Contents | Feedback | Search | DRCNet Home Page | Join DRCNet
DRCNet Library | Schaffer Library | Government Publications | GAO Publications
(Letter Report, 06/19/95, GAO/GGD-95-139). Pursuant to a legislative requirement, GAO reviewed juveniles' access to counsel in state and local juvenile delinquency proceedings, focusing on: (1) juveniles' right to counsel under 15 states' laws; (2) the frequency with which juveniles have counsel in juvenile courts in 3 states; (3) the likely impact of counsel on juvenile justice outcomes; (4) juveniles' access to counsel in adult courts; and (5) the quality of counsel juveniles receive. GAO found that: (1) the 15 states reviewed had laws guaranteeing juveniles' right to counsel in delinquency proceedings which included provisions for indigent juveniles; (2) 11 states specified the circumstances under which juveniles could waive their right to counsel and in 3 states, juveniles could waive their right to counsel although statutes did not specifically provide for such waivers; (3) juvenile representation rates varied between 97 percent to 65 percent and between metropolitan and nonmetropolitan areas; (4) the impact of counsel on case outcomes varied by the state and offense, but in 2 states, juveniles without representation were less likely to receive out-of-home placements; (5) differences in case outcomes between represented and nonrepresented juveniles were due more to other variables than to juveniles' representation status; (6) although there was no existing data on juveniles' access to counsel in adult courts, state laws provide for such representation; and (7) prosecutors and juvenile justice officials were generally satisfied with the quality of counsel juvenile offenders received and believed that public defenders were at least as capable as private attorneys, although they had excessive caseloads and limited resources. --------------------------- Indexing Terms ----------------------------- REPORTNUM: GGD-95-139 TITLE: Juvenile Justice: Representation Rates Varied As Did Counsel's Impact on Court Outcomes DATE: 06/19/95 SUBJECT: Juvenile status offenders Courtroom proceedings State law Right to counsel Right to due process Lawyers Juvenile delinquency IDENTIFIER: California Pennsylvania Nebraska Arizona Florida Maryland Missouri South Carolina Utah Idaho Kansas Louisiana New Mexico New York Texas Ohio ************************************************************************** * This file contains an ASCII representation of the text of a GAO * * report. Delineations within the text indicating chapter titles, * * headings, and bullets are preserved. Major divisions and subdivisions * * of the text, such as Chapters, Sections, and Appendixes, are * * identified by double and single lines. The numbers on the right end * * of these lines indicate the position of each of the subsections in the * * document outline. These numbers do NOT correspond with the page * * numbers of the printed product. * * * * No attempt has been made to display graphic images, although figure * * captions are reproduced. Tables are included, but may not resemble * * those in the printed version. * * * * A printed copy of this report may be obtained from the GAO Document * * Distribution Facility by calling (202) 512-6000, by faxing your * * request to (301) 258-4066, or by writing to P.O. Box 6015, * * Gaithersburg, MD 20884-6015. We are unable to accept electronic orders * * for printed documents at this time. * ************************************************************************** Cover ================================================================ COVER Report to the Committee on the Judiciary, U.S. Senate, and the Committee on Economic and Educational Opportunity, House of Representatives June 1995 JUVENILE JUSTICE - REPRESENTATION RATES VARIED AS DID COUNSEL'S IMPACT ON COURT OUTCOMES GAO/GGD-95-139 Juvenile Justice Abbreviations =============================================================== ABBREV DHRS - Department of Health and Rehabilitation Services DJJ - Department of Juvenile Justice DJS - Department of Juvenile Services MSA - Metropolitan Statistical Area NJCDA - National Juvenile Court Data Archive NCJJ - National Center for Juvenile Justice OJJDP - Department of Justice's Office of Juvenile Justice and Delinquency Prevention PMSA - Primary Metropolitan Statistical Area SAS - Statistical Analysis System Letter =============================================================== LETTER B-259801 June 19, 1995 The Honorable Orrin G. Hatch Chairman The Honorable Joseph R. Biden, Jr. Ranking Minority Member Committee on the Judiciary United States Senate The Honorable William F. Goodling Chairman The Honorable William L. Clay Ranking Minority Member Committee on Economic and Educational Opportunity House of Representatives The 1992 reauthorization (P.L. 102-586) of the Juvenile Justice and Delinquency Prevention Act of 1974 (P.L. 93-415) mandated that we study access to counsel in state and local juvenile delinquency proceedings. Some legal organizations and scholars had raised concern about the access to counsel afforded to juveniles in juvenile court proceedings. For example, the American Bar Association; the Consortium of Children, Families and the Law; and individual law professors testified in September 1992 to Congress that half of all juveniles in the United States waive their constitutionally guaranteed right to counsel without speaking to an attorney. To meet our legislative requirements, we agreed with your committees to review state laws for 15 states that we selected to determine juveniles' right to counsel, determine the frequency with which juveniles have counsel in juvenile courts in three states, determine the likely impact of counsel on juvenile justice outcomes, determine if juveniles who are in adult court have counsel, and develop insights regarding the quality of counsel. RESULTS IN BRIEF ------------------------------------------------------------ Letter :1 Statutes guaranteeing juveniles' right to counsel in delinquency proceedings were present in the 15 states whose laws we examined. For juveniles who could not provide counsel on their own, the states had provisions to provide and compensate counsel for them. Of the 15 states, 11 had laws allowing the waiver of counsel under certain circumstances but generally had rules to ensure that waivers were made only when juveniles were aware of their right and voluntarily gave up that right. According to officials in three other states, juveniles can waive counsel even though the state statutes do not specifically address the waiver issue. In the remaining state, juveniles could not waive counsel. We could find no national data regarding juveniles' representation by counsel. However, the National Center for Juvenile Justice (NCJJ)\1 had such data for some jurisdictions, and we analyzed its data for available counties in California, Pennsylvania, and Nebraska for portions of 1990 and 1991. The results only relate to the jurisdictions for which we were able to obtain data and therefore cannot be generalized to other jurisdictions within these states. Throughout this report, our references to the states apply only to the counties in the states for which we have data. Overall representation rates varied widely, from about 97 and about 91 percent, respectively, in California and Pennsylvania to about 65 percent in Nebraska, and within the states the rates varied between metropolitan and nonmetropolitan areas.\2 Representation rates by offense category (e.g., violent crimes, property crimes, drug crimes, etc.) varied, with minimal variation in California and Pennsylvania (about 83 to almost 100 percent) across all categories and a wider variation across categories in Nebraska (about 54 to 91 percent); however, 76 percent of all unrepresented cases in Nebraska were for less serious property offenses. Overall impact of representation on case outcomes varied according to the state and the offense category. Sometimes juveniles with representation were more likely to have their cases adjudicated as delinquent (i.e., judged to be a delinquent)\3 and sometimes those without representation were more likely to be adjudicated as delinquent. In most cases, juveniles without representation were less likely to receive out-of-home placements (e.g., training school). According to our logistic regression\4 models for California, Nebraska, and Pennsylvania, unrepresented juveniles were generally about as likely to have their cases adjudicated than represented juveniles, but characteristics other than representation (e.g., detention prior to adjudication and prior offense history) were more strongly associated with the likelihood of adjudication. According to our models for placement outcomes for the three states, juveniles' characteristics other than representation (e.g., detention prior to adjudication or prior offense history) were more strongly associated with placement decisions. We could not locate any databases to determine if juveniles in adult court had counsel or to compare access to counsel for juveniles in adult and juvenile court. However, on the basis of our review of criminal law, juveniles in adult court have the same right to counsel as adults. In addition, our survey of prosecutors indicated that juveniles in adult and juvenile court were given the same opportunity as adults to be represented. (We could not determine the extent to which juveniles waived their right to counsel in adult court.) Prosecutors and juvenile justice officials in the eight local jurisdictions we visited were generally favorable concerning the quality of counsel provided to juveniles. They also reported that public defenders were seen to be at least as capable as private attorneys in representing juveniles, and several officials noted that public defenders were more knowledgeable in the laws and rules governing juvenile proceedings. However, these officials noted that while public defenders had the most experience in juvenile law and procedures, they also had very high caseloads. These officials raised concerns over the limited resources and growing caseloads of public defenders. -------------------- \1 NCJJ, located in Pittsburgh, Pennsylvania, is the research division of the National Council of Juvenile and Family Court Judges. \2 NCJJ data from California contained only metropolitan counties. \3 A delinquent is a youth who has been charged with or had a case adjudicated for criminal conduct. \4 Logistic regression analysis is a widely accepted statistical methodology used when the dependent variable is qualitative, such as if a juvenile is adjudicated as delinquent. Regression analysis identifies relationships between the dependent variable and two or more key variables, such as the use of counsel and the juvenile offender's current offense category and prior offense history and supervisory status. BACKGROUND ------------------------------------------------------------ Letter :2 Juvenile justice is primarily the responsibility of state and local authorities because juvenile courts derive their authority from state legislatures. The juvenile courts of all 50 states are part of their states' judicial system but are generally county or city based and vary in structure, policy, and procedure. Thus, juvenile courts' jurisdiction and procedures vary widely across and within states. Appendix I illustrates differences in profiles of the juvenile justice processes in nine states.\5 -------------------- \5 Arizona, California, Florida, Maryland, Missouri, Nebraska, Pennsylvania, South Carolina, and Utah. MAJOR ELEMENTS OF JUVENILE COURT PROCEDURE ---------------------------------------------------------- Letter :2.1 When juveniles commit crimes, they are usually subject to the jurisdiction of the juvenile court as a delinquent offender. Referrals of delinquents come to juvenile court from many sources, including parents, police officers, victims, and schools. Law enforcement agencies provide the vast majority of referrals to juvenile courts. The Department of Justice's Office of Juvenile Justice and Delinquency Prevention (OJJDP) reported in 1991 that law enforcement agencies referred 85 percent of all delinquency cases to the courts. Referrals generally come to the court's intake department to be screened by intake staff.\6 The intake staff examine the referral to determine what type of crime is alleged to have been committed and how to proceed with the handling of the referral. The intake decision process has several potential outcomes. Among other things, intake staff may decide to dismiss the case for insufficient legal evidence or to resolve the matter informally out of court.\7 These informal dispositions may include diverting the juvenile to a social agency for services,\8 informal probation, or the payment of fines or some form of restitution. If intake staff decide that the referral should be handled formally, a petition is drafted and filed to provide notice of the issues that will be adjudicated. The petition charges the juvenile with a delinquency offense violation and requests an adjudicatory or waiver hearing. In some states, petitions may be dismissed before the actual adjudicatory hearing\9 is held for various reasons, such as when the case is not viewed by court officials as a particularly serious offense. For example, in Pennsylvania, less serious petitioned cases may be disposed of by consent decrees--informal adjustments\10 without an adjudicatory hearing. At an adjudicatory hearing, the juvenile may have his or her case adjudicated, i.e., judged as delinquent and, if so, the case proceeds to a disposition hearing.\11 The judge is responsible for presiding over the hearing, including determining placement of the juvenile during the pending proceedings, making any interim orders that are necessary to the conduct of the case, and deciding the disposition of the case if there has been a finding of delinquency. The range of disposition options available to the juvenile court judge generally includes (1) commitment to an institution; (2) placement in a group home or other residential facility; (3) probation; (4) referral to an outside agency or mental health program; or (5) imposition of a fine, community service, or restitution order. A juvenile may be temporarily placed in a detention facility during various stages of a case as it progresses through the juvenile justice system. Generally, juveniles alleged to be delinquent may be placed in secure custody in a physically restrictive facility. Depending on the state's detention laws, the juvenile may be detained to protect the community, the juvenile, or both. In addition, detention may be used to ensure the juvenile's appearance at a hearing or while the juvenile is awaiting long-term placement in another facility. A judicial decision to detain or continue detention may occur before or after adjudication or disposition. OJJDP estimated that in 1991 juveniles were held in detention facilities between the time of referral and case disposition in about 20 percent of the 1.3 million delinquency cases.\12 -------------------- \6 We obtained the information in this section from various sources, such as documents and discussions with juvenile justice officials. \7 Besides insufficient legal evidence, a referral may be dismissed because (1) the offense is petty or seen as a low risk, (2) the referral is a first offense, and the level of its seriousness does not merit formal consideration, (3) the juvenile and his/her family have reimbursed a victim for damages, (4) formal processing is seen as unnecessary, (5) or because family strengths are prominent and the juvenile is perceived to be responsive to parental controls and discipline. \8 Diversion programs were designed to remove first-time offenders and relatively minor offenders from the formal adjudication process. To meet the treatment needs of those juveniles, most of whom have not faced previous juvenile court adjudication, a variety of community-based programs and agencies have been created. \9 A hearing where a judge presides over a juvenile court proceeding to determine if the juvenile actually committed the alleged offense. \10 The juvenile and the court agree to suspend the case with the juvenile remaining at home under probation conditions for up to 6 months. \11 Juveniles can also be adjudicated as status offenders. Juveniles who have come in contact with the juvenile justice system by committing an offense (such as running away or truancy) that would not be a crime if committed by an adult are status offenders. Our review focused on delinquents and did not include status offenders. For more information on status offenders, see our report entitled Noncriminal Juveniles: Detentions Have Been Reduced but Better Monitoring Is Needed (GAO/GGD-91-65, Apr. 24, 1991). \12 At the time of our review, 1991 was the most recent data available. ROLE OF COUNSEL IN JUVENILE COURT ---------------------------------------------------------- Letter :2.2 Historically, juvenile delinquents subject to the jurisdiction of juvenile court were not afforded the same constitutional rights guaranteed to adult criminal defendants. However, in 1967 the U.S. Supreme Court held that juvenile offenders were entitled to the assistance of counsel in certain juvenile delinquency proceedings, In re Gault, 387 U.S. 1 (1967).\13 In Gault, the Supreme Court concluded that the due process clause of the Fourteenth Amendment requires that in proceedings to determine delinquency (e.g., adjudication hearings) that may result in commitment to an institution in which the juvenile's freedom is curtailed, the juvenile and his or her parents must be notified of the juvenile's right to be represented by counsel. The Supreme Court also indicated that if the juvenile and his or her parents could not afford counsel, counsel must be appointed to represent the juvenile. With regard to the waiver of counsel, the Supreme Court indicated in Gault that waiver must be an "intentional relinquishment or abandonment of a fully known right." The Supreme Court also stated that the juvenile and his or her parents have a right expressly to be advised that they may retain counsel. If they are unable to afford counsel, they are entitled, considering the seriousness of the charge and the potential commitment, to be appointed counsel. However, they can choose to waive counsel. Currently, there are three major approaches to the provision of defense counsel in juvenile courts: public defenders, which include private attorneys or law firms under contract with or granted funds by a governmental entity to defend juveniles; private agencies, such as legal aid societies under similar contracts or grants; and government defender agencies; private attorneys appointed by the court to represent individual juveniles and paid from public funds; and privately paid attorneys. -------------------- \13 In an earlier case, the U.S. Supreme Court indicated that a juvenile was entitled to the assistance of counsel in a proceeding to determine whether the juvenile's case should be waived to the jurisdiction of the criminal court. Kent v. United States, 383 U.S. 541, 561-62 (1966). SCOPE AND METHODOLOGY ------------------------------------------------------------ Letter :3 To aid us in identifying the issues regarding access to counsel, we reviewed relevant literature in bibliographies provided by NCJJ and OJJDP. We developed information on juveniles' rights and access to counsel through a combination of methods, including (1) an analysis of statutes, state administrative procedures, and case law in 15 states;\14 (2) an analysis of NCJJ statistical data on 3 states;\15 (3) nationwide surveys of county prosecutors and public defenders; (4) telephone interviews with selected state and local judges in Florida, Louisiana, Michigan, North Carolina, Ohio, Pennsylvania, Texas, and Utah; and (5) visits to four states--Florida, Louisiana, Pennsylvania, and Utah--to meet with juvenile justice officials to understand the processes and frequency of when counsel was provided. We completed our statistical analysis of NCJJ data in two parts. First, using seven general crime categories (see table 1.), we developed statistics on case processing outcomes for certain counties in California, Nebraska, and Pennsylvania for calendar years 1990 and 1991.