Labeling the Labelers

Permanent Link: http://ufdc.ufl.edu/UFE0020159/00001

Material Information

Title: Labeling the Labelers A Theoretical Model and Quasi-Experiment of Organizational Deviance Amplification
Physical Description: 1 online resource (87 p.)
Language: english
Creator: Ward, Jeffrey T
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007


Subjects / Keywords: amplification, deviance, labeling, organizational, theory
Criminology, Law and Society -- Dissertations, Academic -- UF
Genre: Criminology, Law, and Society thesis, M.A.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation


Abstract: Labeling theory research has provided important information about the amplifying effects deviant labels can have on individual behavior. Labeling theorists have needlessly restricted their analysis to individual actors. My study blends organizational and labeling theories to develop a theoretical model of organizational deviance amplification. The model specifies the sequence of events culminating in organization deviance amplification and predicts under which circumstances amplification of deviant behavior is most likely to occur. The theoretical model of organizational deviance amplification is partially tested on a novel form of police organizational deviant behavior, namely citation issuance deviancy. Time-series data and a quasi-experimental design are used to test three organizational deviance amplification hypotheses. The main findings are that the ascription of a 'speed trap' label led to a statistically significantly increase in the number of speeding citations issued but the label neither resulted in the police agency issuing tickets with a lesser speed differential nor affected the percentage of out-of-state drivers receiving traffic citations. The partial support found for the theoretical model of organizational deviance amplification necessitates further empirical testing of the model with both mediating variables and larger sample sizes. Directions for future research in organizational deviance amplification are discussed.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Jeffrey T Ward.
Thesis: Thesis (M.A.)--University of Florida, 2007.
Local: Adviser: Lanza-Kaduce, Lonn M.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2007
System ID: UFE0020159:00001

Permanent Link: http://ufdc.ufl.edu/UFE0020159/00001

Material Information

Title: Labeling the Labelers A Theoretical Model and Quasi-Experiment of Organizational Deviance Amplification
Physical Description: 1 online resource (87 p.)
Language: english
Creator: Ward, Jeffrey T
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007


Subjects / Keywords: amplification, deviance, labeling, organizational, theory
Criminology, Law and Society -- Dissertations, Academic -- UF
Genre: Criminology, Law, and Society thesis, M.A.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation


Abstract: Labeling theory research has provided important information about the amplifying effects deviant labels can have on individual behavior. Labeling theorists have needlessly restricted their analysis to individual actors. My study blends organizational and labeling theories to develop a theoretical model of organizational deviance amplification. The model specifies the sequence of events culminating in organization deviance amplification and predicts under which circumstances amplification of deviant behavior is most likely to occur. The theoretical model of organizational deviance amplification is partially tested on a novel form of police organizational deviant behavior, namely citation issuance deviancy. Time-series data and a quasi-experimental design are used to test three organizational deviance amplification hypotheses. The main findings are that the ascription of a 'speed trap' label led to a statistically significantly increase in the number of speeding citations issued but the label neither resulted in the police agency issuing tickets with a lesser speed differential nor affected the percentage of out-of-state drivers receiving traffic citations. The partial support found for the theoretical model of organizational deviance amplification necessitates further empirical testing of the model with both mediating variables and larger sample sizes. Directions for future research in organizational deviance amplification are discussed.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Jeffrey T Ward.
Thesis: Thesis (M.A.)--University of Florida, 2007.
Local: Adviser: Lanza-Kaduce, Lonn M.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2007
System ID: UFE0020159:00001

This item has the following downloads:

Full Text




2007 Jeffrey T. Ward 2


To my father, the late Edward David Ward, Sr. 3


ACKNOWLEDGMENTS First and foremost, I thank the members of my committee, Dr. Lonn Lanza-Kaduce, Dr. Nicole Leeper Piquero, and Dr. Jodi Lane, for their contributi ons to my study. Additionally, I thank the committee for their patience and belief in me as both a theorizer and a researcher. Moreover, I extend a special thanks to my co mmittee chair, Dr. Lonn Lanza-Kaduce, for his insightful ideas and his empoweri ng mentoring style. I acknowledge my fellow graduate students in the Department of Criminology, Law & Society at the University of Florida for their thoughtful ideas and genuine encouragement. I also thank my wife Kerri for her boundl ess moral support and for being a secondary education English teacher. Hence, I am indebted to my wife for editing my study! 4


TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........7 LIST OF FIGURES.........................................................................................................................8 ABSTRACT.....................................................................................................................................9 CHAPTER 1 INTRODUCTION................................................................................................................. .11 2 THEORETICAL FRAMEWORK..........................................................................................15 Labeling Theory................................................................................................................ ......15 Origins of Labeling..........................................................................................................15 Deviance Amplification Hypothesis................................................................................18 Criticisms of the Labeling Perspective............................................................................20 Organizational Theory.......................................................................................................... ..22 Anthropomorphizing Organizations................................................................................22 Identity-Image Interdependence......................................................................................24 Reputation..................................................................................................................... ...25 Organizational Deviance Amplification.................................................................................27 3 REVIEW OF THE LITERATURE........................................................................................32 Labeling Evidence..................................................................................................................32 Mental Patients................................................................................................................32 Juvenile Delinquents.......................................................................................................34 Police, Citizens, and Traffic Stops.........................................................................................36 Conflicts Regarding Speeding.........................................................................................36 Traffic Stops and Outcomes............................................................................................38 4 DATA AND METHODOLOGY...........................................................................................41 Hypotheses..............................................................................................................................41 Data.........................................................................................................................................42 Traffic Citation Data........................................................................................................42 Traffic Volume Data........................................................................................................44 Town Characteristics Data..............................................................................................45 Police Agency Characteristics Data................................................................................46 Methodology...........................................................................................................................46 Interrupted Time-Series Design......................................................................................46 Intervention Specification and Impact Assessment.........................................................49 5


5 RESULTS...................................................................................................................... .........52 Descriptive Analysis........................................................................................................... ....52 Regional Overview..........................................................................................................52 Population baseline comparisons.............................................................................52 Traffic volume baseline comparisons......................................................................54 Local Details....................................................................................................................55 Intervention Effects................................................................................................................57 Hypothesis One...............................................................................................................57 Hypothesis Two...............................................................................................................61 Hypothesis Three.............................................................................................................64 6 DISCUSSION AND CONCLUSIONS..................................................................................68 APPENDIX ACF/PACF PLOTS..............................................................................................75 LIST OF REFERENCES...............................................................................................................80 BIOGRAPHICAL SKETCH.........................................................................................................87 6


LIST OF TABLES Table page 4-1 Alachua, High Springs, and Waldo town comparison.......................................................48 4-2 Alachua, High Springs, and Waldo police department comparison..................................48 5-1 Demographic characteristics of individuals receiving speeding citations in Alachua, High Springs, and Waldo...................................................................................................56 5-2 Differences in speeding citation issu ance among Alachua, High Springs, and Waldo Police Departments............................................................................................................5 7 5-3 Uncontrolled speeding citation i ssuance intervention effects model.................................60 5-4 Controlled speeding citation issu ance intervention effects model.....................................60 5-5 Uncontrolled speed different ial intervention effects model...............................................63 5-6 Controlled speed differential intervention effects model...................................................63 5-7 Uncontrolled percentage out-ofstate intervention effects model......................................66 5-8 Controlled percentage out-of-s tate intervention effects model..........................................66 7


LIST OF FIGURES Figure page 1-1 Image of the deviant labe l (speed trap billboard)..............................................................12 2-1 Theoretical model of organi zational deviance amplification.............................................31 4-1 Interrupted time-series design............................................................................................4 9 5-1 Population baseline 2003 da ily citation issuance rates......................................................53 5-2 Traffic volume baseline 2003 daily citation issuance rates...............................................54 5-3 Monthly speeding citation issuance sequence plots...........................................................59 5-4 Monthly average speed di fferential sequence plots...........................................................62 5-5 Monthly percentage of out-of-state speeding citations sequence plots.............................65 8


Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Arts LABELING THE LABELERS: A THEORETICAL MODEL AND QUASI-E XPERIMENT OF ORGANIZATIONAL DEVIANCE AMPLIFICATION By Jeffrey T. Ward August 2007 Chair: Lonn Lanza-Kaduce Major: Criminology, Law and Society Labeling theory research has provided important information about the amplifying effects deviant labels can have on indivi dual behavior. Labe ling theorists have n eedlessly restricted their analysis to individual actors. My study blends organizational a nd labeling theories to develop a theoretical model of organizational deviance amplification. The model specifies the sequence of events culminating in organization deviance amplific ation and predicts under which circumstances amplification of deviant behavior is most likely to occur. The theoretical model of organizational deviance amplification is partially tested on a novel form of police organizational deviant behavior, namely citati on issuance deviancy. Time-series data and a quasi-experimental design are used to test three organizational deviance amplification hypotheses. The main findings are that the ascripti on of a speed trap labe l led to a statistically significantly increase in the number of speeding ci tations issued but the label neither resulted in the police agency issuing tickets with a lesser sp eed differential nor affected the percentage of out-of-state drivers receiving tra ffic citations. The partial suppor t found for the theoretical model of organizational deviance amplification necessita tes further empirical testing of the model with 9


both mediating variables and larger sample sizes. Directions for fu ture research in organizational deviance amplificati on are discussed. 10


CHAPTER 1 INTRODUCTION Central to the labeling persp ective is the notion that devi ance is an ongoing process of social interaction (Schur, 1971). The process of deviance studied by labe ling scholars involves interaction between an agent of social control an d, typically, an individu al. There are two main types of social control agents; informal agents are entities that act wit hout official capacity (e.g., family, teachers, peers), whereas formal agents ar e entities that act with official capacity (e.g., police, courts). Despite the fact that there are different types of social control agents, they all possess a common ability to ascribe deviant labels albeit with varying impacts and degrees of formality (Grimes & Turk, 1978). Labeling theo rists tend to overlook the motivations of the deviant individual in favor of ex amining characteristics and actions of both informal and formal agents of social control (Lor ber, 1967, p. 302). The status ch aracteristic hypothesis, derived from the conflict perspective, contends that power, both economic and political, influences which entities are labeled and fo r what types of behavior (Paternoster & Iovanni, 1989). Thus, social control agents necessarily play an important role in the study of deviant behavior. Labeling theorists are not only concerned with th e actions of social control agents; they are also interested in determining the effects devi ant labels have on indi viduals identity and behavior. Labeling theorists attempt to an swer the question: what happens to the individual after being labeled? [emphasis added] (Pater noster & Iovanni, 1989). Stemming from the symbolic interactionist traditi on, the deviance amplification hypothe sis states that individuals who are labeled are likely to be excluded from conventional opportunities, alte r their selfidentity, and acquire supportive deviant others, u ltimately leading to an amplification of their deviant behavior (see Paternos ter & Iovanni, 1989). Significa ntly, empirical tests of the deviance amplification hypothesis ha ve focused on the effects labels have on individuals. As a 11


result, it is unclear whether labeling organizatio ns deviant has parallel effects on organizational behavior. By extending labeling theory to the or ganizational actor, the ce ntral question becomes: what happens to the organization after being labeled? More sp ecifically, do organizations that are labeled amplify th eir deviant behavior? In 2003, a small, rural police agency was soci ally branded for engaging in organizational deviancy in speed enforcement. The labeled poli ce agency has a relatively high full-time officer to population ratio and issues a very high rate of citations compared to other municipalities. The deviant label ascribed to the speed trap town is in the form of publicly visible billboards placed outside the city limits (see Figure 1-1) which se rve two main purposes. Fi rst, the speed trap Figure 1-1. Image of the devian t label (speed trap billboard).1 billboards connote that the police agency is enga ging in citation issuance primarily for revenue generation and, thus, labels the sp eed enforcement behavior of th e police organization deviant. Second, the publicly visible billboard directly warns motorists traveling through the town that the chances of receiving a speeding citation are grea tly heightened. The speed trap label was 1 Image taken in 2007 by Jeffrey T. Ward. 12


ascribed by the American Automobile Associat ion (AAA), which is a nationwide travel and automotive support organization that has considerab le power and reputability with a very large clientele. My study contributes to the la beling literature in two principal ways. First, my study blends organizational and labeli ng theories to derive a theore tical model of organizational deviance amplification. Second, my study begins the process of empirically evaluating the theoretical model of orga nizational deviance amplification by examining the effects of a label on police organizational behavior. More specifically, I test whether the citation issuance behavior of a labeled police agency was amplified by the ascr iption of the speed trap label. In typical labeling theory research, police agencies are seen as agents of social control, part of the criminal justice system that ascribes deviant labels to individuals. The police ag ency of interest has received a deviant label; in essence, this is a case of labeling the labe lers. The extension of labeling theory to the organizationa l actor is potentially of great import. The effects labels have on certain organizations may in fundamental ways emulate those on individuals, which would have important implications for official re sponses to organizational deviant behavior. There are three primary reasons empirical tests of the de viance amplification hypothesis have focused on individual behavior and have ig nored organizational behavior. The first reason is that organizations are generally omitted from typical images of deviant actors (Simmons, 1965; see also Liska, 1987; Ermann & Lundman, 2002) The deviance literature is dominated by an emphasis on nuts, sluts, and preverts, to the neglect of deviance committed by the economic and political elite (Liazos, 1972). This is not to say that organizat ional actors canno t be deviant. It is useful to view organizations as capable of committing deviant acts; in fact, they can and do originate deviant actions (Ermann & Lund man, 2002, p. 49). In studies of deviance, 13


organizational actors are often overlooked in favor of individual actors, bu t this fact alone does not mean they are unwort hy of scholarly attention. The second reason organizational be havior has not been examined in this way is the fact that the deviance amplification hypothesis is derived from the concept of symbolic interaction. Blumer (1962, p. 180) notes that the term sym bolic interaction refe rs, of course, to the peculiar and distinctive character of interaction as it ta kes place between human beings. It is clear that Blumer is specifying interaction betw een people and not organi zations. Organizations are often seen as entities that lack the basic human capabilitie s to think and act (Sherman, 1978, p. 6). However, social-psychological concepts have been utilized to explain organizational behavior (e.g., see Staw et al., 1981). Moreover, certain effect s may very well generalize across different units of analysis (Mille r, 1978) and alternating units a nd levels of analysis can expand and elaborate sociological theory (Vaughn, 1992). In light of these arguments, there appears to be good reason to extend the con cept of deviance amp lification to the or ganizational actor. Perhaps the strongest reason scholars have not tested th e organizational deviance amplification hypothesis is that the hypothesis predicts an effect counter to what much organizational researchers would expect. When an organization is labeled, it tends to offer alternative accounts of its actions and works toward rebuilding goodwill (Ermann & Lundman, 2002). It appears that negatively labele d organizations should lessen, or cease, their deviant behavior in order to rebuild goodwill; at the very least, the unfavorable behavior should not be amplified. While most organizations w ill attempt to dodge deviant labels (Ermann & Lundman, 2002, p. 31), label avoidance is premised on the notion that or ganizational reputation is essential to organizational success and surviv al. This is not necessarily the case for all organizations and, consequently, deviance amplif ication emerges as a th eoretical possibility. 14


