CJPS3

Available online at www.sciencedirect.com

8CI.NC. @DI".CT8

JOlJRNAL OF CRIMINAL JlJSTICE

Pergamon Journal ofCriminaJ Justice 31 (2003) 475-492

Robin Shepard Engel.

Division of Criminal Justice. University of Cincinnati, P. O. Box 210389, Cincinnati. 08 4522/-OJ89, USA

Abstract

The importance of suspects' resistance toward police officers has almost always been described in tenns of its influence on police behavior. Given the cenb'ality of citizens' resistance in the literature on police behavior, it is surprising that so little attention has been focused on explaining suspects' resistance independent of its influence on police behavior. This research examined the factors influencing multiple measures of suspects' resistance using systematic observation data collected in 1977 from twenty-four departments in three metropolitan areas. The findings show that non-White suspects were more likely to be noncompliant toward White officers but were not more likely to show more aggressive forms of resistance (e.g., verbal aggression, physical aggression, or disrespect). In addition, female suspects were more likely to be disrespectful toward offICers compared to male suspects. The relevance of these findings for future research is discussed. C> 2003 Elsevier Ltd. AU rights reserved.

Introduction

The importance of suspects' resistance and de- meanor toward police officers has typically been described in tenJJS of its influence on police behavior. Ethnographic studies of police behavior have consis- tently suggested that suspects who refuse to defer to an officers' authority or definition of the situation were more likely to be sanctioned by police through formal or informal means (Bittner, 1974; Brown, 1981; Skolnick, 1966; Westley, 1953, 1970). Like- wise, the bulk of prior quantitative research examin- ing demeanor analyzed the influence of citizens' disrespect on police behavior. That is, citizens' de- meanor was examined strictly as a predictor variable of police behavior during police-citizen encounters. With only two exceptions (Klinger, 1994; Mastrofski, Worden, & Snipes, 1995), this body of research reported that suspects who acted in disrespectful

ways were more likely to be sanctioned through ~t, citations, or the use of force by police officers (Black, 1980; Engel, Sobol, & Worden, 2000; Lund- man, 1974, 1994, 1996; Smith & Visher, 1981; Worden, 1989; Worden & Shepard, 1996).

Although research examined differences in the influence of separate measures of demeanor on police behavior, to date, little empirical work examined suspects' demeanor and resistance toward police explicitly. Given the centrality of suspects' demeanor in the literature on police behavior, it was surprising that so little attention focused on explaining suspects' displays of disrespect independent of their influence on police behavior. Therefore, it is unknown whether indicators of social stratification and the stakes for establishing and maintaining social identity during police-citizen encounters influence displays of resis- tance and disrespect by suspects toward officers.

In contrast to the scant empirical research examin- ing citizens' resistance and disrespect toward police, the literature examining citizens' attitudes toward police had a rich history. As a result of findings from the Kerner Commission in ) 967, which reported that

. Tel.: +1-513-556-5827; fax: +1-513-556-3303.

E-mail address: [email protected] (R.S. Engel).

;-.7-23S2J1)3/s - see front matter 02003 Elsevier Ltd. All rights reserved.

476 R.S. Engel/Journal of Criminal Justice 31 (2003) 475-492

been established that researchers can examine changes in the predictors of citizens' resistance over time. This study applied both sociological stratification theories and social-psychological perspectives to examine police-citizen encounters. Hypotheses regarding the influence of social stratification. propensity for ag- gression, and the stakes of maintaining social identities during police-citizen encounters over suspects' dis- plays of resistance and disrespect toward police are presented and tested with multivariate analyses.

The size and richness of the PSS data offered a unique opportunity to examine the potential predic- tors of multiple measures of suspects' resistance during police-citizen encounters that were not read- ily available and/or measured in more current data. In the PSS data set, resistance toward police was con- ceptualized and measured in four different forms: noncompliance, verbal aggression, physical aggres- sion, and disrespect. Utilizing these measures of suspects' resistance, multivariate analyses showed that the strongest and most consistent predictor of resistance was the use of alcohol or drugs by sus- pects. In addition, non-White suspects were more likely to be noncompliant toward White officers but were not more likely to display other more aggressive types of resistance toward officers (e.g., verbal ag- gression, physical aggression. or disrespect). Also, contrary to expectations, female suspects were more likely than male suspects to be classified by observers as disrespectful toward officers. The implications of these findings for future research are explored.

Defining resistance during police-citizen encounters

Klinger's (1994, 1996b) null findings regarding the influence of suspects' demeanor on police behav- ior led to his critique of the literature that, in summary, suggested researchers misconceptualized and improperly measured suspects' demeanor. He argued that physical resistance should not be consid- ered a form of disrespect, but rather that it should be considered an illegal act that is likely to influence police behavior. Worden, Shepard, and Mastrofski (1996, p. 328), however, argued that ethnographic research suggested that officers responded to extreme forms of disrespect not necessarily because they were illegal behavior, but because these actions repre- sented affronts to their authority (Bittner, 1974; Van Maanen, 1974, 1978). As such, even affronts of the greatest gravity (e.g., physical aggression) represent a form of negative demeanor toward officers. Worden et al. concluded that many diverse (and sometimes subtle) forms of citizen behavior may be character- ized by police officers as disrespectful (p. 326).

there was a national crisis in race relations and a profound and deep-seeded hostility between minority citizens and police officers (Williams & Murphy, 1990), extensive research was conducted that exam- ined the relationships between citizens' characteristics (particularly citizens' race) and their attitudes and satisfaction with police. This line of research has continued to the present day, with findings over- whelmingly suggesting that minorities, males, juve- niles, and citizens of lower social-economic status have more negative attitudes toward police compared to Whites, females, adults, and citizens of higher socioeconomic status (for a review, see Hurst & Frank, 2000; Taylor, Turner, Esbensen, & Winfree, 2001). Unfortunately, such an extensive body of knowledge does not exist for predicting citizens' actions-and more specifically, their resistance-toward police. While the empirical fmdings regarding citizens' atti- tudes remained relatively consistent across the past forty years, it is unknown if the patterns of citizens' resistance toward police have changed.

The current study attempted to fill this void by establishing baseline measures and predictors of citi- zens' resistance toward officers using data collected in 1977 for the Police Services Study (pSS}-one of the largest and most respected systematic observational studies of police officers to date. The dated nature of the PSS data may somewhat limit its generalizability to current urban policing. Certainly, policing has changed dramatically in the last twenty-five years-but have citizens' reactions toward police changed? There are some indications that the interactions between citizens and police today appear relatively consistent to those experienced in the late I 970s. For example. current observational studies reported similar types of inci- dents and problems that police encounter and consis- tency in the use of arrest, force, and other coercive actions to handle these situations (Mastrofski, Snipes, Parks, & Maxwell. 2000; Novak, Frank. Smith. &

Engel, 2002; Parks, Mastrofski, Dejong, & Gray. 1999). Likewise, as previously noted, citizens' per- ceptions and attitudes toward police have not dramat- ically changed. Larger social issues including poverty and race are as salient today as they were twenty-five years ago, as are the continuing problems between police and citizens perpetuated by social contexts (for example. see Velez, 200 I).

Nevertheless, the purpose of the present research was to explore in)portant issues that had been previ- ously ignored and to provide a contribution to what might become a cumulative body of research examin- ing issues of police-citizen interactions. Findings from data collected in the late 1970s can provide a basis for comparisons with future research utilizing data collected for more current systematic observa- tional studies of police. It is only after this baseline has

R.s. Engel / Joumal of Criminal Justice 31 (1003) 475-491 477

While researchers continue to struggle to define and measure citizens' displays of disrespect, most would agree that each of these foons of behavior on their face represents some form of resistance toward police. Resistance might then be properly conceptualized as including not only physical aggres- sion and other overtly hostile acts, but also verbal aggression (e.g., arguing with or cursing at an offi- cer), noncompliance or statements that challenge officers' authority or legitimacy (e.g., denying an officer's accusation. questioning an officer's judg- ment, or refusing to respond to an officer's questions or requests), and general fonns of disrespect (e.g., sarcasm, body language). Each of these actions represented a challenge (either passive or active) to an officers' authority or definition of the situation (see Bittner, 1974; Van Maanen. 1974, 1978). Just as each of these forms of resistance might be expected to have a varying influence over officers' actions, one might also expect that factors predicting these actions would vary. That is, particular types of persons or situations might be more likely to invoke specific forms of resistance toward officers.

perspectives rooted in social psychology (Brown & Levinson, 1987; Tedeschi & Felson, 1994). These propositions are discussed in further detail below and summarized in Table I.

Resistance as a symbol of perceptions of injustice

Theoretical framework

Scholars speculated about the relationship be- tween citizens' characteristics (e.g., race, sex, age, intoxication. etc.) and their antagonistic, disrespect- ful, and hostile behavior toward police. For example, many suggested that 000- White suspects, particularly young Black males, were more likely to be disre- spectful toward police. Indeed, the pen:eived rela- tionship between demeanor and race was used to explain higher arrest rates for young African-Amer- ican males. As Walker (1999) proposed, "to the extent that officeR stereotype young African-Ameri- can males as potential suspects, they may provoke higher rates of antagonistic behavior that. in turn. results in higher rates of arrest" (pp. 226-227).

