”musmflo.” 3.: L... 34.1“ ‘ ugh...» n. Murmuv . . swwwfiug Hum. . :23 1. m 5 .. 3:97 LIBRARY Michigan State University This is to certify that the thesis entitled THE EFFECT OF PROTEST POLICING ON MASS DEMONSTRATIONS BETWEEN 2004 AND 2006 IN SOUTH KOREA presented by YUNJONG YU has been accepted towards fulfillment of the requirements for the MS. degree in Criminal Justice / " 7 , ’ cQ/fijéf’wfl Majcr Professor’s Signature 61.44/36”, 3 av 7 Date MSU is an affinnative-action, equal-opportunity employer PLACE IN RETURN BOX to remove this checkout from your record. To AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. 1 DATE DUE l DAIEDUE DAIEDUE lam outing t‘r T36 * 6/07 p'ICIRC/DateDueindd-pJ THE EFFECT OF PROTEST POLICING ON MASS DEMONSTRATIONS BETWEEN 2004 AND 2006 IN SOUTH KOREA By Yunjong Yu A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE School of Criminal Justice 2007 ABSTRACT THE EFFECT OF PROTEST POLICING ON MASS DEMONSTRATIONS BETWEEN 2004 AND 2006 IN SOUTH KOREA By Yunjong Yu The purpose of this study is to examine the relationship between police actions to control mass demonstrations and the peaceful or violent behavior of the demonstrators in South Korea. In order to accomplish this objective, this study uses existing police records on mass demonstrations from the Korean National Police Agency (KNPA) as well as data from the agency’s Annual Report. The sample of cases includes all demonstrations with over 1,000 demonstrators occurring over a two and one-half year period (1/1/04-6/30/06). Logistic regression analysis indicated that the characteristics of demonstrations and police measures had a strong effect on the violent behavior of the demonstrators. Organizers with a history of violence, demonstrations with political/military purposes, the close distance between police and demonstrators, the arrangement of police equipment, and the presence of parade were more likely to be involved in violent demonstrations. However, contrary to expectation, organizers with joint-groupS/sub-groups, the number of demonstrators, jurisdiction, the ratio of demonstrators to police officers, the deployment of female officers, and the presence of outside observers at the demonstrations were not Si gnificantly related to the violent behavior of demonstrators. These findings suggest policy implications to the police in finding effective protest policing on mass demonstrations in South Korea. ACKNOWLEDGEMENT There are many kind individuals that I want to be deeply grateful to throughout the thesis process. I could not complete this work without their support and coaching. First, I would like to thank the members of my committee, Dr. Edmund F. McGarrell, Dr. Mahesh Nalla, and Dr. Sheila Royo Maxwell. They not only advised but also encouraged me throughout the entire research process. Especially, I thank for their kindness and patience. They made me feel competent despite my lack of ability. They also allowed me the opportunity to find out the value of continued learning in my life. It was my great honor to work with such fantastic professors. Second, I want to give Special thanks to my police partners, including Youngwoo Lee and Heekyoung Kim working in the Korean National Police Agency (KNPA). Their constant availability and assistance allowed me to complete this thesis. They also continued to provide support throughout my graduate school experience in the US. Third, I would like to extend the gratitude to PhD. candidates, J inseong Cheon g, Daehoon Kwak, Eunseong Jeong, and Wangjun Kim for their willingness to challenge me to learn, interpret, and explain the ins and outs of logistic regression analysis. Fourth, I must also thank Alexis, N ayana, Julia, and Cassie who helped my writing during my thesis process. Their kindness and assistance were much appreciated. Finally, my wife, Youngiu Min and daughter, Doyeon, deserve a special thank for their profound love and support. It was their understanding and motivating encouragement that allowed my graduate study to be completed successfully. They have always revitalized me by showing their lovely smiles during my 2 years in the US. iii TABLE OF CONTENTS LIST OF TABLES ..................................................................... vi LIST OF FIGURES ..................................................................... vii CHAPTER I: INTRODUCTION ...................................................... 1 CHAPTER II: OVERVIEW OF DEMONSTRATIONS IN SOUTH KOREA 7 Background of South Korea ............................................................ 7 Number of Demonstrations .............................................................. 9 Causes and Patterns of Demonstrations ............................................... 11 CHAPTER III: COMPARISON OF PROTEST POLICIN G .................... 16 Protest Policing in South Korea ........................................................ 16 Harsh Penalty Policy ............................................................... 16 Increase of Police Presence Policy ............................................... 17 Stop of Use of Tear Gas Policy ..................................................... l 7 Participation of Outside Observers .............................................. 20 Protest Policing in the US .............................................................. 20 No Intervention ..................................................................... 21 Escalated Force Style ............................................................... 21 Negotiated Management Model ................................................ 22 Command and Control Style ........................ A .............................. 24 Summary of Protest Policing ............................................................ 25 CHAPTER IV: LITERATURE REVIEW .......................................... 27 Effects of Characteristics of Demonstrations ....................................... 27 Organizers of Demonstrations ................................................... 27 History of Violence .............................................................. 29 Purposes of Demonstrations ..................................................... 29 Size of Demonstrations ........................................................... 31 Effects of Police Measures ............................................................ 32 Jurisdiction .......................................................................... 32 Police Size ......................................................................... 33 Police Equipment ................................................................. 35 Police Permit for Parade .......................................................... 36 Female Officers .................................................................... 36 Effects of Outside Observers ......................................................... 40 Research Questions ..................................................................... 43 CHAPTER V: DATA AND METHODS ............................................ 45 Data ....................................................................................... 45 Sample ..................................................................................... 48 Dependent Variable ...................................................................... 49 Independent Variables .................................................................. 50 iv Analytic Strategies ......................................................................... 54 CHAPTER VI: RESULTS ................................................................. 57 Univariate Statistics ...................................................................... 57 Description for the Independent Variables ...................................... 57 Description for the Dependent Variable ......................................... 62 Bivariate Statistics ....................................................................... 63 Multivariate Statistics ................................................................... 71 Multicollinearity Examination ................................................. 71 Logistic Regression Analysis .................................................... 74 Summary ................................................................................. 80 CHAPTER VII: DISCUSSION AND CONCLUSION ............................. 83 Explanations for Results ............................................................... 83 Study Limitations ......................................................................... 90 Future Study .............................................................................. 93 Conclusion and Policy Implications ................................................. 95 APPENDICES Appendix A: Correlation Matrix (N=921) .......................................................... 103 Appendix B: Collinearity Diagnostics With Organizers of Demonstrations. 104 Appendix C: Collinearity Diagnostics Without Organizers of Demonstrations. 105 REFERENCES ............................................................................. l 06 Table 1. Table 2. Table 3. Table 4. Table 5. Table 6. Table 7. Table 8. Table 9. Table 10. LIST OF TABLES Total Number of Demonstrations Between 1995 and 2003 ........... The Number of Illegal Demonstrations, Demonstrations used Fire Bottles, and Number of Fire Bottles ...................................... The Number of Used Tear Gas and Used Fire Bottles during the Demonstrations ............................................................. The Number of Police Injury during the Demonstrations ............. Description of Variables for Analysis .................................. Descriptive Statistics for Mass Demonstrations in South Korea Relationship between Selected Independent Variables and Dependent Variable ...................................................... Tolerance and Variance Inflation Factors (VIF) ...................... Tolerance and Variance Inflation Factors (VIF) With and Without Organizers of Demonstrations ........................................... Logistic Regression Results .............................................. vi 10 18 19 52 58 65 72 73 75 Figure 1. Figure 2. Figure 3. Figure 4. Figure 5 LIST OF FIGURES Total Number of Demonstrations Between 1995 and 2003 ............ Total Number of Participants in Demonstrations Between 1995 and 2003 ........................................................................... Number of Illegal Demonstrations and Demonstrations used Fire Bottles ........................................................................ The Number of Tear Gas and Fire Bottles during the Demonstrations ............................................................... Hypotheses for Relationship between Independent Variables and Dependent Variable .......................................................... vii 10 19 44 CHAPTER I : INTRODUCTION Demonstrations are an individual’s fundamental, constitutional right across space and time (Kim, 2004; Lee, 2006; Smith, 2004; Yu, 2003). The First Amendment of the United States Constitution proscribes the making of any law abridging “the right of people peaceably to assemble.” Similarly, the right to assembly in South Korea is described in article 21 of the Constitution as follows: “All citizens enjoy the freedom of Speech and the press, and of assembly and association.” Such demonstrations as a right to freedom of assembly allow people to gather together and express their sentiments, especially when they intend to express their views about governmental actions and other public issues (Nalla, 2006; Smith, 2004). Due to the increasing frequency of demonstrations, these demonstrations have become a central part of the process of political representation in Western Democracies (Dalton, 1988; Della Porta & Reiter, 1998; Koopmans, 1993; McCarthy, McPhail, & Smith, 1996; Tilly, 1983). However, this right guaranteed by the Constitution is not absolute. The police perform the difficult task of maintaining public order, while protecting first amendment rights during the demonstrations. Thus, the police try to impose “reasonable” restrictions on time, place, duration, or manner of demonstration (Cox v. New Hampshire, 194]). These restrictions on demonstrations often require a balancing of numerous and variable conflicting interests, but in reality, it is difficult to balance conflicts between protecting freedom of demonstration and protecting the peace and tranquility of the community. The police sometimes conflict with the demonstrators in preventing violence, maintaining free traffic, and insuring that public facilities serve the needs and convenience of other citizens (Smith, 2004). From the police’S point of view on controlling demonstrations, the police do not mechanically enforce the law, but vary their responses according to protests. The police often play a careful and strategic role at protests, which is most often designed to maintain control (Della Porta & Reiter, 1998; Earl, Soule & McCarthy, 2003; McCarthy, et al., 1996). The police have a great deal of policing options in the way they deal with demonstrations, ranging from continued presence with no further action to the deployment of “escalated force” (Earl et al., 2003; McCarthy et al., 1996; McPhail, Schweingruber, & McCarthy, 1998). Also, many scholars emphasize the interaction between police and the action of crowds to control demonstrations effectively (Reicher, Stott, Cronin, & Adang, 2004). They‘advised that we can not understand protest repertoires and their evolution without understanding this interaction. Over the last decade, there has been extensive research on the policing of protest in the Western democracies, including cross-national comparative research (Della Porta & Reiter, 1998). For example, in Europe, the police are using communications much more effectively, but are using paramilitary measures particularly when state police play a leading role (Della Porta & Reiter, 1998). In the US, the leading view has also been more conciliatory approach emphasizing communications (McPhail et al., 1998). However, this is being revised because of the emergence of new characteristics of protests such as anti-globalization (Button, John, & Brearley, 2002; Noakes & Gillham, 2005). In Canada, there has historically been a low tolerance for public disorder, but Canada has changed into a hybrid police form in which control and service are practiced Simultaneously (McPhail et al., 1998). However, previous literature has some limitations for understanding mass demonstrations in South Korea. First of all, a few studies examined a relationship between the characteristics of demonstrations and the violent behavior of demonstrators. That is, although the characteristics of demonstrations are primary variables that may strongly affect the violent behavior of demonstrators in practice, not many Studies have described the effect of the characteristics of demonstrations satisfactorily. In reality, many police agencies predict and judge whether or not reported demonstrations transform violent by checking the characteristics of demonstrations in advance. Based on this judgment, the police usually set and decide the protest management practices for the reported demonstrations (Kim, 2004; Lee, 2006; Yu, 2003). Nevertheless, most research mainly focused on the relationships between the styles of policing and demonstrations, or between the characteristics of demonstrations and police responses to demonstrations such as police brutality. The results of these studies suggested that police actions have a Significant influence on the behavior of citizens involved in mass demonstrations (Brown, 2006; Carter, 1987; Della Porta & Reiter, 1998; Dotson, 1974; Earl et al., 2003; Gurr, 1970; Hibbs, 1973; Koopmans, 1993; McPhail et al., 1998; Morgan & Clark, 1973; Nalla, 2006; Schweingruber, 2000; Vitale, 2005). Thus, to understand mass demonstrations precisely in South Korea, in addition to effects of police measures, the effects of the characteristics of demonstrations on the level of violence by demonstrators should be examined. This is one of the aims of this study. Next, previous studies focused on the policing of protest in the US, though there have been recent transformations in the policing of protests across Western democracies (Della Porta & Reiter, 1998). Not many studies have examined the relationship between police measures and the violent behavior of demonstrators with respect to Korean views. In South Korea, both police measures and the violent behavior of demonstrators have been influenced by historical and political factors. For example, in South Korea, many citizens believe in a right to demonstration in response to South Korea’s history of military governments (Hwang, McGarrell, & Benson, 2005). Furthermore, the police are often seen as an extension of authoritarian control. Thus, in general, demonstrators do not recognize the police as legitimate authority when the police control mass demonstrations. Korean scholars argue that police measures do not largely affect the violent behavior of demonstrators. In contrast, the police assume their authority in exerting control and do not recognize how their behaviors may influence crowd behaviors. Therefore, in order to suggest effective policy to control mass demonstrations, the historical background of South Korea Should be considered in the study. The hypothesis of this study results from such reasons that historical factors in South Korea are more likely to affect the violent behavior of demonstrators. Finally, even in South Korea, no studies have examined the effect of protest policing on mass demonstrations by analyzing a large sample of cases of demonstrations in South Korea. This method is a prerequisite for drawing correct results in the study. Thus, this research will become the first attempt to examine the relationship between the policing of protests and mass demonstrations at the incident level of mass demonstrations in South Korea. Therefore, the primary purpose of this study is to understand mass demonstrations in South Korea and find the appropriate police practices to reduce the levels of violence by demonstrators through analyzing and assessing various police measures attempted in South Korea. In detail, this study will examine whether or not a variety of police measures to control mass demonstrations, such as the ratio of demonstrators to police officers, the distance between police and demonstrators, the arrangement of police equipment, and the deployment of female officers have a significant effect on the violent behavior of demonstrators. If police measures do not affect the violent behavior of demonstrators because of unique, historical factors of South Korea, this Study will additionally reveal what is related to the violent behavior of demonstrators. In order to understand this, this study will examine a relationship between the characteristics of demonstrations and the violent behavior of demonstrators. In other words, this study will examine whether organizers of demonstrations, a history of violence, the purposes of demonstrations, and the number of demonstrators affect the violent behavior of demonstrators, regardless of police measures. Furthermore, this study will deal with the effect of the presence of Non Governmental Organization (NGO)SI as an observer on the violent behavior of demonstrators. This strategy is recently recognized as a type of community policing model in South Korea (Hwang et al., 2005). To raise the level of reliability of this study, it will use the police data. Previous researchers heavily relied on using the media sources, such as newspapers and electronic news reports that are often the common source of data on protest events in many places ‘ Since 2004, the Korean National Police Agency has worked with NGOS as an observer to overlook behaviors of the police and demonstrators at the demonstrations. To avoid confusing NGOS as an observer with NGOS as an organizer of demonstrations, it describes NGOS as outside observers in this study. and times (Earl et al., 2003; Olzak, 1989). However, the use of such media sources has some problems, such as selection bias, description bias, and researcher bias2 (Olzak, 1989; McCarthy et al., 1996). These problems may lead researchers to wrong inferences about the characteristics of the demonstrators they hope to describe and understand (Olzak, 1989; McCarthy et al., 1996). For these reasons, this study addressed these limitations by using existing police records on mass demonstrations from the Korean National Police Agency (KNPA) as well as data from the agency’s Annual Report. More specifically, this study uses a relatively objective sample of mass demonstrations (N: 921) drawn from the KNPA. The sample includes all demonstrations with over 1,000 demonstrators occurring in a two and one half year period (1/1/04-6/30/06). In addition, this study focuses on explaining perspectives for understanding demonstrations and measures which may apply to the situations of South Korea. Thus, the first section describes a unique historical background of South Korea. Then the paper reviews various ways in which the KNPA has controlled demonstrations in the past and introduces the policing of protest in the US. The next section reviews the literature on the effects of characteristics of demonstrations, police measures, and outside observers on violence by demonstrators. In the main section, we examine the relationships between various variables: whether the characteristics of demonstrations, police measures, and outside observers have an influence on the level of violence by demonstrators. Finally, we suggest the policy implications that can apply in South Korea in accordance with results of this study. 2 Media biases include: media bias in the selection of a few of the many possible events to observe and report (selection bias), media bias in the descriptions of the events they do select to report (description bias), and the reliability and validity of media trace recovery by researchers (researcher bias) (see Olzak, 1989). CHAPTER II: OVERVIEW OF DEMONSTRATIONS IN SOUTH KOREA Background of South Korea The demonstrations and riots in South Korea reflect an unique historical background (Yu, 2003). Different troubles throughout the nation’s history have caused demonstrations and riots. Historically, South Korea experienced colonization by Japan for more than three decades during the first half of the twentieth century. In the aftermath of Japan’s defeat in World War II in 1945, Korea was divided into South Korea and North Korea. Thus, the history of South Korea formally began in 1948 with Syngman Rhee as the first president. In 1950, the Korean War broke out when North Korea invaded South Korea. After the armistice, South Korea experienced political turmoil under years of autocratic leadership of Syngman Rhee, which was ended by student revolt in 1960 (Nahm, 1996). After the student revolution, a new parliamentary election was held on July 29, 1960. Yun PoSun was elected as the President on August 13, 1960. During the Second Republic, the proliferation of political activity which had been repressed emerged. Much of this activity was from leftist and student groups. Union membership and activity grew rapidly during the later months of 1960. Around 2,000 demonstrations were held during the eight months of the Second Republic. The Second Republic was strongly democratic, but was overthrown in less than a year and replaced by a military coup d’état3 which was led by Major General Park ChungHee on May 16, 1961. During this military rule, many people had suffered because of his long autocratic leadership (Nahm, 1996). For example, he promulgated emergency 3 A coup d’état is the sudden overthrow of a government through unconstitutional means by a part of the State establishment (See American Heritage dictionary of the English Language). It is most common in countries with unstable governments and in countries with little experience of successful democracy. decrees in 1974 and 1975 which led to the jailing of hundreds of dissidents. However, for this time, this country continued dramatic economic growth. Thus, South Korea was hailed as one of the “Four Dragons” in Asia in addition to Hong Kong, Singapore, and Taiwan (Hwang et al., 2005). This country has become an economically fast-develOping country, and has been willing to compete with Japan for world trade. After president Park ChungHee was assassinated in 1979, a vocal civil society emerged that led to strong protests against the authoritarian rule. Composed primarily of university students and labor unions, protests reached a climax after Major General Chun Doo-hwan’s 1979 Coup d’état. Public outrage consolidated nationwide support for democracy, paving the road for the first democratic elections in 1987 (Nahm, 1996). With the Sixth Republic, the country has gradually stabilized into a liberal democracy. Finally, in 1992, Kim Young—3am was elected as the first nonrrrilitary president. In sum, as stated above, South Korea historically experienced a military government for much of the late twentieth century (Fowler, 1999; Lab & Das, 2003; Nalla & Hwang, 2004). The Third, Fourth, and Fifth Republics, which were military government, were nominally democratic, but were widely regarded as the continuation of military rule. During the period of military government, there had been many demonstrations and riots that requested democratization. Law enforcement agencies suppressed the requests of citizens and many people experienced unfairness of law enforcement in the process of regression (Fowler, 1999; Lab & Das, 2003; Nalla & Hwang, 2004). Consequently, the police force was perceived as the negative image by its oppressive role. Thus, the government was partially to blame for some acts of violence during the demonstrations (Hwang et al., 2005). Number of Demonstrations In South Korea, demonstrations are comparatively common and usually are peaceful, and illegal demonstrations still occur. However, among them, there are sometimes violent demonstrations called riots where the level of danger is high and often involves fire bottles during the demonstrations. Since 1995, the total number of demonstrations increases gradually (see Table 1, Figure l and 2). According to the KNPA, the number of demonstrations increased from 11,605 cases in 1995 increased to 24,541 cases in 2003. Similarly, the number of demonstrators also increased from 2,319,000 in 1995 to 2,778,430 in 2003.4 Table 1: Total Number of Demonstrations Between 1995 and 2003. Year 1995 1996 1997 1998 1999 2000 2001 2002 2003 Cases 1 1,605 12,219 9,729 1 1,797 17,209 22,011 23,946 34,138 24,541 Participants 2,319, 1,016, 1,500, 2,108, 2,978, 3,61 1, 3,167, 4,587, 2,778, 000 960 220 330 250 170 360 900 430 *Source: Korean National Police Agency, 2004. Figure 1: Total Number of Demonstrations Between 1995 and 2003. 