CHAPTER 2 THEORETICAL FRAMEWORK Labeling Theory Origins of Labeling Labeling theory has its roots in the writings on symbolic interaction. As suggested by the term symbolic interaction, the theory concerns itself with the meaning actions and words have for the parties involved in the interaction. Char les Cooley (1902) develo ps the concept of the looking-glass self. According to Cooley (1902), the self-id ea has three key elements: the imagination of our appearance to the other person; the imagina tion of his judgment of that appearance, and some sort of self -feeling, such as pride or mortification (p. 152). The self idea is not formed from internal sources alone; instea d, one looks at oneself us ing the perceptions of other people. An individual percei ves his or herself how others perc eive him or her. It should be noted that self-identity is typica lly conceived of as a relatively stable construct (Bailey, 2003). The symbolic interactionist notion of the self most resembles the environmental self in which an individuals sense of self is affected by expe rience and social interac tions (Bailey, 2003). Building on the notion of symbolic intera ction, Frank Tannenbaum became the first scholar to discuss the implications of deviant labels. The first dramatization of the evil which separates the child out of his group for specialized treatment plays a greater role in making the criminal than perhaps any other experience (Tannenbaum, 1938, p. 21). The dramatization of evil describes the definitional tran sfer of the concept of evil from an inappropriate act to the actor. The result of the drama tization of evil is that an in dividual who has been tagged assumes the deviant role (p. 21). This first fo rmulation of labeling th eory is particularly noteworthy because it was the first to implicate so ciety in the deviant and criminal actions of individuals. 15


Building on the works of sociologists before him, Edwin Lemert offers a theoretical distinction between two forms of deviance: primary and secondar y. The key distinction between the two types is centered on how the individual perceives his or her ru le breaking behavior. Primary deviance occurs when an individual engages in deviant behavior without encompassing a deviant identity. Primary deviance is rationali zed or otherwise dealt with as functions of a socially acceptable role (Lemert, 1951, p. 75). Moreover, primary deviance is widespread and individuals engage in it for a vari ety of reasons. In c ontrast, secondary deviance occurs when a person begins to employ his [or her] deviant behavior or a role based upon it as a means of defense, attack, or adjustment to the overt and covert problems created by the consequent societal reaction to him [or her] (p. 76). Primary deviance progresses into secondary deviance once a deviant label has been applied and that la bel becomes part of the individuals identity. Labeling theory is a theory of secondary, as opposed to primary, deviance. This transition from primary to secondary de viance does not occur instantaneously; rather, Lemert (1951) proposes an appr oximate eight-step sequence of interaction process which includes the following steps: (1) primary deviation, (2) soci al penalties, (3) further prim ary deviation, (4) stronger penalties and rejection, (5) fu rther deviation, perhaps with hostilities and resentment beginning to focus upon those doing the penalizing, (6) crisis reached in tolerance quotient, expressed in formal action by the community stigmatizing the deviant, (7) strengthening of the deviant conduct as a react ion to the stigmatization and penalties, (8) ultimate acceptance of deviant and social status and efforts at adjustment on the basis of the associated role (p. 76). As suggested by Lemert, the transition to secondary deviance usually does not simply result from societal reaction to a single prim ary deviation, albeit this is a po ssibility especially for deviant behavior that is highly disdaine d. Typically, labels are ascribed following failed attempts to curb undesirable behavior. In order for deviance amplification to be a likely outcome of societal reaction to behavior, stigmatization, in wh ich community members make no attempt to 16


reintegrate the deviant individual back into society, is necessary (see Braithwaite, 1989). Then, applying a stigmatizing label successfully identifies the individual as a problem but does little to end the deviant behavior. Moreover, reacting to primary deviance with social penalties and stigmatizing labels not only fails to stop the undesirable behavior but can amplify it as the individual transitions from prim ary to secondary deviance. While the concept of secondary deviance is crucial to understa nding the effects labels have on individuals self-concepts and behavior, it does only so mu ch to clarify what exactly constitutes deviant behavior. There are two main definitional approaches used in the study of deviance: the normative and the r eactionist (Liska, 1987). The form er defines deviant behavior as behavior that violat es norms; that is, deviance is rule-b reaking behavior. According to the normative definition, individuals can under-conform or over-confor m to a particular normative expectation. Alternatively, the re actionist approach defines devian ce as the result of a certain reaction to actual, or illusory, be havior. While the two approaches can be utilized together to gain a better understanding of the different type s of deviance (see Hecker t & Heckert, 2002), the reactionist definition is the defi nition employed by labeling theorists. In one of the most recognizable statements of the labeling perspectiv e, Howard Becker contends that deviance is not a quality of the act the pers on commits, but rather a consequence of the application by others of rules and sanctions to an offender. The devian t is one to whom that label has successfully been applied; deviant behavior is behavior that people so label (Becker, 1963, p. 9). It is quite evident from this definition that the objectivity of the act become s, at best, secondary to the societal reaction. Reactionist de finitions of deviant behavior re directed academic attention away from the causes of deviance to the sources of de viance and the effects of deviant labels (Best, 2004, p. 25). 17


Deviance Amplification Hypothesis An early attempt to summarize the assumptions of labeling theory was conducted while the theory was, academically, in full stride during the early 1970s. While not all labeling theorists would necessarily agree w ith all of the assumptions, Clarence Schrag (1971) lays out nine that are believed to be central to the la beling perspective, which include: (1) no act is intrinsically criminal (2) criminal definitions are en forced in the interest of the powerful, (3) a person does not become a criminal by violation of the law but only by the designation of criminality by aut horities, (4) due to the fact that everyone conforms and deviates, people should not become dichotomized into criminal and non-criminal categories, (5) the act of getting caught begi ns the labeling proce ss, (6) getting caught and the decision-making in the criminal jus tice system are a function of offender as opposed to offense characteristics, (7) age, socioeconomic class, and race are the major offender characteristics that establish pattern s of differential criminal justice decisionmaking, (8) the criminal justice system is esta blished on a freewill perspective that allows for the condemnation and rejection of the identi fied offender, (9) labeling is a process that produces, eventually, identification with a de viant image and subculture, and a resulting rejection of the rejectors (Schrag 1 971, pp. 89-91; see also Wellford, 1975, p. 333). Wellford (1975) notes that the fi rst, sixth, and ninth assumptions represent the basis for the labeling perspective. The first statement is taken from the social reactionist definition of deviance (see Becker, 1963). The sixth statem ent is derived from the conflict tradition and contains the crux of the status characteristics hypothesis. The ninth as sumption is taken from symbolic interactionism and re presents the basic tenets of the deviance amplification hypothesis. While all of the assumptions of labeling theory have been laid out for completeness, only the deviance amplification hypot hesis will be explained in detail as it is pertinent to the current research. The deviance amplification hypothesis specif ies the process and conditions under which a labeling event is most likely to lead to s econdary deviance. The most common pathway to secondary deviance involves a label that is made public, which triggers the following chronological events: exclusion from conventiona l activities, alteration of self-identity, the 18


acquisition of supportive deviant others and, finally, second ary deviance (see Paternoster & Iovanni, 1989). In contrast, if a label is kept private, it is unlikely to trigger the events leading to secondary deviance. More specifically, a pr ivate label will usually result in the following chronological events: inclusion in conventional activities, maintenance of identity, maintenance of supportive conventional others and, finally, lack of sec ondary deviance (see Paternoster & Iovanni, 1989). What is clear from the literature on the effects of labeling individuals is that the pathways are not always followed in all cases (Paternoster & Iovanni, 198 9). An individual can dodge or, ultimately, accept a deviant label in inst ances that the two pathways do not suggest. If a label is made public, there are two wa ys through which an individual can escape involvement in secondary deviance. The first focuses on the actions of others whereas the second centers on the actions of the labelee. If social reactions to a public label are inclusive, this can help reintegrate the individual back into society (see Braith waite, 1989). Inclusive responses help to ensure the labelee is not barred from conventiona l opportunities and normal routines, which can ultimately prevent secondary deviance (see Pa ternoster & Iova nni, 1989). If an individual is excluded from conventional activities, that person ma y reject the deviant label in a process known as deviance disavowal (Davis 1961). This results in the maintenance of ones previous identity and an escape from s econdary deviance. While a public label does not always lead to secondary deviance, a private one does not always ensu re an individual will escape secondary deviance either In a process known as dev iance avowal, an individual identifies with a label despite th e fact that it is kept private (Turner, 1972). While a private label does not result in harsh social reactions, the indi vidual nevertheless alters his or her identity through a process of self-labeling (e.g., see Th oits, 1985). The subjecti ve, hidden meaning of the label becomes objective reality and the deviant heads down a path toward secondary 19


deviance. The deviance amplification hypothesis for individuals can be summed up as follows: a labeling occurrence may result in a blockage to conventional opportunities, an alteration of selfidentity, and the acquisition of supportive deviant others, which may ultimately result in deviance amplification (see Pa ternoster & Iovanni, 1989). Criticisms of the Labeling Perspective Labeling theory gained considerable popularity during the 1960s and continued into the 1970s before coming under heavy attack. One major criticism lies in the f act that the ascription of deviant labels does not create deviant behavior to begin with (Akers, 1968). While there is some evidence to suggest that an applied labe l precedes initial acts of deviance in time (see Matsueda, 1992; Triplett & Jarjour a, 1994), this critique remain s an important issue to be acknowledged in labeling research. Interestingly, any particular primary or secondary deviant act can appear behaviorally iden tical. Labeling theory explains secondary deviance but does not address why individuals engage in primary deviance. If deviant behavior is behavior that is labeled (Becker, 1963), explaining primary devian ce for a particular indi vidual is relatively unimportant because that individuals behavior has not yet evoked a societal reaction. Davis (1972, p. 461) notes that labeling theorists who cl aim that reactive processes of society provide the causal factor in deviance are thereby providing only a partial view of a highly complex problem. Employing a reactionist concept of devi ance completely ignores the objectivity of the act. Hence, it is seen as troublesome that an individual who has enga ged in rule-breaking behavior in which no social reactio n occurs is deemed not to be a deviant; but, an individual who is falsely accused of engaging in rule-breaking behavior in whic h a social reaction occurs is deemed to be a deviant (Gibbs, 1966). While one of the strong points of the labeling perspective is that it acknowledges societys role in de viance production, a critical weakness is that it 20


consequently ignores the indivi duals objective deviant behavior, which results in an individual escaping culpability for rule-breaking behavior. Gibbs (1966) argues that labeling theorists h ave never specified exactly what kind of reaction identifies deviant acts. There is a vari ety of potential respon ses to behavior and it remains unclear exactly which reac tion constitutes the ascription of a deviant label. In essence, the concepts used in labeling are confusing a nd ambiguous, resulting in a theory that lacks important scientific value (Gibbs 1972). In a response to these critiques, Scheff (1974) notes both that other sociological theories lack denotativ e concepts and that connotative concepts have value because they challenge researchers everyday assumptions. Still, th e definitional criticisms levied against the labeling perspective continue to linger. Criticisms of labeling did not ju st come from scholars prey ing on definitional weaknesses; attacks also came from scholars in several different theoretical camps, one of which is commonly believed to be in tune with the labeling perspectiv e. Conflict theorists attacked labeling theorists for selecting easy targets (Best, 2004, p. 35). As stated previously, labeling theorists have mainly studied nuts, sluts, and preverts, wh ile failing to examine deviance among the powerful and ruling classes (Liazos, 1972). Focusing on cert ain types of behaviors inadvertently supports and strengthens the status quo (Best, 2004, p. 36). While labeling theorist s utilize notions of power and conflict, conflict theorists criticized the labeling perspective and its scholars for failing to completely deal with power re lations in the study of deviance (p. 36). Attacks did not only come from conflict theorists; ot her camps joined in the criticizing as well. Feminists draw attention to the establis hed power relations in patriarchal societies and examine the impacts on men and, in particular, women. While focusing on the application of labels to women and their subsequent effects me shed well will the focuses of labeling scholars 21


(see Schur, 1983), the interest in the topics of victimization a nd womens social movements did not (Best, 2004, p.39). Additionally, groups of individuals sharing deviant la bels felt that their struggles should be seen in light of identity politics as opposed to deviance per se (Best, 2004). Thus, labeling theory was attacked by scholar s from other theoretical camps because the perspective failed to adequately address salient and important issues in critical, feminist, and identity political thought. While the labeling perspective was highly popular in the 1960s, the theoretical and practical criticisms of the labelin g perspective led, in part to its downfall by the mid-1980s. Organizational Theory Anthropomorphizing Organizations Anthropomorphizing organizations, or the ex tension of human qua lities and socialpsychological concepts to organizational actors, is not something that all scholars accept without reservation, or even at all. While organiza tional theorists often em ploy social-psychological concepts to explain organi zational behavior, many schol ars find anthropomorphizing organizations is inappropriate and unproductive. The debate as to whether organizations and corporations assume human qualities is particular ly strong in the field of criminology. In one view, there can be no social psychology of so-c alled corporate or organizational crime because corporations have no biological or psychologi cal characteristics (Cressey, 1995, p. 418). Moreover, organizations and corporations cannot intend actions, t hus, behavioral theories of crime cannot explain organizational crimina lity (Cressey, 1995). Then, there are some theoretical and practical reasons to oppose anthropomorphizing organizations. Despite this, there are many academics in criminology that study crimes of organizations and corporations and employ various theoretical perspectives to explain them. 22


One of the interesting things about organi zations is that the people who make up the organizations are quite irrelevan t; organizations tend to operate in certain ways despite which individuals are in what positions (Coleman, 200 2). While organizations are comprised of individuals, characteristics of organizations cannot simply be explained by the characteristics of the individual members. In responding to Cresseys critiques ag ainst anthropomorphizing organizations, it is argued that corporations have the ability to modify their goals and policies (Braithwaite & Fisse, 1995, p. 442). The idea that organizations can alter qualities is central to the notion of organizational deviance amplification. Braithwaite and Fisse (1995) note that biological theories and correlates of crime such as race and impul siveness have little, if any, applicability to organizations; however, many theoretical persp ectives (e.g., rational choice, social learning) can be employed to unders tand the behavior of both individuals and organizations. Because organizations have an identity that can be influenced by social interaction (a point that will be discussed in detail later) just as individuals do, deviance amplification may indeed occur for organizations. It is impossible, in a dvance of a theory being developed and put to the test, to rule out any level of genera lity in theory application (Braithwaite & Fisse, 1995, p. 446). Extending labeling concepts to the organizational actor may help model the behavior of labeled organi zations. While some scholars may contest anthropomorphizing organizations, the use of established organizational theory justifies the extension of labeling concep ts to organizational actors.2 2 It is acknowledged that organizations are in some ways different from individuals. Thus, the theoretical model of organizational deviance amplification formulated here will be similar but not entirely identical to the traditional deviance amplification model for individuals. My study anthropomorphizes organizations but adjusts labeling theory to allow for differences in the deviance amplification process for organizations. 23


