<%BANNER%>

Establishing Legitimacy and Democratic Rule of Law in Latin America

Permanent Link: http://ufdc.ufl.edu/UFE0024927/00001

Material Information

Title: Establishing Legitimacy and Democratic Rule of Law in Latin America Their Impact on Victimization in 12 Countries
Physical Description: 1 online resource (181 p.)
Language: english
Creator: Tennyson, Kristin
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: corruption, guardianship, latin, law, legitimacy, police, victimization
Criminology, Law and Society -- Dissertations, Academic -- UF
Genre: Criminology, Law, and Society thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: This dissertation examines the effect of democratization and legitimacy of law and legal authorities on victimization among individuals in 12 Latin American countries, using data from the Latin American Public Opinion Project (LAPOP). Much research in Latin America either focuses on individual countries within the region or lumps a group of dissimilar countries together in order to examine trends and patterns of victimization. However, scholars agree that the political contexts of each country can be vastly different in many cases. Therefore, this dissertation also examines the differences among the countries with regard to legitimacy and democratization as well as the differences in victimization. Furthermore this dissertation compares results for individual victimization for each country to the aggregated results from responses of individuals across the 12 Latin American countries.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Kristin Tennyson.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Lanza-Kaduce, Lonn M.
Local: Co-adviser: Parker, Karen F.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024927:00001

Permanent Link: http://ufdc.ufl.edu/UFE0024927/00001

Material Information

Title: Establishing Legitimacy and Democratic Rule of Law in Latin America Their Impact on Victimization in 12 Countries
Physical Description: 1 online resource (181 p.)
Language: english
Creator: Tennyson, Kristin
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: corruption, guardianship, latin, law, legitimacy, police, victimization
Criminology, Law and Society -- Dissertations, Academic -- UF
Genre: Criminology, Law, and Society thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: This dissertation examines the effect of democratization and legitimacy of law and legal authorities on victimization among individuals in 12 Latin American countries, using data from the Latin American Public Opinion Project (LAPOP). Much research in Latin America either focuses on individual countries within the region or lumps a group of dissimilar countries together in order to examine trends and patterns of victimization. However, scholars agree that the political contexts of each country can be vastly different in many cases. Therefore, this dissertation also examines the differences among the countries with regard to legitimacy and democratization as well as the differences in victimization. Furthermore this dissertation compares results for individual victimization for each country to the aggregated results from responses of individuals across the 12 Latin American countries.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Kristin Tennyson.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Lanza-Kaduce, Lonn M.
Local: Co-adviser: Parker, Karen F.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024927:00001


This item has the following downloads:


Full Text

PAGE 1

ESTABILISHING LEGITIMA CY AND DEMOCRATIC RU LE OF LAW IN LATIN AMERICA: THEIR IMPACT ON INDIVID UAL VICTIMIZATION IN 12 COUNTRIES By KRISTIN MARIE TENNYSON A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009 1

PAGE 2

2009 Kristin Marie Tennyson 2

PAGE 3

To my Susan who inspire me every day. 3

PAGE 4

ACKNOWLEDGMENTS I thank my parents and family for their en couragement and unwavering support throughout my academic career. I particularly thank my Mom for helping with Susan so I could focus on my work. I also thank Tommy and Susan for hanging in there these past few years while Mommy worked on her big paper. Without your love and understanding I could not have accomplished all that I have. I thank my husband for telling me Whatever it takes, you will finish this. You are too good to me. I thank Pat and Tom Graham for supporting me and helping with Susan. All your help truly made a difference. I thank my Dad for all his encouragement and for being so proud of his daughter the doctor. I thank my Un cle Stephen and Aunt Kathy for their patience through this long process and for aiding and guidi ng me safely through graduate school. I thank my Aunt Susan, who I know was on my shoulder in this process, urging me on and giving me strength. Words cannot express how grateful I am to you; I hope I made you proud. I thank my Saskia Santos for her endless pa tience, hours of advice, and many late night chats. I might not always say it, but Saskia, you are truly my best friend and I could not have done this without you. I thank Karen F. Parker and Lonn Lanza-Kaduce for their numerous hours of work on this dissertation project. I thank th em as well for their friendship and invaluable advice. I thank Charles H. Wood for his help and advice and fo r continuing to work w ith me throughout my entire graduate career. I thank Timothy Clark for coming to my aid on this project and for understanding my passion for this ar ea of research. I also thank Su sanna Smith for taking part in my dissertation project and my gra duate career. I am grateful for her participation. Lastly, to all my former professors and fellow graduate students I give my thanks for pushing me to excel and challenging me at every turn. 4

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........9 LIST OF FIGURES.......................................................................................................................11 CHAPTER 1 THE CONTEXT OF VICTIMIZATION...............................................................................13 Introduction................................................................................................................... ..........13 Background and Significance.................................................................................................16 2 DEMOCRACY, LEGITIMACY & CRIMINAL JUSTICE: THE EFFECTs OF CORRUPTION..................................................................................................................... ..22 A History of Violence.............................................................................................................22 Democracy in Latin America..................................................................................................24 The Role of Legitimacy......................................................................................................... .29 Legitimacy & Legal Authorities......................................................................................32 Legitimacy of Police, Courts and Government...............................................................33 Country Levels of Legitimacy................................................................................................34 More Legitimate Countries..............................................................................................35 Costa Rica................................................................................................................35 Uruguay....................................................................................................................37 Chile.........................................................................................................................39 Countries with Lower Levels of Legitimacy...................................................................40 Bolivia......................................................................................................................40 Peru...........................................................................................................................42 El Salvador...............................................................................................................44 Guatemala.................................................................................................................47 Countries of Average Legitimacy...................................................................................50 Mexico......................................................................................................................50 Brazil........................................................................................................................51 Venezuela.................................................................................................................53 Dominican Republic.................................................................................................56 Jamaica.....................................................................................................................58 Conclusions.............................................................................................................................59 3 CRIMINOLOGICAL THEORY: LEGITIMA CY, CRIME AND VICTIMIZATION.........62 Theory.....................................................................................................................................62 Patterns of Victimization...................................................................................................... ..63 Legitimacy and Increases in Crime........................................................................................68 5

PAGE 6

Legitimacy and Crime: The United States......................................................................68 Legitimacy and Crime: Latin America............................................................................71 Empirical H ypotheses.............................................................................................................72 Demographic Variables...................................................................................................72 Theoretical Variables.......................................................................................................73 Democracy...............................................................................................................73 Court justice.............................................................................................................73 Trust of police..........................................................................................................74 Vigilante justice........................................................................................................74 4 RESEARCH DESIGN AND METHODS..............................................................................75 Data and Sample.....................................................................................................................75 More Legitimate Countries.....................................................................................................7 6 Costa Rica Data...............................................................................................................7 6 Uruguay Data...................................................................................................................77 Chile Data..................................................................................................................... ...77 Countries with Lower Levels of Legitimacy..........................................................................77 Bolivia Data.....................................................................................................................77 Peru Data.........................................................................................................................78 El Salvador Data..............................................................................................................7 8 Guatemala Data...............................................................................................................78 Countries of Average Legitimacy...........................................................................................79 Mexico Data....................................................................................................................79 Brazil Data.................................................................................................................... ...79 Venezuela Data................................................................................................................7 9 Dominican Republic Data...............................................................................................80 Jamaica Data................................................................................................................... .80 Advantages.............................................................................................................................80 Limitations.................................................................................................................... ..........81 Measures.................................................................................................................................82 Dependent Variable.........................................................................................................82 Independent Variables.....................................................................................................83 Theoretical Variables.......................................................................................................84 Analytic Strategy.............................................................................................................. ......85 5 PRELIMINARY RESULTS...................................................................................................86 Sample Characteristics............................................................................................................86 Dependent Variable............................................................................................................. ...86 Independent Variables............................................................................................................87 Theoretical Variables.......................................................................................................... ....92 6 VICTIMIZATION IN LATIN AMERICA..........................................................................104 Bivariate Analyses and Multi collinearity Diagnostics.........................................................104 Latin American Region.................................................................................................104 6

PAGE 7

More Legitimate Countries............................................................................................105 Costa Rica..............................................................................................................105 Uruguay..................................................................................................................105 Chile.......................................................................................................................105 Countries with Lower Levels of Legitimacy.................................................................106 Bolivia....................................................................................................................106 Peru.........................................................................................................................106 El Salvador.............................................................................................................106 Guatemala...............................................................................................................107 Countries of Average Legitimacy.................................................................................107 Mexico....................................................................................................................107 Brazil......................................................................................................................108 Venezuela...............................................................................................................108 Dominican Republic...............................................................................................108 Jamaica...................................................................................................................108 Multivariate Analyses.......................................................................................................... .109 Latin America Region...................................................................................................109 More Legitimate Countries............................................................................................110 Costa Rica..............................................................................................................110 Uruguay..................................................................................................................110 Chile.......................................................................................................................111 Countries with Lower Levels of Legitimacy.................................................................111 Bolivia....................................................................................................................111 Peru.........................................................................................................................112 Guatemala...............................................................................................................112 El Salvador.............................................................................................................113 Countries of Average Legitimacy.................................................................................113 Mexico....................................................................................................................113 Brazil......................................................................................................................114 Venezuela...............................................................................................................114 Dominican Republic...............................................................................................114 Jamaica...................................................................................................................115 7 VICTIMIZATION AND LEGITIMACY............................................................................120 Bivariate Analyses and Multi collinearity Diagnostics.........................................................120 Latin American Region.................................................................................................120 More Legitimate Countries............................................................................................121 Costa Rica..............................................................................................................121 Uruguay..................................................................................................................121 Chile.......................................................................................................................122 Countries with Lower Levels of Legitimacy.................................................................122 Bolivia....................................................................................................................122 Peru.........................................................................................................................122 El Salvador.............................................................................................................123 Guatemala...............................................................................................................123 Countries of Average Legitimacy.................................................................................124 7

PAGE 8

Mexico....................................................................................................................124 Brazil......................................................................................................................124 Venezuela...............................................................................................................125 Dominican Republic...............................................................................................125 Jamaica...................................................................................................................126 Multivariate Analyses.......................................................................................................... .126 Latin America Region...................................................................................................126 More Legitimate Countries............................................................................................128 Costa Rica..............................................................................................................128 Uruguay..................................................................................................................128 Chile.......................................................................................................................129 Countries with Lower Levels of Legitimacy.................................................................130 Bolivia....................................................................................................................130 Peru.........................................................................................................................130 Guatemala...............................................................................................................131 El Salvador.............................................................................................................132 Countries of Average Legitimacy.................................................................................133 Mexico....................................................................................................................133 Brazil......................................................................................................................134 Venezuela...............................................................................................................134 Dominican Republic...............................................................................................135 Jamaica...................................................................................................................136 8 DISCUSSION AND CONCLUSIONS................................................................................141 Latin American Region vs. Country Specific Models..........................................................141 Marital Status................................................................................................................. 142 Age................................................................................................................................142 Children.........................................................................................................................144 Sex.................................................................................................................................145 Level of Education........................................................................................................146 Socio-economic Attainment..........................................................................................147 Race...............................................................................................................................148 Discussion.............................................................................................................................149 Legitimacy of Law.............................................................................................................. ..150 Level of Democracy......................................................................................................152 Court Justice..................................................................................................................154 Trust of Police...............................................................................................................155 Vigilante Justice............................................................................................................158 Limitations.................................................................................................................... ........159 Future Research....................................................................................................................163 Theory in the Latin American Contexts...............................................................................164 BIOGRAPHICAL SKETCH.......................................................................................................180 8

PAGE 9

LIST OF TABLES Table page 1-1 Latin America & United States Estimated Homicide Ra te (per 100,000), 2002...............202-1 Mean Level of Each Countrys Four Indicators of Legitimacy Compared with the Grand Means for All Countries..........................................................................................615-1 Descriptive Statistics for Variables us ed in Costa Rica, Uruguay and Chile Models.......965-2 Descriptive Statistics for Victimized Sub-samples from Costa Rica, Uruguay and Chile...................................................................................................................................975-3 Descriptive Statistics for Variables used in Bolivia, Peru, Guatemala and El Salvador Models................................................................................................................................985-4 Descriptive Statistics Victimized Sub-sa mples from Bolivia, Peru, Guatemala and El Salvador.............................................................................................................................995-5 Descriptive Statistics for Variables used in Mexico, Brazil, Venezuela, the Dominican Republic, and Jamaica Models......................................................................1005-6 Descriptive Statistics for Victimized S ub-Samples from Mexico, Brazil, Venezuela, the Dominican Republic, and Jamaica.............................................................................1015-7 Descriptive Statistics for Variable s used in Latin America Total Sample.......................1025-8 Descriptive Statistics for Victimized Sub-sample for Latin America Sample................1036-1 Logistic Regression Analysis based on Victimization for Total Latin America Sample: Non-Theoretical Model......................................................................................1166-2 Logistic Regression Analyses based on Victimization for Costa Rica, Uruguay and Chile: Non-Theoretical Model.........................................................................................1176-3 Logistic Regression Analyses on Victimi zation for Bolivia, Peru, Guatemala and El Salvador: Non-Theoretical Models..................................................................................1186-4 Logistic Regression Analyses on Victimi zation for Mexico, Brazil, Venezuela, the Dominican Republic, and Jamaica: Non-Theoretical Models.........................................1197-1 Logistic Regression Analysis based on Victimization for all Latin American Respondents.....................................................................................................................1377-2 Logistic Regression Analyses based on Victimization for Costa Rica, Uruguay and Chile.................................................................................................................................138 9

PAGE 10

7-3 Logistic Regression Analys is based on Victimization fo r Bolivia, Peru, Guatemala and El Salvador................................................................................................................1397-4 Logistic Regression Analyses on Victimi zation for Mexico, Brazil, Venezuela, the Dominican Republic, and Jamaica...................................................................................140 10

PAGE 11

LIST OF FIGURES Figure page 1-1 Estimated Homicide Rate from Latin America & the United States (per 100,000)..........21 11

PAGE 12

12 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy ESTABILISHING LEGITIMA CY AND DEMOCRATIC RU LE OF LAW IN LATIN AMERICA: THEIR IMPACT ON VICT IMIZATION IN 12 COUNTRIES By Kristin Marie Tennyson December 2009 Chair: Lonn Lanza-Kaduce Cochair: Karen F. Parker Major: Criminology, Law and Society This dissertation examines the effect of de mocratization and legitimacy of law and legal authorities on victimization among individuals in 12 Latin American countries, using data from the Latin American Public Opinion Project (LAPOP). Much research in Latin America either focuses on individual countries within the region or lumps a group of dissimilar countries together in order to examine trends and pattern s of victimization. However, scholars agree that the political contexts of each country can be vastly different in many cases. Therefore, this dissertation also examines the differences among the countries with regard to legitimacy and democratization as well as the differences in victimization. Furthermore this dissertation compares results for individual victimization for each country to the aggregated results from responses of individuals across th e 12 Latin American countries.

PAGE 13

CHAPTER 1 THE CONTEXT OF VICTIMIZATION Introduction Violent crime and property victimization have become synonymous with the Latin American region. However, generally Latin America is either treated as a single unit when analyzing causes of victimization (Fajnzylber, Lederman, & Loayza, 1998; Gaviria & Pags, 1999; Neapolitan, 1994) or researchers examine vict imization within specific countries in Latin America (Gaviria & Velez, 2001; Ibarra, 2003; Klevens, Duque, & Ramrez, 2002; Mitchell & Wood, 1998). Studies of individual countries in Latin America, such as Brazil, Peru or Colombia, are able to account for the nuances of each countrys current context as reflected in individuals perceptions of their political situations. However, studies th at look at Latin America as a single unit are not comparative acro ss the region and may miss key differences. While the entire region of the Americas, in the aggregate, has the highe st rate of violence1 in the world, research strongly indicates that levels of viol ence are not equal across Latin American countries (Buvini & Morrison, 2000; Gaviria & Pags, 1999). Given that violence varies markedly throughout the region, it is neces sary to look closer at the individuals within these countries to determine what key differences, political, historical and so forth, are associated with victimization and how these may vary from country to country in th e region. The context in which victimization occurs may be the most importa nt indicator of why it ex ists. Therefore it is paramount that researchers explor e what predictors contribute to victimization within each country while taking into account the unique politi cal context of one country as compared to others in the region. Although much of the research and analysis of crime and victimization in Latin American has focused on violence, the cu rrent effort will attend to a wider range of 1 When measured through homicide rate s from the World Health Organization. 13

PAGE 14

victimization which may be related to perceptions of the political and le gal contexts. Because victimization is inextricably linked to offenses the research into both violent and nonviolent victimization and crime in Latin American holds re levance. Examining victimization within this larger context permits comparisons of how forces can differentially affect specific areas within Latin America, as well as reveal those factors most associated with victimization across the region. LaFree (2007) argues that histor ic changes, such as war or civil unrest can lead to a perfect storm (LaFree, 2007, p. 507) that crea tes a crime wave and increases victimization rates. In many countries in Latin America, the citizens are still reeling from civil wars (e.g., Guatemala, El Salvador and Ni caragua) or brutal dictatorship s and periods of repression (e.g., Brazil and Argentina). While some of these countr ies have been able to overcome these events and have politically and economi cally stabilized, others have not. These conditions and the unique context of each country need to be ta ken into account when studying victimization in Latin America. Economic or political turmoil has a direct effect on citizens in these countries and are part of the social ecology of victimiza tion (Smith & Jarjoura, 1989, p. 622). Because both macro and micro processes (Smith & Jarjoura, 1989, p. 622) are related to victimization, there is a need to go beyond situational and/or indi vidual characteristics to examine how legal authorities and political stability influence ones risk of being vi ctimized within distinct Latin American countries. LaFree (2007) called for an increase in compar ative research that examines patterns of crime and victimization in non-Western or non-indus trialized nations so th at criminologists can determine if the theories they have developed in one rather small corner of the world (Marsh, 1967, p. 6) apply in other countries. Oftentimes researchers ignore the broader institutional 14

PAGE 15

context in which crime occurs by focusing only on individual motivations and drives (Reed and Yeager, 1996; Willis, Evans, & LaGrange, 1999). While it is very hard to quantify political context even at the national level, attempts should be made to analyze the role of political change and uncertainty and other precursors to victimization in Latin America. Within this broad context, the goal of this dissertation is to provide a systematic examination of the predictors that lead to the likelihood of victimization within 12 Latin American countries that are dissimilar in recen t history, level of demo cratic stability, and legitimacy of legal authorities. First, I seek to determine if citizens perceptions of democracy and the legitimacy of legal authorities influen ce the likelihood of victimization. Second, I will examine the differential effects of predictors from country to count ry. This is critical to the dissertation because any evidence of differences in the effects of key predictors across these countries may be due to the unique political contex t of the country, further justifying the need for research that does not treat Latin America as a single unit. An underlying aspect of this dissertation is using traditional theoretical perspectives within the field of criminology and sociology when identifying predictors of victimiza tion and thus testing the applicability of these perspectives to the study of victimization within various Latin American countries. Accomplishing these specific aims will shed light on several general issues. It will help determine whether or not similar predictors of victimization are found in smaller groups or areas of Latin America. If differences are found, th e research can provide information about how researchers can disentangle Latin American countries into subgroups Finally, this project will help to clarify our understanding of victimizati on in the context Latin America and provide the opportunity to re-examine our theories of victimization. 15

PAGE 16

Background and Significance Latin America is noted as one of the most violent regions of the world (Fajnzylber, Lederman, & Loayza, 1998; Moser, 2006; Rose nberg, 1991; Rotker, 2002; Sherman, 2000). In public opinion polls, violence is at least the second most important policy concern throughout the region and is the primary concern in Argentina, El Salvador and Vene zuela (Latinobarmetro, 2006 as cited in Soares & Narito mi, 2007). Recent findings from the World Health Organization (WHO) (2007) indicate that the re gion of the Americas has the highest homicide rate in the world, eclipsing that of Sub-Saharan Africa2 (Krug, Dahlberg, Mercy, Zwi, & Lozano, 2002). While some countries in Latin American have homi cide rates far greater than those of the United States or European countries, other Latin American nations are relatively peaceful (See Table 1; Figure 1). Colombia has the highest homicide rate in La tin America as well as in the world, with a rate of 79.7 homicides per 100,000 citizens (World Health Organization). El Salvadors homicide rate is 45.6 per 100,000, while Brazils a nd Venezuelas rates ar e in the 30s (World Health Organization, 2007). In contrast, Chile s homicide rate is 5.3 and Perus is 3 per 100,000, which is well below the United States homicide rate of 6.2 (World Health Organization, 2007). This extreme disparity between the most violent na tions and those with rela tive peace dramatizes the need to understand the context of crime and victimization in each Latin American country. Both crime rates and rates of interpersonal vi olence are particularly high in countries that have recently experienced periods of extreme vi olence. Civil wars and periods of unrest have left behind broken families, poverty, unemplo yment and often an abundance of guns and ammunition; under these conditions, violence pr oliferates (Archer & Gartner, 1984; Buvini & 2 The region of the Americas includes Latin America as well as the United States and Canada. 16

PAGE 17

Morrison, 2000). For example, after the end of its 12 year civil war, El Salvadors homicide rate rose 36%. Research has shown that police e ffectiveness and judicial systems are also compromised in post-conflict nations, which adds to the already volatile environment (Morrison, Buvini & Shifter 2003, as cited in Frhling, Tulchin, & Golding, 2003). Political unrest is not the only factor that contributes to high crime ra tes in Latin America; issues of political and economic development, as well as culture and hi story, impact rates of violence in the region (Heinemann & Verner, 2006). Neapolitan (1994) argues that homicide rates in Latin America are driven not only by structural and demographic charact eristics, but also by cultural a nd historic forces. Neapolitan (1994) used homicide data from both the World Health Organization (WHO) and the International Criminal Police Organization (IN TERPOL) from 65 and 107 nations, respectively. These data represent homicide rates from 1986 to 1990. Analysis of the WHO homicide data showed that, at the national level, income in equality, gross domestic product (GDP) per person, population under 15 years of age and urban population were all significantly related to homicide rates. In addition, Neapo litan (1994) created a dummy variable to represent the Latin American region and found that it was the sing le greatest predictor of homi cide in the models. Results from the INTERPOL data indicated that, at the national level, income, GDP per capita, urban populations and the percent of population under 15 were all signifi cantly related to homicide rates, while controlling for the Latin American region (Neapolitan, 1994). The author concludes that there is something unmeasured in the mode l, besides the measures of development that drives homicide rates in Latin America (Neapolit an, 1994). He concludes th at there is something unique about the culture and history in the regi on that drives the high rates of violence. 17

PAGE 18

While Neapolitans (1994) research he lps understand aggregat e level rates of victimization, through homicide st atistics, there are se rious flaws in his research. Primarily, Neapolitan (1994) is unable to a ccount for culture or history in the models and assumes that the unexplained variance in his models is evidence of the cultural and historical effect. Secondly, Neapolitan (1994: 20) treats Latin America as a single unit in the analysis, yet admits that the region does not have uniform homi cide rates when broken down by country. Lastly, Neapolitans (1994) study uses homicide rates as a measure of violence/victimizati on. Homicide data are thought to be unrepresentative of victimization rates in the region because they do not capture the majority of crime, pr operty victimization (Buvini et al., 2005; Heinemann & Verner, 2006). What is important about this study, however, is that at the aggregate level, Neapolitan (1994) shows that context (his tory, politics, and econom ics) does matter when it comes to violence and victimization. There is a need to examine if a nd how these forces play a role in individual victimization. Current research on individual victimi zation does not include these important factors. Using data from the Latinobarmetro, a public opinion poll surveying 17 Latin American countries, Gaviria and Pags ( 1999) studied household victimizat ion in Latin America over a three year period (1996-1998). Household vict imization measured if any member of the respondents family had been assaulted, robbed or victimized in any way during the past twelve months. Consistent with homicide rates, victimization rates va ry across the region with Peru, Ecuador, Mexico, Venezuela, El Salvador and Gu atemala having the highest victimization rates (some as high as 40%) (Gaviria & Pags, 1999) Uruguay, Chile and Panama had the lowest victimization rates in the region. Results from the study also showed that household victimization increases as socioeconomic status city size and city-lev el population growth 18

PAGE 19

19 increase (Gaviria & Pags, 1999). Additio nally, the authors found that urban household victimization was negatively correlated with reported levels of confidence in criminal justice agencies including both the police and judiciary (Gaviria & Pags, 1999) Gaviria and Pags (1999) conclude that urban victimization results from an inability of the judiciary and police in the region to maintain order and public safety. While this study represents an improvement ove r research that relie s solely on homicide data as a measure of victimizati on, it has limitations due to the t ype of data used. First, the Latinobarmetro survey asks respondents if they or any members of their family have been victims of a crime in the past 12 months. Broadening the scope of victimization to the respondents family members may inflate the rates of victimization. Respondents of the survey were also restricted to urban ar eas in Latin America and wealthie r individuals were over-sampled (Gaviria & Pags, 1999). While the authors weighted the data to offset the number of wealthy respondents, there is no representa tion of rural or suburban house holds in the data (Gaviria & Pags, 1999). Existing scholarship is encouraging for La tin American scholars and criminologists; however, more research is needed to address limitations of prev ious research and to advance scholarship in the field. This dissertation project will address some of the limitations of previous research, and also attempts to apply a theoretical perspective to victimi zation in Latin America.

PAGE 20

Source: Pan American Health Organization, Health Su rveillance and Disease Management Area. He alth Statistics and Analysis Unit. PAHO Regional Mortality Database. Estimates based on World Population Prospects 2006 Revision. 2007 Table 1 1. Latin America & United States Estimated Homicide Rate (per 100,000), 2002 Argentina Brazil Chile Colombia Ecuador El Salvador Guatemala Mexico Peru Venezuela United States Homicide Rate (per 100,000) 6.8 31.7 5.3 79.7 16.2 45.6 24 11.4 3 6.2 34.5 20

PAGE 21

21 Figure 1. Estimated Homicide Rate from La tin America & the United States (per 100,000)

PAGE 22

CHAPTER 2 DEMOCRACY, LEGITIMACY & CRIMINAL JUSTICE: THE EFFECTS OF CORRUPTION A History of Violence Even prior to the first Europeans setting f oot on the Western Hemisphere, violence was a part of the social order for those living in what is now Latin America. The Aztec, Inca and Maya are the three most prominent Pre-Colombian soci eties and all establishe d their empires through military repression and the overthrow of neighboring peoples (Sherman, 2000; Wright, 2005). They were known for their aggression and brutal violence. The Spanish and Portuguese also had a long history of conflict; The Reconquest, or reclaiming of the Iberian peninsula from the Moors, lasted for 700 years prior to contact wi th the New World. Only in 1492 did the kingdoms of Spain unite under Ferdinand and Isabella, in celebration of the military expulsion of the Moors. The conquest of the New World became an extension of the pattern of violence experienced during the Reconquest of Iberia (Wiarda, 200, p. 78). Latin America underwent a colonial history typifi ed by violence and a c onsistent pattern of coercion (Sherman, 2000, p. 13). Rule of law has never been the true basi s of political power for most Latin American countries, but rather rule th rough threat and intimidation (R osenberg, 1991; Wiarda, 2001). Rule of law in the New World was absolutist hierarchical, authorit arian, top-down and emphasized the concept of unequa l application of the law (W iarda, 2001, p. 74). A two-class society arose in the Viceroyaltie s of Spain and Portugal. One cl ass was able to manipulate or ignore the law; the other class was burdened by it. A nobleman from Spain or Portugal, was exempted from paying taxes and making tributes, could not be imprisoned for debt or have his horse, arms or residence confiscated (Wiarda, 2001, p. 81). The lower classes were burdened by labor exploitation, heavy taxation, sumptuary laws and even slavery. The rights of the nobles most certainly did not extend to the masses or to the indigenous population in the region. Social 22

PAGE 23

order was based on a racial caste system wher e, each socioracial classification had its own rights, responsibilities, and ev en law courts in which differe ntial treatment was handed out, depending on ones rank (Wiarda, 2001, p. 100). This unequal applica tion of law set up a dualism in society and a lack of trust in these agencies of social order, such as the courts. Colonial society in Latin America was fraught with corrupt ion (Gibson, 1966; Fleischer, 1997; Ratliff & Buscaglia, 1997). A ccording to Gibson (1966, p. 107-108): venality, graft, peculation, and personal us e of public funds atte nded the operation of government at all levelsThe distinction between fees and bribes became blurredThis situation was rather one of normal and expect able corruption, within which an occasional figure stood out for resisting the problem. Obedezco, pero no cumplo is the most famous phrase in the political history of colonial Spanish America (Sherman, 2000, p. 48). This term was coined when colonial viceroys realized that the law, as it was issued from the crown in Spain, was often discordant with the way of life in the colonies. When imprudent or irrelevant laws were issued by the crown, the viceroys would write back in response, I obey, but I do not comply (Sherman, 2000, p. 48). They accepted the authority of the crown to create laws, but emboldened by distance, they exerted their own authority to simply ignore them. This simple response has laid much of the f oundation for the rule of law in Latin America, specifically those countries desc ending from Spanish rule. The crown, far away in Spain, was unable and sometimes unmotivated to enforce all of its decrees; profit was more motivating than rule of law and as long as revenue kept comi ng in from the viceroyalties, the crown would look the other way (Wiarda, 2001). As a result, laws hi storically are regarded as optional guidelines, rather than concrete boundaries for social inte raction. This conceptualiz ation of law as fluid rather than concrete and binding led to a great deal of fricti on between not only the upper and lower classes in the colonies, but between thos e loyal to the crown and those who enjoyed the 23

PAGE 24

legal freedoms and lack of strict legal boundaries in the New World. It was eventually the fight between the creollos and peninsulares the white rulers in the col onies and those born in either Spain or Portugal that helped push the co lonies towards independence (Wiarda, 2001). The history of modern Latin America has reinfo rced the arbitrary nature of law and is no less violent. Country after country has swung like a pendulum between dictatorship and democracy; some, like Peru, have experienced ev ery form of government in between the two. Likewise, country after country (excepting Colo mbia, Costa Rica and Venezuela) fell into the hands of a military, left-wing, or right-wing dict atorship during the 1960s and 70s (Valenzuela, 2004). Brazils elected government was overthrown in 1964 by a military coup. Similarly in 1966 Argentinas government fell in a coup. Chile, one of the more politically open governments in the region known for its rule of law and respec t of human rights, fell to the dictatorship of Augusto Pinochet in 1973 (Rosenberg, 1991; Sher man, 2000). It was a long slow climb for many of these countries to convert or re-convert to democracy. Braz il, for example, remained under military rule for nearly a quarter of a century. Guatemala was embroiled in a thirty year Civil War that lasted until 1996. However, one-byone over the ensuing tw o decades, these same countries in Latin America (re)converted to democratically elected governments. Democracy in Latin America It is important when discussing democracy to have a working definition of the concept, which is easier said than done. On e of few things that political scientists, sociologists, and researchers agree upon concerning democracy is th at there is no universally accepted working definition for scholars (Fischer, 1996; Fierlbec k, 1998; Snyder, 2000; Doorenspleet, 2005). The major problem is that democracy is an abstr act, imprecise idea and that there are varying shades of democracy (Fierlbeck, 1998, p. 11). There are liberal democracies, consolidated democracies, transitioning democracies, electo ral democracies, and so on and so forth. The 24

PAGE 25

definition laid forth here is by no means all inclusive or definitive, but it is a working definition that allows for future discussion of the topic as it related to Latin America. Democracy, stripped down to its most simp listic definition, is the achievement and maintenance of a political regime based on popular sovereignty, which is expressed through free and competitive elections (Dahl, 1971; Sanjun, 2000, p. 89). This definition is the operating definition for most democracies in the Latin American region; they are electoral democracies (Schedler, 1998, p. 93)1. However, casting a ballot is merely one of the many rights afforded to citizens under democratic rule of law. Citizenship, consists of assuring that each person is treated as a full member in a society of equa ls (Sanjun, 2000, p. 89). Therefore the definition should include the notion that in a democracy: The actions of officials are constrained by constitutional pr ovisions and commitments to civilian liberties; and governmen t candidates sometimes lose elections and leave office when they do. Freedom of speech, freedom to organize groups to contest elections, and reasonably equitable representation of varied viewpoints in the media are presumed to be preconditions for free and fair election. (Snyder, 1998, p. 25-26) This means that in addition to universal suffrage, citizens in Latin Ameri ca should feel safe that their governments will protect their civil, poli tical and social rights (Schedler, 1998; Domingo, 1999; Sanjun, 2000). Including this in the defini tion of democracy is important because it excludes the possibility of demo cratically elected officials, following the will of the people, using oppression against minority classes or speci fic racial groups (e.g., National Socialism in 1940s Germany). This protection and respect for citi zens rights is elemental to the democratic process and these rights should be guaranteed through legal processe s and constitutional structures (Fischer, 1996, p. 4). 1 For a discussion on the sub-types of democracy, s ee David Collier and Steven Levitsky, "Democracy with Adjectives: Conceptual Innovati on in Comparative Research," World Politics 49 (April 1997): 430-51. 25

PAGE 26

And so, the working definition of democracy used here is twofold: 1. Democratic governments have periodic free and fair elections in which a good portion of the adult population participates in order to ensu re government accountability and rule by the will of the majority. 2. Democratic governments protect th e civil, political and social rights of citizens in so much that all view points are valid, freedom of sp eech is protected, and minority voices are not silenced. One of their civil and social right s is to be free of victimizatio n by both their fellow citizens and their own governments. Democracy without such freedom is a contradiction in terms (Beetham, 2004, p. 62). However, there is no long standing history of certain governments in the region, democratic or otherwis e, behaving in this way. While the conversion to democracy in Latin America has been a great leap forward towards modernization (Sherman, 2000, p. 5), it has not had the desired stabilizing effect in the region (Baln, 2000; Tedesco & Barton, 2004; Valenzuela, 2004; Rodrig ues, 2006). Political democracy did not automatically generate a demo cratic rule of law (Caldeira & Holston, 1999, p. 694). On the contrary, violence and victimization have thrived in this new democratic era. As noted earlier, Latin America is one of the most violent regions in the world. These fledgling democracies have dealt ineffectively with their citizens needs and have failed to keep them safe from victimization by their fellow citizens (Baln, 2000, p. 5). Some point to the disjuncture between democratic, freely elected governments and a true participatory democratic populous as a root cause of the problem (Caldeira & Hols ton, 1999; ODonnell, 1999: Caldeira, 2000; Davis, 2006). In other words, there exists the first elem ent of democracy as outlined here, but not the second. While elections in Latin America are cleaner and more frequent than fifty years ago (Sherman, 2000, p. 137), citizens of some of these count ries do not feel they are politically a part of the nations in which they live. The lingering authoritarian political culture in many of these 26

