? DECEMBER 2006 Where Do We Go From Here? A REPORT ON SEX OFFENDERS AND SEX OFFENDER HOUSING IN PHOENIX, ARIZONA Where Do We Go From Here? A Report on Sex Offenders and Sex Offender Housing in Phoenix, Arizona Prepared for the Phoenix Police Department and the Phoenix City Council Prepared by Charles M. Katz, Ph.D. Vincent J. Webb, Ph.D. Todd Armstrong, Ph.D. Constance Kostelac, M.S. Center for Violence Prevention and Community Safety College of Human Services Arizona State University December 2006 About the Authors Charles M. Katz is the Interim Director of the Center for Violence Prevention and Community Safety and an associate professor in the School of Criminology and Criminal Justice at Arizona State University. Dr. Katz earned his Ph.D. in criminal justice from the University of Nebraska at Omaha in 1997. His research primarily involves collaborating with agencies to increase their organizational capacity to identify and strategically respond to crime and delinquency affecting local communities. He is the co-author and co-editor of three books: The Police in America: An Introduction (5th edition; McGraw Hill Publishing); Controversies in Criminal Justice (Roxbury Publishing); and On Their Own: Policing Gangs in America (Cambridge University Press). Vincent J. Webb is the Dean and Director of the George J. Beto Criminal Justice Center at Sam Houston State University. Dr. Webb has held a number of top administrative positions in leading criminal justice academic and research programs over his 30-year career. His academic specialties include criminal justice policy analysis and evaluation research, police studies, violence prevention, and criminal justice planning. Dr. Webb’s current research interests involve gangs, human trafficking, and sex-offender residential policy studies. Dr. Webb received his Ph.D. in sociology from Iowa State University. Todd A. Armstrong is an assistant professor in the Administration of Justice program at Southern Illinois University. Professor Armstrong is interested in policy and program evaluation, research on offense type patterns, and criminological theory. His evaluation work has included assessments of programs and policies implemented in schools, communities, and a number of criminal justice system settings. Professor Armstrong’s work on offense patterns has been supported by a grant from the National Institute of Justice; it includes an exploration of the predictors of offense type, a test of the cause of changes in specialization over the criminal career, and an analysis of the impact of age on offense type. His theoretical work includes an assessment of the generality of acts of crime and delinquency and a discussion of the incorporation of learning processes in control theories. Recent research has appeared in Criminal Justice and Behavior, Justice Quarterly, Journal of Criminal Justice, and Advances in Criminological Theory. i Constance Kostelac is the Police Research Supervisor for the Phoenix Police Department’s Crime Analysis and Research Unit (CARU). Ms. Kostelac has been with the department since 1999, serving as a crime analyst. She was promoted to unit supervisor in 2001. During this time, CARU grew to a staff of more than 20 employees. Ms. Kostelac is responsible for overseeing the department’s efforts in administrative, tactical, and strategic crime analysis. She is a former president of the Arizona Association of Crime Analysts (AACA) and helped to develop a certificate program in crime and intelligence analysis. Ms. Kostelac earned her master’s of science degree in criminology and criminal justice from Florida State University in 1997. She is a doctoral candidate in justice and social inquiry at Arizona State University. Her research interests include police organizations with a focus on the process of civilianization, trends in arrests, and the distribution of crime across time and space. Acknowledgements The authors would like to thank members of the Crime Analysis and Research Unit in the Phoenix Police Department for all of their hard work on this project and for providing much of the official data used in the present study. We would also like to thank Jill Levenson at Lynn University and Richard Tewksbury at the University of Louisville for assisting us with the construction of survey instruments. Last, we would like to thank David Choate and Casey Klaus who managed much of the data collection and analysis for the offender portion of this study, and Cher Stuewe-Portnoff who served as the copyeditor for this report. The teal ribbon on the cover of this report is the official awareness ribbon for sexual assault. This project was funded in part by the City of Phoenix and the Center for Violence Prevention and Community Safety at Arizona State University. Opinions contained herein are those of the authors and do not represent the position of the Phoenix City Council nor the Phoenix Police Department. © 2006 by the Arizona Board of Regents for and on behalf of the Arizona State University and its Center for Violence Prevention and Community Safety. ii Table of Contents Page About the Authors i Executive Summary 1 Introduction - Management and Control of Sex Offenders: A Criminal Justice Policy Problem 5 1 - An Exploration of the Offending Behavior of a Random Sample of Sex Offenders from the City of Phoenix 12 2 - An Assessment of the Accuracy of Sex-offender Address Registration in Phoenix, Arizona 39 3 - An Examination of Sex-offender Perceptions, Experiences, and Opinions 55 4 - Citizen Attitudes about Sex Offenders and Sex-offender Housing Policy in Phoenix 78 5 - The Impact of Sex-offender Residential Clustering and Related Policies on Selected Agencies and Organizations: Perceptions and Recommendations 107 6 - A Geographic Analysis of Sex-offender Clustering and Proposed Distancing Requirements 112 iii Executive Summary Charles M. Katz, Ph.D. Vincent J. Webb, Ph.D. Todd Armstrong, Ph.D. Prior to developing and approving new ordinances that would further regulate sexoffender distance restrictions and clustering, the Phoenix City Council approved a study to examine the nature and impact of sex-offender clustering. Their overall goal was to use the study to fill in some important knowledge gaps about the issue of sex-offender residential clustering in order to inform the development of effective policy. To examine this issue on behalf of the city council, we relied on a multi-methodological research design. We wanted to incorporate information from several different points of view, giving policy makers a more comprehensive perspective from which to inform their decision making. For this study, we relied on four resources: official police data, offender interview data, citizen survey data, and key stakeholder interview data. Our major findings and subsequent recommendations are presented below. Major Findings Criminal History Report Section Analysis showed that rates of recidivism were relatively low among registered sex offenders in Phoenix. Fewer than half of all level-2 and level-3 active registered sex offenders were re-arrested for any offense after initial registration. Among those who were re-arrested, charges for sex offenses were rare. About 6.2% of sex offenders were charged with sexual assault and 5.7% were charged with a child-sex crime. Rates of rearrests and charging were based on an average 6-year period following registration. Recidivism rates described above are similar to those reported in several other studies. Address Verification Section Based on information gathered as part of the sex-offender interview process, it is our opinion that between 61.9% and 70.8% of registered sex offenders in Phoenix may not actually reside at their registered addresses. Further, these percentages do not include the roughly 6-7% of sex offenders who failed to register at all and therefore were not included in this analysis. 1 Sex Offender Interview Section As a direct consequence of sex-offender registration and community notification, released ex-offenders faced substantial psychological, social, and economic hardship. Successful re-entry was hampered when residential restrictions severely constrained social ties with family, friends, and others. Under policies in effect when this study was conducted, sex offenders found it difficult to find and hold jobs and to find and maintain affordable housing. Enhanced restrictions, particularly those related to housing selection, could be predicted to escalate those difficulties. The data supported the supposition that public policies to further restrict sex-offender housing options could easily drive more offenders underground, and police would have even less accurate information about offenders’ locations. Citizen Survey Section Regardless of the number of sex offenders living nearby, citizens were two to three times more likely to indicate that non-sex crimes (e.g., robbery, drug sales, assault) were problems in the neighborhood than to indicate that sex crimes (e.g., rape, child molestation) were. Citizens living in neighborhoods with high numbers of sex offenders voiced significantly more concern about sex crimes, however, compared with those living in neighborhoods with few or no sex offenders. Surveys of citizens suggested that strategies such as supervised probation, community notification, sex-offender registration, treatment, distance buffers around places children gather, and limits on sex-offender residential density all were generally viewed as effective. Likewise, a majority indicated support for a number of policies regulating the spatial location and distribution of sex offenders, with a stronger emphasis on regulating distance from key community facilities (e.g., schools, day-care centers, parks) than on the number of offenders living in a particular area. Stakeholder Section Some stakeholders perceived that the general public and politicians overestimated and over-reacted to the threat posed by sex offenders, and that this possibly resulted in overregulation. They believed that over-regulation “pushes” sex offenders into living in certain areas of the community. Opinions varied on the level of risk posed by sexoffender clustering, but some stakeholders cautioned that “cluster-busting” could be a factor in driving sex offenders underground, leading to law enforcement having less accurate information on their locations. Many stakeholders voiced concern about unfunded mandates and mission displacements that could result from new sex offender policies. They mentioned that in the past, costs associated with community notification, reporting, and supervision had not been covered by policy makers when enacting new policies; as a result, affected agencies were required to redirect resources from other priorities. Some stakeholders reported that they had been unable to provide shelter services to their intended clients when they were displaced by sex offenders unable to find other housing under the new policies. 2 Mapping Section If residential distance restrictions pertaining to schools, child-care centers, and parks and restrictions pertaining to other offenders were implemented together, there would be no feasible way to physically disperse the current registered sex-offender population throughout Phoenix. Even if such policies were enacted in part, we believe that this would lead to offenders choosing not to register and to abscond, simply to be able to find housing in the city. Major Recommendations Many of the policy recommendations in this report focus on increasing the accuracy of the information found in the sex-offender registry. Several of the policies and strategies under discussion for responding to the city’s sex-offender problem depend on knowing where registered sex offenders are living, so deriving accurate address information from sex-offender registration records would be critical to their success. This is true whether the policies in question pertained to reducing sex-offender clustering or to enforcing residential distance restrictions around schools, child-care centers, and parks. Below is a summary of our recommendations for increasing the accuracy of sex-offender registration information in the City of Phoenix. 1. An interagency task force should be formed to develop a process map that delineates clear organizational lines of responsibility for sex-offender registration, notification, and address verification. Related, we recommend clarifying policies and assignment of responsibility for sex-offender address verification to prevent unintended gaps in coverage. 2. For the most serious registered sex offenders, require enforcement officials to conduct mandatory, random, in-person interviews with offenders at their registered addresses. The Phoenix City Council should allocate financial resources for the Sex Crimes Unit to hire the additional staff needed to fulfill this recommendation. 3. The city, county, and state should develop and document clear performance measures and the methods for compiling and calculating success criteria to ensure registration compliance. Related, we recommend that Phoenix re-evaluate policies that dictate which offenders are subject to address verification by police officials. 4. Policy makers should consider sex-offender re-integration programs that will encourage sex offenders to register with the police, such as those that would proactively provide them with assistance in finding employment and housing. Above all, we recommend that the city postpone enacting ordinances related to reducing sex-offender clustering or creating residential distance restrictions surrounding schools, child-care centers, and parks until the accuracy of sex offender registration data is improved. If such ordinances are enacted without those improvements, we believe that sex-offender registration data will become 3 even less accurate over time, which in turn could worsen police performance in the investigation of sex crimes. 4 Introduction Management and Control of Sex Offenders: A Criminal Justice Policy Problem Vincent Webb, Ph.D. Charles M. Katz, Ph.D. Management and control of sex offenders has become one of the more perplexing crimerelated policy issues in America. According to the U.S. Department of Justice, nearly 250,000 sexual-offense victimizations occur in the United States each year (Rennison, 2002). The number of persons incarcerated for sexual offending in the United States has been growing steadily: As of the mid-1990s, nearly one in 10 prison inmates had been incarcerated for a sexual offense (Greenfield, 1997:20). The Center for Sex Offender Management reports that about 250,000 sex offenders were under the jurisdiction of correctional agencies in 2002, approximately half of them under some form of community supervision. The concentration of sex offenders is greatest in our largest metropolitan areas, but cities, towns, and villages of all sizes must be prepared with measures for protecting the public from potential offenders and from the fear often associated with the presence of sex offenders. As the Center for Sex Offender Management points out, the irreversible nature of the particular harm caused by sex offending and the public’s fear of such offenses together make this issue of special concern to communities (Center for Sex Offender Management, 2001). The fact is that most persons incarcerated for sexual offenses will eventually re-enter the community. Research shows that about 13% of convicted sex offenders re-offend within 4 to 5 years (Hanson & Bussiere, 1998). Communities are justifiably concerned. In recent years, high profile offenses perpetrated by known sex offenders led to a flurry of new policy proposals, laws, and ordinances, in an attempt to exercise greater control over sex offenders and thereby to increase public safety. The challenge for policymakers, leaders, and criminal justice officials is to thoroughly evaluate alternative strategies for managing sex offenders in the community in order to minimize both the harm caused by re-offending and fear in the community generated by sex offending and the presence of sex offenders. The cornerstone for policy strategies for managing sex offenders is the sexoffender registry, a practice that can be traced back as far as the 1930s. Contemporary sex-offender registration policy, however, is usually associated with the federal Jacob Wetterling Crimes Against Children and Sexually Violent Offender Registration Act passed in 1994 (Center for Sex Offender Management, 5 1999). This act required states to create sex-offender registries, as well as to meet quarterly and annually to address verification requirements. Today, nearly every state has enacted registration laws with three goals in mind: to deter sex offenders from reoffending; to give law enforcement access to relevant information that can be used in sexoffense investigations; and to protect the public from known sex offenders (Center for Sex Offender Management, 1999). Megan’s Law, the most well-known federal amendment to the Wetterling Act, was passed in 1996. It requires states to publicize information about sex offenders, using reasonable discretion. The Pam Lychner Act of 1996 imposed further requirements on states relative to registering and tracking sex offenders. In 1998, still other amendments required states with minimally sufficient registration programs to participate in the National Sex Offender Act. Additional amendments to these acts have been made throughout the past several years. Most recently, the Adam Walsh Child Protection and Safety Act of 2006 added provisions for further increasing the capacity of states to effectively register, track, and manage sex offenders. Registration and community notification usually are considered the principal tools for managing sex offenders in the United States, but recent high-profile recidivistic sex offenses have spawned a new wave of laws and ordinances. These have introduced strategies such as requiring sex offenders to wear electronic tracking devices, imposing civil commitments, and subjecting offenders to periodic polygraph examinations. However, the majority of the newer laws and ordinances focus on enacting spatial policies: They seek to govern the physical spaces that sex offenders may occupy. Often, they will establish buffer zones, prohibiting sex offenders from residing close to certain public places such as schools, day-care centers, parks, and school-bus stops. Typically, offenders are prohibited from residing within 1,000 to 2,000 feet of such places, although in some cases the distance has been as little as 500 feet (Levenson & Cotter, 2005). Sex-offender Registration and Community Notification In Arizona, as in most other states, sex-offender registries and community notification form the foundation of sex-offender management strategies. Typically, these strategies rely on an offender classification system that attributes a certain level of risk to each offender. The Phoenix system, like most others, estimates the risk to the community posed by each offender based on his or her likelihood of re-offending. Most jurisdictions use a system of three risk levels; Phoenix added a fourth (levels 0 through 3), with level 3 being the highest.1 The community-notification process in Arizona is triggered by a sex offender’s release from jail or prison or sentence to probation. When this occurs, the responsible county adult probation agency or the Arizona Department of Corrections is required to enter offender data into a state-wide, accessible database. Offenders at all 1 Level 0 is a unique historical artifact applying to offenders convicted and released prior to June 1, 1996, when Arizona’s sex-offender community notification statute went into effect. In essence, level-0 offenders would have been covered by the current statute, had they offended and been released after June 1, 1996. 6 levels are required by law to register, but only level-2 and level-3 offenders are subject to community-notification laws.2 As of February 3, 2006, the Phoenix Police Department’s sex-offender database listed 939 active registered level-2 and level-3 sex offenders who resided within the city. Of these, 53% (495) were classified as level 2 and 47% (444) as level 3. The database also included 1,008 active level-0 and 516 active level-1 offenders; offenders at these levels Arizona Public Act 13-3821 were required to register, but were not subject to community-notification laws. This act requires sex offenders residing in Finally, the database contained information Arizona to register with the county entries for inactive registered offenders (e.g., sheriff’s office where they are living within those who were deceased or incarcerated, or 10 days of entering the county. All offenders who have been convicted of who had absconded or moved away). Classification of a sex offender begins with an assessment of certain characteristics. The Phoenix Police Department uses a screening tool designed by the Arizona Department of Corrections. This program generates two risk scales: one for sex-offending risk and one for general recidivism. A score for each type of risk is assigned to the offender, based on 19 different factors such as the offense type, prior number of convictions, weapons use, use of force, relationship to victim, alcohol and drug use, employment history, documented behavior indicating sexually deviant interests, and so forth. The assessment produces a total score that places the offender in one of four sex-offending risk categories, ranging from lower risk to very high risk, and in one of five categories of general recidivism risk, ranging from lower risk to ultra-high risk. With the department’s program, for community notification purposes, persons who rank high or very high for sex-offending risk or very high or ultra-high for general recidivism risk are recommended for classification as level-3 offenders. Those ranking intermediate for sex-offending risk 2 violations or attempted violations of sex offenses are subject to this requirement. (The act specifies that sex offenses committed in other jurisdictions and those committed before September 1, 1978, are included when those offenses have the same elements as an offense listed in the act.) When a level-2 or level-3 offender moves into a community, the surrounding neighborhood, area schools, appropriate community groups, and prospective employers are notified. Notification consists of a flyer showing the offender’s photograph and exact address, together with a summary of that person’s status and criminal background. A press release and a level-2 or level-3 flyer is given to the local electronic and print media, as well, so that the information can be published locally. If a level-2 or level-3 offender fails to register or re-register pursuant to section 13-3821 or 13-3822 and a warrant is issued, before issuing the warrant, the law enforcement agency requesting the warrant is to assemble, print, and distribute appropriate flyers regarding the offender. (State of Arizona, Department of Public Safety, Sex Offender Info Center, http://az.gov/webapp/portal/sows.jsp.) Information on sex-offender classification by the Phoenix Police Department was obtained through personal interviews with officials and from documents provided by the department, including the sexoffender risk assessment tool. 7 or intermediate or high for general recidivism risk are recommended for classification as level-2 offenders. Those who rank as a lower sex-offending risk and a lower general recidivism risk are recommended for classification as level-1 offenders. The initial risk scores and recommended offender-classification levels based on this assessment are not binding on law enforcement, which has the discretion to make alternative decisions. Arizona law (ARS 13-3825) states that the law enforcement agency having legal responsibility for an offender has the final authority to determine his or her classification. Note that Arizona has at least 20 statutory crimes identified as sex offenses, some of which are sexually motivated property crimes (e.g., burglary). Note also that the risk assessment tool can take non-sex offenses into consideration when determining an offender’s classification level, so that the final level assigned may or may not be based solely on the sexual offense. Trends in Spatial Regulation of Sex Offenders Phoenix and other cities encounter numerous problems in their attempts to track and manage registered sex offenders. In response, most are considering a new wave of potential laws and ordinances. Residential distributions that could result in high-density clusters of sex offenders in certain neighborhoods are of concern to city officials, community leaders, and criminal justice officials. Phoenix and the State of Arizona, among others, are adopting or considering adopting some of the newer policies and strategies in use elsewhere. For example, the use of Global Positioning Satellite (GPS) technology for tracking sex offenders is under consideration. Also, an Arizona law was recently passed that limits the number of sex offenders in multifamily dwellings to 10% of that dwelling’s total units (Phoenix Insurgent, 2006; Rubiano, 2005).3 Arizona is also increasing its use of civil commitments and treatment resources for incarcerated sex offenders prior to re-entry (Crawford, 2004). In Phoenix, recent proposals for regulating sex offenders’ residences included restricting to 2,000 feet the residential proximity of sex offenders to schools, churches, school bus stops, and other places where children congregate (Sexton, 2005). Nationally, little is known about the effectiveness of such sex-offender management and control regulations, for sex offenders or for their communities. In a recent Iowa case, the courts used the lack of research into the connection between sexoffender proximity and recidivism as the basis for finding distance restrictions (the 2,000foot rule) unconstitutional, although that ruling was later overturned by a higher court 3 The Arizona statute states: “An adult probation officer shall not approve the residence of a probationer, who is required to register, in any multifamily dwelling unless the number of registered probationers who reside is less than ten percent of the number of dwelling units. Not more than one probationer classified as a level three offender…shall reside in a multifamily dwelling. These provisions do not apply to any of the following: a person placed on probation before the effective date of this section until that person changes residence; a person who resides in a residential treatment facility or a person who participates in a supervised program that provides transitional services, including diagnostic evaluation, behavioral, medical, psychiatric, psychiatric, psychological and social service care; a juvenile who resides with a parent or guardian; a multifamily dwelling in an industrial or commercial zone. (Arizona Supreme Court, http://www.supreme.state.az.us/apsd/Sex%20Offenders/Statutory%20Requirements%208-12-05.pdf). 8 (Davey, 2006). Iowa law enforcement officials now see that the unanticipated difficulties posed for offenders by the 2,000-foot rule had likely driven sex offenders underground, making them still more difficult to locate, supervise, and control. A body of research is emerging on the impact of sex-offender restrictions that can, however, inform policy to a certain degree. For example, one study found that the passage of sex-offender registration and notification laws had no systematic impact on the incidence of rape (Walker et. al, 2001). Another study, prepared for the Colorado legislature, concluded that shared living arrangements were successful in containing and treating sex offenders in the community (Colorado Department of Public Safety, 2004). Several studies have examined recidivism rates for sex offenders, for example, comparing child molesters and rapists; these found that over time, child molesters have significantly higher failure rates through re-arrest (Center for Sex Offender Management, 2001). Still, much needs to be learned in order to guide effective policy development. Buffer zones and regulations limiting sex-offender residential densities are increasingly popular approaches, although such policies generally are enacted through laws and ordinances without benefit of research into the nature of the community’s particular sex-offender problems or the potential impacts of regulation. To its credit, in 2005, the Phoenix City Council commissioned a study to examine the nature and impact of sex-offender clustering in Phoenix as a prelude to developing new ordinances that would impose distance restrictions and/or limit residential sex-offender clustering. The goal was to fill some important voids in their knowledge about the problem in Phoenix in order to inform effective policy development. The Present Study Problems associated with the presence of known sex offenders in communities and the potential role of spatial regulation or housing management in policing offenders are multidimensional and complex. It follows that any study attempting to describe and analyze these problems will also be multidimensional. In conducting a comprehensive study of sex offenders and sex-offender clustering, the authors believed that it would be most useful to examine the problem using four distinct information sources: official police data, offender interview data, citizen survey data, and stakeholder interview data. Each of these four data sources could yield critical information and deeper insight into Phoenix’s sex-offender problem. Taken together, these data sources would provide a solid basis upon which to develop effective municipal criminal justice policy. In order to respond successfully to sex crimes in Phoenix, the city’s policymakers must first properly diagnose the problem. Insight into the nature of the problem and an understanding of the community’s capacity to address the problem are necessary precursors to effective policy. In the sections that follow, the authors provide Phoenix policymakers with a comprehensive examination of sex offenders and the problems associated with sex-offender housing in their city. 9 This study consists of six reports. The first examines Phoenix’s sex-offense patterns and trajectories of sex offenders. The second assesses the accuracy of sex offenders’ registered addresses in the city and estimates the number of registered sex offenders who do not reside at their registered addresses. The third report presents our findings on sex-offender perceptions and experiences with sex-offender policy and legislation. The fourth report discusses findings on Phoenix citizens’ perceptions of the sex-offender problem, their fears with respect to such offenses, their perceptions of sex-offender management policies, and their experiences with the sex-offender policy as community residents. The fifth report examines the impact of the sex-offender problem and policies from the perspective of selected organizations in Phoenix and presents their proposals for improving sex-offender clustering policy. The sixth and final report is a geographic analysis of Phoenix that will help determine the feasibility of potential residential restriction policies on sex-offender living arrangements. List of References Arizona Supreme Court, (n.d.) Retrieved July 14, 2006, from http://www.supreme.state.az.us/apsd/Sex%20Offenders/Statutory%20Requiremen ts%208-12-05.pdf. Center for Sex Offender Management (1999). Sex offender registration: Policy overview and comprehensive practices. Washington, DC: U.S. Department of Justice, National Institute of Justice, Center for Sex Offender Management. Center for Sex Offender Management (2001). Recidivism of sex offenders. Washington, DC: U.S. Department of Justice, National Institute of Justice, Center for Sex Offender Management. Colorado Department of Public Safety. (2004). Report on safety issues raised by living arrangements for and location of sex offenders in the community. Denver, CO: Colorado Department of Public Safety, Division of Criminal Justice, Sex Offender Management Board. Crawford, A. (2004, January 20). Corrections will stress treatment-expansion put aside at least a year. Arizona Republic, B7. Davey, M. (2006, March 15). Iowa’s residency rules drive sex offenders underground. The New York Times, A1. Diaz, E. (2004, April 27). Sex offender bill gets push - vote stalled in Senate panel. Arizona Republic, B8. 10 Greenfield, L. A. (1997). Sex offenses and offenders: An analysis of data on rape and sex. Washington DC: U.S. Department of Justice, Bureau of Justice Statistics. Hanson, R. K., & Bussiere, M. (1998). Predicting relapse: A meta-analysis of sexual offense recidivism studies. Journal of Consulting and Clinical Psychology, 63, 50, 802-809. Levenson, J., & Cotter, L. (2005). The Impact of sex offender residence restrictions: 1,000 feet from danger or one step from absurd. International Journal of Offender Therapy and Comparative Criminology, 49 (2), 168-178. Phoenix Insurgent. (2006, January 3). Arizona’s ever-watchful eye: Moving towards a maximum surveillance and deterrence society. http://www.phx.news.com/fullstory.php?article=30207. Rennison, C. M. (2002). Criminal victimization 1001: changes 2000-01 with trends 1993-2001 (NCJ#194610). Washington, DC: U.S. Department of Justice. Rubiano, S. (2005, July 6). Arizona moves to track down 957 missing sexual predators. Arizona Republic, A1. Sexton, C. (2005, July 2). City looks at Iowa law on sex felons. Arizona Republic, B1. Walker, J., Maddan, S., Vasquez, B, VanHouten, A., & Ervine-McLarty, G. (2001). The influence of sex offender registration and notification laws in the United States. Little Rock, AR: Arkansas Crime Information Center. 