<%BANNER%>

Women and crime in context

University of Florida Institutional Repository

PAGE 1

WOMEN AND CRIME IN CONTEXT: EXAMINING THE LINKAGES BETWEEN STRUCTURAL CONDITIONS AND FEMALE OFFENDING WITHIN THE CONTEXT OF PLACE By STEPHANIE ANN HAYS A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS UNIVERSITY OF FLORIDA 2005

PAGE 2

Copyright 2005 by Stephanie Ann Hays

PAGE 3

This document is dedicated to my parents, Larry and Glenda Hays.

PAGE 4

ACKNOWLEDGMENTS I would like to thank my entire committee, Dr. Karen Parker, Dr. Eve Brank, Dr. Lonn Lanza-Kaduce, and Dr. Lawrence Winner for all of their help and support in completing this thesis. In addition, I would like to thank Larry for being a social science friendly statistician and agreeing to be the minor representative, Lonn for being a great department chair who is willing to help with anything and everything and for exchanging tales of Iowa, and Eve for putting up with me for the past two years as a TA and for being a great mentor and friend. Most importantly, I have to thank Karen for being the best committee chair that I could ever ask for, for keeping me on schedule with endless amounts of encouragement and patience, and for all of her support along the way not only as my chair, but also as a mentor, colleague, friend, and even family on some holidays. I also need to thank my parents, Larry and Glenda Hays, and my brother Bryan for supporting me in all of my decisions throughout the years and for spending numerous family vacations either touring colleges across the country or helping me move. iv

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................iv LIST OF TABLES ...........................................................................................................viii LIST OF FIGURES .............................................................................................................x ABSTRACT .......................................................................................................................xi CHAPTER 1 INTRODUCTION........................................................................................................1 Importance of Gender...................................................................................................1 Importance of Place......................................................................................................2 Research Questions.......................................................................................................5 Summary.......................................................................................................................5 2 STRUCTURAL THEORIES: GENDER AND PLACE..............................................6 Feminist Perspective.....................................................................................................6 What Is Patriarchy?...............................................................................................6 Patriarchy and Crime.............................................................................................8 Empirical Findings................................................................................................9 Summary..............................................................................................................13 Social Disorganization Theory...................................................................................13 Empirical Findings..............................................................................................15 Summary..............................................................................................................18 Conclusion..................................................................................................................18 3 WOMEN AND CRIME IN CONTEXT....................................................................20 Patriarchy and Gender................................................................................................20 Social Disorganization and Place...............................................................................22 Summary.....................................................................................................................28 4 DATA AND METHODOLOGY...............................................................................31 Source of the Data......................................................................................................31 Unit of Analysis..........................................................................................................31 v

PAGE 6

Measures.....................................................................................................................33 Dependent Variables...........................................................................................33 Independent Variables.........................................................................................33 Private patriarchy.........................................................................................34 Public patriarchy (economic marginalization).............................................36 Social disorganization..................................................................................39 Controls........................................................................................................41 Methodology...............................................................................................................43 Principal Component Analysis............................................................................46 Private patriarchy.........................................................................................47 Public patriarchy (economic marginalization).............................................47 Variance Inflation................................................................................................49 Analytical Plan............................................................................................................50 5 RESULTS...................................................................................................................56 Violent Crime.............................................................................................................56 Private Patriarchy................................................................................................57 Public Patriarchy.................................................................................................57 Social Disorganization.........................................................................................58 Controls...............................................................................................................59 Property Crime............................................................................................................60 Private Patriarchy................................................................................................61 Public Patriarchy.................................................................................................61 Social Disorganization.........................................................................................62 Controls...............................................................................................................63 Comparing Regression Coefficients...........................................................................63 Violent Crime......................................................................................................64 Property Crime....................................................................................................65 Summary.....................................................................................................................65 6 DISCUSSION AND CONCLUSION........................................................................71 Discussion...................................................................................................................71 Limitations and Future Research................................................................................73 Conclusion..................................................................................................................74 APPENDIX A ABBREVIATIONS AND DEFINITIONS.................................................................75 B RURAL-URBAN CONTIUUM CODES (BEALE CODES)....................................83 C SOUTHERN STATES...............................................................................................84 D WESTERN STATES..................................................................................................85 E CORRELATIONS BEFORE PRINCIPAL COMPONENT ANALYSIS.................86 vi

PAGE 7

F CORRELATIONS AFTER PRINCIPAL COMPONENT ANALYSIS....................93 G VARIANCE INFLATION FACTOR VALUES........................................................96 H URBAN VIOLENT CRIME MODELS.....................................................................97 I URBAN PROPERTY CRIME MODELS..................................................................99 LIST OF REFERENCES.................................................................................................101 BIOGRAPHICAL SKETCH...........................................................................................109 vii

PAGE 8

LIST OF TABLES Table page 2-1 Private and Public Patriarchy.....................................................................................7 4-1 Means and Standard Deviations of Variables before Principal Component Analysis....................................................................................................................53 4-2 Principal Component Factor Matrices after Rotation Total Sample.....................54 4-3 Means and Standard Deviations of Variables after Principal Component Analysis....................................................................................................................55 5-1 Negative Binomial and ZINB Regression Coefficients, (Z Scores), and Standardized Errors for Female Violent Crime Arrest Models, 2001......................67 5-2 Negative Binomial and ZINB Regression Coefficients, (Z Scores), and Standardized Errors for Female Property Crime Arrest Models, 2001....................69 B-1 2003 Beale Codes.....................................................................................................83 B-2 1993 Beale Codes.....................................................................................................83 E-1 Correlation Matrix of Variables before Principal Component Analysis Total Sample......................................................................................................................87 E-2 Correlation Matrix for Variables before Principal Component Analysis Urban Sample......................................................................................................................89 E-3 Correlation Matrix for Variables before Principal Component Analysis Rural Sample......................................................................................................................91 F-1 Correlation Matrix for Variables after Principal Component Analysis Total Sample......................................................................................................................93 F-2 Correlation Matrix for Variables after Principal Component Analysis Urban Sample......................................................................................................................94 F-3 Correlation Matrix for Variables after Principal Component Analysis Rural Sample......................................................................................................................95 viii

PAGE 9

G-1 Variance Inflation Factor Values for Independent Variables Included in Female Violent Crime and Property Crime Arrest Models, 2001........................................96 H-1 Negative Binomial Regression Coefficients, (Z Scores), and Standardized Errors for Urban Female Violent Crime Arrest Models, 2001...........................................97 I-1 Negative Binomial Regression Coefficients, (Z Scores), and Standardized Errors for Urban Female Property Crime Arrest Models....................................................99 ix

PAGE 10

LIST OF FIGURES Figure page 2-1 Relationship between Patriarchy and Crime..............................................................8 2-2 Burgesss Zones for City Growth.............................................................................14 2-3 Basic Theoretical Model of Social Disorganization Theory....................................15 3-1 Proposed Relationship between Patriarchy, Social Disorganization, and Crime.....30 4-1 Simplified Variable Relationship Model..................................................................52 x

PAGE 11

Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Arts WOMEN AND CRIME IN CONTEXT: EXAMINING THE LINKAGES BETWEEN STRUCTURAL CONDITIONS AND FEMALE OFFENDING WITHIN THE CONTEXT OF PLACE By Stephanie Ann Hays May 2005 Chair: Karen F. Parker Major Department: Department of Criminology, Law and Society This study examines the association between social structural conditions and female offending in urban and rural areas. Although studies of female offending have increased in recent years, the majority of this research has focused on individual or situational characteristics to the exclusion of structural predictors. In fact, few empirical studies have examined the structural correlates of female offending. This study draws upon social disorganization and feminist theories to explore the linkages between structural conditions and female offending within the context of place. Using 2000 census data and 2001 UCR arrest data, Poisson-based techniques are used to investigate the influence of structural predictors on female arrests for both violent and property offenses within urban cities and rural towns. Furthermore, this examination allows for the investigation of whether these structural factors exhibit similar influences on female arrest patterns across urban and rural areas. xi

PAGE 12

CHAPTER 1 INTRODUCTION This study examines the impact of social structure on female arrests in both urban and rural areas. More specifically, the current study explores the impact of patriarchy and social disorganization on female arrests for both violent and property offenses. It is the first study to examine the impact of patriarchy on female offending at the macro-level, and is also the first study to use towns as the unit of analysis when applying social disorganization theory to rural crime. This introductory chapter outlines the importance of examining both gender and place within criminological research. Importance of Gender The most prevalent concern of feminists in criminology is that all of the major criminological theories were developed without concern for gender and without even considering female offending (Chesney-Lind, 1989; Daly & Chesney-Lind, 1988). Females account for about 22% of all arrests, 11% of murder arrests (Greenfield & Snell, 1999) and their incarceration rates are growing at a faster rate than males (Beck, 2000). Despite this, however, research on female offending has been limited. While research on females and crime has increased since the second wave of feminism, the majority of the research on female offending has been qualitative or at the individual level (see Acoca, 1998; Bloom, Owen, Rosenbaum, & Deschenes, 2003; Kaker, Friedman, & Peck, 2002; Koons-Witt & Schram, 2003). In addition, while some have suggested that there are different pathways to crime for females than males; most of 1

PAGE 13

2 this research has focused on the link between victimization and female offending (see English, Widom, & Brandford, 2001; McClellan, Farabee, & Crouch 1997; Siegel & Williams, 2003; Widom, 1989; Widom & Maxfield, 2001). Furthermore, when female offending is examined it is still the exceptional case such as battered woman syndrome (Comack, 1999). So long as women are recognized only as victims and not as active agents, there is little need to embrace or integrate feminist analyses into the criminological agenda (Comack, 1999, p. 165). Gender has not been properly investigated in criminology and remains in the margins of criminology because of the focus on victimization, and such a focus has neglected that men are also victims and that women are capable of engaging in violent behavior (Comack, 1999). Very few studies have actually examined female offending at the macro-level with one notable exception being the work by Steffensmeier and Haynie (2000). They examined the structural disadvantage correlates of female and male arrest rates for both violent and property offenses in 178 cities. Their overall findings were that the structural variables associated with predicting male arrest rates can also predict female arrest rates just not as strongly. In addition, they found that the levels of offending were higher in cities with increased levels of disadvantage, but that the effect was greater for violent crimes. Overall there is a lack of gender specific studies in criminology at the macro-level and those at the micro-level primarily focus on female victimization. Importance of Place Much like female offending, rural crime has also largely been ignored by the field even though 25% of the U.S. population lives in rural areas with populations less than 2,500 (Weisheit & Donnermeyer, 2000). While research on rural crime is growing, most

PAGE 14

3 of the research has focused on drug use (see Diala, Muntaner, & Walrath, 2004; Donnermeyer, Barclay, & Jobes, 2002; Weisheit & Fuller, 2004), domestic violence (see Davis, Taylor, & Furniss, 2001; Krishnan, Hilbert, & VanLeeuwen, 2001; Websdale, 1998; Websdale & Johnson, 1998), and community policing (see Jobes, 2003, 2002; Liederbach & Frank, 2003; OShea, 1999;Weisheit, Wells, & Falcone, 1994) At the same time, research regarding structural theories of crime has primarily focused on urban areas. It is unknown whether such theories can explain rural crime. Such macro-level theories need to be examined in rural areas to ensure that they are general theories of crime and not urban specific theories because theories that cannot account for both rural and urban circumstances are limited in scope (Weisheit & Donnermeyer, 2000, p. 310-311) (see Appendix A for urban/rural definitions). Recently in criminology there has been an increasing interest in contextual studies. Place is one example of a context and rural areas may be a different context than urban areas. Cebulak (2004) claims the context of rural crime, its causes and its characteristics, are so different than for urban crime, we need a separate set of theories to account for rural crime and justice (p. 72). Some types of crime, such as theft of farm animals or equipment or wildlife crimes, are limited to only rural areas (Cebulak, 2004; Weisheit & Wells, 1999). Furthermore, there are unique features of rural communities that may influence rural crime such as: physical distance and isolation, more informal social control, low mobility and density, higher density of acquaintanceship, mistrust of government, reluctance to seek outside assistance, and factory farms and processing plants (Weisheit & Donnermeyer, 2000; Weisheit & Wells, 1999). For more on the effect of processing plants on rural communities and crime see Broadway (1990).

PAGE 15

4 Contrary to common thought, rural crime and deviance is an issue. While violent crime in large cities has declined since the early 1990s, rural crime rates have been increasing (Weisheit & Donnermeyer, 2000). Between 1991 and 1997 urban violent crime rates decreased by 531.8 per 100,000 but rural rates increased by 37.9 per 100,000 (Weisheit & Donnermeyer, 2000). DUIs are more common in rural areas, and rural youth are more likely to use cigarettes or smokeless tobacco (Weisheit & Donnermeyer, 2000). In addition over the past 20 years, rural youth alcohol use has matched or exceeded that of urban youth, and nonmetropolitan 12 th graders in 1995 had higher use rates for crack cocaine, stimulants, barbiturates, and tranquilizers than their metropolitan counterparts (Weisheit & Donnermeyer, 2000)(See Appendix A for metropolitan and nonmetropolitan definitions). Drug manufacturing, particularly for methamphetamine, is of concern in rural areas. Missouri had more methamphetamine lab seizures than any other state in 1997, with most of the seizures occurring in rural areas (Weisheit & Donnermeyer, 2000), and there were 300 times more methamphetamine lab seizures in Iowa in 1999 than in New York and New Jersey combined (Eagan, 2002). Suicide rates, by any method, are also higher in rural areas (Butterfield, 2005). In the most rural counties, the incidence of suicide with a gun is greater than the incidence of murder with guns in major cities (Butterfield, 2005). From 1989 to 1999, the risk of dying from a gunshot was the same in rural and urban areas the difference was who pulled the trigger (Butterfield, 2005). Crime in rural America is important, but even though contextual studies in criminology have increased, rural areas are still being neglected.

PAGE 16

5 Research Questions The primary research questions this study seeks to answer are as follows: 1. What are the impacts of structural conditions on female arrests? a. What is the impact of patriarchy on female arrests for violent and property crime? b. What is the impact of social disorganization on female arrests for violent and property crime? 2. Does the impact of these structural conditions on female arrests differ in urban and rural areas? Summary This study fills a gap in the existing literature. The purpose of the current study is to examine the impacts of structural conditions on female arrests within the context of place. More specifically, this study examines the contribution of feminist and social disorganization perspectives to addressing female violent and property crime arrests in both urban and rural areas. While this introductory chapter has primarily focused on the importance of examining gender and place, the following chapter discusses the literature and empirical findings in regards to patriarchy and crime. Chapter 2 also presents a theoretical overview of social disorganization theory as well as empirical findings. Chapter 3 provides an overview on the integration of the feminist perspective with social disorganization and outlines the hypotheses. Chapter 4 presents the data and methodology. Chapter 5 discusses the results, and the final chapter provides an overall discussion and conclusions.

PAGE 17

CHAPTER 2 STRUCTURAL THEORIES: GENDER AND PLACE Feminist Perspective A feminist perspective acknowledges that there are gender inequalities (Belknap, 2001) and places gender at the center of analysis (Atwell, 2002; Hartsock, 1998). According to Daly and Chesney-Lind (1988) there are five elements of feminist thought that separate it from more general types of thought. First, gender is a complex social, historical, and cultural product. Second, gender and gender relations order social life and institutions. Third, gender relations and constructs are based upon the superiority and dominance of men over women. Fourth, our systems of knowledge reflect mens view of the world and finally, women should be at the center of intellectual inquiry. Patriarchy is key to feminist thought. Socialist feminism claims that both class and patriarchy are a dual system of domination that explain womens subordination and radical feminism claims that patriarchy is central to explaining womens position in society (Belknap, 2001). What Is Patriarchy? While scholars agree that gender inequality cannot be truly understood without some understanding of patriarchy (Walby, 1986), there are many definitions of patriarchy throughout the literature. Walby defines patriarchy as a system of social structures, and practices in which men dominate, oppress and exploit women (1990, p. 20; 1989, p. 214). She suggests that patriarchy is composed of six main structures: the patriarchal 6

PAGE 18

7 mode of production in the household, patriarchal relations in paid work, patriarchal relations in the state, male violence, patriarchal relations in sexuality, and patriarchal relations in cultural institutions. Table 2-1 illustrates that the six structures are present in both public and private patriarchy. It is generally agreed upon that patriarchy exists in both a public and a private sphere (Atwell, 2002). The public sphere includes institutional structures such as the government, schools, and churches while the private sphere encompasses the home and family. Table 2-1. Private and Public Patriarchy Private Patriarchy Public Patriarchy Dominant Structure Household Production Employment and the State Other Patriarchal Structures Employment State Sexuality Violence Culture Household Production Sexuality Violence Culture Adapted from Theorizing Patriarchy by Sylvia Walby (1990, p. 24) Similarly, Messerschmidt (1986) defines patriarchy as a set of social relations of power in which the male gender appropriates the labor power of women and controls their sexuality (p. x). Patriarchy is a system of hierarchy where men dominate women. Men control and dominate women in the labor force, in the home, and in economic, religious, political, and military systems. Messerschmidt, however, is specifically concerned with explaining the relationship between patriarchy and crime. His work has been incorporated into the criminological literature.

PAGE 19

8 Patriarchy and Crime According to Messerschmidt (1986) and Simpson (1991), the differences in offending between males and females are due to the gendered social organization of production and reproduction. It is because of this that powerless men engage in violent crime while powerless women engage in property crime. It is therefore necessary to examine measures of both economic inequality in the public (production/work) sphere and patriarchal structures in the private (family/reproduction) sphere and the subsequent effects on crime. Figure 2-1 provides a basic conceptual model of the relationship between patriarchy and crime. Private Patriarchy Figure 2-1. Relationship between Patriarchy and Crime According to Messerschmidt (1986), female offending is a form of resistance and accommodation to the oppressed status of women. He suggests that females are less likely to engage in serious crime because their subordination and powerlessness isolates them from engaging in such crimes. In addition, females have less opportunity to engage in serious crime. He claims, instead, that women are more likely to engage in property crime because of their oppressed economic status. For instance, Messerschmidt (1986) suggests that female teenagers shoplift because they need certain items in order to please Public Patriarchy (production/work) (family/reproduction) Female Crime

PAGE 20

9 men. Larceny and theft by female teenagers is a result of a combination of their marginalized economic position, gender-role socialization, and control over female-sexuality. He goes on to suggest that crime is also necessary for survival among adult women due to female unemployment and the feminization of poverty. Many women turn to gender-specific illegal options, such as fraud, to put food on the table (Messerschmidt, 1986, p. 84). Overall, women engage in nonviolent offenses merely as a means to accommodate their powerless position in society. Empirical Findings Literature examining the impact of patriarchy or gender inequality on female crime has largely focused on the public sphere by using the economic marginalization hypothesis. Economic marginalization refers to the economic disadvantage of women relative to that of men (Heimer, 2000). The general idea is that as the disadvantage of females increases, their involvement in crime will also increase. It proposes that the increased hardship of women relative to men is the reason why the gender gap has been narrowing in the past few decades. Only a few studies have directly tested the economic marginalization hypothesis and they have found limited support for the hypothesis (Box & Hale, 1983, 1984; Hunnicutt & Broidy, 2004; Steffensmeier & Streifel, 1992). Box and Hale (1983, 1984) found limited support for the economic marginalization hypothesis. They examined female conviction rates from 1951-1979 in England and Wales and found that female unemployment was associated with conviction rates for both violent and property offenses but that female unemployment was better at explaining property convictions. Heimer (2000), however, criticized their studies for only

PAGE 21

10 looking at one measure of economic well being and for using variables that measured womens absolute poverty rather than their well-being compared to that of men. Steffensmeier and Streifel (1992) examined U.S. arrest trends from 1960-1985 and found that female arrest was correlated with policing trends and with economic marginalization. Their primary measure of economic marginalization was single-parent female headed households. They found the percent of female headed households to be significant and positively related to major property crime and with burglary and prostitution. Again Heimer (2000) criticized their study for using only limited, absolute measures. They did list the female unemployment rate relative to males as one of their variables; however, they did not discuss any findings in their paper in regards to such a measure. More recently, Hunnicutt and Broidy (2004) examined 1975-1994 adult conviction rates by gender per 100,000 in ten countries (Austria, Denmark, Hungary, Greece, Panama, Portugal, Italy, Chile, Sweden, United Kingdom, and United States) and also found some support for economic marginalization. Contrary to expectations, they found no support for unemployment being related to conviction rates; however, they did find female employment in the service sector to be positively related to female conviction. They also found divorce to be positively associated with female conviction rates. While they originally included divorce as a measure of the liberation hypothesis, they themselves suggest that it may in fact be measuring economic marginalization. Like the previous studies, though, all the measures used were absolute measures and did not look at womens status relative to that of mens.

PAGE 22

11 Few other attempts have been made to try to conceptualize patriarchy at the macro-level. Most studies that look at patriarchy and crime do so at the individual level. One exception is the work by Yllo (1983, 1984) and Yllo and Straus (1984) that examined the relationship between the status of women at the state level and violence against wives. In all three studies data were aggregated at the state level (N = 30). Gender equality was measured on four different dimensions economical, educational, political, and legal. The first study (1983) found that violence against wives was highest in the least egalitarian states as well as the most egalitarian states. In the second study (1984) she found the highest levels of violence to be in male-dominated families in the more egalitarian states. In the third study, Yllo and Straus (1984) again found that violence against wives was highest in the least egalitarian and the most egalitarian states, that more patriarchal norms coincided with more violence, and that violence was highest when there was inconsistency between womens public status and their status within marriage. Like most feminist pieces, however, the focus of these studies was on male offending and female victimization. Subsequently, female offending was neglected. Whaley and Messner (2002) examined gender equality and homicide rates in 191 cities. They used five gender inequality indicators ratio of male to female median income, the percent of males ages 16 and above who were employed in the civilian labor force relative to the percent of females, the percent of executives, managers, and administrators who were male, the percent of those employed in the labor force who were male, and the ratio of men to women ages 25 and above with four or more years of college. They also included sex-specific measures of economic disadvantage percentage black, percentage poor, percentage unemployed, and the Gini index. Gender

PAGE 23

12 equality had no effect on females killing males or on females killing females while the index of economic disadvantage was positively related to females killing both males and females. Several other studies have also examined the effects of gender inequality on offending while not specifically taking a feminist approach. Most notably, Steffensmeier and Haynie (2000) examined the effects of structural disadvantage on female and male arrest rates from 1987-1993 in 178 cities. Some of the independent variables included were female poverty, female joblessness, and female-headed households. Again, these were all absolute measures but they did find that levels of offending were higher for both genders in cities with higher levels of social and economic disadvantage. Furthermore, the structural variables were robust for predicting male and female rates; however, the effect was greater for serious crimes. In addition, they found that structural disadvantage was more strongly related to partner homicides by females than by males. DeWees and Parker (2003) examined the relationship between womens economic and social status and homicide in 162 cities using six gender specific indicators: the percent of females employed in managerial and professional occupations, the percent of females working part-time, female median income, percent of females living below poverty, the percent of females with a bachelors degree or higher, and the percentage of women married. Female poverty was positively significant in the total, family, and acquaintance homicide models. Marriage reduced homicide in the total, intimate, and acquaintance models but was not significant in the family model. Part-time employment was found to reduce female offending in the total and intimate models, but the index of inequality had no significant effect in any of the models. They did find,

PAGE 24

13 however, that the indicators were better at explaining female offending in non-southern cities than in the South. Again, however, there is the concern that the research only used absolute measures of womens status. Summary While scholars distinguish between private and public patriarchy, most of the literature examining the relationship between patriarchy and crime at the macro-level has focused on public patriarchy, particularly economic marginalization. The relationship between private patriarchy and crime has not been tested at the macro-level. While there is some support for the economic marginalization hypothesis, only a few studies have directly tested it and these studies have only used limited, absolute measures of womens status instead of measures that capture womens status relative to that of men. Some other studies have examined the relationship between public patriarchy and crime; however, the majority of this research has focused on patriarchy and female victimization (see Yllo, 1984, 1983; Yllo & Straus, 1984). While DeWees and Parker (2003) and Steffensmeier and Haynie (2000) found gender inequality to be related to female offending, they too only used absolute measures of womens status. Overall, there is some support for the notion that female disadvantage, primarily female poverty as an absolute measure, is related to female offending. Similarly, social disorganization theory also examines the relationship between poor economic status and crime. Social Disorganization Theory In the past 20 years there has been a renewed interest in social disorganization theory. Social disorganization theory has its roots in Chicago. Park and Burgess (1925)

PAGE 25

14 emphasized the importance of looking at natural areas and their characteristics. Areas are considered to be functioning and changing organisms. Burgess (1925) suggested that cities expand and grow outward in concentric circles from the core of the city zone one. Zone two is the transition zone. It is generally the oldest and poorest zone. The third zone consists of workers homes. The fourth zone is the residential zone and consists of single family housing, and finally the fifth zone is the commuter zone or suburbs. Invasion is constantly occurring in all of these zones as new people move to the city and current residents try to move outward into a different zone. The zones are displayed in Figure 2-2. Zone 1 C ity C o re Zone 2 Transiti on Figure 2-2. Burgesss Zones for City Growth In Juvenile Delinquency and Urban Areas, Shaw and McKay (1942) examined juvenile delinquency in Chicago. Their main argument was that poor economic status, population heterogeneity, and residential mobility lead to social disorganization. Social disorganization in turn leads to breakdowns in conventional attachment and informal and Zone 3 Workers Zone 4 Resident ial Zone 5 C o mmut e r

PAGE 26

15 formal control and therefore results in more crime. More recently, it has been suggested that family disruption (Sampson, 1987; Sampson & Groves, 1989) and population size and density (Mayhew & Levinger, 1976) also contribute to social disorganization. The key theoretical concepts of social disorganization theory are displayed in Figure 2-3. Several studies have examined social disorganization theory and have found support for the theory in urban areas (Kposowa, Breault, & Harrison, 1995; Lee, Maume, & Ousey, 2003; Petee & Kowalski, 1993; Sampson, 1991; Sampson & Groves, 1989). These studies are discussed below. Low Economic Figure 2-3. Basic Theoretical Model of Social Disorganization Theory Empirical Findings Sampson and colleagues have conducted several studies on social disorganization theory. Sampson (1987) examined the relationship between family disruption and crime. Using 1980 homicide and robbery rates for 171 cities, he found family disruption, region, Residential Mobility Family Disruption Population Size / Density Population Heterogeneity Status Social Less Attachment Disorganization And Control Crime

PAGE 27

16 and density to be related to juvenile robbery, and family disruption, region, population size, and housing density were related to adult robbery. Family disruption was also found to be related to adult homicide. Furthermore, family disruption and population size were also related to black juvenile homicide, and region and city size were associated with black adult homicide. Sampson and Groves (1989) expanded the research on social disorganization theory. Unlike previous studies, they measured the levels of social organization within communities besides using the standard structural indicators of social disorganization. Based on data from the 1982 British Crime Survey (N =10,905 individuals, N = 238 localities), they found family disruption, urbanization, and ethnic heterogeneity to be associated with more disorderly peer groups. Urbanization was negatively associated with friendship networks, and socioeconomic status was positively related to organizational participation. Further analysis revealed that the structural measures of social disorganization were mediated by the community organization measures. In other words, community social disorganization accounts for much of the effect of socioeconomic status, residential stability, family disruption, and ethnic heterogeneity on crime. In an extension of the previous study, Sampson (1991) analyzed data from the 1984 British Crime Survey (N = 11,030 individuals, N = 526 polling districts) and found that residential stability has a direct effect on community-based social ties which in turn increases the level of community cohesion. In yet another study, Sampson, Raudenbush, and Earls (1997) tested the notion that concentrated disadvantage and residential instability decrease collective efficacy and whether in turn, collective efficacy explains the relationship between neighborhood

PAGE 28

17 disadvantage and crime. They interviewed 8,782 residents of 343 neighborhood clusters in Chicago. The results supported their hypothesis. The effects of concentrated disadvantage and residential instability on violence were mediated in a large part by collective efficacy. Consistent with social disorganization theory, Kposowa, Breault, and Harrison (1995) examined crime in counties with a population larger than 100,000 (N = 408) and found poverty to be a significant predictor of violent crime, and church membership and divorce were significant predictors of property crime. When examining crime in all counties (N = 3,076), the strongest predictor of property crime was urbanity, but percent black, percent Hispanic, population change, and unemployment were also significant. Predictors of violent crime in the total county sample were percent black, percent Hispanic, church membership, urbanity, and population density. When examining homicide in the large counties (N = 408), they found percent black, Gini coefficient, divorce rate, and population change to be the strongest predictors. Contrary to social disorganization theory though, they found poverty to be significantly related to homicide but in the negative direction. Lee, Maume, and Ousey (2003) also examined the relationship between socioeconomic disadvantage and poverty on the homicide rate average from 1990-1992 in 778 metropolitan counties. Unlike Kposowa and colleagues (1995), Lee and colleagues (2003) did find poverty concentration to be significant and positively related to homicide. Disadvantage was also found to be significant and positively related to homicide. Land, McCall, and Cohen (1990) also examined the structural correlates of homicide rates across time 1960, 1970, and 1980 and across space cities, SMSAs

PAGE 29

18 (see MSA in Appendix A), and states. While not specifically testing social disorganization theory, they did find support for the indicators of social disorganization. Resource deprivation was associated with higher homicide rates across all the time periods and locations. In addition, population structure an index of population size and density and percent divorced were also related to homicide across most of the models. Summary Overall, while there is much support for social disorganization theory, there are numerous inconsistencies in terms of which social disorganization indicators are significant and for which types of offenses. Furthermore, some studies have findings opposite of what social disorganization theory would suggest. For instance when looking at the predictors of homicide in metropolitan counties, Kposowa and colleagues (1995) found poverty to be significant but in the negative direction; however, Lee and colleagues (2003) found poverty concentration to have a significant positive effect on homicide in metropolitan counties. Conclusion This chapter has discussed the feminist perspective in regards to patriarchy and crime as well as social disorganization theory and the empirical studies available on both. While there is support for both theories, the feminist perspective has only examined the relationship between public patriarchy and crime at the macro-level, and social disorganization theory has focused on and was developed around urban areas. It is unknown how adequate social disorganization theory is at explaining crime in nonurban areas. The following chapter discusses the need to examine the influence of both private

PAGE 30

19 and public patriarchy as well as the context of place (urban and rural) on female offending.

PAGE 31

CHAPTER 3 WOMEN AND CRIME IN CONTEXT It is necessary to examine both gender and place when investigating female offending. Gender is a fundamental element of criminology that cannot be ignored (Comack, 1999), and patriarchy is essential for capturing the the depth, pervasiveness, and interconnectedness of different aspects of womens subordination (Walby, 1990, p. 2). Likewise, place is also a fundamental aspect of criminology, and it is past time for criminological theories and methods to include the rural context (Weisheit & Wells, 1996, p. 384). In addition, the structures of patriarchy manifest themselves differently in rural and urban areas (Websdale, 1998). For instance, rural women are more involved in household production and less involved in the public sphere, and there is also a more traditional division of labor in rural areas (Websdale, 1998). Patriarchy and Gender The current study is the first research to attempt to measure private patriarchy at the macro-level. In an effort to build upon the feminist literature, the current study includes multiple measures of the feminist perspective. Traditional, absolute measures of female poverty and part-time work are used to measure public patriarchy (economic marginalization). In addition, measures of womens position relative to that of men in the workforce are also used to measure public patriarchy (economic marginalization) in order to address Heimers (2000) concern about only using absolute measure to conceptualize economic marginalization. Such measures allow for us to capture both the economic 20

PAGE 32

21 marginalization of women and public patriarchy in the workforce of male dominance over females in the labor force and market. To date, no studies have examined the influence of private patriarchy on female offending at the structural level. While the liberation hypothesis originally proposed that female equality would lead to an increase in female crime (see Adler, 1975; Simon, 1975), it has essentially been discounted (Belknap, 2001; Heimer, 2000). Similarly, power-control theory also suggests that female juvenile delinquency will increase in more egalitarian families (see Hagan, Gillis, & Simpson, 1985; Hagan, Simpson, & Gillis, 1988, 1987; McCarthy, Hagan, & Woodward, 1999); however, there is mixed and inconsistent support for the theory (Belknap, 2001) and some suggest it is merely the liberation hypothesis reworded. Messerschmidt (1986) suggests that females who are confined and disadvantaged in the home are more likely to hurt themselves instead of hurting others. Because of this, I propose that private patriarchy will have no significant effect on female violent arrests in urban or rural areas. In terms of property crime, Messerschmidt (1986) also argues that females engage in property crime because of capitalism and patriarchy and that in order to curb crime we need to reduce power and class and gender inequality. Therefore, I predict that indicators of private patriarchy will have a positive effect on female arrests for property crime in both urban and rural areas. H 1 : Indicators of private patriarchy (family/reproduction sphere) will have no effect on female violent crime arrests and will have a positive effect on female property crime arrests in both urban and rural areas.

