et h er , p ro v id e su b st an ti v e m ea n in g to
ea ch
fa ct o r.
486 J.R. Lasley et al.
Regression results: Wave 2
Three similar OLS regression equations were estimated for the Wave 2 data to assess the statistical significance and relative importance of predictor variable for officer attitude during the year 2007 (see Table 6). In Model 1, the race effects-only equation, both offi- cer ethnicity predictors failed to achieve statistical significance. Adding the Pro Factor predictor to this model revealed a significant improvement in the fit of this model (R2 = 0.15) over the previous one (R2 = 0.00), indicating that attitudes relating to the Pro Factor measure produced significant improvement in the community-closeness dependent mea- sure. Addition of the final predictor, the Con Factor, contributed significantly to the fit of the model (R2 = 0.25), and suggested that the effect of attitudes related to this measure (b = �0.32) produced virtually the same impact on the community closeness criterion as did the Pro Factor measures (b = 0.32). Taken together, these results suggest that the unique effects of officer race may diminish in a police organizational environment where traditional outlooks towards policing and the community have also diminished. How this change occurs, however, can only be speculated on rather than proven, given the method- ological limitations of this study.
Fixed effects regression results: Wave 1 vs. Wave 2
A series of fixed-effects regression models were fitted to the full data set to test for sig- nificant changes in the minority officers’ community attitudes over time. The first set of fixed-effects regression models, three OLS regression models presented in Table 7, test for statistically significant change in officers’ attitudes between both time periods, Wave 1 (1992) and Wave 2 (2007). The two statistically significant parameter estimates observed for the race effects-only model, Model 1, indicated that the degree of attitude change from Wave 1 to Wave 2 was approximately the same for both African-American and Latino officers (b = �0.20 and �0.18, respectively) – above and beyond that of Cau- casian officers. In Model 2 the change measure for the Pro Factor also achieved statistical significance (p < 0.05), but failed to do so in Model 3. When all effects are added to the
Table 5. Wave 1 (1992) OLS regression equations.
Variable DF Parameter estimate
Standard error t-value Pr > |t|
Standardized estimate R
2
Model 1: race effects-only model Intercept 1 3.23370 0.09270 34.88 <0.0001 0 African-American 1 0.68542 0.14220 4.82 <0.0001 0.24633 Latino 1 0.68059 0.15380 4.43 <0.0001 0.22615 0.06
Model 2: race and Pro Factor effects model Intercept 1 1.58524 0.20943 7.57 <0.0001 0 African-American 1 0.52872 0.13250 3.99 <0.0001 0.19002 Latino 1 0.35003 0.14704 2.38 0.0177 0.11631 Pro index 1 0.16758 0.01943 8.62 <0.0001 0.38767 0.20
Model 3: all effects model Intercept 1 2.68962 0.33001 8.15 <0.0001 0 African-American 1 0.42002 0.13233 3.17 0.0016 0.15095 Latino 1 0.28090 0.14502 1.94 0.0534 0.09334 Pro index 1 0.14651 0.01967 7.45 <0.0001 0.33894 Con index 1 �0.08291 0.01940 �4.27 <0.0001 �0.19331 0.23
Police Practice and Research: An International Journal 487
right-hand side of the regression equation, only officer race and the Con Factor variables remained significant predictors of the police community relations attitude dependent measure. In concert, these findings suggest that, not only were the observed changes in attitudes owing to race statistically significant over time, but they were also above and
Table 6. Wave 2 (2007) OLS regression equations.
Variable DF Parameter estimate
Standard error
Standardized estimate t-value Pr > |t| R2
Model 1: race effects-only model Intercept 1 4.58696 0.06082 75.42 <0.0001 0 African- American
1 0.03804 0.09330 0.41 0.6837 0.02158
Latino 1 0.03209 0.10091 0.32 0.7506 0.01683 0.00
Model 2: race and Pro Factor effects model Intercept 1 3.20442 0.16305 19.65 <0.0001 0 African- American
1 �0.02198 0.08576 �0.26 0.7979 �0.01247 Latino 1 �0.09958 0.09362 �1.06 0.2881 �0.05223 Pro index 1 0.11734 0.01300 9.02 <0.0001 0.40748 0.15
Model 3: all effects model Intercept 1 4.15633 0.20046 20.73 <0.0001 0 African- American
1 �0.01874 0.08078 �0.23 0.8167 �.010673 Latino 1 �0.07590 0.08824 �0.86 0.3902 �0.03981 Pro index 1 0.09105 0.01275 7.14 <0.001 0.31618 Con index 1 �0.14066 0.01904 �7.39 <0.0001 �0.32326 0.25
Table 7. Wave 1 (1992) vs. Wave 2 (2007) fixed-effects OLS regression equations.