40,000 - , -—-— ., 35.000 30,000 A : 25,000 \— 20,000 15,000 // 10,000 , Ar/ .5 5,000 Number of Demonstrations 0 I T l r r r r r i 1995 1 996 1 997 1 998 1999 2000 2001 2002 2003 Year 4 The number of demonstrations in Table l and Table 2 considered all demonstrations occurring in South Korea, including demonstrations with over 1,000 participants as the sample of this study. Figure 2: Total Number of Participants in Demonstrations Between 1995 and 2003. 5000000 4500000 4000000 Number of Participants N 01 O O 8 O 1 995 1996 1 997 1 998 1999 2000 2001 2002 2003 Year According to the KNPA (2004), the number of illegal demonstrations largely decreased from 809 cases in 1995 to 134 cases in 2003 (see Table 2 and Figure 3). In particular, the number of violent demonstrations using fire bottles largely decreased and the number of fire bottles used at the demonstrations also tremendously decreased since 1998, when compared to 1997. There are many reasons for this, most scholars think it was Significant change in police policy to control demonstrations (Kim, 2004; Yu, 2003). This matter will be discussed later. However, violent demonstrations using fire bottles at the demonstrations still occur and result in serious problems in South Korea. Table 2: The Number of Illegal Demonstrations and Demonstrations used Fire Bottles. Year 1995 1996 1997 1998 1999 2000 2001 2002 2003 Illegal Demonstrations 809 8 1 1 664 67 129 105 215 l 1 8 134 Demonstrations 143 190 172 2 7 7 23 8 14 used Fire Bottles *Source: Korean National Police Agency, 2004. 10 Figures 3: Number of Illegal Demonstrations and Demonstrations used Fire Bottles. 900 300 -_|=\ 700 600 1 —-I— Illegal Demonstrations Number 500 \\ 400 + Demonstrations used 300 X Fire Bottles 200 \ 100 m\ JV W1 0 l I l to o ’\ e x (L e» ’99 '99 .39 ’90 ’99" (90° 09° q9o (’90 Year Causes and Patterns of Demonstrations There can be found similarities and differences between South Korea and the US. in terms of causes and patterns of demonstrations. There are a substantial number of theories that attempt to explain the factors that cause a person or a group to protest and act violently during the demonstrations. Harris (1997) described types of these theories as theory of aggression, authoritarian personality theory, frustration-aggression theory, and social learning theory. He also explained that negative social, economic, and political conditions lead people to choose violence as a method to achieve their desired goals. These conditions in detail mean poverty (deteriorating economic conditions), discrimination and conflicts between races or religions, police brutality, massive job losses, reduction of governmental financial aid (dismantled government support), or even the outcome of a Sporting event (Harris, 1997; Myers & Li, 2001). For example, the 1992 ll Los Angeles riot was associated with poverty, discrimination, and police brutality that had persisted for decades (Ong & Hee, 1993). However, Harris’s arguments have limited value for explaining demonstrations and riots in South Korea. Demonstrations and riots in South Korea usually occur because of governmental causes (i. g., one-sided policy decision-making), or global causes (i. g., World Trade Organization, Free Trade Agreement), but as stated earlier, South Korea has additionally historical factors. Thus, when the historical background of South Korea has been considered, “procedural justice (Tyler, 2003) and defiance theories (Sherman, 1993)” have more proper value to explain demonstrations and riots in South Korea. According to Tyler’s (2003) procedural justice theory, the public’s law-related behavior is powerfully influenced by people’s subjective judgments about the fairness of the procedures. In other words, people are more likely to obey and cooperate if they feel that they have been fairly treated, but if they do believe that legal authorities are illegitimate, the likelihood of defiance, hostility, and resistance increase (Tyler, 2003). Similarly, Korean people have still thought of demonstrations including riots, as a kind of right to defiance toward the past illegitimate military government. Thus, most demonstrators do not think of riots as serious even if they are accompanied by illegal means (Hwang et al., 2006; Kim, 2004; Yu, 2003). In addition, in South Korea, demonstrators’ perception toward the police force is largely different from that of other countries. The majority of citizens in Western democracies including the US, from a functional perspective, perceive the police as the agency responsible for crime control, public order maintenance, and citizen safety (Bittner, 1970; Hwang et al., 2005). Also, citizens generally recognize the role of the 12 police, and they tend to respect police authority as a whole. Consequently, police officers enforce the law strictly toward criminals as well as rioters. However, the role of the police in South Korea is complicated by its national history. The perception of the South Korea police force was largely Shaped by its oppressive role at the time of Japanese colonization and by military coup (Pyo, 2001). The police force resulted in serving the ruling powers that wanted to maintain their authority by suppressing the citizens’ opposition (Hwang et al., 2005). Accordingly, the historical image of the South Korean police has been often criticized by the public. Especially, perceptions of demonstrators are problematic. They are less likely to recognize the role of riot police and are not afraid of the police (Kim, 2004). They tend even to make issues through intentional conflicts with the police. Their objective is to gain media attention by attacking riot police officers. Moreover, due to the influence of these historical factors, attitudes of the press toward police enforcement are also unfair (Yu, 2003). If the police tolerate petty violations of demonstrators, they are more likely to blame the police for inaction. Conversely, if a demonstrator is injured during police’s suppression to control public order, they rightly tend to blame the police for oversuppression even if police actions are legitimate. In a survey (N=427 persons) about attitudes toward the press, 23.9% of the respondents answered that the press was fair, while 67.4% of respondents answered that the press was more favorable toward demonstrators than the police (Yu, 2003). Thus, the public appear to tolerate certain forms of public disorder as inherent in a properly functioning democracy. As a result, it is not easy for police officers in South Korea to enforce the law toward violent demonstrators. Arrests which are generally made only in 13 cases involving extreme violence are rare. Social learning theory (Harris, 1977) may apply to explaining demonstrations and riots in South Korea. That is, people who witness the acts, such as throwing bottles at the police and acting in other aggressive ways, may realize that they could also act aggressively to the police. People may perceive such actions as effective ways to settle grievances, and satisfy unmet material needs (Bandura, 1973). Furthermore, types of riots in South Korea are somewhat different from those of general riots. In general, riots are described as some terms, such as race riot, prison riot, student riot, hooliganism, and street fighting. Goode (1992) categorized riots according to the participants’ motives and goals as below: purposive riot, which results from discontent over specific issues and aims to achieve specific goals; symbolic riot, which is meant to express discontent but not to achieve a Specific goal; revelous riot, which occurs after a celebration by a crowd that gets out of hand and is equivalent to the celebration riots; and useless riot, which has no observable motivation or goal. However, in South Korea, most riots occur during the demonstrations. Most riots are classified into purposive riots, but prison riots do not occur. Also, revelous riots and symbolic riots are rare, and useless riots do not occur. Finally, there has been a difference in terms of the degree of danger that rioters cause. Riots in the US. sometimes cause fires, robbery, looting, and attacking general citizens. Many people think of these patterns as dangerous. For example, during the riot in Los Angeles on April 29, 1992, demonstrations turned violent, and full-scale riots had broken out throughout city. Fires, looting, Shootings and beatings raged though the city until May 2. AS a result, 54 people were killed, and more than 2,300 were injured. More 14 than 7,000 fires, looting and attacks on vehicles resulted in an estimated $1 billion in damage (Ong & Hee, 1993). On the contrary, in South Korea, there would be dangerous demonstrations, including throwing rocks and fire bottles, but rioters do not loot, fire, and attack citizens. They attack only police officers because they generally perceive the police force as frontline representatives of government authority (Hwan g et al., 2005; Kim, 2004; Yu, 2003). 15 CHAPTER III: COMPARISON OF PROTEST POLICIN G Protest Policing in South Korea The Constitution and law provides people for freedom of assembly in South Korea, and the Government absolutely respects this right in practice. The Law on Assembly and Demonstrations prohibits only demonstrations that are considered likely to undermine public order. The Law requires that the police Should be notified in advance of demonstrations of all types including political rallies. The police also must notify organizers of demonstrations if they consider an event impermissible under this law. However, most demonstrations routinely are approved. To gain control of the streets during the demonstrations, the KNPA has used a variety of crowd-control measures for a long time, including tear gas, multiple arrests, etc. This section will assess whether community policing or authoritarian policing practices are related to levels of violence by demonstrators. Thus, this section Simply introduces two broad styles of protest policing that the KNPA attempted in the past. In other words, this section deals with authoritarian policing which is based on deterrence theory, such as harsh penalty and increase of police presence (Wright, B. R. E., Caspi, A., Moffitt, T. E., & Paternoster, R., 2004). Also, this section considers community policing which is based on procedural justice theory (Tyler, 2003), such as stopping use of tear gas and participation of outside observers at the demonstrations. Harsh Penalty Policy During the past military government, the KNPA strongly confronted demonstrators without considering procedural justice. AS deterrence theory suggested, 16 the Police Agency used lots of tear gas and arrest for lawbreakers in order to control riots during the demonstrations. Also, the police had reformed laws to introduce harsher penalties on arrested persons at the demonstrations. However, statistically speaking, the number of illegal demonstrations did not remarkably decline. In addition, it had little impact on the number of incidents of throwing fire bottles during the demonstrations. In practice, it seemed that strong confrontation of the police had little impact on preventing violent demonstrations. Increase of Police Presence Policy In addition to harsh penalty, the Police Agency used Simple tactics to control violent demonstrations. These were to increase the number of riot police officers to about 2 or 3 times the number of demonstrators at the demonstrations. The police had focused on enforcement efforts on arresting lawbreakers. Also, police forces were better equipped and better trained to deal with crowds arranged at the demonstrations. However, there was no evidence found that violent demonstrations were reduced. Stop of Use of Tear Gas Policy In 1998 and 1999, illegal demonstrations were extremely reduced. There were many reasons for this, but it was significant change in police policy to control demonstrations. That is, there was significant attempt by the KNPA to control demonstrations and riots in 1998 and 1999. In 1998, high officials at the KNPA declared it necessary to implement fundamental changes to the system, culture, and operation of the police (KNPA, 2002). The intensive reform efforts were initiated as part of the “Police Grand Reform” of 1998. It was directed toward internal issues of the police organization as well as improving the relationship between police and citizens. The focus l7 was toward the changes of the public’s image of the police as fair, which was widely recognized as the primary causes of public discontent with police services (KNPA, 2002). The police initiated a number of strategies to strengthen their relationships with the public. Accordingly, the policy toward demonstrations changed according to these trends. In 1998, the KNPA began to permit the processions on the roads for vehicles which had been prohibited, and in 1999 they stopped the use of tear gas, as a series of “the new protests and demonstrations management policies,” which were implemented in an effort to cope with demonstrations.5 As a result, demonstrators stopped throwing fire bottles (Molotov cocktails) for a long time and the result was a tremendous reduction in illegal demonstrations and the use of fire bottles (see Table 2, Table 3, and Figure 4). A peaceful demonstration culture began to be established more or less successfully, attracting favorable reports from foreign and vernacular press. It seemed that the approach suggested by procedural justice and defiance theories was correct. Table 3: The Number of Tear Gas and Fire Bottles during the Demonstrations. Year 1995 1996 1997 1998 1999 2000 2001 2002 2003 Tear 128,981 213,847 134,405 3,403 0 0 0 0 0 Gas Fire 38,880 80,620 69,160 170 613 749 2,453 457 2,223 Bottle *Source: Korean National Police Agency, 2004. 5 The KNPA has consistently maintained the policy that stops using tear gas during the demonstrations from 1999 up until recently (KNPA, 2006). 18 Figures 4: The Number of Tear Gas and Fire Bottles during the Demonstrations. 250.000 — 200,000 /1\\ 5 150,000 .3 J +Tear Gas 5 + Fire Bottle 2 100,000 \ 50.000 (/\‘\ 0 T I T 1995 1996 1997 1998 1999 2000 2001 2002 2003 Year However, this strategy resulted in unexpected results. There were many physical collisions between police and demonstrators. As time went on, the effects of this strategy began to fade. Some demonstration groups still have staged violent demonstrations and sometimes large-scale demonstrations in downtown areas have caused extreme inconvenience and damage to the ordinary people who were not involved in the demonstrations. AS a result, many injuries happened in the process of the police-citizen collisions. Many police officers and demonstrators were injured during the demonstrations (see Table 4). Finally, two farmers who were demonstrating in front of Congress died in 2005. Table 4: The Number of Police Injury during the Demonstrations. Year 1995 1996 1997 1998 1999 2000 2001 2002 2003 Number of 1 179 1881 1023 161 484 311 304 287 749 police injury *Source: Korean National Police Agency, 2004. 19 Participation of Outside Observers The police have attempted to admit a right of demonstration and emphasize self- control of demonstration organizers. Simultaneously, police leaders have implemented significant reforms based on a community policing model. This strategy is focused on participation of the community. Thus, Since 2004, the KNPA has worked with outside observers who watch and record the actions of the demonstrators and police at the demonstrations. These observers are important because they may deter police brutality and because they see the actual behaviors of demonstrators. The problem of whether the police enforce laws well or the demonstrators really violate laws is always debated in South Korea. Thus, the presence of these observers helps not only keep people safe by discouraging police attacks but also offers exact information that can be useful in the defense of police and protesters or in suing police. This policy of participation of the community to control demonstrations and riots is estimated as a partial success. Protest Policing in the US. In order to evaluate how policing is correlated with protests, there is a need for reviewing policing styles of the US. during the early 19903 to present. This review shows us a variety of police responses used to control demonstrations. This policing can be divided into four Strategies: no intervention of the police in the early-to-mid 1900s, policing practices in the 19608 which were marked by “escalated force,” those in the 19803 and 19905 which have been characterized by “negotiated management,” and “command and control style” Strategies which are based on the “quality of life 20 philosophy of policing.”(Brown, 2006; Della Porta & Reiter, 1998; Earl et al., 2003; McPhail et al., 1998; Nalla, 2006; Schweingruber, 2000; Vitale, 2005). No Intervention The American police forces in the early—to-mid 19003 had little impact on the collective behavior of whether demonstrators became violent or peaceful (Brown, 2006). According to Brown (2006), the police in the early—to-mid 19903 were incapable of controlling any public events because they were uneducated, untrained, and unorganized. In other words, the early police forces had little power to affect whether protests displayed in a peaceful or violent pattern. For example, public events that were fundamentally peaceful were held regardless of police measures to control them. Similarly, the police did nothing to control the collective behaviors which originated in a violent manner or became violent during the demonstrations. A3 for the control of collective violence prior to the 19503, local police forces had little positive impact and in many cases made the Situations worse (Brown, 2006). Similarly, municipal police forces toward the urban riots were poorly prepared to handle the violence and often made no effort to repress the riots. Thus, Brown (2006) explained that in many cases the police simply waited for the riots to subside. Escalated Force Style In the 19503, 19603, and 19703, the police operated under a philosophy of escalated force. The escalated force model of policing represented a major Shift away from the early policing, and instead used far more aggressive tactics. Thus, any Show of violence by the protesters was met with overwhelming police force in return (Brown, 2006; McPhail et al., 1998; Schweingruber, 2000). That is, if protesters used violent 21 against the police, then the police would go one step beyond the resistance of the protesters (McPhail et a1, 1998; Schweingruber, 2000; Vitale, 2005). Even in cases in which were the peaceful demonstrators used non-violent tactics and peacefully attempted to exercise their First Amendment rights, the police would use violent tactics to break up the demonstrations (Della Porta & Reiter, 1998; McPhail et al., 1998). For example, McPhail et a1. (1998) suggested that well-known demonstrations in which the police used the escalated force approach include the Birmingham civil rights campaign ( 1963), the Chicago Democratic Convention (1968), and the confrontation between student protesters and National Guard soldiers at Kent State University (1970). Although there were a number of issues, including public discontent with combative police tactics, police departments utilized the escalated force model of crowd control which became the norm during the 19503 and 19603. Negotiated Management Model During the late 20th century, there was a fundamental change in the style of protest policing from the 19603 to the 19803 (Brown, 2006; McCarthy & McPhail, 1998; McPhail et al., 1998; Nalla, 2006). In contrast to the escalated force style, the new doctrine of negotiated management was based on greater cooperation between police and demonstrators. Many police agencies that use a negotiated management model emphasize the right of people to protest and centers on cooperation between police forces and demonstrators in order to reduce the potential for violence (McPhail et al., 1998). The new approach emphasizes the following factors: the protection of free speech rights, toleration of community disruption, ongoing communication between police and demonstrators, avoidance of arrests, and limitation of the use of force to situations where 22 violence is occurring (Schweingruber, 2000). Thus, under the negotiated management system, the police negotiate with demonstrators before the demonstration so that demonstrators can exercise their First Amendment rights with minimal conflict with the police. For example, Washington, DC. police follows the negotiated management style while dealing with the thousands of demonstrations that occur each year (Della Porta & Reiter, 1998; McPhail et al., 1998). This new philosophy is now in place in much of the US. and Europe (Brown, 2006; Della Porta & Reiter, 1998; McPhail et al., 1998; Nalla, 2006; Waddington, 1994). Many Studies suggest that the negotiated management model of protest policing was employed most and the implement of this approach was successful. Fisher (2001) stated that an “air of cooperation” can be established between the police and demonstrators, and the police can be reasonably confident that the protest leaders will handle disruptive members of their group at their own level. However, the negotiated management model of protest policing has some weaknesses. Small, more confrontational grassroots groups (Vitale, 2005) and newly self- identified political actors (Tilly, 2000) emerged in the late 19803 and early 19903, and they often employed anarchist principles, such as sit-ins, road blockades, traffic stoppages, and lockdowns during their demonstrations (Button et al., 2002; Nalla, 2006; Vitale, 2005). Their innovative actions disrupted existing spatial routines (Nalla, 2006; Noakes et al., 2005). For example, “the disruptive potential of such transgressive contentions was demonstrated at the Battle in Seattle6 during the 1999 meetings of the 6 The Seattle Police Department negotiated numerous agreements with mainstream political groups who would be to protest the polities of WTO. However, the police failed to reach agreement with the grassroots wing of the global justice movement. Consequently, the police lost control of area, and the WTO was forced to cancel the opening day of its 1999 meeting (See Nalla, 2006; Noakes et al., 2005: 240). 23 WTO Conference” (Noakes et al., 2005: 240). Thus, the police felt the necessity for new strategies for controlling protests more effectively. Command and Control Style The most recent police strategy to emerge i3 command and control style, which is based on the “quality of life philosophy of policing,” and relies on zero tolerance. This policing method controls low-level disorder through the use of zero tolerance enforcement of minor crimes (Vitale, 2005). Vitale (2005) explained that the N YPD employed this strategy on February 15, 2003.7 This style emphasizes police’s attempt to micro-manage all important aspects of the demonstration in order to eliminate any illegal activity during the demonstrations (Nalla, 2006; Vitale, 2005). Under this policing style, because the police set clear and strict guidelines on acceptable behavior with demonstration organizers with very little negotiation, this approach is distinguished from negotiated management model (Vitale, 2005). This approach is often maintained through the use of force, even against peaceful protests. Thus, some have interpreted incorrectly that this is a clear shift from a negotiated management model to an escalated force style. Some viewed the N YPD’S response as a “paramilitary response” to protest policing (Jefferson, 1990). However, Vitale (2005) argued that this approach does not represent a return to escalated force because it attempts to avoid the use of force through careful management of the demonstration. The command and control style strategy is vulnerable when confronted by many demonstrators who directly resist police management in controlling demonstrations; however, this style of policing has been effective in reducing disorder at 7 On Feb 15, 2003, anti-war organizers planned a march in midtown Manhattan in New York to go fast the United Nations and end at Central Park, but NYPD denied a march permit, deployed thousands of police officers, charged crowds with horses, and arrested hundreds of protesters (see Nalla, 2006; Vitale, 2005). 