\16 Specifically, we reported the percentages of juveniles petitioned to court who received counsel. Second, we computed and analyzed the probabilities that juveniles with counsel and without counsel were adjudicated or placed. We analyzed these two decision points because the Gault ruling focused on these key decisions in which juveniles were at risk of losing their liberty. Juvenile facilities are not all the same. For example, facilities vary in degrees of security. Data were not readily available to differentiate between types of placement juveniles received or the security of the facility. Therefore, our analysis only recognized that the juveniles were either placed or not placed. Also, the data identified represented and unrepresented juveniles who were adjudicated as delinquents. However, data were not readily available to identify those who were represented but during the process had their charges reduced (e.g., from aggravated to simple assault). Therefore, our analysis of the data may not recognize the impact of being represented in terms of impact on reduction in charges. Table 1 Crime Categories and Types Crime categories Crime types ------------------ ---------------------------------------- Violent crimes Murder and nonnegiligent manslaughter; rape; robbery; aggravated assault; other violent crimes, which include negligent manslaughter, sexual assault, kidnapping, and others but excludes simple assault; and indeterminate violent offenses (i.e., the crime was a violent crime, but the type of violent crime was not known). Property crimes Burglary, fraud, forgery, and embezzlement; larceny; motor vehicle theft; unspecified larceny; other property (e.g., trespassing and vandalism); indeterminate property (i.e., the crime was a property offense, but the type of property offense was not known). Simple assault Generally an attack without a weapon resulting in minor injury. Drug crimes Possession, trafficking, and other or unspecified drug crime. Weapons Unlawful sale, manufacture, transportation, distribution, or possession. Public order or Disorderly conduct, escape, obstruction nonviolent crimes of justice, and other public order crimes (e.g., prostitution). Indeterminate Crime type not known. ------------------------------------------------------------ Note: Definitions of crime types included in the six broader crime categories may vary across states. Further, certain specific crime types are excluded from juvenile court jurisdiction in some states, e.g., murder in Pennsylvania. Source: Derived from Uniform Crime Report crime classifications, NCJJ offense categories, and state definitions of crimes. Because case characteristics, in addition to representation, may affect case outcomes, we developed logistic regression models to control for these characteristics. We looked at the relationships between the type of counsel (i.e., public defender, court-appointed attorney, or private attorney) in juvenile delinquency proceedings and key decision points--adjudicated and placed--in the juvenile judicial process controlling for a number of relevant variables. We estimated the models separately for the 3 states for up to 10 crime types yielding 24 adjudication and 24 placement models. The variables we controlled for in our logistic regression models measured three types of case characteristics: (1) demographic variables, such as age, race, and gender; (2) offense-related variables, such as current crime type, the severity of the current crime type, and offense history; and (3) juvenile court-status variables, such as detention prior to adjudication; source of referral; informal adjustments prior to adjudication; representation by counsel and the type of counsel representing the case, location of the courts, and length of time to dispose of a case. See appendix III for a discussion on how we classified crime types and the variables used in the models. Also, to better understand the underlying processes, we developed profiles of the juvenile justice systems in nine states. We coordinated our work with OJJDP, NCJJ, the American Bar Association, the National Legal Aid Defenders' Association, and the Coalition for Juvenile Justice. Appendix II provides a more detailed discussion of our objectives, scope, and methodology, including the basis that we used to select states and identify officials. The details of the modeling procedures as well as of the results are discussed in appendix III. The data we developed, including the results of our interviews with juvenile justice officials, are not projectable to other locations. We did our work from March 1993 through November 1994 in accordance with generally accepted government auditing standards. Since no federal agency has responsibility for the issues discussed in this report, we did not obtain comments on a draft of this report. However, we did discuss our results with NCJJ and OJJDP officials and, where appropriate, incorporated their comments. -------------------- \14 Arizona, California, Florida, Idaho, Kansas, Louisiana, Maryland, Missouri, Nebraska, New Mexico, New York, Pennsylvania, South Carolina, Utah, and Texas. \15 The three states are California, Pennsylvania and Nebraska. The statistical data for all three states were somewhat limited. California's data only included five counties, albeit five of the largest counties in the state that represented 40 percent of the state's juvenile population between ages 10 and 17. Pennsylvania's data were for 1991 only and did not include Philadelphia county, the most heavily populated county in the state. Philadelphia county accounted for about one-third of all petitioned cases in Pennsylvania; however, Philadelphia petitions all cases to juvenile court. In the other counties, petitioned cases are limited to the more serious ones. Therefore, Philadelphia's data are not comparable with the rest of the counties. In Nebraska, one of the largest counties' data did not consistently indicate whether juveniles had representation at the adjudication hearing. However a state official estimated that about 75 percent of juveniles in this county had representation when they appeared before a judge at an adjudication hearing. \16 In considering the relative severity of crimes, we considered classes of crimes, such as violent or property, rather than specific types of crimes within a class or specific offense behaviors within a type of crime. Therefore, we considered violent crimes as the most serious class of crimes, followed by property crimes, drug crimes, simple assault, weapons, and public order. Because we could not determine what classes of crimes fell into the indeterminate category, we did not consider it in the ranking of offense severity. JUVENILES' RIGHT TO COUNSEL CAN BE WAIVED IN MOST STATES WE REVIEWED ------------------------------------------------------------ Letter :4 To determine juveniles' right to counsel, including the process that juveniles can use to waive their right to counsel and the provision of counsel for indigent juveniles, we reviewed the statutes, case law, and state administrative rules of the court in 15 states.\17 (See app. IV for a table summarizing the state laws.) We found that each of the 15 states had provisions that provide juveniles charged with delinquent offenses the right to counsel. Further, most of the states permitted juveniles, or their parents or guardians, to waive their right to counsel under certain circumstances. -------------------- \17 Arizona, California, Florida, Idaho, Kansas, Louisiana, Maryland, Missouri, Nebraska, New Mexico, New York, Pennsylvania, South Carolina, Texas, and Utah. JUVENILES' ACCESS TO COUNSEL ---------------------------------------------------------- Letter :4.1 Each of the states had provisions that provided juveniles access to counsel. However, the 15 states differed on details of how counsel is to be provided.\18 See appendix IV for additional information on juveniles' access to counsel. -------------------- \18 These provisions go beyond the specific requirements of the Gault decision, which only dealt with a juvenile's right to counsel at an adjudication hearing. JUVENILES' RIGHT TO WAIVE COUNSEL ---------------------------------------------------------- Letter :4.2 Eleven of the 15 state statutes that we reviewed specifically outlined circumstances and procedures under which juveniles or their parents or guardians could waive the right to counsel. In general, the waiver had to be knowing and voluntary, i.e., an "intentional relinquishment or abandonment of a fully known right." In the four remaining state statutes, waiver was not specifically addressed. In three of these four states, juveniles can waive counsel. In the remaining state, juveniles cannot waive counsel. Eleven states had specific requirements for how the waiver of counsel was to occur. For example, in Arizona, the waiver had to be in writing or in the minutes of the courts and could be withdrawn at anytime. In Florida, even if there is a waiver of counsel, the offer of counsel must be renewed at subsequent stages during the proceedings. In Idaho, a juvenile may waive counsel if he or she does so "intelligently," and the court determines that the best interest of the juvenile does not require the appointment of counsel. In Louisiana, the juvenile may waive counsel only after consulting with an attorney or other adult interested in the juvenile's welfare. In Missouri, a juvenile may waive counsel only with the approval of the court. In Pennsylvania, only the juvenile's parent or guardian can waive the right to counsel on behalf of the juvenile. New York and Texas only allow waiver under certain circumstances. In New York, a juvenile can waive counsel only after a law guardian has been appointed and the court holds a hearing. In Texas, a juvenile cannot waive counsel at certain proceedings, such as an adjudication hearing or a disposition hearing. In four states, the statues did not address waiver. In Kansas, the statute does not specifically address waiver, but a recent state Attorney General opinion concluded that juveniles can waive their right to counsel if it is a "knowing and intelligent waiver." In Utah and Nebraska, the state law was silent on waiver of counsel by juveniles. However, Utah and Nebraska officials told us that juveniles are allowed to waive counsel in these states. In New Mexico, the state statute requires that juveniles be represented at all stages of the proceedings. State officials told us that juveniles are not given the option of waiving counsel. REPRESENTATION RATES VARIED IN THREE STATES MORE WITH GEOGRAPHY THAN WITH OFFENSE SEVERITY OR OFFENSE HISTORY ------------------------------------------------------------ Letter :5 We analyzed cases petitioned to juvenile courts in certain California, Pennsylvania, and Nebraska counties. Our analysis showed that the percentage of cases in which juveniles had counsel (1) varied among the states, (2) varied with the location of the court within states, and (3) varied somewhat across offense categories or with offense history within the states. REPRESENTATION RATES VARIED AMONG STATES ---------------------------------------------------------- Letter :5.1 Table 2 shows, for the six major crime categories and the indeterminable crime category, the number of petitioned delinquency cases disposed of in 1990 and 1991 for California and Nebraska, and in 1991 for Pennsylvania. Using the data from table 2, we computed the percentage of petitioned cases in which juveniles had counsel. Table 3 shows that the percentage of petitioned cases represented by counsel varied among states. In California, about 97 percent of the juveniles were represented. In Pennsylvania, juveniles were represented in about 91 percent of the petitioned cases. In Nebraska, juveniles were represented in about 65 percent of the petitioned cases. In addition, table 3 shows that, for each major crime category, the percentage of petitioned cases represented was lowest in Nebraska. For example, about 90 percent of juveniles in Nebraska referred for violent offenses were represented, as compared with almost 100 percent in California and 94 percent in Pennsylvania. Table 2 Number of Cases Petitioned to Juvenile Court for Calendar Years 1990 and 1991 Pennsylvan Type of offenses California ia Nebraska ------------------------ ---------- ---------- ---------- All offenses, overall 60,710 14,584 4,507 Violent offenses 12,930 3,087 163 Property offenses 25,962 7,864 3,208 Drug offenses 6,846 963 112 Simple assault 3,529 1,440 396 Weapons offenses 3,474 153 85 Public order or other 7,105 624 85 nonviolent offenses Indeterminate offenses 864 453 458 ------------------------------------------------------------ Note: Data were available for five counties in California, which represented 40 percent of the juvenile population between ages 10 and 17. According to an NCJJ official, the Pennsylvania data, which were only for 1991, excluded cases from Philadelphia because Philadelphia petitions all cases, not just the more serious cases, to juvenile court. According to a Nebraska official, data were not available on representation rates in one of the state's larger counties. We also omitted these cases from our analysis. However, he added that an estimated 75 percent of the juveniles were represented at adjudication hearings in that county. These limitations apply to all results using California, Pennsylvania, and Nebraska data. Table 1 and appendix III provides information on the specific types of crimes falling into each offense category in each state. The figures in this table were computed after excluding cases in which the type of counsel was missing in the database. In California, this amounted to 5 percent of 64,275 cases; in Pennsylvania 5 percent of the 15,397 cases; in Nebraska 18 percent of the 6,603 cases. Source: GAO analysis of NCJJ data. Table 3 Percentage of Petitioned Cases Represented by Counsel, by State for Calendar Years 1990 and 1991 Pennsylvan Type of offenses California ia Nebraska ------------------------ ---------- ---------- ---------- All offenses, combined 97.2 90.6 64.9 Violent offenses 99.7 94.0 90.2 Property offenses 98.8 88.9 62.6 Drug offenses 99.4 96.4 80.4 Simple assault 98.7 89.4 75.0 Weapons offenses 99.4 90.8 90.6 Public order or other 82.7 86.2 71.8 nonviolent offenses Indeterminate offenses 98.6 94.3 54.1 ------------------------------------------------------------ Note: See note to table 2. Source: GAO analysis of NCJJ data. Our results were consistent with studies conducted by Barry C. Feld\19 and Dean J. Champion,\20 which showed that the rates of representation varied among states. Feld analyzed the representation rates in cases disposed of in juvenile court in 1984 for California, New York, Pennsylvania, Minnesota, Nebraska, and North Dakota. He found that between 85 percent and 95 percent of juveniles were represented in the large, urban states--Pennsylvania, California, and New York. By contrast, he found that between 37.5 percent and 52.7 percent of juveniles were represented in the midwestern states--Minnesota, Nebraska, and North Dakota. Champion analyzed juveniles' representation rates in California, Pennsylvania, North Dakota, Montana, and Nebraska between 1980 and 1989. He found wide variation in representation rates among the states. In California, the representation rates increased from 82 percent to 91 percent between 1980 and 1989. Over the same period, in Pennsylvania, the percentage represented increased from 55 percent to 85 percent. Conversely, in Nebraska and North Dakota, the representation rates varied between 40 percent and 50 percent over the 1980 to 1989 period. In another study which analyzed Missouri's juvenile justice system, the authors found that youths in urban localities were more likely to be represented by counsel, regardless of their race. When controlling for factors such as prior criminal history and demographic characteristics, representation by counsel remained a significant factor in predicting both detention and petition in rural jurisdictions, and a significant factor in detention, petition, and adjudication in urban jurisdictions.\21 We did not evaluate these studies, including any limitations they may have. -------------------- \19 Barry C. Feld, "In Re Gault Revisited: A Cross-State Comparison of the Right to Counsel in Juvenile Court," Crime and Delinquency, Vol. 34, No. 4, Oct. 1988: pp. 393-424. \20 Dean J. Champion, Changing Involvement of Counsel by Juveniles in Five States, 1980 to 1989: A Longitudinal Analysis, Pittsburgh, PA, National Center for Juvenile Justice, Oct. 15, 1992. \21 Kimberly Kempf, Scott Decker, and Robert Bing, An Analysis of Apparent Disparities in the Handling of Black Youth within Missouri's Juvenile Justice Systems: Technical Report. University of Missouri-St. Louis, Nov. 1990. REPRESENTATION RATES VARIED WIDELY WITH LOCATION OF THE COURT WITHIN STATES ---------------------------------------------------------- Letter :5.2 While each of the three states guaranteed juveniles the right to counsel and permitted them, their parents, or guardian to waive their right, in practice juveniles were more likely to be unrepresented in courts located in nonmetropolitan counties than in metropolitan counties.\22 Table 4 shows the representation rates for Nebraska and Pennsylvania by metropolitan and nonmetropolitan counties.\23 Table 4 Metropolitan and Nonmetropolitan County Representation Rates for Cases Petitioned to Juvenile Court in Pennsylvania and Nebraska Number of Percent of Number of Percent of Rate of juveniles petitioned juveniles unrepresented juveniles State petitioned juveniles unrepresented juveniles unrepresented ------- ------------ ------------ ------------- ------------- ------------- Pennsylvania -------------------------------------------------------------------------------- All 14,584 100.0% 1,374 100.0% 9.4 cases Metropo 12,717 87.2 1,017 74.0 8.0 litan Nonmetr 1,867 12.8 357 26.0 19.1 opolit an Nebraska -------------------------------------------------------------------------------- All 4,449\a 100.0 1,559 100.0 35.0 cases Metropo 2,034 45.7 240 15.4 11.8 litan Nonmetr 2,415 54.3 1,319 84.6 54.6 opolit an -------------------------------------------------------------------------------- Note: See note to table 2. \a Excludes 58 cases (1.3 percent of total) that were missing values on variables used in the analysis. Source: Our analysis of NCJJ data. In Pennsylvania, about 13 percent of the 14,584 juvenile court proceedings occurred in nonmetropolitan counties, but 26 percent of the cases involving juveniles without counsel occurred in nonmetropolitan counties. In metropolitan counties, 8 percent of all juveniles were unrepresented, as compared with about 19 percent who were unrepresented in nonmetropolitan counties. Therefore, juveniles were more than twice as likely to go without representation in nonmetropolitan counties as in metropolitan counties in Pennsylvania (about 19 percent compared with 8 percent). In Nebraska, more than half of the 4,449 juvenile court proceedings were disposed in nonmetropolitan counties.\24 However, about 85 percent of all unrepresented juveniles lived in nonmetropolitan counties. In metropolitan counties about 12 percent of juveniles were unrepresented. By contrast, in nonmetropolitan counties about 55 percent of juveniles were unrepresented. Therefore, juveniles in nonmetropolitan counties were more than 4 times as likely to go unrepresented as those in metropolitan counties. In California, our data for the 60,710 juvenile cases came from 5 metropolitan counties. Nevertheless, two of the five counties accounted for all of the unrepresented juveniles. In the other three counties, no juveniles were unrepresented. In the two counties containing the unrepresented juveniles, the percentage of juveniles unrepresented was about 2 and 18. Our findings on the variability in representation rates across nonmetropolitan and metropolitan court settings within states are consistent with Feld's study of representation in juvenile cases disposed in 1986 in Minnesota.\25 He found 63 percent of petitioned juvenile cases in metropolitan counties were represented by counsel, but only 25 percent of petitioned juvenile cases in nonmetropolitan counties had representation. He attributed these differences to the procedural formality and due process orientation of metropolitan courts as compared with the more traditional and informal processing in nonmetropolitan courts. -------------------- \22 For this analysis, we classified counties located within metropolitan statistical areas as "metropolitan counties," and we classified counties outside of metropolitan statistical areas as "nonmetropolitan counties." These distinctions correspond roughly to (1) more densely populated or more urban and (2) less densely populated or more rural locations. \23 NCJJ data from California contained only metropolitan counties. \24 These data exclude one metropolitan county that contained 56 percent of all unrepresented cases in the state of Nebraska. \25 Barry C. Feld, "Justice by Geography: Urban, Suburban, and Rural Variations in Juvenile Justice Administration," Journal of Criminal Law and Criminology, Vol. 82, No. 1, 1991: pp. 156- 210. PROVISION OF COUNSEL TENDED NOT TO BE AFFECTED BY THE TYPE OF CRIME AND PRIOR OFFENSE HISTORY ---------------------------------------------------------- Letter :5.3 The juvenile court data from California, Pennsylvania, and Nebraska indicated that neither the type of offense nor juveniles' offense history were strongly related to representation rates. In California and Pennsylvania, juveniles who committed certain nonviolent offenses, such as property offenses, were generally as likely to receive representation as juveniles who committed other more serious offenses, such as violent crimes or drug trafficking. However, in Nebraska, juveniles who committed less serious crimes (e.g., property and public order) were less likely to be represented than those who committed more serious crimes (e.g., violent offenses). As shown in table 3, in California, juveniles petitioned to the court for violent offenses were about as likely to be represented (99.7 percent) as juveniles sent to the court for property offenses (98.8 percent).\26 In Pennsylvania, juveniles with violent and drug offenses were only slightly more likely to be represented than juveniles with public order offenses (94.0 and 96.4 percent to 86.2 percent). In Nebraska, the juveniles' representation rates varied across offense categories. Juveniles with weapons, violent, and drug offenses were more likely (90.6, 90.2, and 80.4 percent) than those with property offenses (62.6 percent) to be represented. In Nebraska, juveniles with property offenses comprised about 71 percent of all cases and 76 percent of all unrepresented cases. For violent offenses in Nebraska, 90 percent of the juveniles were represented. Offense history was not uniformly related to the likelihood of representation between the states.\27 In Pennsylvania, across all crime types, those with one or more prior juvenile court dispositions were more likely to be represented than those with none. The extent of the difference in representation rates varied across crime type. For example, 94 percent of violent offenders in Pennsylvania with no prior referrals were represented, and 98 percent of violent offenders with one prior offense were represented. For public order or other nonviolent offenders in Pennsylvania, the difference in representation rates for those with no prior offenses and those with one prior offense was more dramatic. Seventy-seven percent of the public order offenders with no prior offenses were represented compared with 93 percent of those with one prior offense. In Nebraska, however, there were no consistent patterns across crime types, which may be due in part to the small number of cases with multiple prior offenses. For example, for property offenders in Nebraska, representation was likely to be increased from 54 percent (for offenders with no prior offenses) to 86 percent (for offenders with three prior offenses) before decreasing to 78 percent for those with five or more prior offenses. -------------------- \26 The results in this paragraph are based on our analysis of the broad offense categories given in table 1. We did a more refined analysis of the more detailed offense types comprising these offense categories. For details on this analysis, see appendix III. \27 Offense history was measured differently across the states because of the way they collected their data. For example, in Nebraska, offense history was measured by the number of prior referrals to juvenile court. In Pennsylvania, however, it was measured by the number of prior juvenile court dispositions. In California, there was no direct measure of offense history, but California did indicate whether a juvenile was under probation supervision at the time of the commission of the current offense. JUVENILE COURT OUTCOMES VARIED AMONG THE THREE STATES ------------------------------------------------------------ Letter :6 Our analysis of adjudication and placement indicated that decisions varied.