11 1 An Exploration of the Offending Behavior of a Random Sample of Sex Offenders from the City of Phoenix Todd Armstrong, Ph.D. Sex crimes and policies targeted at sex offenders are of increasing public concern. Work considering the prevalence of sex offenses indicates that approximately 1-2% of the U.S. adult male population will be convicted of a sexual offense (California Office of the Attorney General, 2003; Marshall, 1997). Studies assessing the rate at which sex offenders recidivate have found that among typical groups of sexual offenders, 10-15% recidivate within 5 years (Hanson & Morton-Bourgon, 2004). This offending occurs against a backdrop in which criminal justice system policy is evolving rapidly. The Jacob Wetterling Act passed by Congress in 1994 required states to create registries of sexually violent offenders. In 1995, the Wetterling Act was amended to require states and law enforcement agencies to release registration information to the public. In a number of cases, this federal legislation had been preceded by state laws requiring registration and providing for community notification (California Office of the Attorney General, 2002; Matson & Lieb, 1996). Unfortunately, there is scant literature with which to evaluate the wisdom of laws that require sex-offender registration and community notification. A recent review of the literature found only 12 empirical investigations of the impact of community notification and registries (Welchans, 2005). Of these, only two explored the potential impact of sexoffender registries on the offending behavior of sex offenders (Matson & Lieb, 1996; Petrosino & Petrosino, 1999). Clearly, additional research on the impact of sex-offender registries is needed. In particular, we need to test the impact of registration on the behavior of sex offenders. Here we use data on a random sample of level-2 and level-3 sex offenders taken from the pool of all sex offenders in the city of Phoenix (N=244) to make an initial contribution to assessing the impact of registration on the offending behavior of sex offenders. Information included in the description of the sample will allow estimation of the rates at which offenders in the sample recidivate and comparison of the types of offenses engaged in prior to and after registration. An assessment of recidivism rates will help policymakers to understand the extent to which the sex-offender population in Phoenix is at risk for re-arrest. Further, we will explore the predictors of recidivism among this population. These predictors may begin to describe the characteristics of individuals with increased risk for recidivism. We will also contrast the types of offenses that sex offenders tend to engage in prior to registration with those that they engage in after 12 registration. This comparison offers an initial look into the impact of registration on offending behavior. Review of the Known Literature A substantial literature addresses the rate at which sex offenders recidivate. (For recent reviews, see Cottle, Lee, & Heilbrun, 2001; Doren, 1998; Hanson & Brussiere, 1998). Estimates of recidivism rates in these studies are generally low, but they vary a great deal. Early work in this area by Furby et al. (1989) found that recidivism rates varied from 0% to 88% among the studies that were considered. Factors influencing the variability of rates include the type of sex offender, the measure of recidivism, and the length of the follow-up period (Langevin et al., 2004). In a recent meta-analysis of the predictors of recidivism among sex offenders based on 95 studies, Hanson and Morton-Bourgon (2004) found the average observed sexual recidivism rate was 13.7%. Among these studies, the average follow-up time was 5-6 years. Hanson and Morton-Bourgon also found that the average violent recidivism rate (including sexual and non-sexual violence) was 25%, while the general (any) recidivism rate was 36.9%. Factors predicting the recidivism of sex offenders are typically divided into static and dynamic risk factors. Static risk factors are those attributes of the offender that cannot be changed, including demographic characteristics, family background, and treatment and offense history. Dynamic risk factors refer to offender attributes that can change with time. Static risk factors related to sex-offender recidivism include prior criminality, prior sex offenses, enduring personality disorders such as psychopathy, age, time spent in custody, and deviant sexual interests (Broadhurst & Maller, 1992; Browne et al., 1998; Hanson & Bussiere, 1998; Hanson & Harris, 2000; McGuire, 2000; Worling & Curwen, 2000). Dynamic risk factors related to sex-offender recidivism include unemployment, substance abuse, and negative social influences (Hanson & Harris, 1998; Hanson & Harris, 2000; McGuire, 2000). Studies of the predictors of recidivism have also considered the extent to which risk measures designed for the sex-offender population predict risk for recidivism. These measures, referred to in the literature as actuarial measures, can include both static and dynamic indicators of risk. Actuarial measures include the Static-99 (Hanson & Thornton, 2000), the SVR-20 (Sexual Violence Scale; Boer et al., 1997), and the SORAG (Sex Offense Risk Appraisal Guide; Quinsey, et al., 1998). Research has shown that these actuarial measures are predictive of recidivism (Hanson & Morton-Bourgon, 2004). In an effort to address sex crimes, federal and state legislation has resulted in a proliferation of sex-offender registries. At this time, however, only a limited literature addresses the potential impact of these registries on sex-offender behavior. A recent literature review identified 12 studies addressing the potential impact of sex-offender registries (Welchans, 2005). Of these, six were based on opinions of sex offenders or community members; they incorporated no direct measure of implementation or outcome. Three addressed aspects of the implementation of sex-offender registries including the percentage of offenders subjected to community notification (Matson & 13 Lieb, 1996), the accuracy of sex-offender registries (Tewksbury, 2002), and the impact of community notification on the work load of parole and probation officers (Zevitz & Farkas, 2000). One of the studies used legal documents and newspaper articles to examine the impact of registration and community notification on community panic and vigilantism (Younglove & Vitello, 2003). The two remaining studies addressed the potential impact of sex-offender registries on the behavior of sex offenders. Petrosino and Petrosino (1999) examined the nature of the sex offenses committed by sexual psychopaths (N=136) housed in a Massachusetts prison facility for sex offenders. The authors identified the group of sex offenders that had committed crimes that would have made them eligible for the sex-offender registry (n=36). They concluded that of the crimes committed by the offenders eligible for the registry, proactive police warnings could have potentially reached victims in 6 of 12 stranger-predatory cases. (Crimes committed against family members, friends, or acquaintances were not considered.) Schramm and Milloy (1995) evaluated the impact of community notification on the recidivism of sex offenders in Washington State. The authors contrasted a group of 139 sex offenders for whom community notifications were undertaken. This group was compared with 90 sex offenders released before the implementation of community notification. In a survival analysis accounting for time at risk in the community, the authors found that after 54 months at risk in the community, the notification group had a slightly lower estimated rate of recidivism for sex offenses (19%) than the comparison group (22%). Differences between the groups were not statistically significant. Groups also were contrasted with regard to recidivism, defined as a new offense of any kind. In this contrast, 57% of the community notification group had recidivated after 54 months at risk, while 47% of the comparison group had recidivated after 54 months at risk. Purpose of the Current Study This study uses a random sample of registered sex offenders from Phoenix to examine a number of issues of importance to policymakers. Analysis will include an assessment of the demographic and offense history characteristics of sex offenders in Phoenix and a review of the offense types that led to registration for a sex offense. We will also contrast the different types of offenses engaged in prior to registration with offense types engaged in after registration. Finally, we will explore the predictors of recidivism among the sample. These analyses have important policy implications. The assessment of recidivism will help to inform the extent to which the registered Phoenix sex offender population is at risk for recidivism and to identify the characteristics associated with increased risk for recidivism. The comparison of the type of offense engaged in prior to registration to those engaged in after registration will help to inform the extent to which registration influences the behavior of sex offenders. 14 Data The data used in this analysis describe the demographic characteristics and criminal justice system involvement of a random sample of 244 active sex offenders registered in the City of Phoenix. Offenders were eligible for inclusion in the sample if they were level-2 or level-3 sex offenders. The data, collected during February 2006, included demographic information describing offender age, sex, and race/ethnicity, and a variety of information detailing the offense history of the sample members. Offense-history information included date of first arrest, arrest frequency, and charge type. Additional offense-history information described the timing and nature of the offense for which the sample member was required to register. Offense-history information was derived from official records describing the criminal justice system involvement of the offender with the City of Phoenix and the State of Arizona. (Data describing criminal justice system involvement in Arizona excluded activity recorded as occurring in the city of Phoenix to avoid duplication.) Combined, these two data sources described the criminal justice system activity related to each offender occurring in Arizona. (Note that this data excluded criminal justice system activity related to the offenders that occurred outside Arizona.) This chapter’s findings are organized into two sections. The first describes offenders and their criminal justice system involvement, and includes the following sub-sections: 1) offender demographic characteristics, 2) criminal justice system involvement prior to registration, 3) the registration offense, and 4) criminal justice system involvement after registration. The second section assesses the relationship between offender characteristics and rates of offending after initial registration. Offender Demographic Characteristics Offender characteristics including race/ethnicity, gender, probation status, and sex offender level are summarized in Table 1.1. Results showed that offenders were predominantly white males. Approximately two thirds were on probation as of February 2006, and slightly more than half were level-2 sex offenders. 15 Table 1.1 Phoenix Sex-offender Race, Gender, Level, and Probation Status (N=244) Race/Ethnicity (%) American Indian Asian Black Hispanic White Gender (%) Male Female Probation status (%) 7.4 .4 11.5 11.1 69.7 98.8 1.2 Not on probation On probation 34.4 62.3 Level (%) Level-2 Level-3 57 43 Note. Data on the probation status of 8 offenders were missing. The age distribution of offenders as of February 2006 is presented in Figure 1.1. There are seven age categories, each encompassing a range of 10 years. The height of the bar represents the number of offenders of a given age. For example, on February 2006, 23 offenders fell into the first age category, 17-26. The majority of offenders ranged in age from 27 to 56 years. The average age was 41.95 years, with a median of 42. Figure 1.1 Age on February 2006 90 80 Frequency 70 60 50 40 30 20 10 0 17-26 27-36 37-46 47-56 57-66 67-76 87 Age category The distribution of age at first arrest is presented Figure 1.2. Age at first arrest is based on the date for the first arrest, considering any arrest including both sex offenses and non-sex offenses. Offenders are grouped into a category according to their age at the first arrest. For example, among the 244 offenders included in the data, 16 were 11-15 years old at their first arrest. Note that these data did not 16 include arrests occurring outside the state of Arizona, and arrests occurring in Arizona would not have been included if they were not in the datasets we used. If a large number of arrests occurred and were not included in the data, we would expect that the actual distribution of age at first arrest would reflect earlier ages than those described. The distribution presented in Figure 1.2 showed that a large number of offenders were arrested for the first time before the age of 21. In addition, a progressively smaller number of offenders were arrested for the first time at later ages, with a small number of offenders being arrested for the first time in their 40s and 50s. In the sample considered here, the average age at first arrest was 26.27 years. Due to the skewed distribution, the average age at first arrest value is slightly higher than the median age at first arrest (23.52 years). Figure 1.2 Age at First Arrest 60 Frequency 50 40 30 20 10 0 11-15 16-20 21-25 26-30 31-35 36-40 41-45 46-50 51+ Age category The distribution of age at conviction for a registration offense is presented in Figure 1.3. The registration offense is the offense for which an offender was required to register as a sex offender. The average age at conviction for a registration offense was 30.78 years. The vast majority of offenders were convicted for their registration offenses between the ages of 16 and 35. Few offenders were convicted for registration offenses before the age of 15. A moderate and decreasing number of offenders were convicted for registration offenses between the ages of 36 and 50 years. 17 Figure 1.3 Age at Conviction for Registration Offense 60 Frequency 50 40 30 20 10 0 11-15 16-20 21-25 26-30 31-35 36-40 41-45 46-50 51+ Age category Criminal Justice System Involvement Prior to Initial Registration This section explores the criminal justice system involvement of sex offenders in the Phoenix sample prior to their initial registration for a sex offense. We have included an assessment of the number of arrests prior to initial registration, a description of charges prior to initial registration, and a description of the number and type of departmental crime reports (DRs) prior to initial registration. Number of Arrests Prior to Initial Registration Sex offenders in the sample accounted for a total of 1,020 arrests prior to initial registration for a sex offense (see Fig. 1.4). The average number of arrests was 4.18, while the median was 2. The majority of offenders had relatively few arrests prior to the date of initial registration. In contrast, increasingly smaller groups of offenders had large numbers of arrests prior to initial registration. 18 Figure 1.4 Distribution of the Number of Arrests Prior to Initial Registration Number of offenders 60 50 40 30 20 10 0 0 2 4 6 8 10 12 14 17 20+ Number of offenses prior to initial registration Charges Prior to Initial Registration by Number and Type When analyzing arrests, it is important to consider both the number of arrests and the number of individual charges associated with each arrest. An offender can be and often is charged with multiple offenses at arrest; each arrest incident can include multiple charges. Also, the arrest charges based on the initial arrest may differ from the charges used if and when the person is convicted. Sample members accounted for 2,139 arrest charges made prior to their initial registration for a sex offense. Of these, 1,322 (61.8%) were for a felony offense; the remaining 817 (38.2%) were for misdemeanor offenses. Table 1.2 describes the type of charges that these offenders accrued prior to initial registration. This information is disaggregated across three categories of sex offenses and four categories of non-sex offenses. The sexual assault category includes sexual assault, rape, and other sexual abuse crimes perpetrated against adults. The child-sex crimes category includes any sex crime in which the victim is a child, regardless of severity. Other sex crimes include indecent exposure, voyeurism, prostitution, and other sex crimes not falling in either the sexual assault or child-sex categories. The violent crimes category includes homicide, robbery, kidnapping, and aggravated assault. Property crimes include burglary, theft/larceny, motor vehicle theft, and arson. The drug crimes category includes crimes of possession, use, and distribution. The other crimes category includes all crimes not included in any of the above categories, as well as charges for failure to appear and failure to pay. Offenders in the sample had 139 charges for sexual assault, representing 6.5% of all charges. In the sexual assault offense category, 27.05% of all offenders had a charge prior to the arrest that led to their initial registration. A large percentage had been arrested for 19 sex offense charges prior to registration; 32.49% of all charges prior to initial registration were for some type of sex offense, with charges for child-sex occurring most often. When considering the values presented in Table 1.2, note that single offenders may be represented in multiple categories. For example, a single offender may have had a charge for both the sexual assault and the other sex categories. Taking this into account, we found that 70.5% of all offenders had been arrested for some type of sex offense prior to registration. Table 1.2 Charge Type Prior to Initial Registration Offenders Offenders (#) (%) Charge type Sex offenses: Sexual assault 66 27.05 Child-sex 110 45.08 Other sex 63 25.82 Non-sex offenses: Violent 53 21.72 Property 84 34.43 Drug 56 22.95 Other 144 59.02 All charges (#) % of all charges 139 387 169 6.50 18.09 7.90 124 237 153 930 5.80 11.09 7.15 43.48 With regard to non-sex offense charges, we found a relatively large percentage of offenders had been arrested for different types of non-sex offenses. Within charges for non-sex offenses prior to initial registration, arrest charges in the other-offense category were most common. Among the remaining non-sex offense charge categories, charges for property offenses occurred most often. Taking into account the potential representation of single offenders in more than one offense charge category, we found 47.5% of the sample had had a charge for a violent, property, or drug offense prior to initial registration. This data implied that at least a portion of the sex offenders had tended to commit other crimes as well and were not just committing sex offenses. Departmental Reports (DRs) Prior to Initial Registration by Number and Type Next we used departmental crime reports (DRs) to explore offending behavior prior to initial registration. DRs are reports of criminal activity recorded by the police in relation to a particular individual. On a DR, a given person may be listed as a witness, victim, suspect, arrested person, investigative lead, owner, and so forth. DRs are filed on an incident basis; a single DR is filed for each crime incident coming to the attention of the police. Drug offenses and a few others are exceptions. For example, when an incident includes a drug offense, a separate report is written for the drug offense. For the purpose of this analysis, only DRs on which the offender was directly a part of the incident (listed as a suspect, indexed suspect, arrested person, or victim) were included in the counts. 20 Of the offenders in the sample, 120 had been listed on a DR in relation to a sex offense crime prior to initial registration; 528 DRs listed such an offender. The distribution of DRs across offense type is presented in Table 1.3. Among sex offenses recorded on DRs, those for child-sex offenses were most common. In the categories of non-sex offenses, DRs for “other” offenses were most common, followed by property offenses. To complement the information presented in Table 1.3, the percentage of offenders listed on DRs in relation to any sex offense and any serious non-sex offense (violent, property, or drug) was calculated. Just more than 28% (n=70) of the offenders in the sample had been listed on a DR for a sex offense prior to registration, and 29.92% (n=73) had been listed on a DR for a violent, property, or other offense. Table 1.3 Departmental Report (DR) Crime Type Prior to Initial Registration Offenders Offenders DRs Crime type (#) (%) (#) % of all DRs Sex offenses Sexual assault 13 5.33 17 3.22 Child-sex 44 18.03 54 10.23 Other sex 19 7.79 26 4.92 Non-sex offenses Violent 26 10.66 37 7.01 Property 55 22.54 133 25.19 Drug 20 8.20 32 6.06 Other 69 28.28 229 43.37 Note. In the DRs before registration, 7 offenders were listed as victims. A quick comparison of Tables 1.2 and 1.3 shows that the number of DRs listing offenders in the sample was smaller than the number of offenses they had been charged with at the time of arrest. This suggested that arrest charging data may be the more reliable indicator of offender behavior. The difference is likely in part attributable to differences in the recording of information. Charge data may reflect multiple charges for a single arrest incident, while a single DR is filed for a given crime incident. In addition, departmental crime reports would only list an offender’s name if he or she were known to the victim or recognized by a witness; otherwise the suspect’s name would be listed as unknown and the crime would be recorded with only a description of the suspect. Although not by name, offenders may actually be associated with additional DR crimes that were not accounted for in this analysis. Registration Offense The types of offenses for which offenders were initially required to register are presented in Table 1.4. The categories occurring most often were sexual assault/rape, molestation of a child, and sexual conduct with a minor. The categories occurring least often were commercial exploitation of a minor, public sex/indecency, and sexual assault of a spouse. 21 Table 1.4 Offenders’ Registration Offense Types Registration offense: Commercial exploitation of a minor Indecent exposure Kidnapping Molestation of a child Other sex offense Public sex / indecency Sexual assault / rape Sexual conduct w/ minor Sexual exploitation of a minor Sexual abuse Sexual assault of spouse Total # of offenders 1 17 3 59 20 2 65 48 3 24 2 244 % of offenders .4 7.0 1.2 24.2 8.2 .8 26.6 19.7 1.2 9.8 .8 100.0 Table 1.5 presents the elapsed time in years between points of criminal justice system involvement. This included the offender’s first arrest, arrest for a registration offense, conviction for a registration offense, and registration based on the available data for Phoenix and Arizona. First arrest was defined as the date of the first arrest recorded either in the Phoenix police database or the State of Arizona database. An average of 4.03 years elapsed between an offender’s first arrest and the arrest for which he or she initially registered as a sex offender. Approximately 38% of the offenders in the sample had a period of 6 years or more elapse between the first arrest and the registration arrest. This relatively long period may reflect incarceration, inactive offenders, or active offenders who avoided a sex-offense charge that would have led to registration. This last group may have included active offenders not committing a sex offense and offenders committing sex offenses who avoided charges from these offenses. In addition to the group with a relatively long period between first arrest and registration arrest, a group of 100 offenders (41%) had first arrests that preceded their registration arrests by less than a year. From the arrest for an offense leading to initial registration as a sex offender to conviction for that offense, an average of 0.70 years elapsed. Combined with the period between the first offense and the registration offense, an average of 4.73 years elapsed between the first arrest and conviction for the offense for which the offender was initially required to register 22 Table 1.5 Average Time Between First Arrest and Registration Years First arrest to registration arrest 4.03 Registration arrest to conviction .70 Conviction to registration Time from first arrest to registration 6.03 10.76 An average of 6.03 years elapsed between conviction for registration arrest and registration. This average is influenced by a large group with relatively short times between conviction for registration arrest and actual registration and progressively smaller groups with longer periods until registration, up to more than 10 years. Of all offenders, 73 had less than 1 year elapse between conviction and registration. Ninety offenders had less than 2 years elapse between conviction and registration. In contrast, 108 offenders had more than 5 years elapse between conviction and registration, while 60 offenders had more than 10 years. It is likely that long periods of time between conviction and registration are in part explained by incarceration. Criminal Justice System Involvement after Initial Registration This section explores the criminal justice system involvement of offenders after their initial registration for a sex offense. We have included an assessment of the number of arrest charges after registration, a description of the types of charges after registration, and a description of departmental reports (DRs) after initial registration. Number of Arrests after Initial Registration An average of 5.57 years elapsed from registration until the end of data collection. Of the 244 offenders included in the sample, 114 (46.7%) were arrested at least once after initial registration. These offenders accounted for 463 arrest charges. The frequency with which offenders were arrested after the registration offense is presented in Table 1.6. Figure 1.5 presents this same information graphically. Among all offenders in the sample, the average number of arrest charges after initial registration was 1.9 and the median was zero. The average number of arrest charges after initial registration increased to 4.06 when we considered only those with at least one arrest charge after registration. Slightly more than half of the offenders were not rearrested. Of those who were, this showed that a relatively large number of those who were arrested after registration were either arrested more than once or were arrested for multiple charges. 23 Table 1.6 Distribution of Recidivism After Registration Arrest charges Percentage of all after registration (#) offenders 0 53.3 1 16.4 2 9.0 3 6.6 4 2.0 5 2.9 6 2.5 7 1.6 8 0.8 9 1.6 10 0.4 11+ 2.8 Figure 1.5 Distribution of the Number of Arrests After Initial Registration 140 Number of offenders 120 100 80 60 40 20 0 0 2 4 6 8 10 16 20+ Number of arrests after initial registration Arrest Charges After Initial Registration by Number and Type Those committing offenses after initial registration for a sex offense accounted for 775 charges. Of these, 370 charges (47.74%) were for a felony offense. The remaining 405 charges (52.26%) were for misdemeanor offenses. Table 1.7 describes the type of offense engaged in by offenders with arrests after initial 24 registration. The information is disaggregated across three categories of sex offenses and four categories of non-sex offenses. The percentage of offenders with a charge after initial registration does not sum to 100% because many of the offenders were not charged subsequent to initial registration. A relatively small percentage of offenders in the sample had been charged with a sexual assault, child-sex offense, or other sex offense after the initial registration. Across these charge types, charges for other sex offenses were most common. Taking into account the representation of a single offender in multiple categories, we found that after registration, 20.1% of offenders in the sample had been charged with an offense falling into one of the sex-offense categories. When considering this percentage, note that offenders may have been charged with more than one type of sex offense; therefore, the percentage of offenders charged with any sex offense (20.1%) was different from the sum of the percentage of offenders charged across the sex-offense categories. Table 1.7 Charge Type After Initial Registration Offenders Offenders Charge Type (#) (%) Sex offenses: Sexual assault 15 6.15 Child-sex 14 5.74 Other sex 32 13.11 Non-sex offenses: Violent 16 6.56 Property 30 12.30 Drug 21 8.61 Other 89 36.48 Total charges (#) % of all charges 20 25 75 2.58 3.23 9.68 40 119 58 438 5.16 15.35 7.48 56.52 We found that the percentage of offenders engaging in each of the different types of nonsex offenses was relatively small, with the largest percentage for other non-sex offenses; slightly more than one third of the offenders had other non-sex offenses. Within charges for non-sex offenses after initial registration, charges for other offenses were most common. Among the remaining non-sex offense charge categories, charges for property offenses occurred most often. Taking into account the potential representation of single offenders in more than one offense charge category, we found that 20.5% of the offenders in the sample had been charged for a violent, property, or drug offense after initial registration. This percentage was different from the sum of the percentage of offenders charged across the violent, property, and drug categories in Table 1.7. The differences were driven by single offenders charged with multiple offense types. For example, a single offender may be included in the percentage of offenders with a charge for a violent offense and included in the percentage of offenders with a charge for a property offense, but will only be counted once in the percentage of offenders with a charge for either a violent, property, or drug offense. 25 Departmental Reports (DRs) After Initial Registration by Number and Type Next, we extended the analysis of offending after initial registration to a consideration of DRs. Of the offenders in the sample, 104 appeared on one or more of 399 DRs as a suspect, indexed suspect, arrested person, or victim after initial registration. The distribution of DRs across offense types is presented in Table 1.8. The information presented in Table 1.8 shows that among DRs for sex offense crimes, DRs for “other sex offenses” were most common. Among all non-sex offense crimes, ‘other’ offenses were most common, followed by property offenses. To complement the information presented in Table 1.8, the percentage of offenders listed on a DR in relation to any sex offense and the percentage of offenders on a DR in relation to any serious non-sex offense (violent, property, or drug) were calculated. Twenty-seven offenders in the sample (11.06%) were listed in relation to a sex offense after registration, while 72 (29.51%) were listed on a DR for a violent, property, or drug offense after registration. Table 1.8 Departmental Report (DR) Type After Initial Registration Crime Type Offenders (#) Offenders (%) DRs (#) Sex offenses: Sexual assault 7 2.87 7 Child-sex 6 2.46 6 Other sex 15 6.15 25 Non-sex offenses: Violent 19 7.79 42 Property 66 27.05 140 Drug 12 4.92 15 Other 57 23.36 164 % of all DRs 1.75 1.50 6.27 10.53 35.09 3.76 41.10 Note. On five of the DRs after initial registration, the offenders were listed as the victim. Comparing Offense Type Prior to Registration to Offense Type After Initial Registration Comparisons between Tables 1.2 and 1.7 (arrest charge type prior to and after initial registration) and Tables 1.3 and 1.8 (DRs prior to and after initial registration) may only be made with regard to the values in the column at the far right of each of the Tables. Tables 1.2 and 1.7 are based on an average of 10.80 years (elapsed time between first arrest and initial registration), while Tables 1.3 and 1.8 are based on an average of 5.57 years (from initial registration until the end of data collection). Differences in the amount of elapsed time directly influence the values in the second, third, and fourth columns of Tables 1.2, 1.3, 1.7, and 1.8. For example, Table 1.2 shows that prior to initial registration, 66 offenders were charged with sexual assault and that an average of 10.8 years elapsed between first arrest and initial registration. If a shorter period had elapsed between first arrest and initial registration, we would expect fewer offenders to 26 accrue a charge of sexual assault. This has direct implications for a comparison between Table 1.2 and Table 1.3: Table 1.3 covers a shorter period and differences in the amount of elapsed time undoubtedly influence values in the tables’ second, third, and fourth columns. In contrast, the values in the column at the far right of each of the tables are not directly influenced by differences in elapsed time and are, therefore, comparable. These values represent the percentage of all offenses accrued during the time period covered by the table that a particular offense type accounts for. By contrasting the number of offenses in a category occurring during a given period with all offenses occurring during that period, we effectively remove the influence of differences in the amount of elapsed time. The results summarized in Tables 1.2, 1.3, 1.7, and 1.8 suggest that the relative frequency of offenses leading to charges for sexual assault and child-sex decreases after initial registration. Frequency distributions based on arrest charge type found that prior-toregistration charges for child-sex crimes comprised 18.02% of all charges incurred by offenders in the sample, while charges for sexual assault comprised 6.5%. After registration, charges for child-sex crimes comprised 3.23% of all charges incurred by the sample and charges for sexual assault comprised 2.58%. Frequency distributions based on departmental report type show that prior-to-registration DRs for child-sex crimes comprised 10.23% of all departmental reports filed for any of the sample offenders, while departmental reports for sexual assault comprised 3.22%. After registration, DRs for child-sex crimes comprised 1.5% of all department records filed regarding sample offenders and departmental reports for sexual assault comprised 1.75% of all departmental reports incurred by the sample. These results indicate that relative to other offense types, offenses leading to arrest charges or to departmental reports for sexual assault and child-sex occur less often after initial registration for a sex offense. Changes in frequency relative to other offense types do not necessarily result in changes in the rate of a particular offense type. Rates are influenced by the number of times an event occurs and the amount of time that elapses. Thus, a particular type of charge may be less common after registration, but the rate at which that charge occurs may actually be higher due to less time having elapsed. Relationship Between Offender Characteristics and Rates of Offending After Initial Registration In this section, we consider rates of offending after initial registration and the relationship between offender characteristics and these rates. In this analysis, rates are based on the number times a particular measure of offending occurs after registration, as compared to the length of time between registration and the end of data collection. To illustrate, the calculation of arrest rates contrasts the number of arrests occurring after initial registration with the time elapsed between initial registration and the end of data collection, or: 27 Arrest rate after = # arrests registration to end datacollection Time registration to end datacollection Rates of offending after initial registration are calculated for arrests and charges across seven categories of charge type. These seven categories are identical to those described earlier in the section on criminal justice system involvement prior to initial registration. Measures of recidivism indexing the rate of charges across the seven categories were calculated in the same way as arrest rates. Table 1.9 describes the distribution of re-offending after initial registration across the sample. The second column of Table 1.9 indicates the number of sample members having a positive value for a particular measure. In the case of arrest, this indicates that they were arrested after initial registration. In the case of the charging categories, this indicates that they had a particular charge type after initial registration. The third column presents the number of sample members who did not re-offend across the different categories. The column on the far right presents the average rate per year for all sample members for arrest and for each of the charge categories. The rates of recidivism among the Phoenix sex-offender sample were shown to be low. Among offenses, child-sex offenses, sexual assault, and violent offenses occurred less frequently while “other” offenses occurred most often. Table 1.9 Re-offending Among the Phoenix Sex-offender Sample Arrests Sex offense charges: Sexual assault Child-sex Other sex Non-sex offense charges: Violent Property Drug Other Yes 114 No 130 Ave. rate per year .35 15 14 32 229 230 212 .03 .01 .06 16 30 21 89 228 214 223 155 .03 .07 .05 .33 Next, we explored the relationship between offender characteristics and re-offending. More specifically, we tested the association between offender gender, race, probation status, and sex-offender level with each of the rates of re-offending shown in Table 1.9. These results are presented in Table 1.10, which displays the rate of offending after initial registration for the different gender and racial groups included in the Phoenix sexoffender sample. A review of the average rates of re-offending for males and females showed that female rates of arrest after initial registration were slightly higher than male rates of arrest. Differences across rates for charging variables suggested that female 28 arrests resulted in charges for other sex offenses and other non-sex offenses. For each of these charge type categories, rates for females were higher than they were for males. In contrast, male rates of arrest after initial registration resulted in charges that were distributed across all of the different charge rate measures. Male rates of charges were higher than female rates for sexual assault, child-sex, violent, property, and drug offenses. Due to the small number of sample members identified as female, we were unable to analyze the statistical significance of differences in rates of offending between males and females. Note that due to the small number of females (n=3) included in the sample analyzed here, the results for females may have been heavily influenced by the behavior of single individuals and as such, may not generalize to the larger population of all female sex offenders. Table 1.10 Rates of Offending After Initial Registration for Gender and Race Groups Gender Race/Ethnicity Hispanic Male Female Amer. Asian Black (N=27) (N=241) (N=3) (N=28) Indian (N=1) White (N=170) (N=18) Arrests .346 .379 Sex offense charges: Sexual .033 .000 assault Child.011 .000 sex Other .056 .353 sex Non-sex offense charges: Violent Property Drug Other .031 .068 .050 .331 .000 .000 .026 .353 .549 .667 .501 .425 .283 n.s. .319 .000 .005 .010 .011 * .000 .000 .005 .042 .008 * .319 .000 .024 .085 .034 n.s. .015 .034 .009 .265 .000 .000 1.667 .00 .094 .054 .019 .433 .026 .159 .037 .341 .022 .059 .052 .321 n.s. n.s. ** n.s. Note. * = statistically significant at p< .05; ** = statistically significant at p <.01; n.s.