24 most political demonstrations (Vitale, 2005). Vitale (2005) explained that this policing style became the dominant philosophy of policing in New York City. Summary of Protest Policing As stated previously, the historical review on protest policing of South Korea and Western countries, including the US, helps us to understand how the different policing styles affect the demonstrations. In sum, there were similarities between South Korea and the US. in terms of protest policing styles. That is, historically, the police in the two countries attempted to employ two reciprocal styles with strong and weaker forces to control demonstrations effectively. There has been a widespread general trend in recent years toward less coercive styles of protest policing in South Korea and the US (Della Porta & Reiter, 1998; McPhail et al., 1998; Nalla, 2006; Waddington, 1994). The police are also trying to find effective models for controlling demonstrations. However, there were differences between the two countries in terms of historical background and its effects. As discussed earlier, various protest policing styles in South Korea were implemented according to historical background. Under the last military government, protest policing mainly focused on deterrence theories. However, harsh penalties and the increase of police presence policies, which were based on these deterrence theories, appeared to have little impact on preventing riots in South Korea. Also, reforming related-laws might have had an impact, but there is little evidence of long term impact. Rather, these empirical results suggested that policies that stopped the use of tear gas and increased fair law enforcement, including the participation of outside 25 observers at the demonstrations, were much more effective than those based on deterrence theory. On the other hand, the first policing in the US. used the escalated force model, in which the militancy of protesters was met by increased militancy of the police; it had even been used against the peaceful demonstrators in the 19503 and early 19603. Then, policing style moved on to the negotiated management model which centers on relations between police and demonstrators (McPhail et al., 1998; Nalla, 2006; Reicher et al., 2004; Vitale, 2005). Finally, as a new strategy based on the philosophy of ‘quality of life,’ the command and control style strategy showed that the police recognize the effects of the policing and are beginning to search for alternatives to control demonstrations. 26 CHAPTER IV: LITERATURE REVIEW Effects of Characteristics of Demonstrations A large concern in policing demonstrations is the relationship between the characteristics of demonstrations and policing. Many studies focused on examining the relationship between characteristics of protest events and police actions, and as discussed earlier, the results of these studies showed that the characteristics of protest have a significant influence on police actions (Della Porta & Reiter, 1998; Earl et al., 2003; Waddington, 1998), although little research directly examined the effects of the characteristics of protest events on the violent behavior of demonstrators. However, there are empirical evidences that the characteristics of demonstrations affect the behaviors of demonstrators (Kim, 2004; KNPA, 2003; Lee, 2006). That is, the police judge the possibility of demonstrator violence by examining the characteristics of demonstrations (i.g., organizers, purposes, and Size) in advance and then set police measures to control demonstrations. Reicher et a1. (2004) emphasized that police officers need to concentrate on understanding the collective identities of the demonstrators to develop new guidelines for public order policing. Thus, the characteristics of demonstrations will be one primary categories discussed in this study. Organizers of Demonstrations Protest policing is affected by organizers of the demonstrations. Noakes et a1. (2005) argued that a protest group’s profile affects how the police perceive them. Also, there are many studies which Show that protest events in which subordinate groups and social movement organizations participated in were more likely to draw police action 27 (Della Porta & Reiter, .1998; Earl et al., 2003; McAdam 1982; McCarthy, 1997; Waddington, 1998). In their analysis of riots in the US. during 1968-1973, Earl et a1. (2003) suggested that the best predictor of police presence at such riots has been presence of sub-groups who posed significant threats to mar the social protest. Furthermore, a few case studies suggested that severe repression is more likely when protests are primarily composed of “socially marginalized participants” (Gamson, 1990; McAdam,1988; Stockdill, 1996), while more middle-class groups are more likely to be perceived as low risk (Della Porta & Reiter, 1998; Noakes et al., 2005; Waddington 1998). McAdam (1988) argued that affluent young whites could more safely engage in threatening protest activities than could blacks in Mississippi. That is, the combination of threatening tactics and black protesters most likely would have led to extreme action by the Mississippi police. Stockdill (1996) found that protest actions predominately undertaken by people of color were more likely to have been repressed, and repressed vigorously, compared with similar protests by whites. In addition, Earl et a1. (2003) found that college student presence reduces the probability of “Calling All Cars”.8 Protests including college students are at relatively lower risk for police presence and action than they would be otherwise. This result contradicts hypothesis suggesting that college student presence increases the probability of more severe forms of police action. Waddington (1998) described young protesters as one of “bad” protesters that are more likely to transgress. In South Korea, during the past military government rule, students and religious organizations were primary groups of demonstrations and riots that requested democratic 8 This approach means police officers combine all the tactics they have available. For example, they use force, make arrests, and frequently use weapons, such as tear gas (See Earl et al., 2003). 28 government (Kim, 2004). However, recently, Civil Society Organizations make up the highest rate of all demonstrations and they protest various social, economic, and military purposes. Thus, their demonstrations are increasing every year in South Korea (Jo, 2006). History of Violence There has been little research showing that the history of violent behavior from demonstrators begets subsequent violence. However, Noakes et al. (2005) argued that the police read certain characteristics as possible indications of transgressiveness. That is, those who have established a reputation for disruptive behaviors and “bad” protesters, which include professional protesters, are more likely to violate the laws (Waddington, 1999). Most demonstrations in South Korea are routinely approved according to the Law on Assembly and Demonstrations (Kim, 2004; Yu, 2003). However, the KNPA has generally judged whether a demonstration turns violent according to organizer’s prior violence (Kim, 2004; Yu, 2003). It means that prior violence of organizers may affect demonstrations again. Purposes of Demonstrations The antagonistic attitudes of demonstrators and purposes of demonstrations are important variables. Many research pointed out that protest groups that use confrontational tactics or those pursuing politically radical goals are more likely to be met with police violence (Earl et al., 2003; McAdam, 1982; Waddington, 1998). McAdam (1982) found that groups using non institutional and confrontational tactics face greater repression, and argued that groups pursuing revolutionary or radical goals will be repressed more strongly than will moderate groups. This is similar to Tilly’s (1978) argument that “accepted groups” with small goals will be repressed less and to Wisler 29 and Giugni’s (1999) finding that counter-cultural groups are repressed more often. Similarly, Waddington (1998) found that groups who are seen as pursuing abstract, diffuse, or radical goals, employing innovative tactics that they do not reveal in advance, and are unwilling to reach agreement with police prior to a demonstration, are deemed likely to challenge police control and are more likely to transgress. Police respond differently to demonstrations based on their assessment of the risk posed by protesters (Waddington, 1998). Earl et a1. (2003) argued that extreme forms of police action are also triggered by threatening characteristics of events. According to Earl et al. (2003), police actions will be the results of the threat posed by protesting groups. Protest policing in the US. is also organized around Shifting philosophies about the nature of the threat posed by demonstrations (Earl et al., 2003). Furthermore, late significant changes in the nature of protesters have posed new challenges to policing organizations (Button et al., 2002). Recently, social movements have occurred across national borders with new contexts such as global issues and are involved with wider national organizations (Button et al., 2002; McCarthy, 1997; Nalla, 2006; O’Neill, 2004). O’Neill (2004) argued that transnational protests have become a hallmark of global activism. These protesters are not interested in promoting specific group interests, but emphasize post-material values, such as social justice and welfare, peace, equity, and equality of the largest groups of people around the world (Nalla, 2006). Also, tactics of protesters have a diverse range of protest activities. For example, “the militant environmental activists (MEA) are generally very active in pursuing policy change and use illegal and unlawful tactics to achieve their objectives” (Button et al., 2002:22). Also, because this is their main occupation, they are called “protesting 30 professionals”. Thus, these characteristics and tactics of protesters have required the police to deve10p efficient and effective tactics to manage them. Size of Demonstrations There is a great deal of research about the relationship between police action and protest size, however, there is little research dealing with the fact that violent demonstrations are more likely to take place when protest size is large. According to Earl et a1. (2003), as protest size rises, both police action and uses of police force become increasingly likely because the police may find it difficult to make arrests in large crowds or to interact with large crowds without inciting protesters. Thus, Earl et al. (2003) argued that the larger the demonstration in size, the greater likelihood the police would use violence. Similarly, Tilly (1978) argued that the larger the size of the protest, the more threatening the event and thus, the higher the likelihood of repression. That is, the police recognize that controlling a large, angry crowd is much more difficult than controlling a small, angry crowd. However, because arrests and significant police intervention often incites anger on the part of protesters, the police may prefer to avoid making arrests or significantly interacting with the protesters until absolutely necessary. Ironically, increasing protest size may encourage police to do nothing up to the point where intervention is so necessary that only serious uses of coercion are seen as possible alternatives (Earl et al., 2003). In sum, the presence of subordinate groups increases the probability of police presence, while the presence of college students dampens it. Large protests, confrontational protests, and protests that endorse radical goals are more likely to draw police presence, and event Size and the use of confrontational tactics escalate police 31 response once police arrived. According to Earl et a1. (2003), situational threats like the use of confrontational tactics and protest size are more important to the police present at the event than other, more diffuse threats, such as advocating goals. Effects of Police Measures As discussed earlier, research shows that a change in policing styles has an effect on the control of demonstrations. To understand in detail the effect of policing, this section focuses on the relationship between the detailed contents of protest policing (e. g., size of police, jurisdiction, arrangement of police equipment, permission of parade, and deployment of female officers) and demonstrations. Jurisdiction Jurisdiction, which is related to police capacity, is a primary variable when measuring the effects of policing styles on the demonstrations. Boudreau (2001) suggested that the overall levels of repressive action are decided by the state’s capacity. In general, the large police departments with access to more resources will have a greater capacity to repress than will smaller forces with lower budgets and fewer officers (Boudreau, 2001; Earl et al., 2003). Earl et a1. (2003) showed that the police capacity affects the types of police action that depends on resources and facilities. Thus, police departments located within jurisdictions of high-capacity are more likely to monitor protests. Also, departments with more resources for training and higher staffing levels are less likely to resort to violence when controlling demonstrations (Earl et al., 2003). For example, Vitale (2005) assessed that since the NYPD has such a large capacity, having 32 largest police force in the country, they effectively controlled a large demonstration on February 15 in 2003. Policing can also be affected by internal characteristics of the police department, which may be related to jurisdiction in a sense because it affects trends in repression (Earl et al., 2003; Stockdill, 1996). Stockdill (1996) argued that police forces with historically high rates of brutality may be more repressive toward protesters. Similarly, Earl et al. (2003) concluded that police forces that have a history of application of violence and brutality in dealing with police functions are more likely to respond violently when policing protests. Police Size Prior studies have examined the effects on presence or size of police as a measure of police action toward demonstrations. Wanderer (1969) and Spilerman (1976) found little correlation between riot severity and police force size. Also, there have been some views that the presence of more police leads to the effect of increasing violence because violators may feel that they are being “oppressed” (Kim, 2004; Lieske, 1978; Reicher et al., 2004). Myers and Li (2001) also argued that having higher numbers of police produces the precipitant for rioting and the repressive activities of police are not influential in rioters’ decisions to take to the streets. However, most prior studies Show positive associations between the level of demonstrations and the relative size of the police force (Brown, 2006; Carter, 1987; Dotson, 1974; Earl et al., 2003; Gurr, 1970; Hibbs, 1973; Koopmans, 1993; Morgan & Clark, 1973; Nalla, 2006; Tilly, 1978; Vitale, 2005). This relationship stresses the deterrent effect of relatively large police forces. That is, a very large police force can 33 minimize violence by overwhelming demonstrators psychologically and physically (Carter, 1987). This relationship also finds support in the observations of many experienced police administrators (Kemer, 1968). In addition, Gurr (1970: 239) argued that “the number of police forces is a major determinant of the extent to which dissidents feel threatened by force.” Hibbs (1973) found that the police force size is a strong predictor of negative sanctions. He observed that “a middle—sized police force is sufficient to antagonize significant segments of the population, but inadequate to effectively suppress violent behaviors of the discontent” (p. 83). Tilly (1978) argued that more police result in few riots because greater numbers of police result in a greater ability to repress rioting. Carter (1987) conducted an analysis about the police force Size and the severity of black rioting in the 19603. He found a curvilinear relationship between police force and rioting: “relative police force tended to be positively correlated with the severity of city rioting up to a point” (Carter, 1987: 610). According to Carter (1987), this relationship combines the positive and negative patterns: small police forces are limited in the damage they can inflict, medium police forces can inflict more damage, and large police forces can minimize violence by overwhelming the opposition both psychologically and physically. In other words, he argued that small to medium sized police presence often enraged rioters and generated additional violence, but this was true only up to a point. A large number of police relative to the number of rioters resulted in lower levels of violence of rioters because the police were better able to impede the black population of an area from forming into “uncontrollable crowds” (Carter, 1987). 34 Similarly, Earl et a1. (2003) tested hypotheses about police presence and police action at protests in New York State between 1968 and 1973. They employed police size to measure threat as a major predictor of police presence, and found that the police attend large protests at a higher rate in order to monitor large protests. Most recently, Vitale (2005) stated that the overwhelming presence of police force is intended to have a “shock and awe” value which deters unlawful, unperrnitted, and violent behavior of demonstrators. Thus, he argued that one reason the NYPD was able to control crowds so effectively is the sheer size of the NYPD. Also, as further evidence that the police size matters, he suggested the cases of urban riots (e.g., 1943 riots in Detroit, 1980 riots in Miami) that the National Guard was called out because police forces were simply not large enough to handle the situation (Nalla, 2006; Vitale, 2005). In addition, Brown (2006) suggested that the Size of the police presence matters and that the larger the number of police at a protest, the better able they are to control the crowd. He argued that police tactics such as increasing police presence seem to be effective at organized demonstrations. Police Equipment A few studies showed that the use of police equipment such as barricades affects the demonstrations (Earl et al., 2003; Nalla, 2006; Waddington, 1994). In many demonstration cases, the police try to prevent disorder by erecting barricades. In their study about various police actions at the protest events in New York between 1968 and 1973, Earl et a1. (2003) found that “preventive and legalistic approaches,” which maintain the public order by using barricades, making arrests, or combining the use of barricades and arrests, predominated. The use of police equipment is especially effective under the 35 negotiated management model of protests where the police work with protest organizers (Nalla, 2006; Waddington, 1994). Waddington (1994) argued that the steel barricades, which he calls “moving wall of steel,” are successful measures of the police keeping the protest orderly in a traffic-free zone. Police Permit for Parade The police’s permission to have a parade affects demonstrations in many ways (Della Porta & Reiter, 1998; Ericson & Doyle, 1999; N alla, 2006; Noakes et al., 2005; Vitale, 2005). Under the negotiated management style of policing, many demonstrations with parades proceeded peacefully and orderly without any disruptions or disorders (Della Porta & Reiter, 1998; Noakes et al., 2005). However, in some demonstrations, permission to have a parade aggravated the unexpected violence or illegal behaviors of demonstrators. Thus, the parade permit reflects the police department’s desire to prevent demonstrations from interfering with the normal functioning of the city (Nalla, 2006). The police deny or restrict a parade because they want to guarantee the quality of life of general citizens. For example, the N YPD denied a parade permit for the demonstration on February 15, 2003 and heavily restricted access to the demonstration area (Vitale, 2005). Female Officers There has been a lot research done on dealing with the effects of female officers in policework (Alpert & Dunham, 1999; Bloch & Anderson, 1974; Block, Anderson, & Gervais, 1973; Grant, 2000; Hale, 1992; Herbert, 2001; Hoffman & Hickey, 2005; Kim, 2004; KNPA, 2000; Koeing, 1978; Kwon, 2006; Lonsway, 2001; Martin, 1997; Paoline & Terrill, 2004; Sherman, 1973; Worden, 1995). The various studies on attitudes toward policewomen are critically reviewed and evaluated, but the studies are not congruent with each other; some research was done on specific gender differences, while other research contradicts the differences between male officers and female officers with respect to citizen’s perception and the use of force. However, there are relatively few studies comparing the effects of policing by female and male police officers on demonstrations. The proportion of female police officers is increasing in all areas of police systems. Female officers are currently assigned to non-patrol duties as well as the patrol divisions. This increase in female police officers raises the question of whether female police officers have a positive impact on crimes at the same rate as male police officers. Many studies on female police officers has noted that their policing style is different from males, based on the traditional assumption that female officers are often perceived as unwilling to display their coercive authority compared to male officers (Lonsway, 2001; Martin, 1997; Paoline & Terrill, 2004). The studies that approve of gender differences contain two inverse statements. The first is that male officers are better than female officers. Male officers do not think that female officers do have a proper role in policing because they are perceived as lacking the physical and emotional strength needed to perform the basic functions of police work (Grant, 2000; Hale, 1992; Herbert, 2001). Thus, female police officers must handle traditional duties, such as searching female prisoners, safe-guarding children, doing clerical work, or working in special units in the detective division. With respect to attitudes toward female officers, citizens also feel that males were better able to handle violent situations and a male/male team was better than a male/female team in handling dangerous, angry, or upset people (Block et. al., 1973). In addition, the Block and Anderson’s survey of citizens’ attitudes in the Washington, DC. area in 1973 showed 37 that citizens approved of equal opportunity for women in the police force, but they were moderately skeptical about the policewoman’s ability to handle violent situations. The second formulation of gender difference emphasizes the positive effects of policewomen in policing (Bloch & Anderson, 1974; Block et al., 1973; Koeing, 1978; Lonsway, 2001; Milton, 1972; Sherman, 1973). Most of these studies have concluded that attitudes toward policewomen in law enforcement have changed markedly since women first entered into the police force. These studies have shown that women may even have a calming effect in potentially violent situations (Sherman, 1973) and policewomen can be just as effective as policemen with proper training (Koeing, 1978). This calming effect is based on as following characteristics: women are associated with sympathy, understanding and warmth, and women officers would arouse less antagonism, fear and violence than male officers (Sherman, 1973, 1975). Also, while it may be heroic and masculine to attack a male officer, it is cowardly to attack a woman, even if She is an officer. This fact may help to reduce the violence towards female officers (Milton, 1972). Women seem to have a greater tolerance of ambiguity and are less authoritarian. Female officers felt that women were more likely to be calm and cool in tough situations and more likely to be decisive, observant, emotionally stable, intelligent, and understanding (Block et al., 1973). According to Koeing (1978), in a study of policewomen in St. Louis County, women were more sensitive to the complainant’s personal needs and citizens felt policewomen handled their service calls and domestic quarrels better than men. In addition, police researchers have noted that women are more skilled at de-escalating violence in encounters with citizens, largely as a result of their communicative verbal Skills (Bloch & Anderson, 1974; Lonsway, 2001). Thus, these 38 gender differences might be welcomed by some police administrators who endorse contemporary community policing philosophies. However, many of the current studies about use of force conclude that the personal characteristics of police officers do not have a substantial impact on the use of force (Alpert & Dunham, 1999; Hoffman & Hickey, 2005; Paoline & Terrill, 2004; Worden, 1995). Alpert and Dunham (1999) reported no statistically significant differences in the level of force used by male and female officers in a study of 546 incidents. Paoline and Terrill (2004) examined the types and amount of coercive force that females exhibit during citizen encounters. This study failed to find substantial differences in the use of force between male and female officers. Other studies found no statistically significant differences between female and male officers (Hoffman & Hickey, 2005; Paoline & Terrill, 2004). Both male and female officers used coercion in similar proportions, and both tended to use verbal force at higher rates when compared to physical force. Also, according to Hoffman and Hickey (2005), the overall rate of force per one hundred arrests was not significantly different for female and male officers, although female officers less frequently used a weapon. With respect to specific types of force, there were no significant differences between female and male officers in the use of unarmed physical force, firearms, flashlights, baton/telescoping batons, other hard objects, or sprays. With respect to injury, suspects arrested by female officers had a lower rate of any injury than persons arrested by male officers, but there were no statistically significant differences between female and male officers in the rate of injury to suspects that resulted in treatment at a hospital (Hoffman & Hickey, 2005). 