Two general explanations are advanced to account for the perceived relationship between suspects' char- acteristics and resistance displayed toward officers. First, it is possible that particular types of citizens (e.g., young minority males) may act in disrespectful or otherwise resistant ways to symbolize their per- ceptions of injustice. This explanation is rooted in sociological stratification and psychological theories that suggest race and socioeconomic status influence actions by both citizens and officeR (Black, 1976; Tyler, 1990). Second. it is possible that dynamics involved during the actual interaction between offi-

cers and citizens lead to displays of resistance. This explanation focuses on the processes involved in police-citizen encounteR and relies on interactionist

Applying Black's (1976) theory of law to explain citizens' deviant behavior toward officers, it is hy- pothesized that increased levels of stratification in police-citizen encounters would result in more dis- plays of disrespect by both officers and citizens. Black theorized that as stratification increased, the quantity of law increased.' Although Black's theory was primarily directed toward describing the quantity of law in society, he argued that his theory of law could also explain other types of social control and deviance (p. 10). Therefore, one could argue that Black's theory of law could be used to explain citizen disrespect toward officers as a fonn of deviant behavior. Thus, this theory predicts that as stratifica- tion increases, other types of social control and deviance increase. That is, greater levels of stratifi- cation between police and citizens may lead to more deviant fonDS of behavior, including citizens' resis- tance of authority.

Sykes and Clark (1975) also suggested that nonns in police-citizen encounters were asymmetrical be- cause the status of the participants was unequal. They argued that given this disparate relationship and the associated nonns and expectations guiding police- citizen encounters, minority citizens might interpret officers' expectations as "indicating the officer's own ethnic group's superordination," while minority citi- zens' refusals to express deference might be inter- preted by officers as refusals "to acknowledge normal social obligations of all citizens and the officer's symbolic status" (p. 590).

Given this disparate relationship, it is hypothe- sized that citizens are resistant toward police officers because they do not recognize officers' legitimacy and use disrespect and/or resistance to symbolize their perceptions of injustice. In his examination of why citizens obey laws, Tyler (1990) described the normative processes associated with compliance (e.g., internalization of nonns regarding values, jus- tice, and commitment to legal authorities). He argued that citizens obey laws either because they believe the laws are legitimate or because they believe legal authorities are legitimate. Analyzing citizen surveys, Tyler reported that "both personal morality and legitimacy were found to have an effect on people's everyday behavior toward the law" (p. 161).

Tyler's (1990) normative theory of citizen compli- ance can also be utilized to explore citizens' resistance toward officers. One could argue that citizens act in

R.S. Engel/Journal of Criminal Justice 31 (2003) 475-492478

Table I Theoretical propositions, applications, and hypotheses

Theoretical propositions Applications and hypothesesTheory

Theory of law (Black, 1980) eThe theory of law can be used to explain other eResistance toward police is a fonD of fonDS of social control and deviance. deviance that can be explained using

the theory of law. eIncreased levels of stnltificalion lead to increases eGreater levels of stnltification lead to in the quantity of law. increases in citizens' resistance toward

police. .Increased levels of stratification lead 10 increases

in deviance.

Nonnative theory of citizen

compliance (Tyler. 1990)

.citizens obey the law because they believe laws and/or legal authorities are legitimate.

.Citizens are more likely to be resistant when they do not recognize the legitimacy of officers' authority.

.Minorities and juveniles are more likely to question officers' authority and therefore more likely to show resistance.

.Minority and juvenile citizens are less likely to believe that the laws and/or legal authorities are legitimate.

Interactive theory of coercive actions (Tedeschi & Felson,

1994)

eThe need to establish and/or protect social identities leads to the use of coercive actions.

.lnteractions between police and citizens often involve challenges to authority and social identities.

.citizens will engage in coercive actions toward police to protect their social identities (particularly in public places).

.citizens will engage in coercive actions toward police in response to violations of politeness norms.

.coercive actions are more likely when a person's social identity has been challenged in public.

.Violations of polite norms provoke coercive responses from persons protecting their social identities.

disrespectful or resistant ways because they do not recognize the legitimacy of officers' authority. Indeed, the police-citizen crisis of the 1960s was exacerbated in part because members of minority communities did not recognize legal authority in American society as fair or just (WilIiams & Murphy, 1990). For many minority citizens, police officers became the ultimate symbol of an unjust society dominated through White supremacy (Walker, 1999). As a result, the interactions between police and minority citizens were enveloped in an atmosphere of distrust and disrespect that has continued to the present day. Citizens who do not respect the legitimacy of officers' authority may be more likely to question that authority, and as a result are more likely to demonstrate resistance toward police officers.

gested that a person's need to establish or protect their social identity often leads to the use of coercive actions. The authors stressed the importance of im- pression management or "saving face" as a predictor of coercive actions. The social dynamics associated with "saving face" are of great concern in situations where there has been a challenge to authority, such as in police-citizen encounters. Indeed, as noted by Tedeschi and Felson, perceptions of disrespect during police-citizen encounters are a frequent source of conflict. According to these scholars, a typical reaction to perceived attacks from others involves some type of coercive behavior, particularly if their social identity has been threatened in a public setting.

While it is clear that this framework could be used to explain officers' coercive actions toward citizens, it is initially less apparent how this theory could be used to explain citizens' actions toward officers. Extending Brown and Levinson's (1987) theory of politeness, Tedeschi and Felson (1994) argued that there were numerous behaviors that could violate "politeness norms" that protect social identities. The violation of these politeness norms often pro- voked a response from persons seeking to protect

their social identities. Tedeschi and Felson described the concept of "negative face" as representing a desire to maintain autonomy and control and not be

Dynamics of police-citizen interactions

While sociological explanations of police-citizen encounters are certainly important, they do not take into account the actual processes or dynamics involved during citizen encounters with police. One theoretical framework that can be utilized to better understand the

dynamics associated with police-citizen conflict is Tedeschi and Felson's (1994) social interactionist theory of coercive actions. Tedeschi and Felson sug-

R.s. Engel/Journal of Criminal Justice 31 (2003) 475-492 479

interfered with by others. Behavior that threatens a person's autonomy (e.g., orders, threats, and wam- ings) is likely to produce some fonn of resistance.

Applying the concepts of "politeness norms" and "negative face" to police-citizen encounters, offi- cers' attempts to control citizens could be construed by citizens as attempts to restrict their freedom and thus may lead to displays of disrespect. This is especially true in situations where citizens do not recognize officers' legitimacy. For example, Tedeschi and Felson (1994) described situations involving interactive justice or confonnity to nonns regarding demeanor, politeness, and respect toward others (p. 219). Tedeschi and Felson suggested that citizens might act in disrespectful ways because they perceive officers to be impolite or disrespectful toward them. Although officers hold the power in these relation- ships, citizens may act in disrespectful ways, often with blatant disregard for the consequences. Thus, citizens might be more likely to challenge officers' authority if they felt their identities were being challenged. their freedoms restricted, or politeness nonns violated-particularly if the encounter took place in front of others.

Prior research

to exhibit the sort of recalcitrant demeanor which police construed as a sign of the confinned delin- quent" (Piliavin & Briar, 1964, p. 212, emphasis added). In contnl,st, Black and Reiss (1970) reported a slightly lower percentage of Black juveniles (15.4 percent) who were antagonistic to police compared toward White juveni1es (17.7 pert:ent) in an observa- tional study of three police departments in the late 1960s. A replication study utilizing systematic obser- vation of police-citizen encounters in a midwestem city. however. showed that the percentage of antag- onistic encounters between police and Black juve- niles was double that of White juveniles-30 percent versus 15 percent. respectively (Lundman et a1., 1978).2 Research utilizing this same data found that in 87.9 percent of encounters with White citizens, complete deference was shown to police compared to 80.9 percent of encounters with non-White citizens (Sykes & Clark, 1975). Furtbennore, these research- ers reported that as the citizens' status declined (measured as "a function of role and seriousness of violation as well as of class and ethnicity"), the leve1 of deference shown to officers declined.

These early research findings paint dramatically different pictures based on widely varying rates of disrespect shown to police by White and non-White suspects. Since studies only reported bivariate asso- ciations, it is unknown whether suspects' displays of disrespect are related to other characteristics after controlling for other relevant factors. In more recent mu1tivariate analyses, Greenleaf and Lanza-Kaduce (1995) applied Turk's (1969) structural conflict the- ory to examine conflicts between police and citizens during domestic disputes. They reported that as the amount of "organization" increased or as the amount of "sophistication" decreased, the likelihood of norm resistance by both citizens and officers significantly increased. Since the variables of interest (i.e., orga- nization and sophistication) were measured as addi- tive indices, however, the direct influence of citizen and situational characteristics was not determined.)