PAGE 27

countries (e.g., Venezuela, Brazil and Peru) mean s that there is not a universality of civil, political and social rights in much of Latin America (Sanju n, 2000; Tedesco & Barton, 2004). Dissident voices are silenced; the poor and marginalized are ignored. In many cases, democracy has been used as a tool to replicate or conti nue authoritarian rule. Fo r example, in Venezuela Hugo Chavez was voted back into power after m ilitary rule in 1999. Similarly, Hugo Benzar was voted into power in Bolivia (1997) and Lucio Gutirrez was voted into power in Ecuador (2003) (Tedesco & Barton, 2004). In Brazil, the military police responsible for the brutal violence and mass disappearances during the military regime maintained their positions as the keepers of public order after the ratification of the cons titution in 1988 (Caldeir a & Holston, 1999). The adoption of a democratic constitution in Brazil actually led to an increase in police violence and torture practices (Anhen, 2003). Augusto Pinochet remained Commander-in-C hief of the Chilean armed forces until 1998, long after the end of hi s brutal dictatorship (Tedesco & Barton, 2004). The transition to democracy in Latin American never signaled a breaking with the old guard, but rather was a calculated response by Latin American elites to cha nged circumstances in order to preserve their power (Wiarda, 2001, p. 327). Under such circumstances, how can the citizens of Latin American countries feel that they can e ffect true change under the political system? Without such power why should they be bound by th eir governments rules? While the progress in the region has represented giant steps towa rds true democratic rule of law, (Lamounier, 2002), the lingering traits of author itarianism make it impossible for the true democratic rule of law to maintain social order among citizens in many of these countries (Rosenberg, 1991, p. 18). There still exists a duality in the applicati on of law whereby rules do not apply to certain members of society. As a result, the state lacks the popular legi timacy it needs to exert social order and stop violence. 27

PAGE 28

In order for citizens to truly feel a part of the political system the state has to enforce laws with uniformity and predictability; they have to institute procedural justice (Sanjun, 2000, p. 90). Few legal systems in Latin America enjoy such legitimacy2. In Caracas, Venezuela, for example, police resources are used to regul ate social conduct su ch as drug addiction, homosexuality and prostitution, none of which are currently cat egorized as crimes under the Venezuelan constitution (Sanjun, 2000, p. 98). Seri ous crimes, however, such as homicide, rape and robbery, only account for 3.59 percent of all arre sts made by the police in Caracas. In Brazil, police routinely invade favelas, breaking down doors, interrogati ng suspects, shooting first and asking questions later, all with little regard for citizens safety or while ignoring their legal rights; and most with the full knowledge and lack of condemnation of their superiors (Mitchell & Wood, 1999; Zaverucha, 2000; Paes-Machado & Noronha, 2002). When the state and its agencies of social control, such as the judiciar y and police, routinely ab use power, they are seen as morally bankrupt and corrupt a nd are generally distrusted; ther e is a breakdown in the social fabric and violence results (Concha-Eastman, 2000). These agencies of social control are repres entative of the state and within the Latin American context, the state, po lice and judiciary are intrinsica lly linked in the minds of the people: Not only is police work carried out with the sup port of a judicial frame-work that promotes arbitrariness and discretionary judgments on th e part of the agents, but also, to make matters worse, the internal regulations they must follow are based on military-style rules that privilege carrying out superiors orders and the loyalty to the institution over respect for citizens and the guaranteeing of free exer cise of their rights. (Sanjun, 2000, p. 97) 2 A notable except is Chile, see Frhling 2007. 28

PAGE 29

A cycle ensues in which a citizen is victimi zed, but feels that the police and judiciary are untrustworthy and turns to a form of vigilante justice, taking the law into his/her own hands (Johnson, 2007). This in turn creates mo re violence, more victimization. The Role of Legitimacy True social control and a reduc tion in social violence require that legitimacy be brought to these democratic governments and their agencies of social control, the legal authorities (Tyler, 1990, p. 57; Weber, 1968). It is important to underst and what legitimacy means in the context of social order. Lipset ( 1983, p. 64) defined legitimacy as the capacity of a political system to engender and maintain the belief that existing political instituti ons are the most appropriate and proper ones for the society. Tyler (2006, p. 375) de fines legitimacy as a psychological property of an authority, institution, or social arrangement that leads those c onnected to it to believe that it is appropriate, proper, and just. According to John Locke, government is not legitimate unless it is carried on with the consen t of the governed (Ashcroft, 1991, p. 524). Classical sociologist Max Weber (1968) argued that legi timacy actually comes in three forms: legitimacy is based on traditional authority (customs and traditions); le gitimacy is based on charismatic authority (the devotion to a charismatic ruler); legitimacy is base d on rational authority (t he process of creating and interpreting laws). It is not essential for a government to be democratic, or for a social institution to be morally corre ct in order for it to have le gitimacy. The practice of African slavery in the American South held a great deal of legitimacy as did the National Socialist Party (NAZI) in pre-World War II Germany (Fierlbec k, 1998; Smith, 2007). However, rule of law and procedural just are essential for democracy to be practiced to its full est definition (Tyler, 1990). In the case of Latin American democracies, legitimacy based on rule of law is the focus here. Tom R. Tyler (2007, p. 10) building off Webe r and the works of Zelditch (2001) and Beetham (1991), states that h aving legitimacy means that t hose in the community being 29

PAGE 30

regulated believe that their aut horities deserve to rule and ma ke decisions that influence the outcomes of the members of the community. In other words, the citizen s of a country believe that the decisions of their govern ment are right and corre ct because the govern ment itself if valid in their eyes. This type of rule is legitimate and the agencies under a government of this nature, such as police and courts, are followed with little resistance because they in turn hold legitimacy. Tyler (1990, p. 57) argues that a sense of legitimacy is the key to success of government, or legal authorities such as the police and courts, to rule effectively. It is this same legitimacy that gives the citizenry of a country a sense of obligation to follow the rules and laws of their county. When legitimacy exits, according to Tyler (2007, p. 10) laws are deferred to voluntarily. In effect, the citizenry will follow th e rules and laws of their government and defer to the will and power of criminal justice agenci es, even if they do not agree with the laws. This type of legitimacy is currently only present in the minority of Latin American countries (e.g., Chile) (Frhling, 2007). Furthermore, in many cases there has not historically been legitimacy of governance in Latin Ameri can countries. The governments preceding many of the democracies in Latin America were au thoritarian in nature, relying on exertion of governmental power and essentially strong arming of the citizenry in order to maintain social control. As Tyler (1990) points out, authoritarian ru le rarely succeeds because it is too economically costly and takes too much effort from the government to maintain (Meares, 2000). Also, because these governments lack legitim acy, they cannot count on the good will of the people to implement social order; rather th ey depend on intimidation and often violence to maintain order. This is the kind of rule that dominated most Latin American countries in the 1960s and 1970s. It is no wonder then that the very necessary component of legitimacy is lacking 30

PAGE 31

in these countries. The deep scar left by author itarian regimes has not ye t healed and therefore true democratic legitimacy cannot and has not taken root (Tadesco & Barton, 2004, p. 1). On the contrary, recent history has proved to the citizens in many of these countries that government is untrustworthy, corrupt and will enact violence on its own people. The citizens of nations in Latin America are deeply scarred by the political violence of au thoritarian regimes that preceded their current democracies (Tedesco & Barton, 2004). Even Colombia which, excepting a brief period in the 1950s, has held free a nd democratic elections since 1936, was racked with political violence and government corruption during the 1970s and 80s (Troyan, 2008). For this reason, the citizenry in many of these countries is detached from the political process and lacks faith in the ability of democracy to effect real change (Rosenberg, 1991; Sherman, 2000). Legitimacy is grown out of the experience of fair and reasoned authority (Smith, 2007). The citizens of most Latin American countries, da ting back to the Coloni al era, have never experienced the exercise of authority as be ing fair (Smith, 2007, p. 31) and therefore their governments lack legitimacy. This lack of legi timacy is responsible for most of the conflict within contemporary Latin American so cieties (Tedesco & Barton, 2004, p. 6). While democracy in name has taken root in Latin American, the legitimacy of many governments has not yet been earned. Tyler (2007, p. 11) argues that democratic societies require legitimacy because in effect, people become self -regulatory when they internalize the values of the government as right and morally correct. Wh en individuals have a normative commitment, a commitment to obeying the law and following le gal authorities (even wh en the citizen does not agree with the law), then clearly it becomes much easier for those legal authorities to first establish order and secondly maintain th at order (Tyler, 1990; 2007). This normative commitment comes only after intensive educatio n and socialization both of which take time 31

PAGE 32

and commitment not only on the part of the peopl e, but more importantly on the part of the government and legal authorities (Tyler, 1990, p. 65). In certain Latin American countries, legitimacy born out of such a tr adition is sorely lacking. Legitimacy & Legal Authorities Elemental to the success of a democratic gov ernment ensuring the rights of its people and securing its own legitimacy is th at the arms of enforcement of the government, legal authorities such as the police and courts, are seen as fair and uncorrupted (Tyler, 2007). When the police and courts focus on retribution and deterrence rath er than procedural fairness, they risk losing their legitimacy; for it is through fair and equitabl e application of the law that citizens accept the rules and regulations of thei r government (Tyler, 2006; Smith, 2007). Even more than their acceptance of laws, citizens will begin to re gulate their own behavior and obey the law. Unfortunately, in a number of Latin American c ountries procedural fairness, like legitimacy, is also wanting. More so than the court system, the police are the most visible legal authorities in most countries and for many come to represent their government and how it ope rates. In some ways, the legitimacy of the police aff ects the legitimacy of the governme nt itself. For there to be a legitimate police force in a democracy, citizens t ypically have to have a sense of fairness and equal application of the law on th e part of the police. This legitimacy is born out of several components. First, personal expe rience with the police is a key component in establishing their legitimacy. Research indicates that a belief in the legitimacy of police is linked to favorable evaluations of encounters with the police (Smith, 2007, p. 31). Secondly, legitimacy is based on the longstanding reputation and the social integration of lower cl asses into a political system. Smith (2007) points out that the po lice force in Great Britain did not initially enjoy widespread legitimacy. The lower and working classes, who were primarily regulated by the police force, 32

PAGE 33

were originally very antagonistic towards the police (Smith, 2007, p. 35). It was not until class struggles and divisions in Great Britain dissipated that the working class gradually grew to see the police as legitimate. This took place because the police exhibited a longstanding pattern of procedural fairness. When we turn our attenti on to Latin American countries and examine the longstanding patterns of interaction with citi zens during police encounters and in the court systems, we find that, the procedur al fairness of these agencies varies across the region and as a result, like democracy, there are varying shades of legal authorities legitimacy across these countries. Legitimacy of Police, Courts and Government If procedural justice and fair application of the law build legitimacy into the legal authorities of a country, corruption is what will tear it down. Seligson (2002) did a study where he explored the role of corr uption at the level of the polic e, municipal government, public officials and other legal authoriti es. Using data from four Latin American countries, El Salvador, Nicaragua, Bolivia and Paraguay, he tested how individuals experien ces with bribery or coercion from legal authorities a ffected their perception of the legitimacy of their governments. Legitimacy was measured as trust of legal authorit ies, respect for the political apparatus in their country, perceptions of fair trials in the country. Roughly 10,000 respondents in the four countries, between approximately 1,500 and 2,500 pe r country, were interviewed regarding the issue. Results from Seligsons (2002) study show ed that corruption ha s a significant negative impact on legitimacy. Even controlling for party affiliation, corruption has a corrosive effect on legitimacy of the government and le gal authorities (Seligson, 2002). We see from survey research that legitimacy and the procedural fairness of police is not universal across Latin American countries. For exam ple, despite the increasing violence in Chile, the police enjoy a good deal of popular support and legitimacy (Fr hling, 2007). In contrast the 33

PAGE 34

legitimacy of Mexican police is nearly non-existe nt. In a recent survey 65 percent of the respondents in Mexico indicated that it was possible to brib e the police (Latinobarmetro, 2004, p. 54). Likewise, in Argentina 57 percent thought this was possible. If you contrast that to Chile, only 22 percent of the population thought the police we re bribe-able. In El Salvador, the number is even lower; only 20 percent of the pop ulation though that it was possible to bribe their nations police (Latinobarmetro, 2004, p. 54). Seei ng that there is such a wide range of corruption across the region, a review of legal aut horities in Latin Ameri ca will shed some light on why legitimacy is lacking in many Latin Amer ican countries. This is meant to give a rough picture of how law operates in certa in Latin American countries. Country Levels of Legitimacy In order to understand how law operates within each specific count ry in this study two approaches were used. First, data from the 12 La tin American countries st udied in this project were used in order to look at the mean scor es on indicators of legitim acy for each country (see Chapter 4 for a description of the data used in this project). We compar e those scores with the grand mean from the entire Latin American regi on. Countries that have scores indicating greater levels of legitimacy on at least three of the four indicators of legitimacy used in this study, level of democracy, court justice, trus t of police and vigilante justice are deemed to have the most legitimate systems of law. Countries that have scor es indicating that they have greater levels of legitimacy on two of the four indica tors of legitimacy have average levels of legitimacy of law. Countries that have scores indicating they have greater levels of legitimacy on one or none of the indicators of legitimacy are deemed to have less legitimate systems of law. In Table 2-1, we see that Jamaica and Uruguay score in the more legiti mate direction on all four indicators (vigilante justice is reverse coded from the other indicators therefore lower scores indicate greater legitimacy). Venezuela and Costa Rica also have more legitimate scores on three of the four 34

PAGE 35

indicators. Countries with averag e levels of legitimacy (two of the four indicators) are Chile, Mexico and Brazil. The Dominican Republic, El Salvador and Guatemala only have one indicator in which the score indi cates greater levels of legitimacy. Bolivia and Peru score in the less legitimate direction on all 4 indicators. If we take these numbers at face value, it w ould indicate that the c ountries where citizens report the highest level of legitimacy of la w and legal authorities are Jamaica, Uruguay, Venezuela and Costa Rica. While high levels of legitimacy in Uruguay and Costa Rica are in keeping with research on the region and a ge neral understanding of ru le of law in those countries, the results for Venezuela and certainly Jamaica are surprising. Given that these scores on legitimacy do not necessarily reflect the general understanding of levels of legitimacy, rule of law and corruption practices in the area, a second methodology is used to understand how the countries in the region operate. A review, albe it brief, of existing l iterature on corruption, legitimacy and legal authorities in the region will help organize th e countries in the region into three categories: more legitimate countries, countri es with lower levels of legitimacy of law and those countries of average legitimacy More Legitimate Countries Costa Rica Costa Rica has long been used as an example of stability both economically and politically in Latin America and is a paragon of democratic consolidation in the region (Seligson, 1987; Seligson & Booth, 1993; Lehoucq, 2005). It is not surprising then that Costa Rica had scores indicating greater than average levels of legitimacy on three of the four indicators of legitimacy of law used in this study (Table 2-1). Costa Ri ca has had free and fair el ections for the majority of the 20th Century and peaceful executive transiti ons since 1950 (Seligson & Booth, 1993). However, research on democracy and legitimacy show s that even this politi cally stalwart country 35

PAGE 36

has a few chinks in its armor. Seligson and B ooth (1993) conducted a study in 1987 where they found that there is a cons iderable lack of support for democra tic political culture in Costa Rica (N=927). According to their findings, roughly a qu arter of the nationally representative sample (24 percent) approved of the government prohibiti ng protest. Another 20 pe rcent agreed that the government should prohibit meetings that critic ize the government. Surp risingly, 19 percent of the sample also supported the idea of government censoring the media outlets such as television, radio and newspapers. This is surprising mostly because in the same study only 13 percent of Nicaraguans, citizens of a historically authoritarian and repressive country, supported such censorship (Seligson & Booth, 1993). Even more astonishing is that nearly 40 percent of respondents of this democracy opposed the right of dissenters to vote in elections (Seligson & Booth, 1993). From these findings it would seem that Costa Rica is not as democratic as scholars once had thought. However, later findings from Booth and Seligson (2003) show that, using a similar data set from data collected in 2002, Cost a Ricans overwhelmingly regard their political system as legitimate (p. 542). This could be due in part to errors in previous research or the overall national context of a longstanding democracy with an untarnished record of respect for human rights (Booth & Seligson, 2003, p. 544). Despite these encouraging findings, in recent years, Costa Rica has indeed experienced a string of scandals at the highest levels of government that have shaken the foundations of democratic rule of law. Lehoucq (2005) writes: In just the past year, two former presidents of this Central American country of 4.3 million have been arrested on corruption charges, while a third has come under investigation. Voter turnout is dropping. Citi zens are unhappy with the tone and content of public life. Agencies and boards responsible for policing the state seem to be working poorly. The public debt is growing to an unhealthy size. The party system, the link between citizens and the state, is di sintegrating. (p. 140) 36

PAGE 37

In a recent survey, 75 percent of Costa Ricans re ported that they believed their public officials were corrupt (Vargas-Cullel & Rosero-Bixby, 2004). However, Lehoucq (2005) points out that whil e the citizens of Costa Rica may have a high level of intolerance for govern mental corruption, the nation is, in fact, not corrupt by regional standards. In addition, the dr op in voter turnout went from a robust level of 80 percent3 in 1998 to 70 percent in 2002, which is still a high voter turnout by American standards where in the 2000 presidential election, only 54.7 percent of registered voters actually voted (Election Assistance Commission, 2007) Moreover, Costa Rica in 2002 was in the 77th percentile in the World Banks governance indicators which Lehoucq (2005, p. 143) points out is a full 22 percentile points higher than othe r nations in Latin America. Cost a Rica also has one of the most level income distributions in the region (Lehoucq 2005). So it would see that despite these small foibles in recent years, Costa Rica still remains one of the most stable de mocracies in the region with fairly low levels of corruption in both the po litical arena and in areas of criminal justice. Uruguay Uruguay is considered to be a beacon of calm in Latin America. While in Sao Paulo police killings are so frequent that they occur, on av erage, ever six hours, in Uruguay, there are only two to three police shootings per year in the entire country (Brinks, 2007). These police shootings, while not always accidental, are of ten the result of poor training and lack of experience (Brinks, 2007). Additionally, while convict ion rates for police killings fall below five percent in Brazilian citie s, in Uruguay 50 percent of all police officers accused of killing a citizen are convicted (Brinks, 2007). When police are convicted by the Uruguayan courts, it is generally due to the legal facts of the case rather than extra-legal factor s that could taint the 3 Voting is compulsory for all Costa Ricans aged 18 and older. 37

PAGE 38

process (Brinks, 2007). Police in Uruguay are relatively free of corruption and abuses and the country itself is among the least violent in the region (Brinks, 2007). Additionally, the legal system runs efficiently and relatively effectively. Latinobarmetro data from 2000 show that Uruguayans have the highest level of confidence in their judiciary of countries in the region. Furthermore, Cao and Zhao (2005) found that among countries in the region, citizens of Uruguay have the highest level of confidence in their police. Over 50 percent of Uruguayans reported having either a great deal or quite a lot of c onfidence in their police force (Cao & Zhao, 2005). These research results are echoed in the data from this study which shows that Uruguayans report the country to have higher scores of legiti macy on all four indicators, including trust of police, used in this study. Evidence of the rule of law and democratic process operating smoothly in Uruguay is the fact that neither the U.S. State Department nor Amnesty International even report on human rights abuses in Uruguay (Brinks, 2007). When this is contrasted with the situations in Mexico, Guatemala and even Brazil, it underscores the real contextual differences among Latin American countries and the reason for varying rates of viol ence. In Uruguay there is a low level of police violence and corruption, and not surprising ly, a high level of regime legitimacy. In a study using 1997 Latinobarmeter survey data from Uruguay, Canche, Modak and Seligson (2001) found that Uruguayans have a hi gh level of satisfaction with their democracy. The authors found that there is both strong support for the curre nt political system, but also strong support for democracy itself. Support fo r the political was measured by gauging the confidence in the armed forces, judiciary, presid ent, congress and police. Support for democracy was measured by asking survey respondents if a democratic government was preferable to an authoritarian government. Both of these measures of satisfaction with de mocracy could also be 38

PAGE 39

considered gauges of the legitimacy of both gove rnment authorities and democracy. From the results we see that there is a good deal of s upport, and therefore legi timacy of the political system and democracy in Uruguay (Canche, Modak & Seligson, 2001). Chile Like Costa Rica and Uruguay, Chile is consid ered by scholars to be a country with a good deal of legitimacy of law. This is evidenced by the level of legitimacy found in the Chilean police. The Carabineros, or Chil ean police, enjoy a good deal of le gitimacy due in part to their professional competence and disciplined structure (Frhling, 2007). Like other police forces in the region (such as Brazils), th ey do have a history of police vi olence, much of which took place during the Pinochet dictatorship.4 And like other countries, Chile also has a high number of reported crimes. However, this in part indicates that Chileans fe el comfortable going to their police when they have been victimized (Frh ling, 2007). Even though, as w ith the rest of Latin America, there has been a recent upswing in crime in Chile, the public still firmly places its trust in the Carabineros and their ability to handle th e escalation of delinquency. This is evidenced in the mean score for trust of police that we see for Chile (Table 2-1) and is due in part to current police practices in Chile. Currently the police in Chile encourage community participation in police and as opposed to countries like Brazil and Me xico and they are more likely to respond to citizen complaints and take them seriously (Frhling, 2007). This is not to say that there is no police co rruption in Chile. Like other Latin American countrys there are issues of police violence a nd misuse of power, however, this is seen as necessary to maintain social order, and theref ore does not erode the polices legitimacy. In a 4 Augusto Pinchet overthrew the presidency of Salvador Allende in 1973 in a military coup. The military repressively maintained power over the country from 1973 to 1990. However, despite its repressive nature, the regime was politically stable. 39

PAGE 40

recent survey, 75 percent of the Chileans polled agreed that some amount of harsh treatment was necessary to control crime rates (Frhling, 2007, p. 127).5 Police also garner a high level of respect because they do not generally use their positions for economic or personal gain. In a recent survey done by the National Institute on Statistics ( Instituto Nacional de Estadsticas, INE) found that 82 percent of the respondents had confidence in the Chilean police force. It is not surprising then that 88 percent of those su rveyed have never witness nor been a victim of police bribery (Frhling, 2007). Also, unlike many police institutions in Latin America, the Chilean police have a longstanding traditi on that helps cement their legitimacy: the Carabineros as well as the armed for ces have come to be seen as permanent institutions of the state whose members ma y have committed serious crimes during the military dictatorship but who since then have rejoined the democratic polity. (Frhling, 2007, p. 127) In comparison to other countries in the regi on, Chileans predominantly believe their legal system, specifically the police, works to ensure that people obey the law (Frhling, 2007). This helps explain why the democratic stability that Chile currently enjoys, despite the high rates of victimization. In contrast to the relative legitimacy of law and lack of corruption that is seen in Costa Rica, Uruguay and Chile, th ere are striking differences found in countries with lower levels of legitimacy such as Bolivia, El Salvador and Guatemala. The lower levels of legitimacy and issues of corruption that is seen in these countries highlight the contextual differences among each Latin American country. Countries with Lower Levels of Legitimacy Bolivia Bolivia is one of the two countries in this study that did not score higher than the grand mean for the region on any of the indicators of legitimacy used in this study (Table 2-1). In 5 Un poco de mano dura del govierno no viene mal 40

PAGE 41

Bolivia, this lack of legitimacy is undermining not only rule of law, but democratic governance as well. In 1982 Bolivia transitioned from military rule to a democratic government with routine elections. However, economic problems in the late 1990s led to a growing lack of legitimacy in Bolivia which is evidenced by rece nt political turmoil and violen t protest. In 2003 these violent protests rocked Bolivias capital, La Paz, when demonstrators took to the streets and left 30 civilians dead (Barr, 2005). What followed were more months of protest and 29 more deaths. In October 2003, after 20 years of succ essful democratic transitions fr om one president to another, the sitting President, Gonzalo Sanchez de Lozada, was forced to resign. The new president, de Lozadas vice president Carlos Mesa, similarly threatened to resign onl y two years later in March 2005. The country had been plagued with economic instability and the economic restructuring of the mid-1980s left the countrys infrastructure weakened and unable to respond to the growing issues of poverty and diminish ed living conditions (Barr 2005). Democracy was the only game in town; however citizens square ly placed all blame for the countrys problems on the political establishment. Democratic rule of law was losing legitimacy and legal authorities were not able to cope with the rising tide of unrest. As a result, demonstrations, strikes, marches, and blockades have become an almost daily occurrence in Bolivia over the past few years (Barr, 2000, p. 71). The effect of this political unrest has undermined the legitimacy of legal authorities in Bolivia. In 2000 results of a public opinion poll showed that the police, courts, congress, armed forces and political parties all had negativ e confidence ratings among the Bolivian people (Encuentas and Estudios, 2000. p. 10) This is also seen in the results found in Table 2-1, where we see that Bolivia scores lower than the mean for the region on both trust in police and court 41

PAGE 42

justice. These lower levels of trust in police a nd confidence in the judiciary are due mainly to issues of corruption. Recent studies show that corruption is rif e and is undermining the states legitimacy (Domingo, 2005, p. 1728). According to Transparency International, in 2001 Bolivia ranked 84th out of 91 countries for its level of corruption (Eyzaguirre, 1999, p. 86). This is due to the fact that in Bolivia there exists chr onic corruption at all levels of the public sector and government (Domingo, 2005, p. 1731). A recent survey showed that 25 percent of Bolivians had either experienced corruption themselves or witnessed it taking place, either by police or government officials (Seligson, 2002). This corruption coupled with the governmen ts inability to respond to recent economic turmoil is undermining the legitimacy of the political process in Bo livia (p. 1731). Public opinion polls show that support for democratic ru le of law is dropping in recent years. In 2004, results from the Latinobar6metro survey showed th at less than half of Bolivian respondents (45 percent) had a preference for democratic rule of law. In this same survey, only 16 percent of respondents reported being satisfied with demo cracy. In 1996, 25 percen t of Bolivians were satisfied with their democratic government (Latinobarmetro, 2004, p. 23). Peru Like Bolivia, the mean scores for all four indica tors of legitimacy of law used in this study indicated lower than average levels of legitimacy when compared with the regional mean. These lower levels of legitimacy might be due in pa rt to recent scandals that the country has experienced. In 2000 President Fujimori of Peru was forced to resign after a videotape surfaced of his personal subordinate, Vladimiro Montes inos, bribing a congress member of the opposing party (McMillan & Zoido, 2004). Montesinos was secretly the head of Peruvian police and during Fujimoris presidency he paid bribes to judges, politicians and the news media. In a 42

PAGE 43

country with operating democratic mechanisms such as a written constitution, presidential term limits, a political party system, routine elec tions and a free press, corruption was able to subvert the democratic process (McMillan & Zoido, 2004). Since 1990 when Fujimori took power in Peru, there had been a string of abuses of power, in cluding the aforementioned bribes and defying the Peruvian constitution to take a th ird term as president. Under his presidency the legal arms of the government were either corrupt or subservient to the executive branch so that there were no checks on Fujimoris power. Wh ile there was an operational or procedural democracy in Peru, it was nothing more than a f aade. Democracy, in its fullest definition as outlined in this paper, was not operating. Fujimor is government fails to meet the conventional minimum standards for democracy (Levitsky & Way, 2002, p. 52). While Peruvians eventually ousted the corru pt Fujimori regime, the history of human rights abuses in Peru is longstanding and corrupt ion rampant. The political violence of the 1980s was coupled with extreme poverty that gave way to a self-co up by Fujimori in 1992. During the 1980s Perus communist part y, the Shining Path emerged a nd stoked the political violence. This violence resulted in the loss of an estim ated 70,000 Peruvians, according to the countrys Truth and Reconciliation Committee. Many of these deaths were at the ha nds of Peruvian police officers (Cao & Zhao, 2005). The police in Peru like those in many Latin American countries, are ineffective, due in part to lack of financing and personnel, but due mostly to their own abusive behavior and corruption (Cao & Zhao, 2 005). It is not surprising then that the mean score of trust of police from respondents in this study is lower than the grand mean for the region (3.35 for Peru versus 3.55 for the region). This is no doubt a reflection of this abusive behavior and corruption by the Peruvian police. 43

PAGE 44

As recent as June 5, 2009, Peruvian National Po lice (PNP) were accuse d of killing twentytwo indigenous protesters in th e Amazonian region of Peru. They were blocking roads in protest of President Alan Garcias economic policies which they claim are exploiting their lands (Bice, 2009). The protestors claim to have been peacefu lly demonstrating, armed with only rudimentary weapons such as sticks and rocks. However, the armed forces and police claim to have been shot upon first and responded only to th e attack. While details of th e conflict are unclear, they do highlight two distinct problems relating to Peru. Th e first is the ever present racial clash between indigenous peoples in Peru and the mestizo class. The second problem that is highlighted here is the clash between citizens and the police/armed fo rces and the governments use of extreme force against its citizens, specif ically when they are protesting government action. This friction between police and citizens has led to similar vigilante justice that is seen in Guatemala. In both indigenous and mestizo comm unities in rural and urban areas, there have been reported incidents of lynchings and mob violence (S nodgrass Godoy, 2004). Victims of lynchings in Peru have been forced to publicly st rip and then been ritualistically crucified. These lynchings are in part a result of the militaristic, violent and repre ssive regimes that have affected the country, but they are also in response to a belief that the governme nt is unable to handle growing crime and violence rates (Snodgrass Go doy, 2004). This underscores a general lack of legitimacy on the part of the PNP. And as scores from Table 2-1 indicate, there are low levels of court legitimacy, democratic legi timacy and higher levels of suppor t for vigilante justice in the country. All these factors contri bute to Peru being deemed a country with lower levels of legitimacy of law. El Salvador El Salvador underwent a brutal civil war last ing 12 years that fina lly ended with peace accords in 1992 (Holiday & Stanley, 1993). The peace accords that were signed between the 44

PAGE 45

Farabundo Marti National Liberation Front (FMLN) and the right wing government gave way to democracy in El Salvador. In an effort to cu rb corruption and human ri ghts abuses in the new government, both sides agreed that the Salvadoria n police force would be made of up 60 percent civilians and no more than 20 percent each of ex -military from both sides of the recent conflict (Perz, 2003; Holiday & Stanley, 1993). Additio nally, military personnel with histories of corruption or human rights abuses were stricken from serving in the police forces (Call, 2002). Great efforts were made to maintain a successf ul cease fire, which was never broken. Yet while some consider the transition to democracy in El Sa lvador to be one of the most successful in the post Cold War era (Call, 2002), it has not been as peaceful or easy a transition as many initially believed. There are still lingering effects of the wa r that can be seen in the countrys police and government. While the international community hailed the Salvadorian transition to democracy as highly successful, the sharp increase in the crime rate in El Salva dor tells a different story. This wave of post-war crime led to police and judicial reforms that were la uded by UN officials, but had the citizens of El Salvador less convinced: A sharp post-war increase in violent crime also helps account for the pattern of judicial reform and sparked creative new strate gies for reducing violence and improving justiceinternational observers consider them a success story, but Salvadorians [sic] are far less enthusiastic (Call, 2003, p. 829) The police reforms that did take place were in response to corruption and scandals that took place after the peace accords had ended (Call, 2003) Despite these efforts to root out corruption and the reforms that demilitarized the police, corruption still lo omed large as a part of the Salvadorian police for several y ears after the peace accords. Lingering authoritarian practices remained a part of policing in El Salvador. Perz (2003) writes that despite democratic transitionspolice and judicial processes remain abusive, 45

PAGE 46

corrupt, and ineffective undermining state le gitimacy (p. 628). The formation of the Policia Nacional Civil (PNC) in El Salvador was supposed to curtail corruption, however between June 1999 and May 2000, the newly established office of the Ombudsman for the Defense of Human rights received over 97 official complaints of human rights abus es. This was over 30 percent of all complaints during that period, and was mo re than the complaints against any other government institution (Perz, 2003). These comp laints ranged from kidnapping to murder (Perz, 2003). Despite these a buses, the PNC managed to overc ome much of the torture and human rights abuses that plague other nations po lice forces in Latin America. Police doctrine in El Salvador emphasizes due process and rule of la w rather than repressive control. So despite previous problems, the Salvadorian police have been part of the successful transition to democracy. Due in part to the rigorous reforms enacted by the newly democratic government, in 1999 El Salvador ranked only 49th on Transparency Internationa ls Corruption Perception Index (Seligson, 2002). While this may seem high, for countries in the region, El Salvador is only considered to be moderately corrupt (Sel igson, 2002, p. 418). In a survey of nearly 3,000 Salvadorians performed by the Institute of Public Opinion (IDUP), only six percent of individuals had experienced gove rnmental corruption including be ing stopped on false pretenses by police, being asked for a bribe by public official s, at work, by police, or by court officials, or observing a bribe (Seligson, 2002). Th is is in comparison with 25 percent of Bolivians who had such experiences with corruption. This lack of co rruption in the PNC is a large part of why trust levels for the police are rather high in El Salvador. In a study conducted between 1996 and 1998, a survey of over 2,900 citizens in El Salvador revealed that on a scale of 1 to 100, police scored a 59 on a trust scale. The PNC was the third most tr usted institution in El Salvador, surpassed only 46

PAGE 47

by the Human Rights prosecutor and the Cathol ic Church (Perz, 2003, p. 634). In this study only three percent of respondents indicated that they had been bribed by a police officer and 12 percent reported witnessing a bribe within the past two years. This is indicative of overall low level of corruption among the PNC (Perz, 2003; p. 634). Despite these findings, results from the data used in this project show that El Salvadorians report lower levels of democracy, lower levels of trust in the court system, and a greater support for vigilante justice than is average for th e region (See Table 2-1). The only area where El Salvador scores higher than the regional mean is in trust for police. This is in keeping with the research that we just outlined from previous studies done on El Salvador. Coupling the results seen in Table 2-1 with previous research finding s, El Salvador ranks among one of the countries with lower levels of legitimacy in the region. Guatemala Guatemalas legacy is very similar to that of El Salvador. However, Guatemalas civil war raged for the greater part of 36 years, ending in 1996. This war left over 200,000 people dead or disappeared (Snodgrass Godoy, 2002). As with El Salvador, the transition to a democratic government was rocky in the years following the peace agreement. It took nearly two years for the legislature to begin amending the constitution and while advances were made in disbanding the military and increasing the civilian police for ce, a general lack of order in the government was reflected in the countrys high crime rates (Kincaid, 2000). Like El Salvador, there was a wave of criminal activity following the end of the civil war. While there is little reliable data on crime rates during this period (Snodgrass G odoy, 2003; Kincaid, 2000), there was a prevalence of drug related crimes, narcotrafficking, kidnappi ng, extortion, bank robberies and motor vehicle thefts following the war (Kincaid, 2000). Much of this crime was related to youth gangs and former guerilla fighters who had an abundance of munitions, no jobs, and little money (Kay, 47