= not statistically significant. An analysis of variance was used to compare rates of re-offending across the different racial groups included in the sample. Statistically significant differences were found for rates of sexual assault charges and child-sex charges. Differences for drug offense charges were also statistically significant. The rate of child-sex charges was highest for Hispanics. A quick review of the mean values of sexual assault charges across the different racial groups showed that the rate of charges for American Indians was far greater than that for the other racial groups. A check of the data found that this elevated rate was the product of the offending behavior of a single individual. Further, we found that the elevated rate of offending for this single individual was heavily influenced by the short amount of time that had elapsed between registration and the end of data collection. The individual in question had a single charge for sexual assault and a single charge for an “other” sex offense, in combination with a 29 short elapsed time from initial registration to the end of data collection (0.17 years). This resulted in an extremely high rate of offending across the sexual assault categories (5.74 charges per year). Racial group differences in drug offense charges were also the product of the behavior of a single individual. In this case, a single offender whose race was recorded as Asian accrued five charges for drug offenses over 3 years after registration, resulting in a group rate of 1.667 charges per year. The strong influence of the behavior of a single individual on the results presented here shows that caution must be used when interpreting these results. Any statements regarding the relationship between race and offending in the sample analyzed here must be tentative. This is particularly true for the groups having the fewest offenders within them. Table 1.11 presents rates of offending after initial registration by probation status and sex-offender level. A comparison of level-2 and level-3 sex offenders shows that level-3 offenders have higher rates of arrest after initial registration. This difference is statistically significant. A review of differences across charge rate variables show that level-3 offenders have higher rates of charges after initial registration for sexual assault, other sex offenses, violent offenses, drug offenses, and other offenses. Many of these differences are relatively large in magnitude, but only differences in rates of charging for other sex offenses are statistically significant. Regarding probation status, those on probation have higher rates of arrest after initial registration. Those on probation are also more likely to have charges for sex assault, child-sex offenses, other sex offenses, property offenses, drug offenses, and other offenses. Differences between the probation status groups are statistically significant for child-sex offenses and drug offenses. Table 1.11 Rates of Offending After Initial Registration by Probation Status and Sex-offender Level Sex-offender level Probation status Three Yes No Statistical Statistical Two (N=139) Significance (N=152) (N=84) Significance (N=105) Arrests .212 .523 ** .333 .295 n.s. Sex offense charges: Sexual assault Child-sex Other sex .012 .011 .010 Non-sex offense charges: Violent .018 Property .068 Drug .043 Other .185 .061 .011 .124 n.s. n.s. * .012 .016 .042 .005 .003 .027 n.s. * n.s. .046 .067 .059 .524 n.s. n.s. n.s. n.s. .029 .074 .074 .338 .036 .056 .010 .289 n.s. n.s. * n.s. Note. * = statistically significant at p< .05; ** = statistically significant at p <.01; n.s. = not statistically significant In the following section, we explore the relationship between criminal justice system involvement prior to initial registration and rates of offending after initial registration. Correlation coefficients representing the magnitude of association between measures of criminal justice system involvement prior to initial egistration and rates of offending after initial registration are presented in Table 30 1.12. Correlation coefficients are a measure of the magnitude of the relationship between two variables. Values around .1 to .2 are generally considered weak; those around .3 are considered moderate; and those above .5 are considered strong. Positive correlation coefficients indicate that two measures trend in the same direction. Put differently, as one increases, so does the other. Similarly, as one decreases, so does the other. Negative correlation coefficients indicate that the two measures trend in opposite directions: as one increases, the other decreases and vice versa. In Table 1.12, age at first arrest and age at conviction represent the age at first arrest for any offense type and age at conviction for the offense which resulted in the offender’s initial registration. Variables representing the number of arrests, the number of charges, the number of felony charges, and the number of misdemeanor charges count the number of times each of these indicators of criminal justice system involvement occurred prior to registration. Table 1.12 shows that the rate of arrest after registration is significantly related to four of the six measures of criminal justice system involvement prior to registration. In contrast, in the strong majority of cases, the relationship between the variables representing the rate of occurrence for the different charge types and the measures of criminal justice system involvement prior to initial registration is not statistically significant. The two exceptions to this pattern are the statistically significant relationship between the number of misdemeanor charges prior to registration and the rate of charges for other offenses after registration, and the number of arrests prior to registration and the rate of charges for other offenses after registration. Note that these correlations are statistically significant, but they are relatively small in magnitude. The relationship between age at first arrest and arrest rate after registration and the relationship between age at conviction and arrest rate after registration are negative. Negative correlations indicate that those who are younger at age of first arrest and age at conviction tend to have higher arrest rates after their registration for a sex offense. The number of arrests and the number of miscellaneous charges are positively related to arrest rate after registration, showing that those with more arrests prior to registration and those with more misdemeanor charges prior to registration have higher arrest rates after registration. 31 Table 1.12 Correlation Between Measures of Criminal Justice System Involvement Prior to Initial Registration and Rates of Offending After Initial Registration After Before initial registration initial Age at first Age at Arrests Charges Felony Misdemeanor registration arrest conviction (#) (#) Charges (#) Charges (#) Arrest rate -.155* -.178** .203** .081 -.062 .206** Sex offense charges: Sexual -.001 assault Child-sex -.049 Other sex .009 Non-sex offense charges: Violent -.079 Property -.068 Drug -.034 Other -.090 -.046 -.053 -.108 -.012 -.028 -.008 -.090 -.092 -.012 -.116 -.037 .085 .119 .150* -.020 -.053 -.029 -.060 -.034 -.067 -.051 -.026 .000 -.052 .013 .031 .072 -.065 -.036 -.011 -.028 -.014 .061 .065 .154* Note. * = statistically significant at p< .05; ** = statistically significant at p <.01. Table 1.13 presents the correlation between counts of arrest charges accrued prior to initial registration and rates of offending after initial registration. In the majority of cases, these correlation coefficients are not statistically significant. There are two exceptions: the correlation between the arrest rate after initial registration and both the number of charges for property offenses prior to initial registration and the number of charges for other offenses prior to initial registration. In both instances, these correlation coefficients are positive, indicating that as the number of charges for property offenses and the number for other offenses prior to initial registration increase, so does the rate of arrest for any offense after initial registration for a sex offense. Again, although statistically significant, these are relatively weak correlations. 32 Table 1.13 Correlation Between Counts of Arrest Charges Accrued Prior to Initial Registration and Rates of Offending After Initial Registration Counts of arrest charges before initial registration After Sexual Child-sex Other sex Violent Property Drug Other initial registration assault Arrest rate -.039 -.080 -.001 .036 .154* .049 .133* Sex offense charges: Sexual assault -.024 Child-sex -.068 Other sex -.043 Non-sex offense charges: Violent Property Drug Other .002 -.029 -.006 -.021 -.024 -.019 -.032 -.025 .007 .064 -.025 -.071 -.042 -.032 -.041 -.016 -.030 -.008 .026 -.033 -.042 -.031 -.029 -.041 -.018 -.055 -.049 -.054 -.039 .002 -.037 -.045 -.050 .033 -.047 .081 -.007 .121 -.032 -.018 .079 .025 -.019 .043 .068 .111 Note. * = statistically significant at p< .05; ** = statistically significant at p <.01. Summary The data used in this analysis were based on a random sample of 244 active sex offenders registered in the City of Phoenix. This sample was predominantly White (64.3%) and male (98.8%). At the time of data collection, sample members averaged 41.94 years of age. The sample contained more level-2 offenders (57%) than level-3 offenders (43%). Of all sample members, approximately two thirds (62.3%) were on probation. The average age at first arrest for any offense of sample members was 26.27 years. The analyses presented above described the offending behavior of sample members and explored the relationship between sample member characteristics and rates of reoffending after initial registration. In the description of the offending behavior of the sample, we explored the timing of offending and the types of offenses committed prior to initial registration, as well as the types of offenses committed after registration. Below, results in each of these areas are summarized. After first arrest, an average of 4.03 years elapsed before arrest for the offense that resulted in registration. Between arrest for registration and actual registration, an average of 6.76 years elapsed, with the bulk of this time lying between conviction for the registration offense and initial registration itself (average = 6.03 years). Using departmental crime reports and arrest charge type information to measure offense type, we found that the most common offense prior to initial registration was an offense categorized as “other non-sex offense.” The next two most common offense types were property offenses and child-sex offenses. When measured by arrest charge type, the three least common offense types were violent offenses, sexual assault offenses, and drug offenses. When measured by departmental report type, the three least common offense types were sexual assault offenses, other sex offenses, and drug offenses. With regard to 33 the offense leading to initial registration as a sex offender, the three most commonly occurring offense types were sexual assault/rape, child molestation, and sexual conduct with a minor. The three least frequently occurring offense types were commercial exploitation of a minor, sexual assault of a spouse, and public sex/indecency. After registration, the two most frequent offense types, as indicated by arrest charge type and DRs, were offenses categorized as “other non-sex offenses” and property offenses. When measured by arrest charge type, the third most commonly occurring offense was other sex offenses. Violent offenses were the third most common offense when using DRs as the indicator of offense type. After registration, the two least frequent offense types as indicated by arrest charge type and DRs were sexual assault offenses and childsex offenses. DRs indicated that the third least frequent offense type was drug offenses, while arrest charge type indicated that the third least frequent offense type was violent offenses. A comparison of the relative frequency of the different types of arrest charges and DRs prior to and after initial registration shows a decline in the relative frequency of sexual assault offenses and child-sex offenses, but an increase in the relative frequency of other sex offenses. When assessing changes in the relative frequency of offense type before and after registration, we must realize that changes in frequency relative to other offense types do not necessarily result in changes in the rate of a particular offense type. Rates are influenced by the number of times an event occurs and the amount of time that elapses. Thus, a particular type of charge may be less common relative to the total number of charges occurring after registration, but the rate at which that charge occurs may actually be higher due to less time having elapsed. In addition to providing a description of the offending behavior of the sample members, the analyses upon which this chapter is based also assessed the extent to which offender characteristics were associated with re-offending after initial registration. To test this, offender characteristics were contrasted across groups based on race, gender, probation status, and sex-offender level. Rates of re-offending included rates of arrests and rates of charges across seven categories: sexual assault, child-sex, other sex, violent, property, drug, and other. Statistically significant differences in rates of re-offending were found across race; however, small numbers of cases within some race groups rendered these differences tentative and differences may not generalize to the larger population of all sex offenders. Sex-offender level was significantly related to rates of arrest after initial registration and rates of charges for other sex crimes after initial registration. In each of these cases, level3 offenders had higher rates of re-offending than level-2 offenders. Probation status was significantly related to rates of charges for child-sex crimes and rates of charges for drug crimes. For both of these measures, those who were on probation had higher rates. The relationship between offender characteristics and rates of re-offending was further explored by testing the association between measures of criminal justice system involvement prior to initial registration and rates of offending after initial registration. Measures of criminal justice system involvement included age at first 34 arrest, age at conviction, number of arrests prior to initial registration, and variables representing the number of arrest charges prior to initial registration across ten different categories: total, felony, misdemeanor, sexual assault, child-sex, other sex, violent, property, drug, and other. The rate of arrest was the only measure of criminal justice system involvement subsequent to initial registration as a sex offender that was consistently related to measures of criminal justice system involvement prior to initial registration. Measures of criminal justice system involvement prior to registration that were significantly related to the rate of arrest after initial registration included age at first arrest, age at conviction, the number of arrests, the number of misdemeanor charges, the number of property charges, and the number of charges in the “other” category. Age at first arrest and age at conviction were negatively related to the rate of arrest after initial registration. Among sample members, as age at first arrest and age at conviction increased, the rate of arrest after initial registration decreased. In contrast, the number of arrests, misdemeanor charges, property charges, and “other offense” charges were positively related to the rate of arrest subsequent to initial registration. Sample members with higher values of these characteristics also tended to have higher rates of arrest subsequent to initial registration. Caution should be taken when interpreting and applying these results. There are two main methodological issues with potential implications for the current work. The first is the potential influence of the behavior of single individuals when group sizes are extremely small. In the analysis of differences in the rates of re-offending across racial groups, the behavior of a single individual had a major impact on group rates and, by extension, influenced the overall results. This demonstrated that results based on groups with few members may not apply to the larger population of all sex offenders. The second methodological issue was the low rate of sexual re-offending based on arrests after registration. Low rates of re-offending made it difficult to detect potential relationships as statistically significant. For example, in Table 1.9, level-2 sex offenders had a rate of charges for other sex crimes that was 0.01, while the level-3 offenders had a rate of 0.124. This difference was more than 10 times in magnitude, but it was not detected as statistically significant, in part as a function of the overall rarity of charges for other sex crimes after initial registration. Thus it is possible that the results presented here were influenced by the relatively rare occurrence among this sample. Policy Implications These findings suggested that registration was related to the type of offenses engaged in by sex offenders. Specifically, we found that after registration, sex offenders showed a decreased tendency to commit sex offenses relative to the total number of offenses that they engaged in. This did not necessarily imply a decrease in the rate of sex offending as a function of registration. Unfortunately, we are unable to comment directly on the impact of registration on rates of recidivism. Such an analysis would require information not present in the data. Further, it is possible that differences in the offense types engaged in prior to registration and those engaged in after registration may be the result of any 35 number of forms of criminal justice involvement co-occurring with registration, including arrest, conviction, and probation. Changes may also be influenced by offender characteristics such as age. Although there was a decreased tendency to commit sex offenses relative to other types of offenses after registration, there is no way to know for sure that the act of registration itself led to the decrease. Given potential competing explanations, strong policy inferences regarding the impact of registration should await more definitive findings. A more definitive analysis, though relatively straightforward, would require additional information describing the offending of the random sample of Phoenix sex offenders analyzed herein, as well as information describing the offending behavior of sex offenders convicted and released prior to the implementation of registration and community notification in Arizona. The analysis finds that offense history and offender demographic and case processing information are poor predictors of risk of recidivism for sex offenses after registration. This is consistent with prior research examining the predictors of sexual recidivism after registration (Schram & Milloy, 1995); however, the broader sex-offender recidivism literature has found a number of static and dynamic offender characteristics that are associated with recidivism. (For a review, see Hanson and Morton-Bourgon, 2004.) Policymakers interested in predicting risk should realize that the preponderance of research indicates that such prediction is extremely difficult and should be only entertained with careful consideration of the ramifications of inaccurate prediction. Nonetheless, risk prediction is an important part of many criminal justice system policies regarding registered sex offenders. Predictions of risk may be improved through the incorporation of actuarial tools with established reliability and validity and through the incorporation of characteristics found to be predictive of sex offending in the literature. In any case, rubrics for risk prediction should be validated through careful research in the Phoenix sex-offender population. List of References Boer, D., Hart, S., Kripp, R., & Webster, C. D. (1997). Manual for the sexual violence risk-20. Vancouver, BC: The Mental Health Law and Policy Institute of the Simon Fraser University. Broadhurst, R. G., & Maller, R. A. (1992). The recidivism of sex offenders in the Western Australian prison population. British Journal of Criminology, 32: 54-77. Browne, K. D., Foreman, L., & Middleton, D. (1998). Predicting treatment dropout in sex offenders. Child Abuse Review, 7: 402-419. Cottle, C. C., Lee, R. J., & Heilbrun (2001). The prediction of criminal recidivism in juveniles. Criminal Justice and Behavior, 28: 367-394. 36 Doren, D. M. (1998). Recidivism base rates, predictions of sex offender recidivism, and the “sexual predator” commitment laws. Behavioral Sciences and the Law, 16: 97-114. Furby, L., Weinrott, M. R., & Blackshaw, L. (1989). Sex offender recidivism: A review. Psychological Bulletin, 105: 3-30. Hanson, R. K., & Bussiere, M. T. (1998). Predicting relapse: A meta-analysis of sexual offender recidivism studies. Journal of Consulting and Clinical Psychology, 66: 348-362. Hanson, R. K., & Harris, A. (2000). Where should we intervene? Dynamic predictors of sexual offense recidivism. Criminal Justice and Behavior, 27: 6-35. Hanson, R. K., & Morton-Bourgon, K. (2004). Predictors of sexual recidivism: An updated meta-Analysis 2002-04. Ottawa: Public Works and Government Services Canada. Hanson, R. K., & Thornton, D. (2000). Improving risk assessments for sex offenders: A comparison of three actuarial scales. Law and Human Behavior, 24: 119-136. Langevin, R., Curnoe, S., Fedoroff, P., Bennett, R., Langevin, M., Peever, C., Pettica, R., & Sandhu, S. (2004). Lifetime sex offender recidivism: A 25-year follow-up study. Canadian Journal of Criminology and Criminal Justice, 46: 531-552. Marshall, P. (1997). The prevalence of convictions for sexual offending. Research Finding No. 55. Research and Statistics Directorate. London: Home Office. Matson, S., & Leib (1996). Community notification in Washington State: 1996 survey of law enforcement. Olympia, WA: Washington State Institute for Public Policy. McGuire, T. J. (2000). Correctional institution based sex offender treatment: A lapse behavior study. Behavioral Sciences and the Law, 18: 57-71. Petrosino, A. J., & Petrosino, C. (1999). The public safety potential of Megan’s Law in Massachusetts: An assessment from a sample of criminal sexual psychopaths. Crime & Delinquency, 45: 140-158. Quinsey, V. L., Harris, G. T., Rice, M. E., & Cormier, C. (1998). Violent offenders: Appraising and managing risk. Washington, DC: American Psychological Association. Schram, D. D., & Milloy, C. D. (1995). Community notification: A study of characteristics and recidivism. Olympia, WA: Washington State Institute for Public Policy. Tewksbury, R. (2002). Validity and utility of the Kentucky Sex Offender Registry. Federal Probation, 66: 21-26. 37 Welchans, S. (2005). Megan’s Law: Evaluations of sexual offender registries. Criminal Justice Policy Review, 16: 123-140. Worling & Curwen (2000). Adolescent sexual offender recidivism: Success of specialized treatment and implications for risk prediction. Child Abuse and Neglect, 24: 965-982. Younglove, J. A., & Vitello, C. J. (2003). Community notification provisions of “Megan’s Law” from a therapeutic jurisprudence perspective: A case study. American Journal of Forensic Psychology, 21: 25-38. Zevitz, R. G., & Farkas, M. A. (2000). The impact of sex-offender community notification on probation/parole in Wisconsin. International Journal of Offender Therapy & Comparative Criminology, 44: 8-21. 38 39 2 An Assessment of the Accuracy of Sex-offender Address Registration in Phoenix, Arizona Charles M. Katz, Ph.D. As one of our responsibilities for the Phoenix Sex Offender Clustering project, we were asked to interview registered sex offenders on a number of topics. This required us to obtain the addresses of registered sex offenders from the Phoenix Police Department and to send teams of interviewers to their residences. Because Arizona sex offenders are required by law to notify the police of any change in address and because of the substantial penalties imposed by law if they fail to register, we fully expected that it would not be difficult to locate and interview a relatively small number (100) of registered sex offenders. During our initial field test of the survey instrument and methodology, however, when the teams went to those registered addresses, they had trouble locating the offenders. The sex-offender address registration information being collected, maintained, and disseminated by the police department was not as accurate as expected. These findings, discussed in this section, were not part of the original study as requested by the Phoenix City Council or the Phoenix Police Department. Still, these are perhaps the most important findings to come from the study. Nearly every policy, procedure, and law pertaining to sex-offender registration, notification, and residency restriction is based on the premise that the police can and do maintain accurate data on where sex offenders reside. To our knowledge, this is one of the first studies in the United States to examine the accuracy of the address data kept by police agencies in sex-offender registries. (For an exception, see Tewksbury, 2002). Literature Review The practice of sex-offender registration arises from the belief that sex offenders are significantly more likely than many other types of offenders to recidivate. This claim has been subject to much debate and research, but sex offenders do constitute a major social problem in America today. Law makers and criminal justice policymakers have responded by enacting sex-offender registration requirements. The Jacob Wetterling Crimes against Children and Sexually Violent Offender Registration Act of 1994 requires states to maintain registries of all offenders convicted of violent sex crimes and crimes committed against children. The law requires offenders to verify their addresses annually for 10 years; 39 40 sexually violent offenders must verify their addresses quarterly for life (Center for Sex Offender Management, 1999). The Adam Walsh Child Protection and Safety Act of 2006 established a national sex-offender database to integrate state registration systems; this act also increased the penalties for sex offenders who failed to register. Sex-offender registration is intended to serve three purposes. First, registration is meant to deter sex offenders from offending again. Sex-offender registries are believed to enable police to monitor the locations and actions of sex offenders, collecting information on their patterns of offending behavior. The presumptions are that registered sex offenders know that police monitor their whereabouts and behavior, and that offenders are deterred by the knowledge that if they re-offend, they will be apprehended. Second, sex-offender registration is believed to increase the effectiveness of police investigations of sex crimes. Information and intelligence gathered through sex-offender registration can help with making arrests for unsolved crimes; for example, information on sex offenders’ residences and prior offending behavior can help match likely suspects with those crimes. Third, citizens in most jurisdictions have access to sex-offender registration information through online resources or community notification programs. Where this is true, sex-offender registration is believed to increase public safety and selfprotection (Rudin, 1996; Center for Sex Offender Management, 1999). Sex-offender registration is required in at least 40 states (Tewksbury, 2006). However, little research has examined offender compliance. Plotkinoff and Woolfson’s examination of sex-offender registration compliance in England and Wales in 2000 was one of the first such attempts. Their study, commissioned by the Research, Development and Statistics Directorate of the Home Office, reported that about 96% of registered sex offenders were in compliance with registry guidelines. Of 8,608 sex offenders required by law to register, “6,262 (73%) had registered; 1,993 (23%) were in custody or were within the allowed 14-day period; and 353 (4%) appeared to be committing an offence by failing to register” (Plotkinoff & Woolfson, 2000: 5). Over the past 3 years, Arizona criminal justice policymakers have begun examining sexoffender registration compliance rates. In 2003, the Arizona Republic reported that about 6% of the state’s sex offenders were not registered as legally required, meaning that police could not account for about 750 of them (Villa, 2003). Two years later, in July 2005, the situation had not improved: The Department of Public Safety announced that roughly 7% of sex offenders had not registered with the police. According to the department, 957 sex offenders were not registered, of which 126 were level-3 offenders -those considered most dangerous by the police. The department, along with the governor, committed that it would allocate additional funds and personnel to locate unregistered sex offenders, and that it would substantially reduce the number of unregistered offenders within 6 months (Rubiano, 2005). The above research was focused on registration compliance rates; it stopped short of examining whether existing registry data was accurate or valid. To our knowledge, only one study has examined the accuracy of registry information in the United States. Tewksbury (2002) examined 537 registered sex offenders from the Kentucky State Police 40 41 online sex-offender registry. He reported finding that much of the information that should have been posted at the Web site was missing. More than 40% of the registry pages were missing offender photos. Tewksbury also examined the authenticity of the addresses provided by registrants. The research did not validate that offenders lived at their registered addresses, but it did examine whether residency was possible: in other words, whether the addresses even existed. He found that about 5% of all addresses in the registry were nonexistent and that more than 10% of the pages listed registrants’ addresses as unknown. His findings raised serious questions about the validity and thus the utility of sex-offender registries. Arizona’s Sex-offender Registration Laws and Phoenix Police Department Policies and Procedures Arizona state law requires that sex offenders register with their county sheriff’s office within 10 days after conviction or after entering and remaining in any Arizona county (Ariz. Rev. Stat. § 13-3821-A). Sex-offender registration is carried out by the records staff of the Maricopa County Sheriff’s Department in conjunction with the sex offender (Phoenix Police Department, 2004: 1). Offenders who are required to register must also register changes of name and address with the sheriff’s office, in person and in writing, within 72 hours of the change (not including weekends and legal holidays; Ariz. Rev. Stat. § 13-3822-A). Offenders without fixed addresses are required to register their general locations; these are classified as transient, and they must re-register every 90 days. According to Phoenix Police Department policy (C.06), a police assistant from the department’s Sex Crimes Unit, Sex Offender Notification Squad, must collect registrations from the sheriff’s office twice a week (Phoenix Police Department, 2004: 12). The unit is staffed with eight detectives and two police assistants. Its standard operating procedures state that if possible, after the unit is made aware of a registered sex offender, it should interview the offender and record the interview on an audiocassette. During such interviews, offenders are asked to sign a Phoenix Police Department sexoffender registration form, and they are told that they are subject to the state’s neighborhood-notification law (Phoenix Police Department, 2004: 2). Offenders are required during the interview to state their name, date of birth, physical description, home address, phone number, university/college address, vehicle information, mailing address, and sex offenses committed.4 Failure to register under the guidelines established under Arizona Revised Statute § 13-3821-A is punishable as a Class 4 felony (Ariz. Rev. Stat. § 13-3824). Detectives are responsible for updating sex-offender files, including address verifications, every 12 months for level-2 offenders and every 6 months for level3 offenders (Phoenix Police Department, 2004). There are no written policies or procedures for address verification, but we were told by the department that it conducts the checks by interviewing the offender by telephone or in person at the offender’s residence. When an offender is on supervised parole or probation, detectives 4 Francis, B. (2006). Personal communication via e-mail on Thursday June 8, 2006. 41 42 may verify the offender’s residence in coordination with Arizona Department of Corrections and/or Adult Probation Department.5,6 The Sex Crimes Unit is responsible for conducting all initial follow-up investigations of sex offenders who fail to register (Phoenix Police Department, 2004:1). This means that the unit must conduct more than 900 6- and 12-month address verifications for level-2 and level-3 offenders, in addition to more than 1,500 18- and 24-month address verifications for level-0 and level-1 offenders. Present Study In this section, we will assess the accuracy of the sex-offender address registration data held by the Phoenix Police Department. As noted, if the sex-offender registry is to fulfill its goals of deterring sex offenders, increasing the effectiveness of police investigations of sex crimes, and increasing public safety and self-protection, it must contain accurate data on sex offenders’ residences. Below, we discuss our methods and present our findings. Methodology The Phoenix Police Department provided Arizona State University with a database of all level-2 and -3 sex offenders registered in the city of Phoenix (N=932).7 The database contained the names, addresses, ethnicities, ages, dates of last address verification, offender classes, and probation status of the sex offenders. All data came from the department’s Sex Crimes Unit, Sex Offender Notification Squad, which is “responsible for the community notification of registered sex offenders, as well as the initial follow-up investigations on ‘fail to register’ violations” (Phoenix Police Department, 2004: 1). Initially, the database was provided to ASU for research purposes other than the examination of the validity of registered sex-offender addresses. Rather, it was provided so that we could locate registered sex offenders at their residences and interview them about their perceptions and experiences with sex-offender registration and notification. During a pre-test of the interview instrument, however, we learned that many of the addresses provided by the department were inaccurate. We then constructed a work sheet for interviewers to use in documenting their attempts to contact the offenders. The research processes developed to determine whether a registered sex offender resided at the address recorded in the Phoenix Police Department registry are discussed below. 