More recently, Lanza-Kaduce and Greenleaf (2000, p. 222) have extended Turk's (1969) theory of norm resistance to incorporate deference reversals where "suspects refuse to acknowledge the officer's position and instead attend to age, race, or gender norms of deference." In an examination of police encounters involving domestic disputes, they de- scribed sixteen combinations of suspects' and offi- cers' age and race that might structure police-citizen interactions. These combinations were collapsed into a three-category deference variable4 and were used to predict norm resistance. Lanza-Kaduce and GreenJeaf (pp. 230- 231) found that the deference variable was a significant predictor of norm resistance and con- cluded that age and race deference patterns that were

Although the importance of police-citizen rela- tions is readily apparent, few studies empirically examined this relationship at its most basic level: police-citizen encounters where police act as street- level bureaucrats (Lipsky, 1980). An abundance of research examined the effects of disrespect over police behavior, yet few studies examined the influ- ence of disrespect and resistance toward police. Much of the research available only reported bivariate relationships between demeanor and suspects' race (Black & Reiss, 1970; Lundman, Sykes, & Clark, 1978; Piliavin & Briar, 1964; Sykes & Clark, 1975). The few multivariate studies that were conducted combined measures of citizen and officer behavior (Greenleaf & Lanza-Kaduce, 1995; Lanza-Kaduce & Greenleaf, 2000) or focused on citizen compliance rather than resistance (Mastrofski, Snipes, & Supina, 1996; McCluskey, Masb'ofski, & Parks, 1999).

Piliavin and Briar's (1964) classic research reported that the differences officers perceived be- tween "serious delinquents" and «good boys" were based on the juveniles' demeanor and race. These scholars concluded that "the tendency for police to give more severe dispositions to Negroes and to youths whose appearance corresponded to that which

police associated with delinquents partly reflected the fact, observed in this study, that these youths also were much more likely than were other types of boys

.J

R.S. Engel! JOUmQI of Criminal Justice 31 (2003) 475-491480

(1990) theory of citizen compliance, Tedeschi and Felson's (1994) theory of violence, and prior resean:h examining citizens' behaviors toward police, three general hypodteses are advanced. First, it is specu- lated that differences in social stratification between police and suspects will influence suspects' reactions toward officers. possibly due to their varying percep- tions of officers' legitimacy. For example, particular types of suspects (e.g., non-White, juveniles) are believed to be more likely to question the legitimacy of officers and therefore more likely to act in a resistant manner. Second, it is speculated that the

stakes involved in maintaining social identity during ~unters with police will lead suspects to question officers' legitimacy and result in displays of resis- tance. Finally, it is argued that suspects or situations that involve a propensity for aggressiveness (e.g., male suspects, intoxicated suspects, situations involv- ing conflict between citizens) are more likely to result

in resistance displayed toward officers.

counter to traditional authority relationships resulted in more resistance. Their measure of the dependent variable (norm resistance), however, included resis- tance by both citizens and officers.s Therefore, the influence of officers' age and race interactions over citizens' behavior is unknown.

Mastrofski et al. (1996) examined the factors influencing citizen compliance to officers' requests through systematic social observation of police offi- cers collected in Richmond, Virginia. 6 They reported

that officers were more likely to gain compliance from male citizens and during encounters in public locations, but less likely to gain compliance from poor citizens, citizens known to police, and encoun- ters where additional officers were present. This study also found that the influence of the citizens' race on the likelihood of compliance was contingent upon the officers' race. Mastrofski et al. reported that White officers encountering minority citizens were significantly more likely to gain compliance, while minority officers encountering White citizens were significantly less likely. The finding that minority citizens were more likely to show compliance to White officers is counter to past research findings (Lundman et al., 1978; Piliavin & Briar, 1964; Sykes & Clark, 1975) and the perspectives of many police officers. Although dated, Piliavin & Briar's (1964) observation that "Negro boys were much more likely than non-Negroes to "give us [officers] a hard time," be uncooperative, and show no remorse for their b'aDsgressions" (pp. 212-213) is an assumption still widely held today by police officers and researchers.

A recent replication of the research ofMastrofski et al. (1996) was conducted by McCluskey et al. (1999) utilizing systematic observational data of officers collected in Indianapolis, Indiana and St. Petersburg, Florida. They reported that few of the original findings held in the subsequent analyses. For example, the only citizen characteristics that significantly predicted com- pliance in the replication study were age and an additive scale measuring "irrational elements," in- cluding intoxication, mental illness, and strong emo- tions (fear, anger, or depression). Both juveniles and citizens with more irrational elements were signifi- cantly less likely to comply with officers' requests. The original finding of Mastrofski et al. regarding the significance of officers' and citizens' race interactions was not supported in the replication study.

Perceptions of officers' legitimacy

It is argued that suspec~ who are non-White or juvenile will be significantly more likely to question offi~' authority and therefore more likely to dis- play fonns of resistance toward officers. Non-White suspec~ are thought to be more likely to question officers' authority based on their perceptions of discrimination. Resean:h showed rather consistently that minorities had more negative attitudes toward the police and were less likely to perceive police officers as legitimate sources of authority (Albrecht & Green,

1977; Bayley & Mendelsohn, 1969; Browning, Cullen, CaD, Kopache, & Stevenson, 1994; Decker, 1981; Dunham & Alpert. 1988; Jacob, 1971; Sca- glion & Condon, 1980; Taylor et al., 200 I; Tucb &

Weitzer, 1997; Webb & Marshall, 1995; Weitzer, 2000; Weitzer & Tuch, 1999). Therefore, it is expected that non-White suspec~ are more likely to display forms of disrespect toward police as com- pared to White suspects. 7

It is also expected that juvenile suspects are less likely to accept police legitimacy and authority. Re- search showed that juveniles had more negative atti- tudes toward police than adults (Hurst & Frank, 2000; Jesilow, Meyers & Namazzi, 1995; Taylor et al., 200 I). In addition, Muir (1977) reported that most police encounters with juveniles involved an element of irrationality. He described this paradox of in'atio- nality as "the more delirious the threatener, the more serious the threat" to police officers, and defined juveniles as irrational because they "simply foresaw

their actions' having consequences different from those which adults anticipated" (p. 127). Given that juveniles are less likely to internalize norms regarding

Hypotheses

Given the ambiguity surrounding the findings reviewed above, it is unclear which factors are likely to influence suspects' resistance toward police. Based

on Black's (1976) theory of social control, Tyler's

R.S. Engel / JoUnlQI of Criminal JlLflice 31 (200J) 475-492 481

Propensity for aggression

Some displays of resistance and disrespect may be based on chamcteristics of the suspect or situation that led to more aggressive forms of behavior. Muir (1977) described situations that involved critical incidents where citizens rebelled against officers' authority. For example, intoxicated citizens were viewed as less rational and more detached that sober citizens interacting with police. Likewise, intoxicated suspects were believed to show more disrespect toward officers based on numerous studies that iden- tified a link between alcohol and/or drug consump- tion and aggressive behaviors (for a review, see Fagan, 1990).

It is also expected that male suspects are more resistant toward officers based on a large body of literature that suggested males are more likely than females to engage in violence and other displays of negativity (for a review, see Tedeschi & Felson, 1994) and are more likely to have negative attitudes toward police (Jesilow et al., 1995). Finally, it is expected that police-citizen encounters that involve conflicts between citizens when the police arrive will result in more displays of resistance. Citizens may be more likely to respond negatively to officers' displays of authority during emotionally charged conflict sit- uations.

values, justice, and commitment to legal authorities, it is expected that juvenile suspects will be more likely to question the legitimacy of officers and therefore more likely to display fonns of resistance and disrespect.

Despite its intuitive appeal, few studies thorough- ly exan1ined the potential influence of officers' char- acteristics on suspects' actions. It is speculated that non-White, female, or younger officers are more likely to encounter resistant suspects because sus- pects perceive these officers as less legitimate. Sus- pects might be more inclined to maintain autonomy and control when the person attempting to restrict their freedom is of a pen;eived lower social status. Likewise, it is argued that interactions between offi- cers' race and suspects' race will significantly influ- ence the behavior of suspects toward officers. Specifically, it is speculated that suspects involved in encounters with police officers of a different race are more likely to display fonns of resistance.

In addition, police entry-whether the encounter was reactive (dispatched or otherwise citizen initiat- ed) or proactive (initiated by the officer)-is believed to influence citizens' reactions toward offICers. Tyl- er's (1990) research found that citizens were more concerned with issues regarding procedural justice when they were proactively stopped by police com- pared to when police were summoned to the scene (see also Tyler & Folger, 1980). Therefore, it is expected that suspects are more likely to question officers' authority in police-initiated encounters and are therefore more likely to react in a disrespectful manner. Likewise, it is hypothesized that suspects who are involved in less serious events leading up to an encounter with police are more likely to show resistance to officers. Officers' use of their authority and citizens' perceptions of officers' legitimacy to use that authority are less likely to be questioned in situations involving serious offenses. In situations involving less serious offenses, however, officers' legitimacy may be questioned, and as a result sus- pects may be more likely to respond negatively.1I

Data

Social identities

As previously noted, Tedeschi and Felson's ( 1994) theory indicated that people were more likely to engage in coercive actions if their social identity was challenged, particularly if the challenge occurred in front of others. Tedeschi and Felson stressed the importance of the social dynamics associated with protective self-presentation when there was a chal- lenge to authority in police-citizen encounters. Therefore, one might expect that suspects are more likely to resist officers if the encounter is in a public location or in the presence of others (bystanders and other officers).