PAGE 48

2001; 2007). Estimates of upwards of 60 youth gangs were operating in Guatemala City alone in 1998 (Kincaid, 2000). Unlike the Salvadorian transition, Guatemala did not place restric tions or screening mechanisms on police hires, so throughout much of the country, the old police and ex-military militia maintain control of policing. Th is meant that over half of the new Polica Nacional Civil (PNC) of Guatemala were made up of corrupt, abus ive, inefficient and incompetent officers from the old national police (Perz, 2003 ; Kincaid, 2000). While 60 percen t of El Salvadors PNC was made up of new civilian recruits, less than 25 percent of Guatemalas PNC had no previous involvement in policing the nation (Kincaid, 2000). Despite these practices the results in Table 21 show that the mean level of trust of the police reported by Guat emalans is only slightly lower than the grand mean for the region. However, ther e are higher than averag e levels of support for vigilante justice, which is in k eeping with research on Guatemala. While in El Salvador the transition to democracy eventually led to a degree of security in the nation, in Guatemala a lack of security was dire ctly related to an incr ease in civilian justice and public lynchings (Snodgrass Godoy, 2002; Kincai d, 2000). Vigilante justice took hold in Guatemala similarly to the way it has been employed in Venezuela, Mexico and Brazil (Snodgrass Godoy, 2004). Between 1996 and 2001, ther e were 421 lynchings recorded in Guatemala by the United Nations (Snodgrass Go doy, 2002). Some of these lynchings involved multiple victims. And many went unpunished by Guat emalan criminal justice officials and took place in broad daylight with mass civilia n participation (Snodgrass Godoy, 2002, p. 642). In 1997 alone, there was an average of one lynchi ng per week (Kincaid, 2000). Lynchings in Guatemala are not only a reflection of the prev ious human rights abuses that took place during 48

PAGE 49

the 36 year civil war, but are also a reflection of a lack of public security and due process of the law. These lynchings were often used as a form of civilian justice in the absence of a competent criminal justice system. With the wave of violen t crime that shook Guatemala after the end of the civil war, the PNC did not have the capacity, re sources and political will to investigate and punish most crimes (Snodgrass Godoy, 2002, p. 645). This lack of state security and an understandable cynicism of a police force ri ddled with war criminals has led many communities to resort to justicia a mano propia6 (Snodgrass Godoy, 2002, p. 645). To add to the grim picture, the victims of lynchings often rece ive no trial, by civilian or state court, and are generally brutally murdered either burned alive or even stoned (Snodgrass Godoy, 2002; 2004). These lynchings are evidence of lack of trust and le gitimacy of legal authorities in Guatemala. In fact, many lynching victims are forcibly taken from police custody because Guatemalan citizens fear that the police or courts will allow them to go free without being punished for their crimes (Snodgrass Godoy, 2002). Evidence of a lack of trust in police is al so found in recent survey data. Surveys done between 1996 and 1998 of 1,200 Guatemalans shows that, on a trust scale of 1 to 100, the average score given to the PNC is only 50 (Perz, 2003). In contrast to the low level of trust for the PNC, 70.5 percent of respondents supported the ar med forces having a role in fighting crime, which is in direct conflict with the terms of the nations Peace Accords. Perz (2003, p. 635) notes that this finding reflects a lack of confidence in the abilities of the National Civil Police to handle the crime problem. However, despite a lack of police legi timacy, the overwhelming majority of Guatemalans (roughly 73 percent) support some level of democratic governance 6 Justice by ones own hand 49

PAGE 50

(Perz, 2003, p. 643). However, it should be noted th at in this current study, the mean score of level of democracy for Guatemala was lower than that for the full Latin American region (2.50 versus 2.89). This highlights the changing nature of democratic rule of law and legitimacy in some countries where legal authorities repr esenting that government lack legitimacy. Countries of Average Legitimacy Mexico Mexicos scores of legitimacy seen in Tabl e 2-1 and research from previous studies of corruption, policing and rule of la w all indicate that Mexico has average levels of legitimacy for the region. Police in Mexico have historically struggled with corruption and maintaining legitimacy (Domingo, 1999; Lpez-Montiel, 20 00; Reames, 2003; Davis, 2006). The current study shows that the score for trust of police in Me xico (3.23) is lower than the regional average (3.55), which is in keeping with research on po licing from the region. Although there have been attempts at rooting out corruption and reformi ng the Mexican police, recent reforms aimed at curbing police corruption have not been able to root out the problem. Corruption among Mexican law enforcement officials is widespread (Dav is, 2006). In 1999, a survey of Mexican citizens showed that 90 percent of the population had little or no confid ence in the police (Reames, 2003, p. 7). This lack of legitimacy is in no small part due to the fact that an estimated eight percent of the income in a Mexican household is spent on police bribes (Reames 2003, p. 7). Police in Mexico have been accused of bribery, violence ag ainst lawful citizens, and even murder (Davis, 2006). Recent years have seen things deteriorat e to the point that mobs have even lynched corrupt officers (Davis, 2006). The lack of rule of law and procedural justice has given way to barbaric street justice (Davis, 2006, p. 57). With such fragile rule of law, there also exists a fragile democracy in Mexico: 50

PAGE 51

One would rate the quality of Mexicos democr acy as relatively low, or mixed at best, precisely because of the human and civil rights abuses associ ated with state and citizen violence as well as unrule of law. (Davis, 2006, p. 58) The relatively low quality of democracy refers to a lack of the second integral part in our definition of democracy, the protection of citizen rights.7 This is missing because the police and legal authorities lack legitimacy, due in part to their own corrupt practices. Despite the overall low levels of legitimacy of police, Mexico remains a country of average levels of legitimacy due to its overall lack of support for vigilante justice, which was one full point below the regional mean. Additionally, respondents from the survey used in this study also indicated that there was a greater than average tr ust in the court system (4.12 versus 3.83 for the region). So while police are generally seen as untrustworthy, the countrys governmental agencies still enjoy some legitimacy. Brazil Brazils police force, like Me xicos, is fraught with corruption which is undermining the ability of citizens to trust thei r police forces and, more importan tly, undermines the legitimacy of democracy itself (Seligson, 2002). The legacy of policing in Brazil has a black mark on it brought about by corrupt and brutal practices du ring the 21 years that the military held power over the country.8 During this period the government, ca rried out widespread repression that included brutality, torture, murder, and dis appearances. (Huggins, 2000, p. 59). Even after democratization, the police have continued their violent practices (Mitchell & Wood, 1999). Some go so far as to say that effective polic ing in Brazil is nonexistent (Paes-Machado & 7 It should be noted that electoral democracy in Mexico is also rather tenuous. Recent elections have been hotly contested and accusations of voter fraud, election tamp ering and corruption have arisen (Domingo 1999). 8 The Lead Years, This was a period in Brazilian history when the military took over the nations government and instituted practices of kidnapping, torture and disappearance s of seditious citizens or any member of society that vocally opposed the government (See Huggins 2000). 51

PAGE 52

Noronha, 2002, p. 74). Brazils record of police violence is one of the wo rst in the Americas (Chevigny, 1995; Huggins, 2000b): In 1992 in the greater metropolitan area of S o Paulo alone, police killed 1,470 civilians almost four times more than the total number of such killings duri ng an entire span of 15 years in Brazils 21-ye ar military dictatorship. (Higgins, 2000b, p. 116) While that number had dropped in So Paulo by the end of the decade to 498 killings in 1998, the number of homicides in Rio di Janeiro wa s at nearly 38 murders per month (Huggins, 2000b, p. 116).9 This rate of police violence is not entirely in response to civilian violence. Recent years have seen multiple national scandals in which uniformed officers have been videotaped, sometimes in broad daylight, beating, punching, torturing and even murdering citizens (The Economist, 1997; 1999). Much of it is based on the sanctioned policing principle of shoot first, investigate later that was adopted by state se cretary of public safety, General Nilton Cerqueira (Zaverucha, 2000). This lack of due process and even violent corruption on the part of the Brazilian police erodes their legitimacy as a le gal authority. Paes-Machado and Noronha (2002) conducted surveys of citizens livi ng in the poorest regions of Sa lvador, Brazil. From their study they found that citizens do not re port police brutality because authorities do not listen to them (Paes-Machado & Noronha, 2002, p. 67). There is a clear lack of due process and procedural justice in Brazilian policing. The Brazilian judiciary does not fare much better when it come s to equal application of the law and a lack of corruption. Even among judges, 77 percent of those surveyed in Brazil believe that their judiciary is in some sort of crisis (Ratliff & Buscaglia, 1997). What is lacking in Brazil are the basic elements of an efficient judici al system (Ratliff & Buscaglia, 1997, p. 63). It is 9 By comparison, New York City police killed only 19 citizens in 1998 and 11 in 1999. Additionally Washington D.C. police killed only 6 citizens in 1996 (Huggins 2000b: 116). It would take approximately 50 years for Washington and 25 years for New York to kill as many citizens as So Paulo Police murdered just in 1999 (Huggins 2000b: 116). 52

PAGE 53

not equally accessible to people of all incomes; th ere are no predictable outcomes; and there is a plague of corruption (Ratliff & Buscaglia, 1997, p. 63). Comparing the research on Brazi ls police and the court system s with the scores in Table 2-1, Brazil fairs moderately when compared w ith the region. The mean score for level of democracy is lower than the region as is the trust in the court system. However, trust of police is higher than the regional grand mean, which is surprising given the abundance of research on police brutality in Brazil. This could, however, be due in part to Brazilians acceptance of police brutality as a necessary evil when it comes to fighting crime. Brazilians also report low levels of support for vigilante justice, which is also su rprising given recent rese arch on public lynchings. Taking previous research into account, Brazil is counted among the countries with average levels of legitimacy. Venezuela The results in Table 2-1 show that Venezuel ans have a lower than average score of trust of police when compared with the grand mean from the region. This is in keeping with research from the region that shows that less than 30 perc ent of Venezuelans supp ort their police force (Cao & Zhao, 2005). This is lower than Brazil ( over 40 percent), Mexico (over 30 percent) and Colombia (49 percent). In a study using public opinion data from Venezuela (N=1,200) when asked about confidence in their police, Venezuelan s overwhelming report that they have not very much or none at all (Cao & Zhao, 2005). This lack of confidence is evidence of a lack of legitimacy in the police. Additional evidence of a lack of trust of cr iminal justice officials is evidence by the lynchings that take place in Venezuela, like other Latin American countries, as a form of vigilante justice. Lynchings in Venezuela are done because citizens do not trust the criminal justice system to enact justice or adequately pun ish criminals. They are also done as a form of 53

PAGE 54

protest and rebellion against insecurity in the country (S nodgrass Godoy, 2004, p. 638). These lynchings are done to demonstrate to the authori ties that they have to do something about the crime problem in the country (Snodgrass Godoy, 2004, p. 638). Recent increases in crime have Venezuelans taking extreme measures to ensure their own safety. In 2001 after the murder of a six year old re sident of Mariara, Venezuela, residents dragged a 41-year old man, Omar Prez Gallardo from his home and severely beat him. Police arrived, saving the mans life. However, a m ob gathered of roughly 400 residents and they stormed the local hospital where Gallardo was be ing held, threatening to burn it down. Upon his release, Gallardo was taken to a local plaza and beaten to death. Citizens removed one of his arms and hung it from a tree; they then burned th e rest of his corpse. Additionally protesters gathered at the scene to denounce the rampant criminal activity in the region (Snodgrass Godoy, 2004, p. 622). It was la ter discovered that: Mr. Prez was innocent of all charges; he ha d apparently been mistaken for a legendary serial killer known as the monster of Mariara. (Snodgrass Godoy, 2004, p. 622) This type of mob justice is not uncommon in Venezuela. Between October 2000 and September 2001, the human rights advocacy group Programa Venezolano de Educacin y Accin en Derechos Humanos estimated that there were 164 lynchings in Venezuela. As noted in Chapter 1 of this text, police resources in Venezuela are often used to counter petty criminal activities or activities that have no real legi slative or constitutional ban (S anjan, 2000). This helps explain why citizens of Venezuela must take the law into their own hands in order to resolve the growing crime problem. Despite the anecdotal evidence and recent c limb in vigilante jus tice in the region, the results shown in Table 2-1 indicate that Venezu elans report a lower le vel of support vigilante justice (3.27) than the mean scor e for the full Latin American re gion (3.76). So while there may 54

PAGE 55

be a rash of recent incidents of vigilante justi ce in the country, it does not necessarily indicate that the country as a whole supports this type of justice. In a study performed by Birkbeck, Gabaldn and LaFree (1993) using face-to-face inte rviews with Venezuelan citizens (N=655) researchers found that 75 percent of the sample thought that when there were problems in their neighborhood it was better for the police to intervene (Birkbeck et al., 1993, p. 34). This is in contrast with respondents from the United States (N=533) who ove rwhelmingly responded that it was better for police to intervene (96 percent) (Birkbeck et al., 1993). However, it is still a good number of respondents that prefer police to interfere with criminal matter, specifically when it is a more serious criminal circumstance (assault). In these instances Venezuelans are more likely to call the police than even U.S. Citizens. The absence of rule of law portrayed in the lynchings and the lack of legal legitimacy in Venezuela has been described as the principl e fault lines of democratic consolidation (Snodgrass Godoy, 2004, p. 640). The lack of legitimacy of state law in Venezuela can be traced to human rights abuses and police misconduct. Cu rrent president, Hugo Chvez was elected in 1998 with the hope that he would clean house in government of the corruption that was rampant in Venezuela (Weyland, 2001). However, despite sw eeping changes in the old guard of politics in Venezuela, Chvez had enacted radical and d ecidedly undemocratic policies in the country (Weyland, 2001, p. 73). Among his changes were cha nging the policy dictating the presidents must wait 10 years to run for re-election, diss olving the countrys sena te (and checks on his power). The Chvez record is rife with violations of human rights, disregard for the rule of law, and contempt for democratic norms and proces ses (Casey, 2005). Despite these abuses of power, Venezuelans still indicate th at their country is fairly democr atic (3.11 out of 4). The mean 55

PAGE 56

score of level of democracy for Venezuela is high er than the regional mean (2.89). It is in other areas of law, however, that Venezuela falls below the mean. The mean score for trust of police in Ve nezuela was 2.75, which is lower than the regional grand mean (3.55). This lack of trust in the police is due to a great deal of corruption and police brutality. Examples of police corruption that had eroded the legitimacy of legal authorities in Venezuela include the killing of peaceful demonstrators by police (Casey, 2005). Additionally, police under the Chvez regime are accused by Human Rights Watch (HRW) and Amnesty International (AI) excessive force agains t anti-Chvez demonstrators as well as illtreatment and even torture of detainees (Casey, 2005). Allega tions by both HRW and AI also claim that the Venezuelan police and National Guard are responsible for d eath squad killings of approximately 100 citizens between 2000 and 2001 (Casey, 2005). Chavista supporters deny these allegations; however the administration has refused to seriously investigate any of these attacks that were predominately done against an ti-Chvez demonstrators. These abuses and the rising political and street violence in Venezuela ar e evidence of lack of regime legitimacy as well as a lack of criminal justice legitimacy. The regime itself under Chvez, while still a skeletal democracy, is lacking many of the democratic principles that have been outlined in this paper. As a result, the ability to evaluate Ven ezuelas democracy is difficult at best. Dominican Republic Scores from respondents from the Dominican Republic indicate that the countrys legal authorities and democratic government have lo wer levels of legitimac y. Although the Dominican Republic has had the guise of democracy since the 1970s, in practice, Dominicans had been under semi-authoritarian rule until the mid-1990s (Finkel, Sabatini, & Bevis, 2000). President Joaqun Balaguer held power from 1978 to 1996 and made executive decisions with few checks and balances from the nations congress or judi ciary (Espinal, 1996; Fink el et al 2000). During 56

PAGE 57

this time, the court system was overrun due to poor funding and widespread corruption (Espinal, 1996). Balaguer himself came to power under a flu rry of accusations of election fraud. His narrow victory margin of 0.7 percent left many in the country feeling disenchanted with democracy and the entire electoral process (Finkel et al., 2000). Exacerbating the situation was the use of repression, harassing political oppone nts, tapping phone lines and tightening state security that marred the Balaguer presidency (Finkel et al., 2000, p. 1856). The end of the Balaguer reign in 1996 brought new hope to Do minicans of true electoral democracy taking hold and an end of government corruption. However, given the events of recent decades the Dominican Republic is considered a country of average legitimacy fo r the region. This is due mostly to the events of the later 1990s which was marked by its efforts to ensure a more stable democracy with free and fair elections, reduced institutional corruption and a more inde pendent judiciary (Espinal, Hartlyn, & Kelly, 2006). And while still quite low, previous research indicates that the overall levels of trust that Dominicans have for their judici ary, congress and president, are higher than regional averages (Espinal et al., 2006). However, the Dominican Republic is still fra ught with institutional corruption. In 2001, Transparency Internationa ls Corruption Perception Index ranked the Dominican Republic ranked 63rd out of 95 countries and 12th among the 18 Latin American countries included for its levels of corruption (Espinal et al., 2006, p. 214) More recently, in 2004 there were a series of bank failures, shortage s of electricity, a hotly contested presidential campaign, and increased governmental corruption that undermined many of the gains made in the late 1990s (USAID, 2005). Police corruption and human rights abuses ar e a persistent problem in the Dominican Republic, as with many Latin American countries. E fforts have been made in recent years to root 57

PAGE 58

out corrupt officers and to stop a buses of power, however the U.S. Department of State lists a battery of human rights abuses and due process violations that take place in the Dominican Republic: Unlawful killings; beatings and other abuse of suspects, detainees, and prisoners; poor to harsh prison conditions; arbitrar y arrest and detention of su spects; severe discrimination against Haitian migrants and their descendant s; widespread perceptions of corruption; violence and discrimination against women; ch ild prostitution and othe r abuse of children; trafficking in persons; and disr egard of fundamental labor right s (U.S. Department of State, 2007). Dominican President, Leonel Fernandez, has made a point to root out governmental corruption, including police corruption, and a ddressed the issue in his 2004 in augural speech. In March of 2009, Fernandez fired 700 police officers who were suspected of corruption and involvement in the drug trade (Brice, 2009). As with many Latin American nations due to the rise in the drug trade in recent years, many police officers who are charged with solving the narcotrafficking problems in the Dominican Republic are now involv ed in it (Brice, 2009). It remains to be seen how successful these efforts will be. Jamaica While Jamaica had more legitimate scores than the regional average for all indicators of legitimacy of law, it is genera lly understood that Jamaica, like Mexico, Brazil, Chile, and other Latin American countries, suffers from issues of police corruption and a re sulting lack of police legitimacy. In Jamaica, corruption is not occasi onal or special instan ces but rather routine practices where law enforcement uses illegal means and the justice system uses unjust practices (Harriot, 2000, p. 65). In 1994, 40 percent of the comp laints to the Police Public Complaints Authority were related to acts of violence, nine percent to harassment and intimidation, and 33 percent to entrepreneurial type of corruption or bribes (Harriot, 2000, p. 58). In 1995 the level of violent corruption increased with 56 per cent of the citizen complaints 58

PAGE 59

relating to excessive use of violence by the Jamaican police (p. 58). In 1993 the general offending rate of the Jamaican police was high er than that for the country as a whole.10 Even more disturbing is that the viol ent crime rate among police officers in Jamaica was more than twice that of the general public.11 All of this corruption and violence negates any of the legitimacy that Jamaican police might hold. The ve ry individuals who are charged with keeping the peace are in fact those who are disturbing it. Du e to its pervasiveness, there is actually a high degree of public acceptance of corruption as part of routine day-to-day transactions with the police (Harriot, 2000, p. 67). Exacerbating the cas e in Jamaica are poor systems of police accountability, a highly politicized police force, a closed bureaucrac y that makes firings difficult, and a complicit public (Harriot, 2000). For th is reason, although Jamaicans indicated higher levels of legitimacy in the data, factoring in what we know from existing research on the country, Jamaica is considered only a country of average levels of legitimacy Conclusions Much of the research on the countries of La tin America outlined in this chapter focuses on issues of corruption among le gal authorities as a measure of legitimacy. While this dissertation project takes into account that individuals may base their perceptions of legitimacy on their experiences with corrupt lega l authorities, this project differs in that it is not specifically looking at corruption. The greater qu estion at hand in this project is how the trust in police acts as a barrier or guardianship against victimizati on. How does a perception of a court system that enacts justice create an atmosphere where indivi duals are less likely to be victimized by their fellow citizens? While a democratic country may be fraught with corrupt ion, if it holds a great 10 2,500 per 100,000 vs. 2,110 per 100,000 11 2,033 per 100,000 vs. 8967 per 100,000 59

PAGE 60

deal of legitimacy in the eyes of the citizenry, how does that affect victimization, if at all? These questions are salient to this study of victim ization in Latin America and go beyond a simple study of corruption in the region. Given the circumstances outlined above for the countries that will be evaluated in this dissertation, it is not su rprising that victimization is a se rious problem many Latin American countries. However, it is also easy to see w hy certain countries (e.g., Costa Rica, Chile and Uruguay) have relatively low rates of both violen t and property victimization. These descriptions give context to each of the countries highlightin g the key differences among these nations. While there are great number of similari ties among countries, civil war in El Salvador and Guatemala, vigilante justice in Mexico and Venezuela, there ar e also distinct contexts in which all of these events occur. These differences, in the opinion of this researcher, make it impossible to lump countries like Uruguay and Costa Rica together with Peru or Brazil. The nuances of each countrys political and legal c ontext necessitate that they be compared and contrasted. 60

PAGE 61

61 Table 2-1. Mean Level of Each Countrys Four Indicators of Legitimacy Compared with the Grand Means for All Latin American Countries Legitimacy Indicator Democracy* Courts* Police* Vigilante# GRAND MEAN 2.89 3.83 3.55 3.76 > Legitimate On 4 Indicators Jamaica 2.98 4.16 3.69 3.20 Uruguay 3.36 4.20 4.17 2.66 > Legitimate On 3 Indicators Venezuela 3.11 3.96 2.75 3.27 Costa Rica 3.14 4.01 5.54 4.11 > Legitimate On 2 Indicators Chile 2.93 3.69 4.93 3.89 Mexico 2.78 4.12 3.23 2.93 Brazil 2.77 3.75 4.59 2.76 More Legitimate On Only 1 of 4 Indicators Dom. Republic 3.09 1.91 2.08 3.78 El Salvador 2.53 3.69 4.11 4.20 Guatemala 2.50 4.23 3.53 4.15 More Legitimate on None of the Indicators Bolivia 2.77 3.64 3.24 3.94 Peru 2.56 3.02 3.35 4.51 *coded from low to high legitimacy # coded from high to low legitimacy

PAGE 62

CHAPTER 3 CRIMINOLOGICAL THEORY: LEGITIMA CY, CRIME AND VICTIMIZATION Theory A debate exists among criminologists as to the applicability of ethnocentric North American criminological theories to non-Wester n and even European societies (Birkbeck, 1993; Encinoza & Del Olmo, 1981; LaFree, 2007; Ma rsh, 1967; Willis, Evans, & LaGrange, 1999: 229). Central to the debate is whether th ese criminological theo ries are based on a conceptualization of crime in the United States or if the tenants of these theories have a broader, global relevance (Willis et al., 1999). Birkb eck (1993) argues that the contributions of criminological research from lesser known regi ons of the world are often ignored among scholars in the field. This cultural insularity (Birkbeck, 1 993, p. 308) limits the applicability of many popular criminological th eories. Comparative studies examining crime outside of the United States, Europe or Canada are necessary to test the applic ability of current theory and discover if and where a theoretica l disjuncture lies (LaFree, 2 007). Adding to the necessity for theoretical studies in Latin Amer ica is a sorely lacking amount of criminological/criminal justice research within the region. Unlike other disciplines, this area of research has yet to fully develop in Latin America and has no solid empirical or behavioral tradit ion (Bergman, 1999). However, despite these deficiencies there is some literature on theory a nd a review of the some relevant theoretical literature is necessary. A commonly cited study that highlighted the disjuncture was DeFleu rs (1970) test of Albert K. Cohens (1955) theory of delinque nt subcultures among 60,000 youths in Argentina (Birkbeck, 1993; LaFree, 2007; Willis et al., 1999). Results from her study showed that crime existed primarily in suburban areas, whereas C ohen (1955) proposed that crime is primarily an urban problem (LaFree, 2007). Additionally, La Fleur (1970) found that there is no shared 62

PAGE 63

consensus of what constitutes middle class values among the youth of Buenos Aires and Caracas. Cohen (1955) attributed delinquency to an inversi on of or reaction against middle class values. Cohen (1955) also proposed that youth were ex posed to the middle class values of success, academic achievement and independence, through their educational system. However, LaFleur (1970) found that the majority of youth in Argen tine urban areas did not attend school. Central to Cohens (1955) argument of delinquent s ubcultures was the contact between youth and societal structures that let them down, or failed to offer them a means to middle class achievement. In the case of Argentine youth, it was specifically the lack of contact with organized societal structures, or marginality th at according to LaFleur (1970), led to delinquency (Birkbeck, 1993). LaFleurs (1970) study highlights the need for comparative theoretica l research. However, this research should not only focus on theories ex plaining the etiology of crime, it should also test the applicability of theories of victimiza tion. Theories that are commonly used to predict victimization in the United States may prove to be a poor fit when applied to Latin American contexts. Patterns of Victimization Victimization research in the United States consistently shows that those who are young, male and unmarried are at a higher risk of victimization (Catalano, 2006; Smith & Jarjoura, 1989). Popular theories of victimization are able explain these findings as they pertain to American or European culture. For example, the lifestyle theory of victimization argues that women are less likely than males to be victims of crime because they spend a greater deal of time in the home. The theory poses that these women have male guardianship from their husbands or fathers which makes them less susceptible to victimization (Hindelang, Gottfredson, & Garofalo, 1978; Meier & Meither, 1993). 63

PAGE 64

Research from Latin America also indicates that young males between the ages of 18 and 24 are most likely to be victims of criminal violence (Buvini & Morrison, 1999). However, this research also shows that women, particularly house wives, are at a high risk of victimization due to the widespread, cultural accepta nce of domestic violence (Buvini Morrison, & Orlando, 2005). Studies show that anywhere between 30 a nd 75 percent of adult women are victims of psychological abuse while between 10 and 40 pe rcent are victims of physical violence (Buvini Morrison, & Shifter, 1999; Morrison & Biehl, 2002). Domestic vi olence is thought to be the most common form of violence in the region (Heinemann & Verner, 2006). In Colombia, for instance, domestic violence affects nearly ha lf of all households while other criminal victimization affects less than 10 percent of those households (G aviria & Velez, 2001). Underreporting is also a widespread problem in Latin America, and most scholars believe that the actual rates of all violen ce, including domestic vi olence, are much higher th an official statistics indicate (Buvini 1999; Soares, 2004). In cont rast, statistics from the United States indicate that while 22.1 percent of women report physical abuse by any partner within their lifetime, only 1.9 percent report physical abuse by a partner within the past 12 months (Tjaden & Thoennes, 2000; 2006). While lifestyle theory may pr edict higher rates of male vict imization in the United States or Europe, lifestyle exposure may not be able to explain the rates of dom estic violence against women in Latin America. Research from the United States also indicates that individuals at higher income levels are less likely to be victimized, due to higher economic resource[s] (Meier & Meither, 1993, p. 469) which allow them to avoid risky situations which might lead to victimization (Hindelang et al., 1978). Academic research consistently support s that, in the United States and the developed world, poorer individuals are more likely both to be victims of violent victimization and 64

PAGE 65

property victimization (Catalano, 2006; Levitt 1999; Williams & Flewelling, 1988) and to be more fearful of victimization (Eisler & Schisse l, 2004). However, in Latin America, wealthy individuals routinely find themselves target s of both property and physical victimization (Heinemann & Verner, 2006). While poorer, less educated individuals report higher rates of violent victimization and have a higher incidence of homicide in Latin Am erica; wealthier residents living in metropolitan areas are most likely to be victims of property crimes (Gaviria & Pags, 1999; Gaviria & Velez, 2001). Wealthy Latin American residents, howev er, in Colombia, Brazil and Mexico have increasingly become targets of kidnapping and extortion schemes that at times result in the beating or death of the victim (Tzanelli, 2006). While kidnapping is on the rise in these countries, it is certainly not as pervasive as other forms of violence. However, this type of violence and high rates of property crimes am ong wealthy Latin American residents illustrate that victimization is not universa lly associated with poverty. Buvini and Morrison (1999) report that in Latin America, recent surveys have not yielded a clear link between poverty and violence when using homicide rates as a measure of victimization. Violent victimization is very high among the poor in the region; th is is interpreted as a result of economic di sparity rather than simply a byproduct of living in poverty (Soares & Naritomi, 2002). This disparity is felt most keenly in by those living in urban areas. Both in the United States and Latin America living in an urban area is predictive of victimization (Catalano, 2006; Gaviria & Pag s, 1999; Gaviria & Velez, 2001; Sampson, 1983). However, a great deal of violence in Latin Amer ica is concentrated in rural and suburban areas (Kay, 2001; Kay, 2007; LaFleur, 1970; LaFree, 2007). Kay (2001, p. 743) notes that, rural violence has been endemic and pe rsistent throughout the history of Latin America. Incidents of 65

PAGE 66

violence are often fueled by issues of land us e, or lingering resentme nt among rural villagers following periods of civil unrest or war (Ka y, 2001; Kay, 2007). Indivi duals living in Latin American countries where they have had recent land law reforms or there are problems with suitable or profitable crops are more prone to victimization (Kay, 2001). Urbanization does promote violence in some Latin American countries, but it is not as explanatory of rural victimization in countries with pr imarily agrarian societies such as Peru or Guatemala. Theories that explain the violence in Latin American may need to take into account the agrarian nature of many of these nations and the history of vi olent struggle for land use when explaining victimization. Given the problems researchers have encount ered in the past when applying traditional criminological theories to Latin American crime and victimization, this dissertation project will not attempt to test a victimization theory, but ra ther will incorporate concepts that are in the tradition of criminological theories but also fit with Latin American scholarship. There will be a conscious effort to fuse the two traditions by borrowing from the routine activities approach to victimization, the North American traditions in legitimacy research and blending them with the Latin American focus on order and political and legal culture. This fusion will allow for a closer examination of victimization while acknowledging contextual differences between countries in the region. The Routine activities approach to the study of victimizati on discusses the importance of understanding how victimization is related to a suita ble target with lack of adequate guardianship intersecting in time and space with a motivated offender (Cohen & Felson, 1979; Lebeau & Castallano, 1987). This approach to crime suggests that individuals who do not have adequate guardianship are more likely to be victimized if they are a suitable target. Guardianship is a 66

PAGE 67

theoretical concept that refers to how protected an individuals pr operty or self is. For example, the presence of friends and family can act as a form of guardianshi p against victimization. Having a family dog also can act as guardianship to the home and family. Carrying mace in your pocket is a form of guardianship. There are many ways to operationalize guardianship; I propose that the concept of guardianship should be expanded to include police guardianship, guardianship of the court system and the guardia nship of a stable democracy and economy. In essence, legitimacy of law is a measure of citize n guardianship in these c ountries. The safety and security that is afforded by a st able government, a criminal jus tice system free of corruption and low levels of mob or vigilante justice cannot be discounted. A legitimate legal system makes individuals more likely to abide by the law and therefore less likely to be victimized. However, instead of discussing the issue as one of guardianship, rather I will discuss it in terms of the legitimacy of these agencies and of democracy in general among the 12 countries in the region. Recall that Gaviria & Pags, (1999) found that urban household victimization was negatively correlated with the degree of confidence in cr iminal justice agencies including both the police and judiciary in their study of 17 countries. Given these findings, the issues of the legitim ization and legitimacy of democracy and of police, courts and legal authorities in Latin American have been accepted among Latin American researchers (Davis Rodrigues 2006; Perez 2003; Lagos 2001). I ex amine the extent to which factors that are related to the st ability of democracies in Latin Am erican also influence individual victimization. More specifically I expect that individuals who have less faith in the police, the courts and/or the stability of their democratic institutions will report more victimization. The two are linked theoretically because the degree of legitimacy that is attri buted to the legal and political institutions indicates the level of guard ianship citizens see its gov ernment as being able 67

PAGE 68

to provide. Those who think that conditions are such that their government is unable to provide its citizens with adequate guardianship from victimization will be more vulnerable to both property and physical victimization. For that reason, this dissertati on turns next to the concepts of legitimacy of legal authorities and of democratic governance in a search to locate measures of guardianship and predictors of victimization. Legitimacy and Increases in Crime Adherence to the law, according to Tyler ( 1990; 2000) occurs because individuals feel a moral obligation to do so, or perc eive their legal system and its entities such as the police and courts to be legitimate (Fagan & Tyler, 2004). In areas where legitimacy is eroded either through police misconduct or corruption, violence and crime increases (Tyler, 1990; Anderson, 1999; Kurbin & Weitzer, 2003; Kane, 2005). Research from the United States that has focused on police legitimacy or effectiveness and crime rates has often focuse d on disadvantaged or marginalized communities (Jacob, 1974; Anderson, 1999; Kurbin & Weitzer, 2003; Sunshine & Tyler, 2003; Kane, 2002; 2003; 2005). However, in the Latin American c ontext the line between a socially organized or disorganized community is often blurred or non-existent. Additionally, police corruption or ill-treatment is not limite d to members of marginalized communities. Therefore, it could be expected that the increases in crime due to lack of police legitimacy that is found in disorganized communities in the Unit ed States (Anderson, 1999; Kurbin & Weitzer, 2003; Kane, 2005) would be found in most Latin American countries as a whole. Legitimacy and Crime: The United States Research on police and criminal justice legitimacy in the United States has consistently shown that as legitimacy increases compliance with the law increases (Tyler, 1990; Kane, 2005). This happens both in the form of cooperation with criminal justice official s but also in the form of obeying the law (Tyler, 1990; Solum, 2004; Fa gan, 2008). As the trust and confidence that 68