39 On the other hand, there are empirical evidences which apply to female officer’s ability to control demonstrations in South Korea. In 1999, the KNPA, for the first time, deployed policewomen to demonstration locations as a series of “new protest policing” to control demonstrations. As a result, many citizens positively evaluated that this new policing tactic changed the culture of demonstrations that was embodied by riot police and tear gas, and contributed to the reduction of the levels of violent behavior of demonstrators (Kim, 2004; KNPA, 2001; Kwon, 2006; Lee, 2006). Also, it attracted favorable reports from internal and foreign press. The press described this new policing tactic as the “lipstick line”, which involves unarming and deploying of all female police officers to control protest marches. The “lipstick line” represents the gentle, non—violent face of the police, in contrast to all males, armed and black—clad riot police; it creates a new image of the police (Kwon, 2006). In this respect, policewomen have made considerable strides in protest policing in South Korea (Kim, 2004; KNPA, 2001). Effects of Outside Observers The final concern is whether the presence of outside organizations, such as legal observers, general citizens, and the press, has an effect on preventing demonstrators from committing violent behaviors. Legal observers in the US. are typically, but not exclusively, law students, legal workers (such as paralegals), and lawyers who may or may not be licensed locally. According to the National Lawyers Guild (2004), the primary goal of a legal observer is to watch and record the actions of the police at the demonstration locations. That is, they watch the actions of all law enforcement officers and particularly note violations of people’s rights. Observers are important because they 40 may deter police brutality and because they collect information that may be helpful in later court proceedings (National Lawyers Guild, 2004). In addition, the information legal observers collect can be useful in jail support of arrestees, criminal defense of protesters, and in lawsuits against police or government agencies. The role of an observer tends to focus on discouraging the police brutality or police repression rather than monitoring or deterring violent behavior of demonstrators. In general, the presence of these observers may make police officers behave differently when they deal with demonstrators. The police may show the limited discretion in controlling demonstrations because of the presence of observers. This resembles the “Hawthorne Effect”, which refers to a phenomenon of observing workers’ behavior or their performance (Rainey, 2003).9 The presence of outside observers may be a burden on police officers. In particular, police officers can be sensitive to the presence of outside observers. The officers may behave differently when they deal with demonstrators as if the officers get more sensitive to the observation. This can effect how the officers respond to situations even if over time the officers get used to the situation. The presence of an observer in the US. may have more of an impact on the behavior of the police than that of demonstrators. Similarly, the National Lawyers Guild today says that they have a long and proud history of fighting police abuse and their role poses the biggest challenge to them (James, 2001). However, there are some differences in the primary role of an observer in South Korea. According to the KNPA (2005), the police have worked with general citizens to serve as an observer who watches the actions of the demonstrators and police at the demonstration scene. Thus, contrary to the role of legal observers in the US, these 9 This name comes from some early work ( 1927-1932) on organizational measurement done at the Western Electric plant in Hawthorne, Ill. The Hawthorne effect appears when people know they are being measured. they modify their behaviors (Rainey, 2003). 41 observers in South Korea play an important role in monitoring the actual behaviors of demonstrators rather than deterring police brutality (Kim, 2004; Yu, 2003). However, there is little research on whether or not the presence of outside observers has a substantial impact on preventing riots. 42 Research Questions The primary research question of this study is whether various measures of policing used to control demonstrations have an influence on the level of violent behavior of demonstrators. To accomplish this objective, this study focuses on three hypotheses that are drawn from the literature review and available data. These hypotheses reflect cultural and historical factors in South Korea and were built upon in the process of various attempts of the police to control demonstrations. First, this study assumes that the characteristics of demonstrations, such as organizers of demonstrations, a violent history of demonstrators, and purposes of demonstrations, are more likely to have an influence on the violent behavior of the demonstrations. That is, regardless of police measures, violent behavior of demonstrators mainly occurs due to the characteristics of demonstrations. For example, if any demonstration is organized by an organizer with a history of violence, or if any demonstration is associated with global purposes such as anti-FT A, a violent demonstration is more likely to take place. Many Korean people agree to this hypothesis even if there is no evidence to prove this fact. This means that the characteristics of demonstrations are strong predictors in explaining the violent behavior of demonstrators during the demonstrations. The second hypothesis states that a variety of police measures to control demonstrations have significant effect on violence by demonstrators. In order to assess this in detail, this study will use police measures, such as the ratio of demonstrators to police officers, the close deployment of the police, the arrangement of police equipment, parade permit, and the deployment of female officers. As stated previously, many prior 43 studies suggested that police actions have an effect on the behavior of citizens involved in demonstrations. However, the results of this study may be different from previous studies because demonstrations and riots in South Korea reflect unique historical factors. The final hypothesis is that outside observers at the demonstration scene have a Significant influence on the behavior of demonstrators. A3 Brown (2006) argued, the police alone can not control all incidents of collective behavior or single-handedly control all situations of collective violence. It is clear that the police alone can not always prevent a peaceful assembly from transforming into a melee or single-handedly control riots. Therefore, the police should be willing to cooperatively work with the citizenry. In this point, the role of outside observers at the demonstrations is important in South Korea. Figure 5: Hypotheses for Relationship between Independent Variables and Dependent Variable. Characteristics of demonstrations 0 Organizers of demonstrations - Joint—group/Sub-group 0 History of violence 0 Purposes of demonstrations - Number of demonstrators Police Measures - Jurisdiction - Ratio of Demonstrators to Police Officers Violent Behavior of 0 Distance between Police and ' Demonstrators Demonstrators ° Arrangement of Police Equipment 0 Presence of Parade 0 Deployment of Female Officers Presence of Outside Observers CHAPTER V: DATA AND METHODS Data The purpose of this study is to examine the relationship between police actions to control mass demonstrations and the peaceful or violent behavior of the demonstrators in South Korea. In order to accomplish this objective, this study uses existing police records on mass demonstrations from the KNPA as well as data from the agency’s Annual Report. The data was collected and reported by police officers in South Korea. The sample of cases includes all demonstrations with 1,000 or more demonstrators occurring over a two and one—half year period (1/1/04-6/30/06). In general, researchers have heavily relied upon media data, such as newspapers and electronic news reports because they are often the common source of data on protest events in many different places and times (Earl et al., 2003; Olzak, 1989). However, since the use of the media reports has some problems with the adequacy of protest event data (McCarthy etal., 1996; Olzak, 1989; Snyder & Kelly, 1977), this study is based on police records on mass demonstrations. First, the reason that this study uses the police reports is that the reports are an important source of evidence for researchers who describe the frequency, form, Size, duration, and intensity of demonstrations. The police agencies manage to keep detailed records of mass demonstrations to set up effective counterplans as well as to notify the public of exact situations (Yu, 2003). Thus, the police reports have much information about policing and the behaviors of demonstrators. They record the situations of almost all protests including the characteristics of the protest, such as the organization 45 sponsoring the protest, purpose, time, location, estimated number of participants, and reports of illegal or violent incidents involved (i. g., rocks thrown, use of fire bottles, sticks, iron sticks, etc.) (McCarthy et al., 1996). Also, the data record characteristics of the police response, such as jurisdiction, number of riot police officers, deployment of female officers, distance between police and demonstrators, whether permission was granted for the demonstration. Furthermore, the data record injuries suffered by the police or civilians, numbers of arrests, and whether there were outside observers. For these reasons, the data of this study are based on police reports on mass demonstrations drawn from the KNPA and try to use a relatively objective sample of mass demonstrations. Unlike police reports, the study based on the media data has limitations in describing and understanding the policing and the demonstrations. Olzak (1989) stated that media biases include selection bias, description bias, and researcher bias. Due to these sources of bias, researchers may not examine the exact characteristics of the population of protest events (Olzak, 1989). Similarly, the media reports on mass demonstrations in South Korea have the problem of selection bias. Most press accounts mainly pay attention to only illegal and violent demonstrations, ignoring peaceful demonstrations that involve political and social issues with many participants. Second, this study deals with only demonstrations having 1,000 or more participants. Although the largest proportion of demonstrations falls into the small Size category, most violent behaviors take place during mass demonstrations with over 1,000 participants. As a result, a variety of police measures to control demonstrations are concentrated on the large size demonstrations. Small size protests tend to be impacted by the threatening power of the police, such as the number of police officers, but because 46 most small size demonstrations are peacefully finished, this study excludes them. In this point, this study using mass demonstrations can effectively measure the relationship between police measures and the behaviors of demonstrators. The unit of analysis of this study is further specified as a demonstration with over 1,000 participants. Thus, the analyses are based on incidents of mass demonstrations in South Korea. Third, the data used in this study are not limited to the special areas of some cities, but represents all jurisdictions of police stations in all cities of South Korea. Mass demonstrations with over 1,000 participants and violent demonstrations not only take place in the metropolitan areas of Seoul, the capital of South Korea, but also occur in other cities. Since South Korea has a national police force system, the police station usually uses consistent measures to control demonstrations. Thus, when this study explains the behaviors of demonstrators, it is more reasonable to examine it by using demonstrations in all cities. Fourth, this study focuses on the 2004 to 2006 time period for two reasons. One is to avoid the effect of the policy stopping the use of “tear gas” at the demonstration scenes. This policy stopped using tear gas as a part of a “new protests and demonstrations management” that was enforced in 1999 (Kim, 2004; KNPA, 2001; Lee, 2006; Yu, 2003). As a result, illegal and violent demonstrations reduced tremendously as shown earlier (see Table 2). Up to now, this policing is consistently maintained as a primary tactic of the police. Thus, this study does not consider this effect of stopping the use of “tear gas.” Another reason for focusing on 2004-2006 is that the new policing making outside observers participate in the demonstration has been recently conducted on a full scale. Although many people recognize it as a new, effective policing, there is no study that 47 examines the effect of outside observers at the demonstrations on the behavior of demonstrators. Thus, this study would be the first to examine the relationship between the two variables. Fifth, because this study focuses on proactive measures of the police to control demonstrations, the reactive measures of the police, such as the use of physical force (i.e., hand-to-hand combat), the use of weapons (i.e., tear gas), and the use of dispersion were eliminated from the analysis. Finally, although the police reports provide a fairly comprehensive portrait of demonstrations in South Korea, there is a potential difficulty with official records of such demonstrations. The potential difficulty with our data sources is that all police reports are not permitted to be published. Therefore, because official constraints prevented us from copying the original data, this study followed procedures that enabled on-site coding of relevant information embedded in the records. Measures to be considered include the number of police, their use of police equipment, the distance between police and demonstrators, the deployment of female officers, as well as the presence of outside observers. However, this study does not deal with detailed records measuring policing, such as the exact number of female officers, the number of outside observers, and exact distance between police and demonstrators at the demonstration scene. Sample There are 943 cases of demonstrations which meet the above criteria. However, 22 cases where the dependent variable was not included were excluded from this study, resulting in 921 cases for the study. More Specifically, the number of cases by year is 417 48 cases (45.3%) in 2004, 356 cases (38.7%) in 2005, 148 cases (16.1%) in 2006, respectively. The number of samples by season is 265 cases (28.8%) in Spring (March to May), 278 cases (30.2%) in summer (June to August), 221 cases (24%) in fall (September to November), and 157 cases (17%) in winter (December to February), respectively. The number of samples by day of the week is 97 cases (10.5%) on Sunday, 86 cases (9.3%) on Monday, 107 cases (11.6%) on Tuesday, 131 cases (14.2%) on Wednesday, 164 cases (17.8%) on Tursday, 180 cases (19.5%) on Friday, and 156 cases (16.5%) on Saturday. In addition, 488 cases (53%) of the sample took place on the weekdays (Monday to Thursday), while 433 cases (47%) occurred on the weekend (Friday to Sunday). Dependent Variable The dependent variable in this study is whether or not the demonstrators committed any violent behaviors at the scene of demonstrations, and this variable is based on police reports of each demonstration. It includes a dichotomous response set of “yes” or “no” describing violent behavior of demonstrators, and these outcomes are coded “0” and “1”: a ‘no violence’ coded as 0 and an ‘at least one act of violence’ coded as 1. Although many demonstrations in South Korea include petty illegal behaviors, these cases are excluded from this study because it is hard to specify the behaviors of demonstrators and police reports do not include their petty behaviors in detail. Therefore, the measures of violent behaviors of the demonstrators are based on behaviors, such as throwing/using dangerous materials (i.e., rocks, fire bottles, wood sticks, and iron sticks), 49 or physically colliding with the police. Thus, the violent behavior of demonstrators is coded into these criteria. Of course, the data can be measured according to the level of violent behavior of demonstrators and be placed along a “continuum” ranging from least to most severe behavior. That is, violent behavior of demonstrators can be ranked in the following manner: “no involvement in violence”, “moderate involvement in violence” (i.e., physical collision with the police), and “serious involvement .in violence” (i.e., use of rocks, fire bottles, sticks and iron sticks, and use of all these things), However, because the number of each type of violence is limited, the typology of this continuum is excluded from this study. Therefore, this study includes only no involvement or involvement in violence by demonstrators. Independent Variables As mentioned earlier, the independent variables have been selected from the police report data that describe the situation of each demonstration.IO These variables are largely classified into three categories as shown in Table 5: characteristics of demonstrations, police measures, and outside observers. Table 5 presents an overview of the independent variables as well as their levels of measurement. The first set of independent variables includes a wide set of characteristics of demonstrations. These data have five variables, including organization Sponsoring the '0 The number of arrest and the injury number of police officers and demonstrators are useful variables to examine the hypothesis in this study. However, police reports do not deal with exact data for the two variables, because most arrests are usually made after demonstrations and the number of injuries is in general calculated after demonstrations. Thus, the two variables are excluded in this study. 50 event, the presence ofjoint-groups or sub-groups, the presence of prior violent behaviors, the purposes of demonstrations, and the number of demonstrators. In detail, the organizers/groups of demonstrations are classified into four types of organizers: labor union, Civil Society Organization, residents, and others. Others include farmers, students, religion group, and booth tellers, etc. The data do not include any individual identifiers and even the variable regarding group or organization sponsoring the demonstration is anonymous in the records. Also, organizers of demonstrations are coded according to whether they include joint-groups or sub-groups. In general, organizers of demonstrations in South Korea tend to prefer to confederate with other groups and have many sub-groups in local areas under them. Thus, these groups mostly participate in the demonstrations to support organizers of demonstrations. This trend tends to be common in South Korea. Furthermore, the presence of a history of violence since 2002 in organizers of demonstrations was checked from the KNPA’s Annual Reports. The agency’s Annual Report is published every year. Many organizers of demonstrations have committed prior violence at the demonstrations. Therefore, this study will focus on the effect of this variable on the violent behavior of demonstrators. In addition, the purposes of demonstrations will be largely classified into four categories: labor purposes, political/military purposes, economic purposes, and others. Economic purposes include environment, pollution, development purposes and other purposes include education, religion, and culture. Also, the number of demonstrators will be used as a continuous variable. It decides the demonstration Size, which may affect the behavior of demonstrators and police measures. 51 Table 5: Description of Variables for Analysis Variable Name Code Dependent Variable Violence by Demonstrators 0 = No violence 1 = At least one act of violence Independent Variables Characteristics of Demonstrations Organizer 0 2 Labor 1 = Civil Society Organization 2 = Residents 3 = , Others J oint-group or Sub-group 0 = No joint-group/Sub-group l Joint-group or Sub-group History of Violence 0 = No history 1 = At least one violence history Purpose 0 = Labor 1 = Political/Military issues 2 = Economic 3 = Others Number of Demonstrators Continuous Police Measures Jurisdiction 0 = Big city 1 = Mid-to-Small city 2 = Rural area Ratio of Demonstrators to Police Continuous Distance between Police and Demonstrators 0 = Far 1 =: (Hose Police Equipment 0 No Arrangement 1 Arrangement Presence of Parade 0 = No Parade 1 = Parade Female Officer 0 = No deployment 1 = Deployment 0 No observers Presence of Observers Observers 52 The second set of the independent variables is a variety of police measures to control demonstrations. Police measures include six variables: jurisdiction, the ratio of demonstrators to police officers, distance between police and demonstrators, the arrangement of police equipment, the presence of parade permit, and the deployment of female officers. In detail, jurisdictions of demonstrations include three categories: big city, mid to small city, and rural areas. This classification was adapted from the study of Hwang, McGarrell, & Benson (2005), who classified cities into three groups: “large city” indicated metropolitan cities with populations of several million or more; “small to mid- sized city” included areas with populations of several hundred thousand; “rural areas” referred to pastoral locations with populations of several thousand or less. Also, the ratio of demonstrators to police officers is the primary independent variable to assess its effect on the violent behavior of demonstrators. This variable will be used as the continuous variable. Distance between police and demonstrators will be considered in two measures: close distance and far distance. Specifically, demonstrations with a close distance between police and demonstrators indicate those where police forces are within the demonstrators’ vision to some extent. They can see police activities and view police size or police equipment at the demonstration scene. Thus, in practice, this generally means demonstrations where police forces are deployed within about 328 feet (100 m) of the demonstrators. On the contrary, demonstrations with a far distance between police and demonstrators indicate those where police forces are beyond the demonstrators’ vision, or are completely out of sight. 53 Furthermore, police equipment will be measured by whether the police arrange barricades or water cannons at the demonstration scene. In addition, this study will measure whether the police permit parades on the vehicle roads and whether the female officers are deployed at the demonstration scene. As stated earlier, most variables of this study include a dichotomous measure with the exception of the number of demonstrators and the ratio of demonstrators to police officers. Finally, this study will consider a variable of the presence of outside observers. This variable measures whether or not outside observers are at the demonstration scene in order to assess their effect on the violent behavior of demonstrators. Analytic Strategies This study uses three methods of analysis (univariate, bivariate, and multivariate statistics) to examine the effect of the independent variables on the dependent variable. First, descriptive univariate statistics will be provided for all independent and dependent variables. This study displays the frequency (N) and percent (%) of each variable in the total sample for violent demonstrations and peaceful demonstrations, respectively. The descriptive statistics for frequency and percent for the characteristics of demonstrations will Show the typical patterns of demonstrations occurring in South Korea. Also, the frequency and percent for police measures will reveal overall patterns of different police approaches to control the situation of demonstrations in South Korea. Second, this study will use bivariate analysis. These statistics will examine the relationship between the independent variables and the dependent variable. Bivariate analysis will mostly use contingency tables to assess the relationship between the 54 independent variables and the dependent variable. Chi-square tests of independence are used to examine bivariate relationships between the two discrete variables containing dichotomous or multinominal distributions (Bachman & Paternoster, 2004). Thus, chi- square analysis will be produced in order to test for the independence of a selected independent variable and the violent behavior of demonstrators. The chi-square analysis will be applied to test the relationship between the characteristics of demonstrations (organizers of demonstrations, join-group/sub-group, history of violence, and purposes of demonstrations) and the violent behavior of demonstrators, or between police measures (distance between police and demonstrators, arrangement of police equipment, parade permit, and deployment of female officers) and the violent behavior of demonstrators. In addition, this analysis will use the lambda (A) or gamma ( r ) statistics that are dependent upon the specific independent variable’s level of measurement. Lambda measures of association are used for nominal variables and gamma measures are used for ordinal variables (Bachman & Paternoster, 2004). Furthermore, a T-test will be used for two continuous variables (the number of demonstrators and the ratio Of demonstrators to police officers). Finally, multivariate statistics will use logistic regression because the dependent variable is a dichotomous variable (Bachman & Paternoster, 2004). This study will examine three models by using logistic regression. The first stage of the process only examines the effect of the characteristics of demonstrations on the behaviors of demonstrators. The second stage of this process is modeled on examining the relationship between different types of police measures and the behaviors of demonstrators. The third stage will use the whole logistic regression, including outside observers. In these 55 processes, this study will check for multicollinearity problems by using collinearity diagnostics in order to increase the reliability of this analysis. In the multiple regression analysis, collinearity problems from the independent variables that are either highly correlated or linearly dependent with each other result in very large standard errors and unreliable or unusual regression coefficients (Bachman & Paternoster, 2004). Thus, variables having these problems will be removed in the analysis. 