PAGE 33

22 Based on the economic marginalization hypothesis, I predict that indicators of public patriarchy will have a positive impact on female arrests for violent and property crime. As indicated in the previous chapter, some studies have found support for the economic marginalization hypothesis. For example, Box and Hale (1983, 1984) found economic marginalization (female unemployment) to be positively associated with female violent and property offenses, and Steffensmeier and Streifel (1992) found economic marginalization (percent female headed households) to be positively related to female arrests for major property crime, burglary, and prostitution. Other studies have also found a relationship between female structural disadvantage and crime. DeWees and Parker (2003) found female poverty to be related to homicide, Whaley and Messner (2002) found female economic disadvantage to be related to females killing, and Steffensmeier and Haynie (2000) found female structural disadvantage to be related to both violent and property offenses. H 2 : Indicators of public patriarchy (economic marginalization), patriarchy in the work and production spheres, will have a positive effect on female arrest for both violent and property crime in urban and rural areas. Social Disorganization and Place Most research on social disorganization has focused on urban areas; however, social disorganization theory can also be applied to rural areas. The structural conditions usually associated with social disorganization are not unique to urban areas. In fact, nonmetropolitan poverty rates have exceeded metropolitan poverty rates every year since poverty was first officially measured in the 1960s (Rural poverty, 2004). In addition, the

PAGE 34

23 2003 unemployment rate was 5.8% in nonmetropolitan areas and 6.0% in metropolitan areas (Rural America, 2004). Albrecht and colleagues (2000) examined how poverty levels in rural America have been affected by industrial transformation. Using 1990 census data on 2,390 nonmetropolitan counties, they argue that Wilsons model for the inner city underclass can be used to understand increased levels of rural poverty and the growth of the rural underclass (for more on rural ghettos see Davidson, 1996). Population mobility is also an issue in rural areas. Over 1,000 nonmetropolitan counties have lost population since 2000, primarily counties in the Great Plains, but there are also fast growing nonmetropolitan recreational counties in the South and West and the growth of the Hispanic population has contributed to nonmetropolitan county population growth in the West, South, and Midwest (Rural America, 2004). Population turnaround in rural counties has been linked to a variety of social problems including problems in education, community solidarity, heath care, social welfare, and crime (Price & Clay, 1980). In addition, minorities comprise 17% of nonmetropolitan residents and 15% of nonmetropolitan families are headed by a single female (Jolliffe, 2003). Snyder and McLaughlin (2004) found that poverty in nonmetropolitan areas closely resembles that in central cities. The risk of poverty for female-headed families and subfamilies with children is significantly higher for those living in nonmetropolitan areas compared to those living in central cities and suburban areas. A few empirical studies of social disorganization in rural areas have provided some support for social disorganization theory (Barnett & Mencken, 2002; Kposowa, Breault, & Harrison, 1995; Lee, Maume, & Ousey, 2003; Osgood & Chambers, 2000;

PAGE 35

24 Petee & Kowalski, 1993). Petee and Kowalski (1993) tested social disorganization theory on violent crime rates in 630 rural counties from 1979-1980. They found residential mobility to have the greatest impact on violent crime followed by single-parent households, and then racial heterogeneity. Barnett and Mencken (2002) applied social disorganization theory to violent and property crime rates circa 1990 in 2,254 nonmetropolitan counties. They found resource disadvantage to have a significant positive effect on violent crime, and population change was significant and positively related to property crime. While Kposowa, Breault, and Harrison (1995) did not specifically test social disorganization theory, they did incorporate standard measures of social disorganization when examining the structural correlates of crime rural counties (N = 1,681). They found population change to be significant and positively related to both violent and property crime. Church membership, divorce rate, and percent Native American were also found to be significant predictors of violent crime, and percent Hispanic was found to be a significant predictor of property crime. Contrary to social disorganization theory though, poverty was not found to be a significant predictor of violent or property crime. They did, however, find poverty, the South, divorce rate, and population change all to be significant when only examining homicide in the small counties. Lee, Maume, and Ousey (2003) however found poverty to not be significant when examining the average homicide rate from 1990-1992 in 1,746 nonmetropolitan counties. They did find a significant and positive relationship between disadvantage and homicide though. Osgood and Chambers (2000) also found no meaningful relationships between indicators of economic status, poverty, or unemployment on the juvenile violent crime

PAGE 36

25 arrest rate in 264 rural counties. Consistent with social disorganization theory they did find residential instability, female headed households, and ethnic heterogeneity to be associated with juvenile arrests. Jobes, Barclay, Weinand, and Donnermeyer (2004) also found none of their economic measures to be significant when examining crime rates in 123 rural LGAs in Australia. In support of social disorganization theory, though, they did find residential instability and family instability to be associated with higher rates of crime. Higher proportions of indigenous people were also associated with higher rates of assault and with break and enter crimes. Arthur (1991) examined the socioeconomic predictors of violent and property crime in 13 rural Georgia counties from 1975-1985. He found the percent of the population below poverty, percent families receiving aid, unemployment, and percent black to predict both violent and property crime. Contrary to most findings at the urban level, however, he found the variables to be better predictors of rural property crime than of rural violent crime. Based on the review of the literature, there are numerous inconsistencies in the findings when examining social disorganization in urban areas. Likewise, it appears these inconsistencies in the findings persist when applying social disorganization theory to rural areas, especially when examining economic disadvantage indicators. Barnett and Mencken (2002) found resource disadvantage to have a significant positive impact on violent crime. Kposowa and colleagues (1995), however, did not find poverty to be a significant predictor of violent crime arrests in rural areas, but they did find poverty to be predictive of homicide rates. When Lee and colleagues (2003) examined rural homicide

PAGE 37

26 rates, they found poverty to be nonsignificant. Osgood and Chambers (2000) also found no meaningful relationships between the indicators of economic status, poverty, or unemployment on juvenile arrests. Furthermore, all of the studies applying social disorganization theory to crime in rural areas have used nonmetropolitan counties as the unit of analysis. This is problematic when applying social disorganization theory because generally counties are comprised of numerous communities. In addition, all of the previous studies employ data before 2000. It is necessary to test social disorganization theory in rural areas using more recent data because in 2003 the OMB changed the definitions of metropolitan/nonmetropolitan and urban/rural, thereby changing the rural-urban continuum codes that these previous studies employed. Furthermore, the work by Lee and colleagues (2003) only examined homicide rates and the work by Osgood and Chambers (2000) only examined juvenile arrest rates from counties in only four states. Also, all of these studies examined either total crime or male crime. The current study addresses these limitations and advances upon these previous efforts by using towns as the unit of analysis, by examining both violent and property arrests for females, by using a national sample, and by using more recent data. Consistent with social disorganization theory, I propose that the measures of social disorganization will have a positive effect on female arrests for violent and property crime. As indicated in this chapter and the previous chapter, numerous studies have found support for social disorganization indicators in predicting crime in both urban cities and in nonmetropolitan counties. H 3 : Indicators of social disorganization will have a positive impact on female arrest for both violent and property crime in both rural and urban areas.

PAGE 38

27 While conditions of female subordination and structural disadvantage are not unique to urban cities, there may be differences in the ability of these indicators to predict female crime within the context of place. Wells and Weisheit (2004) examined the structural covariates of urban and rural crime. It is the only study that has specifically compared the regression coefficients of the structural correlates of metropolitan and nonmetropolitan crime. Overall, they found that the standard structural models for urban crime explain more variance in the urban property and violent crime models than in the rural models. Population change and family instability were significant across all models while racial diversity was consistent across only the models for violent crime. On the other hand, economic resources index was very inconsistent across models. It was inversely related to both property and violent crime in urban counties but in nonmetropolitan rural counties it was unrelated to violent crime and positively related to property crime. When comparing regression coefficients for the urban and nonmteropolitan rural counties, they found the indicators of population change, household instability, cultural capital, economic resources, percent of population 15-24, and unemployment rate to be significantly different in predicting violent crime in large-city counties versus rural, small-town counties. In addition, they found the indicators of population change, urban density, economic stability, cultural capital, economic resources, and unemployment rate to be significantly different in predicting property crime in the large versus small counties. Overall, Wells and Weisheits (2004) research highlights the need to further examine rural crime to determine whether or not structural theories of crime can be applied across communities regardless of population size. Consistent with their findings,

PAGE 39

28 I propose that the impact of some of the structural indicators on female arrests for violent and property crimes will not be the same in urban cities and rural towns. H 4 : While it is expected that these structural indicators will have a positive effect on female crime in both urban and rural areas, the impact of these structural indicators on female arrests will not be the same across the two localities. Summary Both the feminist perspective and social disorganization theory contribute to explaining female offending (see Figure 3-1). The feminist perspective focuses on the influence of patriarchy on crime while social disorganization theory focuses on the ecological conditions that contribute to crime. Both, however, focus on the effects of disadvantage on crime. The feminist perspective suggests that female structural disadvantage relative to males in both the home and in the labor market contributes to female offending by making crime necessary in order for females to survive and overcome their disadvantaged positions. Therefore, concentrated disadvantage of females relative to males within an area leads to an increase in female crime. Similarly, social disorganization theory suggests that concentrated disadvantage (low economic status, ethnic heterogeneity, residential mobility, family disruption, and population size) within an area also leads to an increase in crime due to a breakdown in informal control. Therefore, female offending will also be higher in disorganized localities. Overall, the hypotheses reflect the primary concern that female disadvantage and female offending will be exasperated in areas that exhibit both patriarchy and social disorganization.

PAGE 40

29 By examining female offending by type of offense (property and violent) and by place (urban and rural), this study offers a systematic analysis that hopefully addresses the inconsistencies amongst previous studies. Overall, this study simultaneously addresses the importance of both gender and place. It is the first study to examine female offending by type of offense and by place. It is also the first study to assess the impact of patriarchy on female offending at the macro-level. This study builds upon the feminist literature by quantitatively analyzing the impact of both economic marginalization in the class sphere (public patriarchy) and patriarchy in the family sphere (private patriarchy) on female offending. Furthermore, this study builds upon the literature illustrating the importance of place by looking at both urban and rural areas. Instead of counties, urban cities and small, rural towns are compared as recommended by Wells and Weisheit (2004).

PAGE 41

30 Private Patriarchy Public Patriarchy Low Economic Status Figure 3-1. Proposed Relationship between Patriarchy, Social Disorganization, and Crime. All proposed relationships are positive, except that between the indicators of private patriarchy and female violent crime which is expected to not be significant Female Crime Family Disruption Population Size Controls Residential Mobility

PAGE 42

CHAPTER 4 DATA AND METHODOLOGY Source of the Data Data were obtained from multiple sources. UCR arrest data from 2001 were used to measure the dependent variables. Summary files one and three of the 2000 census were the sources of data for the independent variables (see Appendix A for more information on summary files one and three). The number of police officers employed by the cities and towns were obtained from the 2000 Police Employee (LEOKA) Data (ICPSR 3445). The Law Enforcement Agency Identifiers Crosswalk 2000 (ICPSR 4082) was used to merge the arrest data with the census data. In addition, the 1993 and 2003 Beale Codes (see Appendix B) were employed to help define the rural sample and to calculate a control measure. Unit of Analysis While most studies of social disorganization in urban areas use cities or census tracts as the unit of analysis, all of the studies that have applied social disorganization theory to rural crime have used counties as the unit of analysis. Using counties as the unit of analysis is potentially problematic when applying social disorganization theory because counties generally consist of numerous communities. The current study addresses this concern by using urban cities and rural towns as the unit of analysis instead. For this study, urban cities are defined as those cities with populations larger than 100,000. This is consistent with previous studies that have examined crime in urban 31

PAGE 43

32 areas. It is also consistent with the census in terms that in order for an area to be considered a MSA, it has to be an urbanized area with a population of at least 100,000 (see Appendix A for definitions of urban and urbanized areas). The urban sample consists of 200 cities with populations larger than 100,000. The cities range in size from 99,716 to 8,008,278 with an average population of 330,022 (SD = 677,267.9) and an average female population of 51% (SD = 1.23). Defining rural is more complicated. There is a debate on whether definitions based on population or those based on culture should be used. There is no consensus on a cultural definition of rural however. As a result population size is generally used. The census simply defines rural as territory, population and housing units not classified as urban (see Appendix A for definition of rural). In order to be considered urban by the census the area has to contain at least 2,500 people. Previous studies that have examined crime in rural areas have used counties as the unit of analysis so they have relied on the rural-urban continuum codes, otherwise known as the Beale codes. Because this study is using towns instead of counties, these codes cannot be fully utilized. However, consistent with the census definition and the logic of the Beale codes, for the purposes of this study rural towns are defined as those towns with a population less than 2,500 not located within a MSA (see MSA in Appendix A), and only such towns that are located within counties designated as nonmetropolitan by the 2003 Beale Codes (see Appendix B). As such, the rural sample consists of 787 towns. The rural towns range in size from 80 to 2,595 (M = 1450.69, SD = 588.234) with an average female population of 52.70% (SD = 3.66).

PAGE 44

33 Measures Dependent Variables The two dependent variables are the female violent crime count and the female property crime count. They consist of female arrests for each locality as reported in the 2001 UCR arrest data. The sample is limited to only municipal agencies that reported at least nine months of the year and to only those that were municipal agencies in cities over 100,000 or cities under 2,500 not located within a MSA (see MSA in Appendix A). The female violent crime count is an index of murder, robbery, and aggravated assault. The female property crime count is an index of burglary, larceny, forgery, and fraud. Forgery and fraud are included in lieu of the traditional index offenses of arson and motor vehicle theft due to the nature of female offending. Steffensmeier and colleagues have conducted numerous studies on female arrest trends and have found that female arrests have not increased for serious crimes but that female arrests have increased in terms of minor property offenses such as larceny, forgery, and fraud primarily due to changes in police practices, economic factors, and opportunities for females to engage in such crimes (Steffensmeier, 1980, 1993; Steffensmeier & Cobb, 1981). Similar arrest patterns have also been found when examining female arrest trends in rural areas (Steffensmeier & Jordan, 1978). Independent Variables Essentially both the feminist perspective and social disorganization theory are being tested in this study. As such, numerous measures of private patriarchy, public patriarchy, and social disorganization are utilized as independent variables. These variables are described below.

PAGE 45

34 Private patriarchy Private patriarchy (patriarchy in the family/reproduction sphere) is difficult to conceptualize at the town level with census data. No research has attempted to try to capture the impact of traditional family structure on female offending except for that of DeWees and Parker (2003) which included a measure of the percent of women married. Four measures of private patriarchy are included in this study. These measures are: the percent of families where the husband works and the wife does not, the percent of families married with children, the percent of females working full-time with no income, and the percent of employed females who are unpaid family workers. These variables are defined below. Percent of families where the husband works and the wife does not. The first measure of private patriarchy is the percent of families where the husband works and the wife does not. This measure is an attempt to capture traditional family structure. A wife not in the labor force includes housewives and wives only doing incidental unpaid family work. This measure is calculated by dividing the number of married couple families where the husband is in the labor force and the wife is either not in labor force or unemployed by the total number of married couple families. This is then multiplied by 100 to obtain a percent. The definitions for family, married couple family, not in the labor force, unemployed, and unpaid family work are provided in Appendix A. 100familiescouplemarriedNo.unemployedorforcelabortheinnotwifeandworkinghusbandwithfamiliescouplemarriedNo.

PAGE 46

35 Percent of families married with children. The second measure of private patriarchy is the percent of families that are married with children. This measure is also an attempt to conceptualize traditional family structure. It is calculated by dividing the number of married couple family households (where the householder is 15-64 years of age) with children under 18 years of age by the total number of family households (householder 15-64 years of age). This is then multiplied by 100 to obtain a percent. The definitions for household, household type, householder, and married couple family are provided in Appendix A. 100)householdsfamily of(No.age) of years 18underchildren with householdsfamily marriedof(No. Percent of females working full-time with no income. Another measure of private patriarchy is the percent of females working full-time with no income. This measure is another attempt to capture traditional family structure by capturing women working in the home without income. It is calculated by dividing the number of females ages 15 and over who worked full-time, year round in 1999 with no income divided by the number of females ages 15 and over that worked full-time, year round in 1999. Appendix A provides the census definitions of income, worked in 1999, and full-time, year-round workers. 100 1999)in roundyear time,full worked whoage of years 15 females of (No.income) no with 1999in roundyear fulltime, worked whoage of years 15 females of (No.

PAGE 47

36 Percent of employed females who are unpaid family workers. A final measure of private patriarchy is the percent of employed females who are unpaid family workers. This is yet another attempt to capture the impact of traditional family structure. Unpaid family workers include people who worked 15 hours or more without pay in a business or on a farm operated by a relative. It is hoped that this variable will capture those females that are working for family members for free, especially in rural areas where it may be more likely that females are working on family farms without pay. Refer to Appendix A for definitions of employed and unpaid family workers. 100 age) of years 16 femalescivilian employed No.(kers)family wor unpaid are whoage of years 16 femalescivilian employed (No. Public patriarchy (economic marginalization) Consistent with the works by Steffensmeier and Haynie (2000) and Whaley and Messner (2002), measures of public patriarchy (patriarchy in the work and production sphere) include the ratio of male to female median income, the ratio of males to females ages 25 and above with a bachelors degree or higher, the ratio of males to females ages 16 and above in management and professional occupations, the percent of females ages 16 and above working part-time (1-34 hours per week), and the percent of females living below poverty. Such measures capture economic marginalization but also patriarchy in the class sphere in terms of male dominance. It is important to use ratios and not just absolute measures because the economic marginalization hypothesis is based upon the status of females relative to males (Heimer, 2000). These measures are defined below.

PAGE 48

37 Ratio of male to female median income. One measure of public patriarchy is the ratio of male to female median income. This is consistent with the work by Whaley and Messner (2002). The ratio is calculated by dividing the median male income by the median female income for 1999. The greater the value of the ratio is, the greater the disadvantage of females relative to males. The census definitions for income and median income are provided in Appendix A. 1999)in income with age of years 15 population female for the income(Median 1999)in income with age of years 15 population male for the income(Median Ratio of males to females with bachelors degrees or higher. A second indicator of womens status relative to that of men is the ratio of males to females with a bachelors degree or higher. Consistent with Whaley and Messner (2002), the ratio is calculated as the percent of men ages 25 and over with a bachelors degree or higher divided by the percent of females ages 25 and over with a bachelors degree or higher. higheror degree sbachelor' a with age of years 25 females ofPercent ( higher)or degree sbachelor' a with age of years 25 males ofPercent ( Ratio of males to females in management and professional occupations. Another measure of public patriarchy is the ratio of males to females in management and professional occupations. In order to measure womens status in the workforce, Whaley and Messner (2002) use the percent of executives, managers, and administrators who are male, and DeWees and Parker (2003) use the proportion of females ages 16 and over

PAGE 49

38 employed in management and professional occupations. Both of these measures, however, are absolute measures. This study employs a ratio instead. The ratio is calculated as the percent of civilian employed males ages 16 and over in management, professional, and related occupations divided by the percent of civilian employed females ages 16 and over in such occupations. Appendix A provides the census definition of occupation. s)occupation alprofession andmanagment in age of years 16 females employedcivilian of %(s)occupation alprofession and managementin age of years 16 males employedcivilian of (% Percent of females working part-time. The fourth indicator of public patriarchy is the percent of females working part-time. Consistent with DeWees and Parker (2003) this percentage is based on the number of females ages 16 and over who worked 1-34 hours per week in 1999 divided by the total number of females ages 16 and over that worked in 1999. The result is then multiplied by 100 to obtain a percent. Refer to Appendix A for the census definition of worked in 1999. 100 1999)in worked whoage of years 16 females No.(1999)in per week hours 34-1 worked whoage of years 16 females No.( Percent of females living below poverty. DeWees and Parker (2003), Steffensmeier and Haynie (2000), and Whaley and Messner (2002) all utilized female poverty as an indicator of womens status. For the current study, the percent of females living below poverty is based on the number of females for whom poverty status was

PAGE 50

39 determined in 1999 with incomes below the poverty line divided by the total population for whom poverty status was determined in 1999. The result is then multiplied by 100 to obtain the percent. It was decided to divide by the total population rather than the female population to better capture womens status relative to that of men. Appendix A provides the census definitions of poverty and individuals for whom poverty status is determined. 100 1999)in eddeterermin wasstatuspoverty for whom population Total()levelpoverty below incomes with 1999in determined wasstatuspoverty for whom population (Female Social disorganization Standard measures of social disorganization are used in this study. They include: residential mobility, population change, percent divorced, and population size. The total percent of the population living below poverty, while standard in studies of social disorganization, is not included here because it is highly collinear with the public patriarchy (economic marginalization) measure of female poverty which is more important since the study is taking a gendered approach. Likewise, percent black is not being used as a measure of social disorganization because it too is highly correlated with female poverty. The measures of social disorganization utilized in this study are defined below. Residential mobility. The first measure of social disorganization is residential mobility. Consistent with numerous other studies residential mobility is calculated by dividing the number of people ages 5 and over that lived in a different house in 1995 by

PAGE 51

40 the total number of people ages 5 and over. This is then multiplied by 100 to obtain the percent. 100 age) of years 5 Population(1995)in housedifferent ain lived that age of years 5n (Populatio Population change. Population change is also included as a standard measure of social disorganization. Population change is calculated by subtracting the 1990 population from the 2000 population and then dividing by the 1990 population. The result is then multiplied by 100 to obtain the percent. 100 )population (1990)population 1990 population 2000( Percent divorced. The percent divorced is another standard measure of social disorganization. Divorced refers to those people who are legally divorced and have not remarried. For this study it is calculated as the number of people ages 15 and over who are divorced divided by the number of people ages 15 and over. The result is then multiplied by 100 to obtain a percent. 100 age) of years 15 people No.(divorced) are that age of years 15 people No.(

PAGE 52

41 Population size. Population size is simply measured as the natural logarithm transformation of the 2000 population. )population (2000LN Controls Control measures include whether the county changed from metropolitan to nonmetropolitan or vice versa between 1993 and 2003, the Hispanic population, South, West, and officer rate. The control variables are defined below. County changed. Two dummy variables were created using the 1993 and 2003 Beale Codes to control for if the city or town is located within a county that changed from metropolitan to nonmetropolitan or vice versa between 1993 and 2003. The Beale Codes are provided in Appendix B. It is important to include this measure as a control because in 2003 the OMB released the Census 2000 version of the rural urban continuum codes otherwise known as the Beale Codes. This new version is not fully compatible with the 1993 codes due to OMB and Census Bureau rule changes in the way urban and rural are measured. Nonmetropolitan counties are now divided into micropolitan and noncore. Outlying counties are now considered metropolitan based solely on whether 25% of the workforce commutes into the metropolitan area whereas before outlying counties were considered metropolitan based on if 15% of the workforce commuted and based on population density, urbanization, and population growth. Due to the OMB rule changes approximately 45 metropolitan counties became nonmetropolitan. In addition,

PAGE 53

42 298 nonmetropolitan counties became metropolitan partially from suburbanization but also partially from the rule changes (Economic Resource Service, 2003). The control is unnecessary in the current study however. None of the urban cities or rural towns are located within counties that changed. Percent Hispanic. Controlling for the Hispanic population is consistent with the work of Steffensmeier and Haynie (2000) and with the work of Kposowa and colleagues (1995). It is important to control for the Hispanic population because Hispanics account for 25% of the nonmetropolitan population growth between 1990 and 2000. In addition, around 90% of all nonmetropolitan counties experienced Hispanic population growth in the 1990s, and the Hispanic population is growing faster than all other ethnic and racial groups in rural America thus helping to offset rural population loss (Kandel & Newman, 2004). This measure is computed as the 2000 Hispanic and Latino population divided by the 2000 total population. This is then multiplied by 100 to obtain the percent and then undergoes natural logarithm transformation. See Appendix A for the census definition of Hispanic and Latino. 100 )population (2000 )Population Latino and Hispanic (2000 LN South. A dummy variable for the South was also created to control for regional variation because the study employs a national sample (see Appendix C for list of southern states). This is consistent with previous studies and is especially important for

PAGE 54

43 this study because the rural south has the highest and most persistent poverty rates (Rural poverty, 2004). West. A dummy variable for the West was also created (see Appendix D for list of western states). This is consistent with Steffensmeier & Haynie (2000) and is important for this study because of the fast growing nonmetropolitan recreational counties in the West and South (Rural America, 2004). Officer rate. Controlling for officer rate is consistent with the works of Steffensmeier and Haynie (2000) and DeWees and Parker (2003). Officer data was obtained from the 2000 LEOKA data (ICPSR #3445). The officer rate is calculated by dividing the total number of officers by the 2000 total population. This result is then multiplied by 1,000 to obtain a rate per 1,000. 1000 )population 2000(officers) police ofnumber (Total Methodology The means and standard deviations for the variables are presented in Table 4-1. When examining the indicators of private patriarchy, the percent of families where the husband works and the wife does not has a mean of 24.263 (SD = 3.781) in the urban sample and a mean of 21.744 (SD = 6.480) in the rural sample. The percent of families married with children has a mean of 26.884 (SD = 8.350) in the urban sample and a mean of 27.357 (SD = 7.121) in the rural sample. The urban sample has a mean of .0623 (SD =

PAGE 55

44 .067) in regards to the percent of females working full-time without income, whereas the rural sample mean for this indicator is .113 (SD = .612). The percent of employed females who are unpaid family workers has a mean of .304 (SD = .145) in the urban sample and a mean of .4156 (SD = .699) in the rural sample. In regards to the measures of public patriarchy, the mean for the ratio of male to female median income is 1.582 (SD = .187) in the urban sample and is 1.856 (SD = .367) in the rural sample. The ratio of males to females with a bachelors degree or higher has a mean of 1.153 (SD = .095) in the urban sample and a mean of 1.148 (SD = .472) in the rural sample. The mean for the ratio of males to females in management and professional occupations is .895 (SD = .137) in the urban sample and .694 (SD = .281) in the rural sample. For the percent of females working part-time, the urban sample has a mean of 27.823 (SD = 4.960) and the rural sample has a mean of 28.881 (SD = 8.174). The percent of females living below poverty has a mean of 8.443 (SD = 3.617) in the urban sample and a mean of 10.758 (SD = 5.145) in the rural sample. For the indicators of social disorganization, the urban sample has a mean residential mobility of 52.342 (SD = 6.423) whereas the rural sample has a mean of 42.477 (SD = 8.773). Population change has a mean of 18.643 (SD = 32.714) in the urban sample and a mean of 9.510 (SD = 45.013) in the rural sample. The mean for the percent of the population divorced is 10.596 (SD = 2.093) in the urban sample and 11.159 (SD = 3.249) in the rural sample. The natural log of the population size has a mean of 12.251 (SD = .746) in the urban sample and a mean of 7.168 (SD = .529) in the rural sample. Also reported in Table 4-1 are the t-test values to show whether the variable means are significantly different between the urban and rural samples. All of the

PAGE 56

45 indicators have significantly different means between the two samples except for the percent of married families with children and the ratio of males to females with a bachelors degree or higher. Several of the variable means are significantly higher in the urban sample but several are also significantly higher in the rural sample. For instance, when looking at the measures of private patriarchy, the percent of families where the husband works and the wife does not is significantly higher in the urban cities (t = 7.129, p < .01), while the percent of females working full-time without income (t = -2.265, p < .05) and the percent of employed females who are unpaid family workers (t = -4.149, p < .01) are significantly higher in the rural sample. The same occurs when looking at the measures of public patriarchy. The ratio of males to females in professional and management occupation is significantly higher in urban cities (t = 14.390, p < .01), while the ratio of male to female income (t = -14.706, p < .01), the percent of females working part time (t = -2.320, p < .05), and the percent of females living below poverty (t = -7.356, p < .01) are all significantly higher in the rural sample. When looking at the indicators of social disorganization, residential mobility (t = 17.889, p < .01), population change (t = 2.688, p < .01), and the natural log of the population size (t = 90.705, p < .01) are significantly higher in the urban sample, while percent divorced is significantly higher in rural towns (t = -2.994, p <.01). As for the control variables, percent Hispanic (t = 18.547, p < .01) and the West (t = 7.052, p < .01) are significantly higher in the urban sample whereas South (t = -8.749, p < .01) and officer rate (t = -10.306, p < .01) are significantly higher in the rural sample. These findings indicate that disadvantage is not unique to urban areas and further justify the need to explore female offending within the context of place.