Identity-Image Interdependence Similar to the common usage of the term se lf-identity (see Bailey, 2003), organizational identity has typically been concei ved of as being a stable construc t. Organizational identity is generally defined as that which is central, enduring, and distinctive about an organizations character (Albert & Whetten, 1985). With this definition of or ganizational identity, extending labeling theory to an organizational actor may seem somewhat unwarranted. That is, if an organizations identity is constant, there w ould be no good reason to expect a change in organizational identity following the ascription of a de viant label and, thus, no theoretical reason to expect organizational devian ce amplification. However, it has been argued that, because organizations change and are affected by their environments, organizational identity is better conceived of as a fluid construct as opposed to a rigid one (Gioia et al ., 2000). An alternative conception of organizational iden tity defines it as p recarious and unstablefrequently up for redefinition and revision by organi zational members (Gioia et al., 2000). It should be noted that the changeability of organizational identity is as crucial to the idea of organizational deviance amplification as changeability of self-iden tity is to the traditional notion of deviance amplification among individuals. The definitions of organizational image have ta ken a variety of forms. Most relevant to this research is the notion of th e construed external image of an organization, which is defined as organizational members percepti ons of what outsiders perceive of the organization (Dutton & Dukerich, 1991). Similar to the symbolic interactionist notion of the looking-glass self (see Cooley, 1902), the construed external image of an organization can affect organizational identity (Gioia et al., 2000). Gioia and his colleagues (2000) create a pr ocess model for identity-image interdependence. They contend that the process begi ns when an external event triggers an identity-image comparison. If the organizati on perceives incongruence between organizational 24


identity and construed external image and decide s to do something about it, the organization can act one of two ways in order to restore harm ony. The organization can attempt to change how outsiders perceive the organizati on. Or, the organization can alter its identity so that it matches the organizational image (see Gioia et al., 2000, p. 69). The latter option is the pathway to organizational deviance amplification. Before th e theoretical model for deviance amplification can be specified it is important to know the an swer to the following que stion: if incongruence between an organization's construed external image and identity occurs, what makes an organization change its identity as opposed its image? Reputation Perhaps due to the static concept of organizational identity (see Albert & Whetten, 1985), it is believed that organizations will act to cha nge their image instead of their identity. This reason does not suffice with the notion of a fluid c onstruct of organizational identity (see Gioia et al, 2000). Organizations labeled deviant work to change their organizational image in an attempt to maintain or restore a positive organizational reputation. Organi zational reputation is defined as outsiders stable, long-term judgments of th e actions and achievements of an organization (Fombrun, 1996). Failing to address a negative construed external image can lead to the development, or worsening, of a negative organizational reputation. Earning and maintaining a positive reputati on becomes important when one considers notions of competition and social interaction. One of the ways th at organizations increase their competitiveness is by maintaining a better reputation than the oppositio n (Davies et al., 2002). It follows that organizations that are in compe tition will seek to acquire and maintain positive organizational reputations. Positive reputations are also important for organizations that have repeated social interactions. That is, when inte ractions between an organization and a particular entity are likely to occur again in the future, organizations con cern themselves with reputation 25


effects of organizational decisions and actio ns (Murnigham, 1985, Bettenhausen & Murnigham, 1985; see also Weigelt & Camerer, 1988). While having a positive reputation will likely make accomplishing organizational goals easier, it is not essential for certain organizations to have a positive reputation. If an organization does not have competition, then it need not have or maintain a positive reputation. A lack of competition can occur if an organization or co rporation possesses a monopoly in an otherwise competitive market (e.g., Microsoft) or if a single organization is designated for a particular purpose (e.g., city police agency). When an or ganization is without competition, there is no alternative to get the product or service elsewhere. If the product or service is essential, developing and maintaining a positive reputation becomes a matter of preference as opposed to a necessity. If organizational reputation is nonesse ntial for organizational success and survival, the organization has the option to alter its identity as opposed to altering its image. It has been argued that some organizations may alter their identity when faced with an identity-image comparison. Char acteristics about both th e organizations reputation state prior to the stigmatizing event and the intensity and pervasiveness of the stigmatizing label are important in determining whether an organization will alter its identity when faced with an identity-image comparison in times of crisis. Fiol & Kovoor-M isra (1997, p. 150) contend that organizations with negative reputations will incorporate nega tive attributes more readily and those with positive reputations will resist assimilation of stigma. For the former, the stigmatizing label sets off an identity-image comparison and the organi zation lacks a positive pre-label reputation to offset the effects of the stigmatizing label (i.e., halo effect3). Then, organizations are more likely 3 Having a positive pre-crisis reputation can prevent damage in crisis situations by leading stakeholders to give an organization the benefit of the doubt; alternatively, the po sitive pre-crisis reputation ca n operate as a shield to minimize adverse reputation effects (see Coombs & Holliday, 2006). 26


to alter their identity if the stigmatizing event does not clash with the current state of their organizational reputation. The intensity (i.e., the strength of the disa pproval) and the pervasiveness (i.e., how many disapprove) of the stigma also play an important part in whether orga nizational identity is influenced by a stigmatizing event (Fiol & K ovoor-Misra, 1997). Even organizations with positive pre-crisis reputations can be influenced by intense and pervasive stigmatizing labels. (Fiol & Kovoor-Misra, 1997). The more intens e and pervasive the stigmatizing label is, the more likely it is to lead to a change in organi zational identity. This occurs because intense and pervasive stigmatizing labels work to highlight the disconne ct between the organizational identity and construed external image and, as a result, make changing the construed organizational image a less viable option. Organizational Deviance Amplification It is important to note that there is a continge nt nature of the labeli ng process in that the chances for secondary deviance are greatest if the following all occur: exclusion from conventional opportunities, alteration of self-ide ntity, and the acquisition of supportive deviant others (Paternoster & Iovanni, 1989) An organization can change its identity in order to reduce incongruence between organizational identity and c onstrued external image. So, the alteration of organizational identity is crucial to organiza tional deviance amplification. In developing a theory of organizational deviance amplification we will now explore if the other factors leading to secondary deviance that are traditionally important in labe ling theory are necessary for organizational deviance amplification. According to the process model for identity -image interdependence, a triggering event results in an identity-image comparison that can directly lead to a ch ange in organizational identity (Gioia et al., 2000). An individuals al teration of self-identity is normally contingent 27


upon being excluded from conven tional opportunities; how ever, organizations can alter their identity without necessarily being excluded from conventional opportunities. This is not to say that exclusion from conventiona l opportunities does not have an impact whatsoever. Instead, this impact may be exacerbating as opposed to necessary for organizational deviance amplification. The numerous social interactions between indi viduals and society may be one reason it is relatively easy to imagine how deviant labels can block access to conve ntional opportunities. For example, deviant labels can adversely affect such things as educational attainment and employment (e.g., see Bernburg & Krohn, 2003). Inte ractions between organizations and society make it somewhat harder to envision blocke d access to conventional opportunities. Still, excluding deviant organizations from conventional opportunities can indeed occur. For example, a deviant high-school organization may be denied classroom space for meetings. Importantly, if social reactions to organizational deviant beha vior become more than a simple blocking of access to conventional opportunities, they may have a much different effect. More specifically, if social reaction directly th reatens the existence of an or ganization, deviance reduction may occur. Labeling theory suggests that once an individual takes on a deviant iden tity, he or she will seek out supportive deviant others. However, on ce an organization assumes a deviant identity it may be harder for organizations to acquire suppo rtive deviant organizational others simply due to the fact that they may be few and far be tweenwith the rationale being that nearly all organizations work diligently to escape a deviant label (Ermann & Lundman, 2002, p. 31). Still, there are deviant acts th at require supportive deviant others and certain industries and organization types are criminogenic. For exampl e, heavy electrical equipment industries were 28


involved in a massive pricefixing scandal (see Geis, 2002), in which multiple deviant organizational actors were required. Then, accepting a deviant identity may indeed lead an organization to find supportive devi ant organizational others for certain deviant behaviors. However, many deviant acts can be done without the suppo rt of deviant organizational others. For example, issuing a large number of traffic ci tations does not necessarily require the support of other police organizations. Th e extent to which supportive de viant organizational others are necessary depends on the type of deviant behavior itself. For those organizations that can deviate without the support of others, such support is theori zed to become an exacerbating factor. Through merging theoretical concepts from la beling theory and organizational theory, we can now lay out the theoretical m odel of organizational deviance amplification (see Figure 2-1). The model begins with a labeling event, which triggers an identity-image comparison. If the organization perceives that the organizational identity and construed external image are incongruent, the organization can seek to restor e harmony between the identity and image. If reputation is essential to organizational succe ss, the organization will work to change the construed external image and secondary deviance will not occur. However, an organization that is labeled and does not need to restore or ma intain a positive organizational reputation for organizational success is a candidate for organi zational deviance amplification. Organizations with negative pre-label reputations and those th at experience intense and pervasive stigmatizing labels are most likely to have their identity influenced by a label (Fiol & K ovoor-Misra, 1997). It follows that when a candidate for deviance am plification receives an intense and pervasive label and has a negative pre-label reputation, th e acceptance of a deviant identity and deviance amplification are very li kely to result. 29


Organizations with positive pr e-label reputations are still vulnerable to the effects of intense and pervasive stigmatizi ng labels (see Fiol & Kovoor-Mis ra, 1997) but the alteration of organizational identity becomes only somewhat likely to occur in this case. If the label lacks intensity and pervasiven ess but the organizational has a nega tive pre-label re putation, alteration of organizational identity is also only somewhat likely to occur. If the label lacks intensity and pervasiveness and the organization has a posi tive pre-label reputation, the organization will usually attempt to alter is construed external image and the alteration of organizational identity becomes very unlikely. The concepts of dev iance avowal (Turner, 1972) and deviance disavowal (Davis, 1961) have their organizatio nal counterparts. Organizational deviance avowal occurs if the organization alters its identity despite the fact that the organization has a positive pre-label reputation and is not ascribed an intense and pervasive stigmatizing label. Conversely, organizational deviance disavowal occu rs if the organization refuses to alter its identity in the face of having a negative pre-label reputation and being ascribed an intense and pervasive stigmatizing label. While secondary deviance is generally cont ingent upon the exclusion from conventional opportunities and the acquisition of supportive deviant others for individuals, this is argued to usually not be the case for or ganizations. Instead, these f actors increase the likelihood of deviance amplification but the path way to secondary deviance (for behaviors that do not require multiple parties by definition) requires only the alteration of organizational identity. This represents a major theoretical departure from tr aditional labeling theory. Still, the essence of deviance amplification remains; a deviant label can alter organizat ional identity, which can cause an amplification of organizational deviant behavior. 30


Labeling event Identity-image comparison Negative construed Positive organizational external image identity Figure 2-1. Theoretical model of or ganizational deviance amplification Yes Reputation essential for organizational success? Alteration of organizational identity Exclusion from conventional opportunities Supportive deviant organizational others No Negative pre-label reputation? Stigmatizing label intense & pervasive? No Secondary Deviance Secondary deviance Yes Yes No Stigmatizing label intense & pervasive? No Efforts to alter organizational image Deviance avowal Alteration of identity very likely Discrepancy perceived? No action taken Yes Alteration of identity somewhat likely Deviance disavowal No Yes Alteration of identity very unlikely No 31


CHAPTER 3 REVIEW OF THE LITERATURE Labeling Evidence Gove (1980a, p. 14) argues that evaluati ng the empirical evidence of the labeling perspective should not simply l ook for evidence in support of the theory but should compare the theorys explanatory power relative to other expl anations of deviant behavior. Some scholars have contended that this type of analysis is inappropriate for labeling theory (see Schur, 1971, p. 155). On the one hand, the utility of the theory is lo st if its predictability is not testable. On the other hand, the theory may be inappropriately form ulated when it becomes subject to mainstream tests of empirical verification. This sentiment is most clearly seen in the words of Lemert when he states that labeling theorist s must deny or refute that which they havent said or depend on ideas they dont necessarily share, or accept only with qualification (Lemert, 1976, p. 244). While it is important to assess the empirical evidence supporting key postulations of labeling theory, it should be acknowledged that labeling theorists themselves are not necessarily driving the empirical tests, which can result in theory misspecification. Mental Patients One of the more popular applications of labelin g theory has been to the behavior of the mentally ill. Interestingly, it is difficult to di stinguish between true patients and pseudo-patients within mental hospital settings (Rosenhan, 1973). Once a label is given, behavior is interpreted as being consistent with mental illness. Labeling theory suggests that individuals who are tagged as mentally ill may ultimately fulfill that role, despite their initial ment al state. Gove (1980b) finds that many individuals in the mental health system are seriously ill aside from any effects of secondary deviance. Moreover, the mental heal th system appears to have a comprehensive screening process that occurs befo re the label is ascribed, sugges ting that labeling theorists have 32


overstated the role labels play in mental illness (Gove, 1980b). Acknowledging that mentally ill individuals may have serious psyc hological conditions prior to the ascription of deviant labels, Link and his colleagues (1990) develop a modified labeling approach for the mentally ill. They find that even if mental disorders are not created by the ascription of deviant labels per se labels applied to the mentally ill can a dversely impact the pati ents social networks (Link et al., 1990). Link and Cullen (1990) reviewed the empirical evidence amassed throughout the 1980s and found significant evidence to support the devian ce amplification hypothesis during this time period. Another review of the literature, utilizing 35 studies examining hospita lized and ex-patient attitudes toward mental illness labels and stig ma associated with hospitalization, found mixed support for labeling theory (Weinstein, 1983). Patie nts attitudes toward generalized labels of mental illness were unfavorable but patient s attitudes toward hospitalization and posthospitalization experiences, in many instances, were favorable (Weinstein, 1983). These mixed findings may, in part, be due to the fact that mental patients experience both treatment and stigma. Mental patients quality of life generally increases when treatment services are rendered but the perception of stigma associated with treatment works to decreas e their quality of life (Rosenfield, 1997). The experiences and behavior of mentally ill persons appear to be caused by both real psychiatric problems and labeling effect s. Thus, the literatur e partially supports the deviance amplification hypothesis fo r mental patients. It is argued that labeling theory should still be considered an explanation for the beha vior of mental patients because the empirical evidence for labeling is mixed and the evidence for the psychiatric model remains inconclusive (Scheff, 1999, p. 15). 33