\28 Table 5 shows the number of represented and unrepresented juveniles who had their cases adjudicated and those who were placed. Table 5 Number of Represented and Unrepresented Juveniles Whose Cases Were Adjudicated and Who Were Placed, for Calendar Years 1990 and 1991 Number Not State petitioned Adjudicated adjudicated Placed Not placed ---------- ------------ ------------ ------------ ------------ ------------ California -------------------------------------------------------------------------------- Total 60,710\a 41,985 18,431 21,899 20,086 Represente 58,999 41,039 17,666 21,255 19,784 d Not 1,711 946 765 644 302 represent ed Pennsylvania -------------------------------------------------------------------------------- Total 14,584\b 8,488\c 3,977 2,213 6,272 Represente 13,210 8,183\c 3,296 2,174 6,006 d Not 1,374 305 681 39 266 represent ed Nebraska -------------------------------------------------------------------------------- Total 4,507\d 3,861 645 463 3,398 Represente 2,927 2,393 534 404 1,989 d Not 1,580\d 1,468 111 59 1,409 Represent ed -------------------------------------------------------------------------------- Note: See note to table 2. \a The number includes one case that had a missing value on the variable that indicated whether it was adjudicated. It also included 293 cases that were transferred to criminal court. These 293 cases were excluded from the total number of cases having adjudication hearings (the sum of adjudicated plus not adjudicated) because transfer cases do not go to adjudication hearings. All 293 transfer cases were represented by legal counsel. \b The number includes 251 cases that were transferred to criminal court and were handled in the calculations in the manner described for California. It also includes 1,868 cases that were disposed without an adjudication hearing. These 1,868 cases also were not used in tabulating the adjudication outcomes. \c The numbers include three cases with missing values on the final disposition variable. \d The number includes one case with missing values on the variable indicating the outcome of the adjudication hearing. Source: Our analysis of NCJJ data. In addition to these differences in the likelihood of being adjudicated, overall relatively few adjudicated cases were unrepresented in California and Pennsylvania. Specifically, of the 41,985 cases adjudicated as delinquent in California, only 946 (or 2.3 percent) were unrepresented. In Pennsylvania, only 305 of the 8,488 cases adjudicated as delinquent (or 3.6 percent) were unrepresented. Conversely, in Nebraska, 1,468 of the 3,861 cases adjudicated as delinquent (about 38 percent) were unrepresented. The differences in the likelihoods of adjudication between represented and unrepresented juveniles in California was associated with the location of courts in which cases were heard and in Pennsylvania with the use of informal adjustments after cases were petitioned to the court. The higher likelihood of adjudication for unrepresented cases in Nebraska was associated with juvenile court cases located in more rural counties. Our analysis of the placement decision showed that unrepresented juveniles in Pennsylvania and Nebraska generally were less likely to receive out-of-home placements than represented juveniles for each offense category. In California, the overall likelihood of placement for unrepresented juveniles was higher than represented juveniles. Of the 21,899 juveniles placed in California, 644 (or 2.9 percent) were unrepresented. In Pennsylvania, 39 of the 2,213 juveniles that were placed (or 1.8 percent) were unrepresented. In Nebraska, 59 of the 463 juveniles placed (or 12.7 percent) were unrepresented. -------------------- \28 Our review did not include all possible outcomes of juvenile court proceedings, but focused only on the juvenile being adjudicated as a delinquent and receiving out-of-home placement. OVERALL LIKELIHOOD OF ADJUDICATION FOR JUVENILES WITHOUT COUNSEL VARIED AMONG STATES ---------------------------------------------------------- Letter :6.1 The overall likelihood of a case being adjudicated as delinquent when a juvenile was not represented varied among states. As shown in table 6, these percentages ranged from 93 percent in Nebraska to about 55 percent in California to about 31 percent in Pennsylvania. The differences in the likelihood of adjudication between unrepresented and represented juveniles also varied among states. For example, unrepresented juveniles were less than one-half as likely as represented juveniles in Pennsylvania to be adjudicated. In California, unrepresented juveniles were about 80 percent as likely to be adjudicated as were represented juveniles. (See tables 2 and 3 for the number of juveniles petitioned and the percentage of juveniles who were petitioned and represented in each state.) In California and Pennsylvania, comparatively small percentages of all adjudications (less than 4 percent) involved juveniles that were not represented. In Nebraska, a relatively large percentage of all adjudicated cases involved juveniles who were not represented (about 38 percent). See table 5 for the number of juveniles who were adjudicated and not represented. Table 6 Percentage of Petitioned Cases Adjudicated as Delinquent, by State and Whether Represented or Unrepresented by Counsel, for Calendar Years 1990 and 1991 Type of With Without With Without With Without offense counsel counsel counsel counsel counsel counsel -------- ---------- ---------- ---------- ---------- ---------- ---------- All 69.9% 55.3% 71.3% 30.9% 81.8% 93.0% offense s, overall Violent 67.1 61.4 69.9 26.5 74.1 87.5 offense s 70.4 62.6 73.3 32.1 82.2 93.3 Property offense s Drug 74.6 57.5 74.9 50.0 82.1 100.0 offenses Simple 66.5 53.3 61.0 31.5 75.1 86.9 assault Weapons 71.9 52.4 71.6 20.0 76.6 87.5 offense s Public 69.9 52.9 65.6 21.1 90.2 87.5 order or other offense s Indeterm 65.9 91.7 78.7 33.3 90.7 94.8 inate offense s -------------------------------------------------------------------------------- Note: See note to table 2. Source: GAO analysis of NCJJ data. These aggregate differences in the likelihood of adjudication between the two groups of juveniles (unrepresented and represented) tended to be associated with the location of courts or with the use of informal adjustments by the court to dispose of cases. In California, all of the unrepresented juveniles came from two of the five counties from which we obtained data.\29 In one of these two counties, the overall likelihood of adjudication was lower than in counties where no juveniles were reported as unrepresented. In Pennsylvania, our analysis and discussion with a state official indicated that the aggregate differences in the likelihood of adjudication were due largely to the use of consent decrees to dispose of unrepresented cases. A disproportionate number of cases involving unrepresented juveniles in Pennsylvania were disposed of by a consent decree--an informal disposition in which the juvenile and the court agree to suspend the case providing the juvenile remains at home under probationary conditions for up to 6 months. According to a state official, counsel generally is not present or is waived because juveniles have no risk of adjudication (or placement) if they comply with conditions of the consent decree. For Nebraska, our analysis showed that a disproportionate number of cases involving unrepresented juveniles were processed in nonmetropolitan counties, and cases involving unrepresented juveniles processed in nonmetropolitan counties also were more likely to be adjudicated as delinquent than those processed in metropolitan counties. Thus, unrepresented juveniles in Nebraska were generally more likely to have their cases adjudicated as delinquent in nonmetropolitan counties. -------------------- \29 All five California counties were located in metropolitan areas. UNREPRESENTED JUVENILES GENERALLY WERE LESS LIKELY TO BE PLACED THAN REPRESENTED JUVENILES ---------------------------------------------------------- Letter :6.2 As shown in table 7, unrepresented and adjudicated juveniles faced different likelihoods of receiving out-of-home placements in the three states. California had the highest rate of out-of- home placement for juveniles without counsel. Pennsylvania and Nebraska had lower rates of placement for these juveniles, although unrepresented juveniles in Nebraska were less likely to be placed than those in Pennsylvania. Table 7 Percentage of Cases Adjudicated as Delinquent Receiving an Out-of-Home Placement, by State and Whether Represented or Unrepresented by Counsel, for Calendar Years 1990 and 1991 Type of With Without With Without With Without offense counsel counsel counsel counsel counsel counsel -------- ---------- ---------- ---------- ---------- ---------- ---------- All 51.8 68.1 26.6 12.8 16.9 4.0 offense s, overall Violent 63.2 44.4 29.3 15.4 28.4 0.0 offense s 45.6 36.2 22.9 11.5 17.2 4.0 Property offense s Drug 51.1 56.5 34.0 9.1 30.1 4.5 offenses Simple 41.9 50.0 24.7 11.8 11.2 7.0 assault Weapons 44.3 18.2 19.3 0.0 11.9 0.0 offense s Public 66.1 81.0 35.4 31.3 7.3 4.8 order or other offense s Indeterm 53.7 54.5 47.5 14.3 14.2 3.0 inate offense s -------------------------------------------------------------------------------- Note: See note to table 2. Source: GAO analysis of NCJJ data. Among states, the difference in the likelihood of placement between represented and unrepresented juveniles varied. In California, the overall difference in placement between represented and unrepresented juveniles was about 16 percent. Between offense categories, the differences ranged from about 1 to 26 percent--weapon offenses were about 26 percent, drug offenses were about 5 percent, and indeterminate offenses were about 1 percent. In Pennsylvania and Nebraska, the overall differences between these two groups were about 14 percent in Pennsylvania and 13 percent in Nebraska, and in these two states the differences also varied. In California juveniles' overall likelihood of placement, represented or unrepresented, was the highest of the three states at about 52 percent. However, the likelihood that a juvenile would be unrepresented and placed was about 3 percent. In contrast, juveniles in Nebraska had the lowest likelihood of placement--about 12 percent. However, the risk that a juvenile would be unrepresented and placed was largest in Nebraska. These differences were due to differences in the representation rates. In California, more than 97 percent of juveniles were represented, and the chances that a placed juvenile was unrepresented was comparatively small (about 3 percent). However, in Nebraska, about 35 percent of all cases were not represented. Thus, even though Nebraska placed a smaller fraction of its juveniles than did California, the overall chances that a placed juvenile did not have counsel was much higher (about 13 percent). REGRESSION MODELS FOR THREE STATES SHOWED THAT VARIABLES OTHER THAN REPRESENTATION WERE STRONGLY ASSOCIATED WITH ADJUDICATION AND PLACEMENT OUTCOMES ------------------------------------------------------------ Letter :7 The differences in outcomes between groups of unrepresented and represented juveniles were due more to variables other than the presence of counsel. To control for the variables that were associated with the likelihood of adjudication or placement and which might have been distributed unequally between these two groups of juveniles, we estimated logistic regression models for adjudication and placement. (See app. III for detailed discussion of our approach.) For both the adjudication and placement outcomes, we estimated models for up to 10 specific crime types, e.g., rape, robbery, and larceny. We included in our regressions three classes of independent variables: (1) demographic variables, such as age, race, and gender; (2) offense and offender-behavior variables, such as current offense, offense history, supervisory status, and type of counsel; and (3) court processing variables, such as length of time to disposition, detention prior to adjudication, the location of the court, and the type of representation.\30 Overall we estimated 24 adjudication models, 10 for California, 9 for Pennsylvania, and 5 for Nebraska. We also estimated 24 placement models for the same sets of crime types in each state. For the adjudication models, we found that unrepresented juveniles generally were about as likely as represented juveniles to have their cases adjudicated as delinquents. Also, although we found that the type of counsel had effects on the likelihood of juveniles having their cases adjudicated as delinquent, the effects were not consistent among states or crime types. In addition, our models showed that variables, such as being securely detained prior to adjudication, the juvenile's prior criminal history, and the location of the court, tended to have the largest impacts on the likelihood of adjudication. For the placement models, we found that the absence of counsel was not associated with the likelihood of placement for most crime types. However, there were exceptions to this general finding. Specifically, the absence of counsel reduced the likelihood that juveniles were placed for some crimes in Pennsylvania and Nebraska but not in California. The type of counsel also influenced placement outcomes for juveniles. Finally, variables measuring the severity of the juveniles' offenses and the offense history tended to have the largest impacts on placement outcomes. -------------------- \30 For details on the crime type classifications used for each state's models and on all of the variables analyzed, see appendix III. ADJUDICATIONS ASSOCIATED WITH COURT SUPERVISION, LOCATION OF COURT, AND INFORMAL ADJUSTMENTS ---------------------------------------------------------- Letter :7.1 The results of our modeling of the adjudication decision for the three states showed that for some crime types the absence of counsel had an independent effect on the likelihood of adjudication. The effects of the type of counsel--public defenders, court- appointed attorneys, and privately paid attorneys--on the likelihood of adjudication varied among crime types and states. For example, in California, juveniles represented by court- appointed attorneys were more likely to be adjudicated than juveniles represented by the other attorney types for property, weapons, drug, and minor violent (e.g., simple assault) offenses. Conversely, they were less likely to be adjudicated for serious violent crimes such as rape and aggravated assault. In Pennsylvania, juveniles who were represented by privately paid attorneys were less likely to be adjudicated than juveniles represented by other attorney types for aggravated assault and larceny. In Nebraska, juveniles who were represented by private court-appointed attorneys were less likely to be adjudicated than juveniles represented by other attorney types. The specific variables, other than the presence of counsel, that were associated with the likelihood of adjudication varied, but there were some similarities. For example, in California, whether a juvenile was under court supervision\31 at the time of the offense, the location of the county in which the case was disposed, and whether the juvenile was detained prior to adjudication tended to have larger associations with the likelihood of adjudication, rather than whether a juvenile was represented. For example, juveniles who were detained prior to adjudication were from 1.2 to 1.5 times more likely to be adjudicated than juveniles not detained. In Pennsylvania, petitioned juveniles who had adjudicatory hearings in a court in a metropolitan county and were detained prior to their adjudicatory hearing increased their likelihood of being adjudicated as delinquent. However, the use of informal adjustments (i.e., consent decrees) to dispose of cases resulted in petitioned juveniles not being adjudicated because they did not have adjudicatory hearings. The models showed that unrepresented juveniles were generally as likely as represented juveniles to have their case adjudicated. For some crimes and under some conditions, however, this was not always the situation. For example, unrepresented juveniles with burglary and weapons offenses in California and Pennsylvania and unrepresented juveniles with larceny offenses in Nebraska were less likely to have their cases adjudicated. In addition, in Nebraska juveniles in rural areas, generally were less likely to have their cases adjudicated than those in urban areas. To find out possible reasons for these results, we reviewed relevant literature and talked with an official in Pennsylvania as well as officials representing the American Bar Association, juvenile justice systems, and public defenders to obtain their reactions to these findings. The 1988 and 1991 studies by Barry C. Feld showed that juveniles may be less likely to have their cases adjudicated in rural areas because judges in those areas may have greater familiarity with the juveniles' cases and may handle them less formally or are less likely to adjudicate them as delinquent. Consequently they may be less likely to appoint counsel. The officials whom we interviewed and our review of the literature offered the following possible reasons why juveniles without counsel were less likely to have their cases adjudicated: The judges may be more likely to ensure that juveniles have counsel when they believe that a juvenile has a greater risk of receiving a severe disposition (e.g., placement). The informality of juvenile court processing in some states may be a reason. Juveniles without representation are disproportionately found in rural areas. Judges in these areas may have greater familiarity with cases, and handle them less formally. For example, for these cases, they may be less likely to appoint counsel. Judges may be more lenient to juveniles already under supervision by the court and may be willing to continue their present sentence, rather than adjudicating for a new offense. Recognizing this, the juveniles may decide to waive counsel. The relationships between judges and attorneys may be a factor. For example, some judges may have antagonistic relationships with attorneys. The presence of attorneys, therefore, may result in more severe dispositions. Judges may also be more protective of juveniles without counsel and, therefore, less likely to give them a severe disposition. -------------------- \31 The juvenile was being supervised as a ward of the court immediately prior to the referral. LIKELIHOOD OF PLACEMENT WAS STRONGLY ASSOCIATED WITH DETENTION AND PRIOR OFFENSE HISTORY ---------------------------------------------------------- Letter :7.2 In our regression models, we analyzed variables that were associated with placement outcomes.\32 The models indicated generally that (1) whether the juvenile was detained prior to adjudication and (2) prior offense history had a large impact on the likelihood of placement. In each state, across most crime types, juveniles detained prior to adjudication were more likely to be placed than juveniles who were not detained. In general, detention prior to adjudication was viewed as an indication of the severity of the offense, and juveniles who were part of more severe cases were more likely to be placed. Therefore, the positive association between detention and the likelihood of placement is not surprising. Similarly, juveniles' prior offense history or prior contact with the juvenile justice system had strong and positive effects on the likelihood of placement for most crime types for California and Pennsylvania. In California, juveniles who were under court supervised probation for a previous offense when they committed their current offense were from 2 to 8 times more likely to be placed (for all crimes except rape and robbery) than juveniles who were not under court supervision. In Pennsylvania, juveniles with more than 3 priors were from 4 to 27 times more likely to be placed (for all crimes except weapons) than juveniles with fewer than 3 priors. We also found that the types of representation influenced the likelihood of placement. For example, in California and Nebraska juveniles who were represented by privately paid attorneys were less likely to be placed than juveniles who were represented by other attorney types for larceny and drug-related crimes. In Pennsylvania, juveniles represented by privately paid attorneys were more likely to be placed than juveniles represented by other attorney types for aggravated assault and larceny. -------------------- \32 For a discussion of the constraints imposed on our analysis by the small number of unrepresented and placed cases, see appendix III. JUVENILES IN ADULT CRIMINAL COURT HAD SAME RIGHT TO REPRESENTATION AS ADULTS ------------------------------------------------------------ Letter :8 Under the Sixth Amendment to the U.S. Constitution, all defendants, including juveniles being tried as adults, have the right to "the assistance of counsel for his or her defense." We could not locate a database containing information on how frequently juveniles were represented in adult court. Our review of the law showed that juveniles tried as adults were to be afforded the same right to representation as adults tried in criminal court. In addition, the results of our prosecutor's survey revealed their overall satisfaction with the quality of counsel juveniles were receiving in adult court. QUALITY OF COUNSEL PROVIDED JUVENILES CONSIDERED ADEQUATE BY SURVEY RESPONDENTS AND LOCAL OFFICIALS ------------------------------------------------------------ Letter :9 According to state officials, no criteria existed to determine how well juveniles were being represented by counsel, either in adult or juvenile court. For example, we found no national or state databases designed to define or measure quality within the context of the juvenile justice system. In the absence of such criteria, we obtained some perspective of how well juveniles were being represented in both juvenile and adult court by soliciting the views of juvenile justice officials around the country. To obtain some perspective, we conducted a nationally representative survey of prosecutors that handle juvenile justice matters and interviewed selected juvenile court judges in eight states. In addition, through personal interviews, we obtained the opinions of juvenile justice officials in four states. In our survey of prosecutors and interviews with judges and various officials, we asked questions regarding their perceptions of the quality of counsel across several dimensions, including overall (1) preparation, (2) legal skills, and (3) effectiveness. The questions were targeted to the three types of counsel that represent juveniles in both juvenile and adult courts: public defenders, privately paid attorneys, and court-appointed attorneys. The results of our survey and discussions revealed overall satisfaction with the quality of counsel juveniles were receiving, both in adult and juvenile courts. Some officials, however, raised concerns over nonqualitative considerations that were perceived to affect the quality of counsel, such as high caseloads and limited state funding. PROSECUTORS WERE GENERALLY SATISFIED WITH QUALITY OF ALL THREE TYPES OF COUNSEL PROVIDED JUVENILES ---------------------------------------------------------- Letter :9.1 Our survey showed that while there were some differences in quality of these three types of counsel, the respondents generally viewed all three types of attorneys as being sufficiently prepared, possessing the requisite legal skills, and representing their clients effectively.