PAGE 69

citizens have in their police declines, crime a nd illegal activities increase (Fagan, 2008). This usually results in more heavy-handed policing, which further erodes legitimacy, resulting in more crime to create a spiral of declining legitimacy (Fagan, 2008, p. 127). Tyler (1990) conducted a study on legitimacy and compliance with the law using longitudinal data from approxima tely 1,600 Chicago residents. Re spondents were interviewed in two waves and were asked about their persona l experiences with the police, courts, and correctional authorities as well as their own compliance with th e law. Tyler (1990) measured legitimacy in two ways: support for legal author ities (police and courts ) and the individuals perception of his or her obligation to obey the la w. Results from Tylers (1990) famous Chicago study showed that in the first wa ve, legitimacy was positively associated with compliance even when controlling for personal morality, dete rrence factors and socio-economic factors. Individuals who reported higher leve ls of support for legal authorities were less likely to violate traffic and parking laws, drive while intoxicated or shoplift. This study is important in understanding legal compliance an d its relationship to legitimacy, but it does have a few drawbacks. The sample size is small and due to the method of data collection, phone interviews, the sample may not be representative of the Chicago area. Also, Tyler (1990) admitted that White individuals were over represented in the second wave of the study. Despite these limitations, similar results have been replicated in other cities in the United States. Fagan & Tyler (2004) surveyed a sample of over 1,600 New York City residents that came from a variety of diverse ethnic backgrounds. Respondents were asked about their views on the New York City Police Department as well as their own compliance and cooperation with the law. Measures of police legitimacy that we re used included questions on accepting police authority, ignoring the police, and sharing the values of people in power. Additionally, 69

PAGE 70

respondents were asked about their level of conf idence and trust in the police as well as how honest the police are. Results of the study showed that when it came to obeying the law, including traffic violations, park ing violations as well as crimes such as buying stolen goods or drugs, legitimacy was a significant predictor of both compliance and cooperation with the law. Specifically, when individuals perceived the police to make fair decisions or treat people fairly, they were more likely to be compliant and c ooperative with the law. Procedural fairness was essential for compliance. Fagan & Tyler (2004, p. 10) conclude that people are more accepting of and cooperative with authorities when they ar e treated with fairness a nd respect. The broader implications of this study are limited due to the methodology; phone interviews and a small sample of New York City residents. However, it is important that the findings show that legitimacy in the law inhibits criminality by encouraging compliance with the law. A lack of legitimacy of the police has also been linked to violent crime rates in disadvantaged communities. Kane (2005) combined data from the U.S. Census Bureau and the New York City Police Department to look at police misconduct and violent crime rates in different police precincts. Kane (2005) looked at the number of of ficial incidents of profitmotivated corruption, violence, miscellaneous crimes, administrative misconduct and drug related crimes as a measure of police legitim acy in a given precinct. Violent crime was measured by looking at the annual number of reported homicides, rapes, robberies and aggravated assaults in the give n precinct and dividing that by th e population of the precinct in order to get a crimes rate. This information was tracked from 1975 to 1996. Results from Kanes (2005) study showed that police legitimacy did predict variations in violent offending for both high disadvantage and extreme disadvantage co mmunities indicating that police legitimacy and misconduct is related to violent offe nding at the structural level. 70

PAGE 71

Legitimacy and Crime: Latin America Studies in Latin American have also looke d at the link between legitimacy and illegal activities. Seligson (1980) used a survey of 581 Costa Rican peasan ts to examine the effect of legitimacy on illegal protests. Legitimacy was measured thro ugh trust in the government and illegal activities were both participation in gove rnment strikes and land invasions. Results from Seligsons (1980) study showed th at a low level of trust in the Costa Rican government was associated with both strike participation and land invasions. This shows that in the Latin American context, a low level of legitimacy of government is re lated to a rise in criminal activity. While this is not violent crime, other re search focusing on legitimacy in Latin American government shows that it is asso ciated with victimization. Davis Rodrigues (2006) did a recent study and hypothesized that the violence and victimization seen in Brazil is a direct result of democratiza tion. Davis Rodrigues (2006, p. 243) argues that it is the continued lack of civil democracy and democratic rule of law that is seen in Brazil that drives the perceptions of victimizatio n risk and fear of crime in Brazil. While Davis Rodrigues (2006) discusses the issu e in terms of civil democracy, wh at she is referring to is the systematic exclusion of lower classes in Brazil fro m the judicial process as well as the corruption and capricious, selective, inequitable, or in effective (Hunter, 1985, p. 234) application of the law by Brazils legal authorities. Using data fr om the 2002 Survey of th e Metropolitan Region of Belo Horizonte (n=1,029) Davis Rodrigues (2006) tests how trus t of the police and preference for democratic governance relates to perceived risk of either robbery or assault. While not specifically studying the actual vict imization but rather perceived risk, the study finds support for civil democracy and procedural justice increases feelings of security in ones neighborhoods. Conversely, the findings show that individuals with a high degr ee of distrust for the police actually perceive themselv es to be at more than five times less risk for robbery. Davis Rodrigues 71

PAGE 72

(2006, p. 260) explains that there may in fact be a support for police brutality and corruption in the Belo Horizonte neighborhood because individuals may feel that get-tough policies against crime, including endorsements of the use of violen ce and other illegal or di scriminatory practices by the police are necessary to protect the good citizen. This finding highlights the uniqueness of each countrys experiences and historical context, given that some citizens of the neighborhood talk about the repressive military regime with nostalgia because of its ability to maintain law and order (Davis Rodr igues, 2006, p. 244; Caldeira, 2000). It is the opinion of this researcher that it is not just fear of victimization that results from low levels of legitimacy in the criminal justice system, but rather victimization itself will result. Research supports the notion that when legitimacy declines, vigi lante justice and self-policing results (Johnson, 2007). This vigilante justice of ten takes the form of violent victimization (Kane, 2005; Davis Rodrigues, 2006; Johnson, 2007; Fagan, 2008). Empirical Hypotheses Given the previous discussion on legitimacy of law as it relates to possible victimization, the following empirical hypothese s are proposed. The first set of hypotheses below summarizes the expected empirical relationships between repo rting victimization and e ach of the independent variables, without testing the Legitimacy of La w theory. They provide the framework for the analysis that follows. Demographic Variables H1: The likelihood of victimization will be highe r among male respondents compared to female respondents. H2: The likelihood of victimization will be highe r among respondents who have higher levels of education, as compared to respondents w ho have lower levels of education. 72

PAGE 73

H3: The likelihood of victimization will be hi gher among non-married respondents as compared to married respondents. H4: The likelihood of victimization will be highe r among younger respondents as compared to older respondents. H5: The likelihood of victimization will be highe r among White respondents as compared to nonWhite respondents. H6: The likelihood of victimization will be highe r among respondents with fewer children as opposed to respondents with more children. H7: The likelihood of victimization will be hi gher among respondents with a greater socioeconomic attainment as compared to responde nts with lower levels of socio-economic attainment. Theoretical Variables The following hypotheses demonstrate the expe cted empirical relationships between reporting victimization and each of the independen t variables, testing the Legitimacy of Law theory. Democracy H8: The likelihood of victimization will be high er among respondents who report lower levels of democracy within their country as compared to respondents who re port higher levels of democracy. Court justice H9: The likelihood of victimization will be high er among respondents who report lower levels of court justice as compared to respondents who report higher levels of court justice. 73

PAGE 74

Trust of police H10: The likelihood of victimization will be higher among respondents who report lower levels of trust of police as compared to respondents who report higher levels of trust of police. Vigilante justice H11: The likelihood of victimization will be higher among respondents who report higher levels of approval for vigilante justice as compared to individuals who have lower levels of approval for vigilante justice. 74

PAGE 75

CHAPTER 4 RESEARCH DESIGN AND METHODS Data and Sample Data for this dissertation pr oject were made available by the Latin American Public Opinion Project (LAPOP).1 The data were collected by resear ch teams from AmericasBarometer during 2006 and 2007 as part of the LAPOP, an ongoing 20 year resear ch effort started by a team of researchers at Vanderbilt University. Re searchers conducted interviews with over 27,000 individuals in twenty co untries including all countries in Cent ral America, the majority of South America, parts of the Caribbean, United States, and Canada ( Seligson, 2008). Data from 12 of the countries from LAPOP are included in this study. The countries were chosen based on the availability of the data. In order to ensure uniformity among samp les from differing countries, a team of researchers met in Costa Rica in May of 2006 to set up a common sample design that was employed throughout the region.2 Teams for each country constructed a multi-stage, stratified area probability sample of 1,500 individuals per country. Additionally, a universal informed consent form was administered to all individuals in the stud y and the study was approved for research on human subjects by the Vanderbilt Univ ersity Institutional Review Board (IRB). All researchers on the project were required to take and pass a certifying test on the human subjects protection materials used by Vanderbilt University. Finally, all data for the project were completely anonymous with no identifying inform ation attached, thus each participant is guaranteed complete confidentiality regard ing his/her responses (Seligson, 2008).3 1 Data were made available by the Latin American Studies department at the University of Florida. 2 Funding for this project was made available by a grant from the United Nations Deve lopment Programme (UNDP). 3 Copies of informed consent forms are available in the appendixes of each countries study and are available at www.LapopSurveys.org. 75

PAGE 76

Researchers designed surveys for this project in 10 different local languages commonly used throughout Latin America. In each country of interest, a survey was employed using nearly identical questions that varied primarily in their use of colloqui al vocabulary and/or language. Indigenous populations were interv iewed using surveys translated into their local languages (e.g., Quechua and Mymara in Bolivia). Additiona lly, LAPOP researchers developed different versions of the survey in English, French as well and Portuguese to be employed in various regions where those language s are dominant (Seligson, 2008).4 Survey items consisted of questions pertaini ng to issues of democracy, rule of law, corruption, victimization, as we ll as demographic information such as income, education, occupation, sex, age and additional information. For the purposes of this dissertation project, data analyses will focus on the data collected from Bolivia, Brazil, Chile, Costa Rica, the Dominican Republic, El Salvador, Guatemala, Ja maica, Mexico, Peru, Uruguay, and Venezuela. More Legitimate Countries Costa Rica Data Data from Costa Rica used in this project are from a national probability survey that was employed and collected in 2006. The sample (N= 1,500) was stratified by region (San Jos, the central valley, areas beyond the central valley and metropolitan area) as well as by rural or urban regions (Seligson, 2006c). Researchers conduc ted interviews in 194 sampling units and participants were selected into primary sampli ng units of six respondents in urban regions and 12 respondents in rural regions. The sample is repr esentative of Costa Ricas voting aged population and all respondents are at least 18 years of age. 4 Questionnaire forms from each country can be found in the a ppendixes of each country stud y and are available at www. LapopSurveys.org. 76

PAGE 77

Uruguay Data The data from Uruguay were collected in 2007 with a stratified, probabilistic and multistage sample design (N=1,200) (Seligson, 2007b). The sample is nationally representative of both rural and urban regions and is weighted by the proportion of the population in a given geographic region. The sample is weighted by re gion (Urban Montevideo, inner country cities with less than 2,000 inhabitants, inner country c ities with more than 2,000 inhabitants, and scattered rural populations). Res earchers collected data from a total of 292 sampling points: 118 in urban Montevideo, 149 in citie s with less than 2,000 inhabitants, 15 in cities with more than 2,000 inhabitants, 10 in scattere d rural populations. Researchers conducted 12 interviews in each rural sampling point but only four interviews in all other sampling points (Seligson, 2007b). Chile Data The data from Chile used in this project we re collected in 2006 a nd consist of a national probability sample (N=1,517) drawn from the vo ting aged population. Researchers employed a multistage sampling design stratified by munici pality (Seligson, 2006b). Researchers conducted interviews in 259 sampling units. Survey participants were selected in primary sampling units of approximately 8 in both urban and rural areas. Re searchers selected survey participants within households using a quota system base d on age and sex (Seligson, 2006b). Countries with Lower Levels of Legitimacy Bolivia Data The data from Bolivia used in this dissertation were collecte d using face-to-face interviews in 2006 (Seligson, 2006). The me thodology employed was a national household probability sample stratified by geopolitical departments. One individual from each selected household was chosen to participate in the survey. The resulting sample is representative of the population of Bolivia at the national and department al levels (N=3,013). It is also representative 77

PAGE 78

of all voting aged citizens in the country. The sample is we ighted by department population (Seligson 2006). Peru Data Researchers conducted surveys for the data fr om Peru used in this project in 2006. The survey design employed was a multistage nati onal probability design (N=1,500) stratified by locality (Seligson, 2006i). Partic ipants in the survey were selected from the entire voting-age population of the country. Researchers conducted in terviews in seven different regions of Peru where participants were chosen from primary sampling units of approximately five in urban areas and seven in rural areas. Researchers selected survey part icipants within households using a quota system based on age and sex (Seligson, 2006i). El Salvador Data A national probability design survey was empl oyed in 2006 in order to obtain the data from El Salvador that were used in this diss ertation (N=1,729). Sample stratification was done by both municipalities and rural and urban regi on. Researchers conducted interviews in 222 sampling units and selected six to eight res pondents per sampling unit for urban areas and 12 respondents in each rural primary sampling unit. Th e sample is representative of El Salvadors adult population (S eligson, 2006e). Guatemala Data The data from Guatemala used in this pr ojected were collected in 2006 and the survey method employed was a national probability design (N=1,498) strati fied both by region (northeastern, southeastern, southwestern, northw estern and metropolitan area) and by urban and rural areas. Researchers conducted interviews in 194 sampling units in which respondents were selected into primary sampling units of eight in urban areas and 12 in rural areas (Seligson, 2006f). All participants in the st udy were at least 18 years of age. 78

PAGE 79

Countries of Average Legitimacy Mexico Data The data from Mexico used in this project were collected in 2006 and represent a national stratified and clustered probability sampling design (N=1,560) (Seligson, 2006h). Sample stratification was done regionally (north, south, central, and mid-west regions) as well as by rural/urban location. Researchers conducted interviews in 130 locations in 19 Federated States of Mexico. This sample repres ents individuals from 72 of the 2,445 municipalities in Mexico as well as 127 electoral sections. Each cluster within a samp ling point was comprised of 12 respondents (Seligson, 2006h). All survey resp ondents were at leas t 18 years of age. Brazil Data The population of interest for the survey ad ministered in Brazil co ntained all Brazilian residents with citizenship over th e age of 16. The sample was wei ghted to be representative of countrys voting population in terms of age, sex, and geographic distribution5 (Seligson, 2007). The data from Brazil used in this project we re collected in 2007 a nd consist of a national probability sample (N=1,214) stratified by region (North, Northeastern, Southeastern, South and Mid-west). Researchers conducted interviews in 122 sampling units. Survey participants were selected in primary sampling units of six to ei ght in urban areas and 10 to 12 in rural areas (Seligson, 2007). Venezuela Data The population of interested for the survey administered in Venezuela contained all Venezuelan residents with citizenship over the age of 18 (Seligson, 2007c). Data from Venezuela used in this project were collected in 2007 and the sample was weighted to be 5 Voting in Brazil is allowed for any citizen aged 16 or older, but does not become compulso ry until an individuals 18th birthday. 79

PAGE 80

representative of the countrys voting aged population in terms of age, sex, and geographic distribution (Seligson, 2007c). Researcher s conducting the survey employed a national probability design (N=1,510) stratified by region (w est, mid-west, east, capital, Los Llanos, and Zuliana). The survey was administered in 186 sa mpling units where participants were selected in primary sampling units of eight in urban areas and 12 in rural ar eas (Seligson, 2007c). Dominican Republic Data Data from the Dominican Republic used in this dissertation were collected in 2006. A national probability design was employed in the implementation of the survey (N = 1,519). The sample was stratified regionall y (North, East, South, Santo Do mingo and Metropolitan area) and researchers conducted interviews in 227 sampling united in 67 municipalit ies and six to eight selected respondents per sampling unit for urba n areas and 10 to 12 respondents per rural primary sampling unit. A quota system was used to select respondents from each household based on sex and age (Seligson 2006d). Jamaica Data Researchers collected the data from Jamaica us ed in this dissertation in 2006. The survey methodology employed was a national probability de sign resulting in a sample representative of all Jamaican citizen residents over the age of 18 (N = 1595). Jamaica was stratified into three regions (Kingston Metropolitan Region, Parish Capitals and ma in towns, and Rural Areas). Researchers conducted interviews in 184 sampli ng units and respondents were selected into primary sampling units of eight in urban ar eas and 10 in rural ar eas (Seligson, 2006g). Advantages One of the key advantages of using the LAPOP data in order to study victimization is the uniformity of the data and of th e implementation of the survey de sign across nations. In the past, researchers have relied on homici de data in order to study patte rns of victimization between 80

PAGE 81

Latin American countries (Neapolitan, 1994) becaus e there are a limited number of datasets that contain information on victimization and have been applied uniformly across the region (Gaviria & Pags, 1999). Using homicide data is problematic because it does not account for the majority of victimization in Latin Ameri ca, property victimization (Buvini et al., 2005; Heinemann & Verner, 2006). Homicide data is also considered parti al and fragmentary (Buvini et al., 1999: 2) in part because in this Catholic dominant region, suicides are often listed as homicides. Additionally, unintentional deaths (e.g., car accidents) are sometimes listed as homicides (Heinemann & Verner, 2006). For this reason, it is e ssential to use a dataset that clearly defines victimization and measures it th rough self-reports. Self-reports are the most accurate way to gauge victimization in the regi on as official statistics are plagued with problems of underreporting (Gaviria & Pags, 1999; Heinemann & Ve rner, 2006). Researchers have used selfreported data from across Latin America to st udy victimization but us ing data that assess victimization at the household level, not th e individual level (Gaviria & Pags, 1999). Limitations The LAPOP data have some of the limitations that have applied to previous studies. First and foremost, the LAPOP survey is not a victim ization survey, but rath er a public opinion poll measuring issues of democracy, rule of law, and citizenry. This is problematic because questions regarding the extent or nature of the victimization are not included in the survey. There is also no way to ascertain any information regarding the pe rpetrator of the victimi zation. These are salient questions that help further th e understanding of vict imization, but are missing from this study. Other limitations of these data pertain to the na ture of the data and how they were collected. LAPOP researchers dealt with the majority of missing data in each countrys sample by substituting mean scores of the individual respon dents choices for any index or scale measure where there were missing data. This is a trad itional and commonly accepted method for dealing 81

PAGE 82

with missing data (Acock, 2005; Buhi, Goods on, & Neilands, 2008), and was employed only when the missing data amounted to less than 50 pe rcent of all the respons es for one individual (Seligson, 2008). While mean substitution was used for the majority of missing data in Latin American countries, it was not used for indi viduals who refused to respond to a question. Listwise deletion was used in these cases. Desc riptive statistics for the samples after listwise deletion of cases was performed did not significa ntly alter any of the findings for this study. Although missing data were not a major limitation of the data, the cross-sectional nature of the data was a limitation that needs some considerat ion. Given that the data were collected at one point in time, there is no way to infer causa l ordering with our dependent and independent variables. For this reason, all results indicate only an association between the indicators of legitimacy of law and victimization rather than an indication of causation or prediction of victimization. This is a limitation of using cross-sectional data and will be discussed further in the results secti on of this study. Measures The survey design for each country in this di ssertation project is ne arly identical. In the majority of cases, the measures between countries are also identical. However, there are a few differences in measures between countrie s and these differences are noted below. Dependent Variable The dependent variable in this analysis is victimization Respondents from all countries were asked a question pertaining to their victimization experience in the last year. In Spanish speaking countries respondents were asked: Ha sido usted vctima de algn acto de delincuencia en los ltimos 12 meses? In Brazil individuals were asked: O sr/sra foi vtima de algum ato de deliquncia (assalto, roubo, sequestro relmpago, etc..) nos ltimos doze meses? In the Jamaica respondents were asked: Have you been a victim of any type of crime in the past 12 82

PAGE 83

months? The same question was also asked in na tive languages in Bolivia and Guatemala (e.g., Nahuatl). A follow-up question was then asked to clarify the type of vi ctimization. Respondents indicated if they had been a victim of the following: robbery without physical aggression or threat; robbery with phys ical aggression or threat; physical aggression without robbery; rape or sexual assault; kidnapping; damage to property; home burglary. However, the follow-up question regarding the specific type of victimization was not asked in every country. For this reasons, it was not possible to delineate be tween physical victimization type s and property victimizations types. As a result, the first survey item, whic h was asked of respondents in all countries, was used to create a dichotomous measure of victimization for respondents in each country (1=victimization, 0= no victimization). Independent Variables Several demographic variables will be used to predict victimization in Latin American countries: sex, level of education, marital status age, race, number of children, and socioeconomic attainment. Age is a continuous variable measured in years (Buvini et al., 1999; Klaus & Rennison, 2002; Perkins, 1997). The re spondents number of children is also a continuous variable. Sex, race and marital status are all dichotomous measures with males, whites and married individuals coded as 1 and females, non-whites and unmarried individuals serving as the omitted categories (Kay, 2001; Kay, 2007). Socio-economic attainment is a measure of the individuals attainment of cons umer goods or services and serves as a proxy measure of both income and status. Individuals were asked about owne rship of the following items: television, refrigerator landline phone, cellular phone, vehicle, washing machine, microwave oven, motorcycle, potable running wate r, indoor bathroom and a personal computer. An additive index was created from these item s from 0 to 11 where 0 indicates that the individual has none of these items and 11 indicates that the individual has all of these items. 83

PAGE 84

Level of education is a continuous variable wher e 0 equals no education and 12 equals a college degree. Theoretical Variables Additional independent variable s represent measures of perc eived legitimacy of both the police and court systems. There are also meas ures of the individuals perception of how democratic the government in their country is as well as a measure of whether they believe their government protects citizen righ ts. Both of these measures assess the legitimacy of the government. Lastly, there is a measure assessing an individuals support for vigilante justice which also acts as a measure of the legitimacy of the state. Respondents were asked their levels of trust for the police in th eir country. Respondents answered questions using likert-type responses where 1 means there is no trust at all and 7 indicates that there is a lot of trust in the police. For the purposes of this study we assume that greater levels of trust in the police indicate that the police are seen as more legitimate. In addition to the question concerning trust in the police, respondents were also as ked a likert-type question gaugi ng the extent they believed the courts of justice in their country guarantee a fair trial. Responses ranged from 1 to 7 where 1 indicates not at a ll and 7 indicates a lot. In this study, court justice will be used as a measure of the legitimacy of the c ourt systems in each country. There are two measures of the legitimacy of the state/government in each country used in this study. The first measure is based on the li kert-type question where respondents were asked their level of approval/disapproval of people taking the law into their own hands when the State does not punish criminals. Responses ranged from 1 (strongly disapprove) to 10 (strongly approve). Approval of vigilante justice in this study will be used seen as a sign of lack of state legitimacy. Lastly, the level of democracy in th e country is measured through an individuals perception of how democratic they believe their country to be. Level of democracy is a variable 84

PAGE 85

85 with likert-type responses where individuals were asked if their countr y was very democratic, somewhat democratic, not very democratic, or not at all democratic. The variable was recoded so that higher scores indicate a greater perception of democracy in the respondents country (scores range from 1-4, with 1 being not at all de mocratic and 4 being very democratic). Analytic Strategy Separate logistic regression models will be estimated for each country due to the dichotomous nature of the dependent variables. An additional model will be estimated predicting victimization for the entire Latin American region. Th e use of logistic regression analyses is well-established statistical technique and is appropriately us ed with a dichotomous outcome (Allison, 1999; Long, 1997). The multinomial logistic regression models take the following form (Equation 5-1): (4-1) Log Where represents the odds of an individual experiencing vic timization relative to nonvictimization given any change in X Using the Statistical Package for the Social Sciences (SPSS), I first estimated 12 separate logistic regression models examining victimization in each Latin American country of interest to determine if patterns in victimization predictors emerged among those nations (Bolivia, Brazil, Chile, Cost a Rica, the Dominican Republic, El Salvador, Guatemala, Jamaica, Mexico, Peru, Uruguay, and Venezuela). Additionally I estimated a model for the entire Latin American region, merging data from all 12 countries.

PAGE 86

CHAPTER 5 PRELIMINARY RESULTS Sample Characteristics Table 5-1 presents means and standard devi ations for the variables included in the analysis for the total samples from Costa Rica, Uruguay and Chile, the th ree countries with the highest scores of democracy, legitimacy, and co untries generally agreed upon by scholars to be relatively stable economically and politically. Table 5-2 presents th e descriptive st atistics for the victimized sub-sample of Costa Rica, Uruguay a nd Chiles total sample. Table 5-3 presents the descriptive statistics for all respondents from Bolivia, Peru, Gu atemala and El Salvador. These four countries have relatively similar scores for legitimacy and democracy. The descriptive statistics for Bolivia, Peru, Guatemala and El Salvadors victimized sub-samples are found in Table 5-4. Table 5-5 presents the means and standa rd deviations for the va riables included in the analysis for the samples from Mexico, Brazil, Venezuela, the Dominican Republic and Jamaica. These countries are grouped together because th ey are relatively simila r in their levels of legitimacy and democracy. The descriptive statis tics for the victimized sub-samples from the previous 5 countries can be found in Table 5-6. The descriptive st atistics of the sample of all respondents from all countries combined is in Table 5-7. Table 5-8 presents the means and standard deviations for the victimized res pondents from the entire Latin American region. Dependent Variable Overall, 20 percent of responde nts in from the total Latin America sample report being victimized (victimization =1). When broken dow n by country, the numbers vary in comparison to this overall victimization rate. The lowest ra te of reported victimization was in Jamaica where 10 percent of the sample reported being victimized. The next lowest rates were in Brazil, Bolivia and El Salvador where 16 percent of respondent s reported victimization. In Costa Rica and the 86

PAGE 87

Dominican Republic 17 percent a nd 18 percent of the respondent s from the respective samples reported victimization. In Mexico 21 percent of respondents reported the being victimized. Twenty-two percent of Guatemalan and Uruguayan respondents reported victimization. In Chile just under a quarter of the sample (24 percent) reported being vi ctimized in the last year. The highest rates of victimization in the region were in Venezuela where 26 percent of the sample reported victimization in the previous year a nd in Peru where 27 per cent of respondents also reported victimization. Independent Variables Sex. In the total sample for Latin America, 50 percent of res pondents to the LAPOP survey were male (male =1). Similarly, 50 pe rcent of the respondents from the Dominican Republic were male. Of the Brazilian, Venezuelan, Jamaican, Peruvian and Guatemalan respondents 51 percent were also male. Costa Ri ca, Uruguay and El Salvador all had a sample with 49 percent male respondents. Bolivias samp le had the highest numbe r of male respondents, with 53 percent male. Chiles sample was 46 percent male. The Mexican sample overrepresented females with only 21 percent male respondents. The percentage of males in the victimized samples is slightly higher for the majority of countries. In the total victimized sub-sample fo r Latin America, 53 percent of those victimized were male. The greatest contrast is in Mexico where 63 percent of the victimized sample was male (verses 21 percent of the general Mexican sa mple). In Jamaica 60 percent of the victimized sample were male (verses 51 percent in the total sample for Jamaica). In Bolivia, 58 percent of the victimized sample was male while only 53 percent of the total Bolivian sample was male. Additionally, 57 percen t of Guatemala and El Salvadors vi ctims were male (verses 51 percent for Guatemala and 50 percent for El Salvador in th e total sample). In Peru and Brazil, 56 percent of those victimized were male. This is in c ontrast with the total sa mple populations for both 87

PAGE 88

countries, which were only 51 percent male. In Uruguay and Venezuela, 52 percent of victims were male while 49 percent of the overall sample in Uruguay was male and 51 percent of the Venezuelan total sample was male. In Costa Rica the number of males in both the victimized sample and non-victimized sample were equal wi th 49 percent of both sa mples being male. In the Dominican Republic, 52 percent of the victimi zed sub-sample was male (verses 50 percent male in the total sample). Chiles victimized sample was 47 percent male and roughly equivalent to the proportion of males in the total sample for that country (46 percent). Education Individuals across the Latin American region generally had some education the mean level of education for the region was 4.9, wh ich indicates that the majority of respondents in the region have at least some high-school e ducation (49 percent). Edu cation rates varied for individuals in each country. Per u, Bolivia, Venezuela and Chile had the most educated samples with scores of 6.1, 5.8, 5.6 and 5.6. This indicates that the majority of respondents in these countries have at least a highschool education. Brazil, Jamai ca, Uruguay, and Mexico had the second highest education leve ls with scores of 5.2, 5.0, 4.8, and 4.4. Costa Rica and El Salvadors levels of education were lower with scores of 4.3 and 4.2. In El Salvador 12 percent of the respondents had college training and 11 pe rcent of Mexican participants also reported having some college education. However, El Salv ador also had a high percentage of respondents with no educational training (11 percent). The Dominican Republic had a score of 4.0 years of education. Guatemala had the lowest level of educa tion with a score of 3.4 years. Guatemala also had the lowest number of individuals with colleg e training (6 percent) an d also had the highest number of respondents with no ed ucation at all (12 percent). When overall education levels for individuals among the victimized and the general samples were compared a few key differences are apparent. 88

PAGE 89

Across the region, victims had hi gher levels of education than those in the total sample, with levels of education scores of 4.9 in the total Latin America sample and. 5.7 for the victimized sub-sample from all Latin American c ountries. The largest disparities were seen in El Salvador, Costa Rica and Guatemala where among victims the level of education was between one and two points higher in the former and over full point higher in the latter two countries (4.0 versus 5.8; 4.0 versus 5.5; 4.2 versus 5.8). All ot her countries were within a point difference between the total samples and vic timized samples for each country. However, in all countries the mean level of education was greater among the vi ctimized sample when compared to the mean level of education for th e total country sample. It should be noted that in Mexico only five percent of the total sample reported having no education while among the victimized sample 35 percent reported having no educat ion. Further analysis is need ed to see if there are any statistically meaningful differences between victims and non-victims based on educational attainment. Marital Status. Fifty-seven percent of Latin Americans in the total sample were married (married =1). This number varies among the coun tries of the region. In the Jamaican sample 42 percent self-reported being married as did 48 percent of the Venezu elan sample. Brazils sample was predominately married (52 percent), as were and Uruguays (55 percent) and Chiles sample (57 percent married). In Bolivia, the Dominican Republic and El Salvador, 58 percent of the countries samples were married. In Peru, 60 percent of resp ondents were married while in Guatemala 64 percent were married. Mexico had the highest number of marri ed individuals in its sample with 69 percent of respondents being married. Among the victimized sample, marriage rates did not vary drastically. An exception is Costa Rica, where 59 percent of the victimized sample was married (versus 50 percen t in the un-victimized sample). 89

PAGE 90

Age The average age among all respondents of this survey across Latin America was approximately 38 years of age. The average age for victims from the sub-sample including all countries in the region was 37 years of age. Among the individual countries, the youngest samples were Guatemala where the mean age was 35 and Venezuela and Bolivia where in both countries the mean age was approximately 36. Ho wever, victims in Venezuela and Guatemala were closer to the age of 34 and victims in Bolivia were closer to 35. In El Salvador and Mexico the mean age was approximately 37. While in the victimized sample the mean age was 35 and 33 respectively. The Dominican Republic and Pe rus overall samples had mean ages of approximately 38 years of age while the victimized samples in these countries had mean ages of 36 and 37. In Brazil, respondents in the overall sample were approxi mately 39 years of age while victims were on average 3 years younger (37). Co sta Rican respondents mean age was 40 while, the victimized sample from Costa Rica had an av erage age closer to 38. The oldest samples were Jamaica (43), Chile (43) and Ur uguay (44). However, the average age in the victimized subsamples from these countries was 45, 41, and 45 respectively. In Jamaica the mean age in the victimized sample (45) was greater than the mean age in the country s total sample (43). Generally, with the noted excepti ons, victims were slightly younger than the overall sample. Race The minority of the sample for the total Latin America sample was White (White =1) (33 percent). However, when broken dow n by country these numbers begin to vary drastically throughout the region. Only 4 percent of Jamaicans in the total sample report being White. In Bolivia 11 percent of the overall sample reported being White, which is equal to the percentage of the victimized sample reporting th e same. In Peru, 13 percent of the overall sample was White (verses 12 percent in the victimized sample). In the Dominican Republic, 12 percent of the sample reported being White (verses 11 percen t of the victimized sample). In El Salvador 90

PAGE 91

16 percent of the total sample identified with being White (verses 17 per cent for the victimized sample). Forty-five percent of Brazils sample was white while 51 percent of victims were White. The Venezuelan total sample was 47 percent White and the victimized sample was 48 percent White. Half of Costa Ricas sample was White (50 percent) and 53 percent of the victimized sample was White as well. In Guatemala, a lower number of the overall sample was White (62 percent) verses the victimized sample (67 percent). In Chile and equal number where White in the total sample to those in the vict imized sub-sample. Both were 65 percent White. Lastly in Uruguay the majority of respondents in the sample were White, 80 percent. In the victimized sample a larger majority were also White (86 percent). Children Individuals in the Latin American to tal sample reported having roughly 2 children. This pattern held across the samples for each individual country with two exceptions. In Brazil, respondents had approximately 1 child wh ile in Jamaica, Guatemala and the Dominican Republic the sample averaged between 2 and 3 children per respondent. The number of children for respondents in each country among the victim ized samples did not vary dramatically (See Tables 5-1 thru 5-8). Socio-Economic Attainment The level of socio-economic attainment for the entire Latin American sample was 6 (out of a possible 11), in dicating that most indivi duals had a little more than half of the consumer goods/services. This nu mber was slightly higher among the victimized sub-sample from the total Latin America sample (6.5). Results for the individual countries were not dramatically different. The highest level of socio-economic attainment was in Costa Rica with an average score of 7.7 items per individual. Victims in Costa Rica had a greater level of socio-economic attainment (8.1). Relatively sim ilar scores of socio-economic attainment were found in Uruguay (7.4 for the total sample, 7.5 for victims), Chile (7.2 fo r the total sample, 7.5 91