Juvenile Delinquents There are a number of studies that examine th e effects of labels on juveniles behaviors that show support for the deviance amplifi cation hypothesis (Pfuhl, 1980, p. 234). One reason juveniles are frequently studied is that the identities of adolescen ts are in a state of formation, making labeling effects particular ly relevant. There is evidence that juveniles become more involved in delinquency after being labeled despite the fact they are not initially committed to deviance (Baker, 1973). Gold and Williams (1969) use matched pairs of juveniles and find that apprehension leads to a greater likelihood of further delinquency. Deviance amplification effects have been found with self-reported measures of de linquency in addition to official measures of deviant behavior. Farrington (1977) finds that a group of publicly labeled individuals self-report higher involvement in delinquent beha vior than controls and this effe ct is statistically significant. There is evidence to suggest that labeling theory should consider sex differe nces of labelees. In a sample of male and female adolescents, Ray and Downs (1986) found that the ascription of deviant labels did not precede drug use behavi or for females but did for males. Much of the evidence supporti ng the deviance amplification hypothesis has been called into question. Researchers det ecting links between recidivism and labels often make sweeping generalizations, which cannot be supported by th e data (Meade, 1974). Early empirical tests may have contributed to the demise of labeling theory because they suggested support for the deviance amplification hypothesi s that was unwarranted due to the incompleteness of the labeling models tested (Paternoster & Iovanni, 1989, p. 384). Some of the empirical evidence shows no labeling effects at all. For example, Thomas and Bi shop (1984) note that while there are two theoretical implications of legal sanctions evidence from junior and high school students in Virginia suggests that ne ither are well supported. They conclude by noting that core assumptions of labeling theory, as well as deterrence theory, do not stand up to multivariate 34


statistical designs and long itudinal research (Thomas & Bishop, 1984, p. 1244). Smith and Paternoster (1990) find that ev idence supporting the deviance amplification hypothesis can be accounted for by a selection artifact. Overall, the evidence in support of formal labeling effects for juveniles is mixed. Only more recently has the topic of informal labeling effects been pushed to the forefront of deviance research. The labeli ng effects of informal agents of social control may be as, or more, important than the effects of labels ascribed by formal agents of soci al control. Parental appraisals of their youth directly impact the reflected appraisals of the self (Matsueda, 1992). Hagan and Palloni (1990) find intergenerationa l effects for parent and son labeling, which indicate that deviant labels and their adverse eff ects may be reproduced in families. However, in examining the drug use behavior of adolescent s, Downs and his colleagues (1997) found that parental labels neither had an effect on drug us e nor the individuals se lf-label. Moreover, a particular parental label is likely the result of the way in which an individual presents his or herself to parental figures (Downs et al., 1997). It appears then th at parental labels can influence self-appraisals but that parental labels ar e also affected by self-appraisals. Matsueda (1992) found that reflecte d appraisals of the self that portray the self as deviant can affect future delinquent behavior. Consiste nt with this last statement, Downs and his colleagues (1997) found that self-lab els can indeed affect drug us e behavior among adolescents. However, self-labels also are influenced by drug use behavior (Downs et al., 1997). Hence, the effects of labels and drug use behavior are found to be reciprocal. Taken together, these findings are partially supportive of labeling theorys deviance amplifica tion hypothesis but demonstrate that future research is necessary in uncovering the effects of informal and formal labels on behavior. 35


Police, Citizens, and Traffic Stops Conflicts Regarding Speeding Unlike the enforcement of criminal laws, tra ffic law enforcement places serious strain on the relationship between police a nd middle-class citizens (Vedder & Keller, 1965). Citizens feel frustrated after receiving a traffic citation and feel that the police should redirect their law enforcement efforts to focus on serious crime (Radelet & Carter, 1994). The experience of receiving a traffic citation can adversely affect c itizen attitudes toward police. Correia and his colleagues (1996) find that individuals that r eceive traffic citations evaluate police more negatively than those that do not. This effect may be particularly strong for individuals with low incomes due to the monetary penalty (Smith & Hawkins, 1973). Some evidence suggests that the traffic encounter damages police-citizen re lations but it is importa nt to know why speed enforcement, in particular, is damaging to poli ce-citizen relations. Unde rstanding the motives of citizens and police regarding speed and speed enfo rcement, respectively, is a good place to begin to explore this issue further. Reducing travel time is an important considera tion for motorists. For example, in heavily congested metropolitan areas, motorist s are willing to pay tolls to ut ilize open roads that enable a faster commute (see Brownstown et al., 2002). When traffic is not congested, speeding provides motorists with a simple and straightforward way to reduce travel time. While saving travel time is a major reason motorists exceed posted speed limits, it is not the only one. To a lesser extent, some drivers engage in speeding behavior for excitement (Machin & Sankey, 2006). Speed enforcement, then, is at odds with the trave ling needs and desires of a substantial number of motorists. In a study of police officers motives for tr affic enforcement, Southgate and MirrleesBlack (1991) found the most important goals of en forcement to include: road safety, enforcement 36


of traffic laws, educating drivers, and maintain ing traffic flow. While police may be enforcing the law for safety and traffic management reasons rural police may be responding to fears of the citizens they serve. Residents in rural areas tend to note that speeding is a serious concern (Ball, 2001). Speeding is a concern in urban and subur ban areas too, but the employment of traffic calming strategies helps to reduc e the speeding problem in thes e areas (Walter, 1995). Rural police departments are seen as the only sour ce for protecting small towns from outsiders careless driving. Police are motivated to enfor ce the speed limit for a variety of reasons and strict enforcement can place police-citizen relations into a serious predicament. While there is generally a potential for policecitizen conflict due to contradictory goals, most police agencies are not formally labeled deviant since citizens acknowledge that some traffic enforcement is necessary for safety a nd traffic management reasons. Motorists possess labeling power because drivers constitute a larg e group with politically powerful members and, more importantly, have automotive organizations directly representing these interests. If a particular police agency engage s in excessive speed enforcemen t and generates revenue from doing so, there is the potential for that agency to be labeled deviant. Generating revenue from aggressive speed enforcement alone is typically not enough to be labeled deviant. It should be noted that revenue generation is a byproduct of aggressive speed enforcement and strictly enforcing the speed limit in some locations may be warranted. For example, dangerous curvy roads or high fatality highway segments ma y require aggressive enforcement to prevent accidents and loss of human life. Then, the pub lic will not seek to label aggressive speed enforcement as deviant if it is believed to be don e with genuine intentions of increasing driver and pedestrian safety. The AAA speed trap labe l connotes that the polic e agency intentionally engages in strict enforcement of the speed lim it specifically to generate revenue and without a 37


particular need to address a safety concern. Pu t another way, the deviant label is applied because the police agency is seen to enga ge in speed enforcement not for safety reasons, but instead as a means of increasing police operational and town budgets. Thus, while speed enforcement can have ill-effects on police-citizen relations per se, adding the perception of a revenue generation motive escalates an inherent conflict of interest to a level in which deviant labeling becomes likely. Traffic Stops and Outcomes Traffic stops and speed enforcement can be pr oblematic for police-citizen relationships in general, but profiling can make relations betwee n police and certain groups particularly bad. Profiling is the use of personal characteristics an d contextual cues by a police officer to make a judgment about whether an individual is belie ved to be engaging in criminal activity (Fredrickson & Silijander, 2002). Profiling is a normal part of policing and is often confused with the term racial profili ng (Becker, 2004, p. 183). Racial profiling is a special case of profiling in which officers sing le out individuals to stop and question based on race alone. Several studies have explored the question of whethe r police stop individuals because they are driving while Black or minority. Engel and Ca lnon (2004a) find that Blacks and Hispanics are no more likely to have contraband in their vehicle but are subjected to disproportionate rates of arrests and searches. Similarly, a study c onducted in Maryland found that minorities were inappropriately profiled; the ra tes at which Blacks and Hisp anics were carrying contraband were actually lower than the rate for Whites (Becker, 2004). Not all evidence leads to the conclusion that minorities are more likely to be stopped due to racial profiling. The availability and use of benchmark comparison data may influence the findings of racial profiling studies (see Engle & Calnon, 2002b). For example, Lange and his colleagues (2005) used traffic violating behavior as the benchmark comparison and found th at there is disparity in speeding behavior by 38


race, which is consistent with the racial disparity detected in police stops. The evidence supports the conclusion that Blacks are more likely to speed and are consequently more likely to be stopped (Lange et al., 2005). Whether profiling is the source of the disparity in traffic stops, research nevertheless suggests that minority citizens are stopped more than majority citizens. The question more relevant to my study is whether certain groups are more likely to be issued a citation. Engel and Calnon (2004a) find that there is an increased chance that minorities will receive a traffic citation when all other factors are controlle d. Several studies find th is not to be the case. One study that used census data for benchmar k comparison found minority motorists to be less likely to receive a traffic citation or to be ar rested than their White counterparts (Smith & Petrocelli, 2001). Novak (2004) also found that minorities were significantly less likely to receive traffic citations than Whites. The combination of research findi ngs, a greater likelihood of being stopped and lower likelihood of r eceiving a citation for minorities, suggests that minorities may be stopped for reasons th at do not warrant citation issuance per se (Novak, 2004; Smith & Petrocelli, 2001). The decision to issue traffic citations is not solely dependent upon the individual officers discretion, nor is race/ethnicity th e only factor that is likely to lead to an increased chance of a police-citizen encounter. Organizational norms interact with officer discretion, which can impact which groups of citizens are more likely to receive traffic citations at certain times of the month; minorities and those of the lower class are much more likely to be issued traffic citations during periods of high quota saliency (Lundman 1971). Aside from organizational norm impacts and physical characteristics of the driver s, vehicle characteristics may be important in determining which vehicles are pulled over by police. Novak (2004) finds that residents are less 39


likely to be cited than nonresidents In a town or small city, one qui ck way to tell that a driver is likely to not be a resident is by examining the license plate. Out-of-state li cense plates have been used as a profiling tool for drug runners (see Buerge r & Farrell, 2002). They also can be used as an indicator of the likelihood of traffic citation contestati on, which may be an important consideration for police agencies issu ing traffic citations for profit. 40


CHAPTER 4 DATA AND METHODOLOGY Hypotheses A main goal of my study is to test whet her labeling an organization, which does not require a positive reputation for organizational success and survival, results in an increase in deviant behavior. The author of this study c ontends that a good candi date for organizational deviance amplification is a police agency labeled for deviancy in speed enforcement. First, the speed trap label likely serves as a trigger for an iden tity-image comparison. In deciding how to respond to identity-image incongruence, the organi zation is free to alter its identity because the police organization lacks competition. Additionally, the organization tends not to interact repeatedly with motorists that are issued citati ons, making organizational reputation even less of a concern. As will be discussed later in this ch apter, the organization has a negative pre-label reputation and the stigmatizing la bel is both intense and pervasiv e, which makes it very likely that secondary deviance will result from the labeli ng event. It is hypothesized that the ascription of the speed trap label should lead to an amp lification of the organiza tional deviant behavior. The organizational deviance amplification hypothes es tested are explicitly formulated here. Hypothesis One: The speed trap label will result in an overall increase in the number of speeding citations issued by the labeled police agency. This effect will not be observed for control agencies. Hypothesis Two: The speed trap label will result in citations being issued with a lower speed differential (i.e., posted speed subtracted from actual speed) by the labeled police agency. This effect will not be observed for control agencies. Hypothesis Three: The speed trap label may result in a change of the percentage of out-ofstate drivers that are issued speeding citations by the labeled police agency.4 This effect will not be observed for control agencies. 4 Hypothesis three is explorat ory because the label could have one of two op posite effects. First, the percentage of out-of-state drivers being cited could increase because the police agency may target individuals which may be less likely to contest traffic citations. Second, the percentage of out-of-state drivers being cited could decrease as the 41


Data Data used to test the orga nizational deviance am plification hypotheses comes from four different sources, which are all pub lic records in Florida. The f our types of public record data include: 1) traffic citation data, 2) traffic volume data, 3) town ch aracteristics data, and 4) police agency characteristics data. Traffic Citation Data Traffic citation data are the main source of da ta for my study. The data were obtained in electronic format by written request from the Alachua County Clerks Office located in Gainesville, FL. The data set contains inform ation on all traffic citations issued by police agencies in the county for the study period. Data used for descriptive pur poses and to test the organizational deviance amplifica tion hypotheses are explained here in detail. Five demographic variables are used and are coded as follows: N on-White is coded for White and for nonWhite; Female is coded for male and for female; Hispanic is coded for nonHispanic and for Hispanic; Out-of-State refe rs to whether the license plate on the vehicle is in-state or out-of-state and is coded for in-s tate and for out-of-st ate; and, Age indicates the age of the driver at the time of offense, wh ich was calculated by subtracting the date of birth from the date of offense. The data file was split in order to comp are the citation issuan ce behavior of the intervention and control police agencies. The variab le used to split the data file is Agency and is coded for Alachua, for High Springs, an d for Waldo. The time-series variables used to test the organizational devian ce amplification hypotheses are as fo llows: SpeedCitation is the total number of speeding citations issued in a gi ven month; SpeedDifferential is the average police agency may become less selective and issue citatio ns to any speeders (this assumes some selectively in citation issuance prior to the deviant label). 42


monthly speed above the legal limit. This valu e is calculated by subt racting the posted speed limit from the actual speed for each citation a nd taking the average for each month; Out-ofStateCitation is the monthly pe rcentage of speeding citations i ssued to drivers of out-of-state vehicles; lastly, Officers refers to the number of officers issuing citations each month and is used as a control time-series. The Florida Uniform Traffic Citation Statis tics (2005) were obtaine d directly from the Florida Department of Highway Safety and Motor Vehicles website.5 The data are utilized to gain an understanding of the de viant nature of the speed enfo rcement behavior. These data contain standardized traffic citation informati on for all police agencies in Florida for the following offense categories: 1) criminal, 2) noncriminal moving, and 3) non-moving violations. While my study focuses on the effects of the speed trap label on sp eeding citation issuance behavior, using the Florida Uniform Traffic Citation data allo ws for a look at the citation issuance behavior in a regional context. That is, these data will be used to look at the speeding citation issuance behavior of th e Waldo Police Department as it compares with other police agencies in Florida.6 Because the data collection was labor intensive, the entire state is not utilized for comparison. Instead, data are used on 115 police agencies in 31 contiguous counties, which include: Alachua, Baker, Bradford, Brevard, Citrus, Clay, Columbia, Dixie, Duval, Flagler, Gilchrist, Hamilton, Hernando, La fayette, Lake, Levy, Ma dison, Marion, Nassau, Orange, Osceola, Pasco, Pinellas, Putnam, Semi nole, St. Johns, Sumter, Suwannee, Taylor, Union, and Volusia counties. Data used for regi onal descriptive purposes are coded as follows: 5 http://www.hsmv.state.fl.us/reports/2005UTCStats/UTCStats.html 6 One point of caution in using the Florida Uniform Traffic Citations Statistics should be noted. The data are for the year 2003 and are a measure of speeding citation issu ance behavior that partially captures any deviance amplification effects for the labeled po lice agency. However, the initial base rate before the deviant label was still high prior to the ascription of the deviant label (see Figure 5-3 D). 43