\33 For example, as table 8 shows, most prosecutors responded that all three types of counsel in juvenile court were well prepared. Similarly, as shown in table 9, most prosecutors responded that all counsel in adult court were well prepared. (See app. V for additional details on our survey of prosecutors.) Table 8 Percentage of Prosecutors Responding in the Two Most Favorable Categories to Attorneys' Capabilities in Juvenile Court Court- Public Private appointed Dimensions defenders attorneys counsel --------------- ------------- ------------- ------------- Preparation 77 81 78 Legal skills 73 67 71 Effectiveness 81 83 83 ------------------------------------------------------------ Note: For all three dimensions, we used a 5-point scale, but the scale was different for each dimension. For the preparation dimension, the top two responses were "very well prepared" and "generally well prepared;" for the legal skills dimension, the top two responses were "excellent" and " very good;" for the effectiveness dimension, the top two responses were "very effective" and "generally effective." We omitted those prosecutors who responded "no basis to judge," "not applicable," and those not responding. See appendix V for the tabulation of the questionnaire and appendix II for sampling errors. Source: GAO Survey. Table 9 Percentage of Prosecutors Responding Favorably to the Attorneys' Capabilities in Adult Court Court- Public Private appointed Dimensions defenders attorneys counsel --------------- ------------- ------------- ------------- Preparation 80 90 83 Legal skills 74 74 70 Effectiveness 80 91 79 ------------------------------------------------------------ Note: See note on table 8. Source: GAO Survey. Finally, we asked prosecutors for their views on the impact of representation on juvenile dispositions that were "similar" except for the presence of counsel. Most prosecutors believed that having counsel did not have a big impact on the eventual outcome of the case. About 70 percent said that juveniles received similar dispositions regardless of whether they had counsel. Therefore, in the views of most prosecutors, having representation in juvenile court did not produce large, measurable differences in outcomes for juveniles. -------------------- \33 We did not obtain a national perspective from other juvenile justice officials (e.g., judges or public defenders). JUDGES WERE GENERALLY SATISFIED WITH QUALITY OF COUNSEL PROVIDED TO JUVENILES ---------------------------------------------------------- Letter :9.2 We obtained the views of 16 judges regarding the issues related to the quality of counsel within their jurisdictions in Florida, Louisiana, Michigan, North Carolina, Ohio, Pennsylvania, Texas, and Utah. Overall, the judges indicated that the quality of counsel provided to juveniles in juvenile court was adequate. For example, 10 of the 12 judges who commented on the quality of counsel provided juveniles in juvenile court characterized their perception of quality of counsel with responses such as "acceptable," "good," and "excellent." In addition, one of the 12 judges indicated that the counsel afforded juveniles in juvenile court was as good as the counsel adults receive in adult court. However, when asked to comment on the quality of counsel by attorney type (i.e., public defender, private, and court-appointed), 7 of the 11 judges who responded indicated that quality of counsel varied by type of attorney. For example, four of the seven judges noted that, of the three attorney types, public defenders are generally more knowledgeable about issues such as juvenile law, court proceedings, and disposition options. One judge of the four, in noting the public defender's familiarity with juvenile court law, commented that he perceived public defenders as "specialists." In commenting on the dimensions of quality representation in a juvenile case, judges most often cited (1) legal skills, (2) knowledge of the law, (3) commitment to the client, and (4) preparation. Other dimensions cited were experience, a thorough investigation of the case, knowledge of alternatives, and meeting with the client. OFFICIALS IN FOUR STATES VISITED BELIEVED QUALITY OF COUNSEL WAS GENERALLY ADEQUATE, DESPITE LIMITED RESOURCES ---------------------------------------------------------- Letter :9.3 We discussed the quality of counsel issue with various juvenile court and justice officials--including juvenile prosecutors, judges, attorneys, service workers, and probation officers--located in eight jurisdictions in Florida, Louisiana, Pennsylvania, and Utah. According to state officials, none of these states had developed criteria against which quality of counsel provided could be measured. However, two jurisdictions had established specific qualifications that attorneys must meet to represent juveniles in certain cases.\34 Overall our discussions with juvenile justice officials revealed generally favorable responses concerning the quality of counsel provided to juveniles. Further, several of the officials in Florida, Louisiana, and Pennsylvania indicated that, in terms of effective representation, public defenders are at least as capable as the private attorneys in representing juveniles. The most frequently cited reasons were the public defenders' familiarity and experience with juvenile court procedures and available disposition or service options. For example, several of these officials noted that public defenders may do a better job than private attorneys because they are more knowledgeable of the laws and local rules governing juvenile proceedings. In Utah, one official noted that while some private attorneys may be more aggressive, court-appointed attorneys develop more expertise in handling juvenile cases. However, some of these officials raised nonqualitative considerations that were perceived to affect the quality of counsel, such as high caseloads and limited state funding. They recognized that while public defenders tended to have the most experience in juvenile law and procedures, the officials also noted that public defenders had very high caseloads. For example, one official noted that because of the limited resources, public defenders do not have as much time to prepare and often must seek continuances while others pointed out that the number of public defenders had not kept pace with the increased demand for indigent defense services. Another official commented that the court-appointed attorneys do not have as many resources or the time to devote to a case as private attorneys. -------------------- \34 The two jurisdictions included Tallahassee, Florida, and Philadelphia, Pennsylvania. Tallahassee officials told us that the Second Judicial Circuit required that, when appointing an attorney to handle a case representing a juvenile in a case in which more than the one juvenile is involved, the appointed attorney must (1) be an active member in good standing of the Florida Bar Association; (2) have at least 1 year of juvenile court practice with demonstrated competence; (3) have had, within the past two years, some specialized training in juvenile law or litigation; (4) be familiar with the requisite court system, including specifically the procedural rules regarding timeliness of filings and procedural default; and (5) have demonstrated proficiency in and commitment to quality representation. The Philadelphia Criminal Rules provide guidance on the appointment of counsel. The rules required the juvenile court to maintain a list of attorneys qualified for appointment in both major felony juvenile and nonmajor felony juvenile cases. The rules also specified that, major felony juvenile cases (where the juvenile is likely to be incarcerated as a juvenile or certified to an adult court), the attorney must (1) have qualified for appointment for adult felony cases, (2) have prior experience in at least five juvenile cases, which were tried to completion, and (3) have had at least two major felony juvenile cases in the past 2 years. For all other juvenile cases, however, the attorney must have fulfilled the requirements for appointment of counsel in adult misdemeanor cases. For both major and nonmajor cases, the Rules specify that attorneys must have completed at least one continuing legal education course in juvenile law within the past year, are familiar with the Pennsylvania Juvenile Act and the Philadelphia Court of Common Pleas Juvenile Court Rules, and are reasonably available to accept appointment. ---------------------------------------------------------- Letter :9.4 We are sending copies of this report to the Attorney General; the Administrator, Office of Juvenile Justice and Delinquency Prevention; the Director, Office of Management and Budget; and other interested parties. Copies will also be made available to others upon request. Major contributors to this report are listed in appendix VI. If you have any questions about this report, please contact me on (202) 512-8777. Laurie E. Ekstrand Associate Director, Administration of Justice Issues PROFILES OF STATE JUVENILE JUSTICE PROCESSES =========================================================== Appendix I Table I.1 Profile of Juvenile Justice Processes in Arizona, California, Florida, Maryland, and Missouri Proces ses Arizona California Florida Maryland Missouri ------ ------------ ------------ -------------- -------------- ------------ Courts Superior Superior Circuit court 12 district 45 circuit with court in 15 court in 58 in 20 circuits courts in 24 courts juveni counties counties counties, 8 in 115 le circuit courts counties\a jurisd in 24 counties iction Intake -------------------------------------------------------------------------------- Level Local Local State State Local at which intake is organi zed\b Branch Judicial Executive Executive Executive Judicial\d at which intake is admini stered \c Agency Juvenile County Department of Department of Probation or court probation\e Health and Juvenile service unit probation Rehabilitative Services (DJS) of circuit of Services court govern (DHRS)\f ment respon sible for intake Party Juvenile County DHRS intake DJS intake Juvenile respon court intake probation worker officer officer\g sible staff intake staff for perfor ming intake Party County District State attorney State attorney Juvenile author attorney attorney officer ized to file petiti on Probation -------------------------------------------------------------------------------- Level Local--The Local--the State-- State-- Local at Administrati county juveniles are probation which ve Office of executive provided supervision is probat The Courts funds a probation or administered ion helps to probation aftercare by DJS is fund and department supervision organi monitor that usually through the zed juvenile supports a Delinquency probation separate Case services in probation Management the state division\h Division of DHRS\i Level Judicial Executive Executive Executive Judicial at which probat ion is admini stered Aftercare\j -------------------------------------------------------------------------------- Level State State State State State at which afterc are is organi zed Level Executive Executive Executive Executive Executive at which afterc are is admini stered Agency The The Parole Delinquency DJS The respon Department Services Case Department sible of Youth Branch of Management of Social for Treatment the Division of Services, afterc and Department DHRS Division of are Rehabilitati of the Youth Youth servic on Authority, Services es Institutions and Camps Branch State Institutions for Delinquents -------------------------------------------------------------------------------- Level Executive Executive Executive Executive Executive at which state instit ution servic es is admini stered Respon The The DHRS is DJS operates The sible Department Department responsible all of the Department agency of Youth of the Youth for the state's of Social and Treatment Authority, administration juvenile youth Services, type and Institutions of secure centers, Division of of Rehabilitati and Camps training detention Youth servic on is Branch is institutions, centers, and Services e responsible responsible several secure group homes administers provid for secure for training juvenile institutions ed juvenile centers, detention and institutions youth facilities, community- , community centers, and secure or based corrections training nonsecure treatment contracts, schools, and community- centers and parole conservation based (aftercare camps treatment supervision) camps and services programs -------------------------------------------------------------------------------- \a Circuit sizes range from about 1 to 5 counties. Each judicial circuit has a juvenile court judge, who is appointed by the circuit court. \b "Organized" as defined by the National Center for Juvenile Justice (NCJJ), refers to the level of government from which a service is organized in a state. NCJJ arrived at this definition by determining (1) the level of government in a state from which the authority for the provision of a service originates (e.g., probation services are delivered at the local level but may not be organized from that level) and (2) the level of government in a state from which the delivery of service can be directly affected. \c "Administered" as defined by NCJJ, refers to the branch of government that is vested with the responsibility for the day-to-day provision of a service. \d In all but three counties, (Kansas City, St. Louis City, and St. Louis County), every circuit court has an intake unit administered by probation. \e The county executive funds a probation department that usually supports a separate juvenile probation division. The juvenile probation function often includes intake and investigation services. \f Florida's Department of Health and Rehabilitative Services is an integrated system that delivers social, health, and rehabilitative services to children and youth in 11 Children, Youth and Family districts. Each district maintains a program office that is responsible for integrated intake and program coordination that encompass delinquency, child welfare, and health services. \g In other states, this individual may be referred to as a probation officer. \h On the state level, the Department of the Youth Authority, Community Corrections Branch, has regional offices that work closely with county probation to run prevention and community corrections programs. The Department also administers state funds that flow to county probation. \i The Division employs Community Control Officers to supervise juveniles placed on community control (probation supervision) by circuit courts and to provide aftercare supervision to juveniles released on "furlough" from juvenile institutions. Several Community Control Unit Offices are located throughout each District. \j NCJJ has defined aftercare as the community supervision of youth released from the state delinquent institution. Source: We developed this information from discussions with various state and local juvenile justice officials, court administrators, juvenile justice advocacy groups, judges, and prosecutors. We also reviewed federal, state, and local documents on juvenile justice procedures and processes. We did not verify the accuracy of the information through a review of state statutes and administrative rules. Table I.2 Profile of Juvenile Justice Processes in Nebraska, Pennsylvania, South Carolina, and Utah Processes Nebraska Pennsylvania South Carolina Utah ------------ --------------- --------------- --------------- --------------- Courts with 93 county Court of Common Family court, Separate juvenile courts, Pleas,\b 16 circuit juvenile jurisdiction in 21 60 districts in courts in court, districts; 67 counties\c 46 counties 8 juvenile separate court districts juvenile courts in 3 counties\a Intake -------------------------------------------------------------------------------- Level at Local Local State State which intake is organized\d Branch at Judicial Judicial Executive Judicial which intake is administered \e Agency or County court County juvenile Department of Intake division unit of probation Juvenile of juvenile government department\f Justice (DJJ) court responsible for intake Party Probation Probation DJJ intake Probation responsible officer of officer staff officer for court intake\g performing intake Party County attorney With the With the County attorney authorized court's court's or district to file authority, any authority, any attorney petition individual can individual can file a petition file a petition Probation -------------------------------------------------------------------------------- Level at State\h Local--county State--The State-- which governments Department of probation probation fund juvenile Juvenile officers are is organized probation Justice, employed by the services that Community State Court are operated Division Administrator under the through local to provide administrative field offices services in supervision of judicial local Courts of districts Common Pleas Level at Judicial Judicial Executive Judicial which probation is administered Aftercare\i -------------------------------------------------------------------------------- Level at State Local State State which aftercare is organized Level at Executive Judicial Executive Executive which aftercare is administered Agency The Juvenile County The Department The Department responsible Parole governments of Juvenile of Human for Administration fund aftercare Justice, Services, aftercare of the services that Community Division of services Department of are operated Division Youth Correctional under the through local Corrections is Services, administrative field offices responsible for Division of supervision of the aftercare Juvenile the local supervision of Services Courts of juveniles Common Pleas released from secure institutions by the Youth Parole Authority State Institutions for Delinquents -------------------------------------------------------------------------------- Level at Executive Executive Executive Executive which state institution services is administered Responsible The Department The Department Juvenile The Department agency and of Correctional of Public correctional of Human type of Services, Welfare facilities are Services, service Division of operates youth operated by the Division of provided Juvenile development Department of Youth Services centers, secure Juvenile Corrections treatment Justice, administers the units, and Institutional state's secure youth forestry Division; the confinement camps; the Division's facilities, counties facilities detention administer the include secure centers and detention institutions contracts for facilities for juveniles residential and and an nonresidential assessment/ community- evaluation based services facility -------------------------------------------------------------------------------- \a In much of Nebraska, county courts exercise jurisdiction over juvenile matters. However, in Douglas, Lincoln, and Sarpy counties, separate juvenile courts exercise jurisdiction. Nebraska statute authorizes the establishment of separate juvenile courts in counties with a population greater than 30,000 where authorized by the electorate. In Nebraska, both county courts and juvenile courts exercise their jurisdiction over delinquency matters in concurrence with district courts. \b A Juvenile Court Judges Commission advises juvenile courts on the proper care of delinquents, establishes administrative procedures and standards, examines personnel practices and employment standards used in probation offices, collects and publishes statistics, administers a grant-in-aid program to improve county juvenile probation services, and provides training. \c In Allegheny and Philadelphia counties, the Family Court Division of the Court of Common Pleas exercises jurisdiction. \d "Organized" as defined by NCJJ, refers to the level of government from which a service is organized in a state. NCJJ arrived at this definition by determining (1) the level of government in a state from which the authority for the provision of a service originates (e.g., probation services are delivered at the local level but may not be organized from that level) and (2) the level of government in a state from which the delivery of service can be directly affected. \e "Administered" as defined by NCJJ, refers to the branch of government that is vested with the responsibility for the day-to-day provision of a service. \f Juvenile probation services vary widely among counties. The number of staff and the caseload per probation officer are greatly affected by the size and wealth of the particular county. Smaller counties generally maintain a small staff consisting of one to two people who are likely to handle both adult and juvenile cases. \g Screening varies among counties. In Douglas County, for example, the county attorney screens the juvenile over the phone, having never met the individual. In Sarpy County, on the other hand, Probation Officers screen the juvenile in person. \h In 1985, Nebraska's Juvenile probation services transitioned to a statewide system. Accordingly, juvenile probation services operate under the Nebraska Probation System of the State Judicial Branch. \i NCJJ has defined aftercare as the community supervision of youth released from the state delinquent institution. Source: We developed this information from discussions with various state and local juvenile justice officials, court administrators, juvenile justice advocacy groups, judges, and prosecutors. We also reviewed federal, state, and local documents on juvenile justice procedures and processes. We did not verify the accuracy of the information through a review of state statutes and administrative rules. OBJECTIVES, SCOPE, AND METHODOLOGY ========================================================== Appendix II The 1992 reauthorization (P.L. 102-586) of the Juvenile Justice and Delinquency Prevention Act of 1974 (P.L. 93-415) required us to conduct a study of access to counsel during delinquency proceedings in juvenile court. In discussions with the staffs of the House Committee on Economic and Educational Opportunities (formerly the Committee on Education and Labor) and the Senate Committee on the Judiciary, we agreed to review state laws for 15 states that we selected to determine juveniles' right to counsel, determine the frequency with which juveniles have counsel in juvenile courts in three states, determine the likely impact of counsel on juvenile outcomes, determine if juveniles who are in adult court have counsel, and provide insights regarding the issue of quality of counsel. In addressing these objectives, we met with officials from OJJDP, NCJJ, and other organizations familiar with juvenile justice issues, such as the American Bar Association, National Legal Aid Defenders Association, and the Coalition for Juvenile Justice to determine their perspectives on juveniles' access to counsel. We conducted a review of literature identified in bibliographies provided by NCJJ and OJJDP as well as articles and studies we identified. We used only those studies that were relevant to addressing our objectives. We did not evaluate the studies or their limitations. Also, to better understand state juvenile justice systems, we developed profiles of juvenile justice processes in nine states: Arizona, California, Florida, Maryland, Missouri, Nebraska, Pennsylvania, South Carolina, and Utah. We developed this information from discussions with various state and local juvenile justice officials, court administrators, juvenile justice advocacy groups, judges, and prosecutors. We also reviewed state and local documents on juvenile justice procedure and processes. We did not verify the accuracy of the information through a review of state statutes and administrative rules. These states were selected for their range of urban and rural locations and their diversity of juvenile justice systems. To determine the scope of juveniles' right to counsel in selected states, we analyzed statutes, administrative rules, and case law in 15 states: Arizona, California, Florida, Idaho, Kansas, Louisiana, Maryland, Missouri, Nebraska, New Mexico, New York, Pennsylvania, South Carolina, Texas, and Utah. We selected these states after considering such factors as the diversity of each state's juvenile justice system and the availability of state juvenile court processing data maintained by NCJJ.\1 In addition to our legal analysis of statutes in selected states, we reviewed an analysis of statutes nationwide prepared by NCJJ. To determine the frequency with which juveniles have counsel in juvenile court, we used annual juvenile court data sets and case records compiled and maintained by NCJJ.\2 Specifically, we analyzed data on petitioned delinquency offenders in juvenile courts in California, Pennsylvania, and Nebraska. We identified four other states with NCJJ data on counsel, which we excluded because they were the subject of another study (Minnesota), had a relatively small population (North Dakota), or did not have data we needed for other mandated juvenile justice studies--e.g., Juvenile Justice: Minimal Gender Bias Occurred in Processing Noncriminal Juveniles (GAO/GGD-95-56, Feb. 28, 1995). We used the NCJJ data for the counties in the three states for which data were available to develop statistics for a 2-year period in California and Nebraska covering calendar years 1990 and 1991, and for 1991 for Pennsylvania, the most recent available data at the time of our review. The sample for each state was limited to petitioned delinquency cases only, excluding status cases as well as abuse, neglect, and dependency cases, and routine traffic offenses. We arranged the delinquency offenses into seven broad categories: violent, property, simple assault, drugs, weapons, public order, and indeterminate. To obtain indications of the impact of counsel on dispositions, we did a two-part analysis. First, we developed statistics on case processing outcomes for each state. Second, we developed statistical models that controlled for variables that may affect a juvenile's disposition, e.g., offense severity or prior history. Specifically, for each state, we estimated a series of logistic regression models that investigated the relationships between the type of representation in juvenile delinquency proceedings and key decision points in the juvenile judicial process controlling for a number of relevant variables. We measured four types of representation: (1) no counsel, (2) privately retained counsel, (3) court-appointed private counsel, and (4) public defender. We analyzed two key decision points: (1) whether the juvenile was, in fact, adjudicated as delinquent and (2) whether the juvenile received an out-of-home placement as disposition of the case after adjudication. We estimated the models separately for each of the 3 states for up to 10 crime types. Data were available for five counties in California which represented 40 percent of the juvenile population between ages 10 and 17. The Pennsylvania data, which were only for 1991, excluded cases from Philadelphia because Philadelphia petitions all cases, not just the more serious cases, to juvenile court. These limitations apply to all results using California and Pennsylvania data. Table 1 provides information on the specific types of crimes falling into each offense category in each state. The figures in this table were computed after excluding cases in which the type of counsel was missing in the database. In California, this amounted to about 5 percent of 64,275 cases; in Pennsylvania about 5 percent of the 15,397 cases; in Nebraska about 18 percent of the 6,603 cases. According to a Nebraska official, data were not available on representation rates in one of the state's larger counties. Therefore, we also omitted this county's cases from our analysis, which amounts to 14 percent of all of Nebraska's cases. However, the official added that an estimated 75 percent of the juveniles were represented at adjudication hearings in that county. Regarding juveniles in adult court, we were unable to locate a national database containing information on the frequency of representation. However, we discussed the matter with juvenile justice officials during our field visits and solicited information in our survey of prosecutors. To provide insights regarding the quality of counsel, we surveyed a national sample of local prosecutors; interviewed juvenile justice officials, such as judges at the state and local levels; and surveyed public defender offices. To gather opinions and experiences of prosecutors concerning access to counsel in the juvenile court system, we obtained 226 completed questionnaires from a nationally representative, probability sample of district prosecutor offices that deal with juveniles in juvenile courts. We first identified a stratified probability sample of county prosecutors in 290 of the 3,110 counties in the United States. To gather information about large and small counties, the sample was stratified on the basis of the number of felony convictions in 1985. The 1985 felony convictions were used to first draw this sample of 290 counties in 1986 for the National Judicial Reporting Program. We developed the questionnaire with advice from experts at NCJJ and the Department of Justice's Bureau of Justice Statistics and pretested it with selected prosecutor offices. The survey was mailed in March 1994. Two sets of follow-up telephone calls and two additional mailings were conducted between May and July 1994. We contacted all 290 selected counties to determine whether they had juvenile prosecutors and to determine the counties over which the prosecutors had jurisdiction. We determined that 270 of the 290 counties were eligible members of our study population of prosecutor offices that dealt with juvenile offenders in juvenile court. After weighing the probabilities of selection, we estimated that the study population consisted of approximately 2,118 such juvenile prosecutors in the United States. This number is less than the number of counties (3,110), partly because some counties are consolidated under a single juvenile prosecutor and partly because some jurisdictions do not have prosecutors that appear in juvenile court. Of the 270 sampled prosecutors from the study population, 226 responded, for a response rate of 84 percent. We estimated all figures reported in the text from the returned questionnaires to the estimated 2,118 juvenile prosecutor offices. All sample surveys are subject to sampling error. This may occur because the sample results may differ from what would have been obtained if the entire population had received and returned the questionnaire. The size of sampling errors in any survey depends largely on the number of respondents and the amount of variability in the data. In this report, all estimates, with two exceptions, are made at the 95-percent confidence level with a sampling error of plus or minus 10 percent. This means that, if we drew repeated samples from the entire study population of prosecutor offices, 19 out of 20 samples would produce estimates within 10 percent of the true proportion in the total population. In table 9, the estimates of 80 percent and 74 percent have confidence intervals of plus or minus 10.1 and 10.2 percent, respectively. In addition to the reported sampling errors, any survey may be subject to nonsampling errors as well. For example, differences in how a particular question is interpreted, in the sources of information that are available to respondents or in the types of people who do not respond, can introduce unwanted variability into survey results. We included steps in the data collection and data analysis stages to minimize such nonsampling errors. We selected the sample from a complete list of all counties, and we pretested our questionnaire with experts and members of the target population. Our extensive follow-up efforts were designed to maximize the response rate, and we achieved a final response rate of 84 percent. All data were keyed twice and verified during data entry, and all computer analyses were reviewed by a second, independent analyst. We also administered a semistructured interview to 16 judges to gather information concerning their views on the quality of counsel provided to juveniles. We asked the judges about their perceptions related to quality of counsel for the three types of counsel provided to juveniles--public defenders, private attorneys, and court-appointed counsel. The judges--located in Florida, Louisiana, Michigan, North Carolina, Ohio, Pennsylvania, Texas, and Utah--were judgmentally selected on the basis of their availability, and their views do not necessarily represent those of other judges. We also sent a questionnaire to 26 state-level public defender offices. We selected those states that, according to officials from the National Legal Aid Defenders' Association, comprised the universe of statewide public defender offices that represented juvenile delinquents. We received 19 responses and followed up with telephone calls for clarification. In reporting survey and interview results from juvenile justice officials, we did not evaluate their responses for possible biases. To develop comparative information about the availability of counsel for delinquents, we visited a total of eight jurisdictions within each of four states. Those jurisdictions included Tallahassee and Tampa, Florida; Baton Rouge and West Francisville, Louisiana; Philadelphia, Pennsylvania; and Salt Lake City, Logan, and Ogden, Utah. We judgmentally selected Florida, Louisiana, Pennsylvania, and Utah on the basis of several factors, such as the large concentration of juvenile offenders, the diversity of juvenile justice systems, the range of urban and rural locations, and the availability of state juvenile court processing data from NCJJ. Generally in each of the selected states, we interviewed local judges, prosecutors, and public defenders as well as representatives of juvenile courts, bar associations, and state child welfare agencies. We selected them on the basis of their availability at the time of our visits. Where possible, we obtained statistical information on access to counsel, which was maintained by the jurisdiction, and observed juvenile proceedings in session. -------------------- \1 Each year, NCJJ collects juvenile court processing data for up to 26 states. \2 We used data that were housed in and made available by the National Juvenile Court Data Archive, which is maintained by NCJJ and supported by a grant from OJJDP. These data were originally collected by the Alameda County, CA, Probation Department; the Los Angeles County, CA, Probation Department; the San Francisco County, CA, Juvenile Probation Department; the San Joaquin County, CA, Probation Department; the County of Ventura, CA, Corrections Services Agency; the Nebraska Commission on Law Enforcement and Criminal Justice; and the Pennsylvania Center for Juvenile Justice Training and Research. Neither the original data collectors nor NCJJ bear any responsibility for our analyses or interpretations of the data. METHODS FOR ANALYZING THE FREQUENCY OF REPRESENTATION AND IMPACTS OF COUNSEL ON JUVENILE COURT OUTCOMES ========================================================= Appendix III This appendix discusses the methods we used (1) to review the frequency with which juveniles accused of delinquent activities received counsel and (2) to assess the impact of counsel on two juvenile court outcomes. For each issue, we restricted our analysis to cases petitioned to juvenile court that were disposed in calendar years 1990 and 1991 in California, Nebraska, and Pennsylvania. For the purposes of this analysis, we used the term case to refer to a single juvenile defendant. We used petitioned to refer to defendants who had adjudicatory hearings or who were handled in a formal manner by the juvenile court. APPROACH OF THE REVIEW ------------------------------------------------------- Appendix III:1 To review the frequency of representation, we identified cases of juveniles petitioned to court and then determined if they had legal representation during the processing of their cases. We looked at whether represented juveniles differed from those without representation on three characteristics: (1) the type of crime with which they were charged, (2) their offense history, and (3) the location of the court in more urban or more rural counties. We reviewed the relationship between legal representation and two juvenile court outcomes--adjudication as delinquent and receiving an out-of-home placement after being adjudicated as delinquent. Adjudication refers to the fact-finding hearing in which the juvenile court determines whether or not there is sufficient evidence to sustain the allegations in a petition. Out-of-home placement refers to post-adjudication dispositions that result in the placement of the juvenile outside his or her own home, regardless of whether such placement is in a secure or nonsecure facility. To analyze these relationships we compared the differences in outcomes for juveniles who were represented against those who were not, controlling only for the type of crime. Second, to control for variables other than the type of representation, which may also be associated with juvenile court outcomes, we conducted regression analysis of the relationship between the outcomes and juveniles' case characteristics. Specifically, we estimated separate logistic regressions for up to 10 crime types in each of the 3 states for each of the 2 juvenile court outcomes: (1) whether a juvenile was adjudicated as delinquent and (2) whether an adjudicated juvenile delinquent was placed out-of-home\1 . In estimating these equations, we controlled for three classes of characteristics (1) demographic variables, such as age, race, and gender; (2) offense and offender-behavior variables, such as current offense, offense history, and juvenile court supervisory status; and (3) juvenile court processing variables, such as detention prior to adjudication and the location of the court.\2 We estimated these regressions to answer three general questions: (1) What were the independent or main effects of whether a juvenile was represented on the likelihood either of being adjudicated or of being placed? (2) What difference did the type of counsel--public defender, court appointed, and privately paid--have on these likelihoods? and (3) What independent variables had the largest impacts on the dependent variables? -------------------- \1 These two dependent variables--adjudicated and placed--were dichotomous dependent variables. Each had only two possible values. Juveniles' cases either were adjudicated as delinquent or they were not. Similarly, they either were placed out-of-home or they were not. In our analysis, we looked first at whether a case was adjudicated and second, only for those cases that were adjudicated, we looked at whether they were placed. We did not identify types of placements because of data limitations. \2 The specific crime types and variables used in the models are discussed in more detail in the following sections. APPROACH, DATA, AND METHODS TO THE ANALYSIS ------------------------------------------------------- Appendix III:2 In general, we discuss three types of analyses in this appendix. First, we analyzed the data to determine the number of cases in which juveniles were represented or not. Drawing on issues raised in the literature, we analyzed the representation rates for a number of characteristics, including type of offense, offense history, and location of juvenile courts. Second, we compared juvenile court outcomes--whether adjudicated or placed out-of-home--of juveniles who were represented with those who were not represented. We reported the aggregate differences between these two groups by offense categories. Third, because juvenile court outcomes depend on more than whether a juvenile had counsel, we estimated logistic regressions to determine the independent impacts of the three types of representation on the likelihood of adjudication and placement controlling for other characteristics that affect these likelihoods. In addition, we sought to determine which characteristics were most strongly associated with these likelihoods. DATA ----------------------------------------------------- Appendix III:2.1 We used juvenile court case-level data from California, Nebraska, and Pennsylvania for cases disposed in calendar years 1990 and 1991.