PAGE 57

46 The bivariate correlation matrices are presented in Appendix E. Multicollinearity is a problem if two or more determinants are highly correlated. Generally, if the correlation is .5 or greater multicollinearity could be a problem but there is no definitive rule for this. When examining the correlation matrix for the urban sample, there are noticeably high correlations between some of the indicators of patriarchy. For instance, the bivariate correlation between the percent of married families with children and the percent of families where the husband works and the wife does not is .623. Furthermore, the percent of married families with children is only correlated with female violent crime at -.188 and female property crime at -.189, and the percent of families where the husband works and the wife does not is only correlated with female violent crime at .017 and female property crime at .019. In other words, the correlation is much stronger between the two independent variables than it is between these two independent variables and the dependent variables. The same applies for the correlations between the following variables in the urban sample: the percent of females working full-time with no income and the percent of employed females who are unpaid family workers (.536); the ratio of male to female median income and the ratio of males to females with a bachelors degree or higher (.523); and the ratio of males to females with a bachelors degree or higher and the ratio of males to females in management and professional occupations (.523). One way to deal with this problem is to employ principal component analysis. Principal Component Analysis Principal component analysis was employed on the measures of patriarchy due to the problem of multicollinearity between those measures. The results from the principal

PAGE 58

47 component analysis are presented in Table 4-2. The analyses after varimax rotation yielded four components in each of the models. Private patriarchy Traditional family index. The ratio of male to female median income, the percent of families where the husband works and the wife does not, and the percent of families married with children are loading together in the models. While the ratio of male to female median income was originally conceptualized as a measure of public patriarchy, it is logical that it is loading with these two private patriarchal measures instead. As the percentage of traditional families where the wife is not working increases, the discrepancy between male and female income should also increase, as well as the percent of families married with children. Family unpaid work index: The second dimensional component index includes two private patriarchy measures the percent of females working full-time with no income and the percent of employed females who are unpaid family workers. As the percent of females working full-time without income increases, the percent of employed females who are unpaid family workers also increases. Public patriarchy (economic marginalization) Gender inequality index. The third component includes the ratio of males to females ages 25 and above with a bachelors degree or higher and the ratio of males to females ages 16 and above in management and professional occupations. The ratio of

PAGE 59

48 males to females with a bachelors degree or higher increases as the ratio of males to females in management and professional occupations increases. Percent of females working part-time. While the percent of females working part-time is loading in the gender inequality index in the total sample, it is only at .502. It was decided therefore to let the variable stand alone, especially since it loads alone in the urban sample and loads in the traditional index at only .5 in the rural sample. Percent of females below poverty. The final component is the percent of females who are living below poverty. This measure is loading alone in the total sample. The simplified measurement model is presented in Figure 4-1. The correlation matrices for the simplified models are presented in Appendix F. The matrices illustrate that the high correlations are reduced between the individual measures of patriarchy. This is especially the case for the urban sample; however, the family unpaid work index and the traditional family index still have a correlation of .515, and the gender inequality index and the percent of females living below poverty are correlated at -.506. The means and standard deviations for the simplified models, as well as the t-test results for comparing the urban and rural means are presented in Table 4-3. I will only be discussing the results for the three indexes of patriarchy here, as the means and t-test values for the other indicators are the same as those originally presented in Table 4-1. The urban sample has a mean of 33.103 (SD = 7.056) for the traditional family index whereas the rural sample has a mean of 32.075 (SD = 6.610). For the family unpaid work

PAGE 60

49 index, the urban sample has a mean of .298 (SD = .154) and the rural sample has a mean of .430 (SD = .873). The gender inequality index has a mean of 1.465 (SD = .147) in the urban sample and a mean of 1.312 (SD = .436) in the rural sample. Again there are significant differences between the index means in the urban and rural samples. While the traditional family index is higher in the urban sample (t = 1.863, p < .10), the family unpaid work index is higher in the rural sample (t = -3.999, p < .01). In addition, the gender inequality index is higher in the urban sample (t = 7.985, p < .01). Variance Inflation Since the social disorganization measures and the control variables were excluded from the principal component analysis, it is necessary to determine if multicollinearity is going to be a problem in the regression models. In order to address this issue, the variance inflation factors were calculated for the variables included in the models. Variance inflation factors indicate how much the variance of a coefficient is increased due to collinearity (Ott & Longnecker, 2001, p. 652). The variance inflation factor is designated by VIF = 1 / (1 R 2 ) where R 2 refers to how much of the variation in one independent variable is explained by the others (Ott & Longnecker, 2001, p. 709). It is generally accepted that a variance inflation factor greater than four indicates multicollinearity (Fisher & Mason, 1981, p. 109); however, some suggest that a value of 10 or greater indicates multicollinearity (Pindyck & Rubinfeld, 1998; Ott & Longnecker, 2001, p. 652). Appendix G displays the variance inflation factors for the independent variables included in the models. One of the variables does have a variance inflation factor greater than four. The traditional family

PAGE 61

50 index has a variance inflation factor of 4.327 in the urban sample. It appears that the traditional family index may be multicollinear with officer rate (VIF = 3.722). Therefore, multiple urban models will be run with and without these two measures to determine if multicollinearity between these two variables is problematic in the regression models. Analytical Plan Poisson based regression will be used in the following chapter to examine the structural correlates of female offending in urban and rural areas. Low arrest counts are very common in the data since the study is looking at female counts. Poisson based techniques account for the problem of the relatively low arrest counts (Osgood, 2000). Negative binomial regression models will be used because they allow for overdispersion (Gardner et al, 1995; Osgood, 2000). Count data often have overdispersion, with the variance exceeding the mean (Agresti, 2002). Poisson forces the variance to equal the mean whereas the negative binomial distribution has )/( )Var( )(2k where is called the dispersion parameter. Comparing the sample mean and variance of the dependent count variable provides a simple indication of overdispersion (Cameron & Trivedi, 1998). The means and standard deviations for the dependent variables are provided in Table 4-1 and 4-3. The dependent variables in this study do have overdispersion. For the violent crime count, the urban sample has a mean of 211 and a variance of 256,682, and the rural sample has a mean of .5319 and a variance of 1.337. The same is true for property crime. The urban sample has a mean of 794.11 and a variance of 1,463,247, and the rural sample has a mean of 3.04 and a variance of 35.858. k/1

PAGE 62

51 The negative binomial model, however, usually under predicts the amount of zero counts in the dependent variable. In the current study, the rural sample has 548 ( 71%) zero counts for female violent crime arrests and 304 ( 39%) zero counts for female property crime arrests. Due to the large number of zero counts, zero-inflated negative binomial regression models will also be employed. Zero-inflated models account for overdispersion due to excess zero counts (Cameron & Trivedi, 1998; Min & Agresti, 2002). Since the zero-inflated Poisson model assumes that the mean equals the variance, zero-inflated negative binomial models are more likely to be appropriate (Min & Agresti, 2002). While uncommon in criminology (for an example see Robinson, 2003), zero-inflated negative binomial regression is common in public health research (for an example see Chin & Quddus, 2003).

PAGE 63

52 Figure 4-1. Simplified Variable Relationship Model. All proposed relationships are positive, except that between the indicators of private patriarchy and female violent crime which is expected to not be significant. Traditional Family Index Family Unpaid Work Index Gender Inequality Index % Females Working Par t -Time % Females Living B e l ow P ove r ty Population Change South Residential Mobility % Divorced Population Size (log) Female Violent and Property Crime % Hispanic (log) West Officer Rate Private Patriarch y Public Patriarch y Social Disor g anization Controls

PAGE 64

53 Table 4-1. Means and Standard Deviations of Variables before Principal Component Analysis Urban Rural ^ p < .10 p < .05 ** p < .01 Mean SD Mean SD t Female Violent Crime 211.085 506.638 .5319 1.157 Female Property Crime 794.110 1209.647 3.037 5.988 Private Patriarchy % Families Husband Works, Wife Doesnt 24.263 3.781 21.744 6.480 7.129** % Married Families W/ Children 26.884 8.350 27.357 7.121 -.735 % Females Working FT, No Income .0623 .067 .113 .612 -2.265* % Employed Females Unpaid Family Workers .304 .145 .4156 .699 -4.149** Public Patriarchy Ratio M/F Median Income 1.582 .187 1.856 .367 -14.706** Ratio M/F Bachelors Degree or Higher 1.153 .095 1.148 .472 .249 Ratio M/F Professional Occupations .895 .137 .694 .281 14.390** % Females Working PT 27.823 4.960 28.881 8.174 -2.320* % Females Below Poverty 8.443 3.617 10.758 5.145 -7.356** Social Disorganization Residential Mobility 52.342 6.423 42.477 8.773 17.889** Population Change 18.643 32.714 9.510 45.013 2.688** % Population Divorced 10.596 2.093 11.159 3.249 -2.994** Population Size (log) 12.251 .746 7.168 .529 90.705** Controls % Hispanic (log) 2.521 1.136 .773 1.381 18.547** South .290 .455 .6099 .488 -8.749** West .450 .499 .183 .387 7.052** Officer Rate 2.002 .959 3.513 3.631 -10.306** Offset Female Population (log) 11.579 .747 6.525 .532 90.085**

PAGE 65

54 Table 4-2. Principal Component Factor Matrices after Rotation Total Sample 1 2 3 4 Ratio M/F Median Income 0.761 % Families Husband Works and Wife Doesnt 0.610 % Married Families W/ Children 0.636 Ratio M/F Bachelors Degree or Higher 0.708 Ratio M/F Professional Occupations 0.725 % Females Working PT 0.502 % Females Working FT with No Income 0.807 % Employed Females Unpaid Family Workers 0.815 % Females Below Poverty 0.886 Note: Only factor loadings greater than 0.500 are reported.

PAGE 66

55 Table 4-3. Means and Standard Deviations of Variables after Principal Component Analysis Urban Rural Mean SD Mean SD t Female Violent Crime 211.085 506.638 .5319 1.157 Female Property Crime 794.110 1209.647 3.037 5.988 Private Patriarchy Traditional Families Index 33.103 7.056 32.075 6.610 1.863^ Family Unpaid Work .298 .154 .430 .873 -3.999** Public Patriarchy Gender Inequality Index 1.465 .147 1.312 .436 7.985** % Females Working PT 27.823 4.960 28.881 8.174 -2.320* % Females Below Poverty 8.443 3.617 10.758 5.145 -7.356** Social Disorganization Residential Mobility 52.342 6.423 42.477 8.773 17.889** Population Change 18.643 32.714 9.510 45.013 2.688** % Population Divorced 10.596 2.093 11.159 3.249 -2.994** Population Size (log) 12.251 .746 7.168 .529 90.705** Controls % Hispanic (log) 2.521 1.136 .773 1.381 18.547** South .290 .455 .6099 .488 -8.749** West .450 .499 .183 .387 7.052** Officer Rate 2.002 .959 3.513 3.631 -10.306** Offset Female Population (log) 11.579 .747 6.525 .532 90.085** ^ p < .10 p < .05 ** p < .01

PAGE 67

CHAPTER 5 RESULTS This chapter presents the results from the regression models. The findings from both the negative binomial regression models and the zero-inflated negative binomial regression models are discussed below. In all the models, I offset for the female population thereby essentially creating a rate for the dependent variables. Violent Crime Table 5-1 presents the negative binomial and zero-inflated negative binomial regression coefficients, Z scores, and standard errors for the urban and rural violent crime models. Negative binomial regression provides the better fit for the urban sample because the urban sample has no zero counts. For the urban sample, the percent of females working part-time has a significant and negative effect on violent crime (Z = -2.69, p < .01). The percent of females living below poverty (Z = 5.18, p < .01) and West (Z = 2.21, p < .05) have a significant and positive effect on violent crime in the urban model. The urban models were also run with and without the traditional family index and the officer rate due to the problematic variance inflation factor. The results are presented in Appendix H and are similar to those discussed here. Zero-inflated negative binomial regression is a better fit for the rural violent crime model (Vuong = 28.10, p < .01). Female poverty (Z = 2.21, p < .05) and officer rate (Z = 2.55, p < .05) are significant and positively related to female violent crime arrests in rural towns. In addition, population size is significant and has a negative effect on female 56

PAGE 68

57 violent crime in the rural sample (Z = -3.81, p < .01). A more detailed discussion of these findings follows below. Private Patriarchy Neither of the private patriarchy indexes is significant. This is consistent with the first hypothesis that private patriarchy would not have a significant effect on violent crime. Women confined and disadvantaged in the home are more likely to engage in self-destructive behavior (such as alcohol and drug abuse or suicide) and hurt themselves when they are unable to cope instead of hurting others (Messerschmidt, 1986). Public Patriarchy The findings indicate mixed support for the second hypothesis that the indicators of public patriarchy (economic marginalization) have a positive effect on female violent arrests. The gender inequality index has no significant effect on female violent crime in the urban or rural samples. This is consistent with Whaley and Messner (2002) who also found gender inequality to have no effect on females killing males or females. Contrary to expectations, the percent of females working part-time is negatively related to female violent crime in urban cities. This is, however, consistent with DeWees and Parker (2003) who also found that part-time employment reduced female offending in their total and intimate homicide models. These findings do indicate some support for the economic marginalization hypothesis. The percent of females living below poverty is significant and has a positive effect on female violent crime arrests in both the urban and rural areas. This is consistent with previous studies. Whaley and Messner (2002) found economic disadvantage to be

PAGE 69

58 positively related to females killing males and females. Steffensmeier and Haynie (2000) found female disadvantage to be positively related to homicide, robbery and aggravated assault, and DeWees and Parker (2003) also found female poverty to be positively related to homicide. Social Disorganization Contrary to the third hypothesis, the findings indicate no support for the social disorganization indicators in predicting female violent crime in either urban or rural areas. Neither residential mobility nor population change is significant in any of the violent crime models. Kposowa and colleagues (1995) also found population change to not be a significant predictor of urban violent crime; however, they did find population change to be a significant predictor of homicide in the largest counties. My findings are contrary to several other studies on rural violent crime. Petee and Kowalski (1993) found residential instability to have the greatest impact on rural violent crime. In addition, Kposowa and colleagues (1995) and Wells and Weisheit (2004) found population change to be significant and positively related to rural violent crime, and Osgood and Chambers (2000) found residential instability to be related to rural violent crime. The percent of the population divorced also has no significant effect on either urban or rural violent crime. This is inconsistent with previous research as well. While Kposowa and colleagues (1995) also found divorce to not be a significant indicator of urban violent crime, Wells and Weisheit (2004) did find family instability to be positively related to urban violent crime. In addition, both Kposowa and colleagues (1995) and Wells and Weisheit (2004) found divorce to be positively related to rural violent crime.

PAGE 70

59 Also contrary to social disorganization theory, population size is not significantly related to female arrests for violent crime in the urban sample. This is also inconsistent with the findings of Sampson (1987) in which population size was related to juvenile robbery and black juvenile homicide, and contrary to Land and colleagues (1990) who also found population structure to be related to homicide. In addition, population size is significant but negatively related to rural violent crime which is opposite of what social disorganization theory would suggest; however, Jobes (1999) examined crime in rural Montana towns and also found that smaller towns had proportionally more crime. Controls As for the control measures, percent Hispanic is not a significant predictor of violent crime in either sample. Kposowa and colleagues (1995) also found percent Hispanic to not be a significant predictor of rural violent crime, but they did find it to be a significant predictor of urban violent crime. Wells and Weisheit (2004), however, found their cultural capital index, which included a measure of the Hispanic population, to have a significant and positive effect on violent crime in nonmetropolitan counties. The South is not a significant predictor of female violent crime in urban or rural areas. Kposowa and colleagues (1195) also found South to not be a significant predictor of urban violent crime, but the South was a significant predictor of homicide in small counties. I also find the West to be a significant predictor of urban violent crime, but not rural violent crime. This is consistent with Steffensmeier and Haynie (2000) who also found the West to be a significant and positive predictor of female robbery and burglary in large cities.

PAGE 71

60 Officer rate is not a significant predictor of urban violent crime. This is contrary to the work of Steffensmeier and Haynie (2000). They found officer rate to be significant and positively related to female homicide and robbery arrests. I do, however, find officer rate to be significant and have a positive effect on female violent crime arrests in rural towns. Property Crime Table 5-2 presents the negative binomial and zero-inflated negative binomial regression coefficients, Z scores, and standard errors for the urban and rural property crime models. For the urban sample, the percent of females working part-time (Z = 2.80, p < .01), percent divorced (Z = 3.63, p < .01), South (Z = 2.93, p < .01), and officer rate (Z = 2.25, p < .05) are all significant and positively related to property crime whereas the percent of females living below poverty is significant but negatively related to property crime (Z = -1.65, p < .10). The urban models were also run with and without the traditional family index and the officer rate due to the problematic variance inflation factor. The results are presented in Appendix I and are similar to those discussed here. Again, zero-inflated negative binomial regression models were also employed. Zero-inflated negative binomial regression is not a better fit for the urban property crime model since there are no zero counts in the urban sample. Zero-inflated negative binomial regression is, however, a better fit for the rural property crime model (Vuong = 29.28. p < .01). The traditional family index (Z = -2.21, p < .05) and the gender inequality index (Z = -2.32, p < .05) are significant but negatively related to rural property crime. The South (Z = 2.48, p < .05) and officer rate (Z = 3.42, p < .01) are significant and positively related to female arrests for property crime in rural towns. A more detailed discussion of these findings is presented below.

PAGE 72

61 Private Patriarchy Contrary to the first hypothesis, private patriarchy does not have a significant and positive effect on female property crime. The family unpaid work index is not significant in any of the models. Furthermore, the traditional family index is significant but negatively related to rural property crime. Recall that this index includes the percent of families married with children and the percent of families where the husband works and the wife does not. This finding may be indicating that there is a lack of opportunity for females in rural areas with a higher proportion of traditional families to engage in property crime such as forgery and fraud. Public Patriarchy There are mixed results for the second hypothesis that public patriarchy (economic marginalization) increases female property crime. Gender inequality does not have a significant effect on urban property crime. It does have a significant effect on rural property crime but in the negative direction. Again, while this is contrary to my expectations it may be in part due to the lack of opportunity to engage in such activity. The percent of females working part-time is significant and positively related to property crime in urban cities but is not significant in rural towns. Female poverty is not a significant predictor of rural property crime. It is, however, a significant predictor of urban property crime but contrary to expectations, it too is in the negative direction. This is also contrary to the work of Steffensmeier and Haynie (2000) which found female poverty to be related to female burglary and larceny rates.

PAGE 73

62 Social Disorganization There is also very little support for the third hypothesis that social disorganization variables will have a significant and positive effect on female property crime arrests. Residential mobility, population change, and population size are not significant in any of the property crime models. This is contrary to many other studies. Kposowa and colleagues (1995) found population change to be a significant predictor of property crime in both metropolitan and nonmetropolitan counties. Barnett and Mencken (2002) also found population change to have a significant and positive effect on property crime in nonmetropolitan counties. Jobes and colleagues (2004) found residential instability to be a significant predictor of break and enter crimes and malicious damage offenses in rural Australia. Furthermore, Wells and Weisheit (2004) also found their population change index to be significant and positively related to property crime in both metropolitan and rural counties. Consistent with social disorganization theory, the percent of the population divorced is significant and positively related to urban property crime. This is consistent with the findings of Kposowa and colleagues (1995) and Wells and Weisheit (2004). The percent divorced, however, does not have a significant effect on rural property crime. This is also consistent with the findings of Kposowa and colleagues (1995). It is contrary though to Wells and Weisheits (2004) findings on family instability being a significant predictor of property crime in rural counties, and is also contrary to the work by Jobes and colleagues (2004) which found family instability to be a significant predictor of malicious damage offenses.

PAGE 74

63 Controls As for the control variables, percent Hispanic is not a significant predictor of property crime in the urban or rural sample. This is consistent with Steffensmeier and Haynie (2000) but is contrary to Kposowa and colleagues (1995) who found percent Hispanic to be a significant predictor of property crime in both metropolitan and nonmetropolitan counties. The South is a significant and positive predictor of property crime in both urban cities and rural towns. This also is contrary to the findings of Kposowa et al (1995). My finding that the South is a predictor of property crime may be a result of strain. The South has the highest and most persistent rates of poverty (Rural poverty, 2004). The West is not a significant predictor of urban or rural property crime. This is contrary to Steffensmeier and Haynies (2000) finding that the West was a significant predictor of female burglary in large cities. Officer rate is a significant predictor of female property crime in both the urban and rural samples. This too is contrary to the findings of Steffensmeier and Haynie (2000). They found officer rate to not be a significant predictor of female property crime in large cities. Comparing Regression Coefficients Due to a lack of research on rural crime, it is unknown whether the traditional structural theories of crime that were developed around urban areas can also adequately explain and predict rural crime. In order to test the fourth hypothesis that the impact of the structural indicators will not be the same in urban and rural areas, the regression coefficients must be compared. The formula for comparing regression coefficients is 222121 )( SEbSEbbb

PAGE 75

64 where is the difference between the sample coefficients, and and are the coefficient variances associated with the first and second groups (Paternoster, Brame, Mazerolle, & Piquero, 1998). The Z scores for comparing the urban and rural regression coefficients are presented in Tables 5-1 and 5-2. Comparing the regression coefficients of the best fitting models is essentially the same as comparing the zero-inflated regression coefficients for both samples because the urban zero-inflated negative binomial regression models provide the same results as the urban negative binomial regression models. The results are discussed below. 21bb 21SEb 22SEb Violent Crime The Z scores for comparing the urban and rural regression coefficients for violent crime are presented in Table 5-1. When comparing the best fitting regression model coefficients, the coefficients for the percent of females living below poverty (Z = 3.325, p < .01) and for population size (Z = 3.337, p < .01) are significantly different. In other words, the percent of females living below poverty has a stronger, positive effect on violent crime in urban cities than in rural towns whereas population size has a significantly stronger negative effect on violent crime in rural towns than in urban cities. Wells and Weisheit (2004) are the only researchers who have previously compared the regression coefficients of the structural correlates of crime in urban versus rural areas. While they found their population change index to have a greater effect on rural violent crime, neither of the indicators of population change in the current study the percent of the population that changed from 1990 to 2000, or residential mobility are significantly different between the urban and rural samples. In addition, they also found their cultural

PAGE 76

65 capital index, which included a measure of the Hispanic population, had a greater effect on violent crime in rural counties than in large city counties. I, however, do not find percent Hispanic to have a significantly different effect between the two samples. Property Crime The Z scores for comparing the regression coefficients for the urban and rural property crime models are presented in Table 5-2. When comparing the regression coefficients for the best fitting urban and rural property crime models, the percent of females working part-time (Z = 2.231, p < .05) and the percent of the population divorced (Z = 2.652, p < .05) are significantly different. Again, contrary to Wells and Weisheit (2004), neither of the measures of population change in the current study has a significantly different effect on predicting urban versus rural property crime. In addition, while I find percent divorced to have a greater effect on property crime in urban cities than in rural towns, Wells and Weisheit did not find family instability to have a significantly different effect on property crime in urban and rural counties. Summary Overall, I find mixed support for the hypotheses presented in Chapter 3. Consistent with the first hypothesis, private patriarchy is not related to female violent crime. Contrary to the first hypothesis, though, private patriarchy is not significant and positively related to female property crime. Only the traditional family index is significant; however, it is negatively related to female property crime in the rural sample. In regards to the second hypothesis, I find mixed support for public patriarchy being positively related to either female violent or property crime. Some of the indicators are

PAGE 77

66 positively significant, however, some indicators of public patriarchy are not significant and some are negatively significant. In terms of the third hypothesis, I find no support for the indicators of social disorganization in predicting female violent crime in either the urban or rural samples. The only significant indicator is population size in the rural sample; however, it is negatively related to female violent crime. I also find no support for the social disorganization indicators in explaining female property crime in the rural sample, and the only significant indicator in the urban model is the percent of the population that is divorced. For the fourth hypothesis, I do find that some of the indicators are in fact significantly different between the urban and rural samples. Female poverty has a stronger, positive effect on female violent crime in the urban sample whereas population size has a stronger, negative effect on female violent crime in the rural sample. Likewise, the percent of females working part-time and the percent of the population divorced have a stronger, positive effect on female property crime in the urban sample.

PAGE 78

67 Table 5-1. Negative Binomial and ZINB Regression Coefficients, (Z Scores), and Standardized Errors for Female Violent Crime Arrest Models, 200 1 Urban Model Rural Model Z Private Patriarchy Traditional Families Index SE .004 (0.34) .013 -.006 (-0.68) .0095 .619 Family Unpaid Work SE .068 (0.18) .367 .009 (0.12) .075 .158 Public Patriarchy Gender Inequality Index SE -.519 (-1.36) .382 -.167 (-1.16) .144 -.862 % Females Working PT SE -.036** (-2.69) .013 -.014 (-1.54) .009 -1.391 % Females Below Poverty SE .098** (5.18) .019 .025* (2.21) .011 3.325** Social Disorganization Residential Mobility SE -.0006 (-0.06) .0096 .006 (0.86) .007 -.556 Population Change SE -.001 (-0.92) .002 .0005 (0.30) .002 -.530 % Population Divorced SE .021 (0.77) .028 -.012 (-0.65) .018 .991 Population Size (log) SE -.017 (-0.25) .067 -.541** (-3.81) .142 3.337**

PAGE 79

68 Table 5-1. Continued. Urban Model Rural Model Z Controls % Hispanic (log) SE -.073 (-1.18) .062 .011 (0.25) .047 -1.080 South SE -.169 (-1.21) .140 -.079 (-0.42) .190 -.381 West SE .341* (2.21) .154 .076 (0.35) .219 .990 Officer Rate SE .084 (0.93) .090 .075* (2.55) .029 .095 Constant SE -6.145** (-4.61) 1.331 -2.022^ (-1.65) 1.222 -2.280* Log Likelihood -1090.4095 -351.4 X2 ratio test, alpha = 0 8903.58** Pseudo R2 0.0336 Vuong Test 28.10** N = 199 N = 778 0 Obs = 548 ^ p < .10 p < .05 ** p < .01 Note: The urban model reports the negative binomial regression results. The rural model reports the zero-inflated negative binomial regression results.

PAGE 80

69 Table 5-2. Negative Binomial and ZINB Regression Coefficients, (Z Scores), and Standardized Errors for Female Property Crime Arrest Models, 2001. Urban Model Rural Model Z Private Patriarchy Traditional Families Index SE .002 (0.22) .011 -.018* (-2.21) .008 1.470 Family Unpaid Work SE .195 (0.67) .293 -.065 (-1.05) .062 .871 Public Patriarchy Gender Inequality Index SE -.343 (-1.00) .345 -.313* (-2.32) .135 -.078 % Females Working PT SE .033** (2.80) .012 .002 (0.28) .007 2.231* % Females Below Poverty SE -.026^ (-1.65) .016 -.005 (-0.49) .011 -1.082 Social Disorganization Residential Mobility SE -.0001 (-0.02) .008 .004 (0.63) .006 -.41 Population Change SE .001 (0.82) .002 .0005 (0.64) .001 .224 % Population Divorced SE .087** (3.63) .024 .009 (0.59) .017 2.652* Population Size (log) SE -.093 (-1.54) .060 -.059 (-0.49) .121 -.252

PAGE 81

70 Table 5-2. Continued Urban Model Rural Model Z Controls % Hispanic (log) SE .025 (0.45) .056 -.011 (-0.28) .039 .528 South SE .374** (2.93) .128 .348* (2.48) .140 .137 West SE .083 (0.66) .126 .266 (1.59) .167 -.875 Officer Rate SE .158* (2.25) .070 .069** (3.42) .020 1.223 Constant SE -5.919** (-5.37) 1.103 -4.464** (-4.59) .973 -.989 Log Likelihood -1391.2703 -1239.113 X2 ratio test, alpha = 0 2.9E+04** Pseudo R2 0.0133 Vuong 29.28** N = 199 N = 778 0 Obs = 304 ^ p < .10 p < .05 ** p < .01 Note: The urban model reports the negative binomial regression results. The rural model reports the zero-inflated negative binomial regression results.

PAGE 82

CHAPTER 6 DISCUSSION AND CONCLUSION The purpose of this study was to examine the association between social structural conditions and female offending in urban cities and rural towns. This final chapter outlines the findings presented in chapter five in regards to the research questions posed in chapter one. In addition, this chapter reviews possible limitations of the study as well as directions for future research. Discussion Research question 1a asked: What is the impact of patriarchy on female arrests for violent and property crime? I find that the indicators of private patriarchy have no effect on female violent crime which is consistent with the first hypothesis. Contrary to expectations, however, I also find that the indicators of private patriarchy have no effect on female property crime with one exception. The traditional family index is negatively related to female property crime in the rural sample. While this finding is also contrary to expectations, it may be better explained by lack of opportunity for females to engage in property crime in rural areas that have a high proportion of traditional families (i.e. married couple families with children under 18 and families where the husband works and the wife does not). In regards to the indicators of public patriarchy, I find mixed support for the economic marginalization hypothesis. Female poverty is significant and positively related to female violent crime in urban and rural areas. Contrary to expectations, the percent of females working part-time is negatively related to female violent crime arrests in the 71

PAGE 83

72 urban sample, but consistent with the hypothesis, it is positively related to female property crime arrests in the urban sample as well. Also contrary to expectations, the percent of females living below poverty is negatively related to female property crime arrests in the urban sample and gender inequality is negatively related to property crime in the rural sample. Again, this negative relationship between gender inequality and property crime in rural areas may be better explained by lack of opportunity for disadvantaged females to engage in property crime in rural areas. Research question 1b asked: What is the impact of social disorganization on female arrests for violent and property crime? Contrary to expectations, I find very little support for social disorganization theory in predicting female arrests. The only significant finding in favor of social disorganization theory is that the percent of the population divorced is positively related to female property crime arrests but only in the urban sample. The only other significant finding in terms of the indicators of social disorganization is opposite of what the theory would predict. Population size is negatively related to female violent crime arrests in the rural sample. The second research question posed in chapter one was: Does the impact of these structural conditions differ in urban and rural areas? The findings presented in chapter five illustrate that some of the indicators do have a significantly different impact on female arrests in urban and rural areas. Female poverty is a stronger predictor of urban violent crime than rural violent crime. Population size is a stronger, negative predictor of rural violent crime. In addition, the percent of females working part-time and the percent of the population divorced are both stronger predictors of urban property crime than rural property crime. These findings suggest that there are significant differences in the

PAGE 84

73 indicators of crime in urban and rural areas that need to be further explored. The standard structural predictors of urban crime may not be the best predictors of rural crime. Limitations and Future Research Like all studies, this study is not without limitations. There are problems with using official arrest data regardless, but the primary concern in this study is the validity of the rural arrest data. No studies have examined the validity of arrest statistics in rural jurisdictions (Osgood & Chambers, 2000). This is an issue that future research in rural crime must address. Furthermore, the rural sample is limited to only those small towns that have municipal police agencies. Perhaps this study is missing a large percentage of small towns that do not have their own agencies and have to rely on county and state agencies. A second concern is whether the variables employed are the best measures of the theoretical concepts. For instance, while percent divorced is included here as a standard measure of family disruption for social disorganization theory, some use percent divorced as an indicator of economic marginalization due to the feminization of poverty (Steffensmeier & Streifel, 1992) while others use it as an indicator of female liberation (Hunnicutt & Broidy, 2004). Future research needs to expand upon the variables examined, especially those of private patriarchy, since this is the first study to conceptualize private patriarchy at the city level and the first study to examine its impact on female offending across urban cities and rural towns. Subsequent research should also include measures of opportunity when trying to predict female offending especially when examining female offending within rural areas.

PAGE 85

74 Finally, this study only examines female arrests. In order to be a truly gendered approach, a study must examine and be able to explain both male and female offending (Steffensmeier & Allan, 1996). It would be valuable to examine male arrests as well because previous studies have indicated that male criminality is associated with economic disadvantage and Messerschmidt (1986) claims patriarchy contributes to males engaging in violent crime. With the recent interest in disaggregating, future research should also consider disaggregating by both gender and specific offense at the rural level since this has yet to be done. Conclusion Overall, while indicators of both patriarchy and social disorganization have significant impacts on female arrests, there are numerous inconsistencies in terms of the directions of these relationships, and also in terms of which indicators are significant by type of offense and by place. Furthermore, the impact of some of these indicators is significantly different between the urban and rural samples. My findings justify the need to further examine female offending as well as rural crime. Rural crime can no longer be ignored. Out of the 100 highest ranking counties on homicide, 71 are rural (Kposowa et al., 1995). The finding that population size is negatively related to female violent crime in the rural sample alone justifies the need to further explore the correlates of rural crime. Perhaps Donnermeyer (2004) is correct in suggesting that both social organization and disorganization may lead to crime in rural areas.