The purpose of the cunent study was to examine the above hypotheses using data from a large-scale multisite field study of police behavior-the PSS conducted in 1977. Within this data set, suspect resistance was measured in multiple ways, including measures of suspects' actions toward officers that could be considered resistant (e.g., noncompliance, verbal aggression, and physical aggression) and observers' characterizations of suspects as disrespect- ful. The PSS included twenty-four police departments in three metropolitan areas (Rochester, NY; St Louis, MO; and Tampa/St Petersburg, FL). These depart- ments ranged in size from thirteen sworn officers to over 2,000 and served municipalities with popula- tions that ranged from 6,000 to 499,000. Although not a random sample, the PSS represented a rough cross section of police organizations and service conditions for urban policing in the United States. Compared with other studies of police, the PSS provided a considerably broader base from which to draw generalizations about American policing.

The core methodology of PSS was systematic social observation of patrol officers in the field, along with interviews of citizens and officers focused on sixty neighborhoods that were selected with explicit

R.S. Engel/Journal of Criminal Justice 31 (2003) 475-492482

reference to the race and income of residents. The sample of neighborhoods thus represented a rough cross section of residential service conditions for each department. On a matched sample of fifteen shifts in each neighborhood, trained observers accompanied officers during entire work shifts and unobtrusively took brief field notes about police-citizen encounters. Based on these field notes, observers coded data items about the police-citizen encounters, including the characteristics and behavior of citizens and officers.

Information on 5,688 police-citizen encounters during over 900 shifts was coded on a standardized form (see Caldwell, 1978). Encounters were defined as any event in which there was face-to-face com- munication between a police officer and a member of the public; the communication normally involved verbal exchanges but could have involved only the use of force by police. To address the hypotheses described above, the present work examined a subset of the PSS observational data: 1,461 nontraffic sus- pects encountered by police officers. Citizens were coded as suspects if at the end of the encounter with police officers, they were considered "suspect in a criminal matter or peace disturber" or the "person complained about in a civil matter."

Table 2 Measures of suspects' resistance toward police officers ~~6l suspects) Variables Minimum Maximum Mean S.D.

Suspects' coded actions: Noncompliant 0 Verbally resistant 0 Physically resistant 0 Any disrespectful 0

action

0.09 0.10 0.01 0.14

0.28 0.30 0.12 0.35

0.22 Suspects characterized as disrespectful by observers: At beginning of 0 I 0.05

encounter During encounter 0 I 0.08 At end of encounter 0 I 0.05 Any time during 0 I 0.09

encounter

0.27 0.21 0.29

percent of suspects in this sample. Combined, 14.2 percent of citizens displayed one or more of these types of resistance.

The second general measure of suspects' resis- tance was captured as observers' global assessment of suspects' negative demeanor during police-citizen encounters. Observers were asked to describe sus- pects' demeanor toward officers utilizing a list of "demeanor codes," including one code that was defined as "sarcastic, disrespectful, hostile." The characterization of demeanor by observers was cap- tured at three points in time: when officers arrived on the scene, during the encounter, and when officers were leaving. Previous analyses exploring the effects of suspects' demeanor on officers' behavior used disrespect at the beginning of the encounter (when the officer arrived on the scene) to ensure the proper temporal ordering of the variables (Worden, 1989; Worden & Pollitz, 1984). As noted by Worden and Shepard (1996), this was a conservative approach whereby some suspects who were disrespectful be- fore an arrest might be coded as civil. A less conservative approach would be to measure disre- spect as those characterized as disrespectful at the beginning or during the encounter, with the risk of confounding the effect of suspects' demeanor on police behavior. This was a more salient issue for analyses examining the influence of disrespect on police behavior. For the present analyses, disrespect measured at any time during the encounter will be examined. As shown in Table 2, of the 1,461 suspects examined, 4.9 percent were characterized as disre- spectful at the beginning of the encounter, 7.7 percent during the encounter, and 4.7 percent at the end of the encounter. Combined, 9.2 percent of suspects were coded by observers as "sarcastic, disrespectful, or hostile" at some time during the encounter.

Methods

Measures of resistance

In the PSS data, suspect resistance was measured in two fundamentally distinct ways: (I) as coded actions that might be considered resistant and (2) as observers' characterizations of disrespect. The for- mer, coded actions, represented suspects' behaviors that might be construed as resistant to officers' authority. These actions included noncompliance, verbal aggression, and physical aggression (see Table 2). Observers accompanying officers were prompted on a systematic coding form to indicate if the citizen had engaged in these actions during the police- citizen encounter. Suspects' actions that were consid- ered noncompliant included passive resistance to officers' authority by "refusing to answer questions" or "refusing to otherwise cooperate with officers' requests." In this sample, 9.2 percent of suspects showed noncompliance toward officers. A second coded action, verbal aggression, included more active forms of noncompliance, including "arguing with or cursing at an officer." Verbal aggression was shown by 9.7 percent of suspects. The final form of coded actions included physical aggression, such as "phys- ically fighting with an officer" and "using or attempting to use a weapon against the officer." Physical aggression was only observed from 1.4

R.S. Engel / Joumal of Criminal JIIStic. Jl (100J) 475-492 483

Table 3 CrosstabuIabons of suspect ac1imlS and observa- cbar8Ctcr- iDtions of d~

Characterized by obscrvcn

aadisrespectful

No Yes

Noncompliant (n-13S) 49.6 (67) Verbally resistant (n-141) 3S.S (SO) Physically resistant (n=20) IS.O (3) Any disrespectful action SO.2 (104)

(n=207)

The numbers represent percentages and. in parentheses, the actualvaJues.

physically resistant manner; these actions alone may not constitute "disrespect. "

Clearly, die strength of this argument rests finnly on one's definition of disrespect. Van Maanen (1978) suggested that officers' classifications of disrespect were based on their pen:eptions of a citizen's knowl- edge of their misbehavior and their ability to act differently under the circumstances. He described "die asshole" as someone "who is viewed [by die police] as culpable and blameworthy for his affront- ing action" (p. 231, emphasis added). Therefore, acting in a noncompliant, verbally aggressive, or physically aggressive manner may not necessarily result in an offICer's interpretation of die citizen as

disrespectful. Unfortunately, it was not clear how PSS observers

defined disrespect. They were prompted to select one demeanor code to describe the suspect at three separate points during the encounter and were given only minimal instructions.9 This demeanor measure tapped into what might be considered "typical" interpretations of disrespect and was conditioned by cultural nonns in American society. Yet, given d1at officers were in a position of authority, their expect- ations for citizen behavior might differ somewhat

Table 4 Independent variables (n= 1461 suspects)

Variables Minimum Maximwn MQD S.D.

0.S2 O.SO 0.87 0.34 0.49 0.50

..

I

0.38 0.49

0.03 0.17

0.10 0.30

Cross-tabulations comparing the two general forms of resistance showed significant differences in the interpretation of what constituted resistance toward officers (see Table 3). For example, 23.1 percent of suspects characterized as disrespectful by observers did not display any noncompliance, verbal aggression. or physical aggression toward officers. It was likely that observers coded suspects as disre- spectful based on interpretations of attitudes, body language, and other forms of nonverbal cues or other types of disrespectful actions that were not captured on the standardized coding form.

These measures differed in other significant ways. As shown in Table 3, not all of the suspects who displayed particular actions that might be considered forms of resistance were coded as disrespectful by observers. For example, nearly half of all suspects who were noncompliant with officers' requests, over 35 pen:ent of suspects who verbally argued with or cursed at an officer, and 15 percent of suspects who physically resisted an officer were not considered by observers to be disrespectful. It was clear that as the suspects' actions became more severe, observers were more likely to code them as disrespectful. Nevertheless, there were significant discrepancies between the measures.

One interpretation of these discrepancies is that the observers' characterizations of suspects' demean- or are of questionable validity (see Worden et al., 1996). A second interpretation, however, is that coded actions and observers' global perceptions of disrespect are simply measuring different phenome- na. Perhaps some suspects who act in a passive, noncompliant manner-by refusing to answer ques- tions or refusing to cooperate with officers' requests-are not simultaneously being disrespectful. That is, it may be possible to refuse to cooperate with officers in a respectful manner. For example, suspects may respectfully but firmly assert their perceived rights during encounters with police. The same argu- ment applies to suspects acting in a verbally or

4 53 4 I

1.98 30.47 0.81 0.41

0.66 0.47 3.56 6.34 0.91 1.46

1 30 10

Protection of social identities Public location 0 Number of bystanders 0 Number of other 0

officers present

Propensity for aggression Suspect male 0 Alcohol/drug use 0 Suspect fight with 0

other citizen

f(68) :(91) 1(17) ; (103)

0.73 0.20 0.88 0.49

0.79 0.41 0.23 0.42 0.02 0.13

484 R.S. Engel/Journal of Criminal Juftice 31 (2003) 475-492

from observers' expectations. These differences in expectations might generate differences in officers' and observers' characterizations of citizens as disre- spectful; however, this proposition could not be tested with this data set.