Variable DF Parameter estimate
Standard error t-value Pr > |t|
Standardized estimate R2
Model 1: race effects-only model Intercept 1 1.35326 0.11027 12.27 <0.0001 0 African- American
1 �0.64738 0.16914 �3.83 0.0001 �0.19807 Latino 1 �0.64850 0.18294 �3.54 0.0004 �0.18345 0.03 Model 2: race and Pro Factor effects model Intercept 1 1.18246 0.12950 9.13 <0.0001 0 African- American
1 �0.61019 0.16879 �3.62 0.0003 �0.18669 Latino 1 �0.57384 0.18432 �3.11 0.0020 �0.16233 Pro index 1 0.08779 0.03545 2.48 0.0137 0.11876 0.05
Model 3: all effects model Intercept 1 0.53138 0.15997 3.32 0.0010 0 African- American
1 �0.45482 0.16307 �2.79 0.0055 �0.13916 Latino 1 �0.46122 0.17697 �2.61 0.0095 �0.13047 Pro index 1 0.03867 0.03473 1.11 0.2661 0.05231 Con index 1 �0.11936 0.01859 �6.42 <0.0001 �0.30241 0.13
488 J.R. Lasley et al.
beyond the changes taking place regarding positive and negative outlooks towards increased police–citizen involvement. This model also indicates that reduced officer ‘skepticism’ about community involvement (i.e. the Con Factor) may be the most impor- tant factor (aside from ethnicity) in ‘unfreezing’ traditional policing attitudes. In sum, these findings lend support to the RTB policy position that the increased representation of ethnic minorities in police agencies has provided and will provide a long-term benefit to community-based policing efforts.
Conclusion and policy implications
Findings from this study lend support to the general hypothetical claims of RBT within the context of a major policing organization; namely, the study’s LAPD organization set- ting. Key results are summarized below.
Community attitudes between minority and non-minority officers were significantly different
RTB assumes that minority and non-minority officers will differ in their outlooks towards their role as service providers. Study findings supported this claim. On several levels, atti- tudes of both Latino and African-American officers hired in 1992 by the LAPD under legally mandated affirmative action employment standards were significantly more favor- able towards involvement with the minority community, and the public at large, than were similar attitude measures for Caucasian officers. Thus, the RBT assumptions of passive representation appear to supported as well.
Minority officers’ community attitudes improved over time
A central claim of RBT is that increasing organizational racial diversity will result in active representation, or some form of behavioral change among workers that will positively impact the community that is served by the organization. Although this study did not provide a direct measure of behavioral outcomes, it did provide for measures of attitudinal change among officers over time. Results indicated that Latino and African-American officers not only maintained their more positive outlooks towards the community as compared with their Caucasian counterparts, but also evidenced significant improvement in their attitudes towards the community served during the study’s 15-year time span.
Non-minority officers’ community attitudes improved over time
Perhaps the most unexpected finding of this study was the support of RBT for explain- ing attitude change observed among non-minority officers. That is, Caucasian officers significantly improved their outlooks towards providing service to all community mem- bers to a level that was at a parity with their minority officer co-workers. Thus, it appears as though the infusion of Latino and African-American officers into the tradi- tional policing culture served to ‘unfreeze’ and change many of the pre-existing, less community-oriented attitudes held by non-minority officers. For the RBT perspective, this finding is perhaps the result of active representation in the form of LAPD organi- zational policy changes that are supportive of minority officer viewpoints regarding the police officer’s role in the community.
From a methodological stance, it could be argued that the findings from this study are merely a product of selection bias caused by a good deal of sample attrition from 1992
Police Practice and Research: An International Journal 489
to 2007. This would imply that only those minority officers that (1) have the ability to work with community members and/or (2) are able to cope with a policy changes to community-based policing remained employed with LAPD throughout the 15 study per- iod. Likewise, other less community-minded minority and non-minority officers could have been ‘selected out’ of the study’s second wave panel because they left the LAPD sometime after 1992. Admittedly, it is not known whether officer attrition from these panel data was systematic or random. However, given the nature of positive response pat- terns towards community involvement obtained in 2007 from the now more experienced minority officers, there are strong conceptual grounds for concluding that the study find- ings are indeed valid.
From a policy standpoint, recent attitudinal research has shown that US minority citi- zens generally desire the composition of their police force to have a racial balance between Caucasian and minority officers (Hasisi & Weitzer, 2007; Weitzer, 1996, 2000). Similar research conducted in countries outside of the United States classified as ‘deeply divided’ both politically and racially also suggest similar findings (Hasisi, 2008). Further- more, proponents of ethnic diversification of police argue that such policies may have the potential to reduce ongoing tensions between police and residents within ethnically underrepresented communities whose cultural backgrounds are extremely divergent from those in political power (Weitzer & Hasisi, 2008). This evidence has been produced in Northern Ireland (Weitzer, 1995), Zimbabwe (Weitzer, 1990), Israel (Hasisi, 2008; Weit- zer & Hasisi, 2008) and France (Zauberman & Levy, 2003). Thus, in the case of the present study, the evidenced attitude change among all LAPD officers in 2007 – and par- ticularly among Caucasian officers – may have been due, in part, to LAPD’s traditional policing culture having been transformed through ethnic diversification into a representa- tive bureaucracy.