5 Kostelac, C. (2006). Personal communication via telephone on Tuesday June 13, 2006. The Sex Crimes Unit does not currently track the verification method used in each case; it simply records the date of the most recent verification (Francis, 2006). 7 The police department provided two databases for this study. The first (N=22), received on October 6, 2005, was of limited value due to difficulties encountered contacting the offenders. After discussions with the department and colleagues across the country, we eventually developed a new methodology that allowed both examination of the accuracy of registered addresses and collection of the sex-offender interview data called for in the original research agreement, albeit in a limited manner. The department provided the second database on February 18, 2006; this one contained the data we used for most of the analyses discussed in this report. 6 42 43 Nineteen field interviewers were hired for this project. All were thoroughly screened, including a criminal background check, prior to being hired. The interviewers received approximately four hours of classroom training to learn how to make contact with a sex offender, how to identify themselves as employees of Arizona State University,8 and how to administer the interview instrument. Their training also reviewed issues pertaining to human subjects protections. Special attention was given to maintaining personal safety and to processes for determining whether the sex offender resided at the address, without jeopardizing individual privacy. Last, the interviewers were trained in how to give the $20 incentive to all participants. Field interviewers were directed to attempt to contact registered sex offenders at various times of the day (morning, midday, and evening) and on differing days of the week (week days and weekends).9 Due to budget constraints, interviewers were instructed to attempt contact no more than three times at each address. They were also trained in how to question other persons at the registrant’s address to determine whether the offender lived there. If the interviewer determined that the registered sex offender no longer lived at the address, this was noted on the work sheet and the information was provided to the project manager. If no contact was made at the actual address, interviewers were to contact nearby neighbors, residents, facility managers, and others until they could determine whether or not the registrant lived at the address. When they could confirm an offender’s residency, that information was noted on a work sheet. When they determined from information provided by an independent source that an offender did not live at the registered address, that information also was noted on the work sheet and given to the project manager. When interviewers got their most conclusive information from someone other than the offender, a neighbor, or facility staff (at transitional housing), they recorded this contact as a determination that the offender no longer resided at the registered address. Once we received the database from the police department, we randomly selected 300 registered sex offenders using the SPSS (Statistical Package for the Social Sciences) random-selection function. The resulting information was given to the ASU-trained field interview teams. When the first list was exhausted, we continued to randomly select additional subjects from the database. Eventually, 523 individuals were selected for interviewing in order to complete 100 interviews. (Teams did not visit all addresses three times. When 100 interviews were successfully completed, field activity ended. Interviewers who had received names selected near the end of the project often did not have time to attempt all three contacts. The impact of this is discussed in more detail below.) Work sheets were given to the project manager and the pertinent information was electronically recorded, after which the work sheets were destroyed. None of the individual data collected could be traced back to any individual. We purposely did not construct a link in our database with any individual names (once work sheets were turned in) so that confidentially could not be compromised at a later date. 8 9 Interviewers were required to wear an Arizona State University identification badge. We instructed interviewers to make their contacts only between 9 a.m. and 8 p.m. 43 44 After the project manager determined that there was no reason to make further contact with an individual, that person’s name was removed from the database to prevent it from being reselected and given to another interviewer. Finally, we thought it important for our interviewers to go into the field in teams. Each team consisted of one male and one female. We did this in part to ensure interviewer safety. Also, we wanted to offer interviewees an opportunity to be interviewed by either a male or female to make sure that they felt as comfortable as possible to share sex-related information with the interviewer. Last, interviewing in teams increased the level of interviewer accountability in the field. Findings Table 2.1 charts descriptive data for the population of all registered level-2 and -3 sex offenders in Phoenix, Arizona, for the randomly selected sample, and for the group of offenders not selected. At the time of this study, 932 level-2 and -3 sex offenders were registered as living in Phoenix. We randomly selected 523 (56.1%) for our study sample. Level-2 offenders accounted for 52.4% of the sample; level-3 offenders accounted for 47.6%. With one exception, the characteristics of the randomly selected sample did not differ significantly from those of the unselected group. The exception: offenders in the random sample averaged a longer elapsed time than unselected offenders since their most recent address verifications: 203.3 days for offenders selected compared with 160.7 days for those not selected. 44 45 Table 2.1 Sample and Population Descriptive Data Percent of Population Random Subjects Not All Subjects Selected Sample Gender Male Female Ethnicity Caucasian African American Hispanic Native American Asian Missing Probation No Yes Missing Offender classification Level 2 Level 3 Mean days since address last verified (n) Level 2 Level 3 Levels 2 & 3 n= 98.9 1.1 97.3 2.7 98.2 1.8 60.2 15.1 15.1 8.4 0.6 0.6 62.3 13.9 17.4 4.2 0.5 1.7 61.2 14.6 16.1 6.5 0.5 1.1 30.3 66.3 3.3 35.7 61.1 3.2 32.7 64.1 3.2 52.4 47.6 53.5 46.5 52.9 47.1 237.6* 165.2* 203.3* 523 180.2 138.2 160.7 409 212.1 153.5 184.5 932 *Note. p < .05 Table 2.2 presents the result of all attempted contacts with sex offenders at their registered home addresses. Of the 520 registered sex offenders10 whom we attempted to contact, 27.9% lived at the registered address, 39.6% did not live at the registered address, and 5.8% not only did not live at the registered address, but that address was nonexistent. In 12.1% of the cases, we attempted to make contact with the sex offender on three separate occasions (at different times and on different days of the week), but we were unable to confirm or disprove the offender’s residency there, even after questioning neighbors and others. In about 14% of cases in which at least one attempt at contact was made, interviewers fell short of making three attempts because once the project had met its goal, field operations were ended. 10 Three randomly selected offenders were eliminated to preserve interviewer safety. In one case, the registered address was proximate to a fenced area containing hostile dogs. In another case, interviewers believed that the building at the registered address would be unsafe to enter (abandoned house/drug house). In the final case, the person who answered the door was aggressive and the interviewers ceased contact. 45 46 Table 2.2 Results: Attempted Contacts with Offenders at Registered Addresses Frequency Contact made, address verified 145 Registrant determined not to live at address 206 Address did not exist 30 Attempted contact 3 times, no response 63 < 3 contacts, no response; project ended 76 Total 520 Percent 27.9 39.6 5.8 12.1 14.6 100.0 Table 2.3 shows the results of attempted contacts with sex offenders at their registered addresses, by the number of attempted contacts. Our analysis showed that 85.5% of successful contacts with offenders at their registered addresses occurred on the first attempt; 14.5% occurred on the second attempt. No third attempts at contact were successful. Similarly, when it could be determined that an offender did not live at the registered address, 92.8% of the time that was learned on the first attempt; 7.2% of the time it was learned on the second attempt; and none of the third attempts yielded this information. When the project ended, the teams had attempted to make contact with 76 registered sex offenders for whom no information was obtained during the first or second attempts (65.8% and 34.2%, respectively); the project ended before a third attempt could be made. Table 2.3 Results: Attempted Contacts with Offenders at Registered Addresses by Number of Attempts Percent One Two Three Contact made, address verified 85.5 14.5 0.0 Registrant determined not to live at address 92.8 7.2 0.0 Attempted contact 3 times, no response 0.0 0.0 100.0 < 3 contacts, no response; project ended 65.8 34.2 0.0 Table 2.4 presents the outcomes of our attempted contacts with registered sex offenders by offender characteristics. Offender characteristics were not shown to be significantly associated with successful offender contacts. About 28% of male offenders and 50% of female offenders (n=3) were found to reside at the address where they were registered. Roughly 45% of male offenders and 50% of female offenders (n=3) were determined not to live at the registered address. We did not make contact with 12.3% of the males included in the sample after three attempts, nor with 14.8% of males after one or two attempts. 46 n 145 236 63 76 47 Contact was made and the offender’s address was verified 28.1% of the time when the offender was Caucasian, 21.5% when they were African American, 24.4% when they were Hispanic, 43.2% when they were Native American, 33.3% when they were Asian, and 33.3% of the time for offenders whose ethnicity data was missing. Forty-five percent of registered Caucasian sex offenders were determined not to have lived at the address where they were registered, compared with 51.9% of African Americans, 44.9% of Hispanics, 38.6% of Native Americans, 33.3% of Asians, and 33.3% of those whose ethnicity information was missing. Slightly more than 12% of Caucasians could not be contacted at the registered address after three attempts, compared with 5.1% of African Americans, 14.1% of Hispanics, 15.9% of Native Americans, 33.3% of Asians, and 33.3% of those with missing ethnicity data. Last, 14% of Caucasian offenders had not been contacted and fewer than three attempts had been made when the project ended, compared with 21.5% of African Americans, 16.7% of Hispanics, and 2.3% of Native Americans. The probationary status of registered sex offenders was not associated with the success of the attempted contacts. Contact (and address verification) was made with 29.2% of registered sex offenders on probation and with 24.2% of registered sex offenders not on probation. We found that 43.9% of sex offenders on probation and 46.5% of sex offenders not on probation did not live at their registered addresses. We did not make contact with 10.7% of those on probation or with 16.6% of those not on probation after three attempted contacts, and we did not make contact with 12.7% of non-probationers or 16.2% of probationers after one or two attempted contacts. Neither offender classification nor the number of days since the offender’s address was last verified were significantly related to whether or not contact was made with the offenders (see Table 2.4). 47 48 Table 2.4 Results: Attempted Contacts with Offenders by Offender Characteristics Percent Contact Registrant Attempted Project made, does not live contact 3 ended; 1 or 2 address at address times, no contacts, no verified (confirmed) response response Gender Male 27.6 45.3 12.3 14.8 Female 50.0 50.0 0.0 0.0 Ethnicity Caucasian 28.1 45.0 12.5 14.4 African American 21.5 51.9 5.1 21.5 Hispanic 24.4 44.9 14.1 16.7 Native American 43.2 38.6 15.9 2.3 Asian 33.3 33.3 33.3 0.0 Missing 33.3 33.3 33.3 0.0 Probation No 24.2 46.5 16.6 12.7 Yes 29.2 43.9 10.7 16.2 Offender classification Level 2 29.4 42.3 12.9 15.4 Level 3 26.2 48.8 11.3 13.7 Days since address last verified Mean 191.3 206.5 242.2 177.3 S.D. 159.8 197.9 160.4 181.1 *Note. p < .05 Summary and Conclusions The Phoenix City Council and the Phoenix Police Department did not initially request an analysis of the validity of the addresses in their sex offender registry. The need for such an analysis became apparent only after questions were raised about the data after a number of failed attempts to contact sex offenders at their registered addresses for the information specified in the original request. As it turned out, our findings have serious policy implications both for the city and the state. Nearly every policy, procedure, and law pertaining to sex-offender registration, community notification, and residency restriction is based on the premise that accurate sex-offender residence information is collected and maintained by the police department. In reality, we found that the majority of Arizona’s most violent and dangerous sex offenders did not reside at the addresses at which they were registered. 48 n 514 6 313 79 78 44 3 3 157 346 272 248 49 Due to the methodological strategies used for this study, we were unable to estimate precisely the number of sex offenders who did not live at their registered addresses. However, the data did allow for estimating a range of that number. One approach would be to restrict analysis to those cases in which we could verify whether the registrant did or did not live at the address. Using this approach, we would include only individuals for whom (1) we had made contact and verified the address, or (2) we had determined that the registrant did not live at the address, or (3) we had found that the registrant’s address was nonexistent. As Table 2.5 shows, this approach results in the estimate that 61.9% of registered sex offenders do not live at their registered addresses. This should be considered our most conservative or restrictive estimate. Table 2.5 Conservative Estimate of the Accuracy of Sex-Offender’s Registered Addresses Frequency Registrant contact made, address verified 145 Registrant does not live at address (confirmed) 206 Registered address does not exist 30 Total 381 Percent 38.1 54.1 7.8 100.0 A less conservative or restrictive strategy would be to include cases where multiple contacts were made, but did not result in confirmation of whether the sex offender did or did not live at the registered address. This strategy would require making several probability-based assumptions. • The first assumption would be that those cases that did not result in confirmation of the sex offender living at the registered address after three attempted contacts (cell e) should be recoded as “does not live at address” (cell g). This assumption is based on the belief that after three failed contacts, the registered offender is highly unlikely to be living at that address. • Second, we would make the assumption that those cases that did not result in confirmation of the sex offender living at the registered address after two attempted contacts (cell c) should be recoded as “does not live at address” (cell g). This is justified based on the fact that no third attempted contact with a registered sex offender in this study ever resulted in a confirmation that the offender lived at the registered address (cell d). • Third, we would make the assumption that of those cases that did not result in confirmation of the sex offender living at the registered address after one attempted contact (cell a), 14.5% (n=7) should be recoded as “lives at address” (cell f) and 85.5% (n=43) should be recoded as “does not live at address”(cell g). This assumption is justified based on our 49 50 finding that second attempted contacts (cell b) were successful 14.5% of the time, but third attempted contacts (cell d) were never successful. (See Table 2.6.) Table 2.6 Results: Attempted Contacts with Offenders at Registered Addresses by Number of Attempted Contacts Number of attempted contacts One Two Three N Contact made, address verified N 124 21 b 0d 145 f % 85.5 14.5 0.0 100 Registrant determined not to live at address N 219 17 0 236 g % 92.8 7.2 0.0 100 Attempted contact 3 times, no response N 0 0 63 e 63 % 0.0 0.0 100.0 100 < 3 contacts, no response; project ended N 50 a 26 c 0 76 % 65.8 34.2 0.0 100 Note. Superscripts indicate cells referenced in text. Table 2.7 summarizes the recoded categories and shows the results: Based on the probability-based assumptions discussed above, an estimated 70.8% of Phoenix’s registered sex offenders did not live at their registered addresses. Table 2.7 Recoded Categories Based on Probability-based Assumptions Original Recoded as Contact made, address verified No coding change Registrant determined not to live at No coding change address Attempted contact 3 times, no response Does not live at address Attempted contact 2 times, no response Does not live at address Attempted contact 1 time, no response Lives at Address Attempted contact 1 time, no response Does not live at address Total Frequency 145 236 % 27.9 45.4 63 26 7 43 520 12.1 5.0 1.3 8.3 100.0 Accordingly, based on the above analysis, it is our opinion that between 61.9 and 70.8% of Phoenix’s registered sex offenders do not reside at their registered addresses. Further, these figures do not include the roughly 6-7% of sex offenders who have failed to register with the county at all, and so were not included in this analysis (Rubiano, 50 51 2005; Villa, 2003). We believe that these estimates represent the best-case scenario of the accuracy of sex-offender address registration data.11 Further research would be needed to provide policymakers with a more precise estimate of the validity of sex-offender address registration. Before the interpretation of our findings is complete, we must point out the following potential limitation to the research: Interviewers were instructed to record that an individual did not live at the registered address if that information was provided by any person answering the door, and this was often the case. It is possible that in some instances, it was the sex offender who denied that the offender lived at the address, when in fact he or she did. The reasons for such misrepresentations are many, including fear of verbal or physical harassment by unknown persons. For reasons of confidentiality, interviewers did not give detailed explanations for wanting to make contact with the person named; they stated simply that they were conducting an ASU survey. We believe that deliberate misinformation was possible, but relatively infrequent. We believe that the formal name tags identifying interviewers as ASU employees and the $20 inducement provided ample incentive for individuals to identify themselves and participate in the study. Furthermore, many of the individuals whom we sought to locate lived in group residential facilities including motels, shelters, and group homes, where staff, not the offender, provided interviewers with the relevant information. Recommendations Based on our findings, we make the following recommendations pertaining to Phoenix’s sex-offender address verification policies and practices. Our purpose is to identify actions and priorities that have the most potential for helping to resolve existing problems and to predict and prevent future problems. 1. First, we recommend that a process map be created by an interagency task force to delineate clear lines of responsibility for sex-offender registration, community notification, and address verification. Our study has shown that few stakeholders are confident that they know their respective roles and responsibilities in this area, largely a consequence of organizational and system fragmentation. Members of the task force would represent the Phoenix Police Department, the Maricopa County Sheriff’s Office, Maricopa County Probation, the Arizona Department of Public Safety, and the Arizona Department of Corrections. Other stakeholder groups should participate, as appropriate. At present, we understand that each of the above agencies has unique policies and procedures, bringing about some duplication of effort and lack of coordination; as a result, misinformation is unintentionally being distributed. A joint task force should be 11 One other strategy could be used for estimating the proportion of sex offenders who lived at their registered addresses, but we believed that it would result in a highly inaccurate estimate. If we were to recode cases that did not result in confirmation of the sex offender living at the registered address after one, two, or three attempted contacts as “living at the registered address,” this would result in an estimate of 45.4% of registered sex offenders not residing at the registered address. Such an estimate would be based on the unlikely assumption that registrants did live at all registered addresses where interviewers made numerous attempts without making contact or getting definitive information from others. 51 52 able to clearly identify, coordinate, and document each agency’s roles and responsibilities. 2. We recommend that for the most serious registered sex offenders, verification policy should be that enforcement officials conduct mandatory, random, in-person interviews with offenders at their registered addresses. More than 25% of offenders are living in temporary housing facilities such as motels and shelters; another 13% live with friends or family. Mail and telephone verifications are inappropriate for offenders who are highly likely to recidivate, where there is a strong potential for deception. Clear, common-sense policies and procedures are needed for conducting such address verifications.12 3. We recommend that the city, county, and state clarify and document performance measures that establish success criteria for registration compliance, as well as the methods by which these are compiled and calculated. Current laws and policies rely too heavily on offender compliance with registration requirements, with little attention to the obligation of the city, county, and state to verify that offenders are actually residing at their registered addresses. Including verification in performance measures should elevate verification as a public safety priority. Aggregate level results of address verification checks, based on these performance measures, should be reviewed annually by the city council. 4. We recommend that Phoenix re-evaluate policies that dictate which offenders are subject to address verification by police officials. Currently, more than 900 level-2 and level-3 sex offenders are subject to address verification every 6 to 12 months, and more than 1,500 level-0 and level-1 offenders are subject to address verification every 18 to 24 months. The department’s current resources are inadequate for this task. If current trends continue, the situation can only get worse. Short of a new and substantial investment of resources by the city, the department should be required only to perform in-person address verifications for the subset of offenders most likely to re-offend. This would require that policymakers revisit offender classification guidelines to ensure that they accurately reflect the dangers posed to society, and allocate and distribute resources for monitoring accordingly. 5. We recommend that the Sex Crimes Unit enhance its recordkeeping system to maintain complete address verification data. The system should automatically generate names of individuals who require a sex-offender address registration check. Related, we recommend that the department track prior addresses in its intelligence system, both to monitor sex-offender residential movement patterns and to assist in criminal investigations. 6. We recommend that the Phoenix City Council allocate financial resources for the Sex Crimes Unit to hire needed staff. At present, sex-offender registration 12 University Internal Review Board policies did not permit analysis that related to the registered addresses of individual offenders and their actual places of residence, but our impression was that offenders registered as living in temporary housing facilities were much less likely to be found actually residing at the registered address. 52 53 and address verification requirements are analogous to an unfunded mandate for the department’s Sex Crimes Unit, which cannot reliably deliver rigorous offender address verifications in the field. New staff is needed to conduct random interviews with the most serious registered sex offenders at the physical addresses at which they are registered, and to follow up when offenders are not located at those addresses. Civilians should be hired to conduct these activities; the city council should work with the police department to determine the appropriate staffing level.13 7. We recommend clarifying policies and assignment of responsibility for sex-offender address verification, to assure that there are no unintended gaps in coverage. Among the issues in need of resolution are whether the probation department or local police are responsible for probationary offenders, and what the role of the Department of Public Safety should be. Agencies are cooperating now, but as the monitored sexoffender population grows, clearer lines of responsibility will assure that no offenders fall through the cracks. Also, the City of Phoenix would benefit from a coordinated system for documenting address verifications. All stakeholder agencies (the Department of Public Safety, local police, county probation, sheriff’s office) could use this resource to document each verification date, the method used, and the agency and official that made the verification. An effective approach would be to adopt a shared data management system (using a secured Internet connection) for state, county, and local agencies to use for collecting, maintaining, and disseminating such information. State-wide systems such as Gangnet are already in place and could provide the needed infrastructure. 13 Sworn officers are often preferred for such work. However for this study, teams of students made contact with these individuals. The strategy was successful. Most interviewers reported feeling safe, as proper training and appropriate precautions were administered. 53 54 List of References Center for Sex Offender Management (1999). Sex offender registration policy overview and comprehensive practices. Silver Springs: Center for Sex Offender Management. Phoenix Police Department (2004). Sex offender notification squad standard operating procedures (Policy Number C.06). Phoenix: Phoenix Police Department, Family Investigation Bureau, Sex Crimes Unit. Plotnikoff, J. & Woolfson, R. (2000). Where are they now? An evaluation of sex offender registration in England and Wales. London: Home Office. Rubiano, S. (2005, July 6). Arizona moves to track down 957 missing sexual predators. Arizona Republic. Rudin, J. (1996). Megan’s Law: Can it stop sexual predators—and at what cost to Constitutional rights? Criminal Justice, 11 (3) 2-10, 60-63. Tewksbury, R. (2002, June). Validity and utility of the Kentucky Sex Offender Registry. Federal Probation, 66, 1: 21-27. Tewksbury, R. (2006). Sex offender registries as a tool for public safety: Views from registered offenders. Western Criminology Review 7, 1: 1-12. Villa, J. (2003, February 7). Local section. Arizona Republic. B10. 54 3 An Examination of Sex-offender Perceptions, Experiences, and Opinions Charles M. Katz, Ph.D. Over the past decade, federal and state legislatures have enacted a number of laws designed to address the nation’s sex-crime problem.14 Much of this legislation has focused on the establishment of sex-offender registration and notification. Researchers and policymakers have spent substantial time and effort examining such issues as the socio-demographic characteristics of sex offenders, sex-offender recidivism rates, and the extent to which communities have enacted sex-offender laws. They have also conducted limited examinations of the impact of sex-offender registration and notification on key stakeholder groups such as criminal justice practitioners (e.g., police, probation) and citizens. To date, however, very little research has examined sex-offender registration and notification from the perspective of offenders. Prior research on drug dealing, burglary, robbery, and gun crime has suggested that active offenders (not those in prison, jail, or treatment faculties) have quite different perceptions than the police and general public of offender motives and behavior. Offenders also have been found to have substantially different perceptions about formal and informal socialcontrol mechanisms designed to control their behavior when compared with other stakeholder groups (Decker, 2005). This body of research has repeatedly shown that information obtained from offenders can be extremely useful when designing crimecontrol policy. For example, gang members were interviewed as part of Boston’s Operation Ceasefire; the information obtained from them resulted in specific tactical and strategic policy recommendations that became the framework for the award-winning project. To our knowledge, only two prior studies have collected data directly from sex offenders living in the community for the purpose of examining their perceptions and experiences with sex-offender registration and notification. Tewksbury (2005) used mail surveys to collect data from offenders listed in the Kentucky Sex Offender Registry. He mailed surveys to795 randomly selected registered sex offenders, 390 who lived in urban areas and 405 who lived in non-urban areas; 121 completed surveys were returned, for a response rate of 15.4%. His analysis revealed that sex-offender registration and notification had resulted in a high number of sex offenders losing friends, jobs, and a place to live. The author also found that sex offenders reported substantial shame and social isolation as a consequence of sex-offender registration and notification. 14 For a review of these laws and statutes, see the introduction of this report. 55 Levenson and Cotter (2005) examined the effects of Megan’s Law on sex-offender reintegration into their communities. The authors relied on a non-random sample of 183 outpatient sex offenders who were receiving comprehensive, long-term counseling in Fort Lauderdale and Tampa, Florida. Levenson and Cotter reported a response rate of 85%. Their analysis indicated that job loss and threats and harassment were relatively common problems experienced by sex offenders, and that the social stigma attached to registration and notification resulted in feelings of isolation, embarrassment, hopelessness, and fear of victimization. These studies raise serious concerns about the impact of registration and notification on sex offenders. The policies are enacted for several reasons. Among other things, they are intended through the use of shaming and self-regulation to decrease sex-offender recidivism, and to increase citizens’ self-protective behaviors by making them more aware of sex offenders and their proximity to at-risk populations. However, the research suggests that such public policy has collateral consequences. Sex-offender notification may have such strong psychological and social effects that offenders will opt not to provide criminal justice agencies with accurate information on their living arrangements, or even more problematic, may not register at all (Tewksbury, 2005; 2006). If sex-offender registration and community notification result in released offenders being blocked from achieving basic needs such as finding a home and getting a job, as the above body of research suggests, it might also inadvertently increase the risk of re-offending related to “a maladaptive coping response to environmental stressors and unmet …needs” (Levenson & Cotter, 2005: 62). Present Study This section examines the impact of sex-offender registration and community notification on released, convicted sex offenders currently residing in Phoenix, Arizona. For this study, we examined sex offenders’ perceptions of and experiences with obtaining housing, victimization, and social and emotional consequences as these were related to sex-offender registration and notification. Methods The methods used here are unique, since this is one of only three studies of its kind. (For exceptions, see Sack & Mason, 1980; Tewksbury, 2005.) Prior research examining sex offenders typically has relied on official data (Adkins et al., 2000; Szymkowiak & Fraser, 2002; Tewksbury, 2002), or data collected from treatment populations (Levenson & Cotter, 2005). This study relied on interview data collected directly from 100 sex offenders registered in Phoenix, Arizona. The Phoenix Police Department (PPD) provided Arizona State University with a database of all level-2 and level-3 sex offenders registered in the City of Phoenix (N=932). The data included the offenders’ name, address, ethnicity, age, date of last address verification, offender class, and probation status. Immediately upon receiving the database, using SPSS’s random-selection function, we selected 300 sex offenders as the sample from which we needed to collect 100 completed interviews. Information for 56 locating those offenders was given to ASU-trained field interview teams. When the first list of 300 was exhausted without reaching the interview goal, we continued randomly selecting individuals until 100 individuals had been successfully interviewed. A total 523 sex offenders were selected for interviewing; we made actual contact with 145, and completed 100 interviews. Of the 523 offenders selected, 206 no longer lived at the registered address and 30 were registered at nonexistent addresses. In 139 cases, we were unable to determine whether or not the sex offender lived at the registered address. Of the 145 registered sex offenders with whom we made contact, 100 volunteered to participate in the study. (See Table 3.1.) Table 3.1 Frequency and% of Results of Attempted Contacts with Offenders Freq Contact made, offender volunteered to participate 100 Contact made, offender declined to participate 45 Offender did not live at registered address 206 Address did not exist 30 Contact attempted 3 times, no response 63 Project ended; 1 or 2 contacts attempted, no response 76 Total 520 Percent 19.2 8.7 39.6 5.8 12.1 14.6 100.0 Using a structured interview instrument, we collected data during one-on-one interviews. Interview teams contacted the registrants in-person at their residences, where they were informed about the nature and purpose of the study and told that participation was voluntary and anonymous. In addition, interviewers told eligible participants that they would receive a $20 incentive if they agreed to participate. Interviews were usually conducted at the registrant’s residence and typically lasted 60 minutes. No names or other identifying data were recorded on the interview instrument. For a more thorough discussion of our research methods, see the section titled “An Assessment of the Accuracy of Sex-Offender Address Registration in Phoenix, Arizona.” Description of Sample We collected descriptive data about the registrants in the study sample (see Table 3.2). Ninety-seven percent were males. More than half were Caucasian (60%); 14% were Hispanic, 13% were African American, 4% were Native American, 1% were Asian or Pacific Islander, and 8% identified themselves as Other. Twelve percent of the registered offenders in the sample owned their homes; 49% rented, 13% lived with relatives or friends, 13% lived in a hotel or motel, 7% lived in a dorm, military, or transitional housing facility, 4% lived in a hospital or 57 care facility, and 2% lived in a shelter.15 Eighteen percent of the subjects had not graduated from high school; 25% were high school graduates or had a GED; 31% had some college experience; 20% had an associate’s, vocational, or technical degree; and 6% had earned a bachelor’s degree or more. Most registrants in the sample were single, never married (40%), followed by those who were divorced (38%), married (20%), and widowed (2%). Almost half (48%) lived in a household with annual earned income of less than $20,000; 32% lived in a household with annual income between $20,000 and $39,900; and 20% lived in a household with annual income of $40,000 or more. Forty-eight percent of those interviewed reported being level-2 offenders; 36% reported being level-3 offenders; 3% reported being level-1 offenders. Thirteen percent reported not knowing the level at which they were classified. Almost three quarters of the offenders in the sample were on probation, and about 16% claimed to have committed a sex crime for which they had not been arrested. On average, registered offenders claimed committing 1.1 sex crimes for which they had not been caught. Related, they indicated that, on average, they had been arrested for a sex crime 1.5 times and for a non-sex crime 4.1 times. (See Table 3.2.) 