56 CHAPTER VI: RESULTS Univariate Statistics Table 6 focuses solely on univariate statistics with violent demonstrations used as the dependent variable in the analysis, and in particular the characteristics of demonstrations, police measures, and outside observers used as the independent variables. This table shows the frequency and percentage of each variable in the total samples, which are based on 921 cases of mass demonstrations in South Korea. Description for the Independent Variables The descriptions for the independent variables have three broad categories, as shown in Table 6. The first category of the independent variables was a collection of the characteristics of demonstrations. This table has five variables: organizers of demonstrations, joint groups/sub-groups, history of violence, purposes of demonstrations, and the number of demonstrators involved. This shows the typical patterns of mass demonstrations occurring in South Korea. There were no missing variables. First, with respect to organizers of demonstrations, the majority of the organizers were composed of labor unions. Approximately 40.6% of the samples were demonstrations organized by labor unions, and approximately 28.3% had Civil Society Organizations leading the demonstrations. Also, residents organized 14.4% of the demonstrations, and other organizers, such as farmers (5.0%), students (3.7%), religion organizations (1.8%), and booth tellers (1.0%), etc., made up 16.6% of the demonstrations. This is remarkably different from the composition of demonstration organizers in the past. In the past, student and religious organizations protested for a democratic government in South Korea. 57 Table 6: Descriptive Statistics for Mass Demonstrations in South Korea (N=921). Total sample No violence Violence Variables N % N % N % Dependent variable Violence by Demonstrators 921 100 786 85.3 135 14.7 Independent variables Characteristics of Demonstrations Organizer 921 100 786 100 135 100 Labor Union 374 40.6 327 41.6 47 34.8 Civil Society Organization 261 28.3 217 27.6 44 32.6 Residents 133 14.4 1 17 14.9 16 1 1.9 Others 153 16.6 125 15.9 28 20.7 J oint-group/Sub- group 921 100 786 100 135 100 l=yes 348 37.8 278 35.4 70 51.9 0=no 573 62.2 508 64.6 65 48.1 History of Violence 921 100 786 100 135 100 1=history 311 33.8 231 29.4 80 59.3 0=no history 610 66.2 555 70.6 55 40.7 Purpose 921 100 786 100 135 100 Labor 358 38.9 313 39.8 45 33.3 Political/Military 192 20.8 133 16.9 59 43.7 Economic 149 16.2 134 17.0 15 l 1.1 Others 222 24.1 206 26.2 16 l 1.9 Number of Demonstrators 1,000n - 3129.0 100,000b (5662.2)c Police Measures Jurisdiction 912 100 778 100 134 100 Big city 581 63.7 493 63.4 88 65.7 Mid-to Small city 283 31.0 241 31.0 42 31.3 Rural area 48 5.3 44 5.7 4 3.0 Ratio P/D .00a - .71 (.88)° 10.45 Distance P/D 783 100 664 100 1 19 100 1=close 420 53.6 311 46.8 109 91.6 0=far 363 46.4 353 53.2 10 8.4 Police Equipment 819 100 690 100 129 100 l=arrangement 1 16 14.2 74 10.7 42 32.6 0=no arrangement 703 85.8 616 89.3 87 67.4 Presence of parade 921 100 786 100 135 100 l=parade 191 20.7 139 17.7 52 38.5 0=no parade 730 79.3 647 82.3 83 61.5 Female Officer 806 100 684 100 122 100 1=deployment 127 15.8 92 13.5 35 28.7 0=no deployment 679 84.2 592 86.5 87 71.3 Presence of Observers 822 100 691 100 131 100 l=yes 78 9.5 53 7.7 25 19.1 0=no 744 90.5 638 92.3 106 80.9 “ is minimum, b is maximum, C is Mean (SD) for the Number of demonstrators and Ratio P/D. 58 Now that there is a democratic government, these groups are not protesting as much as before (Kim, 2004). Also, of the 921 cases of peaceful demonstrations and violent demonstrations used in this study, the 37.8% of the organizers protested with joint-groups or sub-groups, and 33.8% of organizers had a history of violence. Second, regarding the purposes of demonstrations, demonstrations with labor purposes had the highest rate (38.9%) of the 921 cases. Political/military purposes, economic purposes, and other purposes made up 20.8%, 16.2%, and 24.1%, respectively. Interestingly, demonstrations with labor purposes appeared approximately twice as much as those with the political/military purposes. Additionally, Table 6 shows the difference between the number of labor unions as organizers of demonstrations (374 cases) and the number of labor purpose demonstrations (358 cases). That is, there was a difference between the two variables, because the labor unions sometimes interfered with some affairs, such as political issues, besides labor works. This reflects unique patterns of demonstrations in South Korea. Third, with respect to the variable for the number of demonstrators, the lowest number of demonstrators participating in the demonstrations was 1,000 participants, while the highest number of demonstrators was 100,000 participants. The number of demonstrators was found to have a mean score of 3129.0 with a standard deviation of 5662.2. The second category of the independent variables includes various variables pertaining to police measures. AS shown in Table 6, police measures to control demonstrations had six variables: jurisdiction, the ratio of demonstrators to police officers, the distance between police and demonstrators, the arrangement of police equipment, the 59 presence of a parade, and the deployment of female officers. As discussed earlier, these variables were constructed based on a dataset and previous research suggesting associations between protest policing and demonstrations. Table 6 shows the total frequency and percentage for each type of response from jurisdiction to deployment of female officers. It also reveals overall patterns in the different police approaches. First, the majority of demonstrations (63.7%) took place in metropolitan cities, including Seoul. Mass demonstrations also occurred in mid-to—small cities (31.0%) and rural areas (5.3%). This percentage shows a remarkable difference between metropolitan cities and other cities. This description indicates that most mass demonstrations in South Korea occurred within urban areas. This variable was missing for 1.0% of the sample. Second, Table 6 shows the ratio of demonstrators to police officers as a measure of police action toward mass demonstrations. The high ratio (10.45) was 1,100 demonstrators to 11,500 police officers, while the lowest ratio (.00) was 30,000 demonstrators to 10 police officers. Although the lowest ratio was .00, the police were deployed in all demonstration scenes, regardless of the demonstration size because the police are responsible for maintaining the public order (Bittner, 1991). The average ratio of demonstrators to police officers was nearly .71 with a standard deviation of .88. Overall, the standard deviation was low, which suggests an approximate normal distribution of the variables as the dispersion of ratio was relatively centered near the mean. This variable was missing for 14.1% of the sample. Third, 53.6% of the sample were related to demonstrations where the police were closely deployed to the demonstrators, while 46.4% were associated with demonstrations where the police were deployed at a far distance. This seems to be a similar rate, but 60 generally, the police have a tendency to deploy their force closely, with the thought of deterrence (Wright et al., 2004). This variable was missing for 15.0% of the sample. Fourth, the presence of the arrangement of police equipment, such as barricades and water cannons, was related to 14.2% of all demonstrations. In most demonstrations (85.8%), police equipment was not arranged at the demonstration scene. Police equipment is used to prevent demonstrators from advancing to special areas to protest. Thus, this equipment is mainly arranged in the areas near government offices and political party buildings. However, Table 6 shows that police equipment was restrictively used at the demonstration scenes. This variable was missing for 11.1% of the sample. Fifth, demonstrations which included a parade made up 20.7% of all demonstrations. In South Korea, parades on public roads are common, but contrary to expectation, the rate for this variable was low. There were no missing cases for this variable. Finally, the deployment of female officers at the demonstration scene comprised 15.8% of all demonstrations. The KNPA has relied on the policy of deploying female officers to control demonstrations since 1999. This policy is mainly employed in the metropolitan city of Seoul because the Seoul Metropolitan Police Agency (SMPA) has a special team of female police officers that control groups of parades during the demonstrations. This variable was missing for 12.5% of the sample. AS the final independent variable category of this study, the presence of outside observers at the demonstration only made up 9.5% of all demonstrations. Contrary to expectation that outside observers play a vital role in overlooking the behavior of both demonstrators and police officers in South Korea, Table 6 shows that most 61 demonstrations (90.5%) did not include outside observers. This variable was missing for 10.7% of the sample in the analysis. Description for the Dependent Variable Regarding the dependent variable, Table 6 displays descriptive statistics for the ordinal dependent variable level of whether a demonstrator committed violent behaviors during the demonstrations. Of the 921 cases used in this study, overall, the majority of the sample (85.3%) was associated with peaceful demonstrations, only approximately 135 (14.7%) cases of the sample were associated with violent demonstrations. In particular, Table 6 reveals the patterns of violent demonstrations in detail by presenting the frequency and percentage of violent demonstration within categories of the independent variables.'1 With respect to the organizers of demonstrations, labor unions included the highest rate (34.8%) of the 135 violent demonstrations, and Civil Society Organization, residents, and other organizers had 32.6%, 11.9%, and 20.7%, respectively. Also, organizers with joint-groups or sub-groups comprised 51.9% of violent demonstrations and organizers with a history of violence made up 59.3% of violent demonstrations. With respect to the variable for the purpose of the demonstrations, political /military purpose had the highest rate (43.7%) of the 135 violent demonstrations, and 33.3%, 11.1%, and 1 1.9% of violent demonstrations were associated with labor purposes, economic purposes, and other purposes, respectively. In addition, with respect to the jurisdiction of police measures, 65.7% of the 134 violent demonstrations occurred in big cities, and 31.3% and 3.0% of violent l' The total violent demonstrations were 135 cases. However, the number of violent demonstrations for each independent variable were somewhat different, because it included the missing cases for each variable. 62 demonstrations took place in mid-to-small cities and rural areas, respectively. The violent demonstrations were related to the following police measures: 91.6% of the 119 violent demonstrations were associated with close distance between police and demonstrators, 67.4% of the 129 violent demonstrations were related to the absence of police equipment at the demonstration, 61.5% of the 135 violent demonstrations were associated with demonstrations that did not include parades, 71.3% of the 122 violent demonstrations were related to the deployment of male officers only. Finally, 80.9% of the 131 violent demonstrations were associated with the absence of outside observers at the demonstration. Bivariate Statistics A bivariate analysis examines the relationship between two different variables, while univariate statistics focus on single measures. It is also possible to test if a relationship between the two variables is significant. Thus, chi-square analysis (X2) was produced in this analysis in order to test for the independence of a selected independent variable and the violent behavior of demonstrators (see Table 7). The chi—square analysis (X2) was applied to test the relationship between four characteristics of demonstrations (organizers of demonstrations, joint-group/sub-group, history of violence, and purposes of demonstrations) and the violent behavior of demonstrators. The chi-square analysis was also used for testing the relationship between five police measures (jurisdiction, distance between police and demonstrators, arrangement of police equipment, presence of parade, and deployment of female officers) and the violent behavior of demonstrators. 63 In addition, the measure of association, found using the chi-square analysis, shows the direction and strength of the relationship between the two variables. The use of the lambda or gamma statistics is dependent upon the specific independent variable’s level of measurement. That is, lambda measures of association (it) are used for nominal variables, and gamma measures (r) are used for ordinal variables (Bachman & Paternoster, 2004). Thus, in this study, nominal independent variables (organizers of demonstrations and purposes of demonstration) resulted in a lambda statistic discussion, while ordinal independent variables (joint—group/sub-group, history of violence, jurisdiction, distance between police and demonstrators, arrangement of police equipment, parade, deployment of female officers, and outside observers) resulted in a gamma statistic discussion. Furthermore, a T-test is used in conducting a hypothesis test where variables are mixed. Thus, the T-test of this analysis was produced for the remaining continuous variables, such as the number of demonstrators and the ratio of demonstrators to police officers, in order to test the relationship between the two independent variables and violence by demonstrators. The following table, Table 7, compared the various independent variables with violent behavior of demonstrators. The bivariate relationships between the occurrence of violent demonstrations and the independent variables in the data are shown in Table 7. Table 7 provides the chi-square (X2), measure of association (lambda and gamma), and T- test statistics for the independent variables and the dependent variable. 64 Table 7: Relationshipbetween Selected Independent Variables and Dependent Variable. Variables No violence Violence Chi-square Measure of T-test Association N % N % x2 Characteristics of Demonstrations Organizer 4.68 .003 Labor Union 327 87.4 47 12.6 Civil Society Organization 217 83.1 44 16.9 Residents 117 88.0 16 12.0 Others 125 81.7 28 18.3 Joint-group/Sub-group 13.32*** .33b l=yes 278 79.9 70 20.1 0=no 508 88.7 65 11.3 History of Violence 45.96*** .56h 1=history 231 74.3 80 25.7 0=no history 555 91.0 55 9.0 Purpose 53.25 *** .02a Labor 313 87.4 45 12.6 Political/Military 133 69.3 59 30.7 Economic 134 89.9 15 10.] Others 206 92.8 16 7.2 Number of Demonstrators 2929.4 4291.2 -2.25* (5453.4)c (6655.8)c Police Measures Jurisdiction 1 .65 -07" Big city 493 84.9 88 15.1 Mid-to Small city 241 85.2 42 14.8 Rural area , 44 9 1 .7 4 8.3 Ratio P/D‘ .62 (.74)c 1.22 (1.32) C -4.75*** Distance P/D2 81 .30*** .85b l=close 31 1 74.0 109 26.0 0=far 353 97.2 10 2.8 Police Equipment 42.61 *** .60b l=arrangement 74 63.8 42 36.2 0=no arrangement 616 87.6 87 12.4 Presence of Parade 30.42*** .49b l=yes 139 72.8 52 27.2 0: no 647 88.6 83 11.4 Female Officer 18.1 1*** .44b 1=deployment 92 72.4 35 27.6 0=no deployment 592 87.2 87 12.8 Presence of Observers l6.70*** .48b 1=yes 53 67.9 25 32.1 0=no 638 85.8 106 14.2 *p<0.05, **p<0.01, ***p<0.001 ‘ is lambda (A), b is gamma (r), and C is Mean (SD). ' indicates the ratio of demonstrators to police officers. 2 indicates the distance between police and demonstrators 65 For the first test of the study hypothesis, tests for the characteristics of demonstrations were first estimated (see Table 7). As shown in Table 7, among the characteristics of demonstrations, four variables (joint-group/sub-group, history of violence, purposes of demonstrations, and the number of demonstrators) were significant. However, the organizers of demonstrations variable was not significant. Although demonstrations organized by labor unions (N=47) and by Civil Society Organizations (N=44) were the most common violent demonstrations, those organized by “other organizers” (18.3%) were relatively more likely to have demonstrator violence than labor union (12.6%), Civil Society Organization ( 16.9%), and residents ( 12.0%). The chi-square for this variable, however, indicated that differences between the organizers were not statistically significant. That is, no real difference between organizers of demonstrations and violence by demonstrators was observed. The lambda statistics also suggested that the organizer variable did little to reduce errors in the prediction of the violent behavior of demonstrators (A=.00). That is, organizers of the demonstrations did not have an effect on violence by demonstrators. However, the remaining variables among the characteristics of demonstrations were significantly related to violence by demonstrators. AS anticipated, joint-groups or sub-groups (p<.001), organizers with a history of violence (p<.001), and the purposes of demonstrations (p<.001) were significantly associated with violence by demonstrators. Indeed, when organizers had joint-groups or sub-groups, the rate of violent demonstrations was 20.1%, while when organizers did not have joint-groups or sub- groups, the rate of violent demonstrations was 11.3%. The difference between rates of violence for this variable was substantial. That is, organizers with joint-groupS/sub- 66 groups were more likely to be violent than not. The gamma statistics for the organizers of demonstrations with joint groups/subgroups (r=.33) suggested a moderate relationship with violence by demonstrators at the demonstrations. Organizers with a history of violence showed a strong relationship with violence by demonstrators. That is, organizers having a history of violence were associated with 25.7% of violent demonstrations, while organizers who do not have a history of violence were related to only 9.0% of violent demonstrations. Also, the gamma statistics for the organizers having a history of violence (r=.56) suggested a strong relationship with violence by demonstrators. That is, organizers with at least one prior violent demonstration were more likely to be associated with violence by demonstrators. AS for the purposes of demonstrations, political/military purposes (30.7%) reported greater demonstrator violence than labor purposes (12.6%), economic purposes (10.1%), and other purposes (7.2%). However, the lambda statistics for the purposes of these demonstrations (A=.02) revealed a relatively weak level of association with violence by demonstrators at the demonstrations. Also, the variable for the number of demonstrators was significantly associated with violence by demonstrators in the protest (p<.05). As shown in Table 7, the mean number of demonstrators in peaceful demonstrations was 2929.4 participants, while that of demonstrators in violent demonstrations was 4291.2. That is, the number of demonstrators in violent demonstrations had a higher mean than that of peaceful demonstrations. It seems to be somewhat consistent with Tilly’s (1978) argument that the larger the size of the protest, the more threatening the event. However, previous literature 67 for this variable focused mainly on the relationship between protest Size and use of police force (Earl et al., 2003), not between protest size and violent demonstrations. The second part of Table 7 shows the significant effects of police measures on the violent behavior of demonstrators. To examine the second part of the hypothesis, the six independent variables of police measures were estimated. All of the variables, with the exception of jurisdiction, were Significantly related to the violent behavior of demonstrators. More specifically, the variable for jurisdiction showed no statistically significant relationship with whether demonstrations were violent. This variable produced insignificant chi-square statistics (p>.05), revealing a lack of a relationship between jurisdiction and the violent behaviors of demonstrators. Table 7 indicates that the ratio of demonstrators to police officers was strongly significant (p<.001). The ratio mean of demonstrators to police officers in peaceful demonstrations was .62, while the mean of this variable in violent demonstrations was 1.22. Contrary to expectation that a large police force can minimize violence by demonstrators, the ratio mean in violent demonstrations was higher than that of peaceful demonstrations. It seems to be somewhat of a contrast with most prior studies that suggested positive associations between the level of demonstrations and the relative size of the police force (Brown, 2006; Carter, 1987; Dotson, 1974; Earl et al., 2003; Gurr, 1970; Hibbs, 1973; Koopmans, 1993; Morgan & Clark, 1973; Nalla, 2006; Vitale, 2005). However, this seems to reflect the fact that in practice, the police deploy more forces toward demonstrations with a higher possibility of violence. 68 As for close distance between police and demonstrators, this variable Showed a strong significant result (p<.001). A close distance between police and demonstrators was associated with 26.0% of violent demonstrations, while a far distance between police and demonstrators was related to only 2.8% of violent demonstrations. It seems that increasing police exposure to demonstrators was more likely to incite violence by demonstrators. Also, gamma statistics (r=.85) for close distance between police and demonstrators revealed strongly positive relationships with the violent behaviors of demonstrators. Also, the variable for the arrangement of police equipment, such as barricades and water cannons, showed a strong significant result (p<.001). That is, the arrangement of police equipment was associated with 36.2% of violent demonstrations, while the absence of police equipment was related to 12.4% of violent demonstrations. Gamma statistics (H60) for the arrangement of police equipment revealed strongly positive relationships with the violent behaviors of demonstrators. In addition, having a parade permit for public road access was strongly significant (p<.001). That is, when demonstrations included a parade, the rate of violent demonstrations was 27.2%, while when demonstrations did not include a parade, the rate of violent demonstrations was 11.4%. Gamma statistics (r=.49) suggest a moderate relationship with the violent behaviors of demonstrators. Likewise, the deployment of female officers at the demonstration was strongly significant (p<.001). The deployment of female officers was associated with 27.6% of violent demonstrations, while the deployment of male officers only was related to 12.8% 69 of violent demonstrations. Also, gamma statistics for the deployment of female officers (r=.44) suggest a moderate relationship with the violent behaviors of demonstrators. Finally, regarding the final research question, the presence of outside observers at the demonstrations was significantly related to violence by demonstrators (p<.001). The presence of outside observers was associated with 32.1% of violent demonstrations, while the absence of the same variable was related to 14.2% of violent demonstrations. Also, gamma statistics (r=.48) for the presence of outside observers revealed moderately positive relationships with the violent behaviors of demonstrators. In sum, significant differences were observed in terms of the characteristics of demonstrations, when simply comparing demonstrators who committed violent behaviors to those who did not at the bivariate level (without controls). That is, Significantly larger proportions of violent demonstrators were related to these variables, such as joint- group/sub-groups, history of violence, purposes of demonstrations, and the number of demonstrators. However, there was no a statistically significant difference for the violent behavior of demonstrators with respect to organizers of demonstrations. Also, significant differences for violence by demonstrators were observed in terms of various police measures, such as the ratio of demonstrators to police officers, the close distance between police and demonstrators, the arrangement of police equipment, the presence of parade, and the deployment of female officers. However, there was no a statistically significant difference for the violent behavior of demonstrators with respect to jurisdiction. Finally, the presence of outside observers at the demonstrations was strongly significant in relation with the violent behavior of demonstrators. 70 Multivariate Statistics The bivariate analyses in Table 7 showed Significant and positive associations between the independent variables and the violent behavior of demonstrators. However, the findings presented in Table 7 were only Simple, bivariate relationships. Even these significant associations failed to account for other more complicated interrelationships between the variables of interest. Therefore, to address this issue, this study utilized multivariate logistic regression to simultaneously measure the effect of all 12 independent variables on the dependent variable. The resulting estimates of coefficients (B) and odds ratios are displayed in Table 10. Multicollinearity Examination Before interpreting the estimates, this analysis considered several diagnostic statistics to assess the reliability of this analysis and to evaluate the overall model. Thus, in order to check multicollinearity problems, correlation matrix, tolerance and variance inflation factors (VIF), as well as collinearity diagnostics were produced for all variables used in the full analysis. First, the correlation matrix was examined in Appendix A. In general, the correlation matrix shows how different variables are related each other. It is true that the correlation matrix is the one way to detect multicollinearity problems. A high correlation indicates multicollinearity problem, but it is not a necessary condition for the existence of multicollinearity, because there can be multicollinearity even if zero order correlations are comparatively low. As shown in Appendix A, the variables were not highly correlated. Unfortunately, however, the correlation matrix may not uncover a linear combination of 71 the independent variables. Thus, other sophisticated methods to diagnose the multicollinearity problems were examined. Second, the tolerance and variance inflation factors (VIF) were examined in this analysis. These were more likely than the correlation matrix to detect a multicollinearity problems, especially when there were multiple correlations between independent variables. Table 8 describes the results. Based on the analysis, most variables did not appear to be problematic, since both tolerance and variance inflation factors (VIF) were close to 1.0. However, although there was not a cutting-point to judge whether the values are problematic or not, labor union as an organizer of demonstrations and labor as the purpose of demonstrations seemed to have problematic low scores of tolerance (.092, .095, respectively) and high scores of VIP (10.90, 10.57, respectively). Table 8: Tolerance and Variance Inflation Factors (VIF). Independent Variables Coefficient (B)* Tolerance VIF Constant .006 Labor Union .006 .092 10.90 Residents .016 .422 2.37 Other organizers .073 .702 1.43 Joint-group/Sub-group -.054 .579 1 .73 History of Violence .116 .588 1.70 Labor Purpose -.086 .095 10.57 Economic Purpose -.183 .418 2.39 Other Purposes -. 