\3 We obtained these data from the National Juvenile Court Data Archive (NJCDA) of the NCJJ. We obtained the data on magnetic tape, and we performed our own analysis of these data. We consulted with research staff at NCJJ to obtain assistance in understanding the data and in classifying certain variables for analysis. For example, we obtained assistance in categorizing the specific statutory offenses or behaviors that were recorded on the data tapes into offense categories and offense types that we could use in the analysis. The data from California were limited to five of the largest counties (Alameda, Los Angeles, San Francisco, San Jaoquin, and Ventura) that reported their data to the NJCDA. These counties represented about 40 percent of the state's juvenile population between ages 10 and 17. For Pennsylvania, we used only the 1991 data, and we did not include data from Philadelphia. In 1991, 14,584 cases were petitioned to juvenile courts in Pennsylvania in counties other than Philadelphia county, where there were 8,102 cases. Those petitioned in Philadelphia county were not directly comparable with those in the rest of the state because Philadelphia petitions all juvenile cases to court--regardless of the severity of the offense. In the rest of the state, petitioned cases are usually limited to the more serious offenses or more serious offenders. In addition, we limited ourselves to the 1991 data because in 1990, Pennsylvania did not report data on the offense history of juveniles, and therefore we could not analyze representation rates by looking at offense history. For similar reasons, we did not include data from Philadelphia county for either year. In addition, Philadelphia also did not report on the prior offense history of juveniles and certain other variables that we used in our analysis. We were able to report, however, the overall representation rates for Pennsylvania in 1990 and Philadelphia in 1991. (See the notes in table III.3.) Nebraska reported data from all of its counties in 1990 and 1991. However, Nebraska had the highest percentage of cases across the three states in which the type of counsel was not reported (about 18 percent). In addition, we had to omit cases from one of the largest counties in Nebraska because of the way it measured legal representation. According to a state official, unlike the other counties, in that one large county, the type or lack of legal representation is recorded when a juvenile arrives at intake. At intake, few juveniles are accompanied by attorneys; therefore, most cases in that county are reported as not having counsel. However, the state official indicated that about 75 percent of juveniles who appeared before judges at adjudicatory hearings in that county were represented by attorneys even though the reporting system indicated that they were unrepresented. Because the data systems did not indicate this change in representation, we deleted these cases from our analysis. There were 905 cases petitioned to juvenile court in this county, or about 16 percent of all petitioned cases in Nebraska. -------------------- \3 These were the most recent data that we were able to obtain from NCJJ at the time we began our analysis. SAMPLE SELECTION ISSUES ----------------------------------------------------- Appendix III:2.2 For all three states, we limited our analysis to those cases that were petitioned to or formally handled by the juvenile courts for delinquency offenses. This choice was made irrespective of the final disposition of cases. CRIME CATEGORIES ----------------------------------------------------- Appendix III:2.3 To analyze these data, we classified detailed information on the offenses for which juveniles were petitioned to court into broader offense categories. For example, the original California data identified offenses by the particular California statute and section that was violated. More than 1,000 potential types of violations were listed. In Pennsylvania, a few hundred specific offenses were listed with which a juvenile could be charged. We created seven broad classes or categories of crimes, which we used in the descriptive analysis to show representation rates and differences in juvenile court outcomes between represented and unrepresented juveniles. Within these broader categories, we defined several specific crime categories. We used these more specific crime types in our regression analyses to control for the wide variation in behaviors that were represented in the general crime categories. See table 1 for our classification scheme. For each state we estimated regressions for different sets of crime types. The specific number and types of crimes were determined by the available data. In California, we estimated regressions for 10 different crime types. In Pennsylvania, we estimated 9 and in Nebraska, we estimated regressions for 5 crime types. The determination as to whether or not a regression could be estimated was based on the number of cases falling into each crime type. REGRESSION METHODS ----------------------------------------------------- Appendix III:2.4 We estimated logistic regressions for two dichotomous dependent variables: (1) whether a case was adjudicated as delinquent and (2) whether a case was placed out-of-home. First, we estimated these regressions to determine the independent effects of representation on juvenile court outcomes when we controlled for relevant variables that also might have been associated with these outcomes. This was a refinement of our earlier analysis in which we simply reported the aggregated differences in whether adjudication or placement was likely for counseled versus uncounseled juveniles. Second, we estimated the regressions to determine whether variables representing the presence of counsel or other factors were more strongly associated with the two outcomes. We did this to determine the relative importance of counsel in predicting juvenile court outcomes. Third, we estimated the regressions to analyze the effects of different types of counsel on these two outcomes. We estimated the independent effects of having no counsel, privately retained counsel, court-appointed private counsel, and public defender representation on the likelihood of adjudication and placement. We were able to note differences in the effects or impacts of the types of counsel for various types of crime across the three states. We fit the data using maximum likelihood techniques.\4 We tested alternative specifications using likelihood ratio tests to arrive at our final models.\5 We used the parameters from the models and the means of the variables in the models to estimate probabilities in order to compare the differences in the likelihood of adjudication or placement resulting from lack of representation and from different types of counsel. We made these comparisons by estimating the overall probabilities from the variables in the models. Then we estimated separate sets of probabilities for each type of counsel. The differences in the estimated probabilities give an indication of the differences in the likelihood of obtaining an outcome for a particular type of counsel, as compared with the overall estimated likelihood of obtaining that outcome. In addition, because the parameters from logistic regressions are not easily interpretable, we exponentiated the estimated parameters to form the odds ratio for each variable in the final models. The odds ratios have a relatively straightforward interpretation. The odds ratio is an estimate of how much more or less likely it is for the outcome of interest to be present among those having a particular characteristic than those not having that characteristic. For example, an odds ratio of four for a variable indicating that a case was unrepresented would be interpreted to show that unrepresented cases were four times as likely as represented cases to have the outcome of interest. Our discussion of the effects of variables on the outcomes is based largely on the use of odds ratios. -------------------- \4 We used the Statistical Analysis System (SAS) logistic procedure to estimate the parameters in the models. SAS uses iteratively reweighted least squares to compute its estimates. See, for example, SAS/STAT User's Guide, Volume 2, GLM-VARCOMP, Version 6, Fourth Edition. Cary, NC: The SAS Institute, 1990. \5 As part of this effort, we used stepwise regression procedures to eliminate variables that did not add to the overall "goodness of fit" of each of the models. VARIABLES USED IN THE REGRESSION ANALYSES ----------------------------------------------------- Appendix III:2.5 As mentioned, we estimated models for two dependent or outcome variables: (1) whether a case was adjudicated as delinquent and (2) whether a case was placed out-of-home. We used three broad categories of independent variables in these regressions and specific variables used in the models fell within these categories. On the basis of existing literature, we used (1) demographic, (2) offense-related, and (3) court-related broad classes of variables. Demographic variables included the juvenile's age, race, and gender. We included these variables in the analysis because the literature on juvenile court dispositions identifies them either as important determinants of outcomes or as indicators of important hypotheses about judicial decisionmaking. For example, age is important in determining outcomes because judges may not want to commit very young juveniles to out-of-home placements. Gender and race may be important indicators of bias. Offense-related variables included the current offense, more or less serious offense behavior within offense types, and prior offense history. These variables give indications of the seriousness of the behavior for which a juvenile is being judged and the seriousness of the offender. More serious offenses and more serious offenders (e.g., "career offenders") generally are viewed as more likely to be adjudicated as delinquent or to receive out-of-home placements. Omitting these variables may lead to overestimates of the effects of types of or lack of counsel on juvenile court outcomes. Finally, court-related variables included those related to judicial processing and those related to location or geography. The court processing variables included whether a case was represented and if so the type of counsel, detention prior to adjudication, the length of time it took to dispose of a case, the source of referral to juvenile court, whether a juvenile was under supervisory status at the time of referral, and whether charges changed between referral and disposition. The location variables included whether the court was in a metropolitan or nonmetropolitan county\6 and the population density of the county in which the case was heard. These variables were included for different reasons. The measures on types of and lack of counsel were related to the key questions of this analysis. Detention prior to adjudication or placement gave another indication of the severity of the case or of the need to ensure that a juvenile would appear for a hearing. The length of time it took to dispose of a case may indicate the lack of evidence in a case, the weakness of a case, or the complexity of a case. Cases disposed of quickly may have been stronger or simpler cases. Change of charges, particularly a decrease in the severity of the charges, indicates the severity of the case. Finally, location variables at a minimum helped to classify courts into types of counties. The literature suggests that rural courts are characterized by less formal processes in which a judge may know the juvenile or juvenile's family personally and use that information in deciding cases. Urban courts are characterized as having more formal mechanisms for deciding cases. Not all variables were available in each state's database and, in some cases, variables were measured differently across the states because of the way events were recorded in their juvenile courts. Table III.1 shows the variables available for use, whether they were used in a state, and if so, how they were measured in that state. Table III.1 Variables Used in the Analysis Variable California Pennsylvania Nebraska -------------------- ------------------ ------------------ ------------------ Dependent -------------------------------------------------------------------------------- Adjudication Whether a Same as in Same as in petitioned California. California. delinquent offender was adjudicated as delinquent. Placement Whether an Same as in Same as in adjudicated California. California. delinquent received an out- of-home placement. Independent -------------------------------------------------------------------------------- Demographic -------------------------------------------------------------------------------- Age Age of the Age of the Age of the juvenile in years juvenile in years juvenile in years at the time the at the time the at the time the case was disposed. case was referred case was referred to juvenile court. to juvenile court. Race Two dummy Two dummy Two dummy variables: variables, as in variables, as in (1) a variable to California. California. indicate whether a juvenile was black or not black and (2) a variable to indicate whether a juvenile belonged to the "other" race category. Gender A dummy variable Same as in Same as in to indicate that California. California. the person in the case was male. Offense-related and juvenile court processing -------------------------------------------------------------------------------- Type of counsel Dummy variables to Same as in Same as in indicate that (1) California. California. the case was uncounseled, (2) the case was represented by privately retained counsel, (3) the case was represented by court-appointed private counsel, and (4) the case was represented by a public defender. (Public defender was the reference category in the models.) Current offense Offenses were Offenses were Offenses were classified into 1 classified into classified into of 10 categories one of nine one of six on the basis of categories on the categories on the the availability basis of the basis of the of data. These availability of availability of were data. These were data. These were --murder --rape --aggravated --rape --robbery assault --robbery --aggravated --burglary --aggravated assault --larceny assault --burglary --simple assault --burglary --larceny --weapons --larceny --simple assault --drugs --simple assault --weapons --weapons --drug possession --drug possession --drug trafficking --drug trafficking Offense severity Within types, Same as in N/A offenses were California. classified as more or less severe, on the basis of characteristics of offense behavior within specific crime types, e.g., attempts versus completed. Offense history N/A Number of prior Number of prior juvenile court referrals to dispositions. juvenile court. These were recoded These were recoded into dummy into dummy variables: (1) no variables: (1) no priors, (2) one to priors, (2) one to two priors, and two priors, and (3) three or more (3) three or more priors. (No priors priors. (No priors was the reference was the reference category in the category used in models.) the models.) Charges changed N/A A dummy variable N/A (dropped) between to indicate that time of referral charges filed at and referral were disposition different from the charges at disposition. (The change usually indicated a decrease in the severity of the offense.) Length of time to N/A Two variables were Two variables were disposition used to measure used to measure the length of time the length of time to disposition: to disposition: (1) the length of (1) the length of time between time between referral and referral and disposition, in disposition, in weeks; and (2) the weeks; and (2) the square of the square of the length of time to length of time to disposition. The disposition. The square of the square of the length of time was length of time was used to measure used to measure nonlinear effects nonlinear effects of time on the of time on the placement placement decision. decision. Detention prior to A dummy variable Same as in Same as in adjudication to indicate that a California. California. case was detained securely prior to adjudication. Source of referral A series of dummy Same as in Same as in to juvenile court variables to California. California. indicate the source of referral to juvenile court. These included variables to indicate that (1) the case was referred by law enforcement officials; (2) it was referred by family members or friends; (3) it was referred by school officials; or (4) it was referred by other officials, such as probation or court officials. ("Other officials" was the reference category.) Supervisory status A dummy variable N/A N/A at the time of to indicate that a referral case was under ward probation at the time of referral. Geographic Indicators -------------------------------------------------------------------------------- Metropolitan status N/A A dummy variable A dummy variable to indicate that to indicate that the court in which the court in which a case was heard a case was heard was located in a was located in a county within a county within a metropolitan metropolitan statistical area. statistical area. Population density The population Same as in Same as in density per 1,000 California. California. residents of the county containing the court in which a case was heard. County indicators A series of dummy N/A N/A variables to indicate the county in which a case was heard. (These were created for California because its data came from five counties, all of which fell in metropolitan areas.) -------------------------------------------------------------------------------- In addition to these variables, we included other variables to indicate missing values for the variables measuring type of counsel and prior offense history for the crime types in which there were comparatively large numbers of missing cases (e.g., greater than 10 percent of the total number of cases). We included these variables to determine if missing values were associated with the dependent variables. -------------------- \6 We defined metropolitan counties as those counties falling within a primary metropolitan statistical area (PMSA) or metropolitan statistical area (MSA), as defined by the U.S. Bureau of the Census. Nonmetropolitan counties fell outside of PMSAs or MSAs. These designations were chosen as proxies for urban and less densely populated areas. The data we obtained did not have measures of location that would have permitted us to identify urban and rural areas, as defined by Census. For California, all of the cases came from metropolitan counties. For Pennsylvania, about 90 percent of the cases were from metropolitan counties. For Nebraska, about 55 percent of all cases were from metropolitan counties. OMITTED VARIABLES ----------------------------------------------------- Appendix III:2.6 As in all modeling efforts, the results are interpretable only in the context of the variables included in the models. Excluding relevant variables from the models, such as a juvenile's demeanor or a judges's personal knowledge of a case, could affect the estimates of the effects of variables in the models. It is not possible, a priori, to determine the direction or magnitude of such effects of all potentially relevant variables that may have been excluded from the models. We tried to produce models that were reasonable, that is, that included at least some of the important variables identified in the literature. Ultimately, we were constrained in our efforts by the variables contained in each state's database. SAMPLE SELECTION AND CASES ANALYZED ----------------------------------------------------- Appendix III:2.7 We estimated separate regressions for up to 10 crime types for the adjudication and placement outcomes in each state. Table III.2 shows the crime types for which we estimated models. We selected these crime types primarily because of the number of cases available for analysis with the exception of public order offenses. We did not estimate models for public order offenses largely because of the relative lack of seriousness of these offenses. We also did not estimate models for indeterminate offenses because we could not ascertain the crime type. We did not estimate models for other offense categories because of the wide variation in crime types falling into that category. For the adjudication models, we selected those cases that were petitioned to juvenile court or otherwise handled more formally at the initial processing of the case.\7 For the placement models, we selected only those cases that were adjudicated as delinquent. Table III.2 Crime Types Used in the Adjudication and Placement Regression Models Crime type California Pennsylvania Nebraska ----------- -- ------------- ------------- ------------- Murder X \a \a Rape X X \a Robbery X X \a Aggravated X X \a assault Burglary X X X Larceny X X X Drug X X X\b trafficking Drug X X X\b possession Simple X X X assault Weapons X X X ------------------------------------------------------------ \a No model was constructed for this crime type generally because of the limited number of cases. \b Drug trafficking and drug possession crimes were combined in Nebraska. -------------------- \7 Because states' procedures for handling cases formally and informally vary, this choice did not necessarily exclude cases whose final disposition was made by an informal adjustment. For example, in Pennsylvania, there are differences between Philadelphia county and the rest of the state in terms of which cases are petitioned to juvenile court. Philadelphia county petitions all cases, regardless of seriousness, and then decides whether to dispose of cases formally through adjudication or informally by another mechanism, such as the consent decree. Further, regardless of the county in Pennsylvania, consent decrees are used as informal adjustments after cases are petitioned to court. In cases disposed of by consent decrees, a juvenile master, rather than a juvenile court judge, hears the case; juveniles do not appear for an adjudicatory hearing; and juveniles tend not to receive legal representation in these cases. RESULTS ------------------------------------------------------- Appendix III:3 The results of our analyses are discussed in two sections. First, we reviewed the findings on the frequency of representation and on the differences in the probabilities of adjudication and placement for unrepresented juveniles compared with represented juveniles in the detailed crime type classifications. Then we discuss three results from the modeling exercise: (1) the effects of lack of representation on the likelihoods of adjudication and placement; (2) the effects of different types of counsel on the estimated probabilities of adjudication and placement; and (3) the variables that had the strongest associations with the adjudication and placement outcomes. REPRESENTATION RATES ----------------------------------------------------- Appendix III:3.1 As table III.3 shows, the representation rates varied among the states. Overall, juveniles in California were more likely to be represented for all serious crime types. More than 99 percent of all violent offenders and more than 98 percent of all property offenders in California were represented. Only for the relatively less serious crime types, public order offenses, did the representation rate in California drop below 90 percent. Besides that one crime category, over 99 percent of juveniles in California would have been represented. Table III.3 Number of Cases Petitioned to Juvenile Courts and Percentage Represented by Counsel, for Calendar Years 1990 and 1991, by State, for Detailed Crime Types Percentage Percentage Percentage Crime Number represente Number represente Number represente type petitioned d petitioned d petitioned d -------- ---------- ---------- ---------- ---------- ---------- ---------- Murder\d 1,097 99.7 5 100.0 1 100.0 Rape 319 99.1 171 99.4 0 \i Robbery 6,003 99.8 508 97.0 32 96.9 Aggravat 4,505 99.6 1,306 93.1 40 85.0 ed assault Other 961 99.1 1,064 92.8 0 \i violent offense s Indeterm 45 100.0 33 93.9 90 90.0 inate violent offense s Burglary 8,099 99.0 2,450 88.6 389 82.3 Fraud\e 438 98.6 91 83.5 49 59.2 Larceny\ 5,496 98.3 3,769 90.4 1,806 59.5 f Motor 6,454 99.2 360 88.9 171 66.1 vehicle theft Arson 228 98.2 167 94.6 25 72.0 Other 5,247 98.4 1,027 83.4 768 58.9 property offense s Drug 1,456 98.8 277 96.0 112\h 80.4 possess