15 Officials with the Phoenix Police Department noted that our sample under-represented Level 2 and Level 3 offenders living in transitional housing (e.g., hotel, motel, or other transitional housing facility). 58 Table 3.2 Descriptive Data for Sample (n=100) Gender (%) Male Female 97.0 3.0 Ethnicity (%) White or Caucasian Black or African American Hispanic or Latino (a) Asian or Pacific Islander Native American Other or Multiple 60.0 13.0 14.0 1.0 4.0 8.0 Residence (past 30 days) (%) Own home Rent home Live with friends or relatives Hospital or other care facility Hotel or motel Dorm, military, other transitional housing facility Shelter 12.0 49.0 13.0 4.0 13.0 7.0 2.0 Education (%) < high school graduate High school diploma or GED Some college Vocational, technical, or associate degree Bachelor degree or above 18.0 25.0 31.0 20.0 6.0 Marital Status (%) Married Single, never married Divorced or separated Widowed 20.0 40.0 38.0 2.0 Est. Annual Household Income (%) < $20,000 $20,000 to $39,900 $40,000 or more 48.0 32.0 20.0 Classification Level (%) Level 1 Level 2 Level 3 Don't know On probation (%) Committed a sex crime, not arrested (%) 3.0 48.0 36.0 13.0 71.0 16.2 Mean S.D. Sex crimes committed, not caught (n) 1.1 3.5 Times arrested (n) Sex crime Non-sex crime 1.5 4.1 2.5 8.2 59 Sex-offender Perceptions: Purpose of Mandatory Sex-offender Registration During our interviews, we asked respondents about their perceptions of the reason for mandatory sex-offender registration (see Table 3.3). More than three quarters of them agreed that the purpose of mandatory registration was to increase the likelihood of quickly arresting a sex offender who re-offends. About two thirds agreed that mandatory sex-offender registration was intended to reduce the chance that sex offenders would recommit the same type of crime and to make neighborhoods safer by informing residents that convicted sex offenders are living nearby. Only about 49% of respondents perceived that sex-offender registration was intended for further punishing convicted sex offenders. Table 3.3 Percent of Respondents Agreeing or Disagreeing With Purpose of Mandatory Sex-offender Registration (n=100) Neither Disagree Agree or agree or strongly nor strongly agree disagree disagree Purpose To further punish convicted sex offenders 48.5 3.0 48.5 To reduce the chance that convicted sex offenders will 63.6 1.0 35.4 commit same type of crime again To increase the chance that sex offenders re-committing 77.6 3.1 19.4 same type of crime will be quickly arrested To make neighborhoods safer by informing residents of 66.7 7.1 26.3 convicted sex offenders living nearby Note. Due to missing data and refusals to answer question, percentages total < 100%. Knowledge of Friends, Family, and Co-workers about Sex Conviction We asked respondents about the proportion of their family, friends, and co-workers who knew that they had been convicted of a sex crime (see Fig. 3.1). Almost 50% stated that nearly all (> 90%) of their friends, family, and co-workers knew about their convictions. Nearly 14% stated that most (60-90%) friends, family members, and co-workers knew, and about 13.1% stated that many of the persons with whom they were close (40-60%) knew. Only 3% of the offenders stated that no one among their friends, families, or co-workers knew about their convictions; 7.1% stated that few (<10%) of their friends, family, or coworkers knew, and 14.1% stated that just some (10-40%) of their friends, family, or coworkers knew about their convictions. 60 Figure 3.1 Proportion of Family, Friends, and Co-workers Who Knew About Sex-Crime Conviction (N=99) 60 48.5 50 Percent 40 30 20 14.1 10 13.1 14.1 7.1 3.0 0.0 0 No One ( 0%) A Few (<10%) Some People (10-40%) A Lot of People (4060%) Most (6090%) Almost All (90%+) Everyone (100%) Frequency of Being Recognized in Public as a Sex Offender Most respondents indicated that in public, they had never been recognized as a sex offender. Specifically, 70.5% stated that they had never been recognized as a sex offender in public; 19% stated that it happened, on average, only once a year; and 5.3% stated that it happened monthly. Approximately 3.2% of respondents stated that they were recognized as a sex offender in public weekly, and 2.1% stated that this occurred daily. (See Fig. 3.2.) 61 Figure 3.2 Frequency of Being Recognized as a Sex Offender in Public (N=99) 80 70.5 70 60 Percent 50 40 30 19.0 20 10 5.3 3.2 2.1 Monthly Weekly Daily 0 Never Annually Frequency of being recognized as a sex offender in public General Perceptions of Sex-offender Registration The vast majority (87.9%) of respondents indicated that they understood why people want a sexoffender registry, and 60.6% agreed that sex-offender registration is a good thing (see Table 3.3). The majority (71.7%) reported feeling ashamed of being a registered sex offender, although half stated that sex-offender registration punished them unfairly. More than 57% of respondents believed that having their photograph posted online at the sex-offender registry Web site was going too far, and 36.2% agreed that people who knew they were on the registry avoided being around them or talking with them. 62 Table 3.3 Percent of Respondents Agreeing or Disagreeing With Statements Representing Potential Perceptions About the Consequences of Sex-offender Registration (N=99) Disagree Agree or Neither agree or strongly strongly agree nor disagree disagree I feel ashamed that I am a registered sex 71.7 4.0 24.2 offender. I understand why people want a sex offender 87.9 3.0 9.1 registry. People avoid being around me or talking with 36.2 11.7 52.1 me when they know that I am on the sexoffender registry. The sex-offender registry is a good thing. 60.6 11.1 28.3 I am being unfairly punished by being on the 50.0 9.2 40.8 sex-offender registry. Having my photograph posted on the online 57.1 12.2 30.6 sex-offender registry is going too far. Note. percentages may not total to 100%, due to missing data or refusals. Positive Effects of Sex-offender Community Notification Our analysis showed that community notification could have positive impacts on registered sex offenders (see Table 3.4). Three quarters of the respondents agreed that most of the people who knew that they were registered offenders were supportive of their recovery. More than two thirds of the respondents (68%) said that they were more motivated to not re-offend in order to prove to others that they were good people. Fiftyfive percent agreed that Arizona's community-notification law had helped them be more honest with people about having committed a sex crime. Only 37% indicated that they were more motivated not to commit another sex crime because their neighbors were being vigilant, however, and only 37% believed that registration and notification were helping to prevent them from re-offending. Likewise, only 27% agreed that their neighbors knowing that they were registered sex-offenders resulted in having less access to potential victims because people would keep their children (or other potential victims) away. 63 Table 3.4 Percent of Respondents Agreeing or Disagreeing With Statements about Positive Impacts of Community Notification (N=100) Agree or strongly agree Neither agree nor disagree Disagree or strongly disagree I am more motivated not to commit another sex crime because I know my neighbors are watching me. 37.0 13.0 49.0 I am more motivated to not re-offend so that I can prove to others that I am a good person. 68.0 11.0 20.0 I think that registration and notification help prevent me from reoffending. 37.0 8.0 54.0 Because my neighbors know that I am a registered sex offender, I have less access to potential victims because people keep their children (or other potential victims) away from me. Arizona's community notification law has helped me to be more honest with people about having committed a sex crime in the past. 27.0 16.0 55.0 55.0 10.0 34.0 I find that most people who know that I am a registered sex offender are supportive of my recovery. 75.0 6.0 19.0 Negative Experiences Resulting From Sex-offender Registration According to the respondents, being on a sex-offender registry could also have negative social and professional consequences (see Table 3.5). Almost two thirds indicated that they had lost or been denied a place to live as a consequence of being registered. Fifty-five percent had been denied a job for this reason; 17% reported being denied a promotion, and 42% had lost a friend. Additionally, 28% reported that people in their neighborhoods would not have contact with them as a result of their registration. Twenty-five percent had been threatened or harassed in person by neighbors, and 20% had been treated rudely in public. Just under one fifth of the respondents reported that a person living with them had been threatened, harassed, assaulted, injured, or suffered property damage as a result of the respondent being a registered sex offender. Fourteen percent had themselves been attacked by someone for this reason; 12% had moved out of their homes because neighbors had complained, and 11% had experienced property damage because someone had learned of the conviction. Eight percent of the respondents had been asked to leave a restaurant, 8% had received threatening phone calls, and 4% had received threatening mail or notes related to being a registered sex offender. 64 Table 3.5 Percent of Respondents Reporting Negative Experience(s) Due to Registration (N=100) Lost or been denied a place to live 63.0 Lost or been denied a job 55.0 Lost a friend 42.0 People in the neighborhood would not have contact 28.0 Threatened or harassed in person by neighbors 25.0 Treated rudely in a public place 20.0 Denied a promotion 17.0 Person living with respondent was threatened, harassed, assaulted, injured, or 17.0 suffered property damage Assaulted/attacked by someone who found out about the conviction 14.0 Moved out of home or apartment because neighbors complained 12.0 Property was damaged by someone who found out about conviction 11.0 Asked to leave a business or restaurant 8.0 Received harassing or threatening telephone calls 8.0 Received harassing or threatening mail, flyers, or notes 4.0 Negative Impact of Sex-offender Community Notification Registered sex offenders in Phoenix believed that community notification had a significant negative impact on their lives. More than two thirds of the respondents indicated that Arizona’s community-notification law had interfered with their recovery efforts by causing added stress. Likewise, more than half agreed that Arizona’s community-notification law made them feel alone, isolated, hopeless, and fearful for their safety. More than 50% agreed that they had less hope for the future due to being a registered sex offender. Forty-three percent agreed that the community-notification law sometimes made them feel that that no one believed they could change, “so why even try?” (See Table 3.6.) 65 Table 3.6 Percent of Respondents Agreeing or Disagreeing With Statements About Negative Impacts of Community Notification Life (N=100) Neither Disagree Agree or agree or strongly nor strongly agree disagree disagree Arizona's community notification law interferes with my 68.0 11.0 21.0 recovery by causing more stress in my life. 56.0 5.0 39.0 I feel alone and isolated because of Arizona's community notification law. 40.0 4.0 56.0 I have lost friends or a close relationship because of Arizona's community notification law. 53.0 6.0 41.0 I am afraid for my safety because of Arizona's community notification law. 53.0 3.0 44.0 I have less hope for the future now that I am a registered sex offender. Sometimes Arizona's community notification law makes me 43.0 7.0 50.0 feel hopeless - "No one believes I can change, so why even try?" Sex Offenders’ Experiences with Victimization About 30% of the respondents reported being verbally harassed since becoming a registered sex offender (see Table 3.7). On average, respondents stated that they had been verbally harassed about four times in the neighborhood and three times outside the neighborhood since becoming registered. Sex offenders who were verbally harassed very rarely (mean=.07) notified the police of the incident. Of particular interest was the finding that of 7.41 mean incidents of verbal harassment, 7.22 were believed to have occurred specifically because the respondent was registered as a sex offender. Just more than 18% of the respondents reported having been threatened with physical harm. On average, respondents stated that they had been threatened 1.17 times in the neighborhood and 1.15 times outside the neighborhood; they very rarely notified the police of these threats. Respondents indicated that most of the threats occurred specifically because they were registered sex offenders; on average, they indicated, of the 2.32 times they were threatened with physical harm, 2.02 of the incidents occurred because they were registered as sex offenders. About 10% of the respondents stated that they had been physically assaulted after becoming registered sex offenders. On average, respondents were physically assaulted in their neighborhoods 0.2 times and reported the assaults 0.08 times. Likewise, on average they were physically assaulted 0.18 times outside their neighborhoods and notified the police 0.02 times. Respondents stated that roughly 70% of the physical assaults resulted from being registered as sex offenders. 66 Only 2% of the respondents indicated that they had been sexually assaulted since becoming a registered sex offender. Sex offenders, on average, had experienced a sexual assault in their neighborhoods 0.02 times since becoming a registered sex offender; they notified the police the same number of times. Related, sex offenders on average were sexually assaulted 0.09 times outside their neighborhoods and notified the police 0.01 times. Of those sexually assaulted, about half stated that the assault occurred because they were registered sex offenders. We asked those interviewed whether they had been robbed since registering as a sex offender; roughly 5% stated that they had been. Respondents indicated, on average, being robbed 0.19 times of which, they stated that 0.16 times occurred because of sex-offender registration. Consistent with their responses to the other kinds of incidents above, the respondents indicated that they very rarely notified the police after being robbed. Last, 18.2% of respondents revealed that they had had property stolen since being registered. On average, they had had property stolen 0.62 times of which, they believed, 0.42 times had occurred because of their sex-offender registration. When the property had been stolen in the neighborhood, they had reported the incident to the police about 20% of the time; when it had been stolen outside the neighborhood, none of the respondents had notified the police. Table 3.7 Experiences with Victimization as a Registered Sex Offender (N=99) Verbally Harassed Percent Ever Prevalence Threatened with Physical Harm 30.3 M Physically Assaulted 18.2 SD M Sexually Assaulted 10.1 SD Robbed Had Property Stolen 5.1 18.2 2.0 M SD M SD M SD M SD In neighbor-hood 4.31 26.70 1.17 8.08 0.20 0.84 0.02 0.20 0.13 0.71 0.51 1.57 Notified police 0.07 0.38 0.04 0.20 0.08 0.53 0.02 0.20 0.01 0.10 0.08 0.31 Out of neighborhood 3.10 26.01 1.15 10.01 0.18 0.95 0.09 0.90 0.06 0.60 0.11 0.65 Notified police 0.00 0.00 0.00 0.00 0.02 0.20 0.01 0.10 0.00 0.00 0.00 0.00 Due to sexoffender registration 7.22 52.31 2.02 11.36 0.27 1.36 0.05 0.36 0.16 1.22 0.42 1.68 Note. M=Mean; SD=Standard deviation. Impact of Sex-offender Registration on Living Arrangements Our findings indicated that nearly half of all respondents’ living arrangements had been adversely affected by being a registered sex offender, and more than half (57%) had had difficulty finding an affordable residence that was not too close, according to regulations, to a school, park, bus stop, or playground (see Table 3.8). Eight percent of the respondents reported having had to move out of homes that they owned because they were too close to a school, bus stop, park, or 67 playground; 20% reported having had to move from “any other residence” for this reason. Approximately one third of the respondents reported having been unable to live with supportive family members or, after being released from prison, having been unable to return to their previous residence for that reason. More than 40% of the respondents stated that they had found it difficult to find an affordable place to live located the required distance from the residences of other registered sex offenders. Table 3.8 Percent of Respondents Indicating that Sex-offender Registration Requirements Adversely Affected Living Arrangements (n=100) Had to move out of a home that offender owned because it was too close to a school, bus stop, park, or playground Had to move out of other residence because it was too close to a school, bus stop, park, or playground After release, unable to return to prior residence because it was too close to a school, bus stop, park, or playground Unable to live with supportive family members because residence was too close to a school, bus stop, park, or playground Difficulty finding affordable place to live that was not too close to a school, bus stop, park, or playground Difficulty finding affordable place to live that was not in close proximity to other registered sex offenders Sex-offender Preferences: Residential Clustering vs. Dispersion We asked respondents their opinions on whether residential clustering or dispersion is the better residential location strategy for registered offenders. Although the majority (59%) stated that it was better for registered sex offenders to be disbursed throughout the city, 30% believed the opposite: that it was better if offenders were concentrated in one area. Eleven percent could not decide or did not respond. (See Fig. 3.3.) 68 Percent 8.0 20.0 32.0 38.0 57.0 41.0 Figure 3.3 Percent of Respondents Favoring Residential Clustering or Dispersion (N=100) 70 59 60 Percent 50 40 30 30 20 11 10 0 Living in one area of town Spread across the city No Response / Could not choose Sex-offender Opinions on Possible Laws Restricting Living Arrangements We asked respondents for their opinions about various laws that, if passed, would restrict the living arrangements of registered sex offenders (see Table 3.9). About two thirds of the respondents disagreed that a law should be passed to regulate the number of sex offenders who can live in close proximity. Three quarters of the respondents believed that registered sex offenders should be able to live in the same residence or complex as other registered sex offenders; however, many agreed that offenders should be able to do so only if a judge (41%) or parole or probation officer (57%) authorizes the arrangement. Although roughly 40% of the respondents believed that sex offenders should be required to move from their homes if located too close to a school, day-care center, park, or playground, only 13.1% believed that offenders should be required to move if they live too close to a bus stop. Likewise, 83% agreed that the city should regulate sex offenders based on the severity of the offense, but only 42% agreed that the city should regulate the distance that sex offenders must live from schools, day-care centers, and parks. 69 Table 3.9 Percent of Respondents Agreeing or Disagreeing With Regulation of Offenders’ Living Arrangements (N=100) Agree or strongly agree Neither agree nor disagree Disagree or strongly disagree There should be a law regulating how many registered sex offenders can live in a particular neighborhood/area. Registered sex offenders should be allowed to reside in the same house, apartment complex, or mobile home park as other registered offenders. Registered sex offenders should be allowed to reside in the same house, apartment complex, or mobile home park as other registered sex offenders only if a judge authorizes them to do so. Registered sex offenders should be allowed to reside in the same house, apartment complex, or mobile home park as other registered sex offenders only if their probation or parole officer authorizes them to do so. Registered sex offenders should be required to move out of their homes or apartments when located too close to a school or day-care facility. Registered sex offenders should be required to move out of their home or apartment when located too close to a park or playground. Registered sex offenders should be required to move out of their homes or apartments when located too close to a bus stop. The city should regulate the distance that registered sex offenders must live from schools, day-care facilities, and parks. 33.7 2.0 64.3 75.0 7.0 18.0 41.4 3.0 55.6 57.0 6.0 37.0 42.7 8.3 49.0 37.8 12.2 50.0 13.1 9.1 77.8 42.9 6.1 51.0 The city should regulate sex offenders based on the severity of their crimes. 83.0 2.0 15.0 Note. Responses of “don’t know” or “no opinion” are not represented, so percentages may not total 100%. Perceived Effectiveness of Various Sex-crime Prevention Strategies About 84% of respondents believed that treatment for sex offenders after release from prison would be the most effective strategy for preventing sex crimes, followed by supervised probation (72.6%) and community notification (61.6%). Roughly 56% of respondents believed that mandatory sex-offender registration and laws to regulate sexoffender proximity to schools, day-care centers, and parks would be effective strategies for preventing sex crimes. Only 35.4% of respondents believed that laws limiting the number of registered sex offenders living together in one area would be effective. (See Table 3.10.) 70 Table 3.10 Percent of Respondents Perceiving Sex-Crime Prevention Strategies to be Effective or Very Effective (N=100) Mandatory registration upon release from prison 55.6 Community notification of registered sex offenders living in area 61.6 Laws to limit the number of registered sex offenders living in one area 35.4 Laws to regulate the distance registered sex offenders live from schools, day56.1 care facilities, and parks Supervised probation upon release from prison 72.6 Treatment or therapy upon release from prison 83.7 Perceived Importance of Various Responses to Combating Sexual Assaults We asked registered sex offenders about their perceptions of the importance of various responses for combating the sexual assault crimes (see Fig. 3.4). They believed that treatment and investigation were the most important responses, followed by sex-offender monitoring and prosecution. Use of prisons and community notification were perceived by the offenders to be the least important responses to combating sexual assault. Figure 3.4 Perceived Importance of Responses to Combating Sexual-assault Crimes (N=99) 6 5 Mean 4 3 2 1 0 Mean Treatment Investigation Monitoring Prosecution Prisons Community Notification 4.94 4.61 3.36 3.35 2.45 2.38 0= Least important response; 6= Most important response 71 Respondents’ Perceptions of the Fairness of Information-sharing as an Element of Sex-offender Community Notification We asked respondents about the fairness of giving neighbors access to the personal information that is commonly made public through sex-offender community notification programs (see Table 3.11). More than 50% of registered sex offenders believed that it was fair for their neighbors to know an offender’s physical appearance; 46% believed that it was fair for their neighbors to know the crime for which they had been convicted. Only a minority of registered offenders believed that it was fair for their neighbors to know an offender’s home address (31%), vehicle description (27%), license plate number (22%), employer’s name and work address (15%), or home telephone number (7%). Table 3.11 Percent of Respondents Believing it is Fair or Unfair for Neighbors to be Advised of Registered Sex Offenders’ Personal Information (N=100) Neither fair Unfair Fair nor unfair Physical appearance (descriptions, photographs) 39 8 53 Home address 64 5 31 Home telephone number 88 5 7 Employer’s name and work address 79 6 15 Crimes for which convicted 46 8 46 Vehicle description 65 8 27 License plate number 71 7 22 Arizona’s Sex-offender Internet Site: Offenders’ Experiences and Perceptions In Arizona, information about registered sex offenders is posted on an Internet site. We asked respondents about the accuracy of that information, in their experience, and what information they would act to correct if it were inaccurate. (See Table 3.12.) More than half (56.3%) of the respondents reported that their registered information was mostly or completely accurate. About 21% said that about half of the information posted about them was accurate, and nearly 17% stated that the information posted about them was completely or mostly inaccurate. Most respondents told interviewers that if they discovered inaccurate information about themselves posted on the sex-offender registry Internet site, they would take action to correct that information. Almost 83% of the offenders indicated that if their address was incorrectly listed on the site, they would contact someone. Ninety-seven percent stated that if they moved, they would contact someone. Likewise, almost 90% said that if their offenses were incorrectly listed, they would report this. Roughly 76% of the offenders said that if they found that the photograph on their registry page was of someone else, they would call to have it corrected. 72 Table 3.12 Arizona’s Sex-offender Internet Site: Offenders’ Perceptions of Information Accuracy and Likelihood of Taking Action to Correct Inaccurate Information Accuracy of posted information (N=48) Percent Completely inaccurate 10.4 Mostly inaccurate 6.3 Half inaccurate & half accurate 20.8 Mostly accurate 29.2 Completely accurate 27.1 Don't know 4.2 Declined to answer 2.1 Percent who would notify someone to correct errors (N=99) If you found out that the address listed for you on the sex offender registry was not 82.8 correct, would you contact someone to have it corrected? If you found out that your picture on the sex offender registry was of someone else, 75.8 would you contact someone to have it corrected? If you found out that the offense(s) for which you were convicted were incorrect 89.5 (listed as less or more serious) on the sex offender registry, would you contact someone to have it corrected? If you move or change addresses, would you contact someone to update your 97.0 information on the sex offender registry? Contact with Police and Probation Officers We asked respondents about the frequency and type of their contacts with police and probation officers (see Table 3.13). About 25% of them reported never being contacted by police or probation/parole officers; 20.2% reported being contacted once or twice a year; 36.4% reported being contacted once or twice a month; and 18.2% indicated that they were contacted once a week or more. We also asked about the frequency of in-person contacts with probation or parole officers. Fourteen percent reported never having had in-person contact; 5.1% reported having had such contacts once or twice a year; 43.4% reported having had such contacts once or twice a month; and more than 37% reported having had in-person contacts once a week or more. Last, we asked respondents for the number of times that they had been contacted by the Arizona Department of Public Safety (DPS) for the purpose of address verification. More than half (55%) of the sex offenders reported never having been contacted for address verification by DPS; 17% reported being contacted for that reason once; 11% reported being contacted twice; and 16% reported having been contacted for address verification by DPS three or more times. 73 Table 3.13 Percent of Respondents Reporting Contact with Law Enforcement, Probation Officers, and DPS (N=99) Contact with police or probation/parole officers One or more contacts a week 18.2 Once or twice a month 36.4 Once or twice a year 20.2 Never 25.3 Face-to-face contact with probation or parole officer One or more contacts a week 37.4 Once or twice a month 43.4 Once or twice a year 5.1 Never 14.1 Contact from Arizona Department of Public Safety for address verification Three or more times 16.0 Twice 11.0 Once 17.0 Never 55.0 Conclusions, Policy Implications, and Recommendations For this study, we examined the impact of registration and notification regulations on sex offenders living in Phoenix, Arizona. We were seeking to understand registered offenders’ experiences with housing and victimization, and the social and emotional impact on them of registration and community notification practices. Of the 145 registered sex offenders we contacted, 100 agreed to participate in our study. Many of the sex offenders in the randomly selected interview sample believed that community notification had had some positive outcomes for them, such as motivating them not to re-offend and helping them to be more honest about the sex offense that they had committed. Most offenders said that they understood why the public wants sex offenders to be registered, and that they believed that sex-offender registration is a good thing. Still, when we asked about particular aspects of sex-offender policy, the respondents described several serious personal consequences that they perceived or had experienced to be a consequence of registration and notification. For example, the majority of sex offenders reported having lost a friend, a job, and/or a place to live because of being a registered sex offender. More than half believed that community notification practices had interfered with their ability to recover, added stress to their lives, and made them feel alone, isolated, and afraid for their personal safety. We found that 30% of sex offenders had been harassed as a consequence of community notification, and that for these offenders, harassment was a fairly frequent event. Ten percent of sex offenders indicated 74 that they had been identified and physically assaulted as a result of community notification. Of particular importance here, many sex offenders reported having had difficulty finding an affordable home located the required distance from other registered offenders. They were generally supportive of regulating sex offenders in accord with the severity of their crimes, but they opposed restrictions on living arrangements including those that dictated the distance that offenders must live from schools, day-care centers, parks, playgrounds, and bus stops. Three quarters of the respondents believed that they should be permitted to live in close proximity to other registered sex offenders, although many agreed that this should be the case only when a probation or parole officer (57%) or judge (41%) authorized them to do so. This finding suggests that sex offenders might be more compliant with policies that allow for guided discretionary decisions on the part of probation and parole officers. City policymakers should consider collaborating with county and state officials to determine whether such an approach is practical and, if so, to develop guidelines for probation and parole officers to use in making such decisions. Our findings in Phoenix were similar to findings from prior research: The released exoffenders faced substantial added psychological, social, and economic hardship as a consequence of sex-offender registration and community notification. Successful re-entry was hampered when residential restrictions severely constrained social ties with family, friends, and others. Sex offenders under current policies found it difficult to find and hold jobs, and to maintain a place to live. Enhanced restrictions, particularly related to housing selection, can be predicted to escalate those difficulties. The extreme social stigma attached to individual sex offenders by way of social policies such as registration and notification heighten the cost to the offender of registration compliance, and upon perceiving or experiencing those consequences, many may fail to register address changes as required. Public policies that further restrict their housing options could well drive even more of them underground, such that police would have still less accurate information about offenders’ locations. Alternatively, policymakers could consider re-integration programs that are intended to alleviate stressors that otherwise can raise the chance of recidivism. Such programs could encourage rather than discourage offenders to register with the police by assisting with finding socially and economically appropriate employment and housing. The evidence suggests that otherwise, even those offenders who register initially will be less likely to update their addresses promptly or to self-report inaccurate information. Without their cooperation, databases may never be much more accurate than they are today. Recall that one third of the sex offenders we interviewed reported that most of their personal information as posted on Arizona’s sex-offender Web site was inaccurate. Also, most offenders believed that it was unfair that their neighbors had access to personal information such as their exact addresses, and that they were at risk of harassment or worse, simply by virtue of being registered. At the least, officials should be proactive in ensuring that sex-offender information is correct – that, as soon as possible, policies and practices should be put into place to assure data quality control and timely correction of inaccuracies. 75 Released, registered sex offenders, like other classes who are subject to victimization, may need social services specific to helping them cope with the consequences of postrelease policies such as registration and community notification. These policies make it more likely that ex-offenders will be subject not merely to shunning, but to illegal acts such as harassment, threats, and even physical attacks. We found that the offenders were unlikely to report such incidents; offering support services in these instances could encourage reporting, prevent the escalation of violence, and prove to be another important step in preventing stress- and despair-related recidivism. It could also reduce sex-offender fear of the public, thereby increasing rates of registration and compliance. For obvious reasons, such measures are not likely to be immediately popular with some policymakers. But they are fair and just, and more, they could solve the problems that result in sex-offender noncompliance and inaccurate registry data. As discussed in a forthcoming section (“Citizen Attitudes About Sex Offenders and Sex-offender Housing Policy in Phoenix”), 30% of surveyed citizens reported a willingness to attend neighborhood meetings to help sex offenders successfully reintegrate into society.16 Policymakers could offer programs that support and complement any such efforts made by local community groups. We also found that many sex offenders reported rarely having had contact with law enforcement or probation and parole officers. More than half of the sex offenders we interviewed stated that they had never been contacted by the Arizona Department of Public Safety (DPS) for address verification; DPS is the agency with the primary responsibility for address verification and for the accuracy of data posted on the state’s sex-offender Web site17. This finding, combined with what we learned about the inaccuracies in sex-offenders’ registered addresses, suggests that many problems exist with the system that is supposed to collect, maintain, and disseminate sex-offender data and intelligence. Policymakers should weigh the benefits of sharing the available information across agencies, thereby improving its completeness and accuracy, against the likely consequences of continuing to disseminate data publicly that is largely inaccurate. In the previous section, we estimated that at the time of our study, between 61.9 and 70.8% of Phoenix registered sex offenders were not residing at their registered addresses. In this section, we have reported that of those who did live at their registered locations, one third reported that most of their personal information posted on Arizona’s sexoffender Web site was inaccurate. Taken together, these two findings suggest that the vast majority of data and intelligence collected on sex offenders may be wrong, in one form or another. The recommendations above are intended to suggest some highly effective ways to turn the situation around. 