PAGE 86

APPENDIX A ABBREVIATIONS AND DEFINITIONS Employed Employed includes all civilians 16 years old and over who were either (1) "at work" -those who did any work at all during the reference week as paid employees, worked in their own business or profession, worked on their own farm, or worked 15 hours or more as unpaid workers on a family farm or in a family business; or (2) were "with a job but not at work" -those who did not work during the reference week but had jobs or businesses from which they were temporarily absent due to illness, bad weather, industrial dispute, vacation, or other personal reasons. Excluded from the employed are people whose only activity consisted of work around the house or unpaid volunteer work for religious, charitable, and similar organizations; also excluded are people on active duty in the United States Armed Forces. The reference week is the calendar week preceding the date on which the respondents completed their questionnaires or were interviewed. This week may not be the same for all respondents. (Census Glossary) Family A group of two or more people who reside together and who are related by birth, marriage, or adoption. Families may be a "Married Couple Family," "Single Parent Family," "Stepfamily," or "Subfamily." (Census Glossary) Full-time, year-round workers Full-time year-round, workers consists of people 16 years old and over who usually worked 35 hours or more per week for 50 to 52 weeks in 1999. The term worker in these concepts refers to people classified as 75

PAGE 87

76 Worked in 1999 as defined by the Census. The term worked in these concepts means worked one or more weeks in 1999 as defined by the census under Weeks Worked in 1999. (Census Glossary) Hispanic or Latino origin People who classify themselves in one of the specific Hispanic or Latino categories in the 2000 Census Mexican, Puerto Rican, Cuban, or Other Spanish, Hispanic, or Latino. Origin can be viewed as the heritage, nationality group, lineage, or country of birth or person or ancestors. People who classify themselves as such may be of any race. (Census Glossary) Household A household includes all the people who occupy a housing unit as their usual place of residence. (Census Glossary) Household type and relationship Households are classified by type according to the sex of the householder and the presence of relatives. Examples include: married-couple family; male householder, no wife present; female householder, no husband present; spouse (husband/wife); child; and other relatives. (Census Glossary) Householder The person, or one of the people, in whose name the home is owned, being bought, or rented. If there is no such person present, any household member 15 years old and over can serve as the householder for the purposes of the census. Two types of householders are distinguished: a family householder and a nonfamily householder. A family householder is a householder living with one or more people related to him or her by birth, marriage, or adoption. The householder and all people in the household related to him are family members. A

PAGE 88

77 nonfamily householder is a householder living alone or with nonrelatives only. (Census Glossary) Income Total income is the sum of the amounts reported separately for wages, salary, commissions, bonuses, or tips; self-employment income from own farm or nonfarm business; interest, dividends, net rental income, royalty income, or income from estates and trusts; Social Security or Railroad Retirement income; Supplemental Security Income; any public assistance or welfare payment from the state or local welfare office; retirement, survivor, or disability pensions; and any other sources of income received regularly such as Veterans payments, unemployment compensation, child support, or alimony. (Census Glossary) Individuals for whom poverty status is determined Poverty status was determined for all people except institutionalized people, people in military group quarters, people in college dormitories, and unrelated individuals under 15 years old. These groups also were excluded from the numerator and denominator when calculating poverty rates. They are considered neither poor nor nonpoor. (Census Glossary) Married-couple family This category includes a family in which the householder and his or her spouse are enumerated as members of the same household. (Census Glossary) Median income The median income for individuals is based on the individuals 15 years old and over with income. Median income is rounded to the nearest whole dollar. (Census Glossary)

PAGE 89

78 Metropolitan area (MA) A collective term, established by the federal Office of Management and Budget, to refer to metropolitan statistical areas, consolidated metropolitan statistical areas, and primary metropolitan statistical areas. The OMB revised the definitions in 2003. There are 1,089 metropolitan counties in the U.S. Under the new system metropolitan areas are defined for all urbanized areas regardless of the population. A metropolitan area is defined as (1) central counties with one or more urbanized areas, and (2) outlying counties that are economically tied to the core counties. Outlying counties are included if 25% of the workers commute to the central counties or the reverse is 25% of the central counties commute out. (Census Glossary and ERS) MSA Metropolitan statistical area. A geographic entity defined by the federal Office of Management and Budget for use by federal statistical agencies, based on the concept of a core area with a large population nucleus, plus adjacent communities having a high degree of economic and social integration with that core. Qualification of an MSA requires the presence of a city with 50,000 or more inhabitants, or the presence of an Urbanized Area (UA) and a total population of at least 100,000 (75,000 in New England). The county or counties containing the largest city and surrounding densely settled territory are central counties of the MSA. Additional outlying counties qualify to be included in the MSA by meeting certain other criteria of metropolitan character, such as a specified minimum population density or percentage of the population that is urban. Referred to as SMA beginning in 1949, changed to SMSA in 1959, and then changed to MSA in 1983. (Census Glossary)

PAGE 90

79 Nonmetropolitan The area and population not located in any metropolitan area (MA).Nonmetropolitan counties are outside the boundaries of the metropolitan area and are divided into two subtypes: micropolitan areas (centered on urban clusters of 10,000 or more) and all remaining noncore counties. (Census Glossary and ERS) Not in labor force Not in labor force includes all people 16 years old and over who are not classified as members of the labor force. This category consists mainly of students, housewives, retired workers, seasonal workers interviewed in an off season who were not looking for work, institutionalized people, and people doing only incidental unpaid family work (less than 15 hours during the reference week). (Census Glossary) Occupation Occupation describes the kind of work the person does on the job. For employed people, the data refer to the persons job during the reference week. For those who worked at two or more jobs, the data refer to the job at which the person worked the greatest number of hours. Some examples of occupational groups include managerial occupations; business and financial specialists; scientists and technicians; entertainment; healthcare; food service; personal services; sales; office and administrative support; farming; maintenance and repair; and production workers. (Census Glossary) OMB Office of Management and Budget Poverty The Census Bureau uses a set of money income thresholds that vary by family size and composition to detect who is poor. If the total income for a family or unrelated individual falls below the relevant poverty threshold, then the family

PAGE 91

80 (and all members of the family) or unrelated individual is classified as being below the poverty level. (Census Glossary) Rural For this study, towns with populations less than 2,500 not located within a MSA, and only such towns located within nonmetropolitan counties. According to the census rural is territory, population and housing units not classified as urban. Summary File 1 This file presents 100% population and housing figures for the total population, for 63 race categories, and for many other race and Hispanic or Latino categories. This includes age, sex, households, household relationship, housing units, and tenure. Also included are selected characteristics for a limited number of race and Hispanic or Latino categories. The data are available for the U.S., regions, divisions, states, counties, county subdivisions, places, census tracts, block groups, metropolitan areas, American Indian and Alaska Native areas, tribal subdivisions, Hawaiian home lands, congressional districts, and zip code tabulation areas. Data are available down to the block level for many tabulations, but only to the census-tract level for others. (Census Glossary) Summary File 3 This file presents data on the population and housing long form subjects such as income and education. It included population totals for ancestry groups. It also included selected characteristics for a limited number of race and Hispanic or Latino categories. The data are available for the U.S., regions, divisions, states, counties, county subdivisions, places, census tracts, block groups, metropolitan areas, American Indian and Alaska Native areas, tribal subdivisions, Hawaiian home lands, congressional districts, and zip code tabulation areas. (Census Glossary)

PAGE 92

81 UCR Uniform Crime Reports Unemployed All civilian 16 years old and over are classified as unemployed if they (1) were neither at work nor with a job but not at work during the reference week, and (2) were actively looking for work during the last 4 weeks, and (3) were available to accept a job. Also included as unemployed are civilians who did not work at all during the reference week, were waiting to be called back to a job from which they had been laid off, and were available for work except for temporary illness. (Census Glossary) Unpaid family workers Includes people who worked 15 hours or more without pay in a business or on a farm operated by a relative. (Census Glossary) Urban For this study, cities with populations larger than 100,000. The census defines urban as all territory, population and housing units in urbanized areas and in places of more than 2,500 persons outside of urbanized areas. Urbanized area (UA) An area consisting of a central place(s) and adjacent territory with a general population density of at least 1,000 people per square mile of land area that together have a minimum residential population of at least 50,000 people. (Census Glossary) Weeks worked in 1999 The data pertain to the number of weeks during the designated calendar year in which a person did any work for pay or profit (including paid vacation, paid sick leave, and military service) or worked without pay on a family farm or in a family business. (Census Glossary) Worked in 1999 People 16 years old and over who did any work for pay or profit (including paid vacation, sick leave, and military service) or worked without pay

PAGE 93

82 on a family farm or in a family business at any time during the past 12 months are classified as "worked in the past 12 months." All other people 16 years old and over are classified as "Did not work in the past 12 months. (Census Glossary)

PAGE 94

APPENDIX B RURAL-URBAN CONTIUUM CODES (BEALE CODES) Table B-1. 2003 Beale Codes Metro Counties 1 Counties in metro areas of 1 million population or more 2 Counties in metro areas of 250,000 to 1 million population 3 Counties in metro areas of fewer than 250,000 population Nonmetro Counties 4 Urban population of 20,000 or more, adjacent to a metro area 5 Urban population of 20,000 or more, not adjacent to a metro area 6 Urban population of 2,500 to 19,999, adjacent to a metro area 7 Urban population of 2,500 to 19,999, not adjacent to a metro area 8 Completely rural or less than 2,500 urban population, adjacent to a metro area 9 Completely rural or less than 2,500 urban population, not adjacent to a metro area Table B-2. 1993 Beale Codes Metro Counties 0 Central counties of metro areas of 1million population or more 1 Fringe counties of metro areas of 1 million population or more 2 Counties in metro areas of 250,000 to 1 million population 3 Counties in metro areas of fewer than 250,000 population Nonmetro Counties 4 Urban population of 20,000 or more, adjacent to a metro area 5 Urban population of 20,000 or more, not adjacent to a metro area 6 Urban population of 2,500 to 19,999, adjacent to a metro area 7 Urban population of 2,500 to 19,999, not adjacent to a metro area 8 Completely rural or less than 2,500 urban population, adjacent to a metro area 9 Completely rural or less than 2,500 urban population, not adjacent to a metro area (Economic Resource Service, 2004) 83

PAGE 95

APPENDIX C SOUTHERN STATES Alabama Arkansas Delaware Florida Georgia Kentucky Louisiana Maryland Mississippi North Carolina Oklahoma South Carolina Tennessee Texas Virginia West Virginia 84

PAGE 96

APPENDIX D WESTERN STATES Alaska Arizona California Colorado Hawaii Idaho Montana Nevada New Mexico Oregon Utah Washington Wyoming 85

PAGE 97

APPENDIX E CORRELATIONS BEFORE PRINCIPAL COMPONENT ANALYSIS

PAGE 98

87 Violent Property Husband Married NoInc Unpaid RatInc RatEd RatOcc PTWork Female Violent Crime 1.000 Female Property Crime .813 1.000 % Families Husband Works, Wife Doesn't .062 .088 1.000 % Married Families w/ Children -.099 -.097 .227 1.000 % Females Working FT, No income -.017 -.023 -.018 .012 1.000 % Employed Females Unpaid Family Workers -.029 -.043 -.032 .053 .347 1.000 Ratio M/F Median Income -.154 -.199 .156 .286 .067 .037 1.000 Ratio M/F Bachelor's Degree or Higher -.012 -.008 -.073 .001 .117 -.003 .116 1.000 Ratio M/F Professional Occupations .086 .140 -.075 -.160 -.014 .016 -.023 .271 1.000 % Females Working PT -.058 -.071 -.120 .106 .066 .132 .358 .137 .189 1.000 % Females Below Poverty .006 -.056 .170 -.341 -.010 -.023 -.079 -.048 -.190 -.167 Residential Mobility .089 .187 .136 -.054 .010 -.029 -.158 .003 .280 .151 Population Change -.004 .031 .050 .056 -.002 .032 .003 -.012 .107 -.005 % Population Divorced -.051 -.030 .009 -.246 .060 -.006 -.054 .020 -.029 .009 Population Size (log) .439 .590 .157 -.021 -.045 -.052 -.313 -.015 .267 -.053 % Hispanic (log) .177 .248 .289 .306 -.02 -.020 -.163 -.012 .007 -.027 South -.115 -.121 .189 -.183 .000 -.013 -.081 -.056 -.172 -.479 West .043 .078 .075 .175 .027 .017 .107 .034 .151 .340 Officer Rate -.012 -.044 -.076 -.251 -.029 -.063 .056 .139 .154 -.082 Table E-1. Correlation Matrix of Variables before Principal Component Analysis Total Sample

PAGE 99

88 Table E-1. Continued Poverty Mobil PopCha Divorced Pop Size Hisp South West Officer Female Violent Cr ime Female Property Cr ime % Families Husband Works, Wife Doesn't % Married Families w/ Chil dren % Females Working FT, No inc ome % Employed Females Unpaid Family Workers Ratio M/F Median Inc ome Ratio M/F Bachelor's Degree or Higher Ratio M/F Professional Occupations % Females Workin g PT % Females Below Poverty 1.000 Residential Mobility -.221 1.000 Population Change -.124 .318 1.000 % Population Divorced .067 .196 .019 1.000 Population Size (log) -.176 .431 .074 -.084 1.000 % Hispanic (log) -.114 .348 .143 -.057 .461 1.000 South .421 -.251 -.054 -.039 -.267 -.222 1.000 West -.261 .341 .188 .138 .229 .410 -.610 1.000 Officer Rate -.019 -.053 .003 -.014 -.274 -.145 .172 -.065 1.000

PAGE 100

Table E-2. Correlation Matrix for Variables before Principal Component Analysis Urban Sample Violent Property Husband Married NoInc Unpaid RatInc RatEd RatOcc PTWork Female Violent Crime 1.000 Female Property Crime .787 1.000 % Families Husband Works, Wife Doesn't .017 .019 1.000 % Married Families w/ Children -.188 -.189 .623 1.000 % Females Working FT, No income -.077 -.081 .274 .351 1.000 % Employed Females Unpaid Family Workers -.044 -.073 .397 .493 .536 1.000 Ratio M/F Median Income -.208 -.206 .429 .468 .279 .285 1.000 Ratio M/F Bachelor's Degree or Higher -.135 -.107 .350 .328 .282 .310 .523 1.000 Ratio M/F Professional Occupations -.082 -.057 .164 -.026 .135 .094 .219 .523 1.000 % Females Working PT -.140 -.167 -.060 -.091 .002 .103 .388 .095 .225 1.000 % Females Below Poverty .229 .135 -.077 -.393 -.200 -.140 -.267 -.364 -.498 .152 Residential Mobility -.209 -.112 .184 .019 -.020 -.062 -.072 .036 .314 .122 Population Change -.106 -.043 .271 .407 .074 .100 .207 .226 .143 -.139 % Population Divorced -.089 .025 -.412 -.442 -.149 -.194 -.098 -.074 -.123 -.103 Population Size (log) .699 .763 .017 -.255 -.116 -.097 -.336 -.145 -.025 -.276 % Hispanic (log) .048 .045 .459 .542 .202 .318 -.105 .012 -.237 -.230 South -.066 .024 .153 -.136 .015 -.079 -.042 -.010 -.057 -.295 West -.092 -.108 .321 .462 .211 .411 .121 .101 .206 .126 Officer Rate .426 .433 -.412 -.675 -.252 -.401 -.448 -.328 -.209 -.200 89

PAGE 101

Table E-2. Continued 90 Poverty Mobil PopCha Divorced Pop Size Hisp South West Officer Female Violent C rime Female Property Cr ime % Families Husband Works, Wife Doesn't % Married Families w/ Chil dren % Females Working FT, No inc ome % Employed Females Unpaid Family Workers Ratio M/F Median Inc ome Ratio M/F Bachelor's Degree or Higher Ratio M/F Professional Occupations % Females Workin g PT % Females Below Poverty 1.000 Residential Mobility -.192 1.000 Population Change -.349 .534 1.000 % Population Divorced -.067 .090 -.024 1.000 Population Size (log) .184 -.082 -.100 .059 1.000 % Hispanic (log) .018 .087 .172 -.393 .057 1.000 South .193 .067 .077 -.060 .068 -.213 1.000 West -.359 .243 .321 -.099 -.107 .457 -.578 1.000 Officer Rate .504 -.281 -.321 .092 .449 -.288 .145 -.520 1.000

PAGE 102

Table E-3. Correlation Matrix for Variables before Principal Component Analysis Rural Sample Violent Property Husband Married NoInc Unpaid RatInc RatEd RatOcc PTWork Female Violent Crime 1.000 Female Property Crime .321 1.000 % Families Husband Works, Wife Doesn't .067 .014 1.000 % Married Families w/ Children -.158 -.173 .177 1.000 % Females Working FT, No income -.044 -.038 -.017 .002 1.000 % Employed Females Unpaid Family Workers -.006 -.014 -.034 .030 .345 1.000 Ratio M/F Median Income -.083 -.070 .205 .280 .056 .009 1.000 Ratio M/F Bachelor's Degree or Higher -.059 -.041 -.089 -.018 .117 -.006 .115 1.000 Ratio M/F Professional Occupations -.028 .011 -.154 -.187 -.005 .038 .068 .279 1.000 % Females Working PT -.112 -.053 -.117 .144 .067 .132 .357 .141 .216 1.000 % Females Below Poverty .182 .054 .238 -.352 -.015 -.035 -.136 -.038 -.112 -.217 Residential Mobility .004 .075 .061 -.063 .031 .004 -.025 .000 .165 .202 Population Change .003 .021 .011 -.016 .000 .037 .015 -.021 .081 .015 % Population Divorced .036 .040 .063 -.219 .063 -.006 -.080 .024 .002 .015 Population Size (log) .170 .215 -.014 .126 -.035 .083 -.013 -.078 -.097 .060 % Hispanic (log) -.028 -.039 .218 .320 -.007 .002 -.013 -.018 -.151 .028 South .119 .069 .259 -.214 -.011 -.032 -.195 -.063 -.111 -.548 West -.027 -.025 -.018 .083 .038 .015 .227 .034 .068 .432 Officer Rate .009 .032 -.034 -.251 -.034 -.072 .014 .149 .239 -.090 91

PAGE 103

Table E-3. Continued Poverty Mobil PopCha Divorced Pop Size Hisp South West Officer Female Violent C rime Female Property Cr ime % Families Husband Works, Wife Doesn't % Married Families w/ Chil dren % Females Working FT, No inc ome % Employed Females Unpaid Family Workers Ratio M/F Median Inc ome Ratio M/F Bachelor's Degree or Higher Ratio M/F Professional Occupations % Females Workin g PT % Females Below Poverty 1.000 Residential Mobility -.154 1.000 Population Change -.080 .283 1.000 % Population Divorced .068 .273 .031 1.000 Population Size (log) -.024 .116 -.015 -.076 1.000 % Hispanic (log) -.038 .200 .109 .024 .052 1.000 South .427 -.200 -.029 -.081 -.116 -.101 1.000 West -.202 .280 .143 .224 -.030 .317 -.592 1.000 Officer Rate -.084 .045 .036 -.034 -.502 -.059 .138 .022 1.000 92

PAGE 104

APPENDIX F CORRELATIONS AFTER PRINCIPAL COMPONENT ANALYSIS Table F-1. Correlation Matrix for Variables after Principal Component Analysis Total Sample Violent Property Tradit Unpaid Inequal WorkPT Poverty Female Violent Crime 1.000 Female Property Crime .813 1.000 Traditional Families Index -.041 -.027 1.000 Family Unpaid Work -.029 -.041 .014 1.000 Gender Inequality Index .034 .063 -.098 .050 1.000 % Females Working PT -.058 -.071 .022 .123 .191 1.000 % Females Below Poverty .006 -.056 -.147 -.021 -.133 -.167 1.000 Residential Mobility .089 .187 .032 -.014 .143 .151 -.221 Population Change -.004 .031 .067 .020 .044 -.005 -.124 % Population Divorced -.051 -.030 -.169 .030 -.003 .009 .067 Population Size (log) .439 .590 .060 -.059 .121 -.053 -.176 % Hispanic (log) .177 .248 .368 -.024 -.004 -.027 -.114 South -.115 -.121 -.026 -.008 -.125 -.479 .421 West .043 .078 .168 .026 .097 .340 -.261 Officer Rate -.012 -.044 -.215 -.058 .181 -.082 -.019 Mobil PopCh Divor Pop Hisp South West Officer Female Violent Crime Female Property Crime Traditional Families Index Family Unpaid Work Gender Inequality Index % Females Working PT % Females Below Poverty Residential Mobility 1.000 Population Change .318 1.000 % Population Divorced .196 .019 1.000 Population Size (log) .431 .074 -.084 1.000 % Hispanic (log) .348 .143 -.057 .461 1.000 South -.251 -.054 -.039 -.267 -.222 1.000 West .341 .188 .138 .229 .410 -.610 1.000 Officer Rate -.053 .003 -.014 -.274 -.145 .172 -.065 1.000 93

PAGE 105

94 Table F-2. Correlation Matrix for Variables after Principal Component Analysis Urban Sample Violent Property Tradit Unpaid Inequal WorkPT Poverty Female Violent Crime 1.000 Female Property Crime .787 1.000 Traditional Families Index -.141 -.140 1.000 Family Unpaid Work -.060 -.085 .515 1.000 Gender Inequality Index -.118 -.087 .197 .236 1.000 % Females Working PT -.140 -.167 -.081 .079 .197 1.000 % Females Below Poverty .229 .135 -.326 -.178 -.506 .152 1.000 Residential Mobility -.209 -.112 .073 -.054 .230 .122 -.192 Population Change -.106 -.043 .399 .103 .201 -.139 -.349 % Population Divorced -.089 .025 -.469 -.201 -.117 -.103 -.067 Population Size (log) .699 .763 -.193 -.115 -.084 -.276 .184 % Hispanic (log) .048 .045 .555 .315 -.156 -.230 .018 South -.066 .024 -.053 -.055 -.043 -.295 .193 West -.092 -.108 .455 .389 .186 .126 -.359 Officer Rate .426 .433 -.651 -.396 -.293 -.200 .504 Mobil PopCh Divor Pop Hisp South West Officer Female Violent Crime Female Property Crime Traditional Families Index Family Unpaid Work Gender Inequality Index % Females Working PT % Females Below Poverty Residential Mobility 1.000 Population Change .534 1.000 % Population Divorced .090 -.024 1.000 Population Size (log) -.082 -.100 .059 1.000 % Hispanic (log) .087 .172 -.393 .057 1.000 South .067 -.060 .077 .068 -.213 1.000 West .243 .321 -.099 -.107 .457 -.578 1.000 Officer Rate -.281 -.321 .092 .449 -.288 .145 -.520 1.000

PAGE 106

95 Table F-3. Correlation Matrix for Variables after Principal Component Analysis Rural Sample Violent Property Tradit Unpaid Inequal WorkPT Poverty Female Violent Crime 1.000 Female Property Crime .321 1.000 Traditional Families Index -.072 -.113 1.000 Family Unpaid Work -.029 -.031 -.004 1.000 Gender Inequality Index -.058 -.026 -.144 .058 1.000 % Females Working PT -.112 -.053 .044 .124 .204 1.000 % Females Below Poverty .182 .054 -.105 -.031 -.085 -.217 1.000 Residential Mobility .004 .075 -.008 .020 .080 .202 -.154 Population Change .003 .021 -.004 .025 .021 .015 -.080 % Population Divorced .036 .040 -.116 .032 .015 .015 .068 Population Size (log) .170 .215 .078 .034 -.102 .060 -.024 % Hispanic (log) -.028 -.039 .349 -.003 -.082 .028 -.038 South .119 .069 .000 -.027 -.097 -.548 .427 West -.027 -.025 .055 .031 .054 .432 -.202 Officer Rate .009 .032 -.191 -.067 .227 -.090 -.084 Mobil PopCh Divor Pop Hisp South West Officer Female Violent Crime Female Property Crime Traditional Families Index Family Unpaid Work Gender Inequality Index % Females Working PT % Females Below Poverty Residential Mobility 1.000 Population Change .283 1.000 % Population Divorced .273 .031 1.000 Population Size (log) .116 -.015 -.076 1.000 % Hispanic (log) .200 .109 .024 .052 1.000 South -.200 -.029 -.081 -.116 -.101 1.000 West .280 .143 .224 -.030 .317 -.592 1.000 Officer Rate .045 .036 -.034 -.502 -.059 .138 .022 1.000

PAGE 107

APPENDIX G VARIANCE INFLATION FACTOR VALUES Table G-1. Variance Inflation Factor Values for Independent Variables Included in Female Violent Crime and Property Crime Arrest Models, 2001 Total Model Urban Model Rural Model Traditional Families Index 1.373 4.327 1.271 Gender Inequality Index 1.146 1.787 1.138 Family Unpaid Work 1.033 1.637 1.031 % Females Below Poverty 1.331 2.519 1.328 % Females Working PT 1.524 2.242 1.615 Residential Mobility 1.646 2.016 1.359 Population Change 1.162 2.000 1.119 % Population Divorced 1.173 1.843 1.196 Population Size (log) 1.771 1.410 1.405 % Hispanic (log) 1.760 2.641 1.345 South 2.301 2.439 2.288 West 2.041 2.942 1.994 Officer Rate 1.286 3.722 1.555 96

PAGE 108

APPENDIX H URBAN VIOLENT CRIME MODELS Table H-1. Negative Binomial Regression Coefficients, (Z Scores), and Standardized Errors for Urban Female Violent Crime Arrest Models, 2001 Full Model Without Trad Family Without Off Rate Private Patriarchy Traditional Families Index SE .004 (0.34) .013 -.002 (-0.22) .010 Family Unpaid Work SE .068 (0.18) .367 .106 (0.30) .350 .082 (0.22) .366 Public Patriarchy Gender Inequality Index SE -.519 (-1.36) .382 -.520 (-1.36) .381 -.508 (-1.33) .382 % Females Working PT SE -.036** (-2.69) .013 -.037** (-2.77) .013 -.040** (-3.18) .013 % Females Below Poverty SE .098** (5.18) .019 .098** (5.18) .019 .104** (5.94) .018 Social Disorganization Residential Mobility SE -.0006 (-0.06) .0096 -.001 (-0.15) .009 -.002 (-0.19) .0096 Population Change SE -.001 (-0.92) .002 -.001 (-0.86) .002 -.001 (-0.78) .002 97

PAGE 109

98 Table H-1. Continued Officer Rate SE .084 (0.93) .090 .067 (0.89) .074 Constant SE -6.145** (-4.61) 1.331 -5.894** (-5.28) 1.116 -5.763** (-4.51) 1.278 Log Likelihood -1090.4095 -1090.4682 -1090.8485 X2 ratio test, alpha = 0 8903.58** 8972.82** Pseudo R2 0.0336 0.0335 0.0332 N = 199 N = 199 N = 199 Full Model Without Trad Family Without Off Rate % Population Divorced SE .021 (0.77) .028 .017 (.069) .025 .012 (0.46) .026 Population Size (log) SE -.017 (-0.25) .067 -.016 (-0.24) .067 .003 (0.04) .064 Controls % Hispanic (log) SE -.073 (-1.18) .062 -.068 (-1.13) .060 -.077 (-1.24) .062 South SE -.169 (-1.21) .140 -.161 (-1.17) .139 -.188 (-1.35) .139 West SE .341* (2.21) .154 .341* (2.21) .154 .299* (2.03) .148 ^ p < .10 p < .05 ** p < .01

PAGE 110

APPENDIX I URBAN PROPERTY CRIME MODELS Table I-1. Negative Binomial Regression Coefficients, (Z Scores), and Standardized Errors for Urban Female Property Crime Arrest Models Full Model Without Trad Family Without Off Rate Private Patriarchy Traditional Families Index SE .002 (0.22) .011 -.009 (-0.94) .0097 Family Unpaid Work SE .195 (0.67) .293 .209 (0.73) .287 .189 (0.65) .292 Public Patriarchy Gender Inequality Index SE -.343 (-1.00) .345 -.339 (-0.99) .344 -.281 (-0.80) .349 % Females Working PT SE .033** (2.80) .012 .033** (2.79) .012 .023* (2.08) .011 % Females Below Poverty SE -.026^ (-1.65) .016 -.026^ (-1.66) .016 -.012 (-0.82) .015 Social Disorganization Residential Mobility SE -.0001 (-0.02) .008 -.001 (-0.09) .008 -.003 (-0.31) .008 Population Change SE .001 (0.82) .002 .002 (0.99) .002 .002 (1.05) .002 99

PAGE 111

100 Table I-1. Continued Full Model Without Trad Family Without Off Rate % Population Divorced SE .087** (3.63) .024 .085** (3.92) .022 .073** (3.09) .024 Population Size (log) SE -.093 (-1.54) .060 -.093 (-1.53) .060 -.040 (-0.71) .056 Controls % Hispanic (log) SE .025 (0.45) .056 .029 (0.55) .053 .009 (0.16) .056 South SE .374** (2.93) .128 .380** (3.06) .124 .333** (2.60) .128 West SE .083 (0.66) .126 .087 (0.70) .124 .059 (0.47) .126 Officer Rate SE .158* (2.25) .070 .151* (2.44) .062 Constant SE -5.919** (-5.37) 1.103 -5.804 (-6.01) .966 -5.469** (-4.94) 1.107 Log Likelihood -1391.2703 -1391.2938 -1393.945 X2 ratio test, alpha = 0 2.9E+04** 3.0E+04** 3.4E+04** Pseudo R2 0.0133 0.0133 0.0114 N = 199 N = 199 N = 199 ^ p < .10 p < .05 ** p < .01

PAGE 112

LIST OF REFERENCES Acoca, L. (1998). Outside/inside: The violation of American girls at home, on the streets, and in the juvenile justice system. Crime and Delinquency, 44(4), 561-589. Adler, F. (1975). Sisters in crime: The rise of the new female criminal. New York: McGraw-Hill. Agresti, A. (2002). Negative binomial regression. In Categorical data analysis (2 nd ed., pp. 559-563). New York: Wiley. Albrecht, D.E., Albrecht, C. M., & Albrecht, S. L. (2000). Poverty in nonmetropolitan America: Impacts of industrial, employment, and family structure variables. Rural Sociology, 65, 87-103. Arthur, J. A. (1991). Socioeconomic predictors of crime in rural Georgia. Criminal Justice Review, 16(1), 29-41. Atwell, M. W. (2002). Definitions, examples and analytical tools. In Equal protection of the law? Gender and justice in the United States (pp. 3-11). New York: Peter Lang. Barnett, C., & Mencken, F. C. (2002). Social disorganization theory and the contextual nature of crime in nonmetropolitan counties. Rural Sociology, 67(3), 372-393. Beck, A. J. (2000). Prison and jail inmates at midyear 1999 (NCJ 181643). Washington, DC: U.S. Department of Justice, Bureau of Justice Statistics. Belknap, J. (2001). The invisible woman: Gender, crime and justice (2 nd ed.). Belmont, CA: Wadsworth/Thomson Learning. Bloom, B., Owen, B., Rosenbaum, J., & Deschenes, E. P. (2003). Focusing on girls and young women: A gendered perspective on female delinquency. Women and Criminal Justice, 14, 117-136. Box, S., & Hale, C. (1983). Liberation and female criminality in England and Wales. British Journal of Criminology, 23, 35-49. Box, S., & Hale, C. (1984). Liberation/emancipation, economic marginalization, or less chivalry: The relevance of three theoretical arguments to female crime patterns in England and Wales, 1951-1980. Criminology, 22, 473-497. 101

PAGE 113

102 Broadway, M. (1990). Meatpacking and its social and economic consequences for Garden City, Kansas in the 1980s. Urban Anthropology, 19(4), 321-344. Burgess, E.W. (1925). The growth of the city. In R.E.Park & E. W. Burgess (Eds.), The city. Chicago: University of Chicago Press. Butterfield, F. (2005, February 13). Social isolation, guns and a culture of suicide. The New York Times, section 1, p. 28. Cameron, A. C., & Trivedi, P. K. (1998). Regression analysis of count data. Cambridge, UK: Cambridge University Press. Cebulak, W. (2004). Why rural crime and justice really matter. Journal of Police and Criminal Psychology, 19(1), 71-81. Census Glossary (n.d.). Retrieved December 1, 2004, from American FactFinder Website: http://factfinder.census.gov Chesney-Lind, M. (1989). Girls crime and womans place: Toward a feminist model of female delinquency. Crime and Delinquency, 35(1), 5-29. Chin, H. C., & Quddus, M. A. (2003). Modeling count data with excess zeroes: An empirical application to traffic accidents. Sociological Methods and Research, 32(1), 90-116. Comack, E. (1999). New possibilities for a feminism in criminology? From dualism to diversity. Canadian Journal of Criminology, 41(2), 161-170. Daly, K., & Chesney-Lind, M. (1988). Feminism and criminology. Justice Quarterly, 5, 101-143. Davidson, O. G. (1996). Broken heartland: The rise of Americas rural ghetto. Iowa City, IA: University of Iowa Press. Davis, K., Taylor, B., & Furniss, D. (2001). Narrative accounts of tracking the rural domestic violence survivors journey: A feminist approach. Health Care for Women International, 22(4), 333-347. DeWees, M. A., & Parker, K. F. (2003). Women, region, and types of homicide: Are there regional differences in the structural status of women and homicide offending? Homicide Studies, 7(4), 368-393. Diala, C. C., Muntaner, C., & Walrath, C. (2004). Gender, occupation, and socioeconomic correlates of alcohol and drug abuse among U.S. rural, metropolitan, and urban residents. American Journal of Drug and Alcohol Abuse, 30(2), 409-428.