PAGE 92

for victims) and Venezuela (7.0 for both samples) Following closely were Brazil (6.6 for the total sample, 6.8 for victims), Mexico (6.5 for th e total sample, 6.8 for victims) and Jamaica (6,1 for the total sample, 6.6 for victims). Countries that fell below the average for the region were the Dominican Republic (5.8 for the total sample 6.0 for victims), Guatemala (5.1 for the total sample, 5.8 for victims), Peru (4.8 for the total sa mple, 5.3 for victims), and El Salvador (4.8 for the total sample, 6.1 for victims). Bolivia (4.4 for the total sample, 5.2 for victims) had the lowest mean score for socio-economic attainment Across the region, the level of socio-economic attainment for victims was higher than that for the total mean of the country, excepting Venezuela, where the means were nearly identical. Theoretical Variables Level of Democracy. Most respondents in the region believed that their country is somewhat democratic, with a level of democr acy score for respondents from the total Latin American sample of 2.9 (with 4 indicating that respondents believed their country to be very democratic). Victims from the sub-sample of from all of Latin America had a mean score of 2.8, also indicating that they believe their country is some what democratic. Levels varied slightly from country to country throughout the region and between victimi zed and total country samples. Uruguayans had the highest mean score for leve l of democracy both amon g those in the total sample (3.4) and the victimized sample (3.3). Th is indicates that most citizens in Uruguay believe their country is between somewhat and very democratic. Costa Rica (3.1), the Dominican Republic (3.1) and Venezuela (3.1) all had scor es indicating that respondents believe their country to be somewhat democratic. Scores for victims in Costa Rica were higher (3.2) while scores in the victimized sub-sample for the Do minican Republic were identical in (3.1). In Venezuela the mean score for level of demo cracy among the victimized sub-sample was 3.0. Jamaica (3.0) and Chile (2.9) had scores indicatin g that individuals believe their country to be 92

PAGE 93

somewhat democratic. These scores did not ch ange among victims in each country. In Mexico (2.8), Brazil (2.8) and Bolivia ( 2.8) individuals responde d that their country was also somewhat democratic. Mean scores among victims in these countries were identical for Mexico and Brazil (2.8 for each country). However the perceived le vel of democracy among individuals from the victimized sub-sample in Bolivia was somewh at lower (2.3). Respondents from Peru (2.6), Guatemala (2.5) and El Salvador (2.5) had the lowest mean scores for level of democracy indicating that individuals in th ese countries believe their government to be between not very democratic and somewhat democratic. Victims in each country had even lower mean scores for level of democracy, 2.3 for respondents from Peru, 2.4 for respondents from Guatemala and 2.3 for respondents from El Salvador. Court Justice The mean score of court justice for the total Latin America sample was 3.8, indicating that there is a modera te level of legitimacy in th e court systems across the region. However, this does not indicate that respondents across the region feel that their court systems provide a great deal of justice (with the highe st possible score being 7). Among the victimized sub-sample for respondents from the total Latin America sample, the number was slightly lower with a mean score of 3.6, which still indicates that most citizens in the region believe there is a fair amount of justice in the court systems in thei r countries. As with trust in the police, score of the level of court justice vary by country acro ss the region. In Guatemala (4.2), Uruguay (4.2), Costa Rica (4.0), and Peru (3.0) the mean score of court justice was the same in both the total samples and the victimized samples for each countr y. The scores however for the total country in Mexico (4.1), Jamaica (4.1) the Dominican Repub lic (3.8) El Salvador (3.7), Chile (3.7) and Bolivia (3.6) indicated that there is a fair expectation of justice, or a fair amount of legitimacy in the court system. These findings were slightly lower among victims in each country, but not 93

PAGE 94

substantially so. The greatest discrepancy was found in Brazil and Venezuela. Among the total sample in Brazil, the trust in the belief that the court system assures justice was roughly 3.8, but among victims it was only 3.3. Likewise in Venezu ela, the mean score for court justice was 4.0 for the total sample but only 3.5 for victims. Later analyses will help distin guish if this variation is statistically significant. Trust in the Police Trust in the police from the total Latin America sample is fairly moderate 3.6 out of 7, indicating that there is also a moderate amount of police legitimacy. The mean score for trust in the police for the victimized sub-sample from the total Latin America sample is only slightly lower at 3.2. This number varies by country and among each countrys victimized sample. The highest level of trust in the police is in Cost a Rica where the mean response was 5.5, indicating a good deal of trust in the police. This number was dramatically lower among victims in Costa Rica where th e mean score of trust in the police was approximately 3. In Chile, the level of trust in the police was also rather high, where the mean score of trust in the police wa s 4.9. Among victims in Chile, the score was also high, 4.9. Brazil had the next highest score for tr ust in the police at 4.6; victims in Brazil had a slightly lower score for trust in the police at 4.3. Uruguayans also scored relatively high for trust in the police at 4.2; victims in Uruguay had a mean score of 3.7. El Salvadors mean score for trust in the police was 4.1 with victims in the sample having a mean trust score of only 3.5. Jamaicas mean score indicated that individuals were fairly trusting of police (3.7) wh ereas victims in Jamaica were slightly less trusting of the police (3.5). Guatemal a had a trust score of 3.5 for the total sample. Among the victimized sub-sample the trust of po lice score was slightly lower; 3.2. Peru (3.4), Mexico (3.2), Bolivia (3.2) and the Dominican Re public (3.1) all had similar scores on trust of police. Their victimized samples had similar scores as well (3.2, 2.8, 3.0 and 2.8 respectively). 94

PAGE 95

The country with the lowest level of trust in the police was Venezuela with a score of 2.8. Victims in Venezuela had a score of 2.3, also the lowest among countries in the region. Vigilante Justice On a scale of 1 to 10 indicating a pproval of citizens taking justice into their own hands when the state fails to do so, mo st respondents in the total Latin America sample show moderate approval of vigilante justice (3.8) where 10 indicates the most support for vigilante justice. Among victims in the total Latin America Sample there is a slightly higher level of approval for vigilante justice (4.1). The highe st level of approval for vigilante justice was found in the Dominican Republic both among the to tal country sample (4.7) and the victimized sample (4.9). Peru also had a high level of s upport for vigilante justic e among the total country sample (4.5) and slightly more approval among th e victimized sample (4.7). There was moderate approval of vigilante justice in Guatemala and El Salvador (a pproximately 4.2 for each country) and this support was higher among victims in each country (4.6 for Guatemala and 4.7 for victims from El Salvador). Costa Rica (4.1) an d Chiles (3.9) total popul ations also had some support for vigilante justice and as with other countries, the support was greater among victims (4.4 and 4.1 respectively). In Bolivia, the level of approval for vigilante justice was identical for both the total sample from the country and the victimized sub-sample (4.0). There were lower levels of approval for vigilante justice in Venezu ela (3.3) and Jamaica (3.2) but again the degree of approval was greater among victims (3.8 for vi ctims in Venezuela and 3.4 for victims in Jamaica). Mexico (2.9) and Brazil (2.8) were am ong the lowest levels of approval for vigilante justice both among the total sample and victim s (3.2 and 3.1 respectively). Among countries in the region, Uruguayans had the lowest level of approval for vigila nte justice both among victims and in the total sample for th e country (2.7 for each group). 95

PAGE 96

Table 5. Descriptive Statistics for Variables used in Costa Rica, Uruguay and Chile Models Costa Rica Uruguay Chile N= 1,430 N= 1,086 N = 1,443 Variable Mean S.D. Mean S.D. Mean S.D. Dependent variable Victimization 0.17 0.37 0.22 0.41 0.24 0.43 Independent variables Sex ( Male = 1 ) 0.49 0.50 0.49 0.50 0.46 0.50 Level of education (years) 4.34 2.97 4.78 2.82 5.64 2.66 Marital status (married =1) 0.59 0.49 0.55 0.49 0.57 0.49 Age (years) 40.38 16.51 44.33 17.49 42.92 16.49 Race (White =1) 0.50 0.50 0.80 0.39 0.65 0.48 Children (range 0-20) 2.28 2.16 2.02 1.94 2.13 1.90 Socio-economic attainment (1-11) 7.67 1.98 7.39 2.15 7.19 2.26 Theoretical variables Level of democracy (1-4) 3.14 0.76 3.36 0.69 2.93 0.67 Court justice (1-7) 4.01 1.68 4.20 1.69 3.69 1.67 Trust of police (1-7) 5.54 1.85 4.17 1.86 4.93 1.70 Vigilante justice (1-10) 4.11 3.40 2.66 2.72 3.89 3.05 Source: AmericasBarometer 2006 96

PAGE 97

97 Table 5. Descriptive Statistics for Victimized Sub-samples from Costa Rica, Uruguay and Chile Costa Rica Uruguay Chile N= 238 N= 238 N = 342 Variable Mean S.D. Mean S.D. Mean S.D. Independent variables Sex (Male = 1) 0.49 0.50 0.52 0.50 0.47 0.50 Level of education (years) 5.47 3.11 5.41 2.84 6.38 2.28 Marital status (married =1) 0.50 0.50 0.53 0.50 0.59 0.49 Age (years) 38.52 16.23 44.85 17.81 41.44 14.96 Race (White =1) 0.53 0.50 0.86 0.34 0.65 0.48 Children (range 0-20) 2.28 2.16 1.99 2.04 2.12 1.84 Socio-economic attainment (1-11) 8.07 2.11 7.50 2.04 7.49 2.22 Theoretical variables Level of democracy (1-4) 3.15 0.73 3.34 0.74 2.88 0.68 Court justice (1-7) 3.96 1.73 4.17 1.76 3.42 1.70 Trust of police (1-7) 3.05 1.78 3.74 1.81 4.85 1.81 Vigilante justice (1-10) 4.41 3.51 2.70 2.80 4.11 3.16 Source: AmericasBarometer 2006

PAGE 98

Table 5. Descriptive Statistics for Variables used in Bolivia, Peru, Guatemala and El Salvador Models Bolivia Peru Guatemala El Salvador N= 2,562 N= 1,370 N = 1,111 N= 1,607 Variable Mean S.D. Mean S.D. Mean S.D. Mean S.D. Dependent variable Victimization 0.16 0.37 0.27 0.44 0.22 0.41 0.16 0.37 Independent variables Sex ( Male = 1 ) 0.53 0.50 0.51 0.50 0.51 0.50 0.49 0.50 Level of education (years) 5.75 3.10 6.14 2.66 3.43 2.74 4.21 3.06 Marital status (married =1) 0.58 0.49 0.60 0.49 0.64 0.48 0.58 0.49 Age (years) 36.24 14.95 37.98 15.02 35.20 13.47 36.79 15.26 Race (White =1) 0.11 0.31 0.13 0.33 0.62 0.49 0.16 0.37 Children (range 0-20) 2.38 2.39 2.08 2.04 2.47 2.44 2.24 2.22 Socio-economic attainment (1-11) 4.43 2.55 4.84 2.38 5.07 2.81 4.77 2.65 Theoretical variables Level of democracy (1-4) 2.77 0.76 2.56 0.77 2.50 0.70 2.53 0.88 Court justice (1-7) 3.64 1.52 3.02 1.48 4.23 1.56 3.69 1.74 Trust of police (1-7) 3.24 1.68 3.35 1.67 3.53 1.77 4.11 1.88 Vigilante justice (1-10) 3.94 2.99 4.51 3.24 4.15 3.07 4.20 3.26 Source: AmericasBarometer 2006 98

PAGE 99

Table 5. Descriptive Statistics Vic timized Sub-samples from Bolivia, Peru, Guatemala and El Salvador Bolivia Peru Guatemala El Salvador N= 406 N= 372 N = 239 N= 258 Variable Mean S.D. Mean S.D. Mean S.D. Mean S.D. Independent Variables Sex (Male = 1) 0.58 0.49 0.56 0.50 0.57 0.50 0.57 0.50 Level of education (years) 6.52 2.96 6.50 2.53 4.37 2.92 5.81 2.99 Marital status (married =1) 0.54 0.50 0.58 0.49 0.64 0.48 0.54 0.50 Age (years) 35.21 14.19 37.06 15.41 33.90 11.89 32.73 13.85 Race (White =1) 0.11 0.31 0.12 0.33 0.67 0.47 0.17 0.38 Children (range 0-20) 1.98 2.10 1.94 1.93 2.39 2.67 1.64 2.65 Socio-economic attainment (1-11) 5.18 2.46 5.28 2.34 5.84 2.81 6.06 2.65 Theoretical variables Level of democracy (1-4) 2.68 0.75 2.60 0.77 2.37 0.78 2.27 0.86 Court justice (1-7) 3.45 1.57 2.98 1.52 4.23 1.62 3.46 1.57 Trust of police (1-7) 2.95 1.68 3.16 1.64 3.23 1.82 3.54 1.82 Vigilante justice (1-10) 4.03 3.12 4.70 3.32 4.57 3.27 4.70 3.36 Source: AmericasBarometer 2006 99

PAGE 100

Table 5. Descriptive Statistics for Variab les used in Mexico, Brazil, Venezuela, the Dominican Republic, and Jamaica Models Dominican Mexico Brazil Venezuela Republic Jamaica N= 1,389 N= 1,080 N = 1,111 N= 1,607 N= 1,055 Variable Mean S.D. Mean S.D. Mean S.D. Mean S.D. Mean S.D. Dependent Variable Victimization 0.21 0.37 0.16 0.37 0.26 0.44 0.18 0.38 0.10 0.30 Independent Variables Sex (Male = 1) 0.21 0.41 0.51 0.50 0.51 0.50 0.50 0.50 0.51 0.50 Level of Education (years) 4.41 2.75 5.23 3.31 5.57 2.84 4.00 2.98 4.97 2.71 Marital Status (married =1) 0.69 0.46 0.52 0.50 0.48 0.50 0.58 0.49 0.42 0.49 Age (years) 37.04 14.03 39.08 16.07 36.01 13.81 38.00 15.36 42.53 15.93 Race (White =1) 0.22 0.41 0.45 0.50 0.47 0.50 0.12 0.33 0.04 0.06 Children (range 0-20) 2.31 2.16 0.82 1.18 2.11 2.13 2.48 2.38 2.65 2.59 Socio-Economic Attainment (1-11) 6.52 2.42 6.57 2.04 7.03 2.01 5.75 2.41 6.08 2.53 Theoretical Variables Level of Democracy (1-4) 2.78 0.68 2.77 0.74 3.11 0.92 3.09 0.79 2.98 0.63 Court Justice (1-7) 4.12 1.68 3.75 1.77 3.96 1.84 3.75 1.91 4.16 1.64 Trust of Police (1-7) 3.23 1.82 4.59 2.01 2.75 1.89 3.11 2.08 3.69 1.83 Vigilante Justice (1-10) 2.93 2.67 2.76 2.71 3.27 3.02 4.67 3.78 3.20 2.80 Source: AmericasBarometer 2006 100

PAGE 101

101 Table 5. Descriptive Statistics for Victim ized Sub-Samples from Mexico, Brazil, Ve nezuela, the Dominican Republic, and Jamaic a Dominican Mexico Brazil Venezuela Republic Jamaica N= 288 N= 177 N = 708 N= 412 N= 100 Variable Mean S.D. Mean S.D. Mean S.D. Mean S.D. Mean S.D. Independent Variables Sex (Male = 1) 063 0.48 0.56 0.50 0.52 0.50 0.48 0.50 0.60 0.49 Level of Education (years) 4.96 2.71 6.09 3.14 6.14 2.75 4.59 3.15 5.11 2.79 Marital Status (married =1) 0.64 0.48 0.47 0.50 0.46 0.50 0.55 0.50 0.47 0.50 Age (years) 34.89 12.91 37.40 14.95 33.75 11.97 36.44 14.53 44.87 15.78 Race (White =1) 0.22 0.41 0.51 0.50 0.48 0.50 0.11 0.32 0.00 0.00 Children (range 0-20) 1.98 1.90 0.73 1.05 1.81 1.84 2.23 2.13 2.29 2.27 Socio-Economic Attainment (1-11) 6.82 2.42 7.11 2.02 7.03 2.07 6.03 2.32 6.64 2.50 Theoretical Variables Level of Democracy (1-4) 2.76 0.73 2.76 0.75 2.98 0.96 3.08 0.82 3.00 0.59 Court Justice (1-7) 3.81 1.74 3.31 1.76 3.51 1.92 3.44 1.87 3.98 1.68 Trust of Police (1-7) 2.78 1.68 4.30 2.11 2.33 1.72 2.77 2.01 3.48 1.86 Vigilante Justice (1-10) 3.20 2.96 3.09 3.06 3.78 3.24 4.91 3.84 3.35 3.08 Source: AmericasBarometer 2006

PAGE 102

Table 5. Descriptive Statistics for Variab les used in Latin America Total Sample Variables Latin America Total Sample (N=19,193) Mean S.D. Dependent Variable Victimization 0.20 0.40 Independent Variables Sex (male =1) Level of Education (years) 0.50 4.94 0.50 3.01 Marital Status (married =1) Race (White =1) Children (range 0-20) Socio-Economic Attainment (0-11) 0.57 0.33 2.20 6.04 0.47 15.48 2.60 Theoretical Variables Level of Democracy (1-4) Court Justice (1-7) Trust of Police (1-7) Vigilante Justice (1-10) 2.89 3.83 3.55 3.76 0.81 1.73 1.95 3.18 Source: AmericasBarometer 2006 102

PAGE 103

103 Table 5. Descriptive Statistics for Victimi zed Sub-sample for Latin America Sample Variables Latin America Victimized Sample (N=3,778) Mean S.D. Independent Variables Sex (male =1) Level of Education (years) 0.53 5.73 0.50 2.92 Marital Status (married =1) Age (years) Race (White =1) Children (range 0-20) Socio-Economic Attainment (0-11) 0.54 36.71 0.37 1.94 6.53 0.50 14.66 0.48 1.99 2.47 Theoretical Variables Level of Democracy (1-4) Court Justice (1-7) Trust of Police (1-7) Vigilante Justice (1-10) 2.83 3.57 3.18 4.05 0.85 1.76 1.93 3.33 Source: AmericasBarometer 2006

PAGE 104

CHAPTER 6 VICTIMIZATION IN LATIN AMERICA Prior to testing the impact of legitimacy on personal victimization in Latin America, we will first look at the socio-demographic pred ictors of victimization among Latin American countries to see how or whether victimization va ries among the relevant countries. First, we will examine the predictors of victimization for the region as a whole and then we will look at the individual countries to see whether or not the f ace of victimization in the region varies from country to country. Bivariate Analyses and Multicollinearity Diagnostics Before conducting the multivariate analyses, we will first examine a series of bivariate correlations. We will look at the region as a whole and then look at each country separately to determine multicollinearity among key variables. Latin American Region For the entire Latin American region, the result s revealed significant bivariate relationships between the dependant variable a nd all of the demographic measures used in this study. In the entire region, there is a significant and positive correlation between victimization and sex (male = 1), race (White = 1), the level of education an d socio-economic attainme nt. There also exists a significant negative correlation between victimi zation and the individuals marital status (married = 1), age and the numbe r of children he/she has. Am ong the independent variables a number of the measures were significantly correlat ed with one another. Given some of the strong bivariate correlations between th e independent variables, I ran a test of the variance inflation factors (VIF) scores to check of multicollinearity. Fisher and Mas on (1981) proposed a cut off of four for VIF scores as a measure of multicollinearity, and this is the cut off used in this project as well. In the results for the full model, no VIF scores exceeded two, therefore, multicollinearity 104

PAGE 105

does not appear to be an issue for the independe nt variables in the full model for all of Latin America. Also, none of the co rrelations exceed a value of r .5, so multicollinearity is not an issue. However, when examining the results for indi vidual countries, there are fewer significant correlations both between the dependent and independent va riables and the independent variables. Therefore each countrys da ta will be examined individually. More Legitimate Countries Costa Rica Bivariate analyses for Costa Rica show that victimization is signi ficantly and positively correlated with only le vel of education ( r = .17) and socio-economic attainment ( r = .09). There is a significant negative correlation betw een victimization and marital status (r = -.08) and number of children ( r = -.07). Due to the significant bivariate corre lation among independent variables in this model VIF scores were esti mated and showed that no score exceeded two, therefore multicollinearity was not an issue. Uruguay The bivariate analyses for Uruguay show that there is a significan t positive correlation between victimization and an i ndividuals race (White = 1) ( r = .07) and level of education ( r = .12). There are no significant nega tive correlations in the bivari ate analyses results for Uruguay. Multicollinearity is not a problem in this m odel as none of the independent variables were correlated at a level of r .05. Chile For Chile, the bivariate analyses reveal a significant positive correlation between the dependent variable and an individuals level of education ( r = .16) and socio-economic attainment ( r = .08). There are no significant negative co rrelations. Age and number of children 105

PAGE 106

were again highly correlated ( r = .54). However, in Chile, despite significant correlations between the independent variables, VIF scores re vealed that none of the variables exceeded values of two, therefore multicollinearity is no t considered a problem for the Chile model. Countries with Lower Levels of Legitimacy Bolivia The results of bivariate analyses for Bo livia show significant positive correlations between victimization and sex (male = 1) ( r = .04), level of education ( r = .11) and socioeconomic attainment ( r = .13). The results also show a si gnificant negative correlation between victimization and num ber of children ( r = -.07). Bolivias bivariate analysis similarly showed positive correlations among the independent variables. As was found in Peru, in Bolivia age and level of education were correlated at a level of r = .62. However, when VIF scores were estimated, none exceeded a value greater than th ree ruling out issues of multicollinearity. Peru Results from Peru show that the dependent variable is significantly and positively correlated with sex (male = 1) ( r = .07), level of education ( r = .08) and socio-economic attainment ( r = .11). The results show no significant negative correlations. The results also showed that among the independent variables there were a number of significant correlations. However only age and level of educa tion were correlated at a level of r = .64. Due to this high level of correlation, VIF scores we re estimated showing that none of the variables had scores greater than three. Therefore, multicollinearity wa s ruled out as a problem in the Peru model. El Salvador For El Salvador, the dependent variable is significantly and positively correlated with an individuals sex (male = 1) ( r = .07), level of education ( r = .23) and socio-economic attainment ( r = .21). The dependent variable is also significantly and ne gatively correlated with an 106

PAGE 107

individuals age ( r = -.12) and number of children ( r = -.12). Among the independent variables level of education and socio-economic attainment were significantly and positively correlated ( r = .60) as were number of children and age (r = .61). As with the other models, VIF scores were calculated to rule out multicollinear ity and no score was greater than two. Guatemala In Guatemala, victimization is significantly and positively correlated with sex (male = 1) ( r = .07), race (White = 1), level of education ( r = .18) and socio-economic attainment ( r = .15). There is a significant and negativ e correlation between victimizati on and the individuals age ( r = -.05) and number of children ( r = -.02). A number of the inde pendent variables are also positively correlated however only the associati on between age and number of children has an r value greater than .5 ( r = .57). To rule out multicollinearity, VI F scores were calculated for the independent variables in this model, and no scores exceeded two. Countries of Average Legitimacy Mexico For Mexico, victimization is significantly and positively correlated with sex (male = 1) ( r = .14), level of education (r = .10) and socio-economic attainment ( r = .07). The results for Mexico also show a significant negative association for age ( r = -.08), marital status (married = 1) ( r = .07) and the number of child ren present in the home ( r = -.08). The correlation between age and number of children was greater than r .50 ( r = .58). However, as with the full model, multicollinearity diagnostics for the independent va riables for Mexico do not reveal VIF scores greater than two and therefore, despite some significant correlations between independent variables, there does not appear to be a problem with multicollinearity between the independent measures. 107

PAGE 108

Brazil Results for Brazil show a positive significant correlation between th e dependant variable and level of education ( r =. 17) and socio-economic attainment ( r = .12). There are no negative significant correlations. As with the other mode ls, many of the independent variables in the bivariate analysis results for Brazil were pos itively correlated, however no variables are correlated at a level of r .05 and upon estimation of the VIF scores, no score exceeded one and a half; therefore multicollineari ty is not an issue among the variables in this model. Venezuela For Venezuela, victimization is has a positive significant corr elation with the individuals level of education ( r = .12). Victimization is significantl y and negatively correlated with an individuals age ( r -.09) and number of children ( r = -.08). Among the independent variables, number of children and age were correlated at r = .61. However an estimation of VIF scores showed that none of the variables had scores gr eater than two, ruling out multicollinearity issues for this model. Dominican Republic The bivariate analyses results for the Dominican Republic show a significant and positive correlation between the dependant va riable and level of education ( r = .09) and socio-economic attainment ( r = .05). The dependant variable is signifi cantly and negatively co rrelated with age (r = -.05) and number of children ( r = -.05). Again in this model, age and the num ber of children were highly correlated ( r = .63). When VIF scores are estimated, however, none exceed two. As a result, multicollinearity is ruled out as an issue in this model. Jamaica For Jamaica, there is a significant positiv e correlation between victimization and sex (male = 1) ( r = .07) and socio-economic attainment ( r = .07). There are no other significant 108

PAGE 109

correlations with the dependent variable. Among the independent variables no two variables are correlated at a level greater than r .05, therefore in the Jamaica m odel, multicollinearity is not considered a problem. Multivariate Analyses Once multicollinearity was ruled out among the models, binary logistic regression models are run on the non-theoretical vari ables to get a sense of the di fferences among models between Latin American countries and between those coun tries and the results for a model including all countries in the region. These models test the hypotheses outlin ed in Chapter 4 (H1, H2, H3, H4, H5, H6 and H7). Latin America Region Results from the logistic re gression model for the entire La tin American region appear in Table 6-1. The table includes the beta and odds ratio (Odds) for each variable. The Model Chisquare ( ) and the Pseudo R are also presented to assess the goodness of fit for the estimated model. The Model Chi-square stat istics indicate that the full model is a better fit than the model with only the constant with all the coefficients being held to zero. Results are presented in Table 6-1 ( =412.08). It was found that the model fits si gnificantly better as a whole model than as an empty model. Table 6-1 shows that, in the model for the entire Latin American region, four demographic variables are signifi cantly predictive of personal victimization: sex (male =1), level of education, race (White =1), and socio-economic attainment. In support of H1, as stated in Chapter 5, male respondents have a greater likelihood of victimiz ation than female respondents (b = .17, p .01). Additionally, H2 is supported since respondents with higher levels of education are more likely to be victimized as compared to respondents with lower le vels of education (b = 109

PAGE 110

.09, p .01). Additionally H5 was supported in our results showing that being White was significantly and positively associated with vict imization, as compared to non-White respondents (b = 0.14, p .01). Lastly, in keeping with H7, individuals with a greater level of socioeconomic attainment had a higher likelihood of being victimized (b = .05, p .01). In terms of percent probability, the likelihood of victimization increases 19 percent for males, for each additional year of education the likelihood of victimizati on increases by nine per cent. The likelihood of victimization increases by 16 percent for Whites. Lastly, for each incremental increase in socioeconomic attainment, the risk of victimization increases by five percent. More Legitimate Countries Costa Rica The assessment of fit for the Costa Rica model (Table 6-2) showed that the full model was significant and a better fi t than an empty model ( =48.11). In this model, education and marital status (married =1) are significantly re lated to victimization. In support of H1, as the level of education increases, the risk of vict imization also increases (b = .13, p .01). Additionally, married individuals have a lower probability of victimization, which is in keeping with H3, which stated that non-married indivi duals would have a lower risk of victimization (b = -.36, p .05). Expressed in odds ratios, this means that among respondents in Costa Rica, as education increases, for each one year increase the odds of victimization increases by 13 percent. Being married decreases the odds of victimization by 30 percent. Uruguay Table 6-2 presents the results for the Urugua y model. This model was significant and a better fit than an empty model ( =25.30). Logistic regression resu lts showed that only level of education and race (White =1) was associated with the risk of victimization. Our first hypothesis is supported showing that in Uruguay, as the leve l of education increases, so does the risk of 110

PAGE 111

victimization (b = .13, p .01). Additionally, H5 is supported as being White is positively and significantly associated with vi ctimization, as compared to non-White respondents (b = .13, p .05). Switching these findings into odds ratios shows us that for every one unit increase in educational attainment the odds of victimization rises by 15 percent. Furthermore, being white in Uruguay increases the odds of victimization by 53 percent. Chile The Chi-Square for the Chilean model, shown in Table 6-2, reveals that the full model holds greater predictive power than an empty model ( =42.03). Logistic regres sion results indicate that education and the number of children are asso ciated with victimization in Chile. In keeping with our first hypothesis, as the le vel of education increases, the ri sk of victimization increases (b = .17, p .01). Contrary to the proposal in H6, results show that the lik elihood of victimization in Chile is higher among individuals with more children. In other words, as the number of children an individual has increases, so does the odds of victimization (b = .10, p .05). When converted to odds ratios, the results indi cate that increases in educat ion also increase the odds of victimization by 19 percent for each one unit that education raises The presence of children also increases the likelihood of victim ization by 10 percent for each ch ild that the individual has. Countries with Lower Levels of Legitimacy Bolivia Logistic regression analyses for Bolivia appear in Table 6-3. Model diagnostics reveal that the full model is a better fit than an empty model ( =58.35). Results show that only H7 is supported in the Bolivia n model. As a socio-economic atta inment increases, the risk of victimization increases (b = .11, p .01). Converting these findings to percent probability, we see that for each one unit increase in a respondents socio-economic attainment the probability of victimization increases by 12 percent. 111

PAGE 112

Peru Results from the logistic regre ssion analysis of data from Pe ru appear in Table 6-3. These results show that the full model is a better fit th an the model with only the constant with all the coefficients being held to zero ( =27.22). In this model, onl y hypotheses are supported. As predicted in H1, males have a greater lik elihood of being victimized in Peru (b = .31, p .01). Additionally, in keeping with the prediction in H7, respondents with higher levels of socioeconomic attainment have a greater likelihood of victimization (b = .09, p .01). When we express these findings in odds ratios, we see th at males in Peru have a 37 percent greater likelihood of being victimized than females. Additionally, as socio-economic attainment increases by one unit, the proba bility of being victimized al so increases by 10 percent. Guatemala The model Chi-Square statistics for Guatemala, presented in Table 6-3, show that the full model is more appropriate than an empty model ( =47.36). Regression statistics show that three socio-demographic indicators are associated with victimization in Guatemala: sex (male =1), education and age. In support of H1, males in Guatemala have a greater probability of victimization than females (b = .30, p .05). Hypothesis 2 is also supported showing that respondents with greater levels of education have a greater like lihood of victimization (b = .12, p .01). Age is inversely related to victimization which supports H4 which predicts that younger respondents will have a greater li kelihood of victimization as co mpared with older respondents (b = -.02, p .05). When converted to probabilities, we see that in Guatemala males have a 35 percent higher likelihood of being victimized. For each incremental incr ease in the level of a respondents education, the likelihoo d of victimization increases by 13 percent. Lastly, there is a 112

PAGE 113

two percent decrease in the lik elihood of victimization for every one unit increase in an individuals age. El Salvador Model diagnostics for the data from El Salvador show that the full model is significant and a better fit than an empty model ( =109.47). Results from this model are in Table 6-3 and show that three of the proposed hypothe ses are supported in this m odel. Education, age and socioeconomic attainment are all significantly related to victimization in El Salvador. As predicted in H2, individuals with higher levels of education are more likely to be victimized in El Salvador (b = .11, p .01). Hypothesis 4 is also s upported; younger individuals in Guatemala have a greater probability of victimization as the odds of vic timization decrease as age increases (b = -.02, p .05). As socio-economic attainment increases, the risk of victimization also increases which is in keeping with the prediction in H7 (b = .14, p .01). The odds ratios of these findings indicate that for every one unit increase in education an individuals risk of vict imization increases by 12 percent. As an individuals age increases, for every one year increase in age there is a two percent decrease in the likelihood of victimization. Finally, for ev ery one unit increase in socioeconomic attainment the possibility of victimization increases by 15 percent. Countries of Average Legitimacy Mexico Results from the logistic regr ession analysis of data from Mexico appear in Table 6-4. These results show that the full model is a better fit than the model with only the constant with all the coefficients being held to zero ( = 42.18). In this model, only H1 was supported; victimization was more likely am ong males in Mexico (b = .65, p .01). The likelihood of victimization for males is 95 percent higher than it is for females. 113

PAGE 114

Brazil Chi-Square statistics indicate that the model for Brazil is significant and a better fit than a model with only the constant ( =25.91). Regression analysis wa s performed, shown in Table 64. In this model from Brazil only H2 and H7 were supported. A respondents level of education was significantly as positively associated with victimization. As education increases so does the odds of victimization (b = .07, p .05). Similarly, as the level of socio-economic attainment increases (1 to 11), the probability of victimization increases (b = .11, p .01). For every one unit increase in both education and socio-economic attainment the likelihood of victimization also increases by seven percent and 11 percent respectively. Venezuela The results for Venezuela are shown in Table 6-4. Model diagnostics for Venezuela reveal a significant model with a bett er fit than the empty model ( =50.45). Logistic regression results indicate that only education and age are predictive of victimizati on (both measure in years). As predicted in H2, individuals with a higher educational leve l are at a greater risk of victimization (b = .08, p .01). Conversely, as age increases the risk of victimization decreases (b = -.01, p .01), which supports the prediction made in H5. The conversion of these fi ndings illustrate that as education rises by one year, the probability of victimization also rises by nine percent. Conversely, as age increases by one year, the probability of victim ization decreases by a slight one percent. Dominican Republic Model diagnostics on the regression for data from the Dominican Republic show that the full model is more appropriate the model with only the constant ( =23.90). Results from the logistic regression analyses, shown in Table 6-4, reveal that only H2 is supported in this model. 114

PAGE 115

As the level of education increases, the risk of victimization also increases (b = .06, p .01). This means that among respondents from the Dominican Republic for each yearly increase in education the probability of victimization incr eases by seven percent. Jamaica In the final country model for Jamaica, shown in Table 6-4, model diagnostics reveal that the full model is a better fit than an empty model ( =18.27). In this model, age, number of children and socio-economic attainment are si gnificantly associated with victimization, supporting H4, H6, and H7. In Jamaica, as age increases so does the risk of victimization (b = .02, p .05). Individuals who have more children are at a lowered risk of victim ization in Jamaica (b = -.13, p .05) Greater levels of socio-economic atta inment are also associated with a greater risk of victimization (b = .09, p .05). For each one year increase in age, the risk of victimization increases by two perc ent. Additionally, for each additional child that an individual has, the risk of victimization decreases by 22 perc ent. Lastly, for each one unit increase in socioeconomic attainment, the odds of victimization rise by 10 percent. 115

PAGE 116

Table 6. Logistic Regression Analysis based on Victimiza tion for Total Latin America Sample: Non-Theoretical Model Variables Latin America Total Sample (N=19,193) Beta Odds Ratio Independent Variables Sex (male =1) Level of Education (years) 0.17 0.09 1.19** 1.09** Marital Status (married =1) Age (years) Race (White =1) Children (range 0-20) Socio-Economic Attainment (0-11) -0.03 -0.01 0.14 -0.00 0.05 0.97 1.00 1.16** 1.00 1.05** Constant -2.09 0.12 Chi-Square R-Square 412.08** 0.03 p .05, ** p .01: Source: AmericasBarometer 2006 116