16 17 See citizen survey section of this report. Kostelac, C. (2006). Personal communication. November 30, 2006. 76 List of References Adkins, G., Huff, D., & Stageberg, P. (2000). The Iowa sex offender registry and recidivism. Des Moines: Iowa Department of Human Rights. Decker, S. (2005). Using offender interviews to inform police problem solving. Washington DC: Community Oriented Policing Services. Levenson, J., & Cotter, L. (2005). The effect of Megan’s Law on sex offender reintegration. Journal of Contemporary Criminal Justice, 21, 1: 49-66. Sack, W.H., & Mason, R. (1980). Child abuse and conviction of sexual crimes. Law and Human Behavior, 4, 3: 211-215. Szymkowiak, K., & Fraser, T. (2002). Registered sex offenders in Hawaii. Honolulu, HI: Department of the Attorney General. Tewksbury, R. (2002). Validity and utility of the Kentucky sex offender registry. Federal Probation, 66, 1: 21-26. Tewksbury, R. (2005). Collateral consequences of sex-offender registration. Journal of Contemporary Criminal Justice, 21, 1: 67-81. Tewksbury, R. (2006). Sex offender registries as a tool for public safety: Views from registered sex offenders. Western Criminology Review, 7. 1: 1-8. (http:/wcr.sonoma.edu/v07n1/01/sexoffender.pdf). 77 4 Citizen Attitudes About Sex Offenders and Sex-offender Housing Policy in Phoenix Charles M. Katz, Ph.D. Vincent Webb, Ph.D. Citizen surveys are often used to assess which issues are of concern to the public and which the public believes are most worthy of attention. Public opinion has the power to shape legislation and public policy, but it typically is invoked only anecdotally in decision-making concerned with sex offenders. Most often, policymakers become aware of public perceptions of sex offenders and their management following incidents that spark neighborhood rallies and petitions. Rarely have policymakers had access to scientific data about the public’s perceptions of sex offenders and related policies. Understanding public attitudes toward sex offenders and sex-offender housing policies is important for several reasons. First, the general hypothesis is that sex offenders have a negative impact on neighborhood quality of life. This perception has been amplified with the enactment in recent years of community-notification ordinances and laws, requiring officials to notify residents when certain kinds of sex offenders move into their neighborhoods. Policymakers and academics often theorize that sex-offender residential patterns are likely to have a negative impact on community stability, fear of crime, and community morale (Zevitz, 2003). If that theory were correct, it would have substantial policy implications. When residents become dissatisfied with a neighborhood, they move away and others may not wish to move in, and that destabilizes and stigmatizes the area. In turn, this can lead to further neighborhood decay (e.g., increases in social and physical disorder) of the type that has been associated with growing neighborhood criminality (Skogan, 1990). This theory rests on the premises that residents are aware of the presence of sex offenders in their neighborhoods, and that the known presence of a sex offender will result in widespread fear and panic in that neighborhood. In fact, however, little research has examined the real social and psychological impacts of the presence of sex offenders on neighborhood residents. Second, understanding public opinion regarding sex offenders and sex-offender housing patterns helps to establish the boundaries of political permission. The public holds policymakers responsible for their decision-making practices, especially those related to crime. Sound public-opinion data inform policymakers regarding what measures the public will or won’t support (Center, 2000). For example, the public’s fear of sex offenders and its dissatisfaction with the criminal justice system’s handling of repeat offenders can lead to a number of policy reforms and statutory enactments “including sex offender registration, 78 community notification, and involuntary civil commitment and lifetime supervision for some sex offender groups” (Center, 2000). Third, public opinion surveys are useful in determining the public’s awareness and understanding of legislation and policies that focus on sex offenders and their management. The purpose of such surveys is to ensure that existing policies and practices are understood by the citizenry, and that future discussion and policy proposals will reflect a shared understanding between citizens and policymakers. The following section presents scientific data about Phoenix residents’ perceptions of sex offenders and the impact of resident offenders on neighborhood life, as well as residents’ opinions about alternative proposals for managing sex-offender housing. We describe our research methodology and present our findings, and then discuss the results and their policy implications for Phoenix, Arizona. Research Design In December 2005 and January 2006, the interview teams completed 793 interviews with Phoenix residents. We used stratified sampling methods to ensure a sufficient sample of residents from five areas, each with a different number of registered resident sex offenders, to determine the social and psychological impact of sex-offender clustering on those living in the neighborhoods. Sample Selection In August 2005, we obtained the addresses of all active registered sex offenders (N=2,381) in Phoenix, Arizona, using the Phoenix Police Department Sex Offender Database. Next, we excluded level-0 and level-1 sex offenders since community notification laws do not pertain to these classes and neighborhood residents may not be aware of their presence. After those exclusions, 857 level-2 and level-3 offenders remained in the study dataset. We geo-coded those and matched the entries to the nearest one-square-mile grid location.18 Due to missing or incomplete data, eight (< 1%) addresses could not be geo-coded to a corresponding grid. We analyzed the data grid by grid and identified those with large numbers of transient persons residing in shelters, halfway houses, treatment facilities, and hotels or motels. Each grid was labeled in accord with the total number of level-2 and level-3 sex offenders residing within it. Areas with no registered resident sex offenders were labeled “low.” Areas with one to three registered resident sex offenders were labeled “medium.” Those with four or more registered resident sex offenders were labeled “high.”19 18 The one-square-mile grids were created by combining the city’s current grids (roughly 0.25 miles each). About 90% of the grids were actually one square mile in size: Four were significantly larger (more than 1.5 square miles) and 10% were significantly smaller (0.5 square miles or less). 19 Natural cut-off points were used to determine whether an area was labeled as low, medium, or high. For example, most areas did not have any sex offenders residing within them; these areas were labeled as low. Very few areas, with the exception of some hyper-clusters (later determined to be transitional areas) had four or more registered sex offenders living within them; these areas were labeled as high. 79 Grids containing large numbers of residential complexes commonly used by transient populations (e.g., homeless shelters, halfway houses, low-end motels) and the grid containing the Tampico apartments were treated and coded separately; a large number of sex offenders resided in these areas and the Tampico apartments had been the subject of substantial press coverage due to the sex offenders residing there. For sampling, we randomly selected two low, two medium, and two high areas, plus one transitional area and the Tampico apartment area.20 There were an insufficient number of households in the medium and high areas to ensure that 100 residents from each would be contacted; we added a half-mile adjoining grid to one of the randomly selected grids in each area to correct for that problem. Likewise, due to a large amount of nonresidential (commercial-industrial) space, the transitional area selected had too few households to sample, so we combined five transitional areas in order to ensure a sufficient sample size. (See the map on the next page.) We randomly selected household phone numbers from each area. Any household member over the age of 17 was allowed to complete the interview. For ease of interpretation, we pooled the data from areas with similar numbers of sex offenders (low, medium, or high). 20 The city has 570 grids, of which 388 (68.1%) were coded as low, 126 (22.1%) were coded as medium, 49 (8.6%) were coded as high, 6 (1.1%) were coded as transitional, and 1 (0.2%) was coded as Tampico. 80 81 In the targeted areas, we randomly selected and called 1,746 phone numbers. Of these, 611 were terminated because the individual contacted did not qualify for the survey (n=426) or the phone number was disconnected (n=185). Of the remaining 1,135 phone numbers dialed, 102 calls were not completed due to no answer, 4 due to busy signals, and 51 due to answering machines. About 1.8% of those contacted were unable to complete the survey because of a language barrier. Of the remaining 1,088 contacts, 72.8% resulted in a completed survey with a qualified respondent. Our overall response rate was 73.4%, high for a telephone survey, particularly for one focused on such difficult subject matter. Response rates varied across study areas. The survey firm, Precision Research, Inc., collected data on the total number of sample records selected and the number of interviews completed for each area.21 The firm advised us that they dialed each number from 3 to 14 times without a response before classifying it as an uncompleted call. The number of attempts made varied in each case depending on the circumstances. In areas having sufficient households to assure a complete sample, interviewers attempted to make contact only three times. In areas with fewer households, interviewers made more efforts to make a contact at each number. Study Area Characteristics Not surprisingly, the transitional area had the greatest number of resident offenders per square mile: 27. Tampico had 13 offenders per square mile, high areas had 9.2, medium areas had 1.6, and low areas had no sex offenders (see Table 4.1). The high density of offenders in the transitional housing area was attributable to homeless shelters, low-end motels, and transient housing, all places where sex offenders commonly congregated. In addition to sex offenders being unevenly distributed across the study areas, there were also differences in the number of calls for police service and of violent crimes reported to the police, with both being highest in the high and transitional areas. Low areas had the fewest calls for service per 1,000 residents and the fewest number of violent crimes reported to the police. We did not attribute great significance to the apparent relationship between the number of sex offenders residing in an area and the area’s level of criminality; the relationship is most likely spurious. The ecological conditions that resulted in greater numbers of sex offenders living in an area were likely to be the same as the conditions that generated criminality. These trends should not be interpreted to mean that sex offenders residing in an area caused increases in criminality. 21 The results: Low area 1 (188 sample records used and 100 interviews completed); Low area 2 (127 sample records used and 100 interviews completed); Medium area 3 (177 sample records used and 100 interviews completed); Medium area 4 (230 sample records used and 93 interviews completed); High area 5 (238 sample records used and 100 interviews completed); High area 6 (204 sample records used and 100 interviews completed); Transitional area 7 (432 sample records used and 100 interviews completed); Tampico area 8 (150 sample records used and 100 interviews completed). 82 Table 4.1 Characteristics of Study Areas Study areas Characteristics Low Medium High Tampico Trans. Registered sex offenders 0 4 23 13 135 Reg. sex offenders per square mile 0 1.6 9.2 13 27 1-mile grids surveyed 2 2.5 2.5 1 5 4,407 7,081 16,843 2,088 24,385 512.08 518.37 789.45 556.95 801.85 3.14 6.08 9.80 3.73 12.04 Calls for service/year Calls for service/year per 1,000 residents Violent crimes/year per 1,000 residents Note. Data were provided by the Phoenix Police Department’s Crime Analysis & Research Unit. Note. Study areas are defined by the number and population density of resident sex offenders. Note. Calls for service and violent crimes are 2005 data. Numbers reflect dispatched and callback calls only. Analyzing the demographic characteristics of the study areas, we noted that the percentage of residents below poverty level was just above 35% in the transitional area and 25% in the high area. These two areas also had the highest unemployment rates (12.5% and 7.6%, respectively). These areas were much more likely than the others to experience resident mobility and turnover, indicated by the high proportions of respondents identifying themselves as renters, foreign-born, and having moved since 1995. Together, these data suggested that areas with the highest numbers of sex offenders also had the highest levels of crime; in fact, these areas were characterized by some criminogenic characteristics known to be associated with crime (e.g., poverty, low education, high neighborhood mobility). One of our challenges was to distinguish residents’ general crime-related public safety concerns from their concerns specifically about sex offenders. The Tampico area has been a focal point of Phoenix’s sex-offender residential controversy and discussion, and yet it is most similar to the low areas with respect to police calls for service and reported violent crime, and similar to the medium and low areas in unemployment rate and the proportion of population living below poverty level. In other words, Tampico is relatively low with respect to crime and related criminogenic characteristics (see Table 4.2). 83 Table 4.2 Socio-demographic Characteristics of Study Areas Characteristics Total residents (N) Gender (%) Male Female Race/ethnicity (%) Caucasian Hispanic African American Other Housing units (N) Housing status (%) Vacancies Renters Owner-occupied Mobility (pop. 5 yrs. and over) (n) Percent who, in 1995, lived in: A different house The same house Foreign-born (%) Unemployment rate (pop. 16 or more yrs.) Income below poverty rate Education (pop. 25 yrs. and over) (n) Education level (%): < High school graduate High school graduate or equivalent Some college, AA degree Bachelor's degree or higher Study areas High Tampico 21,335 3,749 Low 8,606 Medium 13,660 Trans. 30,411 49.7 50.3 48.5 51.5 51.0 49.0 50.1 49.9 59.9 40.1 88.6 7.1 0.6 3.7 3,322 56.4 35.0 3.2 5.4 4,491 37.8 52.9 3.4 5.9 7,807 87.2 7.3 1.9 3.6 1,740 20.7 60.4 15.9 2.9 6,037 1.7 12.9 85.3 8,109 4.3 10.9 84.8 12,598 5.0 56.9 38.1 19,262 6.7 29.8 63.5 3,562 5.8 42.9 51.3 27,927 37.6 62.5 5.4 2.8 3.9 5,920 52.4 47.6 15.1 5.0 9.0 8,481 60.6 39.4 33.5 7.6 25.3 12,078 42.7 57.3 7.7 2.3 7.1 2,686 43.2 56.8 26.4 12.5 36.0 16,936 6.7 15.2 36.6 41.5 26.8 32.5 30.3 10.5 34.7 24.5 29.5 11.4 7.0 13.1 29.0 50.9 49.0 21.6 21.9 7.6 Survey Respondents We found no significant differences across the areas in gender or age of respondents, with roughly 62% of the sample being female. The average respondent was 51.5 years old. However, our analysis showed significant differences across the areas in other sociodemographic characteristics (see Table 4.3). • Respondents living in low areas and the Tampico area were quite similar, particularly when compared with respondents in the other areas. For example, they were more likely than the others to report being Caucasian, married, and homeowners. They were also more likely to have completed a bachelor’s or graduate degree, and to earn more than $30,000 a year. 84 • Respondents living in medium and high areas were about two to three times more likely than those in low areas and Tampico to report being either Hispanic or African American, and were half as likely to report having a bachelor’s degree. Likewise, they were more than 8 to 10 times more likely than those in low areas and Tampico to report that their families earned less than $30,000 a year. • Respondents in the high areas were most likely to report that they rented their homes, were single, and had never been married. Their families earned substantially less income than residents in the other areas. • Respondents residing in the transitional area reported a significantly lower socio-economic status than the others. The majority reported being Hispanic, African American, or members of another minority ethnic group. They were significantly more likely not to have graduated from high school, and were more likely than the others to report renting their home, having lived in it for fewer than 4 years, and earning less than $30,000 a year. Table 4.3 Characteristics of Survey Respondents Mean age Gender Ethnicity* Home ownership* Education* Marital status* Time lived in study area* Family income* Male Female Caucasian AfricanAmerican Hispanic Other Rent Own < HS graduate HS degree Bachelor’s degree Graduate degree Married Single, never married Divorced/sep arated Widowed < 1 year 1-3 years 4+ years < $30,000 > $30,000 Low Medium N=200 N=193 Study areas High Tampico N=200 N=100 Trans. All N=100 N=793 53.6 41.0 59.0 91.8 52.0 37.8 62.2 77.0 48.8 36.0 64.0 78.6 56.2 37.0 63.0 94.0 47.0 36.0 64.0 49.0 51.5 37.8 62.2 79.7 0.5 2.1 3.6 0.0 16.0 3.6 5.6 2.0 8.6 91.4 14.1 6.8 10.8 89.2 12.0 5.7 29.4 70.6 4.0 2.0 6.1 93.9 32.0 3.0 20.6 79.4 12.5 4.2 15.6 84.4 1.0 6.2 4.5 0.0 10.1 4.2 45.2 68.4 65.2 37.4 63.6 57.3 39.6 20.7 21.7 42.4 19.2 28.2 14.2 4.7 8.6 20.2 7.1 10.3 69.4 53.2 39.9 64.6 49.5 54.9 10.2 15.3 35.9 15.2 20.2 19.8 12.8 15.8 15.2 10.1 19.2 14.6 7.7 4.5 19.0 76.5 6.8 93.2 15.8 4.7 19.2 76.2 25.1 74.9 9.1 4.0 24.0 72.0 33.2 66.8 10.1 2.0 10.0 88.0 3.5 96.5 11.1 2.0 31.0 67.0 26.9 96.5 10.7 3.8 20.7 75.5 20.3 79.7 *Note. p ≤ .05 85 Description of the Survey Instrument Respondents were asked several questions to elicit their opinions about and experiences with sex offenders, and the social and psychological impact of having sex offenders residing in their neighborhoods. We also asked for their opinions and perceptions of ordinances and laws targeting sex offenders.22 Most survey questions required respondents to select their answers from a categorized continuum, or Likert Scale. For example, “Do you 1=strongly agree, 2=slightly agree, 3=neither agree nor disagree, 4=slightly disagree, or 5=strongly disagree with the following statement?” For the sake of exposition, throughout the report many categories (e.g., strongly agree and agree; strongly disagree and disagree) are combined into a single category (e.g., agree or disagree.) Respondents were first asked for their perceptions of the quality of life in their neighborhoods. We asked a series of questions about their fear of crime and sex offenders, and about their perceptions of the risk that they and their children could become victims of a sex crime. The next set of questions focused on Arizona’s sex-offender notification law. Respondents were asked for their level of awareness of the state’s sex-offender notification law and their experiences with it. For those who had been notified that a sex offender resided in the neighborhood, we asked how notification processes had affected their quality of life and whether they had responded by changing their behavior. Related, we asked respondents for their perceptions of the effectiveness of the sex-offender notification law and about other strategies that might be effective in educating residents about sex offenders living in the neighborhood. The last section of the survey asked respondents for their perceptions of various residential requirements that might be imposed on registered sex offenders and of the potential effectiveness of such requirements. Demographic information was collected at the end of the interview. Findings Perceptions of Convicted Sex Offenders More than 78% of all respondents believed that convicted sex offenders living near a school, day-care center, or park were more likely to commit future sex crimes than those living away from these locations (see Table 4.4). Slightly fewer than 78% of respondents believed that convicted sex offenders were likely to commit future sex crimes. 22 Contact the principle author for a copy of the survey instrument. E-mail ckatz@asu.edu, or mail request to Charles M. Katz, Ph.D., Associate Professor and Interim Director, Center for Violence Prevention and Community Safety, School of Criminal Justice and Criminology, Arizona State University, PO Box 37100, Phoenix, AZ 85069-7100. 86 About 77% of all respondents believed that having convicted sex offenders living in the neighborhood made people more fearful of crime, but this response differed significantly across areas. Respondents living in the Tampico area were significantly more likely (86.6%) than others to believe that having convicted sex offenders living in the neighborhood would make people more fearful of crime, while those living in high areas were less likely (67.8%) than others to hold that belief. More than 60% of all respondents agreed that having convicted sex offenders living in the neighborhood affected quality of life and home values. However, again, responses varied across study areas. More than 77% of Tampico respondents agreed with this statement, compared with 63.5% of respondents in medium areas, 60% in low areas, 56.3% in high areas, and 51% in the transitional area. About 55% of all respondents agreed with the statement that convicted sex offenders living in close proximity were more likely to commit future sex crimes than those living apart from other offenders. In sum, the public believes that quality of life was affected by sex offenders residing in their neighborhoods. This was most pronounced in the Tampico area, perhaps due to the public focus on sex-offender clustering there. Several factors -- media attention, heightened community awareness, mobilization on the local sex-offender clustering issue -- may have fueled Tampico residents’ concerns about the relationship of sex offenders and neighborhood quality of life. Table 4.4 Percent of Respondents Strongly Agreeing or Agreeing with Statements about Convicted Sex Offenders Study areas Low Medium High Tampico Trans. All Offenders living near park, school, 73.1 83.2 81.1 79.8 72.0 78.2 day-care more likely to re-offend Offenders likely to re-offend 79.6 79.1 76.6 79.0 73.7 77.9 Resident offenders make people more fearful of crime* 79.9 78.7 67.8 86.6 73.2 76.5 Resident offenders affect area quality of life* 60.0 63.5 56.3 77.6 51.0 61.0 Resident offenders affect home values 65.2 55.0 56.0 64.2 60.2 59.6 Offenders living in close proximity increase likelihood of their reoffending 52.1 54.1 52.3 57.0 61.9 54.5 Offenders likely to commit future non-sex crimes* 31.3 35.7 27.4 35.9 49.5 34.2 *Note. p ≤ .05 87 Neighborhood Concern with Sex and Non-sex Crimes To determine whether respondents’ perceptions of sex offending reflected distinct concerns about that issue or were part of a broader concern about crime in general, we asked them to gauge the extent to which both sex and non-sex crimes were a problem in their neighborhoods (see Table 4.5). Child molestation was viewed as a neighborhood problem by 22.6% of all respondents. More than 30% of those living in high areas and in the transitional area reported that child molestation was a problem in the neighborhood. About 17-20% of respondents in the Tampico area and in medium areas viewed child molestation as a problem in the neighborhood. In low areas, only 13% of respondents reported having a problem with child molestation in the neighborhood. About 22% of all respondents reported rape as a problem in their neighborhoods. In high areas, roughly 33% of respondents reported that rape was a problem, followed by 28.3% in the transitional area, about 21% in the Tampico area, 19% in medium areas, and 13% in low areas. Only about 20% of all respondents reported prostitution as a problem in their neighborhoods. About 37% of respondents in high areas and 25% in the transitional area reported that it was a problem in their neighborhoods. Fewer than 15% reported prostitution as a problem in the low and medium areas and in Tampico. Our analysis showed that overall, respondents believed that non-sex crimes were a more serious problem than sex crimes in their neighborhoods. For example, 62% of all respondents believed that robbery was a problem in their neighborhoods; this did not vary significantly across the sampled areas. Likewise, almost 40% of all respondents believed that drug sales were a problem in their neighborhoods. About 61% of respondents living in high areas believed this, followed by 55% of those living in the transitional area, 35.9% living in medium areas, and about 21% and 24% of those living in low areas and the Tampico area, respectively. Assault was viewed as a neighborhood problem by slightly more than 33% of all respondents. More than 52% of the respondents in high areas believed that assault was a problem in their neighborhoods. About 38% of respondents in the transitional area and 31.5% in medium areas believed that assault was a neighborhood problem. Roughly 25% of Tampico area residents and 20% of residents in low areas believed that assault was a problem in their neighborhoods. In addition to inquiring about the respondents’ concerns about sex and non-sex crimes, we asked them about their concerns with sex offenders living in their neighborhoods. About 50% of all respondents believed that convicted sex offenders living in the neighborhood were a problem, but their responses varied significantly across the study areas. More than 71% of Tampico respondents and 68% of respondents living in high areas believed that convicted sex offenders living in the neighborhood were a problem. This compared with 51.2% of respondents living in the transitional area, 40.3% of respondents living in medium 88 areas, and 26.3% of respondents living in low areas who believed that resident sex offenders were a problem in the neighborhood. In sum, we found that the extent to which both sex and non-sex crimes were viewed as problems varied across the study areas. Respondents living in high areas were significantly more likely than respondents in areas with fewer resident sex offenders to state that sex crimes such as child molestation, rape, and prostitution were problems in the neighborhood. Respondents in all areas were 2 to 3 times more likely to identify nonsex crimes as a problem in the neighborhood, compared with sex crimes. Excluding the Tampico area, which had experienced substantial public attention being focused on local sex-offender clustering, the level of concern about sex offenders living in the neighborhood varied in accord with the concentration of resident sex offenders: Respondents living in the high areas expressed the greatest concern, followed by those living in the transitional area; the least concern was expressed by respondents in neighborhoods with the fewest resident sex offenders. Table 4.5 Percent of Respondents Agreeing that Crime is Somewhat of a Problem or a Big Problem in Their Neighborhoods Study areas Crime type: Low Medium High Tampico Trans. Sex Crimes Child molestation* 13.1 20.2 31.8 17.0 34.8 Rape* 13.1 19.1 33.3 20.9 28.3 Prostitution* 13.6 11.0 36.9 8.2 25.0 Non-Sex Crimes Robbery 63.1 54.7 66.7 68.8 57.9 Drug sales* 21.1 35.9 60.9 23.9 55.2 Assault* 19.6 31.5 52.3 25.3 38.3 *Note. p ≤ .05 Perceived Likelihood of Victimization We examined the proportion of respondents who believed that they were moderately likely or very likely to become victims of crime (see Table 4.6). Of all respondents, most believed that they were at higher risk of being robbed or mugged (20.3%) than of being attacked with a weapon (14.6%) or sexually assaulted (6.9%). This pattern was consistent across all study areas, with the exception of the transitional area where respondents were slightly more likely to believe that they would become the victim of an attack with a weapon than to believe that they would be robbed or mugged. Respondents varied across study areas in the proportion that believed that they were likely to become the victim of a crime. Individuals living in high areas and in the transitional area were more likely to believe that they were at risk than those residing in the Tampico area and medium areas; the latter, in turn, were more likely to believe that they were at risk of becoming the victim of a crime than those residing in low areas. The 89 All 22.6 22.3 19.5 62.0 39.3 33.7 pattern of fear of crime found here is generally consistent with observed levels of crime found in each of the study areas. When asked whether they believed that their children were moderately or very likely to be attacked with a weapon, robbed, mugged, kidnapped, or sexually assaulted, 12-16% of all respondents believed that they were. In general, those residing in the transitional area and in high and medium areas were more likely to believe that their children were at risk than those residing in low areas and the Tampico area. Compared with respondents in the other areas, respondents in the transitional area were particularly concerned that their children were likely to become victims of rape or sexual assault. Again, this pattern of concern for their children’s safety was generally consistent with the observed levels of crime (calls for service) found in each study area. Table 4.6 Percent of Respondents Believing They or Their Children are Moderately or Very Likely to Become Crime Victims Study areas Low Medium High Tampico Transitional Self will become victim of: Rape or sexual assault* 2.5 6.3 9.6 6.0 12.5 Attack with weapon* 7.0 15.9 17.7 10.0 26.0 Robbery or mugging* 13.5 17.9 25.6 23.0 25.3 Child will become victim of: Rape or sexual assault 8.6 9.9 13.8 6.7 22.2 Kidnapping 9.9 11.3 19.8 6.7 17.8 Attack with weapon 7.0 19.2 24.7 3.3 17.8 Robbery or mugging 7.0 19.2 17.3 16.7 20.0 All 6.9 14.6 20.3 12.2 13.8 16.0 15.7 *Note. p ≤ .05 Citizen Awareness of Arizona’s Sex-offender Notification Law More than 82% of all respondents were aware of Arizona’s sex-offender notification law before receiving the researcher’s calls (see Table 4.7). Most had learned about it through the media (69.8%). Others reported learning about the law in various ways: when they received a sex-offender notice in the mail (10.1%), from someone telling them about it (8.2%), by other methods (7.5%), on the Internet (2.5%), or from attending a community meeting (1.9%). We found significant variation across study areas in the proportion of respondents who were aware of Arizona’s sex-offender notification law, however. Eightyeight percent of respondents in the Tampico area and in low areas knew about the law, while 78% to 81% of respondents in medium and high areas and in the transitional area knew about it. We also found significant variation across areas in how respondents had learned about the law. Slightly more than 76% of respondents residing in low and medium areas had learned about it through the media, compared with 69.3% of those living in the Tampico area, 61.8% of those living in the high area, and 57.1% of those living in the transitional area. 90 Respondents living in medium and high areas were more likely than those in other areas to have learned about it from another individual; respondents in the Tampico and transitional areas and the high area were more likely than others to have learned about it by receiving a sex-offender notice in the mail. Community meetings helped to generate awareness about Arizona’s sex-offender notification law in the Tampico area, but appeared to have only a small impact in other areas. Table 4.7 Percent of Respondents Aware of Arizona Sex-offender Notification Law and Information Sources Study areas Low Medium High Tampico Trans. All Aware of law* 88.0 78.4 80.7 88.0 78.0 82.6 Source of information:* Word-of-mouth 5.7 9.5 11.5 6.8 6.5 8.2 Media (newspaper or TV) 77.6 76.2 61.8 69.3 57.1 69.8 Sex-offender notice in mail 4.0 5.4 15.9 11.4 19.5 10.1 Community meeting 0.6 0.7 1.9 5.7 2.6 1.9 Internet 2.9 3.4 1.9 1.1 2.6 2.5 Other 9.2 4.8 7.0 5.7 11.7 7.5 *Note. p ≤ .05 Respondents’ Beliefs about the Purpose of Notifying Communities about Released Sex Offenders More than 85% of all respondents agreed that the purpose of sex-offender notification was to “mak[e] neighborhoods safer by letting people know about a convicted sex offender living there” (see Table 4.8). Similarly, 80.1% of all respondents agreed that the purpose of the notification law was to increase the chance of a quick arrest, should a convicted sex offender re-offend. Responses to these two items did not vary significantly across study areas. Nearly 69% of all respondents agreed that community notification reduces the chance that a convicted sex offender will commit another sex crime. However, significant differences emerged on this survey item in accord with the area in which respondents lived. Those in the Tampico area and in low areas were significantly more likely than the others to agree with that statement (76% and 75.1%, respectively). About 69% of those residing in the transitional area and 64.3% of those living in the medium area agreed with it. Respondents living in the high area were least likely to agree that sex-offender notification reduces the chance that a convicted sex offender will commit that type of crime again. About 35% of all respondents agreed with the statement that the notification was intended to punish the sex offender. Those living in the transitional area were the most likely to believe this (45.5%), while those living in the Tampico area were the least likely (25%). Roughly 32% to 37% of respondents in low, medium, and high areas agreed that sex-offender notification was intended to further punish sex offenders. 91 Table 4.8 Percent of Respondents Agreeing or Strongly Agreeing with Purpose of Sex-offender Notification Study areas Purpose of law: Low Medium High Tampico Trans. All Make neighborhoods safer 84.4 85.4 88.1 89.0 79.8 85.6 Facilitate quick arrest for repeat 79.2 79.4 76.3 83.7 86.9 80.1 offense Reduce chance of repeat offense* 75.1 64.3 61.7 76.0 69.1 68.5 Further punish offender* 31.8 36.0 37.2 25.0 45.5 35.1 *Note. p ≤ .05 Respondents’ Experiences with the Arizona Sex-offender Notification Law Slightly more than 30% of all respondents indicated that they had received one or more sex-offender notifications (see Table 4.9). Data analysis showed significant variations across areas, however, with about 47% living in high areas having received a notification, followed by 42% in the transitional area, 33% in the Tampico area, 25.7% in the medium areas, and 10.6% in the low areas. On average, those living in the Tampico and transitional areas received more than twice as many notifications as those in areas with low, medium, and high numbers of sex offenders living in their neighborhoods. These percentages would appear to indicate rather low notification rates, but two qualifications are in order. First, the analysis did not account for transience (how long respondents had been residing in their neighborhoods) nor for the fact that some respondents may not have been living in the neighborhood at the time that a community notification was issued. Second, the policy did not require that everyone living in the neighborhoods within the study grids be notified; the Phoenix Police Department’s policy was that households be notified when located within two blocks of an offender’s residence.23 Table 4.9 Respondents’ Experiences with Arizona Sex-offender Notification Law Study areas Low Medium High Tampico Percent receiving notification* 10.6 25.7 47.2 33.0 Number of notifications/respondent* Mean 1.3 1.4 5.8 12.2 S.D. 0.6 1.0 17.3 23.3 Trans. 