ion Drug 4,885 99.6 484 98.6 \h \h traffic king Other 505 100.0 202 91.6 \h \h drug offense s Simple 3,529 98.7 1,440 89.4 396 75.0 assault Weapons 3,474 99.4 153 90.8 85 90.6 Public 7,105 82.7 624 86.2 85 71.8 order\g Indeterm 864 98.6 453 94.3 458 54.1 inate ================================================================================ Totals 60,710 97.2 14,584 90.6 4,507 64.9 -------------------------------------------------------------------------------- Note: The figures in this table and in tables III.4 and III.5 were computed after excluding cases in which the type of counsel was missing in the database. In California, this amounted to about 5 percent of 64,275 cases and in Pennsylvania about 5 percent of the 15,397 cases. These data are for only those counties for which data were available. \a In California, data were available only from five counties. These included some of the largest counties, and they represented about 40 percent of the juvenile population between ages 10 and 17. \b Data for Pennsylvania were for 1991 only, and they did not include Philadelphia county. Philadelphia accounted for an additional 8,102 cases in 1991, and there were no reported cases of unrepresented juveniles in Philadelphia. (See the text for more details on the data limitations.) \c Data for Nebraska exclude one of the larger counties, which accounted for 906 cases. These cases were excluded because the variable that indicated whether cases were represented by legal counsel was incorrectly recorded. \d The murder category includes the crimes of murder and nonnegligent manslaughter. \e The fraud category includes the crimes of fraud, forgery, and embezzlement. \f The larceny category includes unspecified larcenies in Pennsylvania and Nebraska. \g The public order category includes other nonviolent offenses. \h Drug crimes in Nebraska were not broken down by type; therefore, we could not distinguish between drug possession and drug trafficking there. \i No cases were reported. Source: GAO analysis of NCJJ data. In Pennsylvania, the representation rates were similar, although slightly lower, than California's. Nevertheless, 94 percent of all violent offenders in Pennsylvania were represented. Apart from fraud (with relatively few cases) and other property offenses, more than 88 percent of Pennsylvania's property offenders were represented. As in California, for comparatively less serious offenses, the representation rates were slightly lower in Pennsylvania. In Nebraska, however, representation rates are comparatively low at about 65 percent overall. These rates seem to decrease as crimes become less serious. For example, violent offenders in Nebraska generally are more likely to be represented than property offenders. COMPARATIVE PROBABILITIES OF ADJUDICATION ----------------------------------------------------- Appendix III:3.2 Table III.4 shows the comparative probabilities of adjudication for represented and unrepresented juveniles for the detailed crime type classifications. In general, for California and Pennsylvania, unrepresented juveniles were less likely to be adjudicated than those represented. In Nebraska, conversely, unrepresented juveniles were more likely to be adjudicated as delinquent. Table III.4 Percentage of Petitioned Cases Adjudicated as Delinquent, by State and by Whether Represented or Unrepresented by Counsel, for Calendar Years 1990- 1991 Crime With Without With Without With Without type counsel counsel counsel counsel counsel counsel -------- ---------- ---------- ---------- ---------- ---------- ---------- Murder\d 59.2 100.0\i 80.0\i \k 100.0\j \k Rape 69.2 33.3\i 57.3 0.0\i \k \k Robbery 67.2 58.3\i 75.1 41.7\i 74.2 100.0\j Aggravat 68.2 70.6\i 75.7 23.5 73.5 83.3\i ed assault Other 67.7 44.4\i 62.9 26.5 \k \k violent offense s Indeterm 82.2 \k 57.1 0.0\i 74.1 88.9 inate violent offense s Burglary 71.2 53.8 77.4 31.3 77.2 89.9 Fraud\e 64.6 66.7\i 55.1 25.0\i 79.3 100.0\i Larceny\ 71.0 70.2 72.9 33.8 82.0 93.4 f Motor 72.9 50.9 76.8 44.4 90.3 89.7 vehicle theft Arson 75.3 50.0\i 67.9 75.0\i 83.3 100.0\i Other 65.7 69.9 65.6 26.4 84.1 93.7 property offense s Drug 70.7 61.1\i 68.4 54.5 81.1\h 100.0\h possess ion Drug 76.1 54.5 79.5 66.7 \h \h traffic king Other 71.3 \k 69.1 20.0\i \h \h drug offense s Simple 66.5 53.3 61.0 31.5 75.1 86.9 assault Weapons 71.9 52.4 71.6 20.0\i 76.6 87.5\i Public 69.9 52.9 65.6 21.1 90.2 87.5 order\g Indeterm 65.9 91.7\i 78.7 33.3 90.7 94.8 inate offense s ================================================================================ Total 69.9 55.3 71.3 30.9 81.8 93.0 -------------------------------------------------------------------------------- \a In California, data were available only from five counties. These included some of the largest counties, and they represented about 40 percent of the juvenile population between age 10 and 17. \b Data for Pennsylvania were for 1991 only, and they did not include Philadelphia county. Philadelphia accounted for an additional 8102 cases in 1991, and there were no reported cases of unrepresented juveniles in Philadelphia. (See the text for more details on the data limitations.) \c Data for Nebraska excluded one of the larger counties, which accounted for 906 cases. These cases were excluded because the variable that indicated whether cases were represented by legal counsel was incorrectly recorded. \d The murder category includes the crimes of murder and nonnegligent manslaughter. \e The fraud category includes the crimes of fraud, forgery, and embezzlement. \f The larceny category includes unspecified larcenies in Pennsylvania and Nebraska. \g The public order category includes other nonviolent offenses. \h Drug crimes in Nebraska were not broken down by type; therefore, we could not distinguish between drug possession and drug trafficking there. \i Percentages were based on 20 or fewer total unrepresented cases. \j Based on one case. \k No cases were reported. Source: GAO analysis of NCJJ data. Table III.4 shows the relatively few unrepresented juvenile cases in California and Pennsylvania for many crime types. In California, for four of the nine crime types--all having fewer than 20 unrepresented juvenile cases--the probability of adjudication was higher for unrepresented juveniles. Overall in California, only 946 of the 41,985 adjudicated cases (or 2.2 percent) in table III.3 were unrepresented. In Pennsylvania, of the 8,488 cases that were adjudicated, only 305 (or about 3 percent) were unrepresented. The differences in the likelihood of adjudication between unrepresented and represented cases in California are based on relatively few unrepresented juvenile cases. For Pennsylvania, the same caveat holds true, that is, differences in the likelihood of adjudication between represented and unrepresented juveniles are based on a relatively small number of unrepresented cases. In Pennsylvania, only for arson did the probability of adjudication for unrepresented juveniles exceed that of represented juveniles, and there were fewer than 20 arson cases. This suggests that a higher likelihood of adjudication for unrepresented juveniles as compared with represented juveniles in California and Pennsylvania was a relatively rare event. The higher likelihood in these cases may be due to specific circumstances in the case that may not have been measured or recorded in the state's juvenile court databases. In Nebraska, however, the probability of adjudication is higher for unrepresented juveniles than for represented juveniles for many crime types, whether or not there are relatively few unrepresented juvenile cases. Overall, 2,156 of the 4,580 adjudicated cases (47 percent) were unrepresented in Nebraska. COMPARATIVE PROBABILITIES OF PLACEMENT ----------------------------------------------------- Appendix III:3.3 Table III.5 shows the comparative probabilities of placement for unrepresented and represented juveniles adjudicated in the three states. For most crime types across the three states, unrepresented juveniles were less likely to receive out-of-home placements than represented juveniles, and, as with the adjudication outcomes, there were a number of crime types with relatively few unrepresented cases (i.e., fewer than 20 cases) where juveniles faced placement decisions. Table III.5 Percentage of Adjudicated Cases Receiving Out-of-Home Placements, by State, Whether Represented or Unrepresented by Counsel, for Calendar Years 1990-1991 Crime With Without With Without With Without type counsel counsel counsel counsel counsel counsel -------- ---------- ---------- ---------- ---------- ---------- ---------- Murder\d 89.1 66.7\i 75.0\h \k 100.0\i \k Rape 58.2 0.0\j 37.2 \k \k \k Robbery 67.6 57.1\i 34.4 20.0\i 39.1\i 0.0\i Aggravat 54.4 33.3\i 26.9 12.5\i 28.0\i 0.0\i ed assault Other 54.8 50.0\i 28.2 15.4\i \k \k violent offense s Indeterm 73.0 \k 25.0\h \k 23.3 0.0 inate violent offense s Burglary 44.1 21.4 26.8 10.6 23.9 6.5 Fraud\e 37.1 50.0\i 7.9 0.0\i 21.7\i 10.0\i Larceny\ 44.7 34.8 21.3 13.0 15.6 3.2 f Motor 51.5 48.1 25.5 25.0\i 27.5 21.2 vehicle theft Arson 36.3 50.0\i 11.2 0.0\i 6.7\i 14.3\i Other 42.2 41.4 20.9 6.1 13.7 1.7 property offense s Drug 52.0 54.5\i 32.9 16.7\i 30.1\h 4.5\h possess ion Drug 49.8 58.3 35.7 0.0\i \h \h traffic king Other 62.2 \k 23.7 0.0\i \h \h drug offense s Simple 41.9 50.0 24.7 11.8 11.2 7.0 assault Weapons 44.3 18.2\h 19.3 0.0\i 11.9 0.0\i Public 66.1 81.0 35.4 31.3\i 7.3 4.8\i order\g Indeterm 53.7 54.5\i 47.5 14.3\i 14.2 3.0 inate offense s ================================================================================ Total 51.8 68.1 26.6 12.8 16.9 4.0 -------------------------------------------------------------------------------- \a In California, data were available only from five counties. These included some of the largest counties, and they represented about 40 percent of the juvenile population between age 10 and 17. \b Data for Pennsylvania were for 1991 only, and they did not include Philadelphia county. Philadelphia accounted for an additional 8,102 cases in 1991, and there were no reported cases of unrepresented juveniles in Philadelphia. (See the text for more details on the data limitations.) \c Data for Nebraska excluded one of the larger counties, which accounted for 906 cases. These cases were excluded because the variable that indicated whether cases were represented by legal counsel was incorrectly recorded. \d The murder category includes the crimes of murder and nonnegligent manslaughter. \e The fraud category includes the crimes of fraud, forgery, and embezzlement. \f The larceny category includes unspecified larcenies in Pennsylvania and Nebraska. \g The public order category includes other nonviolent offenses. \h Drug crimes in Nebraska were not broken down by type; therefore we could not distinguish between drug possession and drug trafficking there. \i Percentages were based on 20 or fewer total unrepresented cases. \j Based on one case. \k No cases were reported. Source: GAO analysis of NCJJ data. Some exceptions occurred in California. In most of the cases in which unrepresented juveniles were more likely than those represented to be placed, there were fewer than 20 unrepresented juvenile cases. The exception to this rule was for other property offenses, but these cases tend to include relatively minor property offenses, such as possession of stolen property. For Pennsylvania and Nebraska, arson in Nebraska was the only crime type in which unrepresented juveniles were more likely to be placed than those represented, and that result was based on fewer than 20 unrepresented juvenile cases. The major difference between California and both Pennsylvania and Nebraska was the prevalence with which each state used out-of-home placements as a juvenile court disposition. California used placements at a higher rate than the other two states. Consequently, unrepresented juveniles in California face an overall risk of placement that is somewhat higher than in the other states, despite the fact that there were relatively few unrepresented juvenile cases being placed in California. However, because so few juveniles were unrepresented in California, the likelihood that a juvenile was both unrepresented and placed was lowest there of the three states. RESULTS FROM THE REGRESSION MODELS ------------------------------------------------------- Appendix III:4 Because we found differences in the comparative probabilities of adjudication and placement between unrepresented and represented juveniles, we looked more closely at the likelihood of receiving these outcomes as a function of whether a case received legal representation and the other characteristics of a case that may determine these outcomes. Thus, we used the regressions first to determine if the lack of representation had an independent effect on the likelihood of adjudication and placement. Second, given the relatively small number of unrepresented cases in California and Pennsylvania, we analyzed the effect on adjudication and placement outcomes of the different types of counsel. Third, given the importance of offense severity and offense history to judicial decisions, we used the regressions to determine whether these and other variables had stronger effects on the likelihood of adjudication and placement than did the lack or type of representation. In interpreting the regression results, two caveats should be considered. First, selection effects can influence outcomes at each stage. Selection effects refer to the process whereby juveniles having selected characteristics are the ones who move onto a particular stage of processing. The characteristics used to select the juveniles who move onto the next stage are important if they also are the characteristics used to determine whether or not a juvenile receives representation. For example, if juveniles who commit less serious crimes are less likely to be adjudicated and less likely to receive representation because they are less likely to be adjudicated, then these selection effects may cause our regressions to over-estimate the effects of representation on juvenile court outcomes. Second, although we used relatively narrow crime types, the crime types encompass a relatively wide range of behaviors for which data were not available. For example, data on victim injury in robbery or assault cases were not available, but victim injury may be important in judges' decisionmaking. Similarly, these unmeasured behaviors may also be related to the provision of counsel and thus have some effect on the estimates of counsel's impact on juvenile court outcomes.\8 -------------------- \8 The data in tables III.3, III.4, and III.5 report probabilities of various outcomes. Some of these are quite high (e.g., California's percentage represented). If the grouped data were used in the regressions, we would encounter problems of restricted variances and variables would not covary with the dependent variable. By using the individual-level data and estimating logistic regressions, we eliminated this potential problem. ADJUDICATION OUTCOMES ----------------------------------------------------- Appendix III:4.1 In the following discussion, the results of our analysis of adjudication outcomes are considered on a state-by-state basis. The effects of the lack of representation on adjudication are discussed followed by the effects of different types of representation. The effects of other variables on adjudication concludes the review of each state. CALIFORNIA --------------------------------------------------- Appendix III:4.1.1 In California, almost all adjudicated cases were represented. Specifically, of the 41,985 cases reported as adjudicated on table III.3, only 946 were unrepresented. As a result, there were relatively few unrepresented cases from which to derive estimates of the effects of lacking representation on the likelihood of adjudication.\9 Nevertheless, for 7 of the 10 crime types for which we estimated adjudication models in California, we found that the lack of representation did not exert independent effects on the likelihood of adjudication. For burglary, larceny, and weapons, the lack of representation had an independent effect on the likelihood of adjudication after controlling for the other relevant variables identified in table III.1.\10 For burglary and weapons, the lack of representation was associated with lower likelihoods (as compared with represented cases), but for larceny it was associated with a higher likelihood of adjudication. Specifically, unrepresented burglars were about one-half as likely to have their cases adjudicated as those represented. Unrepresented juveniles with cases involving weapons were about four-tenths as likely to have their cases adjudicated as those represented. Conversely, unrepresented juveniles with larceny cases were about one and one-half times more likely to be adjudicated than those represented. For robbery and drug trafficking cases, the lack of representation was negatively associated with the likelihood of adjudication, but the effects were too small to be considered meaningful. The lack of representation had some independent effects on the likelihood of adjudication for some crime types in California. However, the effects operated in opposite directions across crimes, were not widespread across crime types, and generally were small relative to the size of the effects of other variables. The influence of the type of representation was also difficult to characterize. In California, most cases were represented by either public defenders or by court-appointed private representation. Privately paid attorneys represented juveniles in about 3 percent of the cases. Thus, the main comparisons between types of representation are between public defenders and court-appointed private representation. In general, the effects of these two types of representation on the likelihood of adjudication were mixed. Juveniles represented by court- appointed private attorneys were more likely to be adjudicated than those represented by public defenders, for drugs and simple assault, but less likely for rape and aggravated assault. Thus, court-appointed private attorneys likely decrease the chances of adjudication in the more serious crimes. Of course, this result may be due in part to the public defender workload and other factors. The absence of consistent effects of the lack or type of representation on the likelihood of adjudication indicates that variables other than representation may be more responsible for determining judicial decisions than the type of representation. In fact, this appears to be the case. We used the models to identify those variables that had consistently large effects across equations. We found that three variables had large effects on the likelihood of adjudication across most types of crimes. These variables were whether a case was under court supervision (ward probation) at the time of referral, whether a juvenile was detained prior to adjudication, and the location of the court (i.e., the county in which a case was heard). A fourth variable, the age of the juvenile, also was strongly associated with the likelihood of adjudication, but the effects of age were not as consistent nor strong as these other variables. Juveniles "on ward probation"\11 at the time the current offense was committed were from 1.6 to 3.1 times more likely to have their cases adjudicated than those not on ward probation for the crimes of robbery, aggravated assault, simple assault, burglary, larceny, weapons, drug possession, and drug trafficking. Across crime types, juveniles detained prior to adjudication were from 1.2 to 1.5 times as likely as those not detained to have their cases adjudicated. Across different locations, the likelihood of adjudication varied. Depending upon the location of the court and the type of crime, juveniles could be more likely to have their cases adjudicated (from 1.2 to 3.9 times as likely) or less likely to have them adjudicated (about one-half as likely). Older juveniles were slightly less likely to have their cases adjudicated than younger juveniles. Finally, for some specific crime types, other variables, such as race and gender, were associated with adjudicatory outcomes. For example, males referred for aggravated assault were slightly less likely to have their cases adjudicated than females referred for that crime. In addition, blacks referred for drug possession were less likely than persons of other races to have their cases adjudicated. -------------------- \9 The relatively small number of unrepresented cases constrained our choice of model specification to using a dummy variable to indicate the absence of counsel. We would have preferred to estimate separate equations for unrepresented and represented cases and compare the parameter estimates across the equations for all variables in the models. This latter approach is equivalent to estimating one model with a complete set of interactions with the counsel dummy variable. It has the added advantage, however, of not imposing the assumption of equal variances between the uncounseled and counseled equations, as is implicit in the fully saturated model. The implications of constraining the variances relate mostly to hypothesis tests on individual variables. \10 The final equations for each crime type contained a different subset of the variables listed on table III.1. The final variables in the models were determined by the contribution of individual variables to the overall goodness of fit of the model to the data. The parameter estimates, odds ratios, and variances for variables used in the 28 adjudication models are available upon request from us. \11 Ward probation refers to supervision of a juvenile as a "ward of the court." PENNSYLVANIA --------------------------------------------------- Appendix III:4.1.2 As was the case in California, most adjudicated cases in Pennsylvania were represented. Of the 8,488 adjudicated cases, only about 3 percent were unrepresented. Thus, while we tested for the effects of the lack of representation on adjudication outcomes, unrepresented cases were rare, and unrepresented adjudicated cases were even rarer. For two of nine crime types in our regressions, the lack of representation exerted an independent and negative effect on the likelihood of adjudication. Burglary and simple assault cases that were unrepresented were less likely to be adjudicated than cases in those crime types that were represented.