Acknowledgments The authors wish to thank Dilip Das, Editor of PPR, and anonymous reviewers of PPR for their helpful suggestions regarding substantive and theoretical revisions made to earlier drafts of this study.
Notes on contributors James R. Lasley is Professor of Criminal Justice at California State University, Fullerton.
James Larson is Associate Professor of Criminal Justice Administration at National University and Chair of the Department of Professional Studies.
Chandrika Kelso is Associate Professor of Criminal Justice Administration at California State University, Fullerton.
Gregory Chris Brown is Assistant Professor of Criminal Justice at California State University, Fullerton.
References Alex, N. (1969). Black in blue: A study of the Negro policeman. New York: Appleton-Century-
Crofts. Campbell, V. (1980). Double marginality of Black policemen. Criminology, 17, 477–484. Chappell, A. (2009). The philosophical versus actual adoption of community policing: A case
study. Criminal Justice Review, 34, 5–28. Christopher Commission. (1991). Report of the Independent Commission of the Los Angeles Police
Department. Los Angeles, CA: LAPD. Crank, J. (2004). Understanding police culture. Cincinnati, OH: Anderson.
490 J.R. Lasley et al.
Dowler, K. (2005). Relationship between race and police work. Police Quarterly, 8, 475–489. Eterno, J. (2003). Policing within the law: A case study of the New York Police Department. West-
port, CT: Greenwood. Felkenes, G., & Unsinger, P. (1992). Diversity, affirmative action and law enforcement. Springfield,
IL: Charles C. Thomas. Haarr, R.N., & Morash, M. (1999). Gender, race and strategies of coping with occupational stress
in policing. Justice Quarterly, 16, 303–336. Hasisi, B. (2008). Police, politics, and culture in a deeply divided society. Journal of Criminal
Law & Criminology, 98, 1119–1146. Hasisi, B., & Weitzer, R. (2007). Police relations with Arabs and Jews in Israel. The British Jour-
nal of Criminology, 47, 728–746. Hawkins, H. (2001). Police officer burnout: A partial replication of Maslach’s burnout inventory.
Police Quarterly, 4, 343–360. Hindera, J.J. (1993). Representative bureaucracy: Imprimis evidence of active representation in the
EEOC District Office. Social Science Quarterly, 74(2), 95–108. Kuykendall, J.L., & Burns, D.L. (1980). The black police officer: An historical perspective. Jour-
nal of Contemporary Criminal Justice, 1, 103–113. Meier, K.J., & Stewart, J. Jr. (1992). The impact of representative bureaucracies: Educational sys-
tems and public policies. American Review of Public Administration, 22(3), 157–171. Meier, K.J., Wrinkle, R.D., & Polinard, J.L. (1989). Representative bureaucracy and distributional
equity: Addressing the hard question. Journal of Politics, 61, 1025–1039. Paoline, E.A. (2004). Shedding light on police culture: An examination of officers’ occupational
attitudes. Police Quarterly, 7, 205–236. Paoline, E.A., Myers, S.M., & Worden, R.E. (2000). Police culture, individualism, and community
policing: Evidence from two police departments. Justice Quarterly, 17, 575–605. Seiden, S.C. (1997). The promise of representative bureaucracy: Diversity and responsiveness in a
government agency. Armonk, NY: M.E. Sharpe. Sun, I.Y., & Payne, B.K. (2004). Racial differences in resolving conflicts: A comparison between
Black and White police officers. Crime & Delinquency, 50, 516–541. Weitzer, R. (1990). Transforming settled states: Communal conflict and internal security in North-
ern Ireland and Zimbabwe. Berkeley, CA: University of California Press. Weitzer, R. (1995). Policing under fire: Ethnic conflict and police community-relations in Northern
Ireland. Albany, NY: State University of New York Press. Weitzer, R. (1996). Racial discrimination in the criminal justice system: Findings and problems in
the literature. Journal of Criminal Justice, 24, 309–322. Weitzer, R. (2000). Radicalizing policing: Residents’ perceptions in three neighborhoods. Law and
Society Review, 34(1), 129–155. Weitzer, R., & Hasisi, B. (2008). Does ethnic composition make a difference? Citizens’ assess-
ments of Arab police officers in Israel. Policing and Society, 18, 262–276. Wertsch, J. (1989). Mind as action. New York: Doubleday. Wilkins, V., & Williams, B.N. (2008). Black or blue: Racial profiling and representative bureau-
cracy. Public Administration Review, 68, 654–665. Wilson, J.Q. (1968). Varieties of police behavior. Cambridge, MA: Harvard University Press. Wooldridge, J.M. (2001). Econometric analysis of cross-section and panel data. Cambridge, MA:
MIT Press. Zauberman, R., & Levy, R. (2003). Police, minorities, and the French Republican ideal. Criminol-
ogy, 41, 1065–2001.
Police Practice and Research: An International Journal 491
Copyright of Police Practice & Research is the property of Routledge and its content may not be copied or
emailed to multiple sites or posted to a listserv without the copyright holder's express written permission.
However, users may print, download, or email articles for individual use.