Town/City represents the name of the town, cit y, or municipality. SpeedPop refers to the average number of speeding citations issued pe r day per 1,000 residents in 2003; TotalPop is the average number of traffic citations issued per day per 1,000 reside nts in 2003; SpeedVol represents the average number of speeding citations issued per day per 10,000 vehicles in 2003; finally, TotalVol is the average number of tr affic citations issued per day per 10,000 in 2003. The last two variables, in part, ne cessitate a detailed discussion on tra ffic volume calculation. Traffic Volume Data To supplement the traffic citation data, the 2005 Florida Traffic Information (FTI) Disk published by the Florida Department of Transpor tation was obtained by written request. The traffic volume data include annual average daily traffic (AADT) counts at specified locations throughout towns and cities in Florida. Most town s and cities in the state have count data for one or more sites within the boundaries of the munici pality. AADT count data refer to the average number of cars passing a count site per day in a particular year. Data in time-series format are available for Waldo, High Springs and Alachua whereas 2003 AADT count data are available for 96 of 115 towns. For the 2003 AADT count data, the measure employed to compare traffic volume across cities is the average amount of daily traffic flowing throughout the city limits each day. TrafficVol is calculated by summing up the 20 03 AADT count data from all sites within the boundaries of a city/town and dividing this summa tion by the total number of sites. Because interstate highways in Florida ar e in the jurisdiction of the Florida Highway Patrol, traffic count sites on interstates that pass thr ough municipalities were not included in the an alysis. It should be noted that the average amount of daily traffic calculated above is a function of the location of the counter sites. That is, more counters in hi gh volume areas would tend to increase the traffic volume measure used in this analysis, whereas more counters in low volume areas would 44


decrease this measure. However, TrafficVol is a decent measure to determine the average traffic volume flowing throughout a particular town or city and permits comparison across towns and cities.7 The fact that counter site locations ca n affect the traffic volume measure should be recognized as a limitation of the data. The 2005 FTI disk has a historic feature that permits resear chers to access traffic counts from previous years. While only a single AADT count is given for each location, Seasonality Factors (SF) are included that pe rmit the calculation of traffic volumes for given months. The use of SF is necessary in order to examine tra ffic citation behavior of the three towns from August 2001 through August 2005. There were five count sites that reported SF in Alachua County in 2005. Because there was not a counting site near or within each town, the average of the five available SF was taken and used for all three towns. The monthly average daily traffic (MADT) counts were obtained using the followi ng formula: MADT = AADT / SF. While AADT data are available for 2001 through 2005 the SF was only available for 2005. Thus, the 2005 SF was used to calculate the MADT in all case s. This is not a sign ificant problem because seasonality effects8 are likely to be similar from year to year. The MADT count data are used as a control time-series when testing the organiza tional deviance amplifi cation hypotheses and it is referred to as Traffic. Town Characteristics Data Data on the characteristics of the towns in the study are obtained from United States Census Data (2000), which is accessed online.9 Census data are primarily utilized to examine 7 The speeding citation and total citation issuance rates measur ed with the highest traffic count data site yield similar results. Thus, the measure of traffic volume employed in my study is satisfactory. 8 Traffic volume can be influenced seasonally by a number of factors (e.g., number of vacationers, academic school year, and wet/dry seasons). 9 http://www.census.gov/ 45


the comparability of the towns i nvolved in the study. Census data permit the researcher to see how similar the intervention town and control to wns are across various demographic variables. Variables compared in my study include: population, median income, average family size, median age, percent female, percent non-White, percent Hispanic, percent high school graduate, percent veterans, percent in labor force, and percent below poverty level. Police Agency Characteristics Data Data on police agency characteristics are us ed to determine the comparability of the intervention police agency to the two control agencies. The source of these data is the Criminal Justice Agency Profile Reports10 published by the Florida Department of Law Enforcement, which is accessed online. Variables used in my study include the following: number of officers, minimum age required for employment, length of standard shift in hours, minimum education required for employment, and average salary. Demographic characteri stics of officers for individual police departments are not available on line. For those variables that changed annually during the study period, the averag e value is reported. Methodology Interrupted Time-Series Design Interrupted time-series designs are widely us ed to test policy impacts and are an example of a strong quasi-experimental design (Cook & Campbell, 1979). Despite the frequency of use, interrupted time-series designs are not true e xperiments and thus are subject to methodological criticisms. Britt and his collea gues (1996) note that impact assessments have the following key methodological problems: many studies lack or us e inappropriate contro l series, researchers often improperly specify an impact model, a nd impact assessments may be influenced by the 10 http://www.fdle.state.fl.us/CJST/CJAP/ 46


length of pre-and-post-intervention time periods Attending to the critiques of Britt and his colleagues (1996), this section discusses the methodology employed to test the organizational deviance amplification hypotheses, including: the use of control se ries, impact model, and length of pre-and-post intervention periods. While single time-series designs do not utilize control series, my study employs the use of multiple, interrupted time-series design, allowing for analysis of comparison groups. The police agency choices for comparison are limited by the tr affic citation data and thus a perfect control agency is unavailable. However, by definiti on, quasi-experiments lack strictly comparable control groups (McDowall et al., 1996). Conc lusions made from models employing multiple control series that are thoughtfully selected are often stronger than those made from the use of single, appropriate control (McD owall et al., 1996). While th e police agency choices were limited by the data, those selected for analysis had to meet the following criteria: 1) town has its own police agency, 2) town has a population und er 10,000, and 3) town has at least one major U.S. route passing through the city limits. The sel ection criteria resulted in the identification of two control police agencies. While there are mo derate differences between the intervention and control town characteristic s (e.g., population, median household income, % high school graduate) and the intervention a nd control police depart ment characteristics (e.g., salary, number of officers), overall the three rural towns are si milar across many dimensions (see Tables 4-1 & 4-2). Specifying an impact assessment is often problematic because the researcher does not always know exactly when the effect takes pl ace. In impact assessment studies, the policy change date is often used as the date in which to assess the impact. Due to publicity, true effects of policy changes may happen before or after a part icular law is technicall y enacted (Britt et al., 47


1996). While it may be tempting to search for th e date of the true effect, scholars caution against fishing for intervention effects because on e increases the likelihood of finding effects due to random chance (McDowall et al., 1996). To avoi d fishing for intervention effects, the impact Table 4-1. Alachua, High Springs and Waldo town comparison. Alachua, FL High Springs, FL Waldo, FL Population 6,098 3,863 821 Median household income $38,075 $34,354 $24,028 Avg. family size 2.97 2.98 3.05 Median Age 37.1 39.4 36.9 % Female 53.3% 53.9% 52.3% % Non-White 32.4% 23.6% 23.4% % Hispanic 3.6% 4.1% 2.1% % High school graduate 83.0% 82.5% 72.7% % Civilian veterans 13.3% 13.8% 18.4% % In labor force 64.5% 62.2% 56.4% % Individuals below poverty 16.0% 12.0% 16.7% Table 4-2. Alachua, High Springs, and Waldo police department comparison. Alachua, FL High Springs, FL Waldo, FL Number of officers* 18 11 7 Minimum age 19 19 21 Standard shift 12 hr. 12 hr. 12 hr. Minimum education HS/GED HS/GED HS/GED Salary* $26,557.40 $22,750.00 $19,600.00 *denotes average 2001-2005 month is specified in advance to be August 2003the month in which the deviant label was ascribed. Moreover, it is important to specify the expected relationship of the intervention in advance (Campbell & Stanley, 1966). From the con cept of secondary deviance, it is anticipated that organizational deviance amplification effects will be permanent. That is, amplification of deviant behavior should not regress to lower levels in the future. 11 Because the theoretical model of organizational deviance amplification does not require th e exclusion from conventional opportunities and the acquisition of supportive devi ant others, the alteration of organizational 11 While the effects are thought to be permanent, it should be noted that organizational identity is a fluid construct and could change again in the future. 48


identity is believed to occur relatively quickl y. Thus, an abrupt-permanent impact assessment model will be utilized to test the three organizational devian ce amplification hypotheses. Impact assessments can be affected by increase s or decreases in the lengths of the pre-andpost-intervention periods (Britt et al., 1996). Generally, a model with more time points is perceived to be better; but, the longer a time-ser ies is, the greater the chance the data will be influenced by historical threats (McDowall et al., 1996), some of which may be unknown to the researcher. Thus, short time-series data may fail to account for larger trends that may be occurring, whereas long time-series data may be in fluenced by history. The lengths of the preintervention and post-intervention periods in my study are fixed at two years around the August 2003 impact month. Thus, there ar e a total of 49 time-series m onthly data points broken down into 24 pre-intervention, 1 interven tion, and 24 post-intervention months. It should be noted that the pre-intervention and post-inte rvention periods may be somewhat short, which could influence the results and, thus, is a limitation of the data worth acknowledging upfront. The interrupted time-series design consists of one intervention police agency a nd two control police agencies with 49 monthly observations (see Figure 4-1). Alachua O1O24 O25 O26O49 High Springs O1O24 O25 O26O49 Waldo O1O24 X25 O26O49 Figure 4-1. Interrupted time-series design Intervention Specification and Impact Assessment The intervention being assessed is the presence of the speed trap billboards, which serve as the deviant label. As Lemerts (1951) seque nce of interaction pro cess suggests, the police agencys primary deviance resulted in social penalt ies that failed to curb the deviant behavior. Ultimately, the interaction between organizations and the police department culminated in the ascription of the speed trap label. As an ex ample of a societal reaction to primary speed 49


enforcement deviancy, AAA worked with state o fficials in 1999 to inst all reflective tape along the highway (Associated Press, 1999). This social penalty sought to reduce the speed of motorists so that the police agency would not be able to issue citations at a high rate. Since social penalties failed to stop the deviant speed enforcement behavior, AAA formally labeled the town a speed trap in August of 2003 by pl acing billboards outside the city limits. Recall that an important factor influencing the likelihood of devi ance amplification for individuals is the level of publicity of the deviant label. Specifically, if the deviant label is made public, a contingent model of deviance amplificati on suggests that it is more likely that the individual will transition into secondary deviance (Paternoster & Iovanni, 1989). In this case, the deviant label applied to the organization is certainly made public; unlik e most deviant labels, the one ascribed to the police agency is, quite litera lly, visible (see Figure 1-1) It is important to note the highly stigmatizing nature of the speed tr ap billboards. The police agency is only one of two in the nation to have been ascribed the deviant label by AAA12, which makes the stigma associated with the deviant label particularly intense. Moreover, the stigma is pervasive because the police agency was thrust into the nationa l spotlight for engaging in aggressive speed enforcement. While the ascription of the speed trap label makes it clear that the construed external image of the organizati on is negative, the pre-label reput ation is also negative since the police agency has been in the media for its devian t speed enforcement behavior in the past (see Sack, 1995; Associated Press, 1999). Because the stigmatizing label is intense and pervasive and the Waldo Police Department has a negative pre-labe l reputation, the police agency is very likely to alter its organizational ident ity as opposed to devoting resour ces and efforts to attempt to change its organizational image. 12 Lawtey, FL was also labeled a speed trap by AAA in August 2003. 50


An abrupt-permanent intervention model will be used to determine if the deviant label had any amplifying effects on the citation issuance behavior, which is measured by: (1) number of speeding citations issued per month (SpeedCitati on), (2) average monthly speed differential (SpeedDifferential), and (3) percentage of out -of-state drivers receiv ing speeding citations (Out-of-StateCitation). To test the abrupt-perma nent model, Intervention is defined as a step function with the 24 pre-intervention months code d as and the intervention month and the 24 post-intervention months coded as (i.e., step 25). Time-series analysis will be conducted using SPSS statistical software and auto-regressi ve integrated moving average (ARIMA) models will be used when appropriate. If time-series analysis reveal s the lack of autocorrelation, trending, and moving averages, the intervention imp act will be estimated using OLS regression. 51


CHAPTER 5 RESULTS Descriptive Analysis Regional Overview In 2003, the Waldo Police Department issu ed 4,738 speeding citations and 6,452 total citations, the latter of which include criminal violations, non-criminal moving violations, and non-moving violations. Waldo was one of only six of the 115 towns analyzed in which speeding citations made up more than 70% of the total numb er of traffic citations issued in 2003. Waldos percentage of speeding citations to the total number issued equaled 73.43% whereas the average for all agencies was 25.69% (N =115). The Waldo Police Department averaged 526.44 speeding citations per officer in 2003, ma king Waldo only one of three towns to eclipse the 500 speeding citations per year per officer mark. For agenci es with valid data, th e 2003 mean value was 75.37 speeding citations per officer and the median value was 27.60 sp eeding citations per officer (N=107). Waldo had the highest officer to population ratio in 2003 with a value of 10.96 officers per 1,000 residents. For agencies with valid data, the mean value was 3.22 officers per 1,000 residents and the median value was 2.82 officers per 1,000 residents (N=107). Population baseline comparisons In 2003, the Waldo Police Department issued 15.81 daily speeding citations per 1,000 residents and 21.53 daily total citati ons per 1,000 residents. In the same year, the daily speeding citation issuance rate ac ross all cities was 0.62 per 1,000 resident s and the median value equaled 0.18 per 1,000 residents (N=115). Th e daily total citation issuance rate across all cities equaled 1.55 per 1,000 residents with a median value of 0.88 per 1,000 residents (N=115). The outlying nature of the Waldo Police Departme nts citation issuance behavior ca n best be illustrated with a simple boxplot (see Figure 5-1). 52


A B Figure 5-1. Population baseline 2003 daily citation issu ance rates. A) Speeding citation issuance rates (N = 115), B) Total cita tion issuance rates (N = 115). For both speeding citation issuance and genera l citation issuance, Waldo has the highest rate using town population size as the baseline for comparison. While there are seven extreme outliers for speeding citation issuance, the Waldo and Lawtey Police Departments are substantially farther away from the inter-quartil e range than are the other extreme outlier police departments. The second major outlier, the Lawt ey Police Department, was the only other police department to be given the deviant label by AAA.13 In addition to being an outlier for speeding citation issuance, Waldo maintains its distinction as the greatest extreme outlier for all offenses. When using population data as the baseline for comparison, the Waldo and Lawtey police departments exhibit significantly different citation issu ance behavior when compared to the other 113 police agencies, especially for speeding citation issuance. 13 Time-series data are unavailable for Lawtey. Data were obtained from the Alach ua County Clerks Office and Lawtey is in Bradford County, FL. Thus, deviance am plification effects for Lawtey are not measured. 53