174 .502 1.99 Number of Demonstrators .000 .854 1.17 Mid-to-Small city .058 .725 1.38 Rural area -.075 .762 l .3 1 Ratio P/D .029 .71 1 1.41 Close distance P/D .221 .820 1.22 Police Equipment .135 .799 1.25 Parade .112 .908 1.10 Female Officer -.005 .793 1.26 Observer .040 .857 1 . 17 *3 value is unstandardized coefficients. 72 Third, to further check the magnitude of linear dependency and the amount of variation that each variable contributes to each concept, condition index and variance proportions by the SPSS commands for collinearity diagnostics were obtained. Appendix B contains the results. Most variables appeared to have normal values of the condition index. However, the deployment of female officers appeared to have somewhat abnormal value of the condition index (10.054), but it was not problematic, because the high variance proportions were not observed as shown in Appendix B. However, the value of the variance proportions for outside observers variable, which had high value of the condition index (14.609) indicated that outside observers seemed to measure the same concepts, which was showed the strong correlation between labor union as an organizer of demonstrations and labor purpose as the purpose of demonstrations. Table 9: Tolerance and Variance Inflation Factors (VIF) With and Without Organizers of Demonstrations. Model 1 Model 2 Independent Variables Coefficient Tolerance VIF Coefficient Tolerance VIF (B)* (B)* Constant .006 .021 Labor Union - .006 .092 10.90 Residents .016 .422 2.37 Other organizers .073 .702 1.43 Joint-group/Sub- -.054 .579 1.73 -.059 .597 1.67 group History of Violence .116 .588 1.70 .116 .595 1.68 Labor Purpose -.086 .095 10.57 -.095 .491 2.04 Economic Purpose -.183 .418 2.39 -.184 .523 1.91 Other Purposes -.174 .502 1.99 -. 163 .516 1.94 Number of .000 .854 1.17 .000 .856 1.17 Demonstrators Mid-to-Small city .058 .725 1.38 .061 .747 1.34 Rural area -.075 .762 1.31 -.070 .815 1.23 Ratio PD .029 .711 1.41 .027 .717 1.39 Close distance P/D .221 .820 1.22 .226 .839 1.19 Police Equipment .135 .799 1.25 .134 .801 1.25 Parade .112 .908 1.10 .116 .919 1.09 Female Officer -.005 .793 1.26 .001 .800 1.25 Observer .040 .857 1.17 .036 .863 1 .16 *B value is unstandardized coefficients. 73 Thus, to deal with the multicollinearity problems, the variable for organizers of demonstrations, which appear to be less relevant to the theory, was dropped. After that, the measures to check multicollinearity problems again were examined. Table 9 and Appendix C contain the results. As expected, there was no problem. That is, the value of variance inflation factors (VIP) for labor purpose became much smaller after dropping the organizers of demonstrations in Table 9 (10.57—+2.04). Also, the condition index and variance proportions for outside observers in collinearity diagnostics were acceptable as shown in Appendix C. Therefore, logistic regression analysis in Table 10 included the characteristics of demonstrations, excluding organizers of demonstrations, police measures variables, and outside observers, in order to examine how the independent variables affect the violent behavior of demonstrators. Logistic Regression Analysis Table 10 shows the results of each variable entered into the model. More specifically, Model 1 only examined the effect of the characteristics of demonstrations on the violent behavior of demonstrators. Model 2 included the effect of police measures variables on the violent behavior of demonstrators, Model 3 included three broad categories, such as characteristics of demonstrations, police measures, and outside observers. Thus, Model 3 shows much exact results. Logistic regression analysis also listed the reference group to make it easier to interpret the odds ratios. That is, the ‘political/military purpose’ variable from the purposes of demonstrations category and the ‘big city’ variable from the jurisdiction category were the reference groups, respectively, as shown Table 10. 74 Table 10: Logistic Regression Results. Model 1 Model 2 Model 3 Variables B SE Odds B SE Odds B SE Odds (association with violence) ratio ratio ratio Characteristics of Demonstrations Joint—group/Sub-group -. 15 .24 .86 —.59 .34 .56 -.58 .34 .56 History of violence 1.16*** .24 3.19 1.28*** .34 3.61 1.27*** .34 3.57 Purpose Labor -1.16*** .23 .31 -.65* .31 .52 -.65* .31 .52 Economic —1.08** .34 .34 -1.32** .43 .27 -1.31* .43 .27 Other purposes -1.30*** .32 .27 -1.51*** .40 .22 -1.50*** .40 .22 Political/Military Reference group Number of Demonstrators .00 .00 1.00 .00 .00 1.00 .00 .00 1.00 Police Measures Jurisdiction Mid-to-small city .45 .31 1.56 .43 .31 1.54 Rural area -1.62 .87 .20 -1.61 .87 .20 Big city Reference group Ratio P/D' .21 .13 1.23 .21 .13 1.23 Close distance P/D2 2.93*** .39 18.76 2.93*** .39 18.69 Police Equipment .82** .31 2.26 .81 * .31 2.25 Presence of parade 1.19*** .27 3.30 1.19*** .27 3.27 Female officer .01 .32 1.01 -.01 .33 .99 Presence of Observers .09 .35 1.10 Constant -1.43 .23 -4.40 .51 -4.63 .54 -2 Log Likelihood 689.97 441.60 441.53 Chi-squared (dt) 77.64 (6) 210.57 (13) 210.63 (14) Nagelkerke R2 .14 .42 .42 *p<0.05, **p<0.01, ***p<0.001 I indicates the ratio of demonstrators to police officers. 2 indicates the close distance between police and demonstrators. In Model 1, there were several different changes observed, when compared to the findings of the bivariate analyses noted earlier in Table 7. That is, organizers with joint- groups or sub-groups and the number of demonstrators variables no longer appeared to be significant, although the coefficient was fairly Significant in the bivariate analyses. Especially, the odds ratio (odds multiplier) for the number of demonstrators was equal to 1.0, because the logistic regression coefficient (Beta) value was zero. It indicates that when the number of demonstrators increased by 1 person, the odds of violence by 75 demonstrators did not change (Bachman & Paternoster, 2004). That is, there was no effect of the number of demonstrators on violence by demonstrators. However, two variables of characteristics of demonstrations, such as a history of violence and the purposes of demonstrations, still remained significantly associated with violence by demonstrators. As shown in Table 10, the odds ratio for organizers with a history of violence was 3.19. The value of the odds ratio means that when organizers of demonstrations had a history of violence, the odds of committing violence increased by 3.19 times. That is, it indicates that the odds of violence in organizers with a history of violence were 3.19 times higher than those who did not have a history of violence. Also, regarding the purposes of demonstrations, demonstrations having labor purposes, economic purposes, and other purposes were significantly related to the violent behavior of demonstrators (p<.001, p<.05, p<.001, respectively). As shown in Table 10, the odds ratio for labor purposes, economic purposes, and other purposes were .31, .34, and .27, respectively. The effect of purposes of demonstrations on violence by demonstrators is negative because these values are less than 1.0. The odds ratio for these variables indicates that demonstrations having labor, economic, and other purposes were less likely to be violent than those having political/military purposes. That is, the odds of violence by demonstrators having labor, economic, and other purposes were lower, compared to those having political/military purposes. The R squared (.14) value of Model 1 in Table 10, indicated that only 14 percent of the variation in the violent behavior of demonstrators were explained by the characteristics of demonstrations. 76 Model 2 in Table 10 introduced the characteristics of demonstrations and police measures variables into the logistic regression analysis. The findings were compared to those of Model 1 and bivariate analyses in order to account for the effects of two broad categories in detail. There was no change found for the characteristics of demonstrations, when comparing the results of Model 2 to those of Model 1. Variables for organizers with joint-groups/sub-groups and the number of demonstrators were again found to be insignificant. Therefore, controlling for all other variables, these variables were unrelated to involvement in the violent behavior of demonstrators. However, the other variables from the characteristics of demonstrations category still remained significant. The statistical significance of the regression coefficients indicated that organizers with a history of violence (p<.001), labor purposes (p<.05), economic purposes (p<.01), and other purposes (p<.001) had significant effects on the violent behavior of demonstrators. The odds ratio for a history of violence variable (3.61) tells us that the odds of violence by demonstrators were over 3.6 times higher in organizers with a history of violence than those not having one. That is, those who had experienced violence in the past were more likely to use violence than those who had not experienced violence. This result increased a little, compared to Model 1 (approximately 3.2 times—+3.6 times). Likewise, the purposes of demonstrations were still significant. The values of the odds ratio for labor purposes, economic purposes, and other purposes (.52, .27, .22, respectively) indicate that the odds of violence in demonstrations having labor, economic, and other purposes were lower than those having political/military purposes, consistent with Model 1. 77 Also, there were several changes for the effect of police measures, when comparing the results of Model 2 to those of the bivariate analyses. That is, when Model 2 included police measures, the jurisdiction variable still remained insignificant. Interestingly, the statistical effects of the ratio of demonstrators to police officers and the deployment of female officers variables disappeared when the characteristics of demonstrations and police measures variables were included in Model 2, although these two variables were significant in the bivariate analyses. In particular, most Korean police administrators think that the deployment of female officers is a significant predictor in the context of South Korea, but this statistical effect disappeared in Model 2. In contrast, the statistical significance of the regression coefficients for close distance between police and demonstrators (p<.001), the arrangement of police equipment (p<.01), and the presence of parade (p<.001) still remained significant. The value of the odds ratio for close distance (18.76) indicates that the odds of violence in demonstrations with the close deployment of police officers were approximately 18.8 times higher than those with the far deployment of police officers. Also, the value of the odds ratio for arrangement of police equipment (2.26) indicates that the odds of violence in demonstrations where police equipment were arranged were approximately 2.3 times higher than those where they were not arranged. In addition, the value of the odds ratio for a parade (3.30) indicates that the odds of violence in demonstrations with a parade were approximately 3.3 times higher than those that did not have a parade. On the other hand, the R squared (.42) value indicated that 42 percent of the variation in the violent behavior of demonstrators were explained by the characteristics of demonstrations and police measures. When both characteristics of demonstrations and 78 police measures were included in Model 2, the R-squared value significantly increased from .14 to .42. This increase of the R squared value indicated that police measures variables were stronger predictors of the violent behavior of demonstrators than characteristics of demonstrations variables. Finally, Model 3 as a full model included three categories, characteristics of demonstrations, police measures, and outside observers. However, most of the results in Model 3 were consistent with those of Model 2. That is, regarding the characteristics of demonstrations variables, organizers with joint-groups or sub-groups and the number of demonstrators variables Still remained insignificant. Also, the variable for organizers with a history of violence and purposes of demonstrations were still significant. The odds ratio for a history of violence variable (3.57) indicates that the odds of violence in organizers with a history of violence were approximately 3.6 times higher than those not having one. The odds ratio of labor purposes, economic purposes, and other purposes (.52, .27, and .22, respectively) indicates that the odds of violence in demonstrations with labor purposes, economic purposes, and other purposes were less likely than those with political/military purposes. That is, the political/military purposes were more likely to cause violence by demonstrators than other purposes. Regarding police measures category, the variables for the ratio of demonstrators to police officers and the deployment of female officers were not significant. However, close distance between police and demonstrators (p<.001), the arrangement of police equipment (p<.05), and the presence of a parade (p<.001) variables were still significant. The findings of these variables were consistent with those of Model 2. That is, close distance between police and demonstrators were approximately 18.7 times more likely to 79 be violent than those located at a far distance, demonstrations with an arrangement of police equipment were approximately 2.3 times more likely to be violent than those with absence of police equipment, and demonstrations with a parade were approximately 3.3 times more likely to be violent than those where a parade was absent. In Model 3, the effect of outside observers, included in the logistic regression analysis as a new variable, was not significant, unlike the result of the bivariate analyses. On the other hand, the R squared value (.42) of Model 3 did not change, when compared to that of Model 2. Summary Among the characteristics of demonstrations category, the history of violence and purposes of demonstrations variables were significantly associated with violence by demonstrators. That is, organizers with a history of violence and demonstrations with particular political/military purposes were more likely to be involved in violence by demonstrators, controlling for all other variables across the models. Especially, the history of violence variable was highly significant and maintained a strong significance across the models. When controlling for all other variables, the organizers with a history of violence variable increased the likelihood of involvement in violence by demonstrators. Demonstrations having labor, economic, and other purposes were less likely to be involved in violence by demonstrators than those having political/military purposes, while controlling for all other variables. Therefore, based on these findings, the hypothesis of this study, predicting a relationship between the characteristics of demonstrations and violence by demonstrators, was partially supported. As stated 80 previously, this hypothesis assumed that the characteristics of demonstrations, such as the violent history of demonstrators and purposes of demonstrations are more likely to have an influence on the violent behavior of demonstrators. Therefore, the results of this study would be strong evidence for this hypothesis. However, there was no evidence that organizers with joint-groups or sub-groups and the number of demonstrators were related to violence by demonstrators. Also, in the research questions, this study assumed that a variety of police measures to control demonstrations have a significant effect on the violent behavior of demonstrators. In order to assess that, this study used various police measures, such as the ratio of demonstrators to police officers, the arrangement of police equipment, the parade, and the deployment of female officers. Consequently, highly significant independent variable predictors of involvement in violence by demonstrators were obtained. That is, the close distance between police and demonstrators, the arrangement of police equipment, and the presence of parades were found to be significant predictors of involvement in the violent behavior of demonstrators. In particular, when the distance between police and demonstrators was close, the likelihood of violence by demonstrators increased by 18.7 times. The odds value for close distance between police and demonstrators was much larger than the remaining values. Moreover, the use of police equipment was associated with violence by demonstrators. Besides, under demonstrations having a parade, demonstrators were more likely to employ violence. Therefore, based on these findings, there was strong evidence in support of the hypothesis, predicting a relationship between police measures and the violent behavior of demonstrators. 81 However, the ratio of demonstrators to police officers and the deployment of female officers were insignificant. The final hypothesis of this study was whether outside observers have a significant influence on the violent behavior of demonstrators. As stated earlier, the role of outside observers at the demonstration scene is important in South Korea. However, unfortunately, a relationship between the presence of outside observers and the violent behavior of demonstrators was not found in the analysis, when controlling other variables. Therefore, there was no evidence to accept this hypothesis. 82 CHAPTER VII: DISCUSSION AND CONCLUSION Explanations for Results The primary purpose of this research was to examine the relationship between various police measures and violence by demonstrators. As stated earlier, this research assumed that three broad categories (characteristics of demonstrations, police measures, and outside observers), used as the independent variables, would affect the violent behavior of the demonstrators, used as the dependent variable. Consequently, several significant findings were gleaned about the relationships between the independent variables and the dependent variable. The findings of the Study showed that as expected, the characteristics of demonstrations as well as a variety of police measures have a strong effect on the behavior of the demonstrators. That is, organizers with a history of violence, demonstrations with political/military purposes, a close distance between police and demonstrators, the arrangement of police equipment, and the presence of a parade permit were more likely to be involved in violent demonstrations, controlling for all other variables across models. However, organizers with joint-groups/sub-groups, the number of demonstrators, jurisdiction, the ratio of demonstrators to police officers, the deployment of female officers, and the presence of outside observers at the demonstrations did not have a Significant influence on violence by demonstrators. In order to further understand the effects of protest policing on mass demonstrations in South Korea, a more detailed explanation and comparison of the findings in this study and the findings in the literature review are needed. 83 First, regarding the effect of a history of violence among the characteristics of demonstrations on violence by demonstrators, this analysis showed that whether organizers experienced prior violence had a great deal to do with violence by demonstrators. In other words, when any demonstration was organized by an organizer with a history of violence, the likelihood of involvement in violence was approximately 3.6 times higher than those who did not have a history of violence. It is consistent with police practice that the police have generally judged whether a reported demonstration turns violent according to a history of violence by demonstrators (Kim, 2004; Yu, 2003). Also, the results of this study indicated that when any demonstration was associated with political/military purposes, the likelihood of involvement in violence by demonstrators increased. This is similar to Earl et a1. (2003), McAdam (1982), and Waddington’s (1998) arguments that groups using confrontational tactics or pursuing politically radical goals were related to violence at the demonstrations. However, organizers having sub-groups in this analysis did not show significant results. This is a contrast with previous research findings of the US. that radical organizations with sub—groups have higher rate for violent demonstrations (Earl et al., 2003; McAdam, 1982; Waddington, 1998). This result also reveals current pattern of demonstrations in South Korea. That is, organizers of demonstrations in South Korea tend to confederate with other groups to raise the effect of the protest, but these joint rallies do not necessarily increase the likelihood of violence during the demonstrations. Second, little relationship was found between jurisdiction and the violent behavior of demonstrators in the analysis. This is not consistent with Boudreau (2001) and Earl et a1. (2003) views that a police department located within jurisdictions of high-capacity are 84 more likely to monitor protests. There were seemingly differences according to jurisdiction. That is, 65.7% of violent demonstrations occurred in big cities, 31.3% and 3% took place in mid-to-small cities and in rural areas, respectively. However, the finding for each category was not significant. The reason is related to the fact that South Korea is policed by one national police agency. The KNPA controls demonstrations in all cities with the same level of resources and facilities. Thus, they generally support police stations with lower budgets and fewer officers when great size demonstrations occur within their own jurisdiction. This shows unique police patterns in controlling demonstrations in South Korea. Third, the finding for the ratio of demonstrators to police officers was different from the assumption of deterrence theory, which suggests that an increase of police presence may deter the violent behavior of demonstrators (Wright et al., 2004). There were many previous research studies to suggest significant results for police measures such as police size (Brown, 2006; Carter, 1987; Dotson, 1974; Earl et al., 2003; Gurr, 1970; Hibbs, 1973; Koopmans, 1993; Morgan & Clark, 1973; Nalla, 2006; Tilly, 1978; Vitale, 2005). Conversely, there were opinions that the possibility of violence depends on the characteristics of demonstrations, regardless of police measures. According to Reicher et a1. (2004), most collective events tend to be peaceful and require no substantial police action. However, this was not to suggest that the police do not have the ability to curb collective violence, but that there were limits as to the impact the police can have in a situation of collective violence. These contrasting findings indicate that the police size variable can affect demonstrations positively or negatively. 85 Nevertheless, the finding for the ratio of demonstrators to police officers in the present analysis indicated no relationship with the violent behavior of demonstrators. This may reflect that police measures to control mass demonstrations in South Korea are affected by the demonstrators’ perception toward the police. That is, unlike the perceptions of citizens in Western democracies, many Korean people have a tendency not to recognize the role of the police because of unique historical factors, as stated earlier. South Korean citizens recognize demonstrations and riots as a kind of right to defiance toward past illegitimate military governments (Hwang etal., 2005). Thus, this study finds that even if the number of police officers dispatched to control demonstrations increases at the demonstration scene, it does not have an effect on the violent behavior of demonstrators. Fourth, the primary findings regarding the close distance between police and demonstrators, the arrangement of police equipment, and the presence of parade permit were found in the analysis. At first, this study assumed that various police measures would prevent demonstrators from employing violence during the demonstrations. As expected, this result was that police measures affected the violent behavior of demonstrators. However, most police measures used in this study had the opposite effect on violence by demonstrators. In other words, when police officers were closely deployed to demonstrators (approximately 18.7 times), police equipment was arranged at the demonstration (approximately 2.3 times), and the police permitted for a parade (approximately 3.3 times), the odds of violence by demonstrators were statistically increased rather than decreased. 