PAGE 114

103 Donnermeyer, J. F. (2004, November). Rural crime, poverty, and community. Paper presented at the American Society of Criminology annual meeting, Nashville, TN. Donnermeyer, J. F., Barclay, E. M., & Jobes, P. C. (2002). Drug-related offenses and the structure of communities in rural Australia. Substance Use and Misuse, 37, 631-661. Economic Resource Service. (2003, August) Measuring rurality: New definitions in 2003. Retrieved November 23, 2004 from http://www.ers.usda.gov/Briefing/Rurality/Newdefinitions/ Economic Resource Service. (2004, April 28). Measuring rurality: Rural urban continuum codes. Retrieved November 23, 2004, from http://www.ers.usda.gov/briefing/rurality/RuralUrbCon/ Egan, T., (2002, December 8). The nation: Pastoral poverty; The seeds of decline. The New York Times, section 4, p. 1. English, D. J., Widom, C. S., & Brandford, C. B. (2001). Childhood victimization and delinquency, adult criminality, and violent criminal behavior: A replication and extension, final report (NCJ 192291). Washington, DC: U.S. Department of Justice, National Institute of Justice. Fisher, J. C., & Mason, R. L. (1981). The analysis of multicollinear data in criminology. In J. A. Fox (Ed.), Methods in quantitative criminology (pp. 99-125). New York: Academic Press, Gardner, W., Mulvey, E. P., & Shaw, E. C. (1995). Regression analyses of counts and rates: Poisson, Overdispersed Poisson, and negative binomial models. Psychological Bulletin, 118, 392-404. Greenfield, L. A., & Snell, T. L. (1999). Women offenders (NCJ 175688). Washington, DC: U.S. Department of Justice, Bureau of Justice Statistics. Hagan, J., Gillis, A. R., & Simpson, J. (1985). The class structure of gender and delinquency: Toward a power-control theory of common delinquent behavior. American Journal of Sociology, 90(6), 1151-1178. Hagan, J., Simpson, J., & Gillis, A. R. (1987). Class in the household: A power-control theory of gender and delinquency. American Journal of Sociology, 92(4), 788-816. Hagan, J., Simpson, J., & Gillis, A. R. (1988). Feminist scholarship, relational and instrumental control, and a power-control theory of gender and delinquency. British Journal of Sociology, 39(3), 301-336.

PAGE 115

104 Hartsock, N. C. M. (1998). Fundamental feminism: Process and perspective. In The feminist standpoint revisited and other essays (pp. 32-43). Boulder, CO: Westview Press. Heimer, K. (2000). The nature of crime: Continuity and change. Changes in the gender gap in crime and womens economic marginalization. Criminal Justice 2000, 1, 427-483. Hunnicutt, G., & Broidy, L. M. (2004). Liberation and economic marginalization: A reformulation and test of (formerly?) competing models. Journal of Research in Crime and Delinquency, 4(2), 130-155. Jobes, P. C. (1999). Residential stability and crime in small rural agriculture and recreational towns. Sociological Perspectives, 42(3), 499-524. Jobes, P. C. (2002). Effective officer and good neighbour: Problems and perceptions among police in rural Australia. Policing, 25(2), 256-273. Jobes, P. C. (2003). Human ecology and rural policing: A grounded theoretical analysis of how personal constraints and community characteristics influence strategies of law enforcement in rural New South Wales, Australia. Police Practice and Research, 4(1), 3-19. Jobes, P. C., Barclay, E., Weinand, H., & Donnermeyer, J. F. (2004). A structural analysis of social disorganisation and crime in rural communities in Australia. Australian and New Zealand Journal of Criminology, 37(1), 114-140. Jolliffe, D. (2003). Nonmetro poverty: Assessing the effect of the 1990s. Amber Waves (September). Washington, DC: U.S. Department of Agriculture, Economic Research Service. Kaker, S., Friedemann, M. L., & Peck, L. (2002). Girls in detention: The results of focus group discussion interviews and official records review. Journal of Contemporary Criminal Justice, 18, 57-73. Kandel, W., & Newman, C. (2004). Rural Hispanics: Employment and residential trends. Amber Waves, (June). Washington, DC: U.S. Department of Agriculture, Economic Research Service. Koons-Witt, B. A., & Schram, P. J. (2003). The prevalence and nature of violent offending by females. Journal of Criminal Justice, 31, 361-371. Kposowa, A. J., Breault, K.D., & Harrison, B. M. (1995). Reassessing the structural covariates of violent and property crime in the USA: A county level analysis. British Journal of Sociology, 46(1), 79-105.

PAGE 116

105 Krishnan, S. P., Hilbert, J. C., & VanLeeuwen, D. (2001). Domestic violence and help-seeking behaviors among rural women: Results from a shelter-based study. Family and Community Health, 24(1), 28-38. Land, K. C., McCall, P. L., & Cohen, L. E. (1990). Structural covariates of homicide rates: Are there any invariances across time and social space? American Journal of Sociology, 95(4), 922-963. Lee, M. R., Maume, M. O., & Ousey, G. C. (2003). Social isolation and lethal violence across the metro/nonmetro divide: The effects of socioeconomic disadvantage and poverty concentration on homicide. Rural Sociology, 68(1), 107-131. Liederbach, J., & Frank, J. (2003). Policing Mayberry: The work routines of small-town and rural officers. American Journal of Criminal Justice, 28(1), 53-72. Mayhew, B. H., & Levinger, R. L. (1976). Size and the density of interaction in human aggregates. American Journal of Sociology, 82, 86-110. McCarthy, B., Hagan, J., & Woodward, T. S. (1999). In the company of women: Structure and agency in a revised power-control theory of gender and delinquency. Criminology, 37(4), 761-788. McClellan, D. S., Farabee, D., & Couch, B. M. (1997). Early victimization, drug use, and criminality: A comparison of male and female prisoners. Criminal Justice and Behavior, 24, 455-476. Messerschmidt, J. W. (1986). Capitalism, patriarchy, and crime: Toward a socialist feminist criminology. Totowa, NJ: Rowman & Littlefield. Min, Y. and Agresti, A. (2002). Modeling nonnegative data with clumping at zero: A survey. Journal of the Iranian Statistical Society, 1, 7-33. Osgood, D. W. (2000). Poisson-based regression analysis of aggregate crime rates. Journal of Quantitative Criminology, 16(1), 21-43 Osgood, D. W., & Chambers, J. M. (2000). Social disorganization outside the metropolis: An analysis of rural youth violence. Criminology, 38(1), 81-115. OShea, T. C. (1999). Community policing in small town rural America: A comparison of police officer attitudes in Chicago and Baldwin County, Alabama. Policing and Society, 9(1), 59-76. Ott, R. L., & Longnecker, M. (2001). An introduction to statistical methods and data analysis (5 th ed.). Pacific Grove, CA: Duxbury. Park, R. E. & Burgess, E. W. (1925). The city. Chicago: University of Chicago Press.

PAGE 117

106 Paternoster, R., Brame, R., Mazerolle, P., & Piquero, A. (1998). Using the correct statistical test for the equality of regression coefficients. Criminology, 36(4), 859-866. Petee, T.A., & Kowalski, G. S. (1993). Modeling rural violent crime rates: A test of social disorganization theory. Sociological Focus, 26(1), 87-89. Pindyck, R. S., & Rubinfeld, D. L. (1998). Econometric models and economic forecasts (4 th ed.). Boston: McGraw-Hill. Price, M. L., & Clay, D. C. (1980). Structural disturbances in rural communities: Some repercussions of the migration turnaround in Michigan. Rural Sociology, 45(4), 591-607. Robinson, A. L. (2003). The impact of police social capital on officer performance of community policing. Policing, 26(4), 656-689. Rural America at a glance, 2004. (2004, September). (Agriculture Information Bulletin No. 793).Washington, DC: U.S. Department of Agriculture, Economic Research Service. Rural poverty at a glance. (2004, July). (Rural Development Research Report No. 100). Washington, DC: U.S. Department of Agriculture, Economic Research Service. Sampson, R. J. (1987). Urban black violence: The effect of male joblessness and family disruption. American Journal of Sociology, 93, 348-382. Sampson, R. J. (1991). Linking the microand macro-level dimensions of community social organization. Social Forces, 70(1), 43-64. Sampson, R. J., & Groves, W. B. (1989). Community structure and crime: Testing social-disorganization theory. American Journal of Sociology, 94(4), 774-802. Sampson, R. J., Raudenbush, S. W., & Earls, F. (1997). Neighborhoods and violent crime: A multilevel study of collective efficacy. Science, 277, 918-924. Shaw, C. R., & McKay, H. (1942). Juvenile delinquency and urban areas. Chicago: University of Chicago Press. Siegel, J. A., & Williams, L. M. (2003). The relationship between child sexual abuse and female delinquency and crime: A prospective study. Journal of Research in Crime and Delinquency, 40, 71-94. Simon, R. J. (1975). Women and crime. Lexington, MA: Lexington Books. Simpson, S. S. (1991). Caste, class, and violent crime: Explaining difference in female offending. Criminology, 29(1), 115-135.

PAGE 118

107 Snyder, A. R., & McLaughlin, D. K. (2004). Female-headed families and poverty in rural America. Rural Sociology, 69(1), 127-149. Steffensmeier, D. (1980). Sex differences in patterns of adult crime, 1965-1977. Social Forces, 58, 1080-1109. Steffensmeier, D. (1993). National trends in female arrests, 1960-1990: Assessment and recommendations for future research. Journal of Quantitative Criminology, 9, 413-441. Steffensmeier, D., & Allan, E. (1996). Gender and crime: Toward a gendered theory of female offending. Annual Review of Sociology, 22, 459-487. Steffensmeier, D., & Cobb, M. J. (1981). Sex differences in arrest patterns, 1934-1979. Social Problems, 28, 37-50. Steffensmeier, D., & Haynie, D. (2000). Gender, structural disadvantage, and urban crime: Do macrosocial variables also explain female offending rates? Criminology, 38(2), 403-438. Steffensmeier, D., & Jordan, C. (1978). Changing patterns of female crime in rural America, 1962-1975. Rural Sociology, 43(1), 87-102. Steffensmeier, D. D., & Streifel, C. (1992). Time-series analysis of the female percentage of arrests for property crimes, 1960-1985: A test of alternative explanations. Justice Quarterly, 9, 77-103. Walby, S. (1986). Patriarchy at work. Minneapolis, MN: University of Minnesota Press. Walby, S. (1989). Theorizing patriarchy. Sociology, 23(2), 213-234. Walby, S. (1990). Theorizing patriarchy. London: Blackwell. Websdale, N. (1998). Rural woman battering and the justice system An ethnography. Thousand Oaks, CA: Sage Publications. Websdale, N., & Johnson, B. (1998). An ethnostatistical comparison of the forms and levels of woman battering in urban and rural Kentucky. Criminal Justice Review, 23(2), 161-196. Weisheit, R. A., & Donnermeyer, J. F. (2000). The nature of crime: Continuity and change Change and continuity in crime in rural America. Criminal Justice 2000, 1, 309-357. Weisheit, R. A., & Fuller, J. (2004). From the field: Methamphetamine in the heartland: A review and initial exploration. Journal of Crime and Justice, 27(1), 131-151.

PAGE 119

108 Weisheit, R. A., & Wells, L. E. (1996). Rural crime and justice: Implications for theory and research. Crime and Delinquency, 42(3), 379-397. Weisheit, R. A., & Wells, L. E. (1999). The future of rural crime in America. Journal of Crime and Justice, 22(1), 1-27. Weisheit, R. A., Wells, L. E., & Falcone, D. N. (1994). Community policing in small town and rural America. Crime and Delinquency, 40(4), 549-567. Wells, L. E., & Weisheit, R. A. (2004). Patterns of rural and urban crime: A county-level comparison. Criminal Justice Review, 29(1), 1-22. Whaley, R. B., & Messner, S. F. (2002). Gender equality and gendered homicides. Homicide Studies, 6(3), 188-210. Widom, C. S. (1989). Child abuse, neglect, and violent criminal behavior. Criminology, 27, 251-271. Widom, C. S., & Maxfield, M. G. (2001). An update on the "cycle of violence (NCJ 184894). Washington, DC: U.S. Department of Justice, National Institute of Justice. Yllo, K. (1983). Sexual equality and violence against wives in American states. Journal of Comparative Family Studies, 14, 67-86. Yllo, K. (1984). The status of women, marital equality, and violence against wives. Journal of Family Issues, 5, 307-320. Yllo, K., & Straus, M. A. (1984). The impact of structural inequality and sexist family norms on rates of wife-beating. Journal of International and Comparative Social Welfare, 1, 16-29.

PAGE 120

BIOGRAPHICAL SKETCH Stephanie Hays is originally from rural Iowa. After graduating high school in 1999, Stephanie left Iowa to pursue a college education. She received a B.A. with distinction in sociology-criminology in 2001 and a Master of Human Relations degree in 2003 from the University of Oklahoma. She will receive a M.A. degree in criminology, law and society with a minor in statistics in May 2005 from the University of Florida. Stephanie will be staying at the University of Florida to continue working towards a PhD in criminology. 109


Permanent Link: http://ufdc.ufl.edu/UFE0009880/00001

Material Information

Title: Women and crime in context : examining the linkages between structural conditions and female offending within the context of place
Physical Description: xi, 109 p.
Language: English
Creator: Hays, Stephanie Ann ( Dissertant )
Parker, Karen F. ( Thesis advisor )
Brank, Eve ( Reviewer )
Lanza-Kaduce, Lonn ( Reviewer )
Winner, Lawrence ( Reviewer )
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2005
Copyright Date: 2005

Subjects

Subjects / Keywords: Criminology, Law and Society thesis, M.A   ( local )
Dissertations, Academic -- UF -- Criminology, Law and Society   ( local )

Notes

Abstract: This study examines the association between social structural conditions and female offending in urban and rural areas. Although studies of female offending have increased in recent years, the majority of this research has focused on individual or situational characteristics to the exclusion of structural predictors. In fact, few empirical studies have examined the structural correlates of female offending. This study draws upon social disorganization and feminist theories to explore the linkages between structural conditions and female offending within the context of place. Using 2000 census data and 2001 UCR arrest data, Poisson-based techniques are used to investigate the influence of structural predictors on female arrests for both violent and property offenses within urban cities and rural towns. Furthermore, this examination allows for the investigation of whether these structural factors exhibit similar influences on female arrest patterns across urban and rural areas.
Subject: Crime, patriarchy
General Note: Title from title page of source document.
General Note: Document formatted into pages; contains 120 pages.
General Note: Includes vita.
Thesis: Thesis (M.A.)--University of Florida, 2005.
Bibliography: Includes bibliographical references.
General Note: Text (Electronic thesis) in PDF format.

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0009880:00001

Permanent Link: http://ufdc.ufl.edu/UFE0009880/00001

Material Information

Title: Women and crime in context : examining the linkages between structural conditions and female offending within the context of place
Physical Description: xi, 109 p.
Language: English
Creator: Hays, Stephanie Ann ( Dissertant )
Parker, Karen F. ( Thesis advisor )
Brank, Eve ( Reviewer )
Lanza-Kaduce, Lonn ( Reviewer )
Winner, Lawrence ( Reviewer )
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2005
Copyright Date: 2005

Subjects

Subjects / Keywords: Criminology, Law and Society thesis, M.A   ( local )
Dissertations, Academic -- UF -- Criminology, Law and Society   ( local )

Notes

Abstract: This study examines the association between social structural conditions and female offending in urban and rural areas. Although studies of female offending have increased in recent years, the majority of this research has focused on individual or situational characteristics to the exclusion of structural predictors. In fact, few empirical studies have examined the structural correlates of female offending. This study draws upon social disorganization and feminist theories to explore the linkages between structural conditions and female offending within the context of place. Using 2000 census data and 2001 UCR arrest data, Poisson-based techniques are used to investigate the influence of structural predictors on female arrests for both violent and property offenses within urban cities and rural towns. Furthermore, this examination allows for the investigation of whether these structural factors exhibit similar influences on female arrest patterns across urban and rural areas.
Subject: Crime, patriarchy
General Note: Title from title page of source document.
General Note: Document formatted into pages; contains 120 pages.
General Note: Includes vita.
Thesis: Thesis (M.A.)--University of Florida, 2005.
Bibliography: Includes bibliographical references.
General Note: Text (Electronic thesis) in PDF format.

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0009880:00001


This item has the following downloads:


Full Text












WOMEN AND CRIME IN CONTEXT: EXAMINING THE LINKAGES BETWEEN
STRUCTURAL CONDITIONS AND FEMALE OFFENDING WITHIN THE
CONTEXT OF PLACE















By

STEPHANIE ANN HAYS


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF ARTS

UNIVERSITY OF FLORIDA


2005

































Copyright 2005

by

Stephanie Ann Hays

































This document is dedicated to my parents, Larry and Glenda Hays.















ACKNOWLEDGMENTS

I would like to thank my entire committee, Dr. Karen Parker, Dr. Eve Brank, Dr.

Lonn Lanza-Kaduce, and Dr. Lawrence Winner for all of their help and support in

completing this thesis. In addition, I would like to thank Larry for being a social science

friendly statistician and agreeing to be the minor representative, Lonn for being a great

department chair who is willing to help with anything and everything and for exchanging

tales of Iowa, and Eve for putting up with me for the past two years as a TA and for being

a great mentor and friend. Most importantly, I have to thank Karen for being the best

committee chair that I could ever ask for, for keeping me on schedule with endless

amounts of encouragement and patience, and for all of her support along the way not only

as my chair, but also as a mentor, colleague, friend, and even "family" on some holidays.

I also need to thank my parents, Larry and Glenda Hays, and my brother Bryan

for supporting me in all of my decisions throughout the years and for spending numerous

family vacations either touring colleges across the country or helping me move.















TABLE OF CONTENTS

page

A C K N O W L E D G M E N T S ................................................................................................. iv

LIST OF TABLES ............... ............................. .. .. .. ............ .. viii

LIST OF FIGURES ............................... ... ...... ... ................. .x

ABSTRACT .............. .................. .......... .............. xi

CHAPTER

1 IN TR OD U CTION ............................................... .. ......................... ..

Im portance of G ender ..................................................... .... .............. .
Im portance of Place .................. ........................................... .............. .. .2
R research Q uestions............ .................................................................. ........ .. .... .5
Su m m ary ...................................... .................................. ..................... 5

2 STRUCTURAL THEORIES: GENDER AND PLACE .............................................6

F em in ist P ersp ectiv e ........................................................................... .......... .. .. .
W hat Is P atriarchy? ................. .... ............................ .. ......... ............ ..
Patriarchy and Crim e .................. .......................... .... .... ................. .8
E m pirical F findings ................................................... .............. .......... ...... .
S u m m ary ......................................................................................... ............ 13
Social Disorganization Theory ...........................................................................13
Empirical Findings .................................. .... .... .... .............. 15
S u m m ary ............................................................................... 18
C conclusion ...................................................................................................... ....... 18

3 WOMEN AND CRIME IN CONTEXT......................................... ............... 20

Patriarchy and G ender ...................................................... ...... ...... ... .. 20
Social D isorganization and Place ........................................ .......................... 22
Sum m ary ...................................... ................. ................. .......... 28

4 DATA AND M ETHODOLOGY ........................................ ......................... 31

Source of the D ata .................................... ....................................... ...... .... .... 3 1
U nit of A n aly sis ................................................................. ............................ .3 1



v









M e a su re s ............................................................................................................... 3 3
D ep en d ent V ariab les ........................................ ............................................3 3
Independent V ariables ............................................................. .............. 33
P riv ate patriarchy ........................ ......................... ............... ... 34
Public patriarchy (economic marginalization) ..........................................36
Social disorganization ........................................................................... 39
C o n tro ls ................................................................4 1
M methodology ................. .............................................. ........ .............43
Principal Com ponent A nalysis.................................... ..................................... 46
Private patriarchy .................. .................. ...... ...... ........47
Public patriarchy (economic marginalization) ..........................................47
V ariance Inflation .................. .................................................. 49
A n a ly tic a l P la n ...................................................................................................... 5 0

5 R E S U L T S .............................................................................5 6

V violent C rim e ....................................................... 56
P private P atriarchy ............................................................57
P u b lic P patriarch y ........................................................................................... 5 7
Social Disorganization...................... .......... ......... 58
C o n tro ls ...........................................................................5 9
P property C rim e.....................................................60
P riv ate P atriarchy ............................................................6 1
P u b lic P patriarch y ........................................................................................... 6 1
Social Disorganization...................... .......... ......... 62
Controls ....................................................63
Com paring R egression Coefficients .................................................................... 63
V violent C rim e ...............................................................64
P rop erty C rim e ..............................................................6 5
S u m m a ry ............................................................................................................... 6 5

6 DISCUSSION AND CONCLUSION ...........................................71

D iscu ssio n ......................... .................................................................................... 7 1
Limitations and Future Research ..................................................................... 73
C conclusion ...................................................................................................... ....... 74

APPENDIX

A ABBREVIATIONS AND DEFINITIONS ............................................. 75

B RURAL-URBAN CONTIUUM CODES (BEALE CODES) .................................83

C SO U T H E R N ST A T E S ......................................................................................... 84

D W E ST E R N ST A T E S ............................................................................................ 85

E CORRELATIONS BEFORE PRINCIPAL COMPONENT ANALYSIS .................86









F CORRELATIONS AFTER PRINCIPAL COMPONENT ANALYSIS...................93

G VARIANCE INFLATION FACTOR VALUES .....................................................96

H URBAN VIOLENT CRIM E M ODELS ........... ................................... ............... 97

I URBAN PROPERTY CRIME MODELS................. .........................99

LIST OF REFEREN CES .. ................................................................... ............... 101

BIOGRAPHICAL SKETCH ................................................................ 109
















LIST OF TABLES


Table page

2-1 Private and Public Patriarchy .............................................................................. 7

4-1 Means and Standard Deviations of Variables before Principal Component
A n a ly sis .............................................................................................................. 5 3

4-2 Principal Component Factor Matrices after Rotation Total Sample...................54

4-3 Means and Standard Deviations of Variables after Principal Component
A n aly sis ............................. ............................................................. ............... 5 5

5-1 Negative Binomial and ZINB Regression Coefficients, (Z Scores), and
Standardized Errors for Female Violent Crime Arrest Models, 2001...................67

5-2 Negative Binomial and ZINB Regression Coefficients, (Z Scores), and
Standardized Errors for Female Property Crime Arrest Models, 2001....................69

B-1 2003 Beale Codes ......... ............. ................. .. .... ...... .. ............ 83

B-2 1993 Beale Codes ....... .. .... ........................... .... ...... .. ............ 83

E-1 Correlation Matrix of Variables before Principal Component Analysis Total
S am ple ............. .... .... ............ ..........................................87

E-2 Correlation Matrix for Variables before Principal Component Analysis Urban
Sam ple ............. .... .... ............ ..........................................89

E-3 Correlation Matrix for Variables before Principal Component Analysis Rural
Sam ple ................. ........ ............................. ........................... 9 1

F-l Correlation Matrix for Variables after Principal Component Analysis Total
S am ple ............. .... .... ............ ..........................................93

F-2 Correlation Matrix for Variables after Principal Component Analysis Urban
Sam ple ............. .... .... ............ ..........................................94

F-3 Correlation Matrix for Variables after Principal Component Analysis Rural
S am p le ................................ ......... ................................................9 5









G-l Variance Inflation Factor Values for Independent Variables Included in Female
Violent Crime and Property Crime Arrest Models, 2001 .....................................96

H-1 Negative Binomial Regression Coefficients, (Z Scores), and Standardized Errors
for Urban Female Violent Crime Arrest Models, 2001 ..........................................97

I-1 Negative Binomial Regression Coefficients, (Z Scores), and Standardized Errors
for Urban Female Property Crime Arrest Models..............................................99
















LIST OF FIGURES


Figure p

2-1 Relationship betw een Patriarchy and Crim e ........................................ ...................8

2-2 B urgess's Z ones for C ity G row th.................................................. ..................... 14

2-3 Basic Theoretical Model of Social Disorganization Theory ..............................15

3-1 Proposed Relationship between Patriarchy, Social Disorganization, and Crime.....30

4-1 Simplified Variable Relationship M odel ............................. ............................... 52















Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Arts

WOMEN AND CRIME IN CONTEXT: EXAMINING THE LINKAGES BETWEEN
STRUCTURAL CONDITIONS AND FEMALE OFFENDING WITHIN THE
CONTEXT OF PLACE

By

Stephanie Ann Hays

May 2005

Chair: Karen F. Parker
Major Department: Department of Criminology, Law and Society

This study examines the association between social structural conditions and

female offending in urban and rural areas. Although studies of female offending have

increased in recent years, the majority of this research has focused on individual or

situational characteristics to the exclusion of structural predictors. In fact, few empirical

studies have examined the structural correlates of female offending. This study draws

upon social disorganization and feminist theories to explore the linkages between

structural conditions and female offending within the context of place. Using 2000 census

data and 2001 UCR arrest data, Poisson-based techniques are used to investigate the

influence of structural predictors on female arrests for both violent and property offenses

within urban cities and rural towns. Furthermore, this examination allows for the

investigation of whether these structural factors exhibit similar influences on female

arrest patterns across urban and rural areas.














CHAPTER 1
INTRODUCTION

This study examines the impact of social structure on female arrests in both urban

and rural areas. More specifically, the current study explores the impact of patriarchy and

social disorganization on female arrests for both violent and property offenses. It is the

first study to examine the impact of patriarchy on female offending at the macro-level,

and is also the first study to use towns as the unit of analysis when applying social

disorganization theory to rural crime. This introductory chapter outlines the importance

of examining both gender and place within criminological research.



Importance of Gender

The most prevalent concern of feminists in criminology is that all of the major

criminological theories were developed without concern for gender and without even

considering female offending (Chesney-Lind, 1989; Daly & Chesney-Lind, 1988).

Females account for about 22% of all arrests, 11% of murder arrests (Greenfield & Snell,

1999) and their incarceration rates are growing at a faster rate than males (Beck, 2000).

Despite this, however, research on female offending has been limited.

While research on females and crime has increased since the second wave of

feminism, the majority of the research on female offending has been qualitative or at the

individual level (see Acoca, 1998; Bloom, Owen, Rosenbaum, & Deschenes, 2003;

Kaker, Friedman, & Peck, 2002; Koons-Witt & Schram, 2003). In addition, while some

have suggested that there are different pathways to crime for females than males; most of









this research has focused on the link between victimization and female offending (see

English, Widom, & Brandford, 2001; McClellan, Farabee, & Crouch 1997; Siegel &

Williams, 2003; Widom, 1989; Widom & Maxfield, 2001). Furthermore, when female

offending is examined it is still the "exceptional" case such as "battered woman

syndrome" (Comack, 1999). "So long as women are recognized only as victims and not

as active agents, there is little need to embrace or integrate feminist analyses into the

criminological agenda" (Comack, 1999, p. 165). Gender has not been properly

investigated in criminology and remains in the margins of criminology because of the

focus on victimization, and such a focus has neglected that men are also victims and that

women are capable of engaging in violent behavior (Comack, 1999).

Very few studies have actually examined female offending at the macro-level

with one notable exception being the work by Steffensmeier and Haynie (2000). They

examined the structural disadvantage correlates of female and male arrest rates for both

violent and property offenses in 178 cities. Their overall findings were that the structural

variables associated with predicting male arrest rates can also predict female arrest rates

just not as strongly. In addition, they found that the levels of offending were higher in

cities with increased levels of disadvantage, but that the effect was greater for violent

crimes. Overall there is a lack of gender specific studies in criminology at the macro-

level and those at the micro-level primarily focus on female victimization.



Importance of Place

Much like female offending, rural crime has also largely been ignored by the field

even though 25% of the U.S. population lives in rural areas with populations less than

2,500 (Weisheit & Donnermeyer, 2000). While research on rural crime is growing, most









of the research has focused on drug use (see Diala, Muntaner, & Walrath, 2004;

Donnermeyer, Barclay, & Jobes, 2002; Weisheit & Fuller, 2004), domestic violence (see

Davis, Taylor, & Furniss, 2001; Krishnan, Hilbert, & VanLeeuwen, 2001; Websdale,

1998; Websdale & Johnson, 1998), and community policing (see Jobes, 2003, 2002;

Liederbach & Frank, 2003; O'Shea, 1999;Weisheit, Wells, & Falcone, 1994). At the

same time, research regarding structural theories of crime has primarily focused on urban

areas. It is unknown whether such theories can explain rural crime. Such macro-level

theories need to be examined in rural areas to ensure that they are general theories of

crime and not urban specific theories because "theories that cannot account for both rural

and urban circumstances are limited in scope" (Weisheit & Donnermeyer, 2000, p. 310-

311) (see Appendix A for urban/rural definitions).

Recently in criminology there has been an increasing interest in contextual

studies. Place is one example of a context and rural areas may be a different context than

urban areas. Cebulak (2004) claims "...the context of rural crime, its causes and its

characteristics, are so different than for urban crime, we need a separate set of theories to

account for rural crime and justice" (p. 72). Some types of crime, such as theft of farm

animals or equipment or wildlife crimes, are limited to only rural areas (Cebulak, 2004;

Weisheit & Wells, 1999). Furthermore, there are unique features of rural communities

that may influence rural crime such as: physical distance and isolation, more informal

social control, low mobility and density, higher "density of acquaintanceship," mistrust of

government, reluctance to seek outside assistance, and "factory" farms and processing

plants (Weisheit & Donnermeyer, 2000; Weisheit & Wells, 1999). For more on the effect

of processing plants on rural communities and crime see Broadway (1990).









Contrary to common thought, rural crime and deviance is an issue. While violent

crime in large cities has declined since the early 1990's, rural crime rates have been

increasing (Weisheit & Donnermeyer, 2000). Between 1991 and 1997 urban violent

crime rates decreased by 531.8 per 100,000 but rural rates increased by 37.9 per 100,000

(Weisheit & Donnermeyer, 2000). DUI's are more common in rural areas, and rural

youth are more likely to use cigarettes or smokeless tobacco (Weisheit & Donnermeyer,

2000). In addition over the past 20 years, rural youth alcohol use has matched or

exceeded that of urban youth, and nonmetropolitan 12th graders in 1995 had higher use

rates for crack cocaine, stimulants, barbiturates, and tranquilizers than their metropolitan

counterparts (Weisheit & Donnermeyer, 2000)(See Appendix A for metropolitan and

nonmetropolitan definitions). Drug manufacturing, particularly for methamphetamine, is

of concern in rural areas. Missouri had more methamphetamine lab seizures than any

other state in 1997, with most of the seizures occurring in rural areas (Weisheit &

Donnermeyer, 2000), and there were 300 times more methamphetamine lab seizures in

Iowa in 1999 than in New York and New Jersey combined (Eagan, 2002). Suicide rates,

by any method, are also higher in rural areas (Butterfield, 2005). In the most rural

counties, the incidence of suicide with a gun is greater than the incidence of murder with

guns in major cities (Butterfield, 2005). From 1989 to 1999, the risk of dying from a

gunshot was the same in rural and urban areas the difference was who pulled the trigger

(Butterfield, 2005). Crime in rural America is important, but even though contextual

studies in criminology have increased, rural areas are still being neglected.









Research Questions

The primary research questions this study seeks to answer are as follows:

1. What are the impacts of structural conditions on female arrests?

a. What is the impact of patriarchy on female arrests for violent
and property crime?

b. What is the impact of social disorganization on female arrests
for violent and property crime?

2. Does the impact of these structural conditions on female arrests differ
in urban and rural areas?


Summary

This study fills a gap in the existing literature. The purpose of the current study is

to examine the impacts of structural conditions on female arrests within the context of

place. More specifically, this study examines the contribution of feminist and social

disorganization perspectives to addressing female violent and property crime arrests in

both urban and rural areas.

While this introductory chapter has primarily focused on the importance of

examining gender and place, the following chapter discusses the literature and empirical

findings in regards to patriarchy and crime. Chapter 2 also presents a theoretical overview

of social disorganization theory as well as empirical findings. Chapter 3 provides an

overview on the integration of the feminist perspective with social disorganization and

outlines the hypotheses. Chapter 4 presents the data and methodology. Chapter 5

discusses the results, and the final chapter provides an overall discussion and conclusions.














CHAPTER 2
STRUCTURAL THEORIES: GENDER AND PLACE

Feminist Perspective

A feminist perspective acknowledges that there are gender inequalities (Belknap,

2001) and places gender at the center of analysis (Atwell, 2002; Hartsock, 1998).