PAGE 117

117 Table 6-2. Logistic Regression Analyses based on Victimization for Costa Rica, Uruguay and Chile: Non-Theoretical Model Costa Rica Uruguay Chile N= 1,430 N= 1,086 N = 1,443 Variable Beta Odds Ratio Beta Odds Ratio Beta Odds Ratio Independent Variables Sex (Male = 1) 0.08 1.09 0.21 1.23 -0.01 0.99 Level of Education (years) 0.13 1.14** 0.13 1.14** 0.17 1.19** Marital Status (married =1) -0.36 0.70* -0.08 0.92 0.05 1.05 Age (years) -0.00 1.00 0.00 1.00 -0.00 1.00 Race (White =1) 0.12 1.13 0.43 1.53* -0.00 1.00 Children (range 0-20) 0.02 1.02 0.03 1.03 0.10 1.10* Socio-Economic Attainment (1-11) 0.03 1.03 -0.05 0.95 -0.01 0.99 Constant -2.26 0.10 -2.17 0.11 -2.16 0.12 Chi-Square 48.11** 25.30** 42.03** R-Square 0.06 0.04 0.04 p .05, ** p .01: Source: AmericasBarometer 2006

PAGE 118

Table 6-3. Logistic Regression Analyses on Victimization for Boliv ia, Peru, Guatemala and El Salv ador: Non-Theoretical Models Bolivia Peru Guatemala El Salvador N= 2,562 N= 1,370 N = 1,111 N= 1,607 Variable Beta Odds Ratio Beta Odds Ratio Beta Odds Ratio Beta Odds Ratio Independent Variables Sex (Male = 1) 0.20 1.22 0.31 1.37** 0.30 1.35* 0.23 1.26 Level of Education (years) 0.04 1.04 0.03 1.03 0.12 1.13** 0.11 1.12** Marital Status (married =1) -0.03 0.97 -0.04 0.96 0.16 1.17 0.07 1.07 Age (years) 0.00 1.00 -0.01 0.99 -0.02 0.98* -0.02 0.98* Race (White =1) -0.20 0.82 -0.11 0.90 0.07 1.07 0.02 1.02 Children (range 0-20) -0.06 0.95 0.04 1.04 0.07 1.07 0.00 1.00 Socio-Economic Attainment (1-11) 0.11 1.12** 0.09 1.10** 0.06 1.06 0.14 1.15** Constant -1.54 0.22 -2.00 0.14 -2.50 0.08 Chi-Squared 58.35** 27.22** 47.36** 109.47** R-Squared 0.04 0.03 0.07 0.11 p .05, ** p .01: Source: AmericasBarometer 2006 118

PAGE 119

119 Table 6-4. Logistic Regression Analyses on Victimization for Mexi co, Brazil, Venezuela, the Dominican Republic, and Jamaica: No nTheoretical Models Dominican Mexico Brazil Venezuela Republic Jamaica N= 1,389 N= 1,080 N = 1,111 N= 1,607 N= 1 ,055 Variable Beta Odds Ratio Beta Odds Ratio Beta Odds Ratio Beta Odds Ratio Beta Odds Ratio Independent Variables Sex (Male = 1) 0.65 1.91** 0.25 1.28 0.05 1.05 -0.08 0.92 0.35 1.41 Level of Education (years) 0.05 1.05 0.07 1.07** 0.08 1.09** 0.06 1.07** 0.01 1.01 Marital Status (married =1) -0.10 0.90 -0.18 0.84 -0.03 0.97 -0.04 0.96 0.16 1.17 Age (years) -0.01 0.99 -0.00 1.00 -0.01 0.99* -0.00 1.00 0.02 1.02* Race (White =1) 0.01 1.01 0.13 1.13 0.05 1.05 -0.20 0.82 -19.00 0.00 Children (range 0-20) -0.00 1.00 -0.01 0.99 -0.01 0.99 -0.02 0.98 -0.13 0.88* Socio-Economic Attainment (1-11) 0.02 1.02 0.11 1.11* -0.03 0.98 0.03 1.03 0.09 1.10* Constant -1.58 0.21 -2.73 -1.-2 0.36 -1.70 0.18 -3.61 0.03 Chi-Squared 42.18** 25.91** 50.45** 23.90** 18.27** R-Squared 0.05 0.04 0.03 0.02 0.04 p .05, ** p .01: Source: AmericasBarometer 2006

PAGE 120

CHAPTER 8 VICTIMIZATION AND LEGITIMACY Bivariate Analyses and Multicollinearity Diagnostics Prior to performing logistic regression anal yses, bivariate correlation analyses were performed as well as multicollinearity diagnos tics. This was done both to examine possible relationships in the models, but also to asses if there were any multicollinearity issues that need to be addressed prior to running the regression models. Latin American Region For the entire Latin American region, the result s of bivariate analysis revealed significant relationships between the dependant variable and all of the demographic measures used in this study. When looking at the data from all countri es combined, there is a significant and positive correlation between victimization and sex (male = 1) ( r = .04), race, the level of education ( r = .13), race (White = 1) ( r = .04) and socio-economic attainment (r = .09). Additi onally there are significant negative correlation between the depe ndent variable and the respondents marital status (married = 1) ( r = -.03), age ( r = -.05) and number of children ( r = -.06). The dependent variable is also significantly related to all of the theoretical variables. Results of the analysis show that victimization is significantly and nega tively associated with the level of democracy ( r = -.04), court justice ( r = -.07) and trust of police ( r = -.09). Victimiza tion is positively and significantly associated with a pproval of vigilante justice ( r = .05). Among the variables a number of the measures were significantly correlated with one another. However, none of the bivari ate correlations exceed a value of r .5, so multicollinearity is not an issue. For this reason, additional multicollinearity diagnostics were not performed. 120

PAGE 121

Looking at the bivariate analyses from the i ndividual countries, ther e are fewer significant correlations both between the de pendent and independent variab les and among the independent variables. Below the results of each count rys bivariate analyses are examined. More Legitimate Countries Costa Rica Bivariate analyses for Costa Rica show that victimization is signi ficantly and positively correlated with only le vel of education ( r = .17) and socio-economic attainment ( r = .09). There is a significant negative correlation betw een victimization and marital status (r = -.08) and number of children ( r = -.07). Only one correlation be tween the dependent variable and theoretical variables was present. Victimization was significantly and negatively associated with trust of police ( r = -.11). As with previous models, age and number children were significantly and positively associated ( r = .68). Due to the strong bivari ate correlation between these two independent variables VIF scores were estima ted and showed that no score exceeded two, therefore multicollinearity was not an issue. Uruguay The bivariate analyses for Uruguay show that there is a significan t positive correlation between victimization and the res pondents level of education ( r = .12) and race (.07). The dependent variable only had one significant and negative correlation with a theoretical variable. Bivariate analyses results for Ur uguay reveal that victimization is negatively and significantly associated with trust in police (r =-.12). None of the independent variables in the Uruguay models were correlated at a level of r .5. Consequently VIF scores were not estimated since multicollinearity was not a problem. 121

PAGE 122

Chile The bivariate analyses for the variables in th e Chile model show that there are significant and positive correlations between victimization and an individuals level of education ( r = .16) and socio-economic attainment ( r = .08). Only one theoretical variable shows a significant negative correlation; the depende nt variable is negatively asso ciated with trust in police ( r = .09). None of the independent variables in th is model were correlated with a score of r .5. Therefore, multicollinearity diagnostics were not performed since it is not considered a problem for this model. Countries with Lower Le vels of Legitimacy Bolivia Bivariate analyses for Bolivia show significant positive correlations between the dependent variable and sex (male = 1) ( r = .04), level of education ( r = .11) and socio-economic attainment ( r = .13). The dependent variable was also si gnificantly co rrelated with three of the theoretical variables. The results also show significant negative correlations between victimization and le vel of democracy ( r = -.05), court justice (r = -.06) and trust of police ( r = .08). Similar to previous bivari ate results, when examining correlations between independent variables used in these models Bolivias bivariate analysis showed an individuals age and number of children ( r = .62). However, after VIF scores were estimated it was shown that no scores exceeded a value of three, and multicollinearity was ruled out as a problem for this model. Peru Bivariate analysis of the data from Peru show s that the dependent variable is significantly and positively correlated with sex (male = 1) ( r = .07), level of education ( r = .08) and socioeconomic attainment ( r = .11). The dependent variab le is also significantly correlated with one of the theoretical variables. Bivariate results also show a significant and negative correlation 122

PAGE 123

between the victimization and trust in police ( r = -.07). The results also showed that among the independent variables age and number of child ren was again significa ntly and positively associated at a level of r .5 ( r = .65). Due to this high level of correlation, VIF scores were estimated and revealed that none of the variable s had scores greater than three. Therefore, multicollinearity was excluded as a problem in the Peru model. El Salvador For El Salvador, the dependent variable is significantly and positively correlated with an individuals sex (male = 1) ( r = .07), level of education ( r = .23) and socio-economic attainment ( r = .21). Bivariate analysis also showed the victimization had a significant and negative correlation with the respondents age ( r = -.12) and number of children ( r = -.12). The dependent variable also had a significant association with all of the theo retical variables. There is a significant and negative associ ation between victimization and level of democracy ( r = -.14), court justice ( r = -.06) and trust of police ( r = -.13). Among the independent variables, age and number of children were significantly correlated ( r = .61). Socio-economic status and level of education were also highly correlated ( r = .60) so again VIF scores were calculated to rule out multicollinearity; no score was greater than tw o. For this reason, multicollinearity is not considered an issue for this model. Guatemala In Guatemala, victimization is significantly and positively correlated with sex (male = 1) ( r = .07), level of education ( r = .18) and socio-economic attainment ( r = .18). Among the theoretical variables, there is a significant and negative correlati on between victimization and the level of democracy ( r = -.10) and trust of police ( r = -.10). There is a significant positive correlation between victimization a nd approval of vigilante justice ( r = .08). Among the independent variables, socio-economic attainment and level of education were significantly and 123

PAGE 124

positively associated (r = .59) as were age and number of children ( r = .57). To rule out multicollinearity, VIF scores were calculated for the independent variables in this model, and no scores exceeded two. Therefore it is assumed that multicollinearity is not an issue in this model. Countries of Average Legitimacy Mexico For Mexico, victimization is significantly and positively correlated many of the independent variables, including th eoretical variables. The results of the bivariate analysis show a significant and positive association betw een victimization and sex (male = 1) ( r = .14), level of education ( r = .10), and socio-economic status ( r = .07). There exists a significant and negative association for age ( r = -.08), marital status (married = 1) ( r = -.06) and number of children ( r = .08). The dependent variable is al so significantly associated with two of the theore tical variables used in the analysis. There is a significant a nd negative relationship between victimization and court justice ( r = -.09) and trust in police ( r = -.13). Among the independent variables, age and number of children was positively and significantly correlated (r = .66). To rule out multicollinearity, VIF scores were calculated for the independent variables in this model, and no scores exceeded two. So as with the model for the en tire Latin American region, multicollinearity was not an issue for the Mexico model. Brazil Bivariate analyses on the variables in th e Brazil model reveal a positive significant association between victimization and a respondents level of education ( r = .12) and socioeconomic attainment ( r = .12). The dependent variable is also significantly and negatively correlated with two theoretical variables, court justice ( r = -.11) and tr ust in police ( r = -.06). No correlations between independent variab les in the model exceed a level of r .5. As with the 124

PAGE 125

other models, multicollinearity is not an issue among the variables in this model, therefore no further diagnostics were performed. Venezuela Victimization has a positive and significant correlation with the individuals level of education ( r = .12) in the Venezuela mode l. Victimization is also significantly and negatively associated with a respondents age ( r = -.09) and number of children ( r = -.08). As for the theoretical variables, all are significantly correlated with the dependent variable. Victimization is significantly and negatively associ ated with level of democracy ( r =-.08), court justice ( r = -.14) and trust in police (r = -.13). Victimization has a signi ficant and positive correlation with approval of vigilante justice (r = .10). Among the independent variables in the model age and number of children are correlated at r = .61. Multicollinearity diagnostics reveal that no VIF scores of the variables had scores greater than two, ruling it out as an issue for this model. Dominican Republic For the Dominican Republic, there are si gnificant and positive correlations between victimization and level of education ( r = .09) and socio-econ omic attainment ( r = .06). Victimization is also significantly and negatively associated with age ( r = -.05) and the number of children an individual has ( r = -.05). Bivariate analyses show that two theoretical variables are also significantly corre lated with victimizati on. Both court justice ( r = -.08) and tr ust in police ( r = -.08) have significant and ne gative correlations with the dependent variable. Among the independent variables in the model, only age an d the number of children were correlated at a level that would indicate multicollinearity between the variables ( r = .63). However, when VIF scores are estimated for none exceed two. Thus, mu lticollinearity is not c onsidered a problem in this model. 125

PAGE 126

Jamaica For Jamaica, there is only one signifi cant correlation between the dependent and independent variables. Victimization has a significant and positive association with an individuals socio-ec onomic attainment (r = .07). There are no other significant correlations with the dependent variable. Among the independent variables there are also no significant correlations where r .05. Therefore in the Jamaica model, multicollinearity is not considered a problem. Multivariate Analyses Since multicollinearity is ruled out as an issue for all of the models, binary logistic regression models are able to be run on all independent variables, including theoretical variables. These regression models are run on the combined data for the entire Latin America Region as well as all of the individual Latin American c ountries and test the empirical hypotheses made in Chapter 4 (H8, H9, H10 and H11). Latin America Region Logistic regression results for the entire Latin American region appear in Table 7-1. Presented in the table are the be ta and odds ratio (Odds) for each variable as well as the Chisquare ( ) and the Pseudo R which assess the go odness of fit for the estimated model. The Model Chi-square statistics indicate that the full model is a better fit than the model with only the constant with all the coefficients being held to zero. Results are presented in Table 7-1 ( =651.87). This indicates that the model fits sign ificantly better as a whole model than as an empty model. Table 7 -1 shows that, in the model for the entire Latin American region, four demographic variables are signif icantly predictive of victimization: sex (male =1), level of education, race (White =1), and socio-economic attainment. Being male is predictive of 126

PAGE 127

victimization in this model (b = .17, p .01). Individuals with higher le vels of education are also more prone to victimization (b = .08, p .01). Being White was signi ficantly and positively associated with victimization (b = 0.19, p .01). Finally, individuals w ho have a higher level of socioeconomic attainment have a higher probability of being victimized (b = .05, p .01). When calculated as percent pr obability, the possibility of victimization increases by 19 percent for males. Additionally, each additional year of educ ation increases the likelihood of victimization by nine percent. The probability of victimizati on is 21 percent higher for Whites. Finally, with each incremental increase in soci o-economic attainment, the risk of victimization rises by five percent. Of the theoretical variable s, all were found to be si gnificantly associated with victimization, supporting all of our theore tical hypotheses. As predicted in H8, respondents who report their country to be more demo cratic (on a scale of 1 to 4) are less likely to be victimized (b = -.06). Furthermore, H9 is supported; results indicate that respondents who report a greater trust in the police (on a scale of 1 to 7 where 7 indicates the greatest level of trust) are less likely to be victimized (b = -.10, p .01). Results indicate support for H10 as well; individuals who report a greater level of trust in their nations courts enacting justice (on a scale of 1 to 7 where 7 indicates the greatest level of trust) are less likely to re port victimization (b = -.07, p .01). Lastly, H11 is supported as we find that respondents who approve of vigilante justice (on a scale of 1 to 10, where 10 indicates the highest level of approval for vigilante jus tice) are more likely to be victimized (b = .03, p .01). When expressed as a probabil ity, each increase in perception of democracy decreases the probability of victimi zation by six percent. As trust in court justice increases by one point, the like lihood of victimization decrease s by six percent as well. The likelihood of being victimized decreases by nine percent for each increm ental increase in the 127

PAGE 128

level of trust of the police. Lastly, as approva l of vigilante justice in creases by one point, the probability of victimization increases by three percent. More Legitimate Countries Costa Rica Table 7-2 presents the findings for logistic re gression analyses of the Costa Rica model. It shows that the full mode is significant and a better fit than an empty model ( =73.07). Education and marital status (m arried =1) are significant rela ted to victimization. As an individuals level of education increases, the risk of victimi zation increases as well (b = .16, p .01). Married individuals in Costa Rica also have a lower probability of vi ctimization (b = -.33, p .05). This indicates that in Costa Rica, as education increases by one year the odds of victimization increases by 13 percent. Being ma rried decreases the odds of victimization by 28 percent. Trust of police ( on a scale of 1 to 7) was th e only significant theoretical variable in the Costa Rica model1. As predicted in H10, as a respondents trust of po lice increases, victimization decreases (b = -.19, p .01). For each incremental increase in police trust, the likelihood of victimization decreases by 17 percent. Uruguay The results for the Uruguay model are presente d in Table 7-2. This model was significant and a better fit than an empty model ( =42.86). Results from the logistic regression analysis show that among the demographic variables le vel of education and race (White =1) are associated with the risk of victimization. An increase in an individuals level of education, increases risk of victimization (b = .12, p .01). Additionally, being White is positively and 1 Vigilante justice was approaching significance in this model (b = .04, p = .07). The odds ratio was 1.04. 128

PAGE 129

significantly associated with victimization (b = .45, p .05). Converting the odds ratios shows us that for every one unit increase in level of education, the probabil ity of victimization rises by 13 percent. In addition, White individuals in Urugua y have 58 percent greater probability of being victimized. Only H10 was supported in this model showing that trust in police was si gnificantly related to victimization. As trust in police incr eases, victimization decreases (b = -.18, p .01). This means that as trust in police increases by one unit (on a scale of 1 to 7), the probability of victimization decreases by 17 percent. This finding supports the pr ediction made in H10. Chile Results for the Chilean model are shown in Ta ble 7-2. The Model Chi-S quare shows that the full model holds greater predictive power than an empty model ( =58.84). Logistic regression analysis reveals that education and the numbe r of children are significantly related to victimization. As the level of education increases, the risk of victimization increases (b = .18, p .01). And the number of children an individual has increases the odds of victimization (b = .10, p .05). In terms of percent probability, the resu lts show that for each increase in years of education, the probability of victimization incr eases by 20 percent. For each additional child a respondent has the risk of victimi zation also increases 10 percent. In the Chilean model, the only statistically sign ificant theoretical variable is court justice. As predicted in H9, respondents who report higher levels of trust in court justice (on a scale of 1 to 7) have a lower likelihood of victimization (b = -.13, p .01). This means that for each one unit increase in trust in court justice the proba bility of victimization decreases 12 percent. 129

PAGE 130

Countries with Lower Levels of Legitimacy Bolivia Logistic regression analyses for Bolivia appear in Table 7-3 and show that the full model is a better fit than an empty model ( =80.79). Of the demographic indicators only socio-economic attainment is significantly related to victimization.2 As socio-economic attainment increases (on a scale of 1 to 11), the risk of victimization increases (b = .11, p .01). This means that for each one unit increase in socio-economic attainment the probability of victimization increases by 12 percent. In the Bolivian model two of the theoretical variables are significantly associated with victimization. In support of H8, results showed that individuals who reported higher levels of democracy (on a scale of 1 to 4) had lowe r levels of victimization (b = -.15, p .05). In addition, H10 is supported showing that as trust of police increas es, victimization also decreases (b = -.10, p .01). This means that for every one unit increase in the perception of democracy, the risk of victimization decreases by 14 perc ent. As trust of police increases by one unit, the risk of victimization decreases as well by nine percent. Peru Logistic regression analyses of data from Peru appear in Table 7-3. The full model is a better fit than an empty model ( =42.02). Sex (male =1) and soci o-economic attainment are both statistically associat ed by victimization. Being male increa ses victimization in Peru (b = .31, p .05). As socio-economic attainment increases so does the likelihood of victimization (b = .11, p .01). Males in Peru have a 36 percent greater likelihood of being victimized than females. 2 Both sex (male = 1) (b = .20, p = .08) and level of education (b = .04, p = .07) are approaching significance with respective odds ratios of 1.22 and 1.04. 130

PAGE 131

Additionally, as socio-economic attainment increases by one un it, the probability of being victimized also increases by 11 percent. Results from the Peru model show s upport for two of the proposed hypotheses, H10 and H11. As trust in police increas es, the likelihood of victimi zation decreases (b = -.12, p .01). As the approval of vigilant e justice increases (on a scale of 1 to 10), the likelihood of victimization increases as well (b = .04, p .05). When expressed as percent pr obabilities, this shows that for every one unit increase in trust of police the likelihood of victimization decr eases by 13 percent. Furthermore, individuals who report a greater level of approval for vigilante justice in Peru have a four percent increase in the pr obability of being victimized for every one unit increase in approval. Guatemala The Model Chi-Square statistics for Guatemal a (Table 7-3) show that the full model is more appropriate than an empty model ( =68.68). Logistic regression analysis also shows that sex (male =1) and education are signifi cantly associated with victimization3. Males in Guatemala are at a greater risk of victim ization than females (b = .32, p .05). As the level of education for an individual increases, the likelihood of victimization also increases (b = .12, p .01). Expressed as probabilities, males in Guatemala have a 37 percent greater probability of being victimized. For each one year increase in the leve l of an individuals education, the probability of victimization increases by 12 percent. Among the theoretical variables, level of demo cracy, trust in police and vigilante justice are all significantly related to vic timization. Per the prediction in H8, respondents who reported 3 Age (b = -.01, p =.054) and Socio-economic attainment are approaching significance (b = .06, p = .08) as is the number of children (b = .07, p = .08). The odds ratio for age is 0.986 while the odds ratio for socio-economic attainment is 1.06; the odds ratio for children is 1.08. 131

PAGE 132

higher levels of democracy (on a scale of 1 to 4) increases had lower levels of victimization (b = -.32, p .01). Additionally, H10 was supported; as the reported le vel of trust in police increases (on scale of 1 to 7), the risk of vi ctimization also decreases (b = -.11, p .05). Lastly, in this model there is support for H11, showing that as the level of appr oval of vigilante justice increase (on a scale of 1 to 10), the risk of victimization increases (b = .05, p .05). Interpreted as probabilities, the model reveals th at individuals who report a great er perception of democracy in Guatemala have a 27 percent lower probability of being victimized for each one unit increase in democracy. Furthermore, individuals who report a greater trust in the po lice have a 10 percent lower probability of being victimized for each one unit increase in reported trust. Finally, as the approval of vigilante justice incr eases, the probability of victim ization also increases by five percent, for each incremental increase in approval. El Salvador Model diagnostics for the data from El Salvador show that the full model is significant and a better fit than an empty model ( =137.02). Results from logistic regression analyses (Table 73) show that education, age and socio-economic attainment are all significantly related to victimization in El Salvador. Individuals with hi gher levels of education are more likely to be victimized (b = .10, p .01). As an individuals age increas e, the likelihood of victimization decreases (b = -.02, p .05). Lastly as socio-economic attain ment increases (on a scale of 1 to 11), the risk of victimization increases (b = .13, p .01). The percent prob abilities of these findings demonstrate that for ev ery one unit increase in an indi viduals education the risk of victimization increases by 13 percent. For every one year increase in age there is a two percent decrease in the probability of victimization. Lastly, for every on e unit increase in socio-economic attainment the possibility of vic timization increases by 14 percent. 132

PAGE 133

Both democracy and trust of police are signif icantly associated with victimization in Guatemala.4 In keeping with the prediction from H8, respondents who repor t higher levels of democracy have a lower likelihood of victimization (b = -.29, p .01). There is also support for H10 in the Guatemalan model which shows that as the respondents trust in police increases the likelihood of victimization decreases (b = -.10, p .05). When expressed as percent probabilities, this shows that for every incremental increa se in the perceived level of democracy, the probability of victimization decreases by 25 percent. Additionally, for every one unit increase in trust of the police, the pr obability of victimization increases by nine percent. Countries of Average Legitimacy Mexico Results from the logistic regr ession analysis of data from Mexico are presented in Table 7-4. These results show that the full model is a better fit than the model with only the constant with all the coefficients being held to zero ( = 70.70). Sex was the only demographic variable predictive of victimization.5 Males are more likely to be victimized that females in Mexico (b = .65, p .01). In this model, the likelihood of victimi zation for males is 92 percent higher than it is for females. Two theoretical variables are significant in th e Mexico model, both court justice (1 to 7) and trust in police (1 to 7). As the respondents trust in court justice increases, the risk of victimization decreases, which supports H9 (b = -.09, p .05). Additionally, as the respondents trust in police increases, the ri sk of victimization decreases which also shows support in the Mexico model for H10 (b = -.15, p .01). Expressed in probabilities, this means that for each one 4 Vigilante justice is approaching significance in this model (b = .04, p = .08). The odds ratio is 1.05. 5 Age is approaching significance in the Mexico model (b = -.01, p = .06). The odds ratio for age is 0.99. 133

PAGE 134

unit increase in court justice, th e probability of victimization d ecreases by eight percent. For each one unit increase in an individuals trust of polic e, the probability of victimization decreases by 14 percent. Brazil Results from the Brazil model (Table 7-4) reveal that the full model is significant and a better fit than an empty model ( =40.77). Logistic regression anal ysis reveals that education and socio-economic attainment are statistically related to victimization in Braz il. Level of education was significantly as positively associated with vi ctimization showing that as education increases the likelihood of victimi zation increases (b = .07, p .05). Additionally, as the level of socioeconomic attainment increases, the risk of victimization increases (b = .11, p .05). For every one unit increase in both level of education and socio-economic attainment the likelihood of victimization also increases by seven percent and 11 percent respectively. Of the four theoretical hypotheses proposed, only H9 is supported by the findings from the Brazilian model. As a respondents trust in court justice increases (where 7 indicates the greatest amount of trust), the likelihood of vi ctimization decreases (b = -.14, p .01). For every one unit increase in court justice, th e probability of an individual being victimized decreases by 13 percent. Venezuela Table 7-4 presents the results for Venezuela. This model is statistically significant and a better fit than the empty model ( =146.96). Regression results reveal that education is the only demographic variable significantly associated with victimization6. As the level education increases so does the risk of victimization (b = .08, p .01). The conversion of this finding into a 6 Age is approaching significance in the Venezuela model (b = -.01, p = .07). The odds ratio is 0.99. 134

PAGE 135

percent probability shows that for every one year increase in education, the probability of victimization increases by nine percent. In the Venezuela model, three of the theoretica l variables are statistically significant; court justice, trust in police and vigilante justice. C ourt justice is inversely related to the risk of victimization (b = -.03, p .01). Trust in police is also invers ely related to victimization (b = .13, p .01). Vigilante justice is positively and si gnificantly associated with the risk of victimization in Mexico (b = .06, p .01). From the Venezuela model, results support H9 with, every one unit increase in an individuals perceived level of democracy decreasing the probability of victimization by 12 percent. Furthermore, H10 is supporting showing that every one unit increase in an indivi duals reported trust in police decreases the probability of victimization by 12 pe rcent. Lastly, H11 is supported; as approval of vigilante justice increases by one unit, the likelihood of victimi zation increases by six percent. Dominican Republic The model results for the Dominican Republic are in Table 7-4. Model diagnostics show that the full model is a more appropriate than the model with only the constant ( =41.66). Logistic regression analysis reveal s that level of education is th e only demographic variable that is significantly related to victimization. An increase in the level of an individuals education increases the risk of victimization (b = .06, p .01). For each yearly increase in education the probability of victimization increases by six percent. Of the theoretical variables, court justice is the only statistically significant variable in the Dominican Republic model7. In keeping with H9, as a respondents trust in the countrys court justice increases (on a scale of 1 to 7), the risk of victim ization decreases (b = -.08, p .05). For 7 Trust in police is approaching significance in the Domini can Republic model (b = -.06, p = .06). The odds ratio for trust in police is 0.94. 135

PAGE 136

each one unit increase in the individuals trust that Dominican courts ensure justice, the likelihood of victimization d ecreases by eight percent. Jamaica The final regression model is for Jamaica (Table 7-4) and model diagnostics show that the full model is significant and a bett er fit than an empty model ( =22.02). Age, number of children and socio-economic attainment are all si gnificantly associated with victimization. Older individuals in Jamaica are at a great er risk of victimization (b = .02, p .01). The more children an individual has also decreases the risk of victimization in Jamaica (b = -.13, p .05) Lastly, an increase in socio-economic attain ment is related to a greater ri sk of victimization (b = .09, p .05). For every one year increase in age, the risk of victimization increases by two percent. For each additional child that an individual has, th e risk of victimization decreases by 22 percent. Finally, for every one unit increase in soci o-economic attainment the probability of victimization increases by 10 percent. None of the theoretical hypotheses proposed in this dissertation were supported with the findings from the Jamaica model. 136

PAGE 137

Table 7. Logistic Regression Analysis base d on Victimization for all Latin American Respondents Variables Latin America Total Sample (N=19,193) Beta Odds Ratio Independent Variables Sex (male =1) Level of Education (years) 0.17 0.08 1.19** 1.09** Marital Status (married =1) Age (years) Race (White =1) Children (range 0-20) Socio-Economic Attainment (0-11) -0.03 -0.00 0.19 -0.00 0.52 0.97 1.00 1.21** 1.00 1.05** Theoretical Variables Level of Democracy (1-4) Court Justice (1-7) Trust of Police (1-7) Vigilante Justice (1-10) Constant -0.06 -0.07 -0.10 0.03 -1.56 0.93** 0.94** 0.91** 1.03** 0.21 Chi-Square R-Square 651.87** 0.05 p .05, ** p .01 Source: AmericasBarometer 2006 137

PAGE 138

138 Table 7-2. Logistic Regression Analyses base d on Victimization for Costa Rica, Uruguay and Chile Costa Rica Uruguay Chile N= 1,430 N= 1,086 N = 1,443 Variable Beta Odds Ratio Beta Odds Ratio Beta Odds Ratio Independent Variables Sex (Male = 1) 0.12 1.13 0.22 1.25 0.01 1.01 Level of Education (years) 0.15 1.16** 0.12 1.13** 0.18 1.20** Marital Status (married =1) -0.33 0.73** -0.04 0.96 0.05 1.05 Age (years) 0.00 1.00 0.01 1.01 -0.00 1.00 Race (White =1) 0.13 1.13 0.46 1.58* 0.01 1.01 Children (range 0-20) 0.01 1.00 0.03 1.03 0.09 1.10* Socio-Economic Attainment (1-11) 0.04 1.04 -0.06 0.94 -0.01 0.99 Theoretical Variables Level of Democracy (1-4) 0.06 1.07 -0.00 1.00 -0.11 0.90 Court Justice (1-7) 0.01 1.01 0.04 1.05 -0.01 0.88** Trust of Police (1-7) -0.19 0.83** -0.18 0.83** 0.01 1.01 Vigilante Justice (1-10) 0.04 1.04 0.02 1.02 0.03 1.03 Constant -2.41 0.09 -1.76 0.17 -1.66 0.19 Chi-Square 73.07** 42.86** 58.84** R-Square 0.08 0.06 0.06 p .05, ** p .01: Source: AmericasBarometer 2006

PAGE 139

Table 7. Logistic Regression Analysis based on Victimization for Bolivia, Peru, Guatemala and El Salvador Bolivia Peru Guatemala El Salvador N= 2,562 N= 1,370 N = 1,111 N= 1,607 Variable Beta Odds Ratio Beta Odds Ratio Beta Odds Ratio Beta Odds Ratio Independent Variables Sex (Male = 1) 0.20 1.22 0.31 1.36* 0.32 1.37* 0.24 1.28 Level of Education (years) 0.04 1.04 0.03 1.03 0.12 1.12** 0.10 1.10** Marital Status (married =1) -0.04 0.96 -0.02 0.98 0.12 1.13 0.04 1.04 Age (years) 0.00 1.00 -0.01 0.99 -0.14 0.99* -0.02 0.98** Race (White =1) -0.20 0.82 -0.13 0.88 0.03 1.03 0.01 1.01 Children (range 0-20) -0.05 0.95 0.03 1.03 0.07 1.08 0.01 1.01 Socio-Economic Attainment (1-11) 0.11 1.12** 0.11 1.11** 0.06 1.06 0.13 1.14** Theoretical Variables Level of Democracy (1-4) -0.15 0.86* 0.11 1.12 -0.32 0.73* -0.29 0.75** Court Justice (1-7) -0.06 0.94 0.02 1.02 0.06 1.05* 0.01 1.01 Trust of Police (1-7) -0.10 0.91** -0.12 0.89 -0.11 0.90* -0.10 0.91* Vigilante Justice (1-10) 0.03 1.03 0.04 1.04* 0.05 1.05* 0.04 1.04 Constant -1.58 0.21 -1.76 0.17 -1.31 0.27 -1.48 0.23 Chi-Squared 80.79** 42.02** 68.68** 137.02** R-Squared 0.05 0.04 0.09 0.14 p .05, ** p .01 Source: AmericasBarometer 2006 139

PAGE 140

Table 7-4. Logistic Regression Analyses on Victimization for Mexico, Brazil, Venezu ela, the Dominican Republic, and Jamaica Dominican Mexico Brazil Venezuela Republic Jamaica N= 1,389 N= 1,080 N = 1,111 N= 1,607 N= 1,05 5 Variable Beta Odds Ratio Beta Odds Ratio Beta Odds Ratio Beta Odds Ratio Beta Odds Ratio Independent Variables Sex (Male = 1) 0.65 1.92** 0.22 1.25 0.03 1.03 -0.08 0.93 0.36 1.43 Level of Education (years) 0.05 1.05 0.07 1.06* 0.08 1.09** 0.06 1.06** 0.01 1.01 Marital Status (married =1) -0.10 0.91 -0.17 0.85 -0.02 0.98 -0.05 0.95 0.15 1.17 Age (years) -0.01 0.99 0.00 1.00 -0.01 0.99 0.00 1.00 0.02 1.02** Race (White =1) 0.07 1.08 0.13 1.14 0.02 1.02 -0.18 0.84 -18.96 0.00 Children (range 0-20) -0.00 1.00 -0.01 0.99 -0.01 0.99 -0.02 0.98 -0.13 0.88* Socio-Economic Attainment (1-11) 0.01 1.01 0.11 1.11* -0.03 0.97 0.02 1.02 0.09 1.10* Theoretical Variables Level of Democracy (1-4) 0.01 1.01 0.05 1.05 -0.03 0.97 -0.01 0.99 0.08 1.08 Court Justice (1-7) -0.09 0.92* -0.14 0.87** -0.13 0.88** -0.08 0.92* -0.06 0.94 Trust of Police (1-7) -0.15 0.86** -0.06 0.95 -0.13 0.88** -0.06 0.94 -0.06 0.94 Vigilante Justice (1-10) 0.04 1.04 0.05 1.05 0.06 1.06** 0.02 1.01 0.04 1.04 Constant -0.83 0.44 -2.33 0.10 -0.32 0.72 -1.21 0.30 -3.65 0.03 Chi-Squared 70.70** 40.77** 149.96** 41.66** 22.03* R-Squared 0.08 0.06 0.08 0.03 0.05 p .05, ** p .01: Source: AmericasBarometer 2006 140