86 These results might suggest that police measures seemed not to prevent violence by demonstrators, but to invoke demonstrators to use violence. These results are somewhat consistent with the opinions of researchers who argued negative effects of police forces on demonstrations (Kim, 2004; Reicher etal., 2004). These results may be related to the demonstrators’ perception that they think of violence as an effective way to gain media or government attention, by attacking the police who are deployed at a close distance, or by committing violence during the parade. Thus, the findings of this study seemed to suggest the alternative policy, that in order to reduce violence by demonstrators, the police should not be closely deployed to the demonstrators and should not invoke demonstrators by arranging police equipment, such as barricades and water cannons, at the demonstration. However, an alternative explanation of the findings for police measures is also suggested. The reason that the violent behavior of demonstrators increases when police measures were employed at the demonstration is that various police measures were generally used toward demonstrations with a higher possibility of violence. Thus, the police often employ various measures to control demonstrations with a higher possibility of violence because the police have basic duties to maintain public order (Bittner, 1991). In that respect, the effect of police measures on demonstrations in this study might Show different results from expectation. Therefore, to estimate the effects of police measures exactly, future research may wish to include alternate measures, such as the judgment of the police, or intelligence, to predict the possibility of violent demonstrations. Otherwise, if the findings of this study were true to an extent, there might be differences in the level of violence when significant police measures were implemented 87 at the demonstration scenes. However, this study could not estimate this effect exactly because this study only employed a dichotomous concept for violent demonstrations as the dependent variable. Thus, if future studies incorporate types and degrees of violence as the dependent variable, they can examine the difference of the level of violence by demonstrators in accordance with police measures. Fifth, contrary to expectation, the deployment of female officers had no statistical effect on the violent behavior of demonstrators. This policy was centered on the policy of the KNPA since 1999. Many scholars and police administrators in South Korea evaluated positively the effects of the female officers that it created a new image of the police and made considerable strides in protest policing in South Korea (Kim, 2004; KNPA, 2001; Kwon, 2006). Therefore, they might be somewhat shocked by the unexpected findings of this study. In South Korea, the proportion of female officers has grown gradually every year. Their roles have speedily shifted from one of a caretaker for incarcerated women and comforter of lost children to that of a bona fide law enforcement official. Like other countries, public attitudes have also been changing toward greater acceptance of the female officers in the policing system (Alpert & Dunham, 1999; Hoffman & Hickey, 2005; Hunt, 1999; Koeing, 1978; Martin, 1997; Paoline & Terrill, 2004; Worden, 1995). Thus, regardless of the result of this study, future studies may wish to find new ways to use their roles to control demonstrations continuously. Sixth, contrary to expectation, the presence of outside observers had no statistical effect on violence by demonstrators. No relationship was found in preventing violence by demonstrators at the demonstrations. Although there are no previous studies about reasons for that, it can be inferred that the reason is from the fact that outside observers in 88 South Korea have limited status and their status is largely different from that of legal observers in the US. As stated previously, observers in the US. are typically law students, legal workers, and lawyers, while those in South Korea are mostly composed of general citizens who do not have any legal standing. Also, legal observers in the US. watch and record the actions of the police at the demonstration scene. In addition, they can assist activists who are arrested unexpectedly or who need medical attention, by alerting the appropriate support teams associated with the demonstration (National Lawyers Guild, 2004). Thus, future studies may wish to find new ways to continuously examine the effect of the observer role on demonstrations. Finally, we expect that significant findings for this study strengthen police measures to control demonstrations in South Korea. First of all, this study will contribute to police endeavors to employ new policing tactics to reduce violence during the demonstrations. Although some police measures tested here failed to Show an effect, important other variables, such as distance between police and demonstrators, arrangement of police equipment, and parade permit did. In this respect, to the extent that this is true, it suggests effective ways for police intervention to control demonstrations. Furthermore, this study reinforces the need for comparative criminal justice research (Cao & Hou, 2001; Hwang et al., 2005). Although there were similarities in the findings of this study and previous research, there were also important differences. A particularly interesting point was the finding that police measures were negatively associated with the violent behavior of demonstrators. However, as stated earlier, this study revealed the relationship between police measures and the violent behavior of demonstrators with respect to Korean views. It is an undeniable fact that this study found significant results 89 for demonstrations in South Korea. Thus, as Della Porta and Reiter (1998) argued, it is possible to compare the protest policing in South Korea to that in different Western democracies. Through this comparative study, the Korean police will find new tactics to control demonstrations efficiently and effectively. Study Limitations This study has several major limitations that warrant caution in interpreting the findings. The first problem is the limited selection of the independent variables used in this study. A wide variety of factors should be considered to explain the complex demonstrations in South Korea. Nevertheless, this study relies only on limited variables of the characteristics of demonstrations and police measures. This study did not address important factors influencing the demonstrators directly. For example, among various characteristics of demonstrations, the level of intoxication of demonstrators was not considered sufficiently. Many studies have examined the effect of alcohol on demonstrations. According to previous studies, alcohol may contribute to celebratory riots and high-risk drinking may impact the likelihood of overall disruptive and inappropriate behaviors (Chermack, 1995; Giancola, 2002). Also, this study did not consider the police judgment of whether some reported demonstration turns violent and organizational properties of police departments that may influence police measures toward demonstrations. In particular, according to Stockdill (1996) and Earl et a1. (2003), when these factors of police departments are considered in the context of protest policing, larger, more bureaucratic agencies, or a chief that 90 emphasizes crime-fighting give rise to more police brutality. However, this study did not consider these factors in the analysis. Moreover, this study focused on only one aspect of proactive police measures and did not consider reactive police measures, such as arrests, tear-gas, and dispersion. Earl et al. (2003) emphasized that making arrests at the demonstrations was an important variable in controlling demonstrations. Thus, future development should go beyond this first step to examine the connection between reactive police measures and the violent behavior of demonstrators. In addition, this research did not include functions of the press that may have indirect effects on the violent behavior of demonstrators. Wisler and Giugni (1999) were concerned with the effect of media coverage on protest policing. They found an inverse relationship between the level of media coverage and repression. That is, protests that are less protected by the watchful eye of the media may be “weaker” and thus more prone to repression (W isler & Giugni, 1999). Therefore, future research may wish to incorporate multiple measures of demonstrations, such as intoxication of demonstrators, reactive police measures, and media coverage in order to estimate the exact effects of various variables on demonstrations. Second, as pointed out previously, a dichotomous dependent variable is too simple for analyzing the effects of various variables on demonstrations. The data for the dependent variable can be measured according to the level of the violent behavior and can be placed along a “continuum” ranging from least to most severe behavior. That is, violent behaviors are ranked in the following manner: “no involvement in violence”, 91 “moderate involvement in violence” (e. g., physical collision with the police), and “serious involvement in violence” (e.g., use of rocks, fire bottles, sticks and iron sticks, and use of all these things). However, because the number of each type of violence is limited, the typology of this continuum is excluded from this study. Therefore, this study only includes no involvement or involvement in violence by demonstrators. With respect to the continuum of violence by demonstrators, future studies need to examine the relationship between the types and degrees of violence by demonstrators and policing. Third, although the data used in this study reflected recent policy by using recent data, the study period was relatively short. The period of the study is one of the important factors to test effects of variables exactly. Nevertheless, the data for this study was collected for only two and one-half years (1/1/2004-6/30/2006). In that respect, this is the most serious drawback of the study method. That is, the effects of police measures can not be tested thoroughly for a short period of time. For example, this study employed this period in order to estimate the effects of outside observers on violence by demonstrators because the activities of outside observers started in the early 2004. Contrary to expectation, the result was that outside observers did not have an influence on violence by demonstrators. Therefore, in order to estimate their effects on violence by demonstrators, it can be more reasonable to compare two periods: before conducting observer policy and after conducting the same policy. Thus, future research should incorporate comparison of the period of time for outside observers in order to evaluate their effects on the violent behavior of demonstrators. Finally, the generalizability of findings in this study is somewhat limited with respect to selection bias for variables, data period, and sample size, as discussed above. In 92 particular, the sample included only demonstrations with over 1,000 participants. Since many demonstrations in South Korea include petty illegal behavior, this study does not include them at all. However, in practice, the police use measures toward all demonstrations to maintain public order. Based on this notion, the sample size of this study may be biased toward the independent variables. As a result, problems regarding representativeness can be found. This study had a restricted scope. Therefore, it is difficult to say that the findings of this study represent all effects of police measures on demonstrations in South Korea. Thus, future inquiries should adopt a broader sampling frame in order to generalize the findings. Future study Although this study contributes to an overall understanding of the relationship between police measures and demonstrations, the findings come with several limitations, such as a limited selection of the independent variables, the problem of a dichotomous dependent variable, short data period, and small sample size, as pointed out earlier. Thus, future studies should continue to investigate the effects of various independent variables on violent demonstrations. Future research will enable policy-makers to further find new policies that ultimately produce better police practice in controlling demonstrations. First, future studies may wish to incorporate multiple measures of demonstrations, such as the intoxication of demonstrators, police judgments of whether some demonstrations turn violent, and media coverage that indirectly influences demonstrations, in order to estimate the exact effects of various variables on 93 demonstrations. If future research considers these variables, this would undoubtedly contribute to cumulative knowledge-building in the area of protest policing. Second, researchers should also consider measuring and incorporating all forms of demonstrations into their studies in order to generalize the findings. Thus, future work Should attempt to examine police measures used in controlling all kinds of demonstrations, beyond the sample including demonstrations with over 1,000 participants, used in this study. Third, future studies need to examine the relationship between the types and degrees of the violent behaviors of demonstrators and policing. If future studies incorporate types and degrees of violence as the dependent variable, the study can examine the difference of the level of violence by demonstrators in accordance with police measures. As demonstrated here, expanding the dependent measure to account for various levels of violence by demonstrators provides a clearer picture of what factors influence violence by demonstrators (Terrill & Mastrofski, 2002). Fourth, future studies may wish to continually examine the effect of female officers on the violent behavior of demonstrators. Despite female officers’ substantial presence, the results of this variable in this study were not significant. In South Korea, however, the role of female officers in the policing system has been positively evaluated by police administrators as well as by many citizens. It may be that the sample Size of demonstrations where female officers were deployed (15.8%) was too small to measure impact. Thus, future research needs to estimate their effect on demonstrations again, after concretizing their activities in controlling demonstrations. 94 Finally, future studies may wish to continually find new ways to examine the positive effect of outside observers on violent demonstrations. A significant result for this variable in this study was not found in this analysis, but it may be that the sample size of demonstrations with outside observers (9.5%) was too small to measure their effect, like the variable for deployment of female officers. Thus, in order to exactly estimate the effect of their role on violent demonstrations, additional research needs to incorporate a comparison of the period of time before and after observer policy implementation, to examine the effect of outside observers. Conclusion and Policy Implications This study examined how the characteristics of demonstrations and various police measures have an impact on the violent behavior of demonstrators in South Korea. As stated earlier, there are several limitations, but the real strengths of this study outweigh its potential weaknesses and the knowledge for protest policing on mass demonstrations is improved. The findings of this research suggest that organizers with a history of violence, the close distance between police and demonstrators, and the arrangement of police equipment at the demonstration scenes increased the risk of involvement in violence by demonstrators. These findings give significant implications to the police who seek to develop new tactics in order to manage demonstrations in a manner consistent with the concept of policing by consent and public expectations. Before suggesting policy implications in accordance with the findings of this study, we can see several existing theories that can provide more effective measures to control demonstrations in South Korea. The KNPA has used a variety of crowd-control 95 measures, which are based on related theories: deterrence theory (Wright et al., 2004), procedural justice (Tyler, 2003), and defiance theory (Sherman, 1993). Sherman and Berk (1984) argued the deterrence doctrine held that the pains of punishment deter people from repeating the crimes for which they are punished. According to Wright et a1. (2004), deterrence theory may have some effect on controlling demonstrations and riots. Harsh punishment and an increase of police presence, based on deterrence theory, may prevent the violent behavior of demonstrators at the demonstration scene by making them believe arrests and subsequent sanctions are likely. However, this deterrence research is often criticized for not addressing common third causes (Wright et al., 2004). For example, the unique historical background of South Korea caused lots of riots despite the oppressive responses of police forces. This is because many Korean people recognized demonstrations and riots as a kind of right to defiance toward illegitimate military governments (Hwang et al., 2005). It indicates that causes and patterns of demonstrations in South Korea are different from those of Western countries, including the US. Thus, the effect of coercive police forces based on deterrence theory was not shown in expected actions to control demonstrations in South Korea. South Korea has a unique historical background that can not be explained by the deterrence theory. Therefore, the policing, based on deterrence theory, should consider historical factors to have a real effect of this theory. Procedural justice (Tyler, 2003) and defiance theories (Sherman, 1993) that can be applied to the situation of South Korea have some policy implications. Sherman (1993) and Tyler (2003) argued that people who had experienced contacts with the police obey the law more when they believe it was administered fairly than when they judged 96 that the officials had treated them unfairly. Therefore, to enhance crime prevention, these theories emphasize a fair justice system. The government should invest in law enforcement agents’ education to boost law enforcement fairness. Accordingly, like Tyler’s study (2003), Sherman (1993) argued that the police can gain people’s compliance by engaging in strategies of process-based regulation and by treating community residents in ways that lead them to feel that the police exercise authority in fair ways. Also, when the empirical attempts of the KNPA were evaluated in the context of these theories, as discussed previously, the empirical attempts gained the following results: harsh penalties and the increase of police presence policies, which are based on deterrence theory, appeared to have little impact on preventing violent demonstrations. Rather, the empirical results suggested that policies, such as stopping use of tear gas and increasing fair law enforcement, were much more effective than those based on deterrence theory. It indicated that the fair enforcement of law is important in reducing violence during the demonstrations in South Korea. However, the findings of this study are somewhat different from those of empirical attempts to control demonstrations and riots in South Korea. The empirical attempts of the KNPA have limitations for estimating exactly the effect of attempts by the police. That is, it was a simple numerical comparison of violent demonstrations without analyzing each case. Also, the police did not fully enforce policy based on deterrence theory toward demonstrators. In practice, arrests that centered on deterrence theory were rare and generally were made only in cases involving extreme violence. Therefore, the real effects of deterrence theory for controlling demonstrations did not emerge at the 97 demonstration scene. Thus, based on the findings of this study, the KNPA needs to consider the following views to develop new tactics for controlling demonstrations. The key for new policy and practices on demonstrations is to focus on reducing the violent behavior of demonstrators. First, the policy may wish to consider a deterrence theory much more thoroughly when dealing with organizers having a history of violence or committing violence given these groups’ propensity toward violent behavior. That is, the findings suggested that when organizers with a history of violence led demonstrations, the likelihood of involvement in violence by demonstrators increased. According to Sherman and Berk (1984), the deterrent effects of arrests appeared to depend on the length of the follow up period, the length of detention, and whether the offender perceives that he has been treated fairly. Earl et a1. (2003) also emphasized making arrests at the demonstration scene. Therefore, the police need to respond to lawbreakers strictly during the demonstrations. In that respect, “zero tolerance” tactics suggested in the command and control style of the police in the US. have significant implication (Brown, 2006; Nalla, 2006; Vitale, 2005). Similarly, Nalla (2006) argued that the zero tolerance approach can be applied to all its core missions, including protest policing. Thus, the Korean police need to enforce the law more strictly toward minor violations during the demonstrations according to “zero tolerance” tactics. Of course, this tactic of the police should be based on values embedded in the philosophy of what is called “quality of life policing” (N alla, 2006; Vitale, 2005). The strategy of the NYPD on Feb. 15 , 2003 toward a mass demonstration is a fitting example (Vitale, 2005). At the same time, making arrests 98 toward lawbreakers can be more effective when the arrested suspect perceives that he has been treated in a procedurally fair manner. Therefore, when the police enforce law thoroughly and fairly, as in the opinion of Sherman and Berk (1984), it can be an effective method to employ policies based on deterrence theory in South Korea. Second, imposing a financial charge toward organizers committing violence during the demonstrations or organizers pursuing a parade on the roads for vehicles can be one effective way to control demonstrations. More recently, a local government in South Korea made an ordinance to stop financial support'2 toward organizers committing violence during the demonstrations (Yonhapnews Feb. 09, 2007). Effective ways to manage parades, based on deterrence theory, can be found in the cases of the US. That is to charge a fee for a parade on the roads to reflect the estimated cost of maintaining public order. As for this fee, the Supreme Court in the US has not decided whether nominal charges are constitutionally permissible. However, there were several cases to support this opinion. A federal district court upheld the Kansas City Police Department’s policy of requiring parade sponsors to pay for the cost of traffic control (Schofield, 1994). Similarly, the Six Circuit US Court of Appeals upheld a Columbus, Ohio ordinance that required prepayment of an $85 fee for the cost of processing a parade permit application and prepayment of the cost for traffic control (Schofield, 1994). Also, a California appellate court decided to uphold portions of an ordinance that requires a parade perrnittee to reimburse the city for the actual replacement or repair cost of the destroyed or damaged property, if city property is destroyed or damaged, by reason of permittee’s '2 According to Joongang Ilbo institute for Civil Society in South Korea, since few NGOS are completely independent from financial problems, many of the N GOs in South Korea receive financial support from the government. Since 1997, the government has financially supported more than 1,300 NGOS in order to encourage NGOS to carry out their roles faithfully. 99 use, event or activity. Of course, it is constitutionally significant that indigent groups unable to pay the fees were not precluded from engaging in expressive activity, because an alternative forum was available. For example, sidewalks were free for conducting a parade, because traffic control was not affected and parks were available with costs for related speech activities (Schofield, 1994). Third, in addition to policies based on deterrence theory, the finding of this study may require the police to decide to change the styles for deployment of police forces at the demonstration scene. That is, this policy may be to deploy police officers and police equipment far from the demonstration scene. This policy is drawn from the result of this study that the close distance of the police and arrangement of police equipment at the demonstration scene can cause more violence by demonstrators. The finding of this study suggests that more police deployment at the demonstration scene did not necessarily prevent violence of demonstrators. It is also based on the fact that police officers and police equipment may invoke demonstrators at the demonstration scenes (Kim, 2004). Thus, it should not insist on the deployment of police officers at the demonstration scene. This suggestion can be a sensitive issue in South Korea. Police administrators may also refuse to employ this policy. However, if the police accept this policy, this may lead to a significant change in the policing styles. First of all, this can change the negative image of the police at the demonstration scenes. Also, it may give the opportunity to reduce collisions between the police and demonstrators that have caused lots of injuries. The image of the police is the key influence in determining protest policing. In fact, law enforcement agencies under the past military government were inadequately prepared to deal with violent demonstrations 100 for a long time. The South Korea police, especially, as frontline representatives of government authority produced a negative image in all their works (Hwan g et al., 2005). Therefore, it is important to change citizens’ perception toward the image of the police. Thus, this policy, which deploys police officers and police equipment far from the demonstration scenes and allows protest leaders to handle members of their group at their own level, can be a new strategy to alter the negative image of the police and to gain support from citizens. As Fisher (2001) stated, an ‘air of cooperation’ between police and demonstrators can be established when this policy is employed. This policy that police forces should be deployed at a far distance between police and demonstrators seems to be incompatible with the policy that demonstrations should be controlled according to zero tolerance tactics. However, both policies have different objectives and may be appropriate in different circumstances. The former indicates general tactics for deployment of police forces to avoid provoking demonstrators and to control demonstrations. In other words, it means that police forces, including police officers and police equipment, should be deployed at a far distance from demonstrators at the beginning of demonstrations. On the other hand, when demonstrators break the law or commit violence during the demonstrations, the police should deal with lawbreakers actively, according to zero tolerance tactics. Therefore, when the police manage some demonstrations, police forces should be deployed at a far distance between police and demonstrators if possible, and then if some violence by demonstrators takes place during the demonstrations, the police Should manage lawbreakers according to zero tolerance tactics. 