According to Daly and Chesney-Lind (1988) there are five elements of feminist thought

that separate it from more general types of thought. First, gender is a complex social,

historical, and cultural product. Second, gender and gender relations order social life and

institutions. Third, gender relations and constructs are based upon the superiority and

dominance of men over women. Fourth, our systems of knowledge reflect men's view of

the world and finally, women should be at the center of intellectual inquiry. Patriarchy is

key to feminist thought. Socialist feminism claims that both class and patriarchy are a

dual system of domination that explain women's subordination and radical feminism

claims that patriarchy is central to explaining women's position in society (Belknap,

2001).



What Is Patriarchy?

While scholars agree that gender inequality cannot be truly understood without

some understanding of patriarchy (Walby, 1986), there are many definitions of patriarchy

throughout the literature. Walby defines patriarchy as "a system of social structures, and

practices in which men dominate, oppress and exploit women" (1990, p. 20; 1989, p.

214). She suggests that patriarchy is composed of six main structures: the patriarchal









mode of production in the household, patriarchal relations in paid work, patriarchal

relations in the state, male violence, patriarchal relations in sexuality, and patriarchal

relations in cultural institutions. Table 2-1 illustrates that the six structures are present in

both public and private patriarchy. It is generally agreed upon that patriarchy exists in

both a public and a private sphere (Atwell, 2002). The public sphere includes institutional

structures such as the government, schools, and churches while the private sphere

encompasses the home and family.



Table 2-1. Private and Public Patriarchy

Private Patriarchy Public Patriarchy
Dominant Structure Household Production Employment and the State

Other Patriarchal Structures Employment Household Production
State Sexuality
Sexuality Violence
Violence Culture
Culture

Adapted from Theorizing Patriarchy by Sylvia Walby (1990, p. 24)


Similarly, Messerschmidt (1986) defines patriarchy as "a set of social relations of

power in which the male gender appropriates the labor power of women and controls

their sexuality" (p. x). Patriarchy is a system of hierarchy where men dominate women.

Men control and dominate women in the labor force, in the home, and in economic,

religious, political, and military systems. Messerschmidt, however, is specifically

concerned with explaining the relationship between patriarchy and crime. His work has

been incorporated into the criminological literature.









Patriarchy and Crime

According to Messerschmidt (1986) and Simpson (1991), the differences in

offending between males and females are due to the gendered social organization of

production and reproduction. It is because of this that powerless men engage in violent

crime while powerless women engage in property crime. It is therefore necessary to

examine measures of both economic inequality in the public (production/work) sphere

and patriarchal structures in the private (family/reproduction) sphere and the subsequent

effects on crime. Figure 2-1 provides a basic conceptual model of the relationship

between patriarchy and crime.




Private Patriarchy ++
(family/reproduction)
Female
+ Crime
Public Patriarchy
(production/work)


Figure 2-1. Relationship between Patriarchy and Crime



According to Messerschmidt (1986), female offending is a form of resistance and

accommodation to the oppressed status of women. He suggests that females are less

likely to engage in serious crime because their subordination and powerlessness isolates

them from engaging in such crimes. In addition, females have less opportunity to engage

in serious crime. He claims, instead, that women are more likely to engage in property

crime because of their oppressed economic status. For instance, Messerschmidt (1986)

suggests that female teenagers shoplift because they need certain items in order to please









men. Larceny and theft by female teenagers is a result of a combination of their

marginalized economic position, gender-role socialization, and control over female-

sexuality. He goes on to suggest that crime is also necessary for survival among adult

women due to female unemployment and the feminization of poverty. "Many women

turn to gender-specific illegal options, such as fraud, to put food on the table"

(Messerschmidt, 1986, p. 84). Overall, women engage in nonviolent offenses merely as a

means to accommodate their powerless position in society.



Empirical Findings

Literature examining the impact of patriarchy or gender inequality on female

crime has largely focused on the public sphere by using the economic marginalization

hypothesis. Economic marginalization refers to the economic disadvantage of women

relative to that of men (Heimer, 2000). The general idea is that as the disadvantage of

females increases, their involvement in crime will also increase. It proposes that the

increased hardship of women relative to men is the reason why the gender gap has been

narrowing in the past few decades. Only a few studies have directly tested the economic

marginalization hypothesis and they have found limited support for the hypothesis (Box

& Hale, 1983, 1984; Hunnicutt & Broidy, 2004; Steffensmeier & Streifel, 1992).

Box and Hale (1983, 1984) found limited support for the economic

marginalization hypothesis. They examined female conviction rates from 1951-1979 in

England and Wales and found that female unemployment was associated with conviction

rates for both violent and property offenses but that female unemployment was better at

explaining property convictions. Heimer (2000), however, criticized their studies for only









looking at one measure of economic well being and for using variables that measured

women's absolute poverty rather than their well-being compared to that of men.

Steffensmeier and Streifel (1992) examined U.S. arrest trends from 1960-1985

and found that female arrest was correlated with policing trends and with economic

marginalization. Their primary measure of economic marginalization was single-parent

female headed households. They found the percent of female headed households to be

significant and positively related to major property crime and with burglary and

prostitution. Again Heimer (2000) criticized their study for using only limited, absolute

measures. They did list the female unemployment rate relative to males as one of their

variables; however, they did not discuss any findings in their paper in regards to such a

measure.

More recently, Hunnicutt and Broidy (2004) examined 1975-1994 adult

conviction rates by gender per 100,000 in ten countries (Austria, Denmark, Hungary,

Greece, Panama, Portugal, Italy, Chile, Sweden, United Kingdom, and United States) and

also found some support for economic marginalization. Contrary to expectations, they

found no support for unemployment being related to conviction rates; however, they did

find female employment in the service sector to be positively related to female

conviction. They also found divorce to be positively associated with female conviction

rates. While they originally included divorce as a measure of the liberation hypothesis,

they themselves suggest that it may in fact be measuring economic marginalization. Like

the previous studies, though, all the measures used were absolute measures and did not

look at women's status relative to that of men's.









Few other attempts have been made to try to conceptualize patriarchy at the

macro-level. Most studies that look at patriarchy and crime do so at the individual level.

One exception is the work by Yllo (1983, 1984) and Yllo and Straus (1984) that

examined the relationship between the status of women at the state level and violence

against wives. In all three studies data were aggregated at the state level (N = 30). Gender

equality was measured on four different dimensions economical, educational, political,

and legal. The first study (1983) found that violence against wives was highest in the

least egalitarian states as well as the most egalitarian states. In the second study (1984)

she found the highest levels of violence to be in male-dominated families in the more

egalitarian states. In the third study, Yllo and Straus (1984) again found that violence

against wives was highest in the least egalitarian and the most egalitarian states, that

more patriarchal norms coincided with more violence, and that violence was highest

when there was inconsistency between women's public status and their status within

marriage. Like most feminist pieces, however, the focus of these studies was on male

offending and female victimization. Subsequently, female offending was neglected.

Whaley and Messner (2002) examined gender equality and homicide rates in 191

cities. They used five gender inequality indicators ratio of male to female median

income, the percent of males ages 16 and above who were employed in the civilian labor

force relative to the percent of females, the percent of executives, managers, and

administrators who were male, the percent of those employed in the labor force who were

male, and the ratio of men to women ages 25 and above with four or more years of

college. They also included sex-specific measures of economic disadvantage -

percentage black, percentage poor, percentage unemployed, and the Gini index. Gender









equality had no effect on females killing males or on females killing females while the

index of economic disadvantage was positively related to females killing both males and

females.

Several other studies have also examined the effects of gender inequality on

offending while not specifically taking a feminist approach. Most notably, Steffensmeier

and Haynie (2000) examined the effects of structural disadvantage on female and male

arrest rates from 1987-1993 in 178 cities. Some of the independent variables included

were female poverty, female joblessness, and female-headed households. Again, these

were all absolute measures but they did find that levels of offending were higher for both

genders in cities with higher levels of social and economic disadvantage. Furthermore,

the structural variables were robust for predicting male and female rates; however, the

effect was greater for serious crimes. In addition, they found that structural disadvantage

was more strongly related to partner homicides by females than by males.

DeWees and Parker (2003) examined the relationship between women's

economic and social status and homicide in 162 cities using six gender specific

indicators: the percent of females employed in managerial and professional occupations,

the percent of females working part-time, female median income, percent of females

living below poverty, the percent of females with a bachelor's degree or higher, and the

percentage of women married. Female poverty was positively significant in the total,

family, and acquaintance homicide models. Marriage reduced homicide in the total,

intimate, and acquaintance models but was not significant in the family model. Part-time

employment was found to reduce female offending in the total and intimate models, but

the index of inequality had no significant effect in any of the models. They did find,









however, that the indicators were better at explaining female offending in non-southern

cities than in the South. Again, however, there is the concern that the research only used

absolute measures of women's status.



Summary

While scholars distinguish between private and public patriarchy, most of the

literature examining the relationship between patriarchy and crime at the macro-level has

focused on public patriarchy, particularly economic marginalization. The relationship

between private patriarchy and crime has not been tested at the macro-level. While there

is some support for the economic marginalization hypothesis, only a few studies have

directly tested it and these studies have only used limited, absolute measures of women's

status instead of measures that capture women's status relative to that of men. Some other

studies have examined the relationship between public patriarchy and crime; however,

the majority of this research has focused on patriarchy and female victimization (see

Yllo, 1984, 1983; Yllo & Straus, 1984). While DeWees and Parker (2003) and

Steffensmeier and Haynie (2000) found gender inequality to be related to female

offending, they too only used absolute measures of women's status. Overall, there is

some support for the notion that female disadvantage, primarily female poverty as an

absolute measure, is related to female offending. Similarly, social disorganization theory

also examines the relationship between poor economic status and crime.



Social Disorganization Theory

In the past 20 years there has been a renewed interest in social disorganization

theory. Social disorganization theory has its roots in Chicago. Park and Burgess (1925)










emphasized the importance of looking at natural areas and their characteristics. Areas are

considered to be functioning and changing organisms. Burgess (1925) suggested that

cities expand and grow outward in concentric circles from the core of the city zone one.

Zone two is the transition zone. It is generally the oldest and poorest zone. The third zone

consists of workers' homes. The fourth zone is the residential zone and consists of single

family housing, and finally the fifth zone is the commuter zone or suburbs. Invasion is

constantly occurring in all of these zones as new people move to the city and current

residents try to move outward into a different zone. The zones are displayed in Figure

2-2.


Zone 1 City Core

Zone 2 Transition

Zone 3 Workers

Zone 4 Residential

Zone 5 Commuter











Figure 2-2. Burgess's Zones for City Growth



In Juvenile Delinquency and Urban Areas, Shaw and McKay (1942) examined

juvenile delinquency in Chicago. Their main argument was that poor economic status,

population heterogeneity, and residential mobility lead to social disorganization. Social

disorganization in turn leads to breakdowns in conventional attachment and informal and









formal control and therefore results in more crime. More recently, it has been suggested

that family disruption (Sampson, 1987; Sampson & Groves, 1989) and population size

and density (Mayhew & Levinger, 1976) also contribute to social disorganization. The

key theoretical concepts of social disorganization theory are displayed in Figure 2-3.

Several studies have examined social disorganization theory and have found support for

the theory in urban areas (Kposowa, Breault, & Harrison, 1995; Lee, Maume, & Ousey,

2003; Petee & Kowalski, 1993; Sampson, 1991; Sampson & Groves, 1989). These

studies are discussed below.


Figure 2-3. Basic Theoretical Model of Social Disorganization Theory


Empirical Findings

Sampson and colleagues have conducted several studies on social disorganization

theory. Sampson (1987) examined the relationship between family disruption and crime.

Using 1980 homicide and robbery rates for 171 cities, he found family disruption, region,









and density to be related to juvenile robbery, and family disruption, region, population

size, and housing density were related to adult robbery. Family disruption was also found

to be related to adult homicide. Furthermore, family disruption and population size were

also related to black juvenile homicide, and region and city size were associated with

black adult homicide.

Sampson and Groves (1989) expanded the research on social disorganization

theory. Unlike previous studies, they measured the levels of social organization within

communities besides using the standard structural indicators of social disorganization.

Based on data from the 1982 British Crime Survey (N=10,905 individuals, N= 238

localities), they found family disruption, urbanization, and ethnic heterogeneity to be

associated with more disorderly peer groups. Urbanization was negatively associated

with friendship networks, and socioeconomic status was positively related to

organizational participation. Further analysis revealed that the structural measures of

social disorganization were mediated by the community organization measures. In other

words, community social disorganization accounts for much of the effect of

socioeconomic status, residential stability, family disruption, and ethnic heterogeneity on

crime. In an extension of the previous study, Sampson (1991) analyzed data from the

1984 British Crime Survey (N= 11,030 individuals, N= 526 polling districts) and found

that residential stability has a direct effect on community-based social ties which in turn

increases the level of community cohesion.

In yet another study, Sampson, Raudenbush, and Earls (1997) tested the notion

that concentrated disadvantage and residential instability decrease collective efficacy and

whether in turn, collective efficacy explains the relationship between neighborhood









disadvantage and crime. They interviewed 8,782 residents of 343 neighborhood clusters

in Chicago. The results supported their hypothesis. The effects of concentrated

disadvantage and residential instability on violence were mediated in a large part by

collective efficacy.

Consistent with social disorganization theory, Kposowa, Breault, and Harrison

(1995) examined crime in counties with a population larger than 100,000 (N= 408) and

found poverty to be a significant predictor of violent crime, and church membership and

divorce were significant predictors of property crime. When examining crime in all

counties (N = 3,076), the strongest predictor of property crime was urbanity, but percent

black, percent Hispanic, population change, and unemployment were also significant.

Predictors of violent crime in the total county sample were percent black, percent

Hispanic, church membership, urbanity, and population density. When examining

homicide in the large counties (N = 408), they found percent black, Gini coefficient,

divorce rate, and population change to be the strongest predictors. Contrary to social

disorganization theory though, they found poverty to be significantly related to homicide

but in the negative direction.

Lee, Maume, and Ousey (2003) also examined the relationship between

socioeconomic disadvantage and poverty on the homicide rate average from 1990-1992

in 778 metropolitan counties. Unlike Kposowa and colleagues (1995), Lee and colleagues

(2003) did find poverty concentration to be significant and positively related to homicide.

Disadvantage was also found to be significant and positively related to homicide.

Land, McCall, and Cohen (1990) also examined the structural correlates of

homicide rates across time 1960, 1970, and 1980 and across space cities, SMSA's









(see MSA in Appendix A), and states. While not specifically testing social

disorganization theory, they did find support for the indicators of social disorganization.

Resource deprivation was associated with higher homicide rates across all the time

periods and locations. In addition, population structure an index of population size and

density and percent divorced were also related to homicide across most of the models.



Summary

Overall, while there is much support for social disorganization theory, there are

numerous inconsistencies in terms of which social disorganization indicators are

significant and for which types of offenses. Furthermore, some studies have findings

opposite of what social disorganization theory would suggest. For instance when looking

at the predictors of homicide in metropolitan counties, Kposowa and colleagues (1995)

found poverty to be significant but in the negative direction; however, Lee and colleagues

(2003) found poverty concentration to have a significant positive effect on homicide in

metropolitan counties.

Conclusion

This chapter has discussed the feminist perspective in regards to patriarchy and

crime as well as social disorganization theory and the empirical studies available on both.

While there is support for both theories, the feminist perspective has only examined the

relationship between public patriarchy and crime at the macro-level, and social

disorganization theory has focused on and was developed around urban areas. It is

unknown how adequate social disorganization theory is at explaining crime in nonurban

areas. The following chapter discusses the need to examine the influence of both private






19


and public patriarchy as well as the context of place (urban and rural) on female

offending.














CHAPTER 3
WOMEN AND CRIME IN CONTEXT

It is necessary to examine both gender and place when investigating female

offending. Gender is a fundamental element of criminology that cannot be ignored

(Comack, 1999), and patriarchy is essential for capturing the "...the depth, pervasiveness,

and interconnectedness of different aspects of women's subordination..." (Walby, 1990,

p. 2). Likewise, place is also a fundamental aspect of criminology, and "...it is past time

for criminological theories and methods to include the rural context" (Weisheit & Wells,

1996, p. 384). In addition, the structures of patriarchy manifest themselves differently in

rural and urban areas (Websdale, 1998). For instance, rural women are more involved in

household production and less involved in the public sphere, and there is also a more

traditional division of labor in rural areas (Websdale, 1998).



Patriarchy and Gender

The current study is the first research to attempt to measure private patriarchy at

the macro-level. In an effort to build upon the feminist literature, the current study

includes multiple measures of the feminist perspective. Traditional, absolute measures of

female poverty and part-time work are used to measure public patriarchy (economic

marginalization). In addition, measures of women's position relative to that of men in the

workforce are also used to measure public patriarchy (economic marginalization) in order

to address Heimer's (2000) concern about only using absolute measure to conceptualize

economic marginalization. Such measures allow for us to capture both the economic









marginalization of women and public patriarchy in the workforce of male dominance

over females in the labor force and market.

To date, no studies have examined the influence of private patriarchy on female

offending at the structural level. While the liberation hypothesis originally proposed that

female equality would lead to an increase in female crime (see Adler, 1975; Simon,

1975), it has essentially been discounted (Belknap, 2001; Heimer, 2000). Similarly,

power-control theory also suggests that female juvenile delinquency will increase in more

egalitarian families (see Hagan, Gillis, & Simpson, 1985; Hagan, Simpson, & Gillis,

1988, 1987; McCarthy, Hagan, & Woodward, 1999); however, there is mixed and

inconsistent support for the theory (Belknap, 2001) and some suggest it is merely the

liberation hypothesis reworded. Messerschmidt (1986) suggests that females who are

confined and disadvantaged in the home are more likely to hurt themselves instead of

hurting others. Because of this, I propose that private patriarchy will have no significant

effect on female violent arrests in urban or rural areas. In terms of property crime,

Messerschmidt (1986) also argues that females engage in property crime because of

capitalism and patriarchy and that in order to curb crime we need to reduce power and

class and gender inequality. Therefore, I predict that indicators of private patriarchy will

have a positive effect on female arrests for property crime in both urban and rural areas.



Hi: Indicators of private patriarchy (family/reproduction sphere) will have no
effect on female violent crime arrests and will have a positive effect on female
property crime arrests in both urban and rural areas.









Based on the economic marginalization hypothesis, I predict that indicators of

public patriarchy will have a positive impact on female arrests for violent and property

crime. As indicated in the previous chapter, some studies have found support for the

economic marginalization hypothesis. For example, Box and Hale (1983, 1984) found

economic marginalization (female unemployment) to be positively associated with

female violent and property offenses, and Steffensmeier and Streifel (1992) found

economic marginalization (percent female headed households) to be positively related to

female arrests for major property crime, burglary, and prostitution. Other studies have

also found a relationship between female structural disadvantage and crime. DeWees and

Parker (2003) found female poverty to be related to homicide, Whaley and Messner

(2002) found female economic disadvantage to be related to females killing, and

Steffensmeier and Haynie (2000) found female structural disadvantage to be related to

both violent and property offenses.



H2: Indicators of public patriarchy (economic marginalization), patriarchy in the
work and production spheres, will have a positive effect on female arrest for
both violent and property crime in urban and rural areas.



Social Disorganization and Place

Most research on social disorganization has focused on urban areas; however,

social disorganization theory can also be applied to rural areas. The structural conditions

usually associated with social disorganization are not unique to urban areas. In fact,

nonmetropolitan poverty rates have exceeded metropolitan poverty rates every year since

poverty was first officially measured in the 1960s (Rural poverty, 2004). In addition, the









2003 unemployment rate was 5.8% in nonmetropolitan areas and 6.0% in metropolitan

areas (Rural America, 2004). Albrecht and colleagues (2000) examined how poverty

levels in rural America have been affected by industrial transformation. Using 1990

census data on 2,390 nonmetropolitan counties, they argue that Wilson's model for the

inner city underclass can be used to understand increased levels of rural poverty and the

growth of the rural underclass (for more on rural ghettos see Davidson, 1996).

Population mobility is also an issue in rural areas. Over 1,000 nonmetropolitan

counties have lost population since 2000, primarily counties in the Great Plains, but there

are also fast growing nonmetropolitan recreational counties in the South and West and

the growth of the Hispanic population has contributed to nonmetropolitan county

population growth in the West, South, and Midwest (Rural America, 2004). Population

turnaround in rural counties has been linked to a variety of social problems including

problems in education, community solidarity, heath care, social welfare, and crime (Price

& Clay, 1980).

In addition, minorities comprise 17% of nonmetropolitan residents and 15% of

nonmetropolitan families are headed by a single female (Jolliffe, 2003). Snyder and

McLaughlin (2004) found that poverty in nonmetropolitan areas closely resembles that in

central cities. The risk of poverty for female-headed families and subfamilies with

children is significantly higher for those living in nonmetropolitan areas compared to

those living in central cities and suburban areas.

A few empirical studies of social disorganization in rural areas have provided

some support for social disorganization theory (Barnett & Mencken, 2002; Kposowa,

Breault, & Harrison, 1995; Lee, Maume, & Ousey, 2003; Osgood & Chambers, 2000;









Petee & Kowalski, 1993). Petee and Kowalski (1993) tested social disorganization theory

on violent crime rates in 630 rural counties from 1979-1980. They found residential

mobility to have the greatest impact on violent crime followed by single-parent

households, and then racial heterogeneity. Barnett and Mencken (2002) applied social

disorganization theory to violent and property crime rates circa 1990 in 2,254

nonmetropolitan counties. They found resource disadvantage to have a significant

positive effect on violent crime, and population change was significant and positively

related to property crime.

While Kposowa, Breault, and Harrison (1995) did not specifically test social

disorganization theory, they did incorporate standard measures of social disorganization

when examining the structural correlates of crime rural counties (N= 1,681). They found

population change to be significant and positively related to both violent and property

crime. Church membership, divorce rate, and percent Native American were also found

to be significant predictors of violent crime, and percent Hispanic was found to be a

significant predictor of property crime. Contrary to social disorganization theory though,

poverty was not found to be a significant predictor of violent or property crime. They did,

however, find poverty, the South, divorce rate, and population change all to be significant

when only examining homicide in the small counties.

Lee, Maume, and Ousey (2003) however found poverty to not be significant when

examining the average homicide rate from 1990-1992 in 1,746 nonmetropolitan counties.

They did find a significant and positive relationship between disadvantage and homicide

though. Osgood and Chambers (2000) also found no meaningful relationships between

indicators of economic status, poverty, or unemployment on the juvenile violent crime









arrest rate in 264 rural counties. Consistent with social disorganization theory they did

find residential instability, female headed households, and ethnic heterogeneity to be

associated with juvenile arrests.

Jobes, Barclay, Weinand, and Donnermeyer (2004) also found none of their

economic measures to be significant when examining crime rates in 123 rural LGA's in

Australia. In support of social disorganization theory, though, they did find residential

instability and family instability to be associated with higher rates of crime. Higher

proportions of indigenous people were also associated with higher rates of assault and

with "break and enter" crimes.

Arthur (1991) examined the socioeconomic predictors of violent and property

crime in 13 rural Georgia counties from 1975-1985. He found the percent of the

population below poverty, percent families receiving aid, unemployment, and percent

black to predict both violent and property crime. Contrary to most findings at the urban

level, however, he found the variables to be better predictors of rural property crime than

of rural violent crime.

Based on the review of the literature, there are numerous inconsistencies in the

findings when examining social disorganization in urban areas. Likewise, it appears these

inconsistencies in the findings persist when applying social disorganization theory to

rural areas, especially when examining economic disadvantage indicators. Barnett and

Mencken (2002) found resource disadvantage to have a significant positive impact on

violent crime. Kposowa and colleagues (1995), however, did not find poverty to be a

significant predictor of violent crime arrests in rural areas, but they did find poverty to be

predictive of homicide rates. When Lee and colleagues (2003) examined rural homicide









rates, they found poverty to be nonsignificant. Osgood and Chambers (2000) also found

no meaningful relationships between the indicators of economic status, poverty, or

unemployment on juvenile arrests.

Furthermore, all of the studies applying social disorganization theory to crime in

rural areas have used nonmetropolitan counties as the unit of analysis. This is problematic

when applying social disorganization theory because generally counties are comprised of

numerous communities. In addition, all of the previous studies employ data before 2000.

It is necessary to test social disorganization theory in rural areas using more recent data

because in 2003 the OMB changed the definitions of metropolitan/nonmetropolitan and

urban/rural, thereby changing the rural-urban continuum codes that these previous studies

employed. Furthermore, the work by Lee and colleagues (2003) only examined homicide

rates and the work by Osgood and Chambers (2000) only examined juvenile arrest rates

from counties in only four states. Also, all of these studies examined either total crime or

male crime. The current study addresses these limitations and advances upon these

previous efforts by using towns as the unit of analysis, by examining both violent and

property arrests for females, by using a national sample, and by using more recent data.

Consistent with social disorganization theory, I propose that the measures of

social disorganization will have a positive effect on female arrests for violent and

property crime. As indicated in this chapter and the previous chapter, numerous studies

have found support for social disorganization indicators in predicting crime in both urban

cities and in nonmetropolitan counties.



H3: Indicators of social disorganization will have a positive impact on female
arrest for both violent and property crime in both rural and urban areas.









While conditions of female subordination and structural disadvantage are not

unique to urban cities, there may be differences in the ability of these indicators to predict

female crime within the context of place. Wells and Weisheit (2004) examined the

structural covariates of urban and rural crime. It is the only study that has specifically

compared the regression coefficients of the structural correlates of metropolitan and

nonmetropolitan crime. Overall, they found that the standard structural models for urban

crime explain more variance in the urban property and violent crime models than in the

rural models. Population change and family instability were significant across all models

while racial diversity was consistent across only the models for violent crime. On the

other hand, economic resources index was very inconsistent across models. It was

inversely related to both property and violent crime in urban counties but in

nonmetropolitan rural counties it was unrelated to violent crime and positively related to

property crime. When comparing regression coefficients for the urban and

nonmteropolitan rural counties, they found the indicators of population change,

household instability, cultural capital, economic resources, percent of population 15-24,

and unemployment rate to be significantly different in predicting violent crime in large-

city counties versus rural, small-town counties. In addition, they found the indicators of

population change, urban density, economic stability, cultural capital, economic

resources, and unemployment rate to be significantly different in predicting property

crime in the large versus small counties.

Overall, Wells and Weisheit's (2004) research highlights the need to further

examine rural crime to determine whether or not structural theories of crime can be

applied across communities regardless of population size. Consistent with their findings,









I propose that the impact of some of the structural indicators on female arrests for violent

and property crimes will not be the same in urban cities and rural towns.



H4: While it is expected that these structural indicators will have a positive effect
on female crime in both urban and rural areas, the impact of these structural
indicators on female arrests will not be the same across the two localities.



Summary

Both the feminist perspective and social disorganization theory contribute to

explaining female offending (see Figure 3-1). The feminist perspective focuses on the

influence of patriarchy on crime while social disorganization theory focuses on the

ecological conditions that contribute to crime. Both, however, focus on the effects of

disadvantage on crime. The feminist perspective suggests that female structural

disadvantage relative to males in both the home and in the labor market contributes to

female offending by making crime necessary in order for females to survive and

overcome their disadvantaged positions. Therefore, concentrated disadvantage of females

relative to males within an area leads to an increase in female crime. Similarly, social

disorganization theory suggests that concentrated disadvantage (low economic status,

ethnic heterogeneity, residential mobility, family disruption, and population size) within

an area also leads to an increase in crime due to a breakdown in informal control.

Therefore, female offending will also be higher in disorganized localities. Overall, the

hypotheses reflect the primary concern that female disadvantage and female offending

will be exasperated in areas that exhibit both patriarchy and social disorganization.









By examining female offending by type of offense (property and violent) and by

place (urban and rural), this study offers a systematic analysis that hopefully addresses

the inconsistencies amongst previous studies. Overall, this study simultaneously

addresses the importance of both gender and place. It is the first study to examine female

offending by type of offense and by place. It is also the first study to assess the impact of

patriarchy on female offending at the macro-level. This study builds upon the feminist

literature by quantitatively analyzing the impact of both economic marginalization in the

class sphere (public patriarchy) and patriarchy in the family sphere (private patriarchy) on

female offending. Furthermore, this study builds upon the literature illustrating the

importance of place by looking at both urban and rural areas. Instead of counties, urban

cities and small, rural towns are compared as recommended by Wells and Weisheit

(2004).






30



Private Patriarchy


Public Patriarchy


Low Economic
Status

Residential
Mobility Female Crime

Family Disruption


Population Size


Controls


Figure 3-1. Proposed Relationship between Patriarchy, Social Disorganization, and
Crime. All proposed relationships are positive, except that between the
indicators of private patriarchy and female violent crime which is expected to
not be significant














CHAPTER 4
DATA AND METHODOLOGY

Source of the Data

Data were obtained from multiple sources. UCR arrest data from 2001 were used

to measure the dependent variables. Summary files one and three of the 2000 census were

the sources of data for the independent variables (see Appendix A for more information

on summary files one and three). The number of police officers employed by the cities

and towns were obtained from the 2000 Police Employee (LEOKA) Data (ICPSR 3445).

The Law Enforcement Agency Identifiers Crosswalk 2000 (ICPSR 4082) was used to

merge the arrest data with the census data. In addition, the 1993 and 2003 Beale Codes

(see Appendix B) were employed to help define the rural sample and to calculate a

control measure.

Unit of Analysis

While most studies of social disorganization in urban areas use cities or census

tracts as the unit of analysis, all of the studies that have applied social disorganization

theory to rural crime have used counties as the unit of analysis. Using counties as the unit

of analysis is potentially problematic when applying social disorganization theory

because counties generally consist of numerous communities. The current study

addresses this concern by using urban cities and rural towns as the unit of analysis

instead.

For this study, urban cities are defined as those cities with populations larger than

100,000. This is consistent with previous studies that have examined crime in urban









areas. It is also consistent with the census in terms that in order for an area to be

considered a MSA, it has to be an urbanized area with a population of at least 100,000

(see Appendix A for definitions of urban and urbanized areas). The urban sample consists

of 200 cities with populations larger than 100,000. The cities range in size from 99,716 to

8,008,278 with an average population of 330,022 (SD = 677,267.9) and an average

female population of 51% (SD = 1.23).

Defining rural is more complicated. There is a debate on whether definitions

based on population or those based on culture should be used. There is no consensus on a

cultural definition of rural however. As a result population size is generally used. The

census simply defines rural as territory, population and housing units not classified as

urban (see Appendix A for definition of rural). In order to be considered urban by the

census the area has to contain at least 2,500 people. Previous studies that have examined

crime in rural areas have used counties as the unit of analysis so they have relied on the

rural-urban continuum codes, otherwise known as the Beale codes. Because this study is

using towns instead of counties, these codes cannot be fully utilized. However, consistent

with the census definition and the logic of the Beale codes, for the purposes of this study

rural towns are defined as those towns with a population less than 2,500 not located

within a MSA (see MSA in Appendix A), and only such towns that are located within

counties designated as nonmetropolitan by the 2003 Beale Codes (see Appendix B). As

such, the rural sample consists of 787 towns. The rural towns range in size from 80 to

2,595 (M= 1450.69, SD = 588.234) with an average female population of 52.70% (SD=

3.66).









Measures


Dependent Variables

The two dependent variables are the female violent crime count and the female

property crime count. They consist of female arrests for each locality as reported in the

2001 UCR arrest data. The sample is limited to only municipal agencies that reported at

least nine months of the year and to only those that were municipal agencies in cities over

100,000 or cities under 2,500 not located within a MSA (see MSA in Appendix A). The

female violent crime count is an index of murder, robbery, and aggravated assault. The

female property crime count is an index of burglary, larceny, forgery, and fraud. Forgery

and fraud are included in lieu of the traditional index offenses of arson and motor vehicle

theft due to the nature of female offending. Steffensmeier and colleagues have conducted

numerous studies on female arrest trends and have found that female arrests have not

increased for serious crimes but that female arrests have increased in terms of minor

property offenses such as larceny, forgery, and fraud primarily due to changes in police

practices, economic factors, and opportunities for females to engage in such crimes

(Steffensmeier, 1980, 1993; Steffensmeier & Cobb, 1981). Similar arrest patterns have

also been found when examining female arrest trends in rural areas (Steffensmeier &

Jordan, 1978).