PAGE 141

CHAPTER 8 DISCUSSIONS AND CONCLUSIONS This study responds to the urging of LaFree (2 007) and others (Marsh, 1967; Birkbeck, 1993; Willis et al, 1999) to test theory outside of the U.S. or European context and to acknowledge that theory grounded in the Euro-American culture might apply differently in other cultures. This dissertation project examined the ro le of legitimacy of law in 12 Latin American countries and its effect on i ndividual victimization. A second the purpose of this dissertation project, was to test the theory not only in 12 countries but also using combined data from the entire region to determine if fu ture studies such as this should disentangle data into country specific data in order to test for subtle or ove rt differences among countri es in the region. Results from this study not only show support for the role of legitimacy of law in victimization, but also show that the results from individual countries vary, giving credence to the idea that context matters when studying victimization in Latin America. Latin American Region vs. Country Specific Models One of the interesting findings of this study is that in anal yses containing only demographic variables and includ ing theoretical variables, the model for the entire Latin American region was different from the country specific models. In the model expressed in Table 6-1, we find that sex, level of education, race and socio-economic attainment are all significantly associated with victimization. This supports H1, H2, H5 and H7 made in Chapter 3.1 However, in sub-sequent country specific models there are different st ories being told about victimization. 1 Note that this relationship differs from what victim ization research from the United States tells us about victimization. In the United States, individuals with lower socio-economic status, education and non-Whites are more likely to be victimized. 141

PAGE 142

Marital Status For example, while we do not see support for H3 in the full model, marital-status (married = 1) was not predictive of victimization in the model containing responses from all individuals in the Latin American region, when we look at the model for Costa Rica, being married is significantly and negatively associ ated with victimization. Married individuals in Costa Rica are at a lower risk of victimization than are si ngle individuals. This finding supports previous research from the region, which found that unma rried individuals were at a greater risk of victimization (Catalano, 2006; Smith & Jarjoura, 1989). This might be due to, in part, the guardianship that married individuals have over one another that might make them less susceptible to violent victimizati on. It might also be that they ar e at a lowered risk of property victimization in Costa Rica because there are at least two individuals to guard personal property. This nuance is missed when combining the samples from all countries as some previous studies have done (Fajnzylber, Lederman, & Loayza, 1 998; Gaviria & Pags, 1999; Neapolitan, 1994). Age Surprisingly, we find that in the full model, age is not significantly associated with victimization. This is contrary to previous findings on victim ization in Latin America (Buvini & Morrison, 1999). When we look at other trends in the full-regi onal model, however, the reason for age not being significantly a ssociated with victimization mi ght be that it is so strongly associated with socio-economic stat us in the region, rather than ag e. There might be high rates of victimization among older, more financially su ccessful individuals wh ich is affecting the findings for age. This is in keeping with previ ous research from the re gion (Gaviria & Pags, 1999; Gaviria & Velez, 2001). ). Additionally, the effect of age on victim ization is significant and negative in Guatemala, El Salvador a nd Venezuela, which is in keeping with H4 made in Chapter 2. However, it is signif icant and positive in Jamaica, which is the opposite direction 142

PAGE 143

from what was proposed in the hypothesis. The effect of combining th e responses from all countries might be cancelling out the importance of age in these three countries. If researchers were to only look at the results from the Latin American sample as a whole, they would be missing the effect that age has in specific countrie s in the region. In a few of the subsequent country specific mode ls we see that age is indeed significantly associated with victimization. For example, in Guatemala and El Salvador older individuals are at a lower risk of victimization; we see this in Venezuela as well. Conversely, in Jamaica, older individuals have a higher risk of victimization. These nuanced differences in the findings between the specific countries are missed in models that contain responses fr om all individuals in the Latin American region. Knowing that younger in dividuals in Guatemala and El Salvador are at a greater risk of victimization not only he lps researchers in unders tanding the trends of victimization within these countries, but also may speak to previous resear ch that explains how historical events, such as the 30 year civil war in Guatemala and th e 12 year civil in El Salvador, have disproportionately aff ected the young male population in the country leaving them vulnerable to economic strain and physical and property victimiza tion (Kay, 2001; Kay, 2007). The similar legacies of these two countries also help explain how there are similar victimization trends. Both education and age were significantly associated w ith victimization in these two countries, and while sex was associated with vict imization in Guatemala a nd not El Salvador and socio-economic attainment was associated with vi ctimization in El Salvador and not Guatemala, the two countries still show similarities in the si gnificant variables in th e demographic models as well as similarities in the theore tical models which will be discussed more in a latter section of this chapter. 143

PAGE 144

Children Another variable that is not significant in th e model containing all respondents from all Latin American countries in this study is the num ber of children an indivi dual has. While there is no support for H6 in the Latin American Region model, there is support for the hypothesis in both the Chilean and Jamaican models. However, th e effects that the number of children has on the likelihood of victimization are different in each of these countri es. In keeping with H6 proposed in Chapter 3, the greater number of ch ildren that someone has the greater the odds of victimization in Chile, where each additional child an individual has increases the probability of victimization by 10 percent. This finding might be a further indication of th e effect of wealth or income acting as a buffer to victimization; more a ffluent individuals can a fford security features for themselves, their home, their cars, and so on. It may be that lower income families in Chile have greater numbers of children, as is the case in Peru (Pat rinos & Psacharopoulos, 1997) as well as in the United States (see for example, Becker, Duesenberry & Okum, 1960). We know from previous research that an increase in the nu mber of children in the home adds to the levels of familial violence as previous research has found (Tennyson, 2004; Nikolic-Ristanovic, 2002). The direction of the relations hip between the number of ch ildren an individual has and victimization is opposite of that proposed in H6 in the Jamaican model. Surprisingly, the number of children has the opposite effect in Jamaica. For each additional child in the home, the probability of victimization decreases by 22 percen t. Jamaican families might be an example of how additional family members act as added guard ianship from various forms of victimization, even if those family members are children. Lastly as with age, the effects of children in the home might not be showing up in the full re gional model because the opposing effects of children on victimization in bot h Chile and Jamaica may cancel each other out in the full regional model. While age and the number of child ren were not significant in the Latin American 144

PAGE 145

Region model, there were several variables that were significant in that model as well as country specific models. Sex We find support for H1 in the full regional model and models in Peru, Guatemala and Mexico where victimization is higher am ong male respondents compared with female respondents. Sex (male = 1) was significantly and positively associated with victimization in the Latin American Region model. Males in this model were at a 19 percent greater risk of victimization than corresponding female respondents. This finding is being driven by the effect of sex on victimization in Peru, Guatemala, a nd Mexico. Male respondents in both Peru and Guatemala were more than 30 percent more likely to be victimized than were females. In Mexico, males were over 90 percent more likely to be victimized.2 While we know from previous research that males are more likely to be victimized in the region (Gaviria & Pags, 1999), it is likely that this study, as with pr evious studies, is mi ssing domestic violence victimization against women. This is possibly due to the way the vic timization question was asked. Women who are victims of domestic assault may frame the c oncept differently than they do violent assault from a stranger. It could also have to do with how domestic violence is legally conceptualized. In some Latin American countries, it is not culturally considered a criminal act. In others, certain forms of domestic abuse, such as sexual violence or coercion, are not legally considered criminal (UNIFEM, 2005). Also, ma ny countries did not pass legislation against domestic violence until the 1990s and there mi ght be reluctance culturally and by legal authorities, to accept abuse in the home as a cr ime (UNIFEM, 2005). It is clear that future 2 It should be noted that the sample from Mexico is highly skewed. Only 21 percent of the respondents were male while the vast majority was female. The original sample was 50 percent male, however, after list-wise deletion of missing cases for the variables used in the model, the sample was only 21 percent male. This is a huge limitation of the Mexico sample and severely limits the applicability of the findings. However, even given the skewed data, males are at a higher risk of victimization in Mexico. 145

PAGE 146

research should specifically examine domestic violence and survey questions should be tailored to the unique nature of this crime. Level of Education When we look at the effects of level of education on victimi zation risk in Latin America we find support for H2 in that education is significantly associated with victimization risk in the entire region. We also find support for H2 in Costa Rica, Uruguay, Chile, Guatemala, El Salvador, Brazil, Venezuela, and the Dominican Republic. The exceptions to this trend are Jamaica, Mexico, Peru and Bolivia. Since there is no test specifically for the effect of income on victimization in this study, an individuals level of education may be a proxy for income and this might be what is driving th e higher probabilities of vic timization among more educated individuals across the region and within the specific countries. It is surprising to see that education is significantly associ ated with victimization in Co sta Rica, Uruguay and Chile. These countries not only have lower levels of income disparity than othe r countries in th e region but the income gap is also narrowing (Psacharopoulos Morley, Fiszbein, L ee & Wood, 1997; Lustig, 1995). However, the level of education among the vi ctimized sub-sample is greater than the full samples for those three countries (See Table 5-1 and Table 5-2). Th is might be an indication that respondents with higher levels of education are merely more likely to report victimization, and not an actual risk factor for physical or prope rty victimization. Howeve r, education is also significantly and positively associated with vict imization in other countries that have higher levels of corruption and lower leve ls of stability. It is also significant in countries with higher levels of income disparity (Psacharopoulos et al., 1997; Lustig, 1995) which shows that individuals who have higher levels of educati on are being targeted, pe rhaps for their economic attainment or for their perceived eco nomic attainment, throughout the region. 146

PAGE 147

Socio-economic Attainment The idea that the significance of education is being driven by or is a proxy for income is further supported by the associa tion found between socio-economic attainment and victimization in both the full model containing responses from all Latin American participants and various country specific models. The support for H7 is very strong not only in the full model but in specific country models. Socio-economic attainment is significantly and positively associated with victimization, as predicted in Chapter 3, in both the full model and in Bolivia, Peru, El Salvador, Brazil and Jamaica. This is an increas e of between 10 and 15 percent in the probability of victimization for each additional consumer good that individuals in these countries possess. In the full model, the effect of socio-economic attainment on victimization is less pronounced. There is only a five percent increase in the probability of victimization for each additional consumer good that an individual possess. Howe ver, there is still a significant and positive relationship between socio-economic attainment a nd victimization. The presence of items such as a personal computer or cellular phone might be a measure of how attractive an individual is as a target for property victimization. One of the limitations of using this index is that we are not getting at which consumer good makes someone a more attractive victim, which could be an area of future research. Clearly running water in the home would not be an incentive to break into someones home, but the presence of a computer, on the other hand, might be. The presence of these items may also be a pr oxy for income, which is how the variable was intended, and therefore we are c onfirming previous research findi ngs that show that wealthier individuals are more prone to victimization in Latin America (Gaviria & Pags, 1999; Gaviria & Velez, 2001; Heinemann & Verner, 2006). However, there is nothing to indi cate that the absence of either a home or cellular phone might make someone more attractive as a victim of either property or violent victimizati on. The absence of these items might be a lack of adequate 147

PAGE 148

guardianship, which could increase the probability of both property and physical victimization to the individual. Future research again should address this question. Race Socio-economic attainment is significant in the direction that was predicted in Chapter 3 as is race. Race (White = 1) is significant and positive in the model for the full Latin American region (Table 6-1). This model shows that whites are 16 percent more likely to be victimized than non-White respondents. The hypothesis (H5) was proposed in this direction because the assumption was that White individuals might be of a higher income or socio-economic status than non-White respondents which would make them more attractive victims. This might not be the result expected for all count ries in the region, given what we know about rural violence among indigenous populations in countries such as El Salvador and Guatemala (Kay, 2007). However, taking a closer look at the country models allows for more clarification. What we see in the full-model is being driv en solely by the results from Uruguay, the only country where race is significantly associated with victimizat ion in any direction. In Uruguay, White individuals are 53 percent more likely to report victimization. There might be several reasons for this finding. First, it might be that White individual s in Uruguay are simply more likely to report being victimized. It also might be that they feel more outraged or have more lingering feelings regarding their victimization, which would also make them more likely to report the victimization on the LAPOP survey. While this is all merely conjecture, it also could be simply that White individuals are targeted mo re frequently in Uruguay. This might be due to a higher socio-economic status or income level, or it might be a reflecti on of racial animosity, particularly between White Uruguayans and indige nous groups in the country. In recent years there has been a rapid increase in the indige nous population in Uruguay and non-White citizens in Uruguay have significantly lowe r levels of social development than White citizens (Cabella 148

PAGE 149

& Porzecanski, 2007). Another factor in race (White =1) being a significantly associated with victimization in Uruguay is that the vast majo rity of Uruguayans in this sample identify themselves as White (80 percent). This is not a skewed sample, but rath er a reflection of the actual Uruguayan population, which is predomin ately White. The unique population distribution of Uruguay might be driving the results seen in Table 6-2 and in the full regional model seen in Table 6-1. The results for race in both the full and Uruguayan model really highlight the necessity for researchers to look at individual countries rather than assuming that the results from a full regional model are representative of vi ctimization trends across the region. The issues driving the racial findings in Ur uguay apply only in that country. Discussion Overall when we look at the countries within the region, we do see some patterns that hold across most of the countries in the region, but none are univers al. Age and marital status are not significant in the full Latin American region model. However, they are significant in country specific models. In Costa Rica marital status is significantly and negativ ely associated with victimization. Age has conflicting effects on vict imization. In Guatemala, El Salvador and Venezuela younger individuals have a greater ri sk of victimization while in Jamaica older individuals are more likely to re port being victimized within the previous year. While research that groups nations in Latin Am erica together when performing quantitative research might be more convenient or might create a high degree of significance for the independent variables in the model, it simply misses key differences that are taking place in each country in the region. Nothing highlights this more than the significa nce of race in the full model, which is being driven entirely by what is happening in Urugua y. To look at the full model and assume that Whites across the entire region are being targeted for victimization might lead to gross policy errors, unnecessary racial tens ions, or false assumptions ab out the nature of race and 149

PAGE 150

victimization throughout the region. Instead of taking a step back to look at the region as a whole, researchers really need to take a closer look at individual nations within Latin America in order to appreciate the vast and significant, both statistically a nd contextually, differences among its many nations. Further evidence for the need to examine victimization within each specific country rather than as a whole can be seen in the R-square value for the non-theoretical logistic regression model for the full region (See Table 6-1). The R-Square va lue for this model is rather low, 0.03. In contrast, some of the country specif ic models have higher r-square values (Costa Rica, Mexico and El Salvador specifically). Th e demographic variables used in each model are coming across differently in the individual countr y analyses and are countering each other when used in the full regional model. Th is indicates that something is going on in each specific country model, even when only examining demographic predictors of individual victimization. While many nations in the region have a sh ared Colonial history, the paths that each country has taken since that distant time have b een markedly different. The level of development that each country has reached is also very different, socially, de mocratically, and culturally as well. These divergent paths have effects at the individual level on citizens of each country in the region. For this reason, we now turn to issues of legitimacy of law in the region to help explain what contextually is driving our findings regarding victimization in the countries of Latin America. Legitimacy of Law Results from this study show that there is varying support for the theoretical hypotheses proposed in Chapter 3 (H8, H9, H10 and H11). The impact of indicators of legitimacy of law on the likelihood of individual vic timization in the 12 Latin American countries varies from country to country (Tables 7-1 thru 74). When looki ng at the region as a whole, all aspects of legitimacy of law, including leve l of democracy, court justice, trust of police and support for 150

PAGE 151

vigilante justice3, were significantly associated with vi ctimization. This sh ows that, across the region, using legitimacy of law theo ry to understand risk of victim ization is appropriate. These findings address the ongoing deba te regarding the use of Nort h American criminological and sociological theories to explai n crime and victimization in othe r regions of the world. We see that, similarly to findings in the United States by Tyler and others, that the legitimacy of legal authorities does relate to crime, and in the case of this study, personal victimization. When there are low levels of trust of police, the legitimacy of the court system, and even the overall level of democratic governance across Latin America, there are increases in victimization. Furthermore, support for non-legal or extra-legal efforts to c ontrol crime through vigila nte justice increase the likelihood of victimization across the region. This is in keepi ng with previous research on legitimacy of law done in the United States (See Tyler, 1990; Fagan & Tyler, 2004; Kane, 2005; Fagan, 2008). It is also in keeping with resear ch done previously in Latin America (Seligson, 1980; Davis Rodrigues, 2006). There are several reasons we find support for le gitimacy of law relating to the likelihood of victimization in the entire Latin America region (Tab le 7-1). First, it is possi ble that the effect of combining all data from the 12 countries creates a dataset of sufficient size to tease out the results found in this model. However, elements of legitimacy of law are found to be significant in nearly all of the individual c ountries, which show that the legi timacy of legal authorities does indeed bear some relationship to victimization. What this indicates is that low levels of legitimacy of forms of law enforcement and legal processes are associated with an increase in the risk of victimization to individuals within the region. When the legal authorities fail to provide security to citizens within these countries, 3 Level of democracy, court justice and trust of police are all coded from low to high on legitimacy. Vigilante justice is coded from high to low on legitimacy. 151

PAGE 152

there is an increase in both viol ent and property victimization that, according to the results of this study, is felt most keenly among upper-class and educated members of society. This seems counter-intuitive to not only what we know about victimization in the United States, but also to what we know about how law has historically operated in the region. These are the same classes of people who, in the context of the United States are afforded the greatest levels of security. However, in this region of the world where there is constant flux in the way that law operates, there is no assurance of such secu rity. It seems that the legal aut horities in the region have failed to the point that the upper classes, who historically have had a gr eat deal of legal protection and even preferential treatment, are no longer protected either If a researcher we re only to look at this full regional model including responses from all individuals within the region, again, key contextual differences would be missed. As with the non-theoretical models where there were conflicting results between the full model and co untry specific models, we find that not all indicators of legitimacy of law are signi ficant in each country specific model. Level of Democracy Hypothesis 8 is supported in th ree of the twelve countries us ed in this study. Individuals who report higher levels of democr acy are less likely to be victimized in Bolivia, Guatemala and El Salvador. Conversely then, individuals who report lower le vels of democracy in these countries are more likely to report being victimize d. These are the only three countries out of the 12 where level of democracy is significantly associ ated with victimization which indicates that it is only the responses from these three countries that are driving the results found in the full regional model. As proposed earlie r in this dissertation, it is believed that the individuals who report higher levels of democracy are more likely to be politically insulated from the violence in their country or community or are not victims of political unrest. Keeping with the concept of capable guardianship from the routine activities perspective, individuals who feel that their 152

PAGE 153

country is operating on a highe r level of democratic governa nce are experiencing a greater degree of political protection from their government, as good democratic regimes provide for their citizens. It is the second aspect of democr acy that was outlined in Chapter 2, pertaining to a governments ability to protect the social rights of its citizens that we are discussing here. It may be that those individuals who are reporting higher levels of democracy are reporting not only electoral democracy but rather fu ll democratic citizenship and right s. As we discussed in Chapter 2, it is often this second aspect of democracy that is missi ng from transitioning or newly democratic nations. These countri es have achieved free and fair elections which are the primary requirement for a democratic nation; however they have not secured thei r citizens rights or safety as yet. It is interesting that these th ree countries all have a shared association between the level of democracy and victimization since they have all undergone recent tran sitions to democracy. Bolivia transitioned to democracy in 1982 after a period of military ru le, however, in recent years the country has ex perienced a great deal of unrest a nd even violent protest (Barr, 2005). Similarly, Guatemala transitioned to democracy after the end of their 36 year civil war ended in 1996. The transition in this country to democracy was rocky at best and there has been a great deal of violent and property crime which le ads to vigilante jus tice (Snodgrass Godoy, 2002), public lynchings (Kincaid, 2000) and general gov ernmental instability (Perz, 2003). In El Salvador, the transition to democratic rule of law has also been relatively recent. Peace accords ushered in democratic rule of law in 1992 afte r a 12 year civil war. While the transition to democracy in this country was or iginally heralded for its resounding success, recent years have shown that there is still a great deal of unres t, corruption, and instability (Perz, 2003). And while overall levels of police corruption are low, the democratic government in El Salvador still 153

PAGE 154

seems to lack legitimacy. It would appear that in all three countries where level of democracy is significantly associated with victimization that full democratic rule of law has not been established since the governments have not been able to contai n social violence, political protest or fully protect citizens rights. Where this s econd aspect of democracy is missing, there are lower levels of democratic legitimacy and higher ra tes of victimization. It is not surprising then that Bolivia, Guatemala and El Salvador have th e lowest mean scores on all levels of legitimacy of the countries in the region, ex cepting for Peru (See Table 2-1). Court Justice The level of justice provided by the court system, or their level of legitimacy as an agency of law, is significantly associat ed with individual victimization in half of the countries in the region including Chile, Guatemala, Mexico, Br azil, Venezuela, and the Dominican Republic. Individuals who report lower levels of justice are provided by th eir countrys court system are more likely to also report victimization in these countries. This is in keeping with H9 from Chapter 3. This is an interesting collection of coun tries considering that Chile is considered to be a more consolidated democracy where there are both free and fair elections and protection of citizen rights by the court system. Whereas, th e other countries wher e court justice is significantly associated with indi vidual victimization are consid ered to be less protective of citizen rights, or among the countries of either average or lower levels of legal legitimacy. It could be that respondents in the Chilean sample ar e simply more likely to report victimization. It could also be, since there is no way to determin e causal ordering from the cross-sectional data, that victimization preceded the respondents perception of the court system. A negative experience with the court system in the country af ter being victimized coul d be what is driving the perception of court justice in Chile, and in all the countries for that matter. However, the mean score for all respondents in Chile on court justice was lower than that for the full region 154

PAGE 155

(See Table 2-1) indicating that Chileans in ge neral do not believe their court system provides adequate justice. It is not a belief held only by victims in the country. When we look at the other countries where court justice is predictive of individual victimization, the results are less surprising. Mexico is known for a high level of legal corruption. Davis (2006) found that corruption was not only rampant among law enforcement officials, but also in the court system where brib es are routinely used. In dividuals in Mexico are less protected by their court system. It is possible for individuals who are guilty of either violent or property crimes to go free if they have the ri ght political connections, or the right amount of money (Lawson, 2000): Government supporters and cronies had little to fear from the law, regardless of the crimes they committedRevelations about official corr uption are as stunning as they are frequent. They include everything from government pr otection for criminalsto active involvement in drug trafficking. (Lawson, 2000, p. 283) This leaves individuals in Mexi co vulnerable to victimizati on. The same can be said of individuals in Guatemala, Venezuela, Brazil and the Dominican Republic. In Brazil in particular, there historically have been abuses of the cour t system by wealthy or well connected citizens (Ratliff & Buscaglia, 1997; Buscaglia, Dakolias & Ratliff, 1995). Again, allowing violent or individuals to go free puts citizens at risk of being victimized in countries where these legal abuses take place. If the court system is not ensuring justice for the people in these countries, then they are put at a risk of being victimized due to a lack of guardianship on the part of the courts. This same risk of victimization due to less than capable legal guardianship certainly applies to the police. Trust of Police The support for H10 is the most consistent across th e region. Not only are low levels of trust in the police associated highe r risk of victimization in the full regional model (Table 7-1), 155

PAGE 156

but individuals who report lower levels of trust in the police in Costa Rica, Uruguay, Bolivia, Guatemala, El Salvador, Mexico and Venezu ela also have a highe r likelihood of being victimized. This result is surprising in Costa Ri ca and Uruguay, both countries considered to be highly democratically consolidated where there is a respect for due process and citizen rights. In Table 2-1 we see that Costa Ricans have the gr eatest amount of trust in their police when compared to respondents from the other 11 countri es in this study. The me an score for trust of police was 5.54 (on a scale of 1 to 7), which wa s a nearly a full point higher than any other country in the region (Brazils mean score was 4. 59) and nearly two points higher than the grand mean for the region (3.55). However, the result in Costa Rica might be due in part to recent problems with policing that are affecting public opinions of the trustworthiness of their police (Lehoucq, 2005). This could also be due to the ci tizens reactions to corruption in Costa Rica when it does exist. Lehoucq (2005) found that Costa Ricans are high ly intolerant of governmental corruption, including police abuses of power. However, the recent problems with inadequate policing (Lehoucq, 2005) speak more to the theoretical assumption being made in this dissertation: there is a not ad equate or appropriate police gua rdianship in Costa Rica, which leaves some citizens vulnerable to victimization. The same cannot be said of Uruguay. Uruguay has one of the least corrupt police forces in the region. Uruguayans exhibit a high level of trust in their poli ce, the highest in the region in previous studies (Cao & Zhao, 2005). In this current study, however, while Uruguayans on average had more trust in their police than respondents from ot her countries in the region, the mean score for trust in police (on a scale of 1 to 7) was onl y 4.17 (See Table 2-1). Brazilians exhibited more trust in their po lice force than Uruguayans. It coul d be that, like Costa Ricans, respondents from Uruguay are highl y intolerant of police corruption when it does exist and that 156

PAGE 157

they are more sensitive to the issue, which is w hy they have lower levels of trust in their police force. We cannot at this point ru le out the possibility that causal ordering is an issue in Uruguay. The results from respondents in Uruguay could be a reflection of victim ized individuals having less trust in the police rather than an overall la ck of trust in police ( due to corruption) making individuals more susceptible to victimization. This is an area for future research, and again the cross-sectional nature of the data used in this project make inferences about causation problematic. While the results from Uruguay are surp rising, the results we find from Bolivia, Guatemala, El Salvador, Mexico and Venezuela are in keeping with H10. Both Bolivia and El Salvador rank among the nations in the region with lower levels of legal legitimacy. As noted in Chapter 2, Bolivia is rife with corrup tion (Domingo, 2005, p. 1728). The corruption among the police in Bolivia leave citizen s open to victimization, not on ly by their fellow citizens but also from the police themselves who routinely bribe citizens (Seligson, 2002). While corruption is not as problematic in El Sa lvador, citizens in El Salvador are understandably wary of their police. There are still lingering authoritarian pr actices among the police in El Salvador left over from the military government of the 1990s (Perz 2003). Therefore the citizens of El Salvador may have overall lower levels of trust in th eir police due to their own lingering fears of authoritarian rule. Salvadorans may not feel co mfortable going to their police when they are victimized. In both Mexico and Venezuela, human righ ts abuses by the police are not uncommon. So it is not at all surprising that in dividuals in these countries who report lower levels of trust in their police are more likely to be victimized. In Mexico, bribes by pol ice are routine as are shakedowns (Davis, 2006). The rampant corrupt ion among Mexican police is reflected in the 157

PAGE 158

Mexican respondents in this study reporting a low level of trust in their police (3.23 out of 7). Individuals who report lower leve ls of trust in the police in Mexico are more likely to be victimized and this speaks to the lack of capable guardianship by the police in Mexico. Venezuelans in particular have recently resort ed to mob or vigilante justice as a means of resolving criminal issues. This mob justice rea lly speaks to a lack of safety provided by the police or at the least, the percep tion that the police cannot maintain rule of law in Venezuela. Vigilante Justice Respondents from all 12 countries were aske d about their level of support for vigilante justice in the country. In accordance with H11, we see that individuals in the full regional model who support vigilante justice are more likely to be victimize d. In country specific models, H11 is supported in Peru, Guatemala, and Venezuela. The assumption behind this hypothesis is that where there is a great deal of support for vigilant e justice, there is an inadequate guardianship from legal authorities which makes citizens more likely to support mob or vigilante justice. As previously discussed, Venezuela has seen a recent rise in vigilante or mob justice. Venezuelans feel that their legal system, cour ts and police, cannot ad equately deal with a growing crime problem. Similar to Venezuela, Guat emala has also seen a recent rise in vigilante justice, in the form of public lynchi ngs (Snodgrass Godoy, 2002; Kincaid, 2000). The vast number of lynchings in Guatemala (as many as one per day in 1997) is evidence of a lack of public security and inadequate legal apparatuses to d eal with crime. This also exists in Peru, so it is not surprising that vigilante justice is sign ificant in the Peruvian model. Indigenous and mestizo populations in Peru have resorted to l ynchings and mob violence in the absence of a legal presence in rural communities and even urban areas (Snodgrass Godoy, 2004). While mob justice may make citizens in these countries feel safer, in reality it actually makes them more prone to violent victimization. 158

PAGE 159

As discussed in Chapter 2, innocent indivi duals sometimes fall prey to mob justice (Snodgrass Godoy, 2004; Sanjan, 2000). A lack of due process and investigation can leave citizens open to being victims of retaliatory vi olence and even criminal victimization during a mob event. Therefore, the effect of support fo r vigilante justice on vi ctimization is two-prong. First, individuals who report a support for vig ilante justice probably al so have inadequate guardianship from their legal authorities, leav ing them vulnerable to victimization. Secondly, individuals who support vigi lante justice may take pa rt in these events, wh ich would put them at risk of being victimized while participating in a mob justice event. While this is conjecture about the association between vigilante justice and victimization, it pr ovides a possible explanation for the relationship. As stated earlier, however, if the assumption of causal ordering is wrong in this analysis, it is possible that victimized individuals are more likely to support vigilante justice because they have not been able to obtain jus tice thru their own criminal justice system. However, if this were to be true, it confirms the hypot heses regarding trust of police and court justice, since these individuals are not being adequately guarded by the le gal authorities in their own county. But this is merely speculation. Clearly more research is needed to disentangle the nature of the relationship. Limitations One of the major limitations of the research presented in this dissertation is the crosssectional nature of the data. Id eally, longitudinal data would be used in order to establish the causal ordering of the dependent and independent variables. This is a limitation of all crosssectional data and specifically a limitation of this study. Theoretically, it can be assumed that it is victimization that is dr iving the respondents per ceptions of the democratic rule of law in their country and how well their justice systems operate. It is certainl y possible that respondents who 159

PAGE 160

have been victimized simply have lower levels of trust in both their court and police systems and they have higher levels of support for vigilant e justice. This could be the result of such victimization. However, this projec t shows that, for the most part, in the countries where there is a history of police abuses, low levels of democr atic rule of law, court systems that are not enacting justice and where there have been recent cases of vigilante justi ce, all of these factors are significantly associated with the likelihood of individual vi ctimization. Furthermore, it has not been an assumption of this research that th ese factors predict indivi dual victimization, but certainly that when an individual reports a low le vel of trust in the police, it might very well be that the police that this individual have had cont act with are not providing adequate guardianship and therefore have left this indi vidual open to victimization. This is the basic premise of this research, but it is by no means purported to be the only explanation for the relationships found in this study. A further limitation of this study is the nature of the survey used. As stated in Chapter 4, this survey was not designed as a victimization survey. Preferably, when studying victimization there would be more information about the vict imization experience, including information on time, place, the amount of property damage, the exte nt of physical victimization and so on. Even more importantly to the study at hand, informati on on the perpetrator of such victimization is essential. For example, if responde nts in the survey have been vi ctimized by a police officer, this would certainly have an impact on their per ception of law enforcement and government in general. It would also impact their support for extra-legal justi ce in the country. Not having this information allows for a lot of questions regardin g the nature of individua l victimization in each country. 160

PAGE 161

Out of necessity, in this study victimiza tion types, both physical and property, were combined. While this was done out of necessity not all countries asked the respondents which type of victimization they experienced in previous year, it is not ideal to combine these types of victimization, particularly give n that kidnapping and property damage are collapsed into one form of victimization. This limitation might be why the results show income, education and socio-economic attainment having such an impact on victimization. It might be due in part to fact that the majority of victimization is actually prop erty victimization (theft or property damage). In the countries where we do have information regarding the type of victimization experienced, the majority of victimization is eith er robbery without physical threat4 (7.5 percent) or robbery with physical threat (4.5 percent). The other forms of victimization were al l under 2 percent with property damage (1.2 percent) and home robbery (1.8 percen t) having the next highest frequency. Victimizations such as kidnapping an d sexual assault had very low reporting numbers (less than 0.1 percent each). Due to the infreque nt reporting of violent victimization and the fact that the type of victimization was not asked of respondents in all countries, it was statistically impossible to disentangle the types of victimization. Another limitation of this study is the dichotomous nature of the dependent variable. Respondents were asked if the vic timization had occurred, but ther e is no gauge of the frequency of victimization. It is cert ainly possible given the information we know about repeat victimizations, that respondents may have experienced multiple forms of victimization and multiple experience with the same form of victim ization. There is no way to gauge if individuals who are reporting lower levels of legitimacy of are those individuals are more likely to undergo multiple victimizations. 4 This is how the question was posed to respondents. 161

PAGE 162

Reliability of the data is always an issue when using self-reported survey data. There is always the possibility that individuals may under -report or even over-repo rt victimization. This is entirely possible with some of the data used in this project. For example, in Jamaica only 10 percent of respondents reported be ing victimized in the previous 12 months. However, we know from current research that Jamaica has an ever increasing crime and viol ence problem which has escalated over the past three decades (Moser & Holland, 1997). A recent victimization study found that nearly two-thirds of the respondents of the survey (n= 3, 112) reported being victimized in their lifetime (61.4 percent) and that over a third of the respondents had experienced more than three criminal victimi zations within their lifetime (Wortley, Gartner, Seepersad & McCalla, 2006). While these data were collected on life time experiences with victimization, the same study found that 23.7 percent of respondents had been victimized within the past 12 months and that 11 pe rcent had experienced more than one criminal victimization in that same time period (Wortley et al., 2006). The di sparity between reporting from the Jamaican National Crime Victimization Survey (JNCVS) and the data from the Latin American Public Opinion Project (LAPOP) the two numbers begs the question, how reliable is the LAPOP data from Jamaica? It might be the nature of the su rvey that accounts for the disparity, the survey conducted by Wortley and his associates was a vi ctimization survey; the staff administering the survey was trained specifically to gather info rmation on property and physical victimization. The LAPOP survey was primarily constructed to gath er information regarding political climate, issues of democratization, et cetera. Victimizatio n was not the focus of the survey, and therefore respondents from Jamaica might not have considered it essential to report their victimization to the survey administrators. There ce rtainly exists the possibility that the data from other countries in the survey is either conflated or inflated as well. The list of limitations of this research is 162