101 Fourth, although there was, unfortunately, no evidence found in this study that outside observers affected violent demonstrators, their role in overlooking fair enforcement of the police and violent behavior of demonstrators is still important at the demonstration scene. The desire for fair enforcement drawn from the procedural justice theory can make an effective method to control violent demonstrations in South Korea. Therefore, the police may wish to employ the policies that are based on procedural justice theory, which states that people obey the law more when they believe it is administered fairly. It can be a good way to actively grant legal status in order for observers to strengthen their role in overlooking demonstrations effectively, similar to legal observers in the US. The presence of outside observers would help to preventing violent demonstrations and improving fair enforcement from the police. Finally, regardless of the findings of this study, the police may wish to consider a suggestion to control demonstrations in South Korea. It is the participation of the community that can control demonstrations. As Brown stated (2006), “the police alone can not control all incidents of collective behavior or single-handedly control all situations of collective violence” (p. 108). Citizen support is essential for all police works to succeed. Therefore, as Bauman (1991) argued, the police and government should constantly innovate ways to tap into the community. That is, the police should be willing to cooperatively work with the citizenry. It is a good example that “the police can help defuse public discontent before it escalates to the point that protests become violent, by means of interacting with the citizenry in engaging the citizenry in cooperative projects” (Brown, 20061108). Thus, when the community policing model is implemented as a way of protest policing, violent demonstrations in South Korea will be largely reduced. 102 Beaver... .deo: weave. 00; ***mm. 3.32.2. .2243. no. 31.5. No. timm. .31.. av 305820 028:0 o3 tit. item. .12. is: 3.12. team. 3%. a: waste oaeom 00; No.. $2. :2. mo. .35. .35. AC 033m of 13.2. .512. :2. .138. 13.8. GO 8235.8 8:8 2: ***m _. CO. No.. mo. Amv DE 8556 owe—O co; film—r sacrum. $2.3. Gav DE ovum co; es. .512. AC 20352083 Mo 59:52 Go; .13.. Du. ANV coco—o? me home 9: A: qsouwbamfisocwsfig o w h c n v m N _ 83mg; .23qu 5.522 sous—2.80 H< 5259?. 103 .Eotumno 02300 E .0850 2900.. .«o ~5an.000 2 “E00 0083 m. Jabs—9:00 00:8 h.0 EoEowcaba 3 .DE 00556 0020 m. .QE 0:8 ~— 0003 EB _ _ E0 zaEmOTpE o_ .EofibmcoEow «0 .5950: o £08900 .050 w .3093 2:550... h .3093 .592 0 6000.03 L0 .083; m dzocmbzmhzouwafioq v .BQNanco .050 m .3523: N .302: .022 _ 8. oo. oo. oo. 5. oo. oo. oo. oo. oo. 5. m0. 8. 8. No. 5. mm. 80.3 0.8. 5 oo. oo. oo. oo. 3. 2. S. 3. mo. :4. 2. 3. oo. 2. mo. co. mo. 08.3 So. 3 5. co. mo. 8. mo. mo. co. 5. oo. .3. 3.. 8. mm. m _. S. cm. 00. whmh m _ N. m— .o. oo. oo. co. m _. mo. 2. mo. co. oo. o_. 8. mm. ca. 5. 2. co. 02.0 cum. 3 5. 5. oo. 8. mm. mm. mo. cm. 00. no. vo. co. 3. Z. 3. mo. 5. 2 we gm. 9 oo. oo. 5. mo. 9. 5. oo. 2. 5. mo. 5. co. cm. 8. 3. mo. oo. 30...‘ mom. «— mo. 5. cc. mo. No. mm. 2. mm. 2. co. :. oo. co. 00. mo. 5. oo. mwod wwm. = S. 2. 3. mo. 8. No. 8. 8. No. 00. 8. oo. 3. 3. oo. oo. oo. 2 Wm wmm. 3 No. 3.. oo. 5. oo. _o. co. co. 8. m _. _o. 8. mo. 8. cm. mo. 00. Erma mom. a 2. 0m. 00. _o. 5. co. co. 5. mo. 2. 5. oo. oo. 8. mm. 5. oo. mama mam. a pm. .0. 3. mo. co. mo. 3. co. 5. co. co. co. co. mo. 3. co. co. ~85 «.3. 5 8. mo. mo. 3. 5. o_. 5. 8. cc. 5. co. co. 5. oo. _o. oo. 8. 03am m L w. 0 mm. No. mm. 8. co. oo. 00. cc. co. oo. oo. oo. oo. .o. co. co. co. omod wow. m 0o. 5. oo. 8. oo. oo. mm. 8. co. co. 5. oo. .o. 5. co. co. co. OWN wmo. v S. 3. 00. mo. 8. S. No. 8. 8. 8. oo. oo. 8. S. 3. _o. co. mm _ .N com; m 8. No. oo. oo. co. oo. 8. No. 8. no. 5. oo. oo. 5. no. .8. 8. woo; 30.. N 5. oo. oo. oo. 5. 8. we. 5. oo. .o. 3. co. co. 5. 5. 3. 8. ~32 ommd — S 3 m— 3 2 Nu 3 3 a a h e n v n N u .80:— uoE—Eoo 0=R>=ow§ aux—:52 20:030.:— 0003.:5 .mcogbmsofiom h.0 0.03:5ch 5;? mosmocwflfl $32560 ”m 505%? 104 4.": .III. . I u ll, Dill... £03030 02050 E .2850 22:00 00 50500—000 m_ 05:00 00200 2 40200500 00:00 00 EonwSF—a _ _ .93 000056 3.20 o_ .0». 0:0... 0 .0008 :5: w .36 =aEm-0.-EE h .EofibmcoEoc 00 000:5: o £03050 0050 w 080.50 0:00:80 0 080.50 500— m .800—03 00 000.02 N .0000w-o=w\0=00w-5_0.q _ oo. 33. 3. 3. 3. 03. No. mo. .3. ow. mm. hm. 3. mo. Nomd moo. 3 3. oo. co. co. 3. No. 2 . I. 3. Nb. 3. 3. mm. No. 93.0. mmm. m— 3. oo. oo. 33. mm. 2. 3. mm. 3. no. 2. :. a. co. wind mwm. ~— 3. 00. No. 3. Nm. .3. 3. ON. 33. me. N3. M3. 2. co. mmoé 0mm. 2 3. co. oo. 03. No. vm 03. mo. mm. 03. N _ . we. 3. A: . omwd mhm. 3 m3. No. 2. cc. 3. 3 Z. No. 3. 03. no. No. 3. no. coma 0. m. 0 2. on. mo. 3. 3. 3. 3. 3. mo. 03. 33. No. 3. No. of .m mmm. w mm. 3. 50. oo. 3. No _ _. oo. 3. vo. mo. 3. 33. oo. wmaN 93. 0. 33. No. 03. co. 3. 03. oo. 33. Cu. #3. 3. 3. co. co. Red 000.. 0 mm. 03. No. 03. 3. 00. mo. 3. 3. 33. mo. co. 3. No mvmd 0mm. m E. 3. «3. 2. oo. oo. 3. No. 3 33. No. 03. 3. 3. mhvd 00w. v 3. 3. No mo. 33. 3 Om. no. 33. mo. v3. 3. No. 3 me.N 3— A m mo. mo. 3. 3. 3. co. 3. mo. 3. o_. 3. mo. 3. 3 mm _ .N m3; N No. NO. 3 No. No. 3 mo. .3. 3 3. mo. 3. 3. 3. 3M: man; — um 2 Na 2 3 w b 00“....wa 0:_a>:0w_m 02—052 00050000.:— 0uaata> 000380000005 00 3050030 50:55 0050305 0:805:00 ”0 020000040 105 REFERENCES Alpert, G. P. & Dunham, R. G. (1999). The force factor: measuring police use of force relative to suspect resistance. Washington, DC: Police Executive Research Forum. Bachman, R. & Paternoster, R. (2004). Statistics of criminology and criminal justice (2nd ed.). New York, NY: McGraw-Hill Companies. Bandura, A. (1973). Aggression: A social learning analysis. New Jersey: Prentice-Hall Inc. Bacharch, P. (1961). The Supreme Court, Civil Liberties, and the Balance of Interest Doctrine. The Western Political Quarterly, 14(2), 319-399. Barkan, S. E. & Snowden, L. L. (2001). Collective violence. Needham Heights, MA: Allyn and Bacon. Bittner, E. (1991). Thinking about police contemporary readings. In C. B. Klockars & S. D. Mastrofski (Eds.), The functions of police in modern society (pp. 35-51). New York: McGraw-Hill Inc. Bloch, P. B. & Anderson, D. (1974), Policewomen on Patrol, Final Report. Washington, DC: Police Foundation. Block, R, Anderson, D., & Gervais, P. (1973). Policewomen on Patrol, Major Findings. Washington DC. The Police Foundation. Blum, P. V. (2006). The democratic national convention protests in Los Angeles: August 12-17, 2000. Guild Practitioner, 63(1), 47-56. Boudreau, V. (2001). Contentious interactions and state repression: Toward an interpretive framework. Paper presented at the Conference on Mobilization and Repression: What we know and where we should go from here? June 22, University of Maryland, College Park, MD. Brodeur, J. P. (1998). Introduction: The policing of protest in Western democracies. In D. Della Porta & H. Reiter (eds.), Policing Protest: The control of mass demonstrations in Western democracies. Minneapolis and London: University of Minnesota Press. Brown, B. (2006). Policing crowds and controlling collective violence: An examination of police crowd control in the United States. Paper presented at the 1th International Conference: Toward peaceful mass demonstration and public 106 protection. November 8, The Korean National Police Agency & The Korea Association of Police Science, Seoul, South Korea. Button, M., John, T., & Brearley, N. (2002). New challenges in public order policing: The professionalization of environmental protest and the emergence of the militant environmental activist. International Journal of the Sociology of Law, 30(1), 17-32. Campbell, Shannon, Chidester, Phil, Royer, Jason, Bell, & J arnel. (2004). Remote control: How mass media delegitimize rioting as social protest. Race, Gender & Class, 11(1), 158. Carter, G. L. (1987). Local Police Force Size and the Severity of the 19605 Black Rioting. The Journal of Conflict Resolution. 31(4), 601-614. Chermack, S. T., & Taylor, S. P. (1995). Alcohol and human physical aggression: Pharmacological versus expectancy effects. Journal of Studies on Alcohol, 56(4), 449-456. Cheung, W. (2002). Regulation of public meetings and public processions in San Francisco. New York and London. Christopher, W. (1991). Report of the independent commission on the Los Angeles Police Department. Los Angeles, CA: City of Los Angeles. Clayton, G. (2000). Reclaiming public ground: the right to peaceful assembly. The Modern Law Review Limited. Cox v. New Hampshire, 312 US. 569 (1941). Dalton, R. J. (1988). Citizen politics in Western democracies. Chatham, NJ: Chatham House. DeLint, W. (2005). Public order policing: A touch act to follow? International Journal of the Sociology of Law, 33, 179-199. Della Porta, D. & Reiter, H. (1998). Policing protest: The control of mass demonstrations in Western democracies, Minneapolis and London: University of Minnesota Press. Della Porta, D., Andretta, M., Mosca, L., & Reiter, H. (2006). Globalization from below: Transnational activities and protest networks. Minneapolis and London: University of Minnesota Press. D’Emilio, J. (2005). The 1979 March’s place in history. The Gay & Lesbian Review Worldwide, 12(2), 33-35. 107 Dotson, B. (1974). Social planning and urban violent: An extension of McElroy and Singell. Urban Affairs, 9, 283-301. Dull, P. S. (1948). South Korea Constitution. Far Eastern Survey, 17(17), 205-207. Earl, J ., Soule, S. A., & McCarthy, J. D. (2003). Protest under fire? Explaining the policing of protest. American Sociological Review, 68(4), 581-606. Ericson, R. & Doyle, A. (1999). Globalization and the policing of protest: the case of APEC 1997. British Journal of Sciology, 50(4), 589-608. Farmer, D. J. (1968). Civil disorder control. Public Administration Service, Illinosis. Farrer, D. A. (2004). The First Amendment. The Foundation Press, Inc. Fisher, W. (2001). Personal interview by J. Noakes with commanding officer of the Philadelphia Police Department’s Civil Affairs Unit, 31 May. Fowler, J. (1999). The United Stated and South Korean Democratization. Political Science Quarterly, 114(2), 265-288. Gamson, W. Z. (1990). The strategy of social protest. Reprint, Homewood, IL: Dorsey. Gaubatz, K. T. (1995). Crime in the public mind. Ann Arbor: University of Michigan Press. Giancolar, P. R. (2002). Alcohol-related aggression during the college years: Theories, risk factors, and policy implications. Journal of Studies on Alcohol, 14, 129-139. Gilham, P. F. & Marx, G. T. (2000). Complexity ad irony in policing and protesting: The world trade organization in Seattle. Social Justice, 27(2), 212-236. Goode, E. (1992). Collective behavior. Fort Worth: Harcourt Brace J ovanovich. Grant, D. R. (2000). Perceived gender difference in policing: The impact of gendered perceptions of officer-situation “fit.” Women & Criminal Justice, 12(1), 53-74. Gurr, T. (1967). The conditions of civil violence: First tests of a causal model. Center of Inertational Studies, Priceton University. Gurr, T. (1970). Why men rebel. Princeton, NJ: Princeton Univ. Press. Hale, D. C. (1992). Women in policing. In G. W. Cordner & D. C. Hale (Eds.), What Works in Policing? (pp. 125- 142). Cincinnati: Anderson Publishing Co. and Academy of Criminal Justice Sciences. 108 Harris, C. V. (1997). Causes of civil violence. Berkeley McN air Journal, 5. Herbert, S. (2001). Hard charger or station queen? Policing and the masculinist state. Gender, Place and Culture, 8(1), 55-71. Hibbs, D. (1973). Mass political violence: A cross-national causal analysis. New York: John Wiley. Hoffman, P. B. & Hickey, E. R. (2005). Use of force by female police officers. Journal of Criminal Justice. 33(2), 145-151. Hunt, J. (1999). Police accounts of normal force. In V. Kappeler (Ed.), The police and society (3rd Edition) (pp. 306-324). Prospect Heights, Illinois: Waveland Press Incorporated. Hwang, E. G., McGarrell, E. F., & Benson, B. L. (2005). Public satisfaction with the South Korean Police: The effect of residential location in a rapidly industrializing nation. Journal of Criminal Justice, 33(6), 585-599. James, F. 8., Jr., & Wolfgan, M. E. (1972). Collective violence. Aldine. Atherton Inc. J anowitz, M. (1968). Social control of escalated riots. University of Chicago Press. Jefferson, T. (1990).“ The case against paramilitary policing. Milton Keynes, Buckingham, England: Open University Press. Jeffries, V., Turner, R. H., & Morris, R. T. (1971). The public perception of the Watts riots as social protest. American Sociological Review, 36(3), 443-451. Jo, B. K. (2006). Study on social, economic cost of violent demonstrations. Korean Development Institute Press. Kim, C. J. (2004). Study on Law of demonstrations. Ph.D. dissertation, Don ga University, Pusan, South Korea. Koenig, E. J. (1978). An Overview of Attitudes Toward Women in Law Enforcement. Public Administration Review. 38 (3), 267-275. Koopmans, R. (1993). “The dynamics of protest waves: West Germany, 1965 to 1989.” American Sociological Review, 58(5), 637-658. Korea National Police Agency. (2001). Korea police annual report. Seoul, South Korea. Korea National Police Agency. (2002). Korea police annual report. Seoul, South Korea. Korea National Police Agency. (2003). Korea police annual report. Seoul, South Korea. 109 Korea National Police Agency. (2004). Korea police annual report. Seoul, South Korea. Korea National Police Agency. (2005). Korea police annual report. Seoul, South Korea. Korea National Police Agency. (2006). Korea police annual report. Seoul, South Korea. Kwon, J. B. (2006). Lipstick and teargas: State and labor violence in South Korea in the wake of the Asian financial crisis (1997). Department of Politics and International Relations, Scripps College, Claremont, CA. Lab, S. P., & Das, D. K. (2003). International perspectives on community policing and crime prevention. Upper Saddle River, NJ: Prentice Hall. Lee, H. W. (2006). Study on demonstrations in South Korea. Paper presented at the 1th International Conference: Toward peaceful mass demonstration and public protection. November 8, The Korean National Police Agency & The Korea Association of Police Science, Seoul, South Korea. Lee, K. H. (1996). Comparative study on demonstrations. Korean National Police University. LeGrande, J. L. (1967). Nonviolent civil disobedience and police enforcement policy. The Journal of Criminal Law, Criminology, and Police Science, 58(3), 393-404. Lieske, J. A. (1978). The conditions of racial violence in American cities: A development synthesis. American Political Science Review, 72, 1324-1340. Long, J. S. (1997), Regression models for categorical and limited dependent variables. Sage Publications, Thousand Oaks, CA. Lonsway, K. (2001). Police women and the use of force. Law and Order, 49(7), 109-1 14. Lundman, R. J. (1999). Policing protest: The control of mass demonstrations in Western democracies. Contemporary Sociology, 28(2), 226-. Manddala, G. S. (1983). Limited-dependent and qualitative variables in econometrics. Cambridge: Cambridge University Press. Martin, S. E. (1997). Women officers on the move: An update on women in policing. In R. G. Dunham & G. P. Alpert (Eds.), Critical Issues in Policing (3rd Edition). Prospect Heights, Illinois: Waveland Press Incororated. Marx, G. (1998). Afterword: Some reflections on the democratic policing of demonstrations. In D. Della Porta and H. Reiter (eds.), Policing protest: The control of mass demonstrations in Western democracies. Minneapolis and London: University of Minnesota Press. 110 Mason, A. J ., Stephenson, D. G., & Jr. (1999). American Constitutional Law: Introductory essays and selected cases. Prentice Hall, Inc. McAdam, D. (1982). Political process and the development of black insurgency, 1930-1970. Chicago, IL: University of Chicago Press. McCarthy, J. (1997). The Globalization of Social Movements, In J. Smith, C. Chatfield and R. Pagnucco (eds.), Transnational Social Movements and Global Politics. Syracuse, N.Y.: Syracuse University Press. McCarthy, J. & McPhail, C. (1998). The social movement society. In D. S. Meyer & S. Tarrow (Eds.), The institutionalization of protest in the United States (pp. 83-110). Boulder, CO: Rowman and Littlefield. McCarthy, J ., McPhail, C., & Smith, J. (1996). Images of protest: Dimensions of selection bias in media coverage of Washington demonstrations, 1982 and 1991. American Sociological Review, 61(3), 478-499. McPhail, C. (1971). Civil disorder participation: A critical examination of recent research. American Sociological Review, 36(6), 1058—1073. McPail, C., Schweingruber, D., & McCarthy, J. (1998). Policing protest in the United States: 1960-1995. In D. della Porta and Herbert Reiter (eds.), Policing Protest: The control of mass demonstrations in Western democracies. Minneapolis, MN: University of Minnesota Press. Milton, C. (1972). Women in Policing. Washington, DC. Police Foundation, 13-28. Morgan, W. & Clark, T. (1973). The causes of racial disorders: A grievance-level of explanation. American Sociological Review, 38, 611-624. Morrison, D. & Tumber, H. (1988). Journalists at war: The dynamics of news reporting during the falklands conflict. London: Sage. Myers, D. J. & Li, Y. E. (2001). City conditions and riot susceptibility reconsidered. University of Notre Dame. Nahm, A. c. (1996). Korea: A history of the Korean people (2"d ed.). Seoul: Hollym. Nalla, M. K. (2006). An assessment of the protest and protest policing in the United States. Paper presented at the 1th International Conference: Toward peaceful mass demonstration and public protection. November 8, The Korean National Police Agency & The Korea Association of Police Science, Seoul, South Korea. Nalla, M. K., & Hwang, E. (2004). Relations between police and private security oflicers in South Korea. Unpublished manuscript, East Lansing, MI. 111 National Lawyers Guild. (2004). Legal observer training manual. Available form: http://www.tm{.org/resources/LO Mgnualpdf (Accessed 1 February 2007). Noakes, J. A., Klocke, B. V., & Gillham, P. F. (2005).Whose streets? police and protester struggles over space in Washington, D. C., 29-30 September 2001. Policing & Society, 15(3), 235-254. Olzak, S. (1989). Analysis of events in studies of collective actions. Annual Review of Sociology, 15, 119-141. O’Neill, K. (2004). Transnational protest: states, circuses, and conflict at the frontline of global politics. International Studies Review, 6(2), 233-251. Ong, P. & Hee, S. (1993). Losses in the Los Angeles civil unrest. Los Angeles: Center for Pacific Rim Studies (UCLA). Opal, T. (2005). Toledo community prevents neo-nazi march, fights back against racist police. Turning the Tide. Culver City: 18(5), 5-10. Palmer, K. E. (1998). Constitutional Amendments, 1978 to the present. Gale Group. Paoline, E. A. & Terrill, W. (2004). Women police officers and the use of coercion. Women & Criminal Justice, 15(3/4), 97-119. Pyo, C. W. (2001). Policing: The past. Crime and Justice International, 24, 5-6. Pyo, C. W. (2002). An empirical research on the public perception of the police and various police reform initiatives in Korea: 1999-2001. Paper presented at the Asian Association of Police Studies, Hong Kong. Rainey, H. G. (2003). Understanding and managing public organizations. San Francisco, CA: Jossey-Bass. Reicher, S., Stott, C., Cronin, P., & Adang, O. (2004). An integrated approach to crowd psychology and public order policing. Policing, 27(4), 558-572. Schofield, D. L. (1994). Controlling public protest: First Amendment implications. The FBI Law Enforcement Bulletin, 63(11), 25-32. Schweingruber, D. (2000). Mob sociology and escalated force: sociology’s contribution to repressive police tactics. The sciological Quarterly, 41(3), 371-389. Seaman, J. N. (1939). Constitutional Law: Freedom of the press: Freedom of speech and assembly: Police power. Michigan Law Review, 37(4), 609-613. 112 Sherman, L. J. (1973). A psychological view of women in policing. Journal of Police Science and Administration, 1(4), 383-394. Sherman, L. J. (1975). An evaluation of policewomen on patrol in a suburban police department. Journal of Police Science and Administration, 3(4), 434-438. Sherman, L. W. (1993). Defiance, deterrence, and irrelevance: A theory of the criminal sanction. Journal of Research in Crime and Delinquency, 30(4), 445-473. Sherman, L.W., & Berk, R. A. (1984). The specific deterrent effects of arrest for domestic assault. American Sociological Review, 49(2), 261-272. Sheptycki, J. (1995). Transnational policing and the makings of a postmodern state. British Journal of Criminology, 35, 613-635. Singell, L. D. (1971). The socio-economic causes of the recent urban disorders: some empirical evidence. Land Economics, 47(3), 225-234. Skolnick, J. H. (2005). Democratic policing confronts terror and protest. Syracuse Journal of International Law and Commerce, 33(1), 191—212. Smith, C. E. (2004). Constitutional rights: myths and realities. Wadsworth/Thomson Learning Academic Resource Center. . Snyder, D. & Kelly, W. R. (1977). Conflict intensity, media sensitivity and the validity of newspaper data. American Sociological Review, 42, 105-123. Sorenson, G. P. (1991). The public forum doctrine and its application in school and college cases. Journal of Law and Education, 20(4), 445. Spilerman, S. (1976). Structural characteristics of cities and the severity of racial disorders. American Sociological Review, 41, 771-793. Stockdill, B. C. (1996). Multiple oppressions and their influence on collective action: The case of the AIDS movement. Ph.D. dissertation. Department of Sociology, Northwestern University, Evanston, IL. Stone, G. R., Seidman, L. M., Sunstein, C. R.,& Tushnet, M. V. (1999). The First Amendment. Aspen Law & Business. Stuart, T. J. (Mar 23, 1988). Special to the New York. New York Times: Court Eases Bar to Embassy Protests. New York, NY. A.20. Taeusch, C. F. (1952). Freedom of assembly. Ethics, 63(1), 33-43. Tilly, C. (1978). From mobilization to revolution. Reading, MA: Addison Wesley. 113 Tilly, C. (1983). Speaking your mind without elections, surveys, or social movements. Public Opinion Quarterly, 47, 461-478. Tilly, C. (2000). Spaces of contention. Mobilization, 5(2), 135-151. Tyler, T. R. (2003). Procedural justice, legitimacay, and the effective rule of law. Crime and Justice, 30, 283. Vitale, A. S. (2005a). Innovation and institutionalization: Factors in the development of “quality of life” policing in New York City. Policing & Society, 15(2), 99-124. Vitale, A. S. (2005b). From negotiated management to command and control: How the New York Police Department polices protests. Policing & Society, 15(3), 283—304. Waddington, P. (1994). Liberty and ordersPublic order policing in a capital city, London : UCL Press. Waddington, P. (1998). Controlling protest in contemporary historical and comparative perspective. In D. Della Porta and H. Reiter (eds.), Policing protest: The control of mass demonstrations in Western democracies. Minneapolis and London: University of Minnesota Press. Waddington, P. (1999). Swatting police paramilitarism: A comment on Kraska and Paulsen. Policing & Society, 9(2), 125-141. Wanderer, J. (1969). An index of riot severity and some correlates. American I Sociological Review, 74, 500-505. Weissman, R. (2000). Puppets, protesters and police. Multinational Monitor, 21(5), 24-29. Werner, W. (1938). Public order and the right of assembly in English and the United States: A comparative study. The Yale Law Journal, 47(3), 404-432. Wisler, D., & Giugni, M. (1999). Under the spotlight: The impact of media attention on protest policing. Mobilization, (4), 171-187. Wilson, O.W. (1967). Civil disturbances and the rule of law. The Journal of Criminal Law, Criminology, and Police Science, 58(2), 155-159. Worden, R. E. (1995). The causes of police brutality: Theory and evidence on police use of police. In W.A. Geller and H. Toch (Eds.), Police violence: Understanding and controlling police abuse of force (pp. 23-51). New Haven, Connecticut: Yale University Press. 114 Wright, B. R. E, Caspi, A., Moffitt, T. E., & Paternoster, R. (2004). Does the perceived risk of punishment deter criminally prone individuals? Journal of Research in Crime and Delinquency, 41(2), 180-213. Yonhapnews Online (2007). Stop of financial support for organizations committing violent demonstration. Feb.09. http://app.vonhapnews.co.k¢/YNA/Basic/article/search/YIBW showSearchArticle .aspx?searchpart=article&searchtext=contents id=AKR200702091 14000052 Accessed Feb.09, 2007. Yu, Y. J. (2003). Study on reforming law on assembly and demonstration. MA Dissertation, Yonsei University, Seoul, South Korea. 115 llllfllllllfllllfllllfllllllljlllllll