42.0 All 30.3 11.5 25.3 6.4 17.8 *Note. p ≤ .05 23 The Phoenix Police Department defines minimum notification boundaries as two blocks in either direction of an offender’s residence, including a one-mile radius for all schools, day-care facilities, neighborhood groups, and DHS-licensed facilities. 92 Social and Psychological Consequences of Being Notified That a Sex Offender Lives in the Neighborhood The survey included questions intended to help examine the social and psychological consequences of being notified that a sex offender lives in one’s neighborhood. More than 58% of all respondents indicated that after notification, they were more safetyconscious, and 20.4% stated that they were less likely to go out alone. Responses to these questions did not vary significantly across areas (see Table 4.10). Slightly more than 13% of all respondents stated that they had become more involved with community activities after being notified that a sex offender lived in the neighborhood. Tampico respondents were the most likely (24.6%) to become involved in community activities after notification, followed by transitional-area respondents (16%). About 12% of those living in the medium areas and 10% of those living in the high areas stated that they were more involved in community activities after they were notified. Only 2.4% of respondents living in the low areas stated that they were more involved in community activities after they were notified that a sex offender lived in the neighborhood. Table 4.10 Social and Psychological Consequences of Receiving Sex-offender Notification Study areas % of respondents: Low Medium High Tampico More conscious of safety 48.8 48.5 56.9 75.4 Less likely to go out alone 9.8 18.2 19.6 26.7 More involved with (protective) 2.4 11.9 10.4 24.6 community activities* Trans. 63.5 26.0 All 58.7 20.4 16.0 13.2 *Note. p ≤ .05 The findings described above tended to support the notion that notification might increase public safety in at least two ways. First, notification could lead to awareness, which in turn could lead to citizens being more safety-conscious and cautious. Second, it appears that notification could lead to increased citizen involvement and activism focused on improving neighborhood safety. Notification may also motivate self-protective actions such as those described in the next section. Self-protective Actions Taken After Receiving a Sex-offender Notification Nearly 60% of all respondents stated that they watched their children more closely or restricted their activities after receiving notice that a sex offender was living in the neighborhood; 29.3% installed extra locks, lighting, or doors; 26.2% organized or participated in the local neighborhood watch chapter; 24.8% increased their knowledge of self-defense; 19.4% bought a watch dog. Nearly 2% reported moving to another area. Reported self-protective actions did not vary significantly across the study areas (see Table 4.11). 93 About 61% of all respondents stated that after being notified that a sex offender lived in the neighborhood, they were more careful about locking doors. This change in selfprotective behavior varied across areas. More than 75% of respondents in the Tampico area reported taking this action, compared with about 60% of respondents living in transitional, high, and medium areas, and 37.5% of respondents living in the low areas. More than 16% of all respondents said that they had begun to keep a weapon at home after being notified that a sex offender was living in the neighborhood. This particular self-protective behavior varied significantly by study area. More than 37% of respondents in the transitional area reported taking this action, followed by about 20% of those living in the medium areas, 12% of those living in low and high areas, and 5.2% of those living in the Tampico area. About 26% of all respondents stated that after being notified that a sex offender lived in the neighborhood, they began avoiding particular areas in that neighborhood. More than 35% of respondents in the Tampico and transitional areas stated that they had taken this action. About 24% of respondents in the high areas, 20% of respondents in the low areas, and 15.2% of respondents living in the medium areas stated that they had avoided certain places after being notified that a sex offender lived in the neighborhood. The greatest differences among the study areas occurred among respondents who had tried to force an offender to move out of the neighborhood after notification. More than 15% of respondents in the transitional and Tampico areas stated that they had taken this action, compared with fewer than 2% of respondents in areas labeled low, medium, and high. Table 4.11 Percent of Respondents Acting After Receiving Sex-offender Notification Study areas Low Medium High Tampico Installed new locks, lighting, doors 20.0 23.9 34.9 26.2 Locked doors more often* 37.5 59.1 61.1 75.4 Bought watchdog 20.0 13.6 21.5 13.6 Brought weapon into home* 12.5 19.7 12.3 5.2 Increased knowledge of self-defense 22.5 16.4 24.3 24.6 Watched children more closely, 58.3 60.6 52.4 60.4 restricted their activities Organized/participated in local 27.5 19.4 21.0 32.8 neighborhood watch Attempted to remove offender from 0.0 1.5 1.9 18.6 neighborhood* Moved away from neighborhood 0.0 3.0 1.9 0.0 Avoided certain neighborhood areas* 20.0 15.2 23.9 39.3 *Note. p ≤ .05 94 Trans. 35.3 62.7 28.8 37.3 38.5 All 29.3 60.7 19.4 16.5 24.8 72.5 59.5 37.3 26.2 15.4 6.8 3.8 34.6 1.8 26.2 Social and Behavioral Reactions to a Sex Offender Moving into the Neighborhood Our survey queried respondents about a variety of potential social and behavioral reactions to being notified that a sex offender had moved into the neighborhood. We asked a series of questions about two types of sex offenders: rapists and child molesters. Few significant differences emerged across the study areas. (For area-specific findings, see Table 4.12.) Asked for the actions that they were likely to take upon learning that a rapist had moved into the neighborhood, more than 85% of all respondents stated that they would teach their children how to avoid becoming a rape victim and they would spend more time watching their children playing outside. More than 47% of all respondents said that they would join their neighborhood watch chapter to monitor the rapist’s behavior; 36.4% said that they would sign a petition demanding that the offender not be allowed to live in the neighborhood; 13.2% said that they would move away; and 2.5% said that they would make threatening phone calls to the rapist. Although in the minority, 30% of all respondents said that they would be likely to attend neighborhood meetings to help an offender reintegrate into society; 5.7% stated that they would socialize with the offender. We also asked respondents to answer the same questions, should they learn that a child molester had moved into the neighborhood. Again, well over 85% of all respondents indicated that they would be moderately likely or very likely to teach their children how to avoid becoming a victim and to spend more time watching their children playing outside. More than 51% said that they would be likely to join the neighborhood watch chapter to monitor the offender’s behavior; 36.2% said that they would sign a petition to demand that the offender be prevented from living in their neighborhood. About 13% of all respondents indicated that they would move away; 2.7% said that they would make threatening phone calls to the offender. As with the questions about a rapist moving into the neighborhood, a minority of the respondents had a different perspective: about 26% stated that they would be likely to attend neighborhood meetings to help the offender reintegrate into society, and 6.1% stated that they would socialize with the offender. The respondents’ social and behavioral reactions did not vary substantially based on the type of sex offense. Their response patterns were fairly consistent regardless of whether the hypothetical offender moving into the neighborhood was a rapist or a child molester (see Table 4.12, “all” column). 95 Table 4.12 Respondents’ Likely Social/Behavioral Responses to Sex Offender Moving into Neighborhood Study areas Low Medium High Tampico Trans. All Likely or very likely response if resident offender is a rapist (%): Socialize with offender 3.5 8.3 Attend neighborhood meetings to help offender reintegrate into society Sign petition demanding the offender not be allowed to live in neighborhood Move out of neighborhood Make threatening calls to offender Join neighborhood watch, monitor offender Teach children to avoid becoming rape victim Watch child when playing outside 6.6 3.0 6.0 5.7 26.3 30.7 33.0 28.0 32.0 30.0 34.8 37.2 33.5 40.8 39.4 36.4 9.5 1.0 14.0 4.7 17.2 3.5 9.1 0.0 15.0 2.0 13.2 2.5 45.0 46.1 45.2 48.0 57.0 47.2 92.1 84.4 82.2 81.5 86.6 85.7 88.2 85.1 85.8 83.1 87.4 86.1 Moderately likely or very likely response if resident offender is a convicted child molester (%): Socialize with offender 5.0 5.8 8.0 4.0 7.0 Attend neighborhood meetings to 21.5 29.1 26.1 25.3 29.3 help offender reintegrate into society Sign petition demanding the offender not be allowed to live in 36.2 38.3 34.2 34.3 38.0 neighborhood Move out of neighborhood 9.5 14.7 14.2 6.1 18.0 Make threatening calls to offender 1.5 3.6 3.0 1.0 4.0 Join neighborhood watch, monitor 54.0 46.4 50.0 53.5 57.0 offender Teach children to avoid becoming 91.0 84.9 84.9 84.6 84.2 rape victim Watch child when playing outside 92.0 87.0 89.5 85.7 86.5 *Note. p ≤ .05 6.1 25.9 36.2 12.6 2.7 51.5 86.3 88.6 Respondent Opinions About Issues Related to Sex-offender Community Notification With few exceptions, respondents held remarkably consistent opinions about sexoffender notification, regardless of where they lived or the number of sex offenders living in their neighborhoods (see Table 4.13). Almost 61% of all respondents agreed with the statement that special care should be taken to ensure that citizens do not harass a sex offender who has been released to the community. Related, 35.6% of respondents believed that community-notification laws made it easier for citizens to take the law into their own hands and to harass, threaten, or abuse registered sex offenders. Note that the finding that citizens who lived in the low areas (no registered sex offenders) were significantly more likely to agree with this statement (41.6%) than were those who lived in areas with sex offenders. Even though respondents voiced concern about the potential 96 harassment of sex offenders as a consequence of community notification, almost 59% agreed that people felt safer knowing about convicted sex offenders living in their neighborhoods, and that people learned more about sex offenders and how sex offenders operate because of community-notification laws. Data analysis revealed that many respondents (59.2%) agreed that community notification makes it difficult for convicted sex offenders to establish a new life. Still, only about 39% of all respondents agreed with the statement that convicted sex offenders who are released into the community should be given an opportunity to establish a new life as a law-abiding citizen. Compared with respondents in the low and medium areas, those who lived in the high, Tampico, and transitional areas were significantly more likely to agree that sex offenders should be given that opportunity. Only 45.9% of all respondents agreed with the statement that community notification makes sex offenders less likely to commit another sex crime than they would be if no one in the community knew of their background. Slightly more than 45% agreed that sexoffender registration and community notification were effective tools for preventing future sex crimes. Regardless of which study area respondents lived in, they expressed dissatisfaction with the performance of the police in notifying the public about sex offenders. Only 32.6% of all respondents agreed or strongly agreed with the statement that the police do a good job of notifying citizens about convicted sex offenders living in their communities. 97 Table 4.13 Percent of Respondents Strongly Agreeing or Agreeing with Sex-offender Notification Statements Study areas Low Medium High Tampico Trans. All Care should be taken to protect released offenders from citizen harassment 60.2 58.6 61.6 64.3 61.6 60.9 People feel safer knowing about the offenders in the neighborhood, even if not notified of other kinds of criminals who live there 61.0 62.3 58.2 55.6 55.6 59.3 Notification makes it difficult for offenders to establish new life 66.2 61.6 54.9 56.7 51.5 59.2 Offender registration is effective at preventing future sex crimes 42.7 43.9 54.4 39.1 41.7 45.4 Community notification makes sex offenders less likely to repeat offense 51.6 41.0 44.4 46.5 46.9 45.9 Notification process is effective at preventing future sex crimes 41.1 46.1 46.5 43.2 43.4 44.2 Released convicted sex offenders should have opportunity to establish new life as law-abiding citizen* 34.4 29.1 50.5 40.0 42.4 38.9 Community notification makes it easier for citizens to take law into own hands, harass/threaten/abuse registered sex offenders* 41.6 32.8 35.7 31.3 33.3 35.6 Police do good job of notifying citizens about convicted sex offenders living in community 30.6 32.5 37.5 26.3 32.2 32.6 *Note. p ≤ 0.5 Assessment of the Arizona Sex-offender Notification Law About 97% of respondents stated that the Arizona notification law was very important (74.1%) or somewhat important (22.8%). Likewise, about 91% agreed that the state ought to notify residents when a sex offender lives in the neighborhood. Respondents did not differ significantly across study areas on this question (see Table 4.14). These results are intriguing, given that fewer than half of all respondents believed that sex-offender notifications were effective in preventing repeat sex offenses. The findings in the next section seem to confirm that although citizens doubt the effectiveness of the notification law, they still consider it an important public safety strategy. 98 Table 4.14 Perceived Importance of and Need for Arizona Sex-offender Notification Law (%) Study areas Low Medium High Tampico Trans. Importance of state notification law Not important 3.5 3.2 2.0 4.0 3.0 Somewhat important 26.0 21.1 20.6 25.0 22.0 Very important 70.5 75.8 77.4 71.0 75.0 Should have state notification law Disagree 5.5 4.2 3.5 2.0 4.0 Undecided 5.0 5.8 5.5 2.0 7.0 Agree 89.5 90.1 91.0 96.0 89.0 * p ≤.05 Notification Methods Believed Most and Least Helpful The survey asked respondents to select which notification methods they would find most and least effective for being informed that a sex offender lived in the neighborhood. There were no significant variations across the study areas in the responses to either question. More than 70% of all respondents believed that receiving notices by mail would be the most helpful, while slightly more than 13% selected the media (e.g., newspaper or television). Fewer than 5% selected community meetings, the Internet, word-of-mouth, or other strategies as the preferred method for notification. (See Table 4.15.) Asked which notification method would be least helpful for learning about sex offenders living in the neighborhood, less than a third of all respondents (27.6%) selected word-ofmouth. About 21% believed that the Internet and community meetings were least helpful; roughly 19% believed that the media was least helpful; and 6.1% and 4.4% respectively stated that other methods and notices mailed to homes were least helpful. 99 All 3.0 22.8 74.1 4.0 5.2 90.8 Table 4.15 Most and Least Helpful Notification Methods for Notifying Communities of Resident Sex Offenders % of all Most helpful: Word-of-mouth 3.7 Media: Newspaper or TV 13.3 Sex offender notice in mail 70.9 Community meetings 4.6 Internet 4.9 Other 2.6 Least helpful: Word-of-mouth Media: Newspaper or TV Sex offender notice in mail Community meetings Internet Other 27.6 19.0 4.4 21.1 21.9 6.1 Public Perceptions of the Effectiveness of Strategies in Preventing Sex Crimes We listed six strategies that might help prevent repeat sex crimes (see Table 4.16). Respondents viewed all of them to be very effective or somewhat effective, with no significant differences across study areas. More than 90% of the respondents living in each area thought that supervised probation and community notification were effective preventive strategies. More than 85% believed that mandatory sex-offender registration, sex-offender treatment and therapy, and laws regulating the distance that registered sex offenders must live from schools, day-care centers, and parks were effective. More than 80% of all respondents thought that laws limiting the number of registered sex offenders living in a particular area would be effective; even so, analysis showed that of the strategies offered, respondents had the least confidence in the prevention value of this measure. Table 4.16 Percent of Respondents Believing Prevention Strategies are Somewhat or Very Effective % of all Supervised probation for released convicted sex offenders Community notification of resident registered sex offenders Mandatory registration for released convicted sex offenders Treatment or therapy for released convicted sex offenders Laws regulating distance of registered sex offender’s residence from schools, day care facilities, parks Laws limiting number of registered sex offenders living in close proximity to one another 100 91.7 90.7 89.7 86.4 86.3 81.1 Public Opinions on Sex-offender Residential Restriction Proposals We asked respondents for their opinions about several sex-offender residential proposals being discussed throughout the United States (see Table 4.17). Regulation of the number of sex offenders living in a specific area received the greatest overall support, with 72.2% of all respondents agreeing that this should occur. Likewise, the vast majority (73.4%) believed that sex offenders should not be allowed to reside in the same dwelling as other sex offenders. No significant differences appeared on these items across the study areas. Respondents were substantially less likely to support other proposals governing where a sex offender could live. For example, only about one third agreed that sex offenders should be allowed to reside in the same house, apartment complex, or mobile home park as other sex offenders only if a judge, probation officer, or parole officer authorizes them to do so. Table 4.17 Percent of Respondents Strongly Agreeing or Agreeing with Regulations of Sex-offender Residency Study areas Low Medium High Tampico Trans. Regulate number of offenders living in one community/area Regulate number of offenders living without judicial order in one dwelling, apartment complex, or mobile home park Regulate number of offenders living without probation/parole officer authorization in one dwelling, apartment complex, or mobile home park Do not regulate All 65.3 72.8 74.7 74.7 77.0 72.2 31.6 33.7 39.0 34.0 40.2 35.4 28.1 37.1 36.4 29.9 42.4 34.4 29.4 25.1 28.6 18.8 27.3 26.6 *Note. p ≤.05 Public Preferences for Regulating the Residential Location of Sex Offenders Our analysis indicated that roughly 75% of all respondents supported at least some restriction on where a sex offender may reside (see Table 4.18). About 83% agreed that the city should regulate how far away registered sex offenders must live from schools, day-care centers, and parks; about 75% of respondents agreed that sex offenders should be required to move out of their homes if they are too close to such places. Nearly three quarters of all respondents (74.1%) agreed that the city should regulate the number of registered sex offenders who are permitted to live in the same neighborhood. Very few significant differences surfaced among respondents across the study areas. 101 Table 4.18 Percent of Respondents Strongly Agreeing or Agreeing with City Sex-offender Residence Location Policies Study areas Low Med High Tampico Trans. All Regulate distance of offender residences from schools, day-care facilities, parks Require offender to move if residence is too close to school or day-care facility* Require offender to move if residence is too close to park or playground Regulate number of registered offenders allowed to live in neighborhood 79.1 87.4 83.5 82.0 81.0 82.8 71.4 81.6 79.6 81.4 74.7 77.6 67.3 78.3 75.1 77.6 73.7 74.1 71.5 76.1 73.9 76.5 73.7 74.1 *Note. p ≤.05 Preferred Sex-offender Housing Density We asked respondents about their preferences for the density distribution of sex offenders within the city: whether sex offenders should be concentrated in one area or distributed across the city (see Table 4.19). The majority (47.3%) stated that they preferred for sex offenders to be spread across the city; however, roughly 20% preferred that sex-offender residences be concentrated. One third of all respondents did not know which would be better. Analysis showed no significant differences in how respondents from different areas answered this question. Table 4.19 Preferred Sex-offender Residence Density % (all) 19.5 47.3 33.2 Concentrate in one area Distribute across city Do not know *Note. p ≤.05 Conclusions and Policy Implications The survey findings described in this report provided insight into the concerns of Phoenix citizens regarding sex offending and sex offenders’ residential distribution, as well as their policy preferences for managing sex offenders in the community. Overall, whether they resided in areas with greater numbers of resident offenders or few (or no) offenders, the majority of respondents perceived the known presence of offenders as having a negative impact on the quality of life in their neighborhoods. Citizens’ level of concern about sex crimes and sex offenders who lived in their neighborhoods was generally no greater than their concern about non-sex crimes, e.g., robbery, drug sales, or assault. When developing public safety strategies with 102 the goal of reducing the public’s fear of crime, it would be well to note that reducing fear of sex offending alone is unlikely to reduce fear of non-sex crime, nor will reducing fear of non-sex crimes necessarily reduce fear of sex offending. The concentration of known sex offenders living in their neighborhoods did make a difference in the respondents’ perceptions of the neighborhood as having a problem with sex crimes and resident sex offenders. Overall, about half of all respondents indicated that having convicted sex offenders living in the neighborhood was indeed a problem; Tampico residents were the most concerned, and those living in low areas were the least concerned. Across the board (Tampico being somewhat of an exception), respondents were more concerned about robbery in their neighborhoods than about sex crimes or resident sex offenders. Even in the Tampico area, respondents reported less concern about problems such as child molestation, rape, and prostitution than about non-sex crimes such as robbery, drug sales, and assault. When asked about the likelihood that they or their children would become sex crime victims, all respondents except those in the transitional areas reported that the likelihood was relatively low. Overall, respondents were about three times more likely to think that they would become a victim of a non-sex crime than of a sex crime. They also estimated the likelihood of their children becoming victims of non-sex crimes as greater than the likelihood of them becoming victims of sex offenses. These findings tended to reinforce the conclusion that Phoenix citizens are more fearful about non-sex crimes than about sex crimes. Thus, policies and practices aimed at fear reduction should focus on more than one type of crime. Community notification has become one of the most important tools used to protect the public from registered sex offenders. The vast majority of Phoenix citizens were aware of Arizona’s notification law, with most reporting the media as their source of information. They appeared to have a reasonably good understanding of the law’s affirmative purposes and tended not to perceive its primary purpose as punitive. Their actual experiences with notification depended on where they lived, but even in those areas with the greatest numbers of offenders, fewer than half of the respondents reported having ever received an official notification. This finding could in part be an artifact of the transient nature of the study population or of the current requirement to notify only those residing within two blocks of a registered offender. Yet only about one third of the respondents in the Tampico area and in the high areas reported ever receiving such notification. Artifact or not, further investigation into the notification process is needed to assure that residents close to registered offenders are being properly notified and to identify ways to improve the notification process. The probability that any particular individual, adult or child, will become the victim of a registered sex offender is low. According to our survey findings, the social and psychological impacts and other consequences of learning through notification that a registered sex offender was living in the neighborhood could be both negative and positive. The negative consequences that respondents reported included avoiding specific places and being less likely to go into the neighborhood alone. Few citizens reported having taken the drastic action of moving away when notified of a resident sex offender. 103 The majority of our respondents reported a positive outcome: becoming more safety conscious after notification. A substantial segment reported taking self-protective actions aimed at increasing their own and their children’s safety. Likewise, a substantial segment reported increasing involvement in community organizations to improve public safety in their neighborhoods. Notification serves to strengthen public safety to the extent that citizens take appropriate self-protective action. Avoiding specific places and not going out alone can reduce the risk of victimization, but it also reduces the natural guardianship that is so important for crime prevention and healthy neighborhoods. Notification presents a dilemma. The policy and practice should incorporate strategies for increasing self-protective action, but at the same time, send a positive message that promotes normalcy in the day-to-day affairs of neighborhood residents. Survey findings also indicate that notification would lead to a variety of responsible social and behavioral actions taken by citizens who are informed that a registered sex offender has moved into the neighborhood. Such actions most frequently reported by respondents included educating their children, supervising children outside the home, and joining a neighborhood watch chapter. In general, few respondents indicated a preference for actions such as threatening offenders with telephone calls or moving out of the neighborhood. However, a substantial number indicated that they would sign petitions demanding that the offender not be allowed to reside in the neighborhood. Especially noteworthy is the number of respondents (more than 25%) who indicated that they would attend meetings to help sex offenders reintegrate into society. This finding suggested that restorative justice strategies that promote the reintegration of offenders should be considered as part of a comprehensive strategy for managing sex offenders in the community. In general, a majority of the Phoenix citizens who participated in the survey demonstrated an understanding of the potential negative impact of community notification on sex offenders, including making it difficult for the offender to establish a new life. These respondents indicated that harassment of sex offenders was inappropriate, and more than 35% of them agreed that notification could make harassment of offenders easier. Phoenix citizens offered mixed assessments of the effectiveness of community notification. Although a slight majority indicated that notification made people feel safer, fewer than half thought that it was an effective tool for the control or prevention of sex offending. Furthermore, less than a third indicated that the police were doing a good job of notifying the public. Even though the majority might not have believed that notification was effective, a sizeable majority thought that Arizona’s notification law was important and necessary. They also thought that the most effective method of notification was the U.S. mail. The vast majority of citizens surveyed gave positive assessments for a variety of strategies aimed at addressing the sex-offender problem and controlling sex offenders. Strategies such as supervised probation, community notification, registration, treatment, distance buffers from places where children gather, and 104 limits on sex-offender residential density all were generally viewed as effective. Likewise, a solid majority of citizens indicated support for a number of policies regulating the spatial location and distribution of sex offenders. For example, the majority of Phoenix residents surveyed did support policies restricting multiple sex offenders from living in the same residence, apartment complex, or trailer park, and they tended to prefer having sex offenders spread throughout the city rather than living in high numbers in one area. However, many residents were uncertain exactly what the distribution pattern should be. These views stand in contrast to those of some experts who maintain that an effective strategy for supervising and controlling sex offenders is to have them live together. Such an arrangement can make supervision more effective, and the offenders can help control one other’s behavior. With respect to policies and proposals for addressing the sex-offender residential problem and controlling sex offenders, our survey findings indicated that community notification, while not always thought to be an effective prevention measure, does serve the purpose of increasing the self-protective behaviors of residents once they are notified. Although it could be an artifact of geography and population turnover, it appeared that large numbers of citizens, even in areas with a high number of resident sex offenders, were not always receiving the notification. This issue needs additional analysis, with consideration of strategies for increasing the extent and effectiveness of notification. However, although it might appear that the notification process can easily be expanded, it cannot be done without considerable expense. The survey findings demonstrated widespread support for contemporary strategic proposals aimed at controlling the spatial location and density of sex offenders. Every proposal that appeared to place increased residential restrictions on sex offenders had or would have public support, at least as indicated by these survey results. However, note that residents probably equated increased restrictions with increased control. They were unaware of the difficulty in implementing these proposals and of some of the negative impacts such as making offender supervision more difficult and costly. One finding that should not be overlooked is the possibility of substantial public support for restorative justice strategies that focus on community engagement in reintegrating sex offenders into the community. Such strategies, included as part of a comprehensive model that includes community notification and education, increased supervision, treatment, and manageable residential location practices for offenders, would receive adequate public support. The sex-offender residency issue is a complex problem. Strategies that focus on a single dimension are not likely to solve it. Comprehensive models that engage the community along with professionals in problem-solving are much more likely to be effective. 105 List of References Center for Sex Offender Management (2000, April), retrieved on May 3, 2005, from http://www.csom.org/pubs/pubpinion.html. Skogan, W. (1990). Disorder and decline: Crime and the spiral of decay in American neighborhoods. New York: Free Press. Zevitz, R. (2003). Sex offender community notification and its impact on neighborhood life. Crime Prevention and Community Safety: An International Journal, 5, 4: 41-61. 