Traffic volume baseline comparisons Using the 2003 AADT count data as the ba seline for comparison, Waldo averaged 10.17 daily speeding citations per 10,000 vehicles in 2003. The Waldo police also issued 13.85 daily total citations per 10,000 ve hicles in 2003. For agencies with va lid data, the mean value for daily speeding citations issued was 1.54 per 10,000 vehicl es, and the median daily value was 0.79 per 10,000 vehicles (N=96). For agencies with valid da ta, the mean daily value for all offenses was 5.99 per 10,000 vehicles and the median daily va lue was 3.57 per 10,000 vehicles. For speeding A B Figure 5-2. Traffic volume baseline 2003 daily c itation issuance rates. A) Speeding citation issuance rates (N = 96), B) Total c itation issuance rates (N = 96). citation issuance, Waldo has the second, not the first, highest rate and the gap between Waldo and other extreme outlier police agencies is somewhat narrowed. Still, the Waldo Police Department is still an extreme outlier for speedin g citation issuance (see Figure 5-2 A). For total citation issuance, the Waldo Police Department is not an extreme outlier (see Figure 5-2 B). This finding is likely due to the fact that speeding citations make up a very large percentage of 54


the total number of traffic citations issued by the Waldo Police Department. While the population baseline data show the Waldo Police Department to be an extreme outlier, the traffic volume baseline data reveal that the specificity of the speed trap label ascribed to the Waldo Police Department may be somewhat justified. Local Details Between August 2001 and August 2005, Waldo issued 28,286 traffic citations, while Alachua and High Springs issued 12,140 and 3,879 traffic citations, respectively. During the same time span, Waldo issued 22,645 speeding c itations while Alachua and High Springs issued 6,987 and 1,188 speeding citations, resp ectively. The percentages of speeding citations issued to the total number are as follows: Waldo = 80.01%, Alachua = 57.55%, and High Springs = 30.62%. The demographic characteristics (e.g., non-White, Female, Hispanic, Out-of-State, Age) of individuals receiving speeding citations in the three towns during the study period are similar across some aspects and different across ot hers. It appears that one major difference between the intervention and control agencies is th at Waldo issued a much greater percentage of its speeding citations to out-of-st ate drivers (see Table 5-1). This may be the reflection of the driving population traveling through the town of Waldo. The major highway cutting through Waldo provides for a more direct route from nor theast Florida (and points north) to north central and west central Florida than does the interstate system. Hence, Waldo may have more out-ofstate drivers trave ling through its town. In order to determine whether the demographics of individuals receivi ng traffic citations in the three towns significantly differ from one another, t -tests are conducted. Since there are a large number of cases (N=30,820), differences am ong the towns are likely to be statistically significant. As a result, Cohens d values are reported in addition to the t -test statistic. Cohens d is a measure of effect size, which enables one to compare differences in the demographic 55


Table 5-1. Demographic characte ristics of individuals receiving speeding citations in Alachua, High Springs, and Waldo (N=30,820). Alachua High Springs Waldo Frequency (N) Frequency (N) Frequency (N) Non-White 23.12% (1615) 6,984 10.19% (121) 1,188 16.38% (3705) 22,620 Female 36.74% (2566) 6,985 36.62% (435) 1,188 30.27% (6852) 22,633 Hispanic 2.13% (149) 6,984 2.02% (24) 1,188 1.82% (412) 22,620 Out-of-State 11.12% (774) 6,958 4.65% (55) 1,184 32.23% (7202) 22,348 Mean (SD) Mean (SD) Mean (SD) Age 33.68 13.25 36.29 14.14 39.31 15.95 characteristics of individuals receiving speeding citations in the three towns. For out-of-state speeding citation percentage, the High Springs -Waldo effect size has a magnitude of d=0.76 and the Alachua-Waldo effect size has a magnitude of d=0.53. The out-of-state differences are statistically significant and the e ffect sizes are moderate to larg e. The Waldo Police Department issued a smaller percentage of its speeding tick ets to women drivers comp ared to the other two towns but these effect sizes are relatively small. The Waldo Police Department issued a greater percentage of its speeding tickets to older drivers and the eff ect sizes range from small to moderate. There were no significant differe nces among the three police agencies speeding citation issuance for Hispanic drivers. Additio nally, there were minimal significant differences between Waldo and each of the control agencies for non-White drivers (see Table 5-2). With the lack of data for other comparison agencies, it cannot be determined if these statistical differences are substantial in a regiona l or national context. If the differences were deemed to be substantial in a regional or nationa l context, without a tra ffic violating population baseline for comparison, it cannot be determined if any observed differences are the result of divergent policing pr actices or are simply the reflection of driving populations. Within a local 56


Table 5-2. t -Test and Cohens d: Differences in speeding cita tion issuance among Alachua, High Springs, and Waldo Police Departments. 1. Non-White; 2. Female; 3. Hispanic; 4. Out-of-State; 5. Age Alachua High Springs Waldo Alachua ------------High Springs 1. t (2062) = 12.78, p<0.001 d = 0.35 2. n.s. 3. n.s. 4. t (2198) = 9.01, p<0.001 d = -0.24 5. t (1562) = -5.94 p<0.001 d = -0.19 --------Waldo 1. t (10516) = 12.02, p<0.001 d = 0.17 2. t (11176) = 9.89, p<0.001 d = 0.14 3. n.s 4. t (17273) = -43.09, p<0.001 d = -0.53 5. t (13777) = -26.76, p<0.001 d = -0.38 1.t (1380) = -6.79, p<0.001 d = -0.18 2. t (1303) = 4.43, p<0.001 d = 0.13 3. n.s 4. t (1874) = -40.14, p<0.001 d = -0.76 5. t (1351) = -6.382, p<0.001 d = -0.20 ----context, it is important to note that the demographi cs of individuals receiv ing traffic citations in the three towns are significantly different in so me ways. This does not mean that Alachua and High Springs are necessarily inap propriate control towns; the qua si-experimental design utilizes each towns pre-intervention cita tion issuance behavior and compares it to the post-intervention citation issuance behavior. Intervention Effects Hypothesis One Recall hypothesis one predicted that speeding ci tation issuance would increase as the result of the ascription of the deviant label to the Wa ldo Police Department. The sequence plots reveal potential effects of the devian t label on the speeding citation i ssuance behavior for Waldo (see 57


Figure 5-3 D). More specifica lly, the number of speeding cita tions issued per month by the Waldo Police Department increases after the ascrip tion of the deviant label. The sequence plots reveal that Alachua an d High Springs are relatively unaffected by the intervention. Notably, Alachua has a significant shift in the number of citations issu ed from month 13 (August 2002) to month 17 (December 2002). Examination of the autocorrelation (ACF) and partial autocorrelation (PACF) plots of the pre-intervention months reveals autocorrelation in the citation issuance behavior of the Alachua and Waldo Police Departments (see Appendix). Autocorrelation is confirmed for Alachua and Wa ldo with a constant onl y pre-intervention AR1 model yielding significance at the = 0.01 level. There appears to be periodicity for Waldo with significant values at lag 7 and lag 14 of the AC F plot. However, lags 7 and 14 do not coincide with annual seasonality and there is little theoretical reason to exp ect a 14 month cycle. The best fit to the time-series data is an ARIMA(1,0,0), in which 1 autoregressive term is used and neither differencing nor moving average terms are used. The ARIMA(1,0,0) model is first run with only the Intervention (i.e ., step function; step 25) as the independent va riable. OLS regression results are reported for High Springs because the time-series data lacked autocorrelation and did not need an AR1 correction term. The Intervention is significant at the = 0.001 level for Waldo but is insignificant for the control agenci es (see Table 5-3). Th e sequence plots reveal that the timing of the intervention corresponded we ll with the incr ease in citation issuance by the Waldo police department but this increase is not abrupt and appears to occur over time. To determine if the intervention likely cause d an increase in citation issuance, it is necessary to control for other factors that could have caused an increase in citation issuance. Two of the most likely factors include increase in traffic volume and hiring of additional police officers. The latter could serve as a mediati ng variable whereas the former could negate 58


A B C D Figure 5-3. Monthly speeding cita tion issuance sequence plots. A) All Towns, B) Alachua, C) High Springs, D) Waldo. 59


Table 5-3. Uncontrolled speeding citati on issuance intervention effects model. Estimates Std. Err. t Sig. Non-seasonal Lags (AR1) .674 .107 6.312 .000 Intervention (Step 25) 48.617 32.031 1.518 .136 Alachua Constant 117.117 25.408 4.609 .000 OLS Intervention (Step 25) 2.603 5.489 0.474 .638 High Springs Constant 22.917 3.921 5.845 .000 Non-seasonal Lags (AR1) .301 .139 2.166 .036 Intervention (Step 25) 116.012 31.711 3.658 .001 Waldo Constant 404.122 22.920 17.632 .000 Table 5-4. Controlled speeding citation issuance intervention effects model. Estimates Std. Err. t Sig. Non-seasonal Lags (AR1) .625 .113 5.543 .000 Intervention (Step 25) 27.551 30.790 .895 .376 Officers 2.920 2.384 1.225 .227 Traffic .026 .016 1.647 .107 Alachua Constant -255.880 197.372 -1.296 .202 OLS Intervention (Step 25) -5.176 6.787 -.763 .450 Officers 3.053 1.574 1.940 .059 Traffic .007 .008 .923 .361 High Springs Constant -63.435 64.969 -.976 .334 Non-seasonal Lags (AR1) .177 .152 1.169 .249 Intervention (Step 25) 148.049 30.764 4.812 .000 Officers 16.753 11.089 1.511 .138 Traffic -.032 .028 -1.109 .273 Waldo Constant 661.301 375.194 1.771 .084 observed intervention effects. The number of police officer s is believed to be a factor that would increases citation issuance. Tra ffic volume could either increase or decrease citation issuance. Increased traffic volume would likely increase the number of speeders traveling through the towns, which could result in more speeding cita tions being issued. However, heavy traffic volume could cause congestion and actually lim it the opportunity to sp eed. The latter is somewhat unlikely since the towns are rural. As expected, Alachua and High Springs still have non-significant intervention e ffects. For Waldo, the inte rvention effects found in the 60


uncontrolled model remain significant for the controlled model at the = 0.001 level (see Figure 5-4). The traffic volume in a given month does not significantly impact citation issuance in any of the towns. There appears to be sufficient individuals to cite each month despite monthly fluctuations in traffic volume. The number of police officers for High Springs is marginally significant at the = 0.1 level with a p-va lue of 0.059. Hypothesis Two The second organizational deviance amplificatio n hypothesis predicts that individuals will be issued traffic citations with a lesser speed differe ntial in Waldo followi ng the intervention. By looking at the sequence plots, it does not appear that hypothesis two is supported by the data. That is, there is no abrupt and permanent effect on speed differential in the expected direction (see Figure 5-4). Despite what l ooks to be linear trends in th e data during the pre-intervention months for Alachua and Waldo, the autocorrelation and partial autocorrela tion plots reveal that none of the series needs to be differenced to remove a linear trend (see Appendix). Moreover, differencing the 24 pre-intervention months over-differenced both tim e-series, which resulted in significant negative auto correlation. Autocorrelation is c onfirmed for only Waldo with a constant only pre-intervention AR1 model yielding significance at the = 0.05 level. Because only Waldo needs an AR1 correcting term, OLS regression will be used for both Alachua and High Springs in examining hypothesis two. The results of the uncontrolled OLS regression models indicate that both the Alachua and High Springs police departments were unaffected by the intervention (see Table 5-5). That is, the two control police departments did not issue speeding citations with a greater or lesser speed differential after the deviant label was ascribed to the Waldo Police Department. However, the results of the ARIMA(1,0,0) model indicate that the Waldo Police Department may have issued speeding citations with a larger speed differentia l after the deviant label (see Table 5-5). 61


A B C D Figure 5-4. Monthly average spee d differential sequence plots. A) All Towns, B) Alachua, C) High Springs, D) Waldo. 62


Table 5-5. Uncontrolled speed diffe rential intervention effects model. Estimates Std. Err. t Sig. OLS Intervention (Step 25) -.134 .308 -.435 .665 Alachua Constant 18.920 .220 86.039 .000 OLS Intervention (Step 25) -.338 .524 -.644 .522 High Springs Constant 17.896 .374 47.792 .000 Non-seasonal Lags (AR1) .517 .130 3.976 .000 Intervention (Step 25) 1.289 .278 4.634 .000 Waldo Constant 17.144 .207 82.901 .000 Table 5-6. Controlled speed differe ntial intervention effects model. Estimates Std. Err. t Sig. OLS Intervention (Step 25) .175 .418 .418 .678 Officers .034 .059 .568 .573 Traffic .000 .000 -1.559 .126 Alachua Constant 24.466 3.755 6.515 .000 OLS Intervention (Step 25) .227 .668 .340 .735 Officers -.015 .155 -.097 .923 Traffic -.001 .001 -1.425 .161 High Springs Constant 27.149 6.399 4.243 .000 Non-seasonal Lags (AR1) .581 .129 4.508 .000 Intervention (Step 25) 1.240 .322 3.857 .000 Officers .120 .075 1.594 .118 Traffic .000 .000 1.340 .187 Waldo Constant 12.887 2.673 4.821 .000 The results of the controlled model yields insi gnificant intervention effects for Alachua and High Springs but significant intervention effects at the = 0.001 level for Waldo (see Table 5-6). This effect is in the direction counter to wh at was predicted in hypothesis two. That is, the Waldo Police Department was issu ing traffic citations with a s lightly higher monthly average speed differential after the ascription of the devi ant label. These results should be interpreted with caution. They could be influenced by the relatively high speed differential for month 8 (March 2002). Specifically, this high value may have resulted in the statistical software treating 63


the pre-intervention months as a stationary series, when it does appear as if the series is trending upward prior to the ascrip tion of the deviant label.14 Hypothesis Three The third exploratory organiza tional deviance amplification hypot hesis states that out-ofstate drivers may make up a larger or smaller percentage of dr ivers ticketed for speeding in Waldo. The intervention appears to have no effect on the targeting of out-of-state motorists (see Figure 5-5). One potential permanent change in the percentage of out-of-state drivers ticketed occurs in Alachua (see Figure 5-5 B).15 However, the deviant label does not appear to be responsible for possible changes in the percentage of out-of-state drivers being cited for speeding in Alachua. The shape of the sequence plot (see Figure 5-5 B) resembles the shape noted previously for speeding citation issuance (see Figure 5-3 B). One interpretation is that a policy change occurred in Alachua, which resulted in an overall increase in the number of traffic citations issued, with many of these additional traffic citations being issued to out-of-state drivers. The sequence plot for Waldo (see Figure 5-5 D) reveals annual seasonality in the data. Adding an annual seasonal differencing term to the model eliminates the seasonality. The ACF and PACF plots reveal autocorrelation effects for Waldo but not for Alachua or High Springs. Autocorrelation is confirmed for only Waldo with a constant only pre-intervention AR1 model yielding significance at the = 0.01 level. Thus, OLS regressi on will be used for Alachua and High Springs and an ARIMA(1,0,0)x(0,1,0)12 model will be used for Waldo. The results of the uncontrolled model reveal a significant before and after difference for Alachua but no such effect for High Springs or Waldo (see Figure 5-7). This result should be 14 Because there was a potential trend in the data, an add itional differenced model was ru n (i.e., ARIMA (1,1,0)). In this model, the intervention was found to have no effect on speed differential for any of the police agencies. 15 A longer pre-spike period is needed to establish a baseline in order to determine if in fact a change occurred. 64