Independent Variables

Essentially both the feminist perspective and social disorganization theory are

being tested in this study. As such, numerous measures of private patriarchy, public

patriarchy, and social disorganization are utilized as independent variables. These

variables are described below.









Private patriarchy

Private patriarchy (patriarchy in the family/reproduction sphere) is difficult to

conceptualize at the town level with census data. No research has attempted to try to

capture the impact of traditional family structure on female offending except for that of

DeWees and Parker (2003) which included a measure of the percent of women married.

Four measures of private patriarchy are included in this study. These measures are: the

percent of families where the husband works and the wife does not, the percent of

families married with children, the percent of females working full-time with no income,

and the percent of employed females who are unpaid family workers. These variables are

defined below.



Percent of families where the husband works and the wife does not. The first

measure of private patriarchy is the percent of families where the husband works and the

wife does not. This measure is an attempt to capture traditional family structure. A wife

not in the labor force includes housewives and wives only doing incidental unpaid family

work. This measure is calculated by dividing the number of married couple families

where the husband is in the labor force and the wife is either not in labor force or

unemployed by the total number of married couple families. This is then multiplied by

100 to obtain a percent. The definitions for family, married couple family, not in the labor

force, unemployed, and unpaid family work are provided in Appendix A.



(No. married couple families with husband working and wife not in the labor force or unemployed)
(No. married couple families) 100
(No. married couple families)









Percent of families married with children. The second measure of private

patriarchy is the percent of families that are married with children. This measure is also

an attempt to conceptualize traditional family structure. It is calculated by dividing the

number of married couple family households (where the householder is 15-64 years of

age) with children under 18 years of age by the total number of family households

(householder 15-64 years of age). This is then multiplied by 100 to obtain a percent. The

definitions for household, household type, householder, and married couple family are

provided in Appendix A.



(No. of married family households with children under 18 years of age)
(No. of family households)



Percent of females working full-time with no income. Another measure of

private patriarchy is the percent of females working full-time with no income. This

measure is another attempt to capture traditional family structure by capturing women

working in the home without income. It is calculated by dividing the number of females

ages 15 and over who worked full-time, year round in 1999 with no income divided by

the number of females ages 15 and over that worked full-time, year round in 1999.

Appendix A provides the census definitions of income, worked in 1999, and full-time,

year-round workers.



(No. of females > 15 years of age who worked fulltime, year round in 1999 with no income) 100
(No. of females > 15 years of age who worked full time, year round in 1999)









Percent of employed females who are unpaid family workers. A final measure

of private patriarchy is the percent of employed females who are unpaid family workers.

This is yet another attempt to capture the impact of traditional family structure. Unpaid

family workers include people who worked 15 hours or more without pay in a business or

on a farm operated by a relative. It is hoped that this variable will capture those females

that are working for family members for free, especially in rural areas where it may be

more likely that females are working on family farms without pay. Refer to Appendix A

for definitions of employed and unpaid family workers.



(No. employed civilian females > 16 years of age who are unpaid family workers) 100
--- ------------- --- ----------- ---- x lO O
(No. employed civilian females > 16 years of age)



Public patriarchy (economic marginalization)

Consistent with the works by Steffensmeier and Haynie (2000) and Whaley and

Messner (2002), measures of public patriarchy (patriarchy in the work and production

sphere) include the ratio of male to female median income, the ratio of males to females

ages 25 and above with a bachelors degree or higher, the ratio of males to females ages

16 and above in management and professional occupations, the percent of females ages

16 and above working part-time (1-34 hours per week), and the percent of females living

below poverty. Such measures capture economic marginalization but also patriarchy in

the class sphere in terms of male dominance. It is important to use ratios and not just

absolute measures because the economic marginalization hypothesis is based upon the

status of females relative to males (Heimer, 2000). These measures are defined below.









Ratio of male to female median income. One measure of public patriarchy is the

ratio of male to female median income. This is consistent with the work by Whaley and

Messner (2002). The ratio is calculated by dividing the median male income by the

median female income for 1999. The greater the value of the ratio is, the greater the

disadvantage of females relative to males. The census definitions for income and median

income are provided in Appendix A.



(Median income for the male population > 15 years of age with income in 1999)
(Median income for the female population > 15 years of age with income in 1999)



Ratio of males to females with bachelor's degrees or higher. A second

indicator of women's status relative to that of men is the ratio of males to females with a

bachelor's degree or higher. Consistent with Whaley and Messner (2002), the ratio is

calculated as the percent of men ages 25 and over with a bachelor's degree or higher

divided by the percent of females ages 25 and over with a bachelor's degree or higher.



(Percent of males > 25 years of age with a bachelor' s degree or higher)
(Percent of females > 25 years of age with a bachelor' s degree or higher



Ratio of males to females in management and professional occupations.

Another measure of public patriarchy is the ratio of males to females in management and

professional occupations. In order to measure women's status in the workforce, Whaley

and Messner (2002) use the percent of executives, managers, and administrators who are

male, and DeWees and Parker (2003) use the proportion of females ages 16 and over









employed in management and professional occupations. Both of these measures,

however, are absolute measures. This study employs a ratio instead. The ratio is

calculated as the percent of civilian employed males ages 16 and over in management,

professional, and related occupations divided by the percent of civilian employed females

ages 16 and over in such occupations. Appendix A provides the census definition of

occupation.



(% of civilian employed males 16 years of age in management and professional occupations)
(% of civilian employed females > 16 years of age in management and professional occupations)



Percent of females working part-time. The fourth indicator of public patriarchy

is the percent of females working part-time. Consistent with DeWees and Parker (2003)

this percentage is based on the number of females ages 16 and over who worked 1-34

hours per week in 1999 divided by the total number of females ages 16 and over that

worked in 1999. The result is then multiplied by 100 to obtain a percent. Refer to

Appendix A for the census definition of worked in 1999.



(No. females > 16 years of age who worked 1- 34 hours per week in 1999)100
(No. females > 16 years of age who worked in 1999)



Percent of females living below poverty. DeWees and Parker (2003),

Steffensmeier and Haynie (2000), and Whaley and Messner (2002) all utilized female

poverty as an indicator of women's status. For the current study, the percent of females

living below poverty is based on the number of females for whom poverty status was









determined in 1999 with incomes below the poverty line divided by the total population

for whom poverty status was determined in 1999. The result is then multiplied by 100 to

obtain the percent. It was decided to divide by the total population rather than the female

population to better capture women's status relative to that of men. Appendix A provides

the census definitions of poverty and individuals for whom poverty status is determined.



(Female population for whom poverty status was determined in 1999 with incomes below poverty level)100
L (Total population for whom poverty status was deterermined in 1999)



Social disorganization

Standard measures of social disorganization are used in this study. They include:

residential mobility, population change, percent divorced, and population size. The total

percent of the population living below poverty, while standard in studies of social

disorganization, is not included here because it is highly collinear with the public

patriarchy (economic marginalization) measure of female poverty which is more

important since the study is taking a gendered approach. Likewise, percent black is not

being used as a measure of social disorganization because it too is highly correlated with

female poverty. The measures of social disorganization utilized in this study are defined

below.



Residential mobility. The first measure of social disorganization is residential

mobility. Consistent with numerous other studies residential mobility is calculated by

dividing the number of people ages 5 and over that lived in a different house in 1995 by









the total number of people ages 5 and over. This is then multiplied by 100 to obtain the

percent.



(Population 2 5 years of age that lived in a different house in 1995)100

(Population 2 5 years of age)



Population change. Population change is also included as a standard measure of

social disorganization. Population change is calculated by subtracting the 1990

population from the 2000 population and then dividing by the 1990 population. The result

is then multiplied by 100 to obtain the percent.



(2000 population -1990 population)-
(1990 population)



Percent divorced. The percent divorced is another standard measure of social

disorganization. Divorced refers to those people who are legally divorced and have not

remarried. For this study it is calculated as the number of people ages 15 and over who

are divorced divided by the number of people ages 15 and over. The result is then

multiplied by 100 to obtain a percent.



(No. people > 15 years of age that are divorced) x100
(No. people > 15 years of age)









Population size. Population size is simply measured as the natural logarithm

transformation of the 2000 population.



LN (2000 population)



Controls

Control measures include whether the county changed from metropolitan to

nonmetropolitan or vice versa between 1993 and 2003, the Hispanic population, South,

West, and officer rate. The control variables are defined below.



County changed. Two dummy variables were created using the 1993 and 2003

Beale Codes to control for if the city or town is located within a county that changed

from metropolitan to nonmetropolitan or vice versa between 1993 and 2003. The Beale

Codes are provided in Appendix B. It is important to include this measure as a control

because in 2003 the OMB released the Census 2000 version of the rural urban continuum

codes otherwise known as the Beale Codes. This new version is not fully compatible

with the 1993 codes due to OMB and Census Bureau rule changes in the way urban and

rural are measured. Nonmetropolitan counties are now divided into micropolitan and

noncore. Outlying counties are now considered metropolitan based solely on whether

25% of the workforce commutes into the metropolitan area whereas before outlying

counties were considered metropolitan based on if 15% of the workforce commuted and

based on population density, urbanization, and population growth. Due to the OMB rule

changes approximately 45 metropolitan counties became nonmetropolitan. In addition,









298 nonmetropolitan counties became metropolitan partially from suburbanization but

also partially from the rule changes (Economic Resource Service, 2003). The control is

unnecessary in the current study however. None of the urban cities or rural towns are

located within counties that changed.



Percent Hispanic. Controlling for the Hispanic population is consistent with the

work of Steffensmeier and Haynie (2000) and with the work of Kposowa and colleagues

(1995). It is important to control for the Hispanic population because Hispanics account

for 25% of the nonmetropolitan population growth between 1990 and 2000. In addition,

around 90% of all nonmetropolitan counties experienced Hispanic population growth in

the 1990's, and the Hispanic population is growing faster than all other ethnic and racial

groups in rural America thus helping to offset rural population loss (Kandel & Newman,

2004). This measure is computed as the 2000 Hispanic and Latino population divided by

the 2000 total population. This is then multiplied by 100 to obtain the percent and then

undergoes natural logarithm transformation. See Appendix A for the census definition of

Hispanic and Latino.

LN [(2000 Hispanic and Latino Population) x100
(2000 population) I





South. A dummy variable for the South was also created to control for regional

variation because the study employs a national sample (see Appendix C for list of

southern states). This is consistent with previous studies and is especially important for









this study because the rural south has the highest and most persistent poverty rates (Rural

poverty, 2004).



West. A dummy variable for the West was also created (see Appendix D for list

of western states). This is consistent with Steffensmeier & Haynie (2000) and is

important for this study because of the fast growing nonmetropolitan recreational

counties in the West and South (Rural America, 2004).



Officer rate. Controlling for officer rate is consistent with the works of

Steffensmeier and Haynie (2000) and DeWees and Parker (2003). Officer data was

obtained from the 2000 LEOKA data (ICPSR #3445). The officer rate is calculated by

dividing the total number of officers by the 2000 total population. This result is then

multiplied by 1,000 to obtain a rate per 1,000.



(Total number of police officers)
.-- x 1000
(2000 population)



Methodology

The means and standard deviations for the variables are presented in Table 4-1.

When examining the indicators of private patriarchy, the percent of families where the

husband works and the wife does not has a mean of 24.263 (SD = 3.781) in the urban

sample and a mean of 21.744 (SD = 6.480) in the rural sample. The percent of families

married with children has a mean of 26.884 (SD = 8.350) in the urban sample and a mean

of 27.357 (SD = 7.121) in the rural sample. The urban sample has a mean of .0623 (SD =









.067) in regards to the percent of females working full-time without income, whereas the

rural sample mean for this indicator is .113 (SD = .612). The percent of employed

females who are unpaid family workers has a mean of .304 (SD = .145) in the urban

sample and a mean of .4156 (SD = .699) in the rural sample.

In regards to the measures of public patriarchy, the mean for the ratio of male to

female median income is 1.582 (SD= .187) in the urban sample and is 1.856 (SD= .367)

in the rural sample. The ratio of males to females with a bachelor's degree or higher has a

mean of 1.153 (SD = .095) in the urban sample and a mean of 1.148 (SD = .472) in the

rural sample. The mean for the ratio of males to females in management and professional

occupations is .895 (SD = .137) in the urban sample and .694 (SD = .281) in the rural

sample. For the percent of females working part-time, the urban sample has a mean of

27.823 (SD = 4.960) and the rural sample has a mean of 28.881 (SD = 8.174). The

percent of females living below poverty has a mean of 8.443 (SD = 3.617) in the urban

sample and a mean of 10.758 (SD = 5.145) in the rural sample.

For the indicators of social disorganization, the urban sample has a mean

residential mobility of 52.342 (SD = 6.423) whereas the rural sample has a mean of

42.477 (SD = 8.773). Population change has a mean of 18.643 (SD = 32.714) in the urban

sample and a mean of 9.510 (SD = 45.013) in the rural sample. The mean for the percent

of the population divorced is 10.596 (SD = 2.093) in the urban sample and 11.159 (SD =

3.249) in the rural sample. The natural log of the population size has a mean of 12.251

(SD = .746) in the urban sample and a mean of 7.168 (SD = .529) in the rural sample.

Also reported in Table 4-1 are the t-test values to show whether the variable

means are significantly different between the urban and rural samples. All of the









indicators have significantly different means between the two samples except for the

percent of married families with children and the ratio of males to females with a

bachelor's degree or higher. Several of the variable means are significantly higher in the

urban sample but several are also significantly higher in the rural sample. For instance,

when looking at the measures of private patriarchy, the percent of families where the

husband works and the wife does not is significantly higher in the urban cities (t = 7.129,

p < .01), while the percent of females working full-time without income (t = -2.265, p <

.05) and the percent of employed females who are unpaid family workers (t = -4.149, p <

.01) are significantly higher in the rural sample. The same occurs when looking at the

measures of public patriarchy. The ratio of males to females in professional and

management occupation is significantly higher in urban cities (t = 14.390, p < .01), while

the ratio of male to female income (t = -14.706, p < .01), the percent of females working

part time (t = -2.320, p < .05), and the percent of females living below poverty (t = -

7.356, p < .01) are all significantly higher in the rural sample. When looking at the

indicators of social disorganization, residential mobility (t = 17.889, p < .01), population

change (t = 2.688, p < .01), and the natural log of the population size (t = 90.705, p < .01)

are significantly higher in the urban sample, while percent divorced is significantly

higher in rural towns (t = -2.994, p <.01). As for the control variables, percent Hispanic (t

= 18.547, p < .01) and the West (t = 7.052, p < .01) are significantly higher in the urban

sample whereas South (t = -8.749, p < .01) and officer rate (t = -10.306, p < .01) are

significantly higher in the rural sample. These findings indicate that disadvantage is not

unique to urban areas and further justify the need to explore female offending within the

context of place.









The bivariate correlation matrices are presented in Appendix E. Multicollinearity

is a problem if two or more determinants are highly correlated. Generally, if the

correlation is .5 or greater multicollinearity could be a problem but there is no definitive

rule for this. When examining the correlation matrix for the urban sample, there are

noticeably high correlations between some of the indicators of patriarchy. For instance,

the bivariate correlation between the percent of married families with children and the

percent of families where the husband works and the wife does not is .623. Furthermore,

the percent of married families with children is only correlated with female violent crime

at -.188 and female property crime at -.189, and the percent of families where the

husband works and the wife does not is only correlated with female violent crime at .017

and female property crime at .019. In other words, the correlation is much stronger

between the two independent variables than it is between these two independent variables

and the dependent variables. The same applies for the correlations between the following

variables in the urban sample: the percent of females working full-time with no income

and the percent of employed females who are unpaid family workers (.536); the ratio of

male to female median income and the ratio of males to females with a bachelor's degree

or higher (.523); and the ratio of males to females with a bachelor's degree or higher and

the ratio of males to females in management and professional occupations (.523). One

way to deal with this problem is to employ principal component analysis.



Principal Component Analysis

Principal component analysis was employed on the measures of patriarchy due to

the problem of multicollinearity between those measures. The results from the principal









component analysis are presented in Table 4-2. The analyses after varimax rotation

yielded four components in each of the models.



Private patriarchy

Traditional family index. The ratio of male to female median income, the

percent of families where the husband works and the wife does not, and the percent of

families married with children are loading together in the models. While the ratio of male

to female median income was originally conceptualized as a measure of public

patriarchy, it is logical that it is loading with these two private patriarchal measures

instead. As the percentage of traditional families where the wife is not working increases,

the discrepancy between male and female income should also increase, as well as the

percent of families married with children.



Family unpaid work index: The second dimensional component index includes

two private patriarchy measures the percent of females working full-time with no

income and the percent of employed females who are unpaid family workers. As the

percent of females working full-time without income increases, the percent of employed

females who are unpaid family workers also increases.



Public patriarchy (economic marginalization)

Gender inequality index. The third component includes the ratio of males to

females ages 25 and above with a bachelor's degree or higher and the ratio of males to

females ages 16 and above in management and professional occupations. The ratio of









males to females with a bachelor's degree or higher increases as the ratio of males to

females in management and professional occupations increases.



Percent of females working part-time. While the percent of females working

part-time is loading in the gender inequality index in the total sample, it is only at .502. It

was decided therefore to let the variable stand alone, especially since it loads alone in the

urban sample and loads in the traditional index at only .5 in the rural sample.



Percent of females below poverty. The final component is the percent of females

who are living below poverty. This measure is loading alone in the total sample.



The simplified measurement model is presented in Figure 4-1. The correlation

matrices for the simplified models are presented in Appendix F. The matrices illustrate

that the high correlations are reduced between the individual measures of patriarchy. This

is especially the case for the urban sample; however, the family unpaid work index and

the traditional family index still have a correlation of .515, and the gender inequality

index and the percent of females living below poverty are correlated at -.506.

The means and standard deviations for the simplified models, as well as the t-test

results for comparing the urban and rural means are presented in Table 4-3. I will only be

discussing the results for the three indexes of patriarchy here, as the means and t-test

values for the other indicators are the same as those originally presented in Table 4-1.

The urban sample has a mean of 33.103 (SD = 7.056) for the traditional family index

whereas the rural sample has a mean of 32.075 (SD = 6.610). For the family unpaid work









index, the urban sample has a mean of .298 (SD = .154) and the rural sample has a mean

of .430 (SD= .873). The gender inequality index has a mean of 1.465 (SD = .147) in the

urban sample and a mean of 1.312 (SD = .436) in the rural sample. Again there are

significant differences between the index means in the urban and rural samples. While the

traditional family index is higher in the urban sample (t = 1.863, p < .10), the family

unpaid work index is higher in the rural sample (t = -3.999, p < .01). In addition, the

gender inequality index is higher in the urban sample (t = 7.985,p < .01).



Variance Inflation

Since the social disorganization measures and the control variables were excluded

from the principal component analysis, it is necessary to determine if multicollinearity is

going to be a problem in the regression models. In order to address this issue, the

variance inflation factors were calculated for the variables included in the models.

Variance inflation factors indicate how much the variance of a coefficient is increased

due to collinearity (Ott & Longnecker, 2001, p. 652). The variance inflation factor is

designated by

VIF = 1 / (1 R2)

where R2 refers to how much of the variation in one independent variable is explained by

the others (Ott & Longnecker, 2001, p. 709). It is generally accepted that a variance

inflation factor greater than four indicates multicollinearity (Fisher & Mason, 1981, p.

109); however, some suggest that a value of 10 or greater indicates multicollinearity

(Pindyck & Rubinfeld, 1998; Ott & Longnecker, 2001, p. 652). Appendix G displays the

variance inflation factors for the independent variables included in the models. One of the

variables does have a variance inflation factor greater than four. The traditional family









index has a variance inflation factor of 4.327 in the urban sample. It appears that the

traditional family index may be multicollinear with officer rate (VIF = 3.722). Therefore,

multiple urban models will be run with and without these two measures to determine if

multicollinearity between these two variables is problematic in the regression models.



Analytical Plan

Poisson based regression will be used in the following chapter to examine the

structural correlates of female offending in urban and rural areas. Low arrest counts are

very common in the data since the study is looking at female counts. Poisson based

techniques account for the problem of the relatively low arrest counts (Osgood, 2000).

Negative binomial regression models will be used because they allow for overdispersion

(Gardner et al, 1995; Osgood, 2000). Count data often have overdispersion, with the

variance exceeding the mean (Agresti, 2002). Poisson forces the variance to equal the

mean whereas the negative binomial distribution has

E(Y)= u Var(Y)= / + (,2/k)

where 1/ k is called the dispersion parameter. Comparing the sample mean and variance

of the dependent count variable provides a simple indication of overdispersion (Cameron

& Trivedi, 1998). The means and standard deviations for the dependent variables are

provided in Table 4-1 and 4-3. The dependent variables in this study do have

overdispersion. For the violent crime count, the urban sample has a mean of 211 and a

variance of 256,682, and the rural sample has a mean of .5319 and a variance of 1.337.

The same is true for property crime. The urban sample has a mean of 794.11 and a

variance of 1,463,247, and the rural sample has a mean of 3.04 and a variance of 35.858.









The negative binomial model, however, usually under predicts the amount of zero

counts in the dependent variable. In the current study, the rural sample has 548 (- 71%)

zero counts for female violent crime arrests and 304 (z 39%) zero counts for female

property crime arrests. Due to the large number of zero counts, zero-inflated negative

binomial regression models will also be employed. Zero-inflated models account for

overdispersion due to excess zero counts (Cameron & Trivedi, 1998; Min & Agresti,

2002). Since the zero-inflated Poisson model assumes that the mean equals the variance,

zero-inflated negative binomial models are more likely to be appropriate (Min & Agresti,

2002). While uncommon in criminology (for an example see Robinson, 2003), zero-

inflated negative binomial regression is common in public health research (for an

example see Chin & Quddus, 2003).






52




Traditional Family
__ Index

Family Unpaid
SWork Index

Gender Inequality
Index

% Females Working
.o Part-Time

S% Females Living
Below Poverty


Residential Mobility Female Violent and
Property Crime

Population Change

% Divorced

C Population Size (log)


% Hispanic (log)

South

SWest

Officer Rate



Figure 4-1. Simplified Variable Relationship Model. All proposed relationships are
positive, except that between the indicators of private patriarchy and female
violent crime which is expected to not be significant.











Table 4-1. Means and Standard Deviations of Variables before Principal Component
Analysis


Urban


Rural


Mean SD Mean SD t


Female Violent Crime
Female Property Crime

Private Patriarchy
% Families Husband Works,
Wife Doesn't

% Married Families W/
Children

% Females Working FT, No
Income

% Employed Females Unpaid
Family Workers

Public Patriarchy
Ratio M/F Median Income

Ratio M/F Bachelor's Degree
or Higher

Ratio M/F Professional
Occupations

% Females Working PT

% Females Below Poverty

Social Disorganization
Residential Mobility

Population Change

% Population Divorced

Population Size (log)

Controls
% Hispanic (log)

South

West

Officer Rate

Offset
Female Ponulation (log)


211.085
794.110


506.638
1209.647


24.263 3.781


26.884 8.350


.0623


.304



1.582

1.153


.895


27.823

8.443


52.342

18.643

10.596

12.251


2.521

.290

.450

2.002


.067


.145



.187

.095


.137


4.960

3.617


6.423

32.714

2.093

.746


1.136

.455

.499

.959


.5319
3.037


1.157
5.988


21.744 6.480 7.129**


27.357 7.121 -.735


.113


.4156



1.856

1.148


.694


28.881

10.758


42.477

9.510

11.159

7.168


.773

.6099

.183

3.513


11.579 .747 6.525


.612 -2.265*


.699



.367

.472


.281


8.174

5.145


8.773

45.013

3.249

.529


1.381

.488

.387

3.631


-4.149**



-14.706**

.249


14.390**


-2.320*

-7.356**


17.889**

2.688**

-2.994**

90.705**


18.547**

-8.749**

7.052**

-10.306**


.532 90.085**


*p<.05 **p< .01


A p<.10










Table 4-2. Principal Component Factor Matrices after Rotation Total Sample

1 2 3 4

Ratio M/F Median Income 0.761

% Families Husband Works and Wife Doesn't 0.610

% Married Families W/ Children 0.636

Ratio M/F Bachelor's Degree or Higher 0.708

Ratio M/F Professional Occupations 0.725

% Females Working PT 0.502

% Females Working FT with No Income 0.807

% Employed Females Unpaid Family Workers 0.815

% Females Below Poverty 0.886


Note: Only factor loadings greater than 0.500 are reported.











Table 4-3. Means and Standard Deviations of Variables after Principal Component
Analysis


Urban


Rural


Mean SD Mean SD t


Female Violent Crime

Female Property Crime

Private Patriarchy

Traditional Families Index

Family Unpaid Work

Public Patriarchy

Gender Inequality Index

% Females Working PT

% Females Below Poverty

Social Disorganization

Residential Mobility

Population Change

% Population Divorced

Population Size (log)

Controls

% Hispanic (log)

South

West

Officer Rate

Offset
Female Population (log)

p< .10 *p< .05


211.085

794.110



33.103

.298



1.465

27.823

8.443



52.342

18.643

10.596

12.251


2.521

.290

.450

2.002


506.638

1209.647



7.056

.154



.147

4.960

3.617



6.423

32.714

2.093

.746


1.136

.455

.499

.959


11.579 .747


.5319

3.037



32.075

.430



1.312

28.881

10.758



42.477

9.510

11.159

7.168


.773

.6099

.183

3.513


1.157

5.988


6.525 .532


** p .01


6.610

.873



.436

8.174

5.145



8.773

45.013

3.249

.529


1.381

.488

.387

3.631


1.863^

-3.999**



7.985**

-2.320*

-7.356**



17.889**

2.688**

-2.994**

90.705**


18.547**

-8.749**

7.052**

-10.306**


90.085**














CHAPTER 5
RESULTS

This chapter presents the results from the regression models. The findings from

both the negative binomial regression models and the zero-inflated negative binomial

regression models are discussed below. In all the models, I offset for the female

population thereby essentially creating a rate for the dependent variables.



Violent Crime

Table 5-1 presents the negative binomial and zero-inflated negative binomial

regression coefficients, Z scores, and standard errors for the urban and rural violent crime

models. Negative binomial regression provides the better fit for the urban sample because

the urban sample has no zero counts. For the urban sample, the percent of females

working part-time has a significant and negative effect on violent crime (Z = -2.69, p <

.01). The percent of females living below poverty (Z = 5.18, p < .01) and West (Z = 2.21,

p < .05) have a significant and positive effect on violent crime in the urban model. The

urban models were also run with and without the traditional family index and the officer

rate due to the problematic variance inflation factor. The results are presented in

Appendix H and are similar to those discussed here.

Zero-inflated negative binomial regression is a better fit for the rural violent crime

model (Vuong = 28.10, p < .01). Female poverty (Z= 2.21, p < .05) and officer rate (Z=

2.55, p < .05) are significant and positively related to female violent crime arrests in rural

towns. In addition, population size is significant and has a negative effect on female









violent crime in the rural sample (Z= -3.81, p < .01). A more detailed discussion of these

findings follows below.



Private Patriarchy

Neither of the private patriarchy indexes is significant. This is consistent with the

first hypothesis that private patriarchy would not have a significant effect on violent

crime. Women confined and disadvantaged in the home are more likely to engage in self-

destructive behavior (such as alcohol and drug abuse or suicide) and hurt themselves

when they are unable to cope instead of hurting others (Messerschmidt, 1986).



Public Patriarchy

The findings indicate mixed support for the second hypothesis that the indicators

of public patriarchy (economic marginalization) have a positive effect on female violent

arrests. The gender inequality index has no significant effect on female violent crime in

the urban or rural samples. This is consistent with Whaley and Messner (2002) who also

found gender inequality to have no effect on females killing males or females. Contrary

to expectations, the percent of females working part-time is negatively related to female

violent crime in urban cities. This is, however, consistent with DeWees and Parker (2003)

who also found that part-time employment reduced female offending in their total and

intimate homicide models.

These findings do indicate some support for the economic marginalization

hypothesis. The percent of females living below poverty is significant and has a positive

effect on female violent crime arrests in both the urban and rural areas. This is consistent

with previous studies. Whaley and Messner (2002) found economic disadvantage to be









positively related to females killing males and females. Steffensmeier and Haynie (2000)

found female disadvantage to be positively related to homicide, robbery and aggravated

assault, and DeWees and Parker (2003) also found female poverty to be positively related

to homicide.



Social Disorganization

Contrary to the third hypothesis, the findings indicate no support for the social

disorganization indicators in predicting female violent crime in either urban or rural

areas. Neither residential mobility nor population change is significant in any of the

violent crime models. Kposowa and colleagues (1995) also found population change to

not be a significant predictor of urban violent crime; however, they did find population

change to be a significant predictor of homicide in the largest counties. My findings are

contrary to several other studies on rural violent crime. Petee and Kowalski (1993) found

residential instability to have the greatest impact on rural violent crime. In addition,

Kposowa and colleagues (1995) and Wells and Weisheit (2004) found population change

to be significant and positively related to rural violent crime, and Osgood and Chambers

(2000) found residential instability to be related to rural violent crime.

The percent of the population divorced also has no significant effect on either

urban or rural violent crime. This is inconsistent with previous research as well. While

Kposowa and colleagues (1995) also found divorce to not be a significant indicator of

urban violent crime, Wells and Weisheit (2004) did find family instability to be positively

related to urban violent crime. In addition, both Kposowa and colleagues (1995) and

Wells and Weisheit (2004) found divorce to be positively related to rural violent crime.









Also contrary to social disorganization theory, population size is not significantly

related to female arrests for violent crime in the urban sample. This is also inconsistent

with the findings of Sampson (1987) in which population size was related to juvenile

robbery and black juvenile homicide, and contrary to Land and colleagues (1990) who

also found population structure to be related to homicide. In addition, population size is

significant but negatively related to rural violent crime which is opposite of what social

disorganization theory would suggest; however, Jobes (1999) examined crime in rural

Montana towns and also found that smaller towns had proportionally more crime.



Controls

As for the control measures, percent Hispanic is not a significant predictor of

violent crime in either sample. Kposowa and colleagues (1995) also found percent

Hispanic to not be a significant predictor of rural violent crime, but they did find it to be a

significant predictor of urban violent crime. Wells and Weisheit (2004), however, found

their cultural capital index, which included a measure of the Hispanic population, to have

a significant and positive effect on violent crime in nonmetropolitan counties.

The South is not a significant predictor of female violent crime in urban or rural

areas. Kposowa and colleagues (1195) also found South to not be a significant predictor

of urban violent crime, but the South was a significant predictor of homicide in small

counties. I also find the West to be a significant predictor of urban violent crime, but not

rural violent crime. This is consistent with Steffensmeier and Haynie (2000) who also

found the West to be a significant and positive predictor of female robbery and burglary

in large cities.









Officer rate is not a significant predictor of urban violent crime. This is contrary

to the work of Steffensmeier and Haynie (2000). They found officer rate to be significant

and positively related to female homicide and robbery arrests. I do, however, find officer

rate to be significant and have a positive effect on female violent crime arrests in rural

towns.

Property Crime

Table 5-2 presents the negative binomial and zero-inflated negative binomial

regression coefficients, Z scores, and standard errors for the urban and rural property

crime models. For the urban sample, the percent of females working part-time (Z = 2.80,

p < .01), percent divorced (Z = 3.63, p < .01), South (Z = 2.93, p < .01), and officer rate

(Z= 2.25,p < .05) are all significant and positively related to property crime whereas the

percent of females living below poverty is significant but negatively related to property

crime (Z = -1.65, p < .10). The urban models were also run with and without the

traditional family index and the officer rate due to the problematic variance inflation

factor. The results are presented in Appendix I and are similar to those discussed here.

Again, zero-inflated negative binomial regression models were also employed. Zero-

inflated negative binomial regression is not a better fit for the urban property crime model

since there are no zero counts in the urban sample. Zero-inflated negative binomial

regression is, however, a better fit for the rural property crime model (Vuong = 29.28. p <

.01). The traditional family index (Z = -2.21, p < .05) and the gender inequality index (Z

= -2.32, p < .05) are significant but negatively related to rural property crime. The South

(Z = 2.48, p < .05) and officer rate (Z = 3.42, p < .01) are significant and positively

related to female arrests for property crime in rural towns. A more detailed discussion of

these findings is presented below.