None of the measures of suspects' resistance in the PSS data were without flaws. Each measure was based on observers' interpretations of police-citizen encounters rather than officers' interpretations. Nev- ertheless, these measures-though imperfect-were representative of the measures used throughout the literature to explain police decision making. The strength of these measures was that each captured a slightly different form and level of resistance toward police. Therefore, the following analyses explore each measure of resistance independently in an effort to better explain suspects' behavior toward police.

average age of officers encountering suspects was 30.5 years old. A set of dichotomous interaction terms simultaneously captured the race of the officer and the suspect. Specifically, four variables were created that measured the following relationships: White officer/White suspect, White officer/non- White suspect, non-White officer/White suspect, and non-White officer/non-White suspect. A dichot- omous variable captured whether the encounter was proactive (initiated by the officer) or reactive (dis- patched or otheJWise citizen initiated). Following Klinger (1994), a five-point ordinal scale was used to measure the seriousness of "preintervention crime," where 0 = no crime, I = minor property crime, 2 = minor violent or major property crime, 3 = moderate violent crime, and 4 = major violent crime. I I

Table 4 also describes the variables believed to be associated with social identities and the need to "save face" during police-citizen encounters. A dichoto- mous variable captured whether the location of the encounter was public or private. In addition, the number of bystanders (those citizens who were pres- ent but did not participate in the encounter) and the number of additional officers present were measured as continuous variables.

Independent variables

Variables hypothesized to relate to perceptions of officers' legitimacy are displayed in Table 4. The majority of suspects were White (52.3 percent) and adults nineteen years and older (73.5 percent), while the majority of officers encountering suspects were White (87.1 percent) and male (98.2 percent).IO The

Table 5 Crosstabulations of ~~ect characteristics and measures of ~~t

Verbally resistant

(n= 141) Noncompliant (n= 135)

Physically resistant (n = 20)

Observer characterize as

disrespectful (n= 134) Suspect male Suspect female

9.0 (104) 10.0 (31)

10.3 (118) 7.4 (23)

1.2 (14) 1.9 (6)

8.6 (99) 11.3 (35)

Suspect White Suspect non-White

6.5 (SO). . .

12.2 (85) 7.6 (58)..

11.9(83) 1.2 (9) 1.6 (II)

7.1(54).. 11.5 (80)

Officer White Officer non-White

t.'); (117) ~,., (i8)

9.4 (120) 11.1 (21)

1.4 (18) 1.1 (2)

8.7(111) 12.2 (23)

Officer male Officer female

9.4 (135) 0.0 (0)

9.8 (140) 3.7 (I)

1.4 (20) 0.0 (0)

9.3 (13j) 3.7 {I)

Police initiated Citizen initiated

7.2 (43)* 10.6 (92)

7.7 (46)- 11.0 (95)

1.3 (8) 1.4 (12)

10.1(87) 7.9 (47)

Public location Private location

8.7 (84) 10.4 (51)

9.S (92) 10.0 (49)

1.2 (12) 1.6 (8)

9.4(91) 8.8 (43)

Suspect juvenile Suspect adult

8.5 (33)

9:S(1.~2)

6.5 (25). . 10.8(1.16)

.3{S)

:~ (IS) 7.5 (29) 9.8 (105)

Not intoxicated AlcohoVdrug use

6.3 (71)... 18.7 (64)

7.0 (78)* * *

18.4 (63) 2.6 (9). 1.0 (II)

6.5 (73)- - - 17.8 (61)

The numbers represent percentages and, in parentheses, the actual values. . P<.05. .. P< .01. ... P<.OOI.

R.S. Engel/Journal of Criminal Justice 3 I (2003) 475-492 485

Variables hypothesized to influence resistance based on their aggressive nature (e.g., suspect sex, intoxication, and conflict with other citizens) are also described in Table 4. The majority of suspects were male (78.7 percent), while nearly one fourth were under the influence of drugs or alcohol (23.4 percent). Conflicts during encounters with police were mea- sured with a dichotomous variable that captured whether or not the suspect physically fought with other citizens in the officer's presence.

Findings

Bivariate

Table 5 reports the bivariate relationships between selected independent variables and different types of resistance displayed toward police. Bivariate relation- ships show that non-White suspects are significantly more likely to show all types of resistance with the exception of physical aggression. Contrary to expect-

Perceptions of legitimacy White suspect - 0.63* *

(0.S3) 0.17 (1.18)

-0.33

(0.72) -0.08

(0.92)

0.68

(1.97) -0.17

(0.8S)

White officer

0.67. . (1.96) 0.22

(1.25) 0.42

(1.51) -0.09

(0.91) -0.00

(1.00) 0.17

(1.18) -0.15

(0.86)

0.36 (1.44) 0.37

(1.45) 0.37

(1.44) 0.21

(1.23) -0.00

(1.00) 0.01

(1.01) -0.13

(0.87)

-0.77

(0.46) -4.89

(0.01) -0.32

(0.73) 0.20

(1.22) 0.13- -

(1.14) - 0.27

(0.76) 0.95

(2.59)

White officer! non-White suspect

Non-White officer! White suspect

Non-White officer! non-White suspect

Age of suspect -0.09

(0.91) -0.00

(1.00) 0.17

(1.18) -0.16

(0.85)

0.21 (1.23)

-0.00

(1.00) 0.01

(1.01) -0.14

(0.87)

0.23 (1.25)0.13. . (1.13)

-0.27 (0.76) 0.95 (2.58)

Age of officer

Seriousness

Police initiate

Protection of social identity Public location -0.24

(0.79) 0.01

(1.01) 0;30.. . (1.34)

-0.24

(0.79) 0.01

(1.01) 0.30.. . (1.34)

-0.

(0. O.

(I.' O.

(I.

-0.08

(0.93) 0.04

(1.04) 0.30** *

(1.35)

-0.59

(0.56)0.10.. . (1.11) 0.55.. . (1.73)

-0.63

(0.53) 0.10-. .

(1.11) O.SS.. . (1.73)

Number of bystanders

Number of officers

Prone to aggression Male suspect -0.22

(0.80)

1.17... (3.21)

~O.76 (2.13)

118.8... .17

-0.22

(0.80)

1.17...

(3.22) 0.77

(2.16) 119.2.. .

.17

0.34 (1.41) 0.93.. .

(2.54) 0.56

(J.74) JI8.9*. .

.17

0.34 (1.40) 0.94** *

(2.55) 0.56

(1.76) 119.3** *

.17

-0.84

(0.43) 1.27.

(3.56) 1.31

(3.70) 88.9.. .

.44

-0.85

(0.43) 1.26.

(3.51) I.2S

(3.49)89.S.. . .44

AlcohoVdrug use

Suspect fight with citizen

Model chi-square Nagelkerke K

Entries are unstandardized coefficients and, in parentJIeses, log odds. * P<.OS.

08

93) 04

04) 30... 35)

.. P<.OI.

... P<.OOI.

R.S. Engel/Journal of Criminal Justice 3 J (2003) 475-492486

ations, juveniles are not more likely than adult suspects to be resistant toward police-in fact, adults are significantly more likely to engage in verbal aggression. Somewhat surprisingly, male suspects are not significantly more likely than female suspects to display any type of resistance. Consistent with the hypotheses, suspects under the influence of drug or alcohol are significantly more likely to engage in all fortnS of resistance.

Table 7 Logistic regression of observers' characterizations of suspects' disrespect (11= 1,461) Variable Disrespect-

Constant

Multivariate

0.4S. (I.S6)

0.44

(1.56)

0.84

(2.31)

1,14

(3.14)

0.60

(1.82)

0.63

(1.88)

-0.07

(0.93) 0.00

(1.00) -0.13

(0.87) -0.03

(0.96)

-0.

(0. -0.

(0. -0,

(0. -0,

(1.

Protection of social identity Public location 0.06

(1.06) Number of 0.03 -

bystanders (1.03) Number of 0.32-- -

officers (1.38)

0.06

(1.06) 0.03.

(1.03) 0.33...

(1.38)

0.31 (1.36)

-0.01 (0.99) 0.27.. . (1.31)

Prone to aggression Male suspect - 0.47.

(0.63)1.10.. . (2.99)1.37. . (3.95)

- 0.48. (0.62) 1.10*..

(3.02) 1.38. .

(4.01)

-0

(0 0

(2 I

(5

To further examine these relationships, several logistic regression equations exploring the predictors of different fonns of suspects' resistance are dis- played in Tables 5 and 6.12 In Table 6, the regression coefficients, log odds, and significance levels of characteristics that predict coded actions of resistance are reported. For each dependent variable, two mod- els are estimated: (1) the base model and (2) the base model with officer/race interaction tenns included.

Contrary to expectations, the majority of in de pen- dent variables did not have a significant influence across all fonDS of resistance. The results displayed in Table 6 suggest that the only consistent predictors of all suspects' resistant actions toward police (i.e., noncompliance, verbal resistance, and physical resis- tance) are suspects' use of alcohol or drugs and the number of officers present. Specifically, the odds of noncompliance, verbal resistance, and physical resis- tance toward police are 3.2, 2.5, and 3.6 times more likely when the suspect is under the influence of drugs or alcohol, respectively. The significant find- ings for the number of officers present must be interpreted carefully because additional officers may have been called to the scene to provide assistance after suspects initially resisted police. That is, tem- poral ordering with this variable cannot be measured, and therefore the predictive validity of the presence of officers over suspect resistance is unknown.