106 5 The Impact of Sex-offender Residential Clustering and Related Policies on Selected Agencies and Organizations: Perceptions and Recommendations Vincent J. Webb, Ph.D. As one component of the larger sex-offender clustering study, we conducted a series of interviews with key informants from selected Phoenix agencies and organizations. These were for the purpose of collecting information from the stakeholders’ perspectives about issues related to the residential distribution of sex offenders within the community. The interviews were limited in number, and thus were not representative of all agencies or organizations affected by the sex-offender problem. In addition, not all respondents were representing their agencies’ official positions on the issues spoken to. Nonetheless, their responses shed light from yet another important perspective on the problems and possibilities associated with alternative management and control strategies. Interviewees were selected from a list provided by the Phoenix Police Department. The list identified agency and organization representatives known to have a special role in addressing the Phoenix sex-offender problem. We excluded elected officials in order to maintain a common thread among those interviewed. All those remaining were from governmental or nonprofit agencies that had direct contact with or responsibility for sexoffender policy implementation. Those interviewed had supervised, treated, housed, and been responsible for assuring compliance with laws related to sex offenders in the community, or they had delivered programs and services to sex-offender victims. The small number of interviews (N=10) makes it difficult to maintain appropriate levels of anonymity for those participating. Therefore, what follows are summaries of their perceptions. Sizing Up the Sex-offender Problem and Sex-offender Policy Every key informant interviewed agreed that sex-offender residential clustering was a major issue for the City of Phoenix. However, their opinions differed on the nature of the issue, and not all informants limited their assessments to the perceived threat that sex offenders posed for Phoenix residents. Several informants suggested that members of the public were often confused about the nature of sex offenses and had a tendency to portray all sex offenders as the same, that is, as posing the same degree of risk to the community. Some informants believed that citizens tend to overreact when they hear about sex offenders, overestimating the threat that they pose. Community and family education was suggested as a strategy for countering such overreaction. Nearly all 107 agreed that better information on the levels of sex-offender classification needed to be communicated to residents and policymakers alike. Some thought that the constant negativity of the sex-offender policy debate communicated wholly negative messages to the community; they would prefer that some positive steps for addressing the sex-offender residential problem be included in that debate. Some thought a more factual description of the sex-offender population might generate some measure of compassion for the plight of offenders, paving the way for development of more positive residential policies. A focal point of this study was sex-offender residential “clustering,” and several informants provided detailed assessments of this dimension of the sex-offender issue in Phoenix. Some attributed the development of residential clusters of sex offenders to the increased use of notification, which funnels sex offenders into certain residential areas out of necessity. Not all informants assessed clustering in negative terms. Some reported that sex-offender residential clustering does not necessarily increase risk to the community; they suggested that the level of risk remains the same, whether it is one offender of a certain type or for a group of offenders. Some informants indicated that residential clustering can be a positive strategy, making supervision and potential treatment more efficient and cost-effective. According to some, since sex offenders need a place to live and will reside somewhere in the community, clustering can actually make them easier to track and monitor. Some indicated that strategies for breaking up residential clusters and policies that impose stringent residential restrictions can have the undesirable effect of driving sex offenders underground, making their whereabouts unknown to criminal justice officials and consequently increasing rather than decreasing the risk to the community. Several informants commented on a range of policy options for controlling sex offenders in the community. Some called for the increased use of Global Positioning Satellite (GPS) technology for tracking offenders, which they thought could result in more effective and less costly supervision. Some thought that the state needed to take more responsibility for housing and supervision sex offenders and advocated increased use of civil commitment. Civil commitment was viewed as one way of increasing control over offenders, keeping them off the street while taking advantage of available bed space in state facilities and providing treatment in a residential setting. This option would require careful review and modification of existing laws to enable such expanded use of these facilities. Some informants called for improving technology in order to make community notification more timely. They also called for improving databases to ensure accuracy in sex-offender address information. Still others advocated improving consistency in the classification of sex offenders and expanding group treatment programs prior to their reentry into the community. All of the informants recognized that sex offenders entering the community needed a place to live. Some reported more success in placing sex offenders in jobs that led to self-support than in finding appropriate housing for them. Several informants indicated that policymakers needed to 108 look for positive solutions such as special districts or residential complexes licensed for the purpose of housing such offenders. Some informants expressed concerns about proposed distance restrictions aimed at reducing sex-offender density and/or providing buffer zones around selected areas such as schools or day-care centers. They were supportive of buffer zones around childcentered facilities, but some thought that distance restrictions would make it difficult for sex offenders to obtain housing and would drive them into hiding, making them difficult to locate and control. Others expressed concerns about the ability to enforce compliance with increased distance restrictions; for example, ensuring compliance with distance restrictions from licensed in-home day-care providers would depend on overcoming the difficulty of pinpointing the addresses or geographic locations of such providers. Some interviewees voiced concern about the geographic feasibility of implementing distance restrictions; they thought that it might be impossible to disperse offenders broadly enough across the city to comply with such ordinances. Informants suggested bolder and more creative approaches to housing sex offenders. Pilot programs using self-paid GPS ankle bracelets and group housing situations with intensive supervision were given as examples. Some recommended establishing a community-wide task force to initiate a dialogue and to find positive housing solutions for sex offenders. Unfunded Mandates and Mission Displacement Several informants expressed concern that sex-offender residential policies, existing or proposed, could become unfunded mandates or displace the existing missions of the agency or organization. From their perspective, policymakers, while appropriately concerned about maximizing public safety, tended to propose policy changes without due consideration of the financial impact on public safety agencies. As an example, informants cited the existing sex-offender notification practices. By one estimate, notification flyers alone cost in excess of $160,000 annually, but there is no budget line item for this cost. Policy changes that increase the frequency of notification cause increases in the costs for material, personnel, and other items, usually without an appropriate budget increase. New policies that would impose distance restrictions or buffers would similarly require funds to ensure offender compliance. Any policy that increases the supervision of sex offenders in the community without a corresponding budget increase is likely to result in reallocation of resources away from other important public safety activities. In sum, most key informants were more concerned about the unfunded cost impact of policies on their agencies than they were optimistic about the potential efficacy of such policies. For some, sex-offender policy was seen to have as much impact on the mission of their organization as on their agency’s budget, although that impact was still considerable. For these, the sex-offender problem and lack of positive housing policy was seen as threatening to displace the mission of the organization. For example, a major organization with the formal mission of providing shelter and services to the homeless has over time become a residential center for sex 109 offenders unable to find housing elsewhere. Many of these offenders are reported to be employed and financially stable, but because they cannot find housing, they have no choice but to live at the shelter. The beds taken up by the sex offenders are beds taken away from the homeless that the shelter was originally established to serve. Summary and Recommendations There was consensus among the key informants that we interviewed that sex offender residency was a major issue for the city of Phoenix. There was substantial consensus that the public reaction to sex offenders had become highly emotional and was without a basis in clear and accurate information. When all sex offenders, regardless of classification, are seen as one high-risk group in the public mind, it becomes harder to align policy with effective public safety practice. Sex-offender residential clustering within the community is not seen negatively by all key informants; some view it as facilitating supervision and control and as more cost-effective than policies that disperse offenders. Some informants found the public policy debate on sex-offender residency to be overly negative in tone. From their perspective, policies that impose residential restrictions may initially soothe public fear and concern, but they are not well-informed by the facts surrounding patterns of sex offending or the characteristics of different types of sex offenders. Furthermore, although restrictive policies aim to increase the buffer between offenders and potential victims, they are also difficult to implement and they make it unnecessarily difficult for sex offenders to find housing. Several key informants were leery of policies that imposed harsher restrictions on sex offenders, believing that many would react by going underground or absconding, making it even more difficult for officials to supervise and control them. Informants expressed concern that sex-offender residential policy was frequently developed without adequate consideration of the budget impact on agencies responsible for policy implementation and compliance. Additionally, some concern was expressed over the displacement of organizational missions that occurs when policy focus narrows to residential restrictions instead of broadening to consider positive steps toward solving the sex-offender housing problem. We compiled the following list of specific recommendations from the key informants’ interviews for consideration by policymakers. 1. Develop a community education program to better inform the public about the nature of the sex-offender residential problem emphasizing the differences in levels and offenses and the risks associated with these levels. 2. Accurately inform the public about the potential negative consequences of restrictive housing policies that produce absconders and make it more difficult to locate and control sex offenders. 3. Explore, implement, and test pilot programs using state-of-the-art technology to speed up notification, to reduce the cost of notification, and to track more serious offenders. 110 4. Review, design, and improve integrated databases that increase the accuracy of offender address information and the proximity of child-centered facilities to those addresses. 5. Conduct budget impact studies for proposed policies (new or modifications), and make provision for adequate funding of policy implementation. 6. Review state laws and policies to determine the feasibility of expanding the use of state facilities to house Phoenix sex offenders. 7. Form an interagency non-profit task force to initiate dialogue about the sexoffender residency problem; charge it with developing affirmative housing opportunities for sex offenders. 111 6 A Geographic Analysis of Sex-offender Clustering and Proposed Distancing Requirements Constance Kostelac, M.S. Aimee Cardenas Karen Kontak, M.A. A primary function of the Sex Offender Cluster Study was to understand residential housing patterns of registered sex offenders in the city of Phoenix. This included an examination of current housing patterns and clustering of offenders, as well as an analysis of the feasibility of mandating limits on the proximity of sex-offender residences to key facilities such as schools, child-care centers, and parks. The goal of restricting the proximity of offenders’ residences to such locations rests on the assumption that if the area where offenders lived provided direct access to potential victims, or if they lived close to other offenders who might encourage the type of behavior that leads to future offenses, then sex offenders would be more likely to commit sex crimes again. In addition, regulation of offender housing responds to the concern that a high number of offenders in a particular area not only might commit more offenses, but also might cause high levels of fear among those living near them. Finally, sex-offender clustering is related to the perceived disproportionate impact on neighborhoods with a high number of sex offenders in terms of fear, lowered property values, crime, and related issues. Although the focus of this analysis was not specifically on the re-offending patterns of registered sex offenders, some consideration was given to the relationship between sex offender housing patterns and reported crime in Phoenix. Methodology In order to understand the feasibility of various proposed distancing requirements in the city of Phoenix, as well as the current patterns of residential housing for registered sex offenders, we used geographic information system (GIS) tools using ESRI’s ArcGIS software. We gathered various data sets from a variety of sources including the Phoenix Police Department (PPD), the Phoenix Information Technology Department (ITD) for various GIS layers, the InfoUSA database for locations of child-care centers, and the SexTrak information maintained by PPD’s Family Investigations Bureau for the address information for registered sex offenders residing in Phoenix. The specific methodology used for each section is discussed in more detail below. 112 Residency Restrictions from Key Facilities States or local municipalities are increasingly acting to regulate the distance that registered sex offenders may live from key social facilities such as schools, child-care centers, and parks. The object of such residency limits rests in part on the assumption that released offenders are more likely to commit further sex crimes if they live in close proximity to one of these locations. Neither Phoenix nor Arizona now has a distancing requirement, but recent discussions have centered on new legislation or city codes requiring that sex offenders not reside within 1,000 or 2,000 feet of a school, child-care center, or park. The belief is that this will limit sex offenders’ exposure to children, which in turn will help to decrease the likelihood of recidivism. To begin the analysis, we obtained from ITD information on school locations that included both physical addresses and parcel areas. The analysis was limited to schools physically located within or on Phoenix city boundaries. A total of 491 school locations with an associated 485 parcels were included. School locations were geo-coded primarily from a street centerline file, so the specific locations could differ slightly when geo-coded to the parcel. Also, only schools with available information that was tracked in the ITD file were included. Some locations showed up more than once when multiple types of facilities (e.g., an elementary and a middle school) were located on the same site. The file of child-care center locations was created based on InfoUSA data. This information may not be a complete representation of all child-care locations, particularly those that are home-based or not licensed, since the data set was based on business licenses and contained only available data. We found 795 addresses, and selected 689 that intersected the city boundary. These locations were then geo-coded to the street centerline file with a 93% geo-coding rate. At the time of the study, parcel information on the child-care center locations was not available. In some cases, child-care centers and schools might overlap if both kinds of facilities were present at the same location. Finally, we obtained the physical location of park boundaries for the various types of city parks from ITD. (This information did not account for non-city parks or playground areas such as those that some residential associations maintain in their neighborhoods.) We mapped all of these locations and created 1,000-foot and 2,000-foot buffers around each address or parcel. Maps 6.1 through 6.6 show the locations of the individual types of facilities and their associated 2,000-foot buffers. (The 2,000-foot distance, most commonly used in other jurisdictions, has been the focus of the distancing discussion in Phoenix.) We then considered the remaining area to be available for sex offenders to reside. We also combined the buffered areas for all three kinds of facilities (schools, child-care centers, and parks) to get a picture of the amount of land city-wide that would be unavailable to sex offenders for residential housing (see Map 6.6). We could have further refined the analysis by separating the remaining area by land-use type, removing non-residential areas (current or planned) from the “available” area. However, challenges with the land-use data prohibited us from incorporating this information into the report. A complete file would be necessary to determine the actual residential area remaining after the various distance requirements are imposed. 113 Table 6.1 shows the area remaining within city boundaries after the various buffers were put into place around the protected locations, as well as the area remaining when land use was considered. The city boundary for Phoenix was approximately 515 square miles, which reduced to 478 square miles when the mountain areas were removed. The 478square-mile area was used for the calculations below, but other undeveloped or desert areas within the city were not excluded from the analysis. When all three types of locations were combined, 231 square miles or 48% of the current city area was shown to be off-limits to registered sex offenders, regardless of land-use type.24 The 2,000-foot buffer regulations would make it difficult for registered sex offenders to find housing that was not in violation of at least one distancing requirement within the city boundary. If not otherwise prohibited, this would likely cause offenders by necessity to cluster in whatever areas were available. Policymakers could consider alleviating this problem by imposing different distancing restrictions on offenders based on their offenses and risk levels. For example, offenders who had committed crimes against children could be subject to distancing requirements related to schools and child-care facilities, while other kinds of offenders might not. Of course, this would increase the complexity of managing and monitoring offenders and compliance with their applicable distancing requirements. (See Table 6.1 and Maps 6.1 through 6.6.) Table 6.1 Analysis of Geographic Impact of Distancing Requirements from Key Community Facilities Location type Locations (#) Square miles in Area of city Area of city 2000-ft. buffers removed (%) remaining (%) School locations 491 145 30 70 School parcels 485 186 39 61 Child-care centers 689 165 35 65 Parks 154 106 22 78 School and child258 buffers 195 41 59 care centers School, child-care 600 buffers 231 48 52 centers, and parks 24 Some jurisdictions consider bus or other public transportation stops as key community facilities to be protected by distancing requirements. Bus stops were not included in this analysis. They are present in the city approximately every mile along the main streets; clearly, imposing distance limits from bus stops would render much of the city unlivable for registered sex offenders. 114 Map 6.1 Locations of Schools with 2000-ft. Buffer 115 Map 6.2 Locations of School Parcels with 2000-ft. Buffer 116 Map 6.3 Locations of Child-care Centers with 2,000-ft. Buffer 117 Map 6.4 Location of Parks with 2,000-ft. Buffer 118 Map 6.5 Locations of Child-care Centers and Schools with 2,000-ft. Buffer 119 Map 6.6 Locations of Child-care Centers, Schools, and Parks with 2,000-ft. Buffer 120 Residency Restrictions from Other Offenders (Anti-clustering) Current Arizona anti-clustering statutes apply only to sex offenders on probation. The statutes limit the number and/or percentage of level-2 and level-3 offenders who can reside at a particular location, including multi-family housing. However, the discussion of distancing requirements in Phoenix has included the potential for regulating where offenders may live in relation to one another. Offenders were not, at the time of this study, distributed equally across the city. Rather, they tended to be grouped into particular geographical areas, where a significant number lived in transitional housing locations (e.g., shelters, hotel/motels, treatment facilities, and related). This tended to cluster offenders in particular areas of the city. (A more detailed discussion of the clustering pattern in Phoenix is included below.) The debate has centered in Phoenix on the possibility of dispersing clusters by restricting how far offenders must live from one another. Distributing offender residences across the city is generally intended to reduce the impact on individual neighborhoods. This section of the analysis considered the impact if a distancing requirement for sex offenders similar to the one related to schools, child-care centers, and parks were to be applied with respect to other offenders. When we collected the data, approximately 930 active level-2 (n=490) and level-3 sex offenders (n=440) were living in Phoenix. We calculated the area of the circle for a 2,000-foot residential buffer to be 0.45 miles. To determine the overall area that would be covered by the buffers, this area was multiplied by the number of offenders. We then subtracted that area from the total area of the city (approximately 515 square miles, reduced to 478 square miles after mountains are removed). As shown in Table 6.2, if only level-3 offenders were required to live at least 2,000 feet from one another, 41% of the land in the city would be taken up by the residences of these offenders. Level-2 offenders would take up approximately 46% of the city area. If distancing requirements were imposed both on level-2 and level-3 offenders, the buffers surrounding their locations would encompass approximately 88% of the entire area of the city, meaning that any additional offenders would only have 12% of the city available for finding residences. Of course, this is only a numeric calculation, and it leaves out a number of important considerations for how such distancing regulations would actually be applied. First, our calculation of the area of the city taken up by offenders subject to distancing assumed that they would be evenly distributed. This is not likely to be the case, in large part due to the way housing is distributed across the city. In addition, assuming the distance would be measured from one location to another, they might not fit the circular pattern that was used in the buffer measurement. This particularly would be the case if the measurement included the parcel rather than the physical address of the property. In some cases, two offenders could be living in the same large apartment complex, residing 2,000 feet apart at opposite ends of the complex. In addition, the calculation did not account for how much of the land in the city contained residential housing. As described in an earlier section, it would be necessary to consider the area of the city that was zoned residential to have a complete picture of the impact of 121 the distancing requirement. A final consideration is that this calculation was based on the number of offenders at a particular moment in time, but that number fluctuates almost daily. It has the potential to increase over time, making distancing requirements even more difficult to comply with. Table 6.2 Analysis of Geographic Impact of Distancing Requirements For Active Level-2 and Level-3 Registered Sex Offenders Level Buffer Offenders Sq. mi. within Area of city Area of city distance (#)* buffer areas removed (%) remaining (%) (ft.) 2/3 2000 930 418.5 88 12 2 2000 490 220.5 46 54 3 2000 440 198.0 41 59 Note. Number of offenders is approximate since it has the potential to change on a daily basis. Analysis of the Relationship Between Sex-offender Clustering and Crime One concern with sex-offender clustering is its potential impact on crime and safety in neighborhoods and communities. Various aspects of this issue from the perspective of citizens, offenders, and key stakeholders are discussed elsewhere in this report. However, we are including a brief discussion of the geographic relationship between these two concerns here. Analysis of the geographic correlation between the residential housing patterns or clustering of registered sex offenders in Phoenix and the density of crime was not conducted for a variety of reasons. First, the density of crime and the clustering of sex offenders could be related for reasons not directly tied to the actions of the offenders themselves. As discussed in the offender interview section of this report, finding affordable housing was a major challenge for registered sex offenders, particularly those re-entering the community after a period of incarceration. With registration and notification requirements, the residential options for offenders were limited. In Phoenix, at any given time, a large number of offenders are registered at some form of transitional housing (e.g., shelter, hotel/motel, treatment facility) or as transients. A sample of the distribution of level-2 and -3 offenders in Phoenix is shown in maps 6.7 and 6.8. Map 6.7 demonstrates the relative density of level-2 and level-3 offenders across the city. Although some offenders lived in most areas throughout Phoenix, the highest concentration was in the center of the downtown corridor. The second map depicts locations where more than one offender had listed the same physical address; these are often shelters, treatment facilities, hotels or motels, or apartment buildings. As shown in the maps, at the time we collected data for this project, nearly 300 active, registered level2 and level-3 offenders in Phoenix listed addresses that were considered transitional housing situations. Therefore, the clustering pattern of offenders was in part related to where they could live under the current conditions. 122 Unfortunately, many of the clusters of sex offenders, particularly in transitional housing, corresponded with areas that had overall higher crime levels. Although this suggested that statistically, high densities of sex offenders were related to high densities of crime, this could have had little to do with the presence of the sex offenders. These areas tended to have a variety of other challenges and factors that are associated with higher levels of crime (e.g., high levels of poverty, unemployment, resident mobility and turnover, and lower educational levels). In other words, the presence of criminogenic factors may have been related to both the higher levels of crime in the area and to the density of registered sex offenders living there. There is also an issue of timing: Since offenders can and do move frequently, any analysis of sex-offender residential housing patterns in relation to crime would be a snapshot in time. Unfortunately, information is not currently maintained on the various addresses for a particular offender over time. We recommend that this information be collected; in the future, it could be possible to conduct additional analyses on sex-offender housing patterns and crime, focusing on offender movement into and out of areas and tracking crime patterns at a more individual level. In addition to concerns with the analysis itself, as discussed elsewhere in this report, the address issues encountered during data collection for sex offender interviews pointed to potential limitations with an in-depth analysis based on known offender addresses. It also raised the question of whether more restrictive distancing requirements would increase the complexity of tracking these offenders and reduce their likelihood of complying with registration requirements. 123 Map 6.7 Density of Level-2 and Level-3 Sex Offenders 124 Map 6.8 Repeat Address Locations for Level-2 and Level-3 Sex Offenders 125 Additional Considerations Several decisions have yet to be made with respect to the specific application of distancing requirements for registered sex offenders in Phoenix. To begin with, it has yet to be determined whether the distance from a school (or child-care center or park) refers only to the address or to the parcel itself. For entities on large parcels of land, using only the address could allow sex offenders to live directly across the street from the edge of the park or school because the address point is on the other side of the parcel. Discussions regarding the appropriate distancing requirement (2,000 feet vs. 1,000 feet) for regulating where offenders can live in relation to one another, as well as from key community facilities, are also underway. Another unresolved consideration is the concept of “grandfathering.” If new anticlustering regulations were enacted, sex offenders could find themselves suddenly in violation, regardless of how long they had been living in their current residences. It has yet to be determined whether, for example, a sex offender who has resided in the same home for many years, but is in violation based on new regulations, would be forced to move or would be grandfathered. A related issue is deciding which offender would be asked to move when more than one lived in the same residential neighborhood, should new anti-clustering restrictions be imposed. Enacting new anti-clustering regulations would also raise issues related to managing and verifying sex-offender data and the operational needs of law enforcement agencies tasked with offender management. Enforcing anti-clustering distance regulations would require modifying the current methods of tracking offenders, a fairly mobile and transient population. Funding and manpower are significant issues in this area. In particular, a procedure for determining actual distances from schools, child-care centers, and parks would need to be developed; what constitutes a school, child-care center, and park would need to be agreed upon; and regulations for enforcement with offenders who were not in compliance would need to be established. Determining the distance from a school, child-care center, or park seems a straightforward task in theory, but in order to actually enforce distance requirements, numerous details would need clarification. In particular, the city code or legislation would need to specify whether such measurement should be from the street address, the edge of the parcel, or the midpoint of the parcel, and whether Euclidean distance (as the crow flies) or Manhattan distance (street segment) should be used. If Euclidean distance were used, methods for completing the measurements would need to be established; one cannot physically measure through residents’ backyards and other obstructions, so GIS or related tools would be needed. If Manhattan distance were used, multiple paths along street segments would need to be measured to ensure that the shortest path between the offender’s residence and the school, child-care facility, or park would meet the distance requirement; whenever new streets were developed, distances would need to be remeasured. If parcel boundaries were used and a parcel was an odd shape, distances from multiple edges would need to be taken to determine the shortest distance. Detailed parcel maps and clear identification of parcel boundaries would be crucial for 126 ensuring accurate distance measurements. Further, it would need to be determined which criminal justice agency (local, county, or state law enforcement, probation or parole, etc.) would be responsible for completing and updating the measurements; significant numbers of new staff would be required to handle this task. In addition to establishing an agreed-upon method for determining distance measurements, the city would also need to define what constitutes a school, child-care center, and park. Openings and closings of these facilities and new developments of park areas and demolitions or changes in land use of existing park lands would all need to be tracked. Again, theoretically, this seems like an easy task, but there would be complications. For example, if in-home child-care centers were included, it would be extremely challenging to track openings and closings of such sites, and a definitive description of what constitutes a child-care center would need to be established. Similarly, it would need to be determined whether only city parks were to be protected or if common play areas in residential communities (and similar types of areas) would also qualify. Even with clear definitions of what constitutes a school, child-care center, and park, and with unambiguous, established guidelines for completing distance measurements, questions remain to be answered regarding how actual enforcement would be handled when offenders are not in compliance. As mentioned earlier, it would need to be decided whether or not to grandfather offenders already living in an area who are determined to be out of compliance once new distance requirements come into effect. For example, what happens if a new school, child-care center, or park is opened within 2,000 feet of a sex offender’s existing residence? Will that school, child-care center, or park be allowed to open or would the offender be required to move? If so, would there be mandates about notifying people who may attend the school, child-care center, or park? Also, if two or more offenders were found to be residing too close to each other (and neither one qualifies for grandfathering), an agreed-upon procedure for legally determining which offender may stay and which must go would need to be established before any forced relocation could occur. Further, if relocation of an offender required the sale of his/her home, how long would he or she have to complete the sale? What would happen if the home did not sell? Procedures for handling the variety of possible situations would need to be outlined in detail before enforcement would be possible. Finally, as with the need for additional staff to handle distance measurements, it would also be necessary to allow for extra personnel to handle the enforcement of new policies regulating distances; the number of staff needed to handle the increased work load would need to be determined. 127 Summary In summary, this analysis examined the geographic feasibility of mandating a 2,000-foot buffer that would limit where level-2 and level-3 sex offenders could reside in relation to schools, child-care centers, and parks in Phoenix, Arizona. We also examined the feasibility of mandating a 2,000-foot buffer that would determine how close sex offenders could reside to other sex offenders, and we discussed some of the methodological issues that would need to be addressed before these types of mandates could be implemented. In particular, the amount of remaining available area for potential residency was examined after restricting a 2,000-foot area around schools, child-care centers, and parks. We determined that with this restriction, nearly half of the city’s area would become offlimits for sex offenders. This was the result even before taking into account what portion of the remaining area was residential in character. Similar, and even more striking, is what would happen if sex offenders were not allowed to cluster, if they were required to live at least 2,000 feet from any other sex offender. In this situation, to accommodate the number of offenders registered in Phoenix and still meet the distance requirement, more than 80% of the city’s area would be needed; this would force any new offenders to find residences within only the remaining 20% of the city. Thus, if sex offenders were prohibited from living closer than 2,000 feet from one another, the vast majority of the city’s neighborhoods would have a resident sex offender, and any new offenders would be limited to living in the remaining areas of the city, whether or not those areas were affordable. Further, if residential distance restrictions pertaining to schools, child-care centers, and parks and distance restrictions pertaining to other offenders were implemented together, there would be no feasible way to disperse the current registered sex offender population throughout the city. This raises strong concerns about the possible unintentional consequences of adopting such mandates. It could lead to offenders choosing not to register and to abscond, just in order to find a residence in Phoenix. Considering the outstanding issues with registration compliance and obtaining and updating accurate address information, extreme care needs to be taken before these types of policy changes are made to ensure that they do not increase the tendency for offenders to go underground. Even if implementing distance requirements were geographically feasible, many logistical issues would need to be defined and addressed to enable enforcement of such mandates. In particular, a method for determining actual distances between locations would need to be specified; definitions of what constitutes a school, child-care center, and park would need to be agreed upon; and a reliable method for tracking openings and closings of these facilities would be needed. In addition, officials have several decisions to make: whether or not to grandfather offenders already living in an area, what the legal procedure would be for determining the length of time at a residence in order to identify which offenders are out of compliance and must relocate, how much time offenders are 128 allowed for complying, and how much staffing levels must be increased in the agency responsible for enforcement. 129