Private Patriarchy

Contrary to the first hypothesis, private patriarchy does not have a significant and

positive effect on female property crime. The family unpaid work index is not significant

in any of the models. Furthermore, the traditional family index is significant but

negatively related to rural property crime. Recall that this index includes the percent of

families married with children and the percent of families where the husband works and

the wife does not. This finding may be indicating that there is a lack of opportunity for

females in rural areas with a higher proportion of traditional families to engage in

property crime such as forgery and fraud.



Public Patriarchy

There are mixed results for the second hypothesis that public patriarchy

(economic marginalization) increases female property crime. Gender inequality does not

have a significant effect on urban property crime. It does have a significant effect on rural

property crime but in the negative direction. Again, while this is contrary to my

expectations it may be in part due to the lack of opportunity to engage in such activity.

The percent of females working part-time is significant and positively related to property

crime in urban cities but is not significant in rural towns. Female poverty is not a

significant predictor of rural property crime. It is, however, a significant predictor of

urban property crime but contrary to expectations, it too is in the negative direction. This

is also contrary to the work of Steffensmeier and Haynie (2000) which found female

poverty to be related to female burglary and larceny rates.









Social Disorganization

There is also very little support for the third hypothesis that social disorganization

variables will have a significant and positive effect on female property crime arrests.

Residential mobility, population change, and population size are not significant in any of

the property crime models. This is contrary to many other studies. Kposowa and

colleagues (1995) found population change to be a significant predictor of property crime

in both metropolitan and nonmetropolitan counties. Barnett and Mencken (2002) also

found population change to have a significant and positive effect on property crime in

nonmetropolitan counties. Jobes and colleagues (2004) found residential instability to be

a significant predictor of "break and enter" crimes and "malicious damage" offenses in

rural Australia. Furthermore, Wells and Weisheit (2004) also found their population

change index to be significant and positively related to property crime in both

metropolitan and rural counties.

Consistent with social disorganization theory, the percent of the population

divorced is significant and positively related to urban property crime. This is consistent

with the findings of Kposowa and colleagues (1995) and Wells and Weisheit (2004). The

percent divorced, however, does not have a significant effect on rural property crime.

This is also consistent with the findings of Kposowa and colleagues (1995). It is contrary

though to Wells and Weisheit's (2004) findings on family instability being a significant

predictor of property crime in rural counties, and is also contrary to the work by Jobes

and colleagues (2004) which found family instability to be a significant predictor of

"malicious damage" offenses.









Controls

As for the control variables, percent Hispanic is not a significant predictor of

property crime in the urban or rural sample. This is consistent with Steffensmeier and

Haynie (2000) but is contrary to Kposowa and colleagues (1995) who found percent

Hispanic to be a significant predictor of property crime in both metropolitan and

nonmetropolitan counties. The South is a significant and positive predictor of property

crime in both urban cities and rural towns. This also is contrary to the findings of

Kposowa et al (1995). My finding that the South is a predictor of property crime may be

a result of strain. The South has the highest and most persistent rates of poverty (Rural

poverty, 2004). The West is not a significant predictor of urban or rural property crime.

This is contrary to Steffensmeier and Haynie's (2000) finding that the West was a

significant predictor of female burglary in large cities. Officer rate is a significant

predictor of female property crime in both the urban and rural samples. This too is

contrary to the findings of Steffensmeier and Haynie (2000). They found officer rate to

not be a significant predictor of female property crime in large cities.



Comparing Regression Coefficients

Due to a lack of research on rural crime, it is unknown whether the traditional

structural theories of crime that were developed around urban areas can also adequately

explain and predict rural crime. In order to test the fourth hypothesis that the impact of

the structural indicators will not be the same in urban and rural areas, the regression

coefficients must be compared. The formula for comparing regression coefficients is

(bi b2)
Z= V(SEb2 +SEb22











where bi b2 is the difference between the sample coefficients, and SEbi2 and SEb22 are

the coefficient variances associated with the first and second groups (Paternoster, Brame,

Mazerolle, & Piquero, 1998). The Z scores for comparing the urban and rural regression

coefficients are presented in Tables 5-1 and 5-2. Comparing the regression coefficients of

the best fitting models is essentially the same as comparing the zero-inflated regression

coefficients for both samples because the urban zero-inflated negative binomial

regression models provide the same results as the urban negative binomial regression

models. The results are discussed below.



Violent Crime

The Z scores for comparing the urban and rural regression coefficients for violent

crime are presented in Table 5-1. When comparing the best fitting regression model

coefficients, the coefficients for the percent of females living below poverty (Z = 3.325, p

< .01) and for population size (Z = 3.337, p < .01) are significantly different. In other

words, the percent of females living below poverty has a stronger, positive effect on

violent crime in urban cities than in rural towns whereas population size has a

significantly stronger negative effect on violent crime in rural towns than in urban cities.

Wells and Weisheit (2004) are the only researchers who have previously compared the

regression coefficients of the structural correlates of crime in urban versus rural areas.

While they found their population change index to have a greater effect on rural violent

crime, neither of the indicators of population change in the current study the percent of

the population that changed from 1990 to 2000, or residential mobility are significantly

different between the urban and rural samples. In addition, they also found their cultural









capital index, which included a measure of the Hispanic population, had a greater effect

on violent crime in rural counties than in large city counties. I, however, do not find

percent Hispanic to have a significantly different effect between the two samples.



Property Crime

The Z scores for comparing the regression coefficients for the urban and rural

property crime models are presented in Table 5-2. When comparing the regression

coefficients for the best fitting urban and rural property crime models, the percent of

females working part-time (Z = 2.231, p < .05) and the percent of the population divorced

(Z = 2.652, p < .05) are significantly different. Again, contrary to Wells and Weisheit

(2004), neither of the measures of population change in the current study has a

significantly different effect on predicting urban versus rural property crime. In addition,

while I find percent divorced to have a greater effect on property crime in urban cities

than in rural towns, Wells and Weisheit did not find family instability to have a

significantly different effect on property crime in urban and rural counties.



Summary

Overall, I find mixed support for the hypotheses presented in Chapter 3.

Consistent with the first hypothesis, private patriarchy is not related to female violent

crime. Contrary to the first hypothesis, though, private patriarchy is not significant and

positively related to female property crime. Only the traditional family index is

significant; however, it is negatively related to female property crime in the rural sample.

In regards to the second hypothesis, I find mixed support for public patriarchy being

positively related to either female violent or property crime. Some of the indicators are









positively significant, however, some indicators of public patriarchy are not significant

and some are negatively significant. In terms of the third hypothesis, I find no support for

the indicators of social disorganization in predicting female violent crime in either the

urban or rural samples. The only significant indicator is population size in the rural

sample; however, it is negatively related to female violent crime. I also find no support

for the social disorganization indicators in explaining female property crime in the rural

sample, and the only significant indicator in the urban model is the percent of the

population that is divorced. For the fourth hypothesis, I do find that some of the

indicators are in fact significantly different between the urban and rural samples. Female

poverty has a stronger, positive effect on female violent crime in the urban sample

whereas population size has a stronger, negative effect on female violent crime in the

rural sample. Likewise, the percent of females working part-time and the percent of the

population divorced have a stronger, positive effect on female property crime in the urban

sample.







67


Table 5-1. Negative Binomial and ZINB Regression Coefficients, (Z Scores), and
Standardized Errors for Female Violent Crime Arrest Models, 2001.


Private Patriarchy
Traditional Families Index
P
Z
SE

Family Unpaid Work
P
Z
SE

Public Patriarchy
Gender Inequality Index
f3
Z
SE

% Females Working PT
f3
Z
SE

% Females Below Poverty
P
Z
SE

Social Disorganization
Residential Mobility
P
Z
SE

Population Change
f3
Z
SE

% Population Divorced
f3
Z
SE

Population Size (log)
P
Z
SE


Urban Model


.004
(0.34)
.013


.068
(0.18)
.367



-.519
(-1.36)
.382


-.036**
(-2.69)
.013


.098**
(5.18)
.019


-.0006
(-0.06)
.0096


-.001
(-0.92)
.002


.021
(0.77)
.028


-.017
(-0.25)
.067


Rural Model


-.006
(-0.68)
.0095


.009
(0.12)
.075



-.167
(-1.16)
.144


-.014
(-1.54)
.009


.025*
(2.21)
.011


.006
(0.86)
.007


.0005
(0.30)
.002


-.012
(-0.65)
.018


-.541**
(-3.81)
.142


Z

.619




.158





-.862




-1.391




3.325**


-.556


-.530


3.337**











Table 5-1. Continued.


Urban Model


Controls
% Hispanic (log)
P
Z
SE


South

Z
SE

West
f3
Z
SE

Officer Rate
f3
Z
SE

Constant

Z
SE

Log Likelihood
X2 ratio test, alpha = 0
Pseudo R2
Vuong Test


Ap< .10


*p<.05


-.073
(-1.18)
.062


-.169
(-1.21)
.140


.341*
(2.21)
.154


.084
(0.93)
.090


-6.145**
(-4.61)
1.331

-1090.4095
8903.58**
0.0336


N = 199


Rural Model


.011
(0.25)
.047


-.079
(-0.42)
.190


.076
(0.35)
.219


.075*
(2.55)
.029


-2.022^
(-1.65)
1.222

-351.4


28.10**

N = 778
0 Obs = 548


**p< .01


Note: The urban model reports the negative binomial regression results. The rural model
reports the zero-inflated negative binomial regression results.


Z

-1.080




-.381




.990




.095




-2.280*







69


Table 5-2. Negative Binomial and ZINB Regression Coefficients, (Z Scores), and
Standardized Errors for Female Property Crime Arrest Models, 2001.


Private Patriarchy
Traditional Families Index
f3
Z
SE

Family Unpaid Work
f3
Z
SE

Public Patriarchy
Gender Inequality Index
P
Z
SE

% Females Working PT
P
Z
SE

% Females Below Poverty
P
Z
SE

Social Disorganization
Residential Mobility
f3
Z
SE

Population Change


% Population Divorced
P
Z
SE

Population Size (log)
P
Z
SE


Urban Model


.002
(0.22)
.011


.195
(0.67)
.293


Rural Model


-.018*
(-2.21)
.008


-.065
(-1.05)
.062


-.343
(-1.00)
.345


.033**
(2.80)
.012


-.026^
(-1.65)
.016


-.0001
(-0.02)
.008


.001
(0.82)
.002


.087**
(3.63)
.024


-.093
(-1.54)
.060


-.313*
(-2.32)
.135


.002
(0.28)
.007


-.005
(-0.49)
.011


.004
(0.63)
.006


.0005
(0.64)
.001


.009
(0.59)
.017


-.059
(-0.49)
.121


Z

1.470


-.078


2.231*


-1.082


2.652*


-.252











Table 5-2. Continued


Urban Model


Controls
% Hispanic (log)
P
Z
SE


South

Z
SE

West

Z
SE

Officer Rate

Z
SE

Constant

Z
SE

Log Likelihood
X2 ratio test, alpha = 0
Pseudo R2
Vuong


Ap< .10


*p<.05


.025
(0.45)
.056


.374**
(2.93)
.128


.083
(0.66)
.126


.158*
(2.25)
.070


-5.919**
(-5.37)
1.103

-1391.2703
2.9E+04**
0.0133


N = 199


Rural Model


-.011
(-0.28)
.039


.348*
(2.48)
.140


.266
(1.59)
.167


.069**
(3.42)
.020


-4.464**
(-4.59)
.973

-1239.113


29.28**


N = 778
0 Obs = 304


**p< .01


Note: The urban model reports the negative binomial regression results. The rural model
reports the zero-inflated negative binomial regression results.


-.875


1.223




-.989














CHAPTER 6
DISCUSSION AND CONCLUSION

The purpose of this study was to examine the association between social structural

conditions and female offending in urban cities and rural towns. This final chapter

outlines the findings presented in chapter five in regards to the research questions posed

in chapter one. In addition, this chapter reviews possible limitations of the study as well

as directions for future research.

Discussion

Research question la asked: What is the impact of patriarchy on female arrests for

violent and property crime? I find that the indicators of private patriarchy have no effect

on female violent crime which is consistent with the first hypothesis. Contrary to

expectations, however, I also find that the indicators of private patriarchy have no effect

on female property crime with one exception. The traditional family index is negatively

related to female property crime in the rural sample. While this finding is also contrary to

expectations, it may be better explained by lack of opportunity for females to engage in

property crime in rural areas that have a high proportion of traditional families (i.e.

married couple families with children under 18 and families where the husband works

and the wife does not).

In regards to the indicators of public patriarchy, I find mixed support for the

economic marginalization hypothesis. Female poverty is significant and positively related

to female violent crime in urban and rural areas. Contrary to expectations, the percent of

females working part-time is negatively related to female violent crime arrests in the









urban sample, but consistent with the hypothesis, it is positively related to female

property crime arrests in the urban sample as well. Also contrary to expectations, the

percent of females living below poverty is negatively related to female property crime

arrests in the urban sample and gender inequality is negatively related to property crime

in the rural sample. Again, this negative relationship between gender inequality and

property crime in rural areas may be better explained by lack of opportunity for

disadvantaged females to engage in property crime in rural areas.

Research question lb asked: What is the impact of social disorganization on

female arrests for violent and property crime? Contrary to expectations, I find very little

support for social disorganization theory in predicting female arrests. The only significant

finding in favor of social disorganization theory is that the percent of the population

divorced is positively related to female property crime arrests but only in the urban

sample. The only other significant finding in terms of the indicators of social

disorganization is opposite of what the theory would predict. Population size is

negatively related to female violent crime arrests in the rural sample.

The second research question posed in chapter one was: Does the impact of these

structural conditions differ in urban and rural areas? The findings presented in chapter

five illustrate that some of the indicators do have a significantly different impact on

female arrests in urban and rural areas. Female poverty is a stronger predictor of urban

violent crime than rural violent crime. Population size is a stronger, negative predictor of

rural violent crime. In addition, the percent of females working part-time and the percent

of the population divorced are both stronger predictors of urban property crime than rural

property crime. These findings suggest that there are significant differences in the









indicators of crime in urban and rural areas that need to be further explored. The standard

structural predictors of urban crime may not be the best predictors of rural crime.



Limitations and Future Research

Like all studies, this study is not without limitations. There are problems with

using official arrest data regardless, but the primary concern in this study is the validity of

the rural arrest data. No studies have examined the validity of arrest statistics in rural

jurisdictions (Osgood & Chambers, 2000). This is an issue that future research in rural

crime must address. Furthermore, the rural sample is limited to only those small towns

that have municipal police agencies. Perhaps this study is missing a large percentage of

small towns that do not have their own agencies and have to rely on county and state

agencies.

A second concern is whether the variables employed are the best measures of the

theoretical concepts. For instance, while percent divorced is included here as a standard

measure of family disruption for social disorganization theory, some use percent divorced

as an indicator of economic marginalization due to the feminization of poverty

(Steffensmeier & Streifel, 1992) while others use it as an indicator of female liberation

(Hunnicutt & Broidy, 2004). Future research needs to expand upon the variables

examined, especially those of private patriarchy, since this is the first study to

conceptualize private patriarchy at the city level and the first study to examine its impact

on female offending across urban cities and rural towns. Subsequent research should also

include measures of opportunity when trying to predict female offending especially when

examining female offending within rural areas.









Finally, this study only examines female arrests. In order to be a truly gendered

approach, a study must examine and be able to explain both male and female offending

(Steffensmeier & Allan, 1996). It would be valuable to examine male arrests as well

because previous studies have indicated that male criminality is associated with economic

disadvantage and Messerschmidt (1986) claims patriarchy contributes to males engaging

in violent crime. With the recent interest in disaggregating, future research should also

consider disaggregating by both gender and specific offense at the rural level since this

has yet to be done.

Conclusion

Overall, while indicators of both patriarchy and social disorganization have

significant impacts on female arrests, there are numerous inconsistencies in terms of the

directions of these relationships, and also in terms of which indicators are significant by

type of offense and by place. Furthermore, the impact of some of these indicators is

significantly different between the urban and rural samples. My findings justify the need

to further examine female offending as well as rural crime. Rural crime can no longer be

ignored. Out of the 100 highest ranking counties on homicide, 71 are rural (Kposowa et

al., 1995). The finding that population size is negatively related to female violent crime in

the rural sample alone justifies the need to further explore the correlates of rural crime.

Perhaps Donnermeyer (2004) is correct in suggesting that both social organization and

disorganization may lead to crime in rural areas.














APPENDIX A
ABBREVIATIONS AND DEFINITIONS

Employed Employed includes all civilians 16 years old and over who were either (1)

"at work" -- those who did any work at all during the reference week as paid

employees, worked in their own business or profession, worked on their own

farm, or worked 15 hours or more as unpaid workers on a family farm or in a

family business; or (2) were "with a job but not at work" -- those who did not

work during the reference week but had jobs or businesses from which they were

temporarily absent due to illness, bad weather, industrial dispute, vacation, or

other personal reasons. Excluded from the employed are people whose only

activity consisted of work around the house or unpaid volunteer work for

religious, charitable, and similar organizations; also excluded are people on active

duty in the United States Armed Forces. The reference week is the calendar week

preceding the date on which the respondents completed their questionnaires or

were interviewed. This week may not be the same for all respondents. (Census

Glossary)

Family A group of two or more people who reside together and who are related by

birth, marriage, or adoption. Families may be a "Married Couple Family," "Single

Parent Family," "Stepfamily," or "Subfamily." (Census Glossary)

Full-time, year-round workers Full-time year-round, workers consists of people 16

years old and over who usually worked 35 hours or more per week for 50 to 52

weeks in 1999. The term "worker" in these concepts refers to people classified as









"Worked in 1999" as defined by the Census. The term "worked" in these concepts

means "worked one or more weeks in 1999" as defined by the census under

"Weeks Worked in 1999." (Census Glossary)

Hispanic or Latino origin People who classify themselves in one of the specific

Hispanic or Latino categories in the 2000 Census Mexican, Puerto Rican,

Cuban, or Other Spanish, Hispanic, or Latino. Origin can be viewed as the

heritage, nationality group, lineage, or country of birth or person or ancestors.

People who classify themselves as such may be of any race. (Census Glossary)

Household A household includes all the people who occupy a housing unit as their

usual place of residence. (Census Glossary)

Household type and relationship Households are classified by type according to the

sex of the householder and the presence of relatives. Examples include: married-

couple family; male householder, no wife present; female householder, no

husband present; spouse (husband/wife); child; and other relatives. (Census

Glossary)

Householder The person, or one of the people, in whose name the home is owned,

being bought, or rented. If there is no such person present, any household member

15 years old and over can serve as the householder for the purposes of the census.

Two types of householders are distinguished: a family householder and a

nonfamily householder. A family householder is a householder living with one or

more people related to him or her by birth, marriage, or adoption. The

householder and all people in the household related to him are family members. A









nonfamily householder is a householder living alone or with nonrelatives only.

(Census Glossary)

Income Total income is the sum of the amounts reported separately for wages, salary,

commissions, bonuses, or tips; self-employment income from own farm or

nonfarm business; interest, dividends, net rental income, royalty income, or

income from estates and trusts; Social Security or Railroad Retirement income;

Supplemental Security Income; any public assistance or welfare payment from the

state or local welfare office; retirement, survivor, or disability pensions; and any

other sources of income received regularly such as Veterans' payments,

unemployment compensation, child support, or alimony. (Census Glossary)

Individuals for whom poverty status is determined Poverty status was determined

for all people except institutionalized people, people in military group quarters,

people in college dormitories, and unrelated individuals under 15 years old. These

groups also were excluded from the numerator and denominator when calculating

poverty rates. They are considered neither "poor" nor nonpoorr". (Census

Glossary)

Married-couple family This category includes a family in which the householder and

his or her spouse are enumerated as members of the same household. (Census

Glossary)

Median income The median income for individuals is based on the individuals 15

years old and over with income. Median income is rounded to the nearest whole

dollar. (Census Glossary)









Metropolitan area (MA) A collective term, established by the federal Office of

Management and Budget, to refer to metropolitan statistical areas, consolidated

metropolitan statistical areas, and primary metropolitan statistical areas. The

OMB revised the definitions in 2003. There are 1,089 metropolitan counties in the

U.S. Under the new system metropolitan areas are defined for all urbanized areas

regardless of the population. A metropolitan area is defined as (1) central counties

with one or more urbanized areas, and (2) outlying counties that are economically

tied to the core counties. Outlying counties are included if 25% of the workers

commute to the central counties or the reverse is 25% of the central counties

commute out. (Census Glossary and ERS)

MSA Metropolitan statistical area. A geographic entity defined by the federal Office of

Management and Budget for use by federal statistical agencies, based on the

concept of a core area with a large population nucleus, plus adjacent communities

having a high degree of economic and social integration with that core.

Qualification of an MSA requires the presence of a city with 50,000 or more

inhabitants, or the presence of an Urbanized Area (UA) and a total population of

at least 100,000 (75,000 in New England). The county or counties containing the

largest city and surrounding densely settled territory are central counties of the

MSA. Additional outlying counties qualify to be included in the MSA by meeting

certain other criteria of metropolitan character, such as a specified minimum

population density or percentage of the population that is urban. Referred to as

SMA beginning in 1949, changed to SMSA in 1959, and then changed to MSA in

1983. (Census Glossary)









Nonmetropolitan The area and population not located in any metropolitan area

(MA).Nonmetropolitan counties are outside the boundaries of the metropolitan

area and are divided into two subtypes: micropolitan areas (centered on urban

clusters of 10,000 or more) and all remaining "noncore" counties. (Census

Glossary and ERS)

Not in labor force Not in labor force includes all people 16 years old and over who are

not classified as members of the labor force. This category consists mainly of

students, housewives, retired workers, seasonal workers interviewed in an off

season who were not looking for work, institutionalized people, and people doing

only incidental unpaid family work (less than 15 hours during the reference

week). (Census Glossary)

Occupation Occupation describes the kind of work the person does on the job. For

employed people, the data refer to the person's job during the reference week. For

those who worked at two or more jobs, the data refer to the job at which the

person worked the greatest number of hours. Some examples of occupational

groups include managerial occupations; business and financial specialists;

scientists and technicians; entertainment; healthcare; food service; personal

services; sales; office and administrative support; farming; maintenance and

repair; and production workers. (Census Glossary)

OMB Office of Management and Budget

Poverty The Census Bureau uses a set of money income thresholds that vary by family

size and composition to detect who is poor. If the total income for a family or

unrelated individual falls below the relevant poverty threshold, then the family









(and all members of the family) or unrelated individual is classified as being

"below the poverty level." (Census Glossary)

Rural For this study, towns with populations less than 2,500 not located within a MSA,

and only such towns located within nonmetropolitan counties. According to the

census rural is territory, population and housing units not classified as urban.

Summary File 1 This file presents 100% population and housing figures for the total

population, for 63 race categories, and for many other race and Hispanic or Latino

categories. This includes age, sex, households, household relationship, housing

units, and tenure. Also included are selected characteristics for a limited number

of race and Hispanic or Latino categories. The data are available for the U.S.,

regions, divisions, states, counties, county subdivisions, places, census tracts,

block groups, metropolitan areas, American Indian and Alaska Native areas, tribal

subdivisions, Hawaiian home lands, congressional districts, and zip code

tabulation areas. Data are available down to the block level for many tabulations,

but only to the census-tract level for others. (Census Glossary)

Summary File 3 This file presents data on the population and housing long form

subjects such as income and education. It included population totals for ancestry

groups. It also included selected characteristics for a limited number of race and

Hispanic or Latino categories. The data are available for the U.S., regions,

divisions, states, counties, county subdivisions, places, census tracts, block

groups, metropolitan areas, American Indian and Alaska Native areas, tribal

subdivisions, Hawaiian home lands, congressional districts, and zip code

tabulation areas. (Census Glossary)









UCR Uniform Crime Reports

Unemployed All civilian 16 years old and over are classified as unemployed if they (1)

were neither "at work" nor "with a job but not at work" during the reference

week, and (2) were actively looking for work during the last 4 weeks, and (3)

were available to accept a job. Also included as unemployed are civilians who did

not work at all during the reference week, were waiting to be called back to a job

from which they had been laid off, and were available for work except for

temporary illness. (Census Glossary)

Unpaid family workers Includes people who worked 15 hours or more without pay in

a business or on a farm operated by a relative. (Census Glossary)

Urban For this study, cities with populations larger than 100,000. The census defines

urban as all territory, population and housing units in urbanized areas and in

places of more than 2,500 persons outside of urbanized areas.

Urbanized area (UA) An area consisting of a central places) and adjacent territory

with a general population density of at least 1,000 people per square mile of land

area that together have a minimum residential population of at least 50,000

people. (Census Glossary)

Weeks worked in 1999 The data pertain to the number of weeks during the designated

calendar year in which a person did any work for pay or profit (including paid

vacation, paid sick leave, and military service) or worked without pay on a family

farm or in a family business. (Census Glossary)

Worked in 1999 People 16 years old and over who did any work for pay or profit

(including paid vacation, sick leave, and military service) or worked without pay






82


on a family farm or in a family business at any time during the past 12 months are

classified as "worked in the past 12 months." All other people 16 years old and

over are classified as "Did not work in the past 12 months." (Census Glossary)















APPENDIX B
RURAL-URBAN CONTIUUM CODES (BEALE CODES)

Table B-1. 2003 Beale Codes
Metro Counties
1 Counties in metro areas of 1 million population or more
2 Counties in metro areas of 250,000 to 1 million population
3 Counties in metro areas of fewer than 250,000 population
Nonmetro Counties
4 Urban population of 20,000 or more, adjacent to a metro area
5 Urban population of 20,000 or more, not adjacent to a metro area
6 Urban population of 2,500 to 19,999, adjacent to a metro area
7 Urban population of 2,500 to 19,999, not adjacent to a metro area
8 Completely rural or less than 2,500 urban population, adjacent to a
metro area
9 Completely rural or less than 2,500 urban population, not adjacent
to a metro area


Table B-2. 1993 Beale Codes
Metro Counties
0 Central counties of metro areas of Imillion population or more
1 Fringe counties of metro areas of 1 million population or more
2 Counties in metro areas of 250,000 to 1 million population
3 Counties in metro areas of fewer than 250,000 population
Nonmetro Counties
4 Urban population of 20,000 or more, adjacent to a metro area
5 Urban population of 20,000 or more, not adjacent to a metro area
6 Urban population of 2,500 to 19,999, adjacent to a metro area
7 Urban population of 2,500 to 19,999, not adjacent to a metro area
8 Completely rural or less than 2,500 urban population, adjacent to a
metro area
9 Completely rural or less than 2,500 urban population, not adjacent
to a metro area


(Economic Resource Service, 2004)















APPENDIX C
SOUTHERN STATES

Alabama

Arkansas

Delaware

Florida

Georgia

Kentucky

Louisiana

Maryland

Mississippi

North Carolina

Oklahoma

South Carolina

Tennessee

Texas

Virginia

West Virginia















APPENDIX D
WESTERN STATES

Alaska

Arizona

California

Colorado

Hawaii

Idaho

Montana

Nevada

New Mexico

Oregon

Utah

Washington

Wyoming















APPENDIX E
CORRELATIONS BEFORE PRINCIPAL COMPONENT ANALYSIS



















Table E-1. Correlation Matrix of Variables before Principal Component Analysis Total Sample


Violent Property Husband Married NoInc Unpaid RatInc RatEd RatOcc PTWork


Female Violent Crime 1.000
Female Property Crime .813 1.000
% Families Husband Works, Wife Doesn't .062 .088
% Married Families w/ Children -.099 -.097
% Females Working FT, No income -.017 -.023
% Employed Females Unpaid Family Workers -.029 -.043
Ratio M/F Median Income -.154 -.199
Ratio M/F Bachelor's Degree or Higher -.012 -.008
Ratio M/F Professional Occupations .086 .140
% Females Working PT -.058 -.071
% Females Below Poverty .006 -.056
Residential Mobility .089 .187
Population Change -.004 .031
% Population Divorced -.051 -.030
Population Size (log) .439 .590
% Hispanic (log) .177 .248
South -.115 -.121
West .043 .078
Officer Rate -.012 -.044


1.000
.227 1.000
-.018 .012 1.000
-.032 .053 .347 1.000
.156 .286 .067 .037 1.000
-.073 .001 .117 -.003 .116 1.000
-.075 -.160 -.014 .016 -.023 .271 1.000
-.120 .106 .066 .132 .358 .137 .189 1.000
.170 -.341 -.010 -.023 -.079 -.048 -.190 -.167
.136 -.054 .010 -.029 -.158 .003 .280 .151
.050 .056 -.002 .032 .003 -.012 .107 -.005
.009 -.246 .060 -.006 -.054 .020 -.029 .009
.157 -.021 -.045 -.052 -.313 -.015 .267 -.053
.289 .306 -.02 -.020 -.163 -.012 .007 -.027
.189 -.183 .000 -.013 -.081 -.056 -.172 -.479
.075 .175 .027 .017 .107 .034 .151 .340
-.076 -.251 -.029 -.063 .056 .139 .154 -.082



















Table E-1. Continued


Poverty Mobil PopCha Divorced Pop Size Hisp South West Officer

Female Violent Crime
Female Property Crime
% Families Husband Works, Wife Doesn't
% Married Families w/ Children
% Females Working FT, No income
% Employed Females Unpaid Family Workers
Ratio M/F Median Income
Ratio M/F Bachelor's Degree or Higher
Ratio M/F Professional Occupations oo
% Females Working PT
% Females Below Poverty 1.000
Residential Mobility -.221 1.000
Population Change -.124 .318 1.000
% Population Divorced .067 .196 .019 1.000
Population Size (log) -.176 .431 .074 -.084 1.000
% Hispanic (log) -.114 .348 .143 -.057 .461 1.000
South .421 -.251 -.054 -.039 -.267 -.222 1.000
West -.261 .341 .188 .138 .229 .410 -.610 1.000
Officer Rate -.019 -.053 .003 -.014 -.274 -.145 .172 -.065 1.000




















Table E-2. Correlation Matrix for Variables before Principal Component Analysis Urban Sample


Violent Property Husband Married Nolnc Unpaid RatInc RatEd RatOcc PTWork


Female Violent Crime 1.000
Female Property Crime .787
% Families Husband Works, Wife Doesn't .017
% Married Families w/ Children -.188
% Females Working FT, No income -.077
% Employed Females Unpaid Family Workers -.044
Ratio M/F Median Income -.208
Ratio M/F Bachelor's Degree or Higher -.135
Ratio M/F Professional Occupations -.082
% Females Working PT -.140
% Females Below Poverty .229
Residential Mobility -.209
Population Change -.106
% Population Divorced -.089
Population Size (log) .699
% Hispanic (log) .048
South -.066
West -.092
Officer Rate .426


1.000
.019
-.189
-.081
-.073
-.206
-.107
-.057
-.167
.135
-.112
-.043
.025
.763
.045
.024
-.108
.433


1.000
.623 1.000
.274 .351 1.000
.397 .493 .536 1.000
.429 .468 .279 .285 1.000
.350 .328 .282 .310 .523 1.000
.164 -.026 .135 .094 .219 .523 1.000
-.060 -.091 .002 .103 .388 .095 .225 1.000
-.077 -.393 -.200 -.140 -.267 -.364 -.498 .152
.184 .019 -.020 -.062 -.072 .036 .314 .122
.271 .407 .074 .100 .207 .226 .143 -.139
-.412 -.442 -.149 -.194 -.098 -.074 -.123 -.103
.017 -.255 -.116 -.097 -.336 -.145 -.025 -.276
.459 .542 .202 .318 -.105 .012 -.237 -.230
.153 -.136 .015 -.079 -.042 -.010 -.057 -.295
.321 .462 .211 .411 .121 .101 .206 .126
-.412 -.675 -.252 -.401 -.448 -.328 -.209 -.200