Slightly different patterns emerge that explain suspects' resistance when different types of actions are examined-the most important of these differ- ences is that non-White suspects are significantly more likely to show resistance through noncompli- ance, but not through verbal or physical aggression. When the suspect/officer race interaction tenDS are entered into the equations, it becomes clear that only non-White suspects encountered by White officers are significantly more likely to be noncompliant. The race interactions terms, along with the main race effects, are nonsignificant for more active forms of resistance (e.g., verbal and physical resistance).

Slightly different patterns emerge when the char- acterization of suspects as disrespectful is examined. Table 7 displays the regression coefficients, log odds,

AlcohoVdrog use

Suspect fight wi citizen

Suspect actions Noncompliant 1.64** *

(5.14) 3.42** *

(30.69) 2.39* *

(10.92) 412.S*. .

0.54

Verbal resistance

Physical resistance

Model chi-square Nagelkerke K

113.2*** 0.16

114.4.. . 0.16

Entries are unstandardized coefficients and, in parentheses, log odds.

* P< .05... P<.OI.... P<.OOI.

34

71) 01

99) 25

78) 04

04)

.99

.37 '.78 .17 .69 .39

.. )... ) .

)

R.S. Engel/Journal of Criminal Justice J 1 (100J) 475-491 487

and significance levels for variables predicting observers' characterizations of suspects as disrespect- ful. The results include three different models: (I) the base model, (2) the base model with officer/race interaction tenns included, and (3) the base model with race interaction terms and citizens' actions included. The third model (including citizens' actions) attempts to examine observers' character- izations of disrespect controlling for the actions of the suspect. As with suspects' resistant behaviors, intoxication and the presence of additional officers significantly influences disrespect. Unlike the previ- ous models, female suspects and suspects who are in conflict with other citizens are significantly more likely to be characterized as disrespectful by observ- ers. Even after controlling for suspects' actions, suspects who are female, under the influence of alcohol or drugs, or involved in a physical fight with other citizens are significantly more likely to be characterized as disrespectful toward officers. As with verbal and physical aggression, the suspect/ officer race interaction terms are not significant predictors of suspects' disrespect as characterized by the observer.

To control for the possibility that citizens were reacting to officers' punishment/control through dis- plays of resistance, the actions taken by officers toward citizens (including arrest, use of force, and displays of disrespect) were captured as dichotomous variables and included in separate models estimating suspects' resistance and disrespect. If citizens' resis- tance was determined by officers' actions, then one would expect that the coefficients for the other variables in the model would have been substantially reduced. The inclusion of these variables, however, did not alter the main findings reported. Although the models including officer behavior explained more variance, the coefficients for the other independent variables remained relatively stable. These results are not displayed in tabular form but are available from the author upon request.

Discussion

variables related to the protection of social identities (e.g., the presence of bystanders, location of the encounter, etc.) and measures of suspects' propensity for aggression (e.g., gender, intoxication) could also predict suspect behavior. As reported above, howev- er, with the exception of suspects' propensity for aggression, most of these relationships withstood

empirical scrutiny. The proxy measures of citizens' perceptions of

officers' legitimacy have a mixed influence over suspects' displays of resistance toward officers. Of particular importance is the fmding that non-White suspects are significantly more likely to display noncompliance (i.e., passively resisting officers' au- thority by refusing to answer questions or otherwise cooperate) but not other more aggressive forms of resistance toward officers' authority. More refmed analyses of the officer/suspect race relationship show that the odds of a non-White suspect acting in a noncompliant manner toward White officers are 2.0 times greater compared to White suspects. Perhaps non-White suspects are more likely to passively challenge White officers' authority simply by assert- ing their constitutional rights in situations that Tede- schi and Felson (1994) have described as involving interactive justice. For example, non-White suspects might be more likely to passively resist White offi- cers' authority by refusing to answer questions or refusing to otherwise cooperate with officers' requests because they perceive (correctly or incor- rectly) that police are unfairly targeting them and that they have the right not to respond. Yet, when more serious forms of resistance are examined-verbal aggression. physical aggression, and characterizations of disrespect-it is clear that non-White suspects are not significantly more likely to engage in such activities compared to White suspects. Other varia- bles thought to reflect perceptions of officers' legit- imacy did not have a significant influence over suspects' resistance toward officers. Contrary to expectations, juvenile suspects are not more likely to resist officers' authority. Likewise, the seriousness of the offense and police initiation of the encounter are not significant predictors of resistance.

The fmding that non-White suspects were signif- icantly more likely to be noncompliant toward offi- cers contradicted findings reported by Mastrofski et al. (1996). Specifically, Mastrofski et al. ~ that non-White suspects were significantly more likely to comply with requests issued by White officers. There are several plausible explanations for the discrepancy in findings. First, the measure of compliance used in the study of Mastrofski et al. differed significantly from the measure of noncompliance in the CUn'ent study. 1:1 Second, Mastrofski et al. examined the

compliance of all citizens (e.g., victims, wiblesses,

Applying Black's (1976) dteory of social control, Tyler's (1990) theory of citizen compliance, and Tedeschi and Felson's (1994) theory of coercive actions, it was hypodtesized that particular types of suspects, situations, and officers would lead to greater displays of resistance during police-citizen encoun- ters. Specifically, it was hypothesized that proxy measures of citizen's pe~eptions of legitimacy (i.e., SUSpect and officer race, age, etc.) could predict resistance and disrespect displayed by suspects to- ward police. Likewise, it was hypothesized that

488 R.S. Engel/Journal of Criminal Justice 31 (2003) 475-492

and suspects) whom the police instructed in some way, while the current study examined noncompli- ance of only those considered suspects. Thus, both the sample and the phenomena that these studies sought to explain were different. A final explanation for the discrepancy in the findings may be that citizens' responses to police have changed signifi- cantly over time. The Richmond data set utilized by Masb"ofski et al. was collected in 1990, while PSS data were gathered in 1977. Note, however, that a replication effort with data collected from 1996 to 1997 in two additional cities also did not reproduce the core finding ofMasb"ofski et al (McCluskey et al., 1999). Perhaps, Richmond's commitment to problem solving and community policing policies represented a unique and more innovative form of policing, or police-minority relations in Richmond were better than those in other American cities. 14

The current findings also show that variables measuring the protection of social identities do not have consistent influences over most forms of resis- tance. There is limited support for the proposition that suspects are more resistant when the number of bystanders increases; however, this relationship only applies to displays of physical aggression. The nature of the location (public or private) does not influence displays of resistance by suspects. Finally, the pres- ence of other officers was consistently and signifi- cantly related to all forms of resistance displayed by suspects; but due to the issues of temporal ordering previously noted, the predictive validity of this var- iable was unknown.

Measures of suspects' propensity for aggression (i.e., suspects' gender and intoxication) both have a significant influence over resistance toward police. The strongest and most consistent predictor of suspects' resistance is the use of drugs and/or alcohol. As hypothesized, suspects' use of drugs or alcohol increases the likelihood of all forms of resistance-noncompliance, verbal aggression, phys- ical aggression, and disrespect. This is consistent with Muir's (1977) observation that intoxicated citizens were more likely to be irrational and detached during their interactions with police. In- toxicated suspects were less aware of the conse- quences of their actions and were more likely to rebel against officers' authority.

Contrary to expectations, males were not more likely to be disrespectful toward police. Compared to male suspects, the odds of being disrespectful for female suspects were increased by a factor of 1.6 (odds ratio=O.63 for male suspects). Although fe- male suspects were significantly more likely than male suspects to be classified as disrespectful by independent observers, they were not more likely to engage in resistant actions (i.e., noncompliance, ver-

bat aggression, or physical aggression) during their encounters with police. It is possible that female suspects were more likely to engage in less-threaten- ing forms of disrespect (e.g., complaining, sarcasm. nonverbal cues of disrespect including body lan- guage) that were captured in observers' character- izations of disrespect, but not captured in coded actions of resistance. This finding is supported by a larger body of research that suggests women are less likely to comply with officers' requests (Mastrofski et al., 1996), more likely to complain (Kowalski, 1996), and more likely to show anger and other emotions in conflict situations (Brody, 1999) compared to men.

Alternatively, one could argue that observers' characterizations of disrespect have less validity than measures of resistance. Researchers speculated that women who acted in a nonstereotypically feminine manner were more likely to be sanctioned by police. That is, women who do not fit the stereotypical female role in one way or another (e.g., by being part of a minority group, engaging in prostitution, or acting in a hostile or aggressive fashion) are more likely to be arrested (Chesney- Lind, 1978). Perllaps trained observers-who are typically middle-class college students-are more likely to interpret nontraditional behavior from female suspects as disrespectful compared to simi- lar behavior from male suspects. To disentangle these relationships. additional studies examining differences between coded actions and observers' characterizations of disrespect are needed.