A B C D Figure 5-5. Monthly percentage of out-of-state speeding citations sequence plots. A) All Towns, B) Alachua, C) High Springs, D) Waldo. 65


Table 5-7. Uncontrolled percentage out -of-state intervention effects model. Estimates Std. Err. t Sig. OLS Intervention (Step 25) .037 .013 2.924 .005 Alachua Constant .081 .009 9.028 .000 OLS Intervention (Step 25) -.022 .015 -1.455 .152 High Springs Constant .053 .011 4.956 .000 Non-seasonal Lags (AR1) .548 .143 3.829 .001 Intervention (Step 25) -.016 .021 -.751 .458 Waldo Constant .001 .014 .057 .955 Table 5-8. Controlled percentage outof-state intervention effects model. Estimates Std. Err. t Sig. OLS Intervention (Step 25) .005 .016 .294 .770 Officers .003 .002 1.176 .246 Traffic .000 .000 2.317 .025 Alachua Constant -.302 .144 -2.093 .042 OLS Intervention (Step 25) -.019 .019 -1.029 .309 Officers .006 .004 1.466 .150 Traffic .000 .000 -1.030 .308 High Springs Constant .184 .180 1.017 .314 Non-seasonal Lags (AR1) .528 .148 3.577 .001 Intervention (Step 25) -.021 .024 -.889 .381 Officers .001 .005 .140 .890 Traffic .000 .000 1.093 .283 Waldo Constant -.004 .015 -.258 .798 interpreted with caution because a spike may have occurred in the data before the intervention. While the billboard may have had some effect in Alachua, the OLS m odel may be picking up significant effects for some other intervention (e .g., crack down on drug runners) that occurred only a few months prior. There are potential effects of control variables on the percen tage of out-of-state drivers receiving traffic citations. Traffic volume can be affected by the weather and seasons. Moreover, Florida is a popular va cation destination and more out-o f-state traffic are anticipated 66


to be on the roads during peak vacation months. Thus, increased traffic volume may be related to increased percentage of out-of-state driver s receiving speeding cita tions. The number of officers is likely unrelated to the percentage of out-of-state drivers recei ving speeding citations. The intervention effects found for Alachua in the uncontrolled model are eliminated in the controlled model. Thus, the inte rvention did not have an effect on the citation issuance behavior of any of the three police departments. In the co ntrolled model, traffic volume is related to the percentage of out-of-state drivers receiving speed ing citations in Alachua but not in High Springs or Waldo (see Figure 5-8). As expected, the numb er of officers issuing citations has no effect on the percentage of out-of-state drivers receiving traffic citati ons for these towns. 67


CHAPTER 6 DISCUSSION AND CONCLUSIONS The theoretical model of organizational devi ance amplification combines organizational and labeling theories to specify the pathway of organizational transition from primary to secondary deviance. The process begins when a de viant label is applied to an organization. This sets off an identity-image comparison in which the organization compares what they think they are to what outsiders think they are. In most cases, when a deviant label causes this comparison and the incongruence is perceived, the organi zation will respond by attempting to alter its construed external image. However, when reput ation is not essential for organizational success and survival, deviance amplification can occu r. Organizations w ith negative pre-label reputations and those that are ma rked with intense and pervasive stigmatizing labels are most likely to alter their organizational identity as opp osed to attempting to change their construed external image. Supportive deviant others a nd exclusion from conven tional opportunities are typically exacerbating, not necessary, conditions for organizational deviance amplification. The proposed theoretical model of organizational deviance amp lification was tested on the Waldo Police Department, which was labeled deviant in August 2003 for their aggressive citations issuance behavior. Regi onal comparisons were made to examine the legitimacy of the applied deviant label. When using populati on as the baseline for citation issuance rate comparison, Waldo was clearly the greatest extreme outlier. When utilizing traffic volume as the baseline comparison, Waldo stands out as an extr eme outlier but is eclipsed by one agency and several others are comparatively much closer. This begs the question: why are other police departments that emerge as extreme outliers not labeled speed traps? One potential explanation for the reason why other police departments that are extreme outliers are not labeled deviant is that these agencies may issue traffic citations in order to 68


address particular safety issu es (e.g., curvy roads, high fatal ity highway segments). AAA does identify a number of areas acro ss the country as strict enfo rcement areas. These police departments are noted for their aggr essive citation issuance behavior in the name of safety. If the other police agencies in Florida that are extreme outliers are not addressing specific safety issues, an alternative explanation as to why Waldo and Lawtey were la beled deviant is needed. An account that fits the data well is that comparisons of traffic citation behavior among police agencies are made on the basis of town size and not on the basis of traffic volume. If this is the case, high traffic volume may actually make certain towns more likely to be categorized as deviant. This could occur because as more i ndividuals travel through the town, more persons have the opportunity to become aware of the abnormal citation issuance behavior. Whether the police department is actually issuing speeding citations at an elev ated rate, when using the best available baseline comparison data, may be irrele vant as long as there is the perception of deviant behavior. The results of the quasi-experiment part ially support the theoretical model of organizational deviance amplification. The inte rvention analysis lend s support to hypothesis one. Increased citation issuance by the Waldo Police Department followed the ascription of the deviant label. Interes tingly, this effect occurre d despite the fact that the speed trap billboards warn motorists of the citation i ssuance behavior of the labeled po lice agency. The deviant label may deter many drivers from speeding through the labeled town, whic h would make this organizational deviance amplification effect ev en more notable. That is, the Waldo Police Department may be writing more speeding citations despite the fact that there are less speeders traveling through the tow n. While an abrupt-permanent imp act was anticipate d, it appears that the deviance amplificatio n may have occurred over time (s ee Figure 5-3 D). Abrupt-permanent 69


impact assessment models can be sufficient es timates for gradual-permanent impact assessment models (see McDowall et al., 1996). A gradual-permanent impact model may have fit the data better and the support found for the abrupt-permanent impact may be an estimation of the true gradual-permanent impact. Perhaps, the alteration of organizational identity is a slower process than what was hypothesized. Or, the alteration of organizational identity may occur relatively quickly but the deviant behavior may continue to amplify over time. Additional time series data points would be necessary to examine if this is th e case. Future labeling research that utilizes interrupted time-series designs should maximi ze the lengths of both the pre-intervention and post-intervention periods whenever possible. This enables a good trend line to be established and allows for a long-term impact assessment. The analysis did not lend support to hypotheses two or three. That is, the deviant label neither caused the Waldo Police Department to i ssue tickets with a lesser speed differential nor to target out-of-state drivers. Hypotheses two and three can be conceived of as testing for changes in the quality of the deviant behavior, whereas hypothesis one can be conceived of as testing changes in the qu antity of it. The results suggest that the Waldo Police Department increased the quantity of the citation issuance de viant behavior, but did not alter the quality; simply put, the deviant label led to more of the same behavior. Fu ture research in organizational labeling should focus on both quantity and quality measures of amplification effects. Because data were only available for one labelee and a quasi-experimental design was utilized, my study necessarily tested a determ inistic view of organizational labeling. The theoretical model of or ganizational deviance amplification is a probabilistic model. Quasiexperimental designs are not well suited for the empirical testing of probabilistic theories of crime and deviance. Since the concept of organi zational deviance amplification is in its infancy, 70


quasi-experiments may be utilized to test th e theoretical model of organizational deviance amplification in the near future. This is not necessarily a bad thing be cause testing the model with a quasi-experimental design will reveal one of two us eful pieces of information. If there is no support found for the model, attention can be turned to exploring probabilistic models. Alternatively, if evidence is f ound in support of the model, then it tells us that organizational deviance amplification is possibl e and that further research on the subject is warrantedthe latter is the case with my study. While quasi-exper iments are useful in early stages of empirical testing, future research should focus on testing the theoretical model w ith larger samples of labeled and non-labeled organizations. Formal pr ocessing in the criminal justice system has been used to examine the effects of deviant labe ls on individual behavior While obtaining large samples of labeled and non-labeled organizations may be more difficult, one potential way to obtain a suitable sample may be to draw from co urt records to obtain a set of organizations that have been formally processed and a comparis on group that has escaped formal processing. While this would only permit the effects of deviant labels ascribed by the criminal justice system to be examined, it would nevertheless allow for a more complete test of the proposed probabilistic theoretical mode l of organizational deviance amplification. Unfortunately the data in my study did not perm it a test of mediating variables. Testing the complete theoretical model of organizational deviance amplif ication requires the ability to identify certain organizational concepts. There is a distinction in organi zational theory between what an organization believes others think about the organization and what others actually do think about the organization (see Brown et al., 2006). While the ascription of a deviant label gives some insight into what the construed external image is likely to be, this may not be entirely accurate. An organization could genuinely believ e that others view the entity as legitimate 71


despite the fact that others in reality do not. Then, measuring the construed external image is best done by gathering information directly from the organization. Intense and pervasive deviant labels make it very clear what others think a bout the organization. In these instances, only a refusal to acknowledge this information would result in any discrepancy. Thus, information conveyed by a particular stigmatizing label can be a good approximation for what the construed external image actually is. From the process model of identity-image interdependence (see Gioia et al., 2000), the other key concept to organiza tional deviance amplification is organizational identity. Determining whether an organization has assume d a deviant identity may be a much more difficult task. Another distinc tion in organizational theory is between what the organization thinks they are and what they want others to think about them (see Brown et al., 2006). Most organizations want to be viewed in a positive light and thus their projected image (see Gioia et al., 2000) may not be indicative of their true or ganizational identity. Moreover, organizations may not reveal their true identity if there are an y constituents in which the organizations seek to satisfy. This may be the case for the Waldo Police Department since they ar e directly serving the local townspeople and have repeated interactions with them. Their projected image may be to maintain credibility with the local townspeopl e whereas the organizatio n could not care less about its reputation to outsiders and society at large. An organization can accept a deviant identity but not openly acknowledge it. Undisclo sed organizational identities create a serious problem for measuring the mediating variable,16 which is central to th e notion of organizational deviance amplification. Future research should draw from orga nizational theory in order to operationalize key theoretical concepts. Quantitative measures of organizational identity and 16 If an organization was reluctant to ever reveal what it is a whistleblower would be necessary to reveal its identity. Otherwise, its true identity may remain unknown. 72


construed external image are, in many cases, idea l. If a measurement dilemma occurs, in which organizational identity differs from projected image, settling for qualitative data from a whistleblower may be a decent option. However, the accuracy of information obtained from whistleblowers should be scrutinized because these individuals may have personal motives for divulging information. It is possible that neither a whistl eblower steps forward nor the organizations reveals its true identity. Scholarly attention should attend to the issue of incongruity between projected image and organizational identity. With the preceding discussion on measuring ke y concepts imbedded in the theoretical model of organizational deviance am plification, perhaps future research should begin to focus on organizations that are likely to prevent a m easurement quandary. While my study examined a public service organization, this type of organization is likely to project a positive image in most cases. Some organization types may be particularly good candidates to examine the effects of deviant labels because they may be unlikely to project positive organizational images when labeled. For example, civilly disobedient organizations labeled deviant could accept their deviant identity and become overtly resistant. If the deviant amplification were to progress to acts of terrorism, the projected image would likel y match the organizational identity because the organization would be attempting to instill fear Future research should focus on organization types that are most likely to confirm or refute organizational deviance amplification hypotheses instead of selecting labeling occurrences that may lead to a measurement quandary. The concept of organizational deviance amp lification has important implications for official and societal responses to deviant behavior. At least theoretically, the ascription of deviant labels to organizations that need not maintain positive reputations can lead to an amplification of deviant behavior. As for indivi duals, ascribing deviant labels to organizations 73


can make the behavior worse. This is not to say that society should a void sanctioning deviant organizations, but that society is cognizant that intervening may have the opposite of the intended effect. There are three societal reactions that may re duce or prevent organizational deviance amplification. The firs t societal reaction which may pr event deviance amplification is the imposition of a severe sanction. Wher eas simply blocking access to conventional opportunities may contribute to deviance amplifi cation, sanctions that directly threaten organizational survival may have a deviance reducing effect. This occurs because organizations seek to prosper and, if no other options are available, will likely alter thei r behavior to avert organizational death. The second societal react ion which could lead to the prevention of deviance amplification is the la ck of a societal reaction. Do ing nothing will stop any deviant label from being ascribed; wit hout a deviant label, secondary deviance is not a theoretical possibility. However, doing nothi ng may result in the worsening of the deviant behavior over time because the organization may realize there are no penalties for its acti ons. Neither of these two societal reactions is ideal because both violate the practice of sanctioning proportionality. The third societal reaction which may reduce the chance fo r organizational deviance amplification is a restorative one. A restorat ive societal reaction (see Braithwaite, 1989) to deviant organizations is one in which the sanctioning entity work s with the deviant organization to eliminate the deviant behavi or and repair the organization s construed external image and reputation. This is an attrac tive option because both parties gain some benefit from the interaction. If research rev eals that organizational deviance amplification is more than a theoretical possibility, furt her attention should be given to the di fferent types of societal reactions and the ways in which these reactions can li mit and prevent devian ce amplification. 74












LIST OF REFERENCES Akers, R. L. (1968). Problems in the sociology of deviance: Social de finitions and behavior. Social Forces, 46 (4), 455-465. Albert, S., & Whetten, D. (1985). Organizational identity. In L. L. Cummings & B. M. Staw (Eds.), Research in organizational behavior (vol. 7, pp. 263-295). Greenwich, CT: JAI Press. Associated Press. (1999, November 26). Wheres waldo? On the radar screen. The New York Times p. D1. Bailey, J.A. (2003). Self-image, self-c oncept, and self-identity revisited. Journal of the National Medical Association, 95 (5), 383-386. Baker, R. P. (1973). A concomitant look at commitment and labeling theory: Divergent but compatible accounts of delinquency causation Unpublished doctoral dissertation, Washington State University, Pullman, WA. Ball, C. (2001). Rural perceptions of crime. Journal of Contemporary Criminal Justice, 17 (1), 37-48. Becker, H. S. (1963). Outsiders: Studies in th e sociology of deviance. New York, NY: The Free Press. Becker, S. (2004). Assessing the use of profili ng in searches by law enforcement personnel. Journal of Criminal Justice, 23 (2), 183-193. Bernburg, J. G., & Krohn, M. D. (2003). Labeling, lif e chances, and adult crime: The direct and indirect effects of official intervention in adolescence on crime in early adulthood. Criminology, 41 (4), 1287-1318. Best, J. (2004). Deviance: Career of a concept. Bellmont, CA: Wadsworth. Bettenhausen, K., & Murnigham, J. K. (1985). The emergence of norms in competitive decision making groups. Administrative Science Quarterly, 30 (3), 350-372. Blumer, H. (1962 ). Society as symbolic interaction. In A. M. Rose (Ed.), Human behavior and social processes: An interactionist approach Boston, MA: Houghton Mifflin. Braithwaite, J. (1989). Crime, shame, and reintegration. New York, NY: Cambridge University Press. Braithwaite, J., & Fisse, B. (1995). On the plausibility of corporate crime control. In G. Geis, R. R. Meier, & L. M. Salinger (Eds.), White collar crime: Classic and contemporary views (3rd ed., pp. 432-449). New York, NY: The Free Press. 80