PAGE 163

certainly not exhaustive, but it does note that th ere are some problems with this study that should be addressed in future research. Future Research This study is a preliminary step in toward understanding crim inal victimization in Latin American and within specific c ountries in the region. The possibilities for fu ture research are infinite. However, first and foremost, future re search should address the limitations of this current study. Ideally, a thorough study of victimization across 12 Latin American countries would use data from a victimization survey th at was designed to coll ect uniform information from all countries. The scope of such a project would be immense, however, some of the differences highlighted in this dissertation show that the need for more in-depth information regarding victimization across th e region is also great. A su rvey that asked information regarding repeat victimizations, perpetrator victimization, domestic violence assault and so forth, would certainly add greatly to our current knowledge of victimiza tion trends and how they differ throughout the region. Additionally, if more questions were asked of respondents in all countries regarding specific issues of law and order that affect perceptions of legitimacy of law, more theoretical variables could be incl uded in a study such as this. Future research should also address issues of legitimacy of law and how that impacts victimization. A longitudinal study that examines these issues and also looks at victimization experiences would help solve the causal ordering problem of this current study. Does legitimacy of law impact victimization or does victimization impact perceptions of legitimacy of law? Or is it cyclical? It could be entirel y possible that individuals who hold already low opinions of the legal authorities in their countr y, once victimized, hold even lower opinions of these agencies. A longitudinal study comparing opinions across time and accounting for the impact of victimization would help answer this question. 163

PAGE 164

Theory in the Latin American Contexts This project had two specific aims; the first was to test Tylers legitimacy of law theory and determine how much citizens perceptions of democracy and the legitimacy of legal authorities is associated with indi vidual victimization. This project di d in fact show that there is a strong relationship between the theoretical variab les and victimization. In the country specific models there is also support fo r aspects of legitimacy of law. However, this comparative study shows that individuals in each of the countries in this study experience victimization differently and that not all aspects of legitimacy of law aff ect individuals in all countries. The support in some countries was greater than in others. For ex ample, in Guatemala all aspects of legitimacy of law were significant in the directions predicted. In Venezuela as well three of the four theoretical variables were significan tly associated with victimization. However, in other countries only one or two of the theoretical variable s were significant. In Jamaica, none of the theoretical variables were associated with individual victimization. Undoubtedly this st udy shows that theory of legal legitimacy is not applicable in all Latin Ameri can countries. It does show nonetheless, that the theory can be applied in divergen t cultural contexts. Future research testing legitimacy of law, as well as other theories in Latin American countri es and globally is warranted. Moreover, future research needs to focus on specific cultural or social conditions within a country and adapt current criminological theory to the context of that country. The second aim of this dissertation was to ex amine how predictors of victimization, both demographic and theoretical, vary from country to country. In this aim, this dissertation project showed that individual victimi zation varies, often dramatically, from country to country. Certain demographic variables had competing affects from country to country. One of the most interesting findings of this study was how the va riables associated with victimization from country to country did not fall into even patt erns among the countries deemed to have more 164

PAGE 165

legitimacy of law, lower legitimacy of law and av erage levels of legitimacy. This is important because it highlights one of the main points of th is project: these countries are more dissimilar than similar and need to be treated as suc h. The idea that there is a Latin American victimization and violence problem is erroneou s. There are specific pr oblems within each country that need addressing. Some of these pr oblems certainly can be addressed within the context of the region, but at the end of the day, these are autono mous countries with their own governmental agencies, and levels of democratiz ation. That which impacts victimization in one country does not necessarily impact victimization in another. Beyond these specific aims, this project attempted to answ er the call by LaFree (2007), Marsh (1967), Reed and Yeager (1996) and ot hers who have argued for more comparative research on crime and victimization, including non-Western or non-industrialized nations. There is a need to take into account the impact of governmental stability on issues of crime and victimization. As citizens of the United States, Canada or European countries often criminologists take for granted a stable democracy, a relative respect for due process and a criminal justice system that operates, for the mo st part, without pervasive corruption. These basic assumptions cannot be made in all Latin American countries. While Costa Rica, Uruguay and Chile have in re cent history been ab le to provide these things for their citizens, other nations, such as Guatemala and Peru, are far from achieving such stability. Ignoring these contextual differen ces among these countries does not allow for researchers to that take into account the broader contex t of individual victim ization. The level of guardianship that an individual in Uruguay is prov ided by its police is ce rtainly different from the guardianship provided to a citizen in Mexico. To ignore the impact that this has on physical or property victimization misses a huge piece of th e puzzle as to what affects an individual being 165

PAGE 166

victimized in one country versus another. Current theory, such as routine activities, needs to be broadened to include the guardians hip of a legitimate system of law, that cannot be assumed when doing cross-national research. Conceptualiz ing guardianship in a way that includes the level of state guardianship provide d to citizens, allows for the po ssibility that these theoretical concepts may operate on a grander scale. It is not just the daily guardianship provided by a security system, a dog, mace, friends or family that impacts ones personal security, it is the government, the police, the courts, and rule of law in general. The study of theory only in Canada, the United States or Europe can only push the field forward to a certain extent. Comparative studies that test current criminol ogical theory are essent ial to development of criminology as a discipline. Tests of criminological theory in non-Western nations are not just warranted, but essential. This includes tests of theories of law, but also shou ld include testing victimization theories such as routine activitie s, offending theories such as soci al learning or even structural level theories would help not only to advance our understanding of what is happening with crime and victimization within the countries of Latin America, but advance the theories themselves. Understanding the cultural limitations of current criminological theory allows for the creation of new theoretical perspectives a nd still advances our knowledge of how law operates in society. Even if research does find that current criminological theory is culturally limited, it still propels the discipline forward because it encourages the creation of new theo ries or adaptation of existing theories to cultu rally diverse societies. 166

PAGE 167

LIST OF REFERENCES ___. (1997). Policing the police: Brazil. (police corruption) (International). The Economist 343.8012, (April 12, 1997), 42(2). ___. (1999). Police crimes.(punishing Brazilian policemen who commit crimes)(Brief Article). The Economist 352.8133 (August 21, 1999), 28. ___. (2005). UNIFEM. Dossier sobre violencia domestic a en Amrica Latina y el Caribe Comit de Amrica Latina y el Carive para la defense de los derechos humanos. Lima, Peru. http://www.unifem.org/materials/it em_detail.php?ProductID=58&lang=spn (accessed November 16, 2009). ___. (2005). USAID, Dominican Republic Congressional Budget Justification, 2005. http://www.usaid.gov/policy/ budget/cbj2005/lac/dr.html (accessed July 10, 2009). ___. (2007). Election Assistance Commission, 2000 Voter Registration and Turnout Review www.eac.gov/clearinghouse/voter -turnout-and-statistics/ (accessed July 10, 2009). ___. (2007). Pan American Health Organization, H ealth Surveillance and Disease Management Area. Health Statistics and Analysis Unit. PAHO Regional Mortality Database http://www.paho.org/English/SHA/core data/tabulator/newTabulator.htm (accessed August 21, 2009). ___. (2007). U.S. State Department, Bureau of Democracy, Human Rights and Labor. Dominican Republic (March 6, 2007). http://www.state.gov/g/drl/rls/hrrpt/2006/78889.htm (accessed July 10, 2009). Acock, A. C. (2005). Working with missing values. Journal of Marriage and Family 67:4, 1012-1028. Ahnen, R. (2003). Between tyranny of the majority and liberty: The persistence of human rights violations under democracy in Brazil. Bulletin of Latin American Research 22 (3), 31939. Allison, P. (1999). Logistic regression using the SAS system: Theory and application Cary, NC: SAS Institute Inc. Anderson, E. J. (1999). Code of the street: Decency, violence and the moral life of the inner city New York: W.W. Norton & Co. Archer, D., & Gartner, R. (1984). Violence and crime in cr oss-national perspective New Haven, CT: Yale University Press. Ashcraft, Richard (ed.). (1991). John Locke: Critical assessments. London: Routledge. 167

PAGE 168

Baln, J. (2002). Introduction. In S. Rotker, Citizens of fear: Urban vi olence in Latin America (1-6). New Brunswick, NJ: Rutgers University Press. Bar, R. B. (2005). Bolivia: Another uncompleted revolution. Latin American Politics and Society 47(3), 69-90. Becker, G. S., Duesenberry, J. S., & Okun, B. ( 1960). An economic analysis of fertility in Becker, G.S., Duesenberry, J. S. & Okun, B. (Eds). Demographic and Economic Change in Developed Countries, (225-256). Washington, D.C.: UMI Publishers. Beetham, D. (2004). Freedom as the foundation. Journal of Democracy 15(4), 61-75. Bergman, M. (1999). Untitled Paper presented in Se ssion I: Explanations a nd analyses of rising crime rates in Latin America, presented at th e Research Workshop held at the University of Texas at Austin. Birkbeck, C. (1993). Against ethnocentrism: A cr oss-cultural perspective on criminal justice theories and policy. Journal of Criminal Justice Education 4 (2), 307-324. Birkbeck, C., Gabaldn, L. G., & LaFree, G. (1993). The decision to call the police: A comparative study of the United States and Venezuela. International Criminal Justice Review 3, 25-43. Booth, J. A. & Seligson, M. A. (2005). Political legitimacy and participation in Cost Rica: Evidence of arena shopping. Political Research Quarterly, 58(4), 537-550. Brice, A. (2009). 30 die as i ndigenous protestors, police clas h in Peru. CCN.com/World, http://www.cnn.com/2009/WORLD/americas/0 6/05/peru.indigenous.clash/index.html?ire f=newssearch (accessed July 10, 2009). Brice, A. (2009). Dominican leader keep s up crackdown on corruption. CNN.com/World, www.tklms.com/2009/WORLD/americas /03/02/dominican.drug.war/index.html. (accessed July 10, 2009). Brinks, D. M. (2007). The judicial response to police killin gs in Latin America: Inequality and the rule of law. New York, NY: Cambridge University Press. Buhi, E. R., Goodson, P. & Neilands, T. B. (2008). Out of sight, not out of mind: Strategies for handling missing data. American Journal of Health Behavior 31(1), 83-92. Buscaglia, E., Dakolias, M. & Ratliff, W. E. (1995) Judicial reform in Latin America: A framework for national development. Stanford, CT: The Hoover Institution. Buvini M. & Morrison, A. (1999). Causes of violence. Technical Note 3. Washington, D.C., Inter-American Development Bank. 168

PAGE 169

Buvini M., Morrison, A. & Orlando, M. B. (2005). Vi olencia, crimen y desarrollo social en Amrica Latina y el Caribe Chapter 5 in FLASCO, Desarrollo social en Amrica Latina: temas y desafos para las polticas pblicas Costa Rica: FLASCO. Buvini M., Morrison, A. & Shifter, M. (1999). Violence in Latin America and the Caribbean: A framework for action. Washington, D.C., Inter-American Development Bank. Cabella, W. & Porzecanski, R. (December 6, 2007) The growth of ethnic minorities in Uruguay: Ethnic renewal or measurement problems? Paper presented at an annual conference Social Statistics and Ethnic Diversity. Montreal, CA. http://www.ciqss.umontreal. ca/Docs/SSDE/pdf/Cabella.pdf (accessed November 16, 2009). Caldeira, T. P. R. (2000). City without walls: Crime, segregat ion and citizenship in So Paulo. Berkeley: University of California Press. Caldeira, T. P. R., & Holston, J. (199 9). Democracy and violence in Brazil. Comparative Studies in Society and History 41 (4), 691-729. Call, C. T. (2002). Assessing El Salvadors trans ition from civil war to peace. in Stedman, S. J., Rothchild, D. & Cousens, E. (Eds.). Ending Civil Wars (543-602). Boulder, CO: Lynne Rienner Press. Call, C. T. (2003). Democratisation, war and statebuilding: Constructing the rule of law in El Salvador. Journal of Latin American Studies 35, 827-862. Canache, D., Mondak, J. J., & Seligson, M. A. (2001). Meaning and measurement in crossnational research on satis faction with democracy. Public Opinion Quarterly, 65, 506-528. Cao, L. & Zhao, J. S. (2005). Confiden ce in the police in Latin America. Journal of Criminal Just ice, 33, 403-412. Casey, L. (2005). Venezuela under Chvez: Some truths are not all that complicated. Dissent (Summer 2005), http://www.dissentmagazine.org/article/?article=220 (accessed July 11, 2009). Catalano, S. M (2006). Criminal victimization, 2005. Washington, DC: Bureau of Justice Statistics, U.S. Department of Justice. Chevigny, P. (1995). Edge of the knife: Police violence in the Americas. New York; New Press. Cohen, A. K. (1955). Delinquent boys: The culture of the gang New York: Free Press. Cohen, L. & Felson, M. (1979). Social change and crime rate trends : A routine activities approach. American Sociological Review 44, 588-608. 169

PAGE 170

Concha-Eastman, A. (2002). Urban violence in Latin American and the Caribbean: Dimensions, explanations, actions. In S. Rotker (ed), Citizens of fear: Urban violence in Latin America (37-54). New Brunswick, NJ: Rutgers University Press. Dahl, R. A. (1971). Polyarchy: Participation and opposition New Haven, CT: Yale University Press. Davis, D. E. (2006). Undermining the rule of law: Democratization and the dark side of police reform in Mexico. Latin American Politics & Society 48(10), 55-86. Davis Rodrigues, C. (2006). Civil democracy, pe rceived risk, and insecu rity in Brazil: An extension of the systemic social control model. The ANNALS of the American Academy of Political and Social Science 605, 242-263. DeFleur, L. B. (1970). Delinquency in Argentina: A study of Crdobas youth. Pullman: Washington State University Press. Domingo, P. (1999). Rule of Law, citizen ship and access to justice in Mexico. Mexican Studies/Estudios Mexicanos 15(1), 151-191. Domingo, P. (2005). Democracy and new social forces in Bolivia. Social Foces 83(4), 17271743. Doorenspleet, R. (2005). Democratic transitions: Exploring the structural sources of the fourth wave. Boulder, CO; Lynne Rienner Publishers. Eisler, L. & Schissel, B. (2004). Privation and vulnerability to victimization for Canadian youth. Youth Violence and Juvenile Justice 2 (4), 359-373. Encinoza, A. R., and Del Olmo, R. (1981). The vi ew from Latin America against transnational criminology: A call for democratic international cooperation. Crime and Social Justice, 15, 61. Encuestas & Estudios (2000). Multiple years. As ipiensan los bolivianos. Annual survey. La Paz: Encuestas & Estudios. February. Espinal, R. (1996). The Dominican Republic: An ambiguous democracy. in Dominguez, J. & Lowenthal, A. (Eds.), Constructing democratic governanc e: Mexico, Central American and the Caribbean in the 1990s (118-134). Baltimore, MD: Johns Hopkins University Press. Espinal, R., Jartlyn, J. & Kelly, J. M. (2006). Performance still matters: Explaining trust in government in the Dominican Republic. Comparative Political Studies 39(2), 200-223. 170

PAGE 171

Eyzaguirre, G. (1999). Percepcione s de la prensa sobre el parl amento y la corte nacional electoral. In C. Q. Soruco and G. Eyzaguirre (Eds). El derecbo a la informaci6n y percepciones sobre instituciones ILDIS. 75-138. Fagan, J. (2008). Legitimacy and criminal justice. Ohio State Journal of Criminal Law 6, 123140. Fagan, J. & Tyler, T. R. (2004). Policing, order maintenance and legitimacy. in G. Mesko, M. Pagon and B. Dobovsek (Eds). Policing in Central and Eastern Europe: Dilemmas of contemporary criminal justice University of Maribor, Slovenia. Fajnzylber, P., Lederman, D., & Loayza, N. (1998). Determinants of crime rates in Latin America and the world: An empirical assessment. Washington, D.C.: The World Bank Fierlbeck, K. (1998). Globalizing democracy: Power, legi timacy and the interpretation of democratic ideas. New York; Manchester University Press. Finkel, S. E., Sabatini, C. A., & Bevis, G. G. (2000). Civic education, civ il society, and political mistrust in developing democracy: The case of the Dominican Republic. World Development, 28(11), 1851-1874. Fischer, M. E. ed. (1996). Establishing democracies. Boulder, CO: Westview Press. Fisher, J., & Mason, R. (1981). The analysis of multicollinear data in criminology. In J.A. Fox (Ed.), Methods in quantitative criminology (pp. 99-125). New York, NY: Academic Press. Fleischer, D. (1997). Politi cal corruption in Brazil. Crime, Law & Social Change 25, 297-321. Frhling, H. (2007). Police legitimacy in Chile. in T. R. Tyler Legitimacy and criminal justice (115-145). New York: Russell Sage. Frhling, H., Tulchin, J. S. & Golding, H. A. (Eds) (2003). Crime and violence in Latin America: Citizen security, democracy and the state. Washington, D.C.: Woodrow Wilson Center Press. Gaviria, A. & Pags, C. (1999). Patterns of crime victimization in Latin America Washington, D.C.: Inter-American Development Bank. Gaviria, A. & Velez, C. E. (2001). Who Bears the Burden of Crime in Colombia? Fedesarrollo Bogot. Gibson, C. (1966). Spain in America New York: Harper Collins. Harriot, A. (2000). Police and crime control in Jamaica: Problems with reforming ex-Colonial constabularies Kingston, Jamaica: University Press of the West Indies. 171

PAGE 172

Heinemann, A. & Verner, D. (2006). Crime and violence in developmen t: A literature review of Latin America and the Caribbean World Bank Policy Research, Washington, D.C.: The World Bank. Hindelang, M. S., Gottfredson, M. & Garofalo, J. (1978). Victims of personal crime. Cambridge, MA: Ballinger. Holiday, D. & Stanley, W. (1993). Building the p eace: Preliminary lessons from El Salvador. Journal of International Affairs 46(2), 1-28 Huggins, M. K. (2000). Legacies of Authoritarianism: Brazilian torturers and murderers reformulation of memory. Latin American Perspectives. 27(2), 57-78. Huggins, M. K. (2000b). Urban violence and police privatization in Brazil: Blended invisibility. Social Justice, 27(2), 113-134. Hunter, A. J. (1985). Private, parochial and publi c orders: The problem of crime and incivility in urban communities. in G. D. Suttles & M. N. Zald (eds.) The challenge of social control: Citizenship and institution building in modern society Norwood, NJ: Ablex. Ibarra, M. (2003). Crime victim ization and reporting in Peru. http://www.terry.uga.edu/~mustard/gc/ibarra.pdf (accessed September 17, 2008). Jacob, H. (1974) The potential for reform of criminal justice Beverly Hills, CA: Sage Publications Johnson, J. L. (2007). When the poor police themse lves: Public insecurity and extralegal criminal-justice administration in Mexico. in T. R. Tyler (ed). Legitimacy and Criminal Justice (167-185). New York: Russell Sage Foundation. Kane, R. J. (2002). The sociology ecology of police misconduct. Criminology 40, 867-896. Kane, R. J. (2003). Social Control in the metropolis: A community-level examination of the minority group-threat hypothesis. Justice Quarterly, 20, 401-431. Kane, R. J. (2005). Compromised police legitimacy as a predictor of violent crime in structurally disadvantaged communities. Criminology 43(2), 469-498. Kay, C. (2001). Reflections on rura l violence in Latin America. Third World Quarterly 22 (5), 741-755. Kay, C. (2007).Land, conflict, a nd violence in Latin America. Peace Review: A Journal of Social Justice, 19, 5-14. 172

PAGE 173

Kincaid, D. A. (2000). Demilitarization and security in El Salvador and Guatemala: Convergences of success and crisis. Journal of International Studies and World Affairs 42(4), v-58. Klaus, P., & Rennison, C. M. (2002). Age patterns in violent victimization, 1976-2000 Washington, DC: Bureau of Justice Stat istics, U.S. Department of Justice. Klevens, J. Duque, L. F., Ramrez, C. (2002) The victim-perpetrator overlap and routine activities: Results from a cross-se ctional study in Bogot, Colombia. Journal of Interpersonal Violence 17(2), 206-216. Kurbin, C. & Peterson, R. D. (2003). Retaliat ory homicide: Concentr ated disadvantage and neighborhood culture. Social Problems 50, 157-180. Kurbin, C. E. & Weitzer, R. (2003). New dir ections in social disorganization theory. Journal of Research in Crime and Delinquency 40, 374-400. Krug, E. G., Dahlberg, L. L., Mercy, J. A., Zwi, A. B. & Lozano, R. (Eds). (2002) World report on violence and health Geneva: World Health Organization LaFree, G. (2007). Expanding criminologys doma in: The American Society of Criminology 2006 presidential address. Criminology 45 (1), 501-532. Lamounier, B. (2002). Introduction: Globalization, in equality and democracy. in J. S. Tulchin and A. Brown (eds). Democratic Governance and social inequality (1-10). Boulder, CO: Lynne Rienner Publishers. Latinobarmetro. (2004). Summary Report, Latinobarmetro 2004: A decade of measurements. Santiago, Chile: Corp oracin Latinobarmetro. Accessed at http://www.latinbarometro.org/uploads/media/2004_01.pdf. Lawson, C. (2000). Unfinished tran sition: Democratization and aut horitarian enclav es in Mexico. Mexican Studies/Estudios Mexicanos 16(2), 267-287. LeBeau, J. & Castellano, T. (1987). The routine activities approach: An inventory and critique. Carbondale, IL: Center for the Studies of Cr ime, Delinquency, and Corrections, Southern Illinois University. Lehoucq, F. (2005). Costa Rica: Paradise in doubt. Journal of Democracy 16(3), 140-154. Levitt, S. D. (1999). The changing relationshi p between income and crime victimization. Economic Policy Review 5 (3), 87-98. Lipset, S. M. (1983). Political man: The social bases of politics (2nd ed.) London: Heinemann. 173

PAGE 174

Long, S. (1997). Regression models for categorical and limited dependent variables. Thousand Oaks, CA: Sage Publications. Lpez-Montiel, A. G. (2000). The military, political power, and police relations in Mexico City. Latin American Perspectives 27(2), 79-94. Lustig, N. (1995). Coping with austerity: Poverty and inequality in Latin America Washington, D.C.: The Brooking Institution. Marenin, O., & Reisig, M. D. (1995). A genera l theory of crime and patterns of crime in Nigeria: An exploration of methodological assumptions. Journal of Criminal Justice, 23, 501. Marsh, Robert M. 1967. Comparative Sociology New York: Harcourt. McMillan, J. & Ziodo, P. (2004). How to s ubvert democracy: Mentesinos in Peru. CESifo Working Paper No. 11730 Meier, R. F. & Miethe, T. D. (1993). Unders tanding theories of cr iminal victimization. Crime and Justice, 17, 459-499. Mitchell, M. J. & Wood, C. H. (1998). Ironies of c itizenship: Skin color, police brutality, and the challenge to democracy in Brazil. Social Forces 77(3), 1001-1020. Morrison, A. & Biehl, L. (eds.) (1999). Too close to home: Domestic violence in the Americas Inter-American Development Bank. Baltimore, MD: Johns Hopkins University Press. Morrison, A., Buvini M. & Shifter, M. (2003). The violent Americas: Risk factors, consequences, and policy implications of soci al and domestic viol ence. in Frhlong et al. (Eds.), Crime and violence in Latin America: Citizen security, democracy, and the state (93-122). Washington, D.C.: Woodrow Wilson Center Press. Moser, C. & Holland, J. (1997). Urban poverty and violence in Jamaica Washington, D.C.: The World Bank. Neapolitan, J. L. (1994). Crossnational variation in homicides: the case of Latin America. International Criminal Justice Review 4, 4-22. Nikolic-Ristanovic, V. (2002). Social Gender and Violence: Po st-Communist War Affected Societies. Kluwer Academic Publishers. ODonnell, G. (1999). Polyarchies and the (un )rule of law in Latin America: A partial conclusion. In J. E. Mendez, G. ODonnell and P. S. Pinheiro (eds) The (un)rule of law and the underprivileged in Latin America South Bend, IN: Norte Dame University Press. 174

PAGE 175

Paes-Machado, E. & Noronha, C. V. (2002). Piers Armstrong ( trans. ). Policing the Brazilian poor: Resistance to and acceptance of police br utality in urban popular classes (Salvador, Brazil). International Criminal Justice Review 12, 53-76. Patrinos, H. A. & Psacharopoulos, G. (1997). Family size, and child labor in PeruAn empirical analysis. Journal of Population Economics 10: 387-405. Perz, O. J. (2003). Democratic legitimacy and public insecurity: Crime and democracy in El Salvador and Guatemala. Political Science Quarterly 118(4), 627-644. Perkins, C. A. (1997). Age patterns of victims of serious violent crime Washington, DC: Bureau of Justice Statistics, U.S. Department of Justice. Psacharopoulos, G., Morley, S., Fiszbein, A., Lee, H. & Wood, B. (1997). Poverty and income distribution in Latin Americ a: The story of the 1980s. Washington, D.C.: The World Bank. Ratliff, W. & Buscaglia, E. (1997) Judicial reform: The neglected priority in Latin America. Annals of the American Academy of Political and Social Science 550, 59-71. Reames, B. (2003). Police forces in Mexico: A pr ofile. La Jolla, CA: Center for U.S. Mexican Studies. Accessed at http://usmex.ucsd.edu/justice. Reed, G. E., & Yeager, P. C. (1996). Organiza tional offending and neoclassical criminology: Challenging the reach of a general theory of crime. Criminology 34, 357. Rosenberg, T. (1991). Children of Cain: Violence and the violent in Latin America. New York, NY: Penguin Books. Rotker, S. (2002). Citizens of fear: Urban violence in Latin America New Brunswick, NJ: Rutgers University Press. Sampson, R. J. (1983). The effects of urbanizatio n and neighborhood characteristics on criminal victimization. in R.M. Figlio, S. Hakim & G. F. Rengert (Eds.) Metropolitan crime Patterns, (3-26). Monsey, NY: Criminal Justice Press. Sanjun, A. M. (2002). Democracy, citizenship and violence in Venezuela. In S. Rotker (ed) Citizens of fear: Urban violence in Latin America (87-101). New Brunswick, NJ: Rutgers University Press. Schelder, A. (1998). What is democratic consolidation? Journal of Democracy 9(2), 91-107. Seligson, M. A. (1980). Trust, efficacy, and modes of political participat ion: A study of Costa Rican peasants. British Journal of Political Science 10, 75-98. 175

PAGE 176

Seligson, M. A. (1987). Costa Rica and Jamaica. in M. Weiner and E. Ozbudun (Eds.) Competitive Elections in Developing Countries Durham, NC: Duke University Press. Seligson, M. A. (2002). The impact of corrup tion on regime legitimacy: A comparative study of four Latin American countrie s The Journal of Politics 64(2), 408-433. Seligson, M. A., (2006). Democracy audit Bolivia, 2006: Technical information. http://sitemason.vanderbilt.edu/lapop/BOLIVIABACK (accessed September 2, 2008). Seligson, M. A., (2006b). The political culture of democracy in Chile, 2006: Technical information. http://sitemason.vanderbilt.edu/lapop/CHILEBACK (accessed August 25, 2008). Seligson, M. A., (2006c). The political culture of democracy in Costa Rica, 2006: Technical information. http://sitemason.vanderbilt.edu/lapop/COSTARICABACK (accessed September 2, 2008). Seligson, M. A. (2006d). The political culture of democracy in Dominican Republic, 2006: Technical information. http://sitemason.vanderbilt.edu/lapop/DOMINICANREPUBLICBACK (accessed September 26, 2008). Seligson, M. A., (2006e). The political culture of democracy in El Salvador, 2006: Technical information. http://sitemason.vanderbilt.edu/lapop/ELSALVADORBACK (accessed September 2, 2008). Seligson, M. A., (2006f). The political culture of democracy in Guatemala, 2006: Technical information. http://sitemason.vanderbilt.edu/lapop/GUATEMALABACK (accessed August 28, 2008). Seligson, M. A. (2006g). The political culture of democracy in Jamaica, 2006: Technical information. http://sitemason.vanderbilt.edu/lapop/JAMAICABACK (accessed September 26, 2008. Seligson, M. A., (2006h). The political culture of democracy in Mexico, 2006: Technical information. http://sitemason.vanderbilt.edu/lapop/MEXICOBACK (accessed August 25, 2008). Seligson, M. A., (2006i). The political culture of democracy in Peru, 2006: Technical information. http://sitemason.vanderbilt.edu/lapop/PERUBACK (accessed August 25, 2008). Seligson, M. A., (2007). The political culture of democracy in Brazil, 2007: Technical information. http://sitemason.vanderbilt.edu/lapop/BRAZILBACK (accessed August 25, 2008). 176

PAGE 177

Seligson, M. A., (2007b). The political cultur e of democracy in Uruguay, 2007: Technical information. http://sitemason.vanderbilt.edu/lapop/URUGUAYBACK (accessed September 2, 2008). Seligson, M. A., (2007c). The political culture of democracy in Venezuela, 2007: Technical information. http://sitemason.vanderbilt.edu/lapop/VENEZUELABACK (accessed August 25, 2008). Seligson, M. A. (2008). LAPOP Latin American Public Opinion Project, 2004-2008 Retrieved August 25, 2008, from web site: http://www.LapopSurveys.org. Seligson, M. A. & Booth, J. A. (1993). Political culture and regime types: Evidence from Nicaragua and Costa Rica. The Journal of Politics 55(3), 777-792. Sherman, J. W. (2000). Latin America in crisis. Boulder, CO: Westview Press. Smith, D. J. (2007). The foundations of legitimacy. in T. R. Tyler (ed.) Legitimacy and Criminal Justice, (30-58). New York: Russell Sage Foundation. Smith, D. A. & Jarjoura, G. R. (1989). Hous ehold characteristics, neighborhood composition and victimization risk. Social Forces 68(2), 621-640. Snodgrass Godoy, A. (2002). Lynchings and democrat ization of terror in postwar Guatemala: Implications for human rights. Human Rights Quarterly 24(3), 640-661. Snodgrass Godoy, A. (2004). When justice is cr iminal: Lynchings in contemporary Latin America. Theory and Society 33(6), 621-651. Snyder, J. (2000). From voting to violence: Democr atization and nationalist conflict New York; W. W. Norton & Company. Soares, Rodrigo R. (2004). Crime reporting as a measure of institutional development. Economic Development and Cultural Change 52 (4), 851-871. Soares, R. R. & Naritomi, J. (2007). Understanding high crime rates in Latin America: The role of social and policy factors Paper presented at the conf erence Confronting Crime and Violence in Latin America: Crafting a pub lic policy agenda. Boston, MA: Harvard University. Solum, L. B. (2004). Procedural justice. Southern California Law Review 181, 277. Sunshine, J. & Tyler, T. R. (1997). Social order and disorder of the street blocks and neighborhoods: Ecology, microecology, a nd the systemic model of social disorganization. Journal of Research in Crime and Delinquency 3, 113-155. 177

PAGE 178

Tedesco, L. & Barton, J. R. (2004). The state of democracy in Latin America: Post-transitional conflicts in Argentina and Chile New York: Routledge. Tennyson, K. M. (2004). Battering in Brazil: Femi nist and Family Sociology Perspectives on Domestic Violence. MA Thesis University of Florida, 2004. Tjaden, P., & Thoennes, N. (2000). Full report of the prevalence, incidence, and consequences of violence against women: Findings from the National Vi olence Against Women Survey Research Report. Washington, D.C.: National Institute of Justice and the Centers for Disease Control and Prevention. Tjaden, P. & Thoennes, N. (2006). Extent, nature, and consequences of rape victimization: Findings from the National Violence Against Women Survey. Special Report. Washington, D.C.: National Institute of Jus tice and the Centers for Disease Control and Prevention. Tulchin, J. S. & Brown, A. (2002). Democratic governance and social inequality Boulder, CO: Lynne Rienner Publishers. Tyler, T. R. (1990). Why people obey the law New Haven; Yale University Press. Tyler, T. R. (2006). Psychological persp ectives on legitimacy and legitimation. Annual Review of Psychology 57, 375-400. Tyler, T. R. ed. (2007). Legitimacy and criminal justice New York: Russell Sage Foundation. Tzanelli, R. (2006). Capitalizing on value: Towards a sociological understanding of kidnapping. Sociology 40 (5), 929-947. Valenzuela, A. (2004). Latin Amer ican presidencies interrupted. Journal of Democracy 15(4), 5-19. Vargas-Cullel, J. & Rosero-Bixby, L. (2005). La cultura poltica de la democracia en Costa Rica, 200, San Jos: EUCR. Weber, M. (1968). Economy and society Berkeley: University of California Press. Weyland, K. (2001). Will Chvez lose his luster? Foreign Affairs 73, 73-87. Wiarda, H. J. (2003). The soul of Latin America: Th e cultural and political tradition New Haven: Yale University Press. Williams K. R. & Flewelling, R. L. (1988). Th e social production of criminal homicide: A comparative study of disaggregated rates in American cities. American Sociological Review, 53, 421-431. 178

PAGE 179

179 Willis, C. L., Evans, T. D., & LaGrange, R. L. (1999). Down home Criminology: The place of indigenous theories of crime. Journal of Criminal Justice, 27 (3), 227-238. World Health Organization. (2007) Pan American Health Organization, Health analysis and statistics unit. Regional core health data initiative; Technical Health Information System. Washington, DC. Wortley, S., Gartner, R,. Seeper sad, R., & McCalla, A. (2006). The Jamaican national crime victimization survey: Final report. Centre of Criminology, University of Toronto: http://www.mns.org.jm/Portals/0/JNCVSfinal_2006%5B1%5D.pdf (accessed November 23, 2009). Wright, R. (2005). Stolen continent: 500 years of conque st and resistance in the Americas New York: Mariner Books. Zaverucha, J. (2000). Fragile democracy and th e militarization of public safety in Brazil. Latin American Perspectives 27(3), 8-31.

PAGE 180

BIOGRAPHICAL SKETCH Kristin M. Tennyson was born in a small town in northern New York. She attended a boarding school in Asheville, NC. She went on to earn a degree in history and Latin American studies from Colgate University in Hamilton, NY Upon completion of her degree, Kristin came to the University of Florida where she earned a Ma ster of Arts in Latin American studies. Kristin then began her doctoral studies in the Department of Criminology at the University of Florida. She lives with her husband, Thom as and their daughter Susan. 180