Collectively. the current fmdings suggest that the factors that influence displays of resistance and disrespect toward police differ based on the type of actions being considered. This supports Klinger's (1994, 1996a) suggestion that researchers need to be more careful in their measures and conceptualiza- tion of demeanor. Future research should consider how to better measure suspect disrespect from patrol officers' rather than observers' perspectives. What is considered "disrespectful" behavior by an observer might be a typical-even expected-response to an experienced police officer. Future research should also explore these conceptualization and measure- ment issues through the use of debriefing techniques and protocol analysis (Mastrofski & Parks, 1990; Worden & Brandl, 1990).

These fmdings have other important implications for future research. It is clear that establishing tem- poral ordering for research-examining police-citizen encounters is critical. Tedeschi and Felson (1994) stressed the importance of the processes of interac- tions. Likewise, Sykes and Clark (1975) cautioned against intetpretations of police and citizen behavior that only consider the personal traits of the individ- uals rather than focusing on the interaction between

R.S. Engel/Journal of Criminal Justice 3 J (1003) 475-491 489

ings in Atlanta. November 8-10, 2001. Special thanks to Richard Felson, Eric Silver, Robert Worden. and Stephen Mastrofski for their helpful comments on an eartier draft of this article.

Notes

citizens and police. Unfortunately, the PSS data set does not establish time ordering of police-citizen interactions. Therefore, it is possible that suspects were acting in noncompliant, resistant, or disrespect- ful ways in response to officers' actions. Future research should examine the interaction between officers and citizens as it unfolds in time-this approach is the next step for the study of suspect and police demeanor. This may require the use of qualitative data to aid in the interpretation of coded information gathered during systematic obsclVation of police officers (Mastrofski & Parks, 1990).

Analyses of more current data that examine the relationship of suspects' perceptions of officers' le- gitimacy and suspects' displays of resistance toward police are needed. Due to the dated nature of the data utilized in this research, generalizability of the find- ings to current police-citizen relationships is ques- tionable. These analyses, however, have created baseline comparisons for future analyses using more current data. Furthennore, these analyses suggest that there are differences in the predictors of different fonDS of resistance and disrespect that researchers should further examine.

Of particular interest is whether variables mea- suring sb'atification and propensity for aggression continue to predict suspects' actions toward police. In the past decade. concern regarding the relation- ship between minority citizens and police has increased. Community perceptions of officers en- gaging in racial profiling during b'affic and inter- rogation stops, coupled with several high profile uses of excessive force (e.g., Rodney King, Abner Louima, Arnadou Diallo, Timothy Thomas), have led to increased tension between minority citizens and police and race riots. Findings from data collected in the late 1970s suggested that minority suspects showed passive fonDS of noncompliance toward officers but did not engage in more active fonDS of resistance including verbal and physical aggression. Applying Tyler's theory of citizen com- pliance, it was speculated that minority suspects who showed resistance toward officers did so, in part, because they questioned officers' legitimacy and authority. As minority citizens' discontent with police practices continues to grow, will more active fonDS of citizens' resistance and disrespect toward police be the result? Additional empirical research based on contemporary studies of policing is need- ed to further explore these relationships.

I. Black defined stratifICation as "any uneven dis1ri- bution of the material conditions of existence" (p. II).

2. Note, however, that only twenty encounters were observed with Black juveniles, six of which were considered

antagonistic. 3. Social organization was measured as additive

index including multiple arrests during a single encounter, the nature of the relationship between suspects and victims, and number of bystanders present. Sophistication was measured as an additive index including officers' experience, suspects' level of intoxication, if the dispute between citizens was in progress when police arrived. if the encounter was visible to others, amount of evidence, use of a weapon, and injury to the victim (Greenleaf &

Lanza-Kaduce, 1995). 4. The deference variable included the following three

categories: "( I) cases in which reversals predominated, (2) cases in which neither reversals nor reinforceR predomi- nated, and (3) cases in which reinforcers predominated" (Lanza-Kaducc & Greenleaf, 2000, p. 227).

5. The measUTe of norm resistance included "verbal attacks on officers, physical assaults on officers, refusal by citizens to obey a lawful order, resisting arrest or interfering with officers, and the use of force by police" (Lanza-Kaduce &. Greenleaf, 2000, p. 227).

6. Citizen compliance was measured as citizens who complied or indicated a willingness to comply in the future to three types of officers' requests: (I) to discontinue illegal behavior, (2) to leave another person alone or leave the premises, or (3) to calm down or cease disorderly behavior (Mastrofski et aI., 1996).

7. Note however, that citizens' actual perceptions of police will not be measured in the analyses to follow. Rather, suspects' characteristics are utilized as proxy measures.

8. Two alternative arguments could be made. First, it could be argued that suspects in more serious encounters with police are more likely to be deferential, even extremely deferential, because they recognize the gravity of the situation and the strong possibility of police coen:ive action. Second. some ethnographic resean:h reported that officers deflDed citizens as disrespectful based on the immediate interaction between them rather than on the facts that initially led to the police-citizen encounter. As explaincd by Van Maanen (1978), "the police view of the asshole as deviant is a product of the immediate transaction between the two and nol a product of an act preceding the transaction" (p. 228; emphasis added).

9. The list of possible demeanor codes included: (I)

businesslike, (2) friendly, (3) apologetic, (4) pleading, tlying to enlist officer's aid, sympathy, (5) frightened, aftaid of officers, (6) cool, detached. could not care less, (7) sarcastic, disrespectful, hostile. (8) other, (9) do not know, not

Acknowledgements

A previous version of this article was presented at the American Society of Criminology annual meet-

490 R.s. Engel! Journal of Criminal J/Lftice 3/ (2003) 475-492

ascertained, (10) unconscious, (II) citizen not present or uninvolved. Observers were given the following instruc- tions:

Note that these items specify demeanor toward the officer. When in doubt code as other (08). As with the emotional state codes, we are trying to capture clearly recognizable demeanors in these items. Somewhat submissive behavior on the part of the citizen partic- ipants can be coded as businesslike (0 I ) and this may be a part of nomIaI police-citizen roles. (Police Services

Study, coding manual). 10. Of the 1.8 percent of suspects encountered by

female officers, only one was coded as disrespectful. Due to this limited sample size, the sex of the officer was excluded from the multivariate analyses that follow.

II. For each encounter, the observer recorded at least one and up to three codes, selected from a list of 247 problem codes, to characterize the nature of the problem at each of three tirnes: (I ) as it was initially presented to police (usually by a dispatcher), (2) upon the officer's arrival at the scene, and (3) at the end of the encounter. The seriousness scale was based on the most complete and accurate information available to the observer at the conclusion of the encounter.

12. The independent variables were measured at different levels of analysis: officer and suspect. Although some officers did encounter multiple suspects, only 17 percent of the officers were involved in five or more encounters, the minimum number necessary for the use of hierarchical linear modeling (Bryk & Raudenbush, 1992). Therefore, nonhierarchical logistic regression was the appropriate technique to use for these analyses.

13. Mastrofski et al. measured citizen compliance as citizens who complied or indicated a willingness to comply in the future to three types of officers' ~ts: (I) discontinue illegal behavior, (2) leave another person alone/leave the premises, or (3) calm down/cease disorderly behavior. In contrast. the current measure of nonco"1'liance was a more general measure that included "refusing to answer questions" and "refusing to otherwise cooperate with officers' requests."

14. The data collected from Mastrofski's study also revealed other rather unique findings that suggested Richmond was different than other police departments. For example, Mastrofski et al. (1995) were the only researchers who reported a statistically significant relationship between offICers' ani tudes and behavior. In addition, this study was one of only two that found suspects' demeanor did not influence officers' behavior (see also Klinger. 1994).

References

Albrecht, S. L., & Green, M. (1977). Attitudes toward the poIic:c and the larger attitude complex, implications for police-community relationships. Criminology, J 5, 67-86.

Bayley, D. H., & Mendelsohn, H. (1969). Minorities and the police. confrontation in America. New York: Free Press.

Bittner, E. (1974). Florence Nightingale in pursuit of Willie

Sutton: A theory of the police. In: H. Jacob (Ed.), 77re potential for reform of criminal justice (pp. 17-43). Beverly Hills, CA: Sage.

Black, D. (1976). 77re behavior of law. New York: Academic Press.

Black, D. (1980). The manners and customs of the police. New York: Academic ~.

Black, D., .t. Reiss, A. J. (1970). Police control of juveniles. American Sociological Review, 35, 63 - 77.

Brody, L. (1999). Gender; emotion and the family. Cam- bridge, MA: Harvard University ~.

Brown, M. K. (1981). Working the street: Police discretion and the dilemmas of reform. New York: Russell Sage Foundation.

Brown, P., &. Levinson, S. C. (1987). Politeness: Some uni- versals in language usage. New York: Cambridge Uni- versity Press.

Browning, S. L., Cullen, F. T., C~, L., Kopacbc. R., .t. Stevenson, T. J. (1994). Race and getting hassled by the police: A research note. Police Studies, 17, I-II.

Bryk, A. S., &. Raudenbush, S. W. (1992). Hierardtical linear models: Applications and data analysis methods. Newbury Park, CA: Sage.