Britt, C. L., Kleck, G., & Bordua, D. J. (1996). A reassessment of the D.C. Gun Law: Some cautionary notes on the use of interrupte d time series designs for policy impact assessment. Law & Society Review, 30 (2), 361-380. Brown, T. J., Dacin, P. A., Pratt, M. G., & Wh etten, D. A. (2006). Identity, intended image, construed image, and reputation: An in terdisciplinary framework and suggested terminology. Journal of the Academy of Marketing Science, 34 (2), 99-106. Brownstown, D., Ghosh, A., Golob, T. F., Kazimi C., & Van Amelsfort, D. (2003). Drivers willingness-to-pay to reduce travel time: Evidence from the San Diego I-15 congestion pricing project. Transportation Research Part A: Policy and Practice, 37 (4), 373-387. Buerger, M. E., & Farrell, A. (2002). The eviden ce of racial profiling: Interpreting documented and unofficial sources. Police Quarterly, 5 (3), 272-305. Campbell, D. T., & Stanley, J. C. (1966). Experimental and quasi-experimental designs for research. Chicago, IL: Rand McNally. Coleman, J. S. (2002). Organizational actors and th e irrelevance of person s. In M. D. Ermann & R. J. Lundman (Eds.), Corporate and governmental deviance: Problems of organizational behavior in contemporary society (6th ed., pp. 95-104). New York, NY: Oxford. Cook, T. D., & Campbell, D. T. (1979). Quasi-experimentation: Design and analysis issues for field settings. Boston, MA: Houghton Mifflin. Cooley, C. H. (1902). Human nature and the social order. New York, NY: Scribners Sons. Coombs, W. T., & Holliday, S. J. (2006). Unpack ing the halo effect: Reputation and crisis management. Journal of Communication Management, 10 (2), 123-137. Correia, M. E., Reisig, M. D., & Lovrich, N. P. (1996). Public perceptions of state police: An analysis of individual-level and contextual variables. Journal of Criminal Justice, 24 (1), 17-28. Cressey, D. R. (1995). Poverty of theory in corpor ate crime research. In G. Geis, R. R. Meier, & L. M. Salinger (Eds.), White collar crime: Classic and contemporary views (3rd ed., pp. 413-431). New York, NY: The Free Press. Davies, G., Chen, R., Da Silva, R., & Roper, S. (2002). Corporate reputation and competitiveness. London, UK: Routledge. Davis, F. (1961). Deviance disavowal: The manage ment of strained inte raction by the visibly handicapped. Social Problems, 9 (2), 120-132. 81


Davis, N. J. (1972). Labeling theory in devi ance research: A critique and reconsideration. The Sociological Quarterly 13 (4), 447-474. Downs, W. R., Robertson, J. F., & Harrison, L. R. (1997). Control theory, labeling theory, and the deliverance of services for drug abuse to adolescents. Adolescence, 32 (125), 1-24. Dutton, J. E., & Dukerich, J. M. (1991). Keeping an eye on the mirror: Image and identity in organizational adaptation. The Academy of Management Journal, 34 (3), 517-554. Engel, R. S., & Calnon, J. M. (2004a). Examining the influence of drivers characteristics during traffic stops with police: Re sults from a national survey. Justice Quarterly, 21 (1), 49-90. Engel, R. S., & Calnon, J. M. (2004b). Comparing benchmark methodologies for police-citizen contacts: Traffic stop data collection for the Pennsylvania State Police. Police Quarterly, 7 (1), 97-125. Ermann, M.D., & Lundman, R. J. (2002). Corpor ate and governmental deviance: Origins, patterns, and reactions. In M. D. Ermann & R. J. Lundman (Eds.), Corporate and governmental deviance: Problems of organizational behavior in contemporary society (6th ed., pp. 3-49). New York, NY: Oxford. Farrington, D. P. (1977). The effects of public labeling. British Journal of Criminology, 17 (2), 112-125. Fiol, C. M., & Kovoor-Misra, S. Two-way mirrori ng: Identity and reput ation when things go wrong. Corporate Reputation Review, 1 147-151. Fombrun, C. J. (1996). Reputation: Realizing value from the corporate image. Boston, MA: Harvard Business School Press. Fredrickson, D. D., & Siljander, R. P. (2002). Racial profiling: Elim inating the confusion between racial and criminal profiling and clarifying what constitutes unfair discrimination and persecution. Springfield, IL: Charles Thomas Publisher Ltd. Geis, G. (2002). The heavy electrical equipmen t antitrust cases: Pricefixing techniques and rationalizations. In M. D. Er mann & R. J. Lundman (Eds.), Corporate and governmental deviance: Problems of organizational behavior in contemporary society (6th ed., pp. 111130). New York, NY: Oxford. Gibbs, J. P. (1966). Conceptions of deviant behavior: The old and the new. The Pacific Sociological Review, 9 (1), 9-14. Gibbs, J. P. (1972). Issues in defining deviant be havior. In R. A. Scott & J. D. Douglas (Eds.), Theoretical perspectives on deviance (pp. 39-68). New Yor k, NY: Basic Books. 82


Gioia, D. A., Schultz, M., and Corley, K. G. ( 2000). Organizational ident ity, image, and adaptive instability. The Academy of Management Review, 25 (1), 63-81. Gold, W. J., & Williams, J. R. (1969). Nationa l study of the aftermath of apprehension. Prospectus: A Journal of Law Reform, 3 3-38. Gove, W. R. (1980a). The labe ling perspective: An overview. In W. R. Gove (Ed.), The labeling of deviance: Evaluating a perspective (2nd ed., pp. 9-26). Beverly Hills, CA: Sage. Gove, W. R. (1980b). Labeling and mental illn ess: A critique. In W. R. Gove (Ed.), The labeling of deviance: Evaluating a perspective (2nd ed., pp. 53-99). Beverly Hills, CA: Sage. Grimes, R. M., & Turk, A. T. (1978). Labeling in context: Conflict, power and self-definition. In M. Krohn & R. Akers (Eds.), Crime, law, and sanctions: Theoretical Perspectives (pp. 39-58). Beverly Hills, CA: Sage. Hagan, J. & Palloni, A. (1990). Th e social reproduction of a crim inal class in working-class London, circa 1950-1080. The American Journal of Sociology, 96 (2), 265-299. Heckert, A., & Heckert, D. M. (2002). A new typology of deviance: Integrating normative and reactionists definitions of deviance. Deviant Behavior, 23 449-479. Lange, J. E., Johnson, M. B., Voas, R. B. ( 2005). Testing the racial profiling hypothesis for seemingly disparate traffic stops on the New Jersey Turnpike. Justice Quarterly, 22 (2), 193-223. Lemert, E. M. (1951). Social pathology: A systematic appr oach to the theory of sociopathic behavior New York, NY: McGraw-Hill. Lemert, E. M. (1976). Response to critics: Feedback and choice. In L. Coser & O. Larsen (Eds.), The uses of controversy in sociology (pp.244-249). New York, NY: Macmillan. Liazos, A. (1972). The poverty of the sociology of deviance: Nuts, sluts, and preverts. Social Problems, 20 (1), 103-120. Link, B. G., & Cullen, F. T. (1990). The labeling th eory of mental disorders: A review of the evidence. Research in Community and Mental Health, 6, 75-105. Link, B. G., Cullen, F. T., Struening, E., Shrout, P. E., & Dohrenwend, B. P. (1989). A modified labeling theory approach to mental disorders: An empirical assessment. American Sociological Review, 54 (3), 400-423. Liska, A. E. (1987). Perspective on deviance (2nd ed.). Englewood Cliffs, NJ: Prentice Hall. Lorber, J. (1967). Deviance as pe rformance: The case of illness. Social Problem, 14 (3), 302310. 83


Lundman, R. J. (1979). Organizational norms a nd police discretion: An observational study of police work with traffic law violators. Criminology, 17 (2), 159-171. Machin, M. A., & Sankey, K. S. (2006). Predictin g young drivers risk perceptions and speeding behaviour. In M. King (Ed.), Proceedings of the 2006 Austra lasian road safety research, policing and education conference. Gold Coast, Australia: Able Video & Multimedia. Matsueda, R. L. (1992). Reflected appraisals, parental labeling, and de linquency: Specifying a symbolic interactionist theory. The American Journal of Sociology, 97 (6), 1577-1611. McDowall, D., Wiersema, B., & Loftin, C. (1996). Using quasi-experiments to evaluate firearms laws: Comment on Britt et al.s reassessment of the D.C. Gun Law. Law & Society Review, 30 (2), 381-392. Meade, A. C. (1974). The labeling approach to de linquency: State of the th eory as a function of method. Social Forces, 53 (1), 83-91. Miller, J. G. (1978). Living systems. New York, NY: McGraw-Hill. Murnigham, J. K. (1985). Coalitions in d ecision-making groups: Organizational analogs. Organizational behavior and human performance, 35 (1), 1-26. Novak, K. J. (2004). Disparity and racial profiling in traffic enforcement. Police Quarterly, 7 (1), 65-96. Paternoster, R., & Iovanni, L. (1989). The labeli ng perspective and delinquency: An elaboration of the theory and an assessment of the evidence. Justice Quarterly, 6 (3), 359-394. Pfuhl, E. H. (1980). The deviance process. New York, NY: D. Van Nostrand. Radelet, L., & Carter, D. (1994). The police and the community (5th ed.). New York, NY: Macmillan. Ray, M. C., & Downs, W. R. (1986). An empiri cal test of labeling th eory using longitudinal data. Journal of Research in Crime and Delinquency, 23 (2), 169-194. Rosenfield, S. (1997). Labeling mental illness: Th e effect of received services and perceived stigma on life satisfaction. American Sociological Review, 62 (4), 660-672. Rosenhan, D. L. (1973). On being sane in insane places. Science, 179 (4070), 250-258. Sack, K. (1995, September 30). For speeder s in Florida, 20 miles of bad road. The New York Times p. 6. Scheff, T. J. (1974). The labeli ng theory of mental illness. American Sociological Review, 39 (3), 444-452. 84


Scheff, T. J. (1999). Being mentally ill: A sociological theory (3rd ed.). Hawt horne, NY: Aldine de Gruyter. Schrag, C. (1971). Crime and justice: American style Washington, D.C.: U.S. Government Printing Office. Schur, E. M. (1971). Labeling deviant behavior: It s sociological implications. New York, NY: Harper & Row. Schur, E. M. (1983). Labeling women deviant: Gender, stigma, and social control. Philadelphia, PA: Temple University Press. Sherman, L. W. (1978). Scandal and reform: Controlling police corruption. Berkeley, CA: University of California Press. Simmons, J. L. (1965). Public stereotypes of deviants. Social Problems, 13 (2), 223-232. Smith, D. A, & Paternoster, R. (1990). Formal processing and future delinquency: Deviance amplification as selection artifact. Law & Society Review, 24 (5), 1109-1132. Smith, M. R., & Petrocelli, M. (2001). Racial prof iling? A multivariate analysis of police traffic stop data. Police Quarterly, 4 (1), 4-27. Smith, P. E., & Hawkins, R. O. (1973). Victim ization, types of citizen -police contacts, and attitudes toward the police. Law & Society Review, 8 (1), 135-152. Southgate, P., & Mirrlees-Black, C. (1991) Traffic policing in changing times. Home Office Research St udy No. 124, London, UK: Her Majestys Stationary Office. Staw, B. M., Sandelands, L. E., & Dutton, J. E. (1 981). Threat rigidity effects in organizational behavior: A multilevel analysis. Administrative Science Quarterly, 26 (4), 501-524. Tannenbaum, F. (1938). Crime and the Community Boston, MA: Ginn. Thoits, P. A. (1985). Self-labeling processes in me ntal illness: The role of emotional deviance. The American Journal of Sociology, 91 (2), 221-249. Thomas, C. W., & Bishop, D. M. (1984). The e ffect of formal and informal sanctions on delinquency: A longitudinal comparison of labeling and deterrence theories. Journal of Criminal Law and Criminology, 75 (4), 1222-1245. Triplett, R. A., & Jarjoura, G. R. (1994). Theore tical and empirical speci fication of a model of informal labeling. Journal of Quantitative Criminology, 10 (3), 241-246. 85


Turner, R. H. (1972). Deviance avowal as neutralization of commitment. Social Problems, 19 (3), 308-321. Vaughan, D. (1992). Theory elaboration: The heuris tics of case analysis. In C. Ragin & H. S. Becker (Eds.) What is a case? Exploring th e foundation of social inquiry (pp. 173-202) New York, NY: Cambridge University Press. Vedder, C., & Keller, O. (1965). Th e police and middle class conflicts. Police, 9 6-8. Walter, C. E. (1995). Suburban residential traffic calming. ITE Journal, 65 (9), 44-48. Weigelt, K., & Camerer, C. (1988). Reputation and corporate strategy: A re view of recent theory and applications. Strategic Management Journal, 9 (5), 443-454. Weinstein, R. M. (1983). Labeling theory and th e attitudes of mental patients: A review. Journal of Health and Social Behavior, 24 (1), 70-84. Wellford, C. (1975). Labeling theory and criminology: An assessment. Social Problems, 22 (3), 332-345. Zimring, F. E., & Hawkins, G. J. (1973). Deterrence. Chicago, IL: University of Chicago Press. 86


BIOGRAPHICAL SKETCH Jeff Ward was born on April 19, 1983 in Takom a Park, Maryland. After graduating from Wheaton High School (MD) in 2001, he attended Rutg ers, The State University of New Jersey. In May 2005, he graduated summa cum laude fr om Rutgers and earned a B.S. degreedouble majoring in physics and psychology with certification in criminology. In the fall of 2005, Jeff enrolled in the Department of Criminology, Law & Society graduate program at the University of Florida. While at the halfway point of his Masters progra m, Jeff married Kerri Shimshock on August 4, 2006. He earned an M.A. degree from UF in 2007 and plans to pursue a Ph.D. degree at UF as well. 87