Caldwell, E. (1978). Patrol observation: The patrol encoun- ter; potrol nOITative, and general shift information forms (police Services Study Report MR-Ol). Bloomington: Indiana University, Workshop in Political Theory and

Policy Analysis. Chesney-Lind, M. (1978). Chivalry re-examined: Women

and the criminal justice system. In: L. Bowker (Ed.), Women Crime and the Criminal Justice System. Lexing- ton, MA: D.C. Health.

Decker, S. H. (1981). Citizen attiwdes toward the police: A review of past findings and suggestions for future policy. Journal of Police Science and Admini.ttration, 9, 80- 87.

Dunham, R. G., &. Alpert, G. P. (1988). Neigbborhocxl dif- ferences in attitudes toward policing: Evidence for a mixed strategy model of policing in a multi-ethnic set- ting. Journal of Criminal Law and Criminology, 79, 5O4-S23.

Engel, R. S., Sobol, J. J., & Worden, R. E. (2000). Further exploration of the demeanor hypothesis: The interaction effects of suspects' characteristics and demeanor on p0- lice behavior. Justice Quarterly, 17, 23S-2S8.

Fagan, J. (1990). Intoxication and aggression. In: M. Tonry, & J. Q. Wilson (Eds.), Drugs and crime, crime and justice, (Vol. 13, pp. 241-320). Chicago: University of Chicago Press.

Greenleaf, R. G., & Lanza-Kaduce, L. (I99S). Sophistica- tion, organization and authority-subject conflict: Redis- covering and unraveling Turk's theory of norm resis lance. Criminology, 33, S6S-S8S.

Hurst, Y. G., & Frank, J. (2000). How kids view cops: The nature of juvenile attitudes toward the police. JoUnlaI of Criminal Justice, 18, 189-202.

Jacob, H. (1971). B lack and White perceptions of justice in the city. Law and Society Review, 6,69-89.

Jesilow, P., Meyer, J.. &. Namazzi, N. (I99S). Public atti-

tudes toward police. American Journal o[ Police, 14, 67-89.

Klinger, D. A. (1994). Demeanor or crime? Why 'hostile'

R.S. Engel / JournQI of Criminal Justice jJ (200J) 475-492 491

citizens are more likely to be ~. Criminology, 12,

475-493.

Klinger, D. A. (1996a). Bringing crime back in: Toward a better understanding of police an-est decisions. Joumal

of Researc;}, in Crime and Delinquency, 11, 333-336. Klinger, D. A. (l996b). More on demeanor and arrest in

Dade County. Criminology, 34, 61-82. Kowalski, R. M. (1996). Complaints and complaining:

Func~ antcce<blts, and consequences. Psydtologi- calBllilelin, //9,179-196.

Lanza-Kaduce, L., & Greenleaf. R. G. (2000). Age and race deference reversals: Extending Turk on police-citi7a1 conflict. Journal of Researr:h in Crime and Delinquency, 37,22/-236.

Lipsky, M. (1980). Street-level bureaucracy. dilemmo.\' of the individual in public services. New York: Russell Sage Foundation.

Lundman. R. J. (1974). Routine police arrest practices: A commonweal perspective. Social Problems, 22, 127 -141.

Lundman. R. J. (1994). Demeanor or crime? The Midwest

city police-citizen encounters study. Criminology, 32,

631-656.

Lundman, R. J. (1996). Demeanor and anat: Additional evidence from ~iously unpublished data. Journal of Research in Crime and Delinquency, 33, 306-323.

Lundman, R. J.. Sykes. R. E., & Clark:, J. P. (1978). Police control of juveniles: A replication. Journal of ReseQlrh in Crime and Delinquency, 15, 74-91.

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

Mastrofski, S. D., Snipes, J. B., Parks, R. B., & Maxwell, C. D. (2000). The helping hand of the law: Police con- trol of citizens on request. Criminology, 38. 307 - 342.

Mastrofski. S. D., Snipes, J. B., & Supina, A. E. (1996). Compliance on demand: The public's response to spe- cific police requests. Journal of Researr:h in Crime and

Delinquency, 33, 269-305. Mastrofski, S. D., Women, R. E., & Snipes, J. B. (1995).

Law enforcement in a time of community policing.

Criminology, 33, 539-563. McCluskey, J. D.. Mastrofski. S. D., & Parks, R. B. (1999).

To acquiesce or rebel: Predicting citizen compliance with police requests. Police Quarterly. 2, 389-416.

Muir, W. K. Jr. (1977). Police: Streetconler Politicians.

Chicago: University of Chicago PreJs. Novak, K. J., Frank, J., Smith, B. W., & Engel, R. S. (2002).

Revisiting the decision to arrest. Crime and Delin-

quency, 48(1). 70-98. Parks. R. B., Mastrofski, S. D., DeJong. C., & Gray, M. K.

(1999). How officers spend their time with the commun- ity. Juslice Quarterly, /6.483-518.

Piliavin, I., & Briar, S. (1964). Police encounters with juve- niles. American Journal of Sociology, 70,206-214.

Scaglion, R., & Condon, R. G. (1980). The Structure of Black and White attitudes toward police. Human OIgan-

izalion, 39, 280-283. Skolnick, J. H. (1966). Justice without trial: Law enforce-

ment in a democratic society. New York: Wiley. Smith, D. A., & VISbcr, C. A. (1981). Street-level justice:

Situational determinants of p>lice arrest decisions. S0- cial Problems, 29, 167-177.

Sykes, R. E., & Clark, J. P. (1975). A theory ofdefcrence

exchange in p>lice-civilian encounters. American Jour-

nolofSociology, 8/, 584-600. Taylor, T. J., Turner, K. B., Esbensen, F., & Winfree, L. T.

(2001). "Coppin" an attitude: Attitudinal differences among juveniles toward police. Journal of Criminal Jus- tice, 29, 295-305.

Tedeschi, J. T., & Felson, R. B. (1994). J-lolellce. aggres- sioll. alld coercive actiollS. Washington, DC: American

Psychological Association. Tuch, S. A., & Weitzel', R. (1997). The poils-trends, racial

differences in attitudes toward the police. Public Opill- ion Quarterly, 61, 642~663.

Turk, A. T. (1969). Criminolily and legal onler. Chicago: Rand McNally.

Tyler, T. R. (1990). Why people obey the law. New Haven, CT: Yale University Press.

Tyler, T. R., & Folger, R. (1980). Distributional and proce- dural aspects of satisfaction with citizen - police encoun- ters. Ba.fic alld Applied Social Psychology, I, 281 - 292.

Van Maanen, J. (1974). Working the street: A developmental view of police bd\avior. In: H. Jacob (Ed.), 77Ie potelltial

for rFfonll of criminal justice (pp. 83 -129). Beverly Hills, CA: Sage.

Van Maanen, J. (1978). The assbole. P. K. Manning, & J. Van Maaneo (Eds.), Policing: .4. view from the strFet (pp. 221 - 238). Santa Monica, CA: Goodyear Publishing.

Velez, M. B. (2001). The role of public social control in urban neighb<Xboods: A multilevel analysis of victim- ization risk. Criminology, 39, 837-863.

Walker, S. (1999). The police in America. (3rd ed.). Boston; McGraw Hill.

Webb, V. J., & Marshall, C. E. (1995). The relative impor- tance of race and ethnicity on citizen attitudes toward the

police. American Journal of Police, 14(2),45-66. Weitzer, R. (2(KX). Racialized policing: Residents' pen:ep-

tions in three neighborhoods. Law alld Society, 34, 129-155.

Weitzer, R., & Tuch, S. A. (1999). Race, class, and pen:ep- tions of discrimination by the p>lice. Crime alld Delin-

qumcy, 45, 494-507. Westley, W. A. (1953). Violence and the police. Americall

Journal of Sociology, 59, 34-41. Westley, W. A. (1970). Violence and the police: A socio-

logical stlldy of law, custom, and morality. Cambridge, MA; MIT Press.

Williams, H., & Murphy, P. (1990). The evolving strategy of police: A minority view. Penpectives Oil Policing,

(Vol 13, pp. I -15). Washington, DC: National Insti- tute of Justice.

Worden, R. E. (1989). Situational and attitudinal explana- tions of police behavioc A ~ reappraisal and

empirical assessment. Law and Society Review, 23, 667 - 711.

Worden, R. E., & Brandl, S. G. (1990). Protocol analysis of p>lice decision-making: Toward a theory of police

behavior. American Joumal of Criminal Justice, 14, 297-318.

Worden, R. E., & Pollitz, A. A. (1984). Police am:sts in

R.s. Engel / JOIInIal of Criminal JrlStiu 3 J (1003) 475-491492

Worden, R. E., Si1q>ard, R. L., & Mastrofski, S. D. (1996). On the ~ing and measurement of suspects' demeanor toward the police: A comment on "Demeanor and Ar- rest". Journal of Research in Crime and Delinquency,

33,324-332.

domestic disturbances: A further look. Law and Society Review, 18, 105-119.

Worden, R. E., & Shepard, R. L. (1996). Demeanor, crime, and police behavior: A reexamination of the police serv- ices study data. Criminology, 34, 83 - 105.