\12 Unrepresented juveniles in burglary cases were about four-tenths as likely to have their cases adjudicated as those who were represented; unrepresented juveniles in simple assault cases were about half as likely. There were generally no large differences in the effects of the various types of representation on the likelihood of adjudication across the crime types. Thus, type of representation also did not have consistent effects on adjudicatory outcomes. In Pennsylvania, the variables that were most strongly associated with the likelihood of adjudication were whether a consent decree was ordered and whether a case was detained prior to adjudication. The type of offense was strongly associated with the likelihood of adjudication, but its effects were not consistent across crime types. Consent decrees are used in place of adjudications for certain, relatively less serious cases in Pennsylvania. Most unrepresented juvenile cases for robbery, aggravated assault, burglary, larceny, and simple assault were disposed of by consent decrees. Juveniles detained prior to adjudication increased the likelihood of having their cases adjudicated, relative to juveniles not detained, for all crimes except weapons. Juveniles detained were from two to seven times more likely to have their cases adjudicated than those not detained. In simple assault and burglary cases--where the gap between unrepresented and represented cases persisted--variables other than the lack of representation had larger effects on the likelihood of adjudication than did the lack of representation. Cases disposed of by consent decrees; cases in which juveniles have been detained prior to adjudication; and cases where either the location of the court (for simple assault), the gender, and the length of time before disposition (for burglary) all produced larger changes in the likelihood of adjudication than did the lack of representation. In addition, race exerted independent influences on the likelihood of adjudication in robbery and rape cases and gender exerted independent influences on the likelihood of adjudication in burglary, larceny, and drug possession cases. Black juveniles referred for rape or robbery were less likely (0.102 and 0.255 times) than other juveniles to have their cases adjudicated. Males referred for burglary, larceny, and drug possession were more likely than females (2.63, 1.5, and 2.5 times) to have their cases adjudicated. Thus, as in California, variables associated with case processing were more strongly associated with the likelihood of adjudication in Pennsylvania than with the lack of representation for most crime types. -------------------- \12 Unrepresented juveniles in weapons cases also were less likely to have their cases adjudicated than those who were represented; however, this finding was based on fewer than 10 cases, 2 of which were adjudicated. In our view, this was too few cases for a substantive finding. NEBRASKA --------------------------------------------------- Appendix III:4.1.3 Unlike California and Pennsylvania, juveniles in Nebraska had about half (47.1 percent) of their adjudicated cases unrepresented. Further, Nebraska adjudicated about one-ninth as many cases as California, but Nebraska also had more than twice the number of unrepresented adjudicated juvenile cases as California. Similar discrepancies occurred in comparing Nebraska to Pennsylvania. Thus, Nebraska's case is very different from the other two states. The results from the regressions show that Nebraska differs from California and Pennsylvania. First, after controlling for juveniles' characteristics, unrepresented juveniles were less likely to be adjudicated only for larceny--a relatively less serious property offense. In larceny cases, unrepresented juveniles were about four-tenths as likely to have their cases adjudicated as those who were represented. For all other crime types, unrepresented and represented juveniles were about equal in the probability of having their cases adjudicated, after controlling for juveniles' characteristics. Second, the type of representation was not systematically related to the likelihood of adjudication across crime types. Juveniles represented by court-appointed private attorneys were about half as likely to have their cases adjudicated than those represented by public defenders for simple assault and larceny crimes. Juveniles represented by privately paid attorneys were half as likely to have their cases adjudicated in simple assault cases as compared with those represented by public defenders. In weapons cases, however, juveniles represented by privately paid attorneys were about 10 times more likely to have their cases adjudicated than those represented by public defenders. In general, juveniles represented by public defenders were more likely to have their cases adjudicated than those represented by other types of attorneys or those that were unrepresented (except for weapons cases). Third, measures of the location of the court in which a juvenile's case was heard had the largest effect on the likelihood of adjudication across crime types. The population density of the county where the court was located had the strongest association with the likelihood that a case was adjudicated for the crimes of simple assault, burglary, larceny, and weapons. The association was negative, that is, as the population density of the county increased, the likelihood of adjudication decreased within crime categories. In other words, in less densely populated counties juveniles were more likely to have their cases adjudicated as delinquent than in more densely populated counties, controlling for other characteristics. However, burglary and larceny cases heard in metropolitan counties were more likely to have their cases adjudicated than those heard in nonmetropolitan counties. This apparent contradiction may stem from the fact that within metropolitan counties there may be two classes of counties in which the outcomes differ. For example, in the most densely populated metropolitan counties (e.g., those containing the central city of the PMSA or MSA), the likelihood of adjudication may be lowest. (This accounts for the negative relationship between population density and adjudication.) Such a relationship would not invalidate the fact that, on the whole, juveniles in metropolitan counties are more likely than those in nonmetropolitan counties to have their cases adjudicated. For aggravated assault and drug cases, the length of time before disposition also had a strong association with the likelihood of adjudication. For both of these crimes, the longer it took to dispose of a case, the less likely it was to be adjudicated. However, in aggravated assault cases, the type of attorney had an effect. Cases with no attorneys and those with court-appointed attorneys were less likely to have their cases adjudicated, whereas in drug cases, there were no differences in the relative likelihood of adjudication across types of attorneys. Other variables having large, positive effects on the likelihood of adjudication included whether a juvenile was detained (for burglary), prior record (for larceny), and age (for weapons). PLACEMENT OUTCOMES ----------------------------------------------------- Appendix III:4.2 In analyzing out-of-home placements, we restricted the sample to cases adjudicated as delinquent. We estimated equations for the same crime types as in the adjudication analysis. In general, we found that the presence or lack of representation was not associated with the likelihood of placement for most crime types. In part, this was due to the relatively few cases involving unrepresented juveniles that were placed in the states, especially in California and Pennsylvania. For example, in California only 6 of the 2,852 cases adjudicated for aggravated assault in which a juvenile faced a placement decision were unrepresented. In Pennsylvania, only 61 of 1,283 cases adjudicated for burglary in which a juvenile faced a placement decision were unrepresented. In general, where there were sufficient cases to produce a robust estimate of the effect of representation on placement outcomes, we estimated models that included representation variables. There were exceptions to this general finding of no effect of the presence or lack of representation on the likelihood of placement. Specifically, in California, unrepresented juveniles were three to five times more likely to be placed, in larceny and drug trafficking cases, than were represented juveniles. In Pennsylvania, unrepresented juveniles were two and one half to three times less likely to be placed than those who were represented in burglary and simple assault cases. In Nebraska, unrepresented juveniles were half as likely to be placed in burglary cases and about two-fifths in larceny and one-tenth as likely to be placed in drug cases than were represented juveniles. The influence of the type of attorney on the likelihood of placement varied across the states. In California for all crimes except murder, juveniles represented by court-appointed private attorneys were more likely to be placed than those represented by private attorneys or by public defenders. Juveniles represented by privately paid attorneys were less likely to be placed than those represented by public defenders in robbery, larceny, weapons, and drug possession and trafficking cases. In Pennsylvania, for all crimes except robbery, simple assault, and drug trafficking, juveniles represented by private representation were from slightly more than one to six times more likely to be placed (depending upon the crime type) than those represented by public defenders. In addition, juveniles represented by court-appointed private attorneys were one and one half to two and one half times more likely to be placed than those represented by public defenders in burglary, larceny, and simple assault cases. In Nebraska, juveniles represented by privately paid attorneys tended to have lower chances of receiving placements than those represented by public defenders for burglary, larceny, and drug cases. A few variables tended to have relatively consistent effects on the likelihood of placement. These included whether a juvenile was detained prior to adjudication, whether charges were changed between the time of referral and disposition, offense history and court supervisory status, and the location of the courts. In California, for example, for the violent crimes of murder, rape, robbery, and aggravated assault, the likelihood of receiving out-of-home placement was highest for juveniles detained prior to adjudication, as opposed to those not detained. These juveniles were from 3.1 to 10.5 times as likely to be placed out-of-home. The likelihood was lowest for juveniles whose charges were changed between referral and disposition.\13 In such cases, these juveniles were 0.08 to 0.38 times less likely to be placed than those whose referral charges were not changed. For the less serious violent crime of simple assault, for property crimes, such as burglary and larceny, and for weapons and drug possession the odds of receiving out-of-home placements were highest for juveniles on ward probation at the time the referral offense was committed and for juveniles detained prior to adjudication. Juveniles on ward probation were from 4.1 to 5.8 times more likely to be placed as compared with those not on ward probation. Juveniles detained were from 2.4 to 3.2 times as likely to be placed as compared with juveniles who had not been detained. In Pennsylvania, across all crime types except rape, detention, prior dispositions, and population density were consistently among the variables most likely to influence the likelihood of a juvenile receiving out-of-home placement. Juveniles detained, for example, were from 2.6 to almost 10 times more likely to be placed; juveniles with priors also were from about 3 to more than 10 times more likely to be placed than juveniles without priors. As the population density of the county in which the juvenile's case was handled increased, juveniles tended to be less likely to be placed and more densely populated counties within metropolitan areas--such as the county containing the central city of the metropolitan area--were generally less likely to use out-of-home placements than the other metropolitan counties. In Nebraska, across all crime types, prior record, detention prior to adjudication, location of the court, and the population density of the county where the court was located were the variables having the largest impacts on the likelihood of placement. Juveniles with priors were from 2 to 6 times more likely to be placed than those without priors, depending on the type of crime, and juveniles detained prior to adjudication were generally from 5 to 10 times more likely to be placed than those not detained. In addition in Nebraska, the effects of the location of the court influenced the likelihood of receiving an out-of-home placement for aggravated assault, simple assault, burglary, and larceny. In nonmetropolitan counties, juveniles who committed these crimes were less likely to be placed, reflecting, perhaps, the relative scarcity of placement options. In addition, as the population density of the county where the court was located increased, juveniles were more likely to be placed. This is consistent with the view that there may be more placement options in urban areas and that the availability of these options influences the likelihood of placement. Finally, a few other variables had noticeable impacts on the likelihood of placement in some cases. For example, in Pennsylvania, race influenced the likelihood of receiving an out-of-home placement for rape. Blacks were 0.044 times less likely than whites and others to be placed for rape. Changes (reductions) in charges reduced the likelihood of placement for rape, robbery and aggravated assault offenses, but changes in charges had no effect on the likelihood of placement for property, weapons, or drug offenses. In California, race and gender affected outcomes for some crime types. Specifically, for drug possession and drug trafficking, blacks were 2.1 and 1.5 times as likely as other persons to be placed. For drug possession, males were 2.5 times as likely as females to be placed. -------------------- \13 In almost all cases, charges were reduced between referral and disposition. JUVENILES' RIGHT TO COUNSEL IN 15 SELECTED STATES ========================================================== Appendix IV Legal State authority Right to counsel Waiver of counsel -------------- -------------- ----------------------- ----------------------- Arizona Ariz. Rev. A juvenile has the Waiver is authorized by Stat. Ann. right to representation statute when the sec. 8-225; by an attorney in juvenile and the parent Ariz. R. Juv. delinquency or guardian waive it. Ct. 6(c). proceedings. The court must find that the child knowingly, intelligently, and voluntarily waives the right to counsel, considering the minor's age, education, presence of parents, and apparent maturity. Waiver may be withdrawn at any time. California Cal. Welf. & If a minor appears in a The statute provides Inst. Code delinquency proceeding for waiver by a minor sec. 634; Cal. without an attorney, if the waiver is R. Juv. Proc. the court must appoint intelligent. The rules 1412(g),(h). one. If the court further state that determines that the waiver must be knowing parent or guardian has and intelligent. the ability to pay for counsel, the court appoints counsel at the expense of the parent or guardian. Florida Fla. Stat. A juvenile's right to A juvenile may waive Ann. sec. counsel applies at all counsel if it is done 39.041; Fla. stages of delinquency freely, knowingly, and R. Juv. Proc. proceedings. intelligently. If 8.165. waiver is accepted at any stage of the proceedings, the offer of counsel must be renewed by the court at each subsequent stage of the proceedings at which the juvenile appears without counsel. Idaho Idaho Code As early as possible in A juvenile may waive sec. 16- the proceedings, and in counsel if he or she 1809A; Idaho any event before the does so intelligently, Juvenile Rules hearing of the petition and the court 3, 4. on the merits, a determines that the juvenile must be best interest of the advised of the right to juvenile does not counsel. If the require the appointment juvenile or his parents of counsel. are financially unable to afford counsel, the court shall appoint one. Kansas Kan. Stat. A juvenile is entitled The statute does not Ann. sec. 38- to the assistance of an specifically address 1606. attorney at every stage waiver. However, in of delinquency April of 1994, the proceedings. If the Attorney General of juvenile and his or her Kansas issued a parents do not retain nonbinding legal an attorney, the court opinion which concluded must appoint an that a juvenile may attorney to represent waive his or her right the juvenile. The court to counsel provided may assess the expense that it is a knowing of the appointed and intelligent waiver attorney against the on the basis of the juvenile or his/her totality of parent as part of the circumstances. expenses of the case. Louisiana La. Rev. Stat. A juvenile has a right A juvenile may waive Ann Ch. C. to an attorney at every counsel at any point in art. 809, stage of delinquency the proceedings if the 810(A). proceedings. court determines that: (1) the child has consulted with an attorney or other adult interested in the child's welfare; (2) both the child and the adult have been informed of the juvenile's rights by the court, and the possible consequences of waiving the right to an attorney; and (3) the waiver is voluntary. Maryland Md. Code Ann. A minor is entitled to The right to an Cts. & Jud. representation at every attorney may be waived Proc. sec. stage of a juvenile by the child or his or 3-821; Md. R. proceeding. Unless her parent if the court Juv. Proc. counsel is knowingly finds that the person 906. and intelligently waiving the right waived, or otherwise understands: (1) the provided, an indigent nature of the child whose parents are allegations and either indigent or proceedings, and the unwilling to employ range of allowable counsel, shall be dispositions; (2) that entitled to be counsel may be of represented by the assistance in Office of the Public determining and Defender. presenting any defenses to the allegations, or other mitigating circumstances; (3) that the right to counsel includes the right to the prompt assignment of an attorney, without charge to the party if he is financially unable to obtain private counsel; (4) that even if the party intends not to contest the charge or proceeding, counsel may be of substantial assistance in developing and presenting material that could affect the disposition; and (5) the other procedural rights available. Missouri Mo. Ann. Stat. A child has a right to A child may waive sec. 211.211; be represented by counsel only with Mo. Juv. Ct. counsel in all approval of the court. Rule 116. delinquency Waiver of counsel may proceedings. A lawyer be withdrawn at any shall be appointed stage of the prior to the filing of proceeding. a petition if the child is indigent and has made a request for an attorney. After the petition has been filed, the court shall appoint an attorney if it feels that the attorney is necessary for a full and fair hearing. New Mexico N.M. Stat. The child shall be The statute does not Ann. sec. 32A- represented by counsel expressly address 2-14; New at all stages of the waiver of counsel by a Mexico Child proceedings on a juvenile. However, the Ct. Rule 10- delinquency petition. statute requires that 205. Unless counsel has if counsel is not entered an appearance retained, counsel shall on behalf of the be appointed for the juvenile within 5 days child. In addition, the after petition is filed statute indicates that or at the conclusion of the child shall be the detention hearing, represented at all the public defender stages of the will be appointed to proceeding. Therefore, represent the juvenile. the juvenile does not have the option of waiving counsel. Nebraska Neb. Rev. When a juvenile is There are no specific Stat. sec. 43- brought before a statutory provisions 272. juvenile court, he or for the waiver of she must be informed of counsel by or for a his or her right to juvenile. However, retain an attorney. If Nebraska officials told the juvenile desires to us that juveniles can have counsel, and waive the right to neither the juvenile counsel. nor his or her parent or guardian can afford to hire an attorney, then the court shall appoint an attorney to represent the juvenile in all proceedings. New York N.Y. Jud. Law Juveniles who are A juvenile can waive secs. 241, subjects of delinquency counsel only after a 249, 249-a. proceedings should be law guardian has been represented by counsel appointed, and the of their own choosing court holds a hearing. or by a law guardian The law guardian must (required by statute to appear and participate be an attorney). The at the hearing and the court shall appoint a court must find by law guardian for a clear and convincing juvenile who is the evidence that (1) the subject of delinquency minor understands the proceeding if nature of the charges, independent legal the possible representation is not dispositional available. alternatives, and defenses, (2) the minor possesses the maturity, knowledge, and intelligence necessary to conduct his own defense, and (3) the waiver is in the best interest of the minor. However, in a proceeding to extend or continue placement of a juvenile delinquent, the court shall not permit waiver of counsel or of a law guardian. Pennsylvania 42 Pa. Cons. A juvenile is entitled A juvenile may not Stat. Ann. to an attorney at all waive his or her right sec. 6337. stages of delinquency to counsel. However, proceedings. An the juvenile's parent, attorney will be guardian or custodian appointed if the court may waive the right if determines that the their interests do not juvenile is unable to conflict with that of hire counsel. If the the juvenile. juvenile appears without a lawyer, the court must ensure that the juvenile knows of his or her right to have an attorney. South S.C. Code Ann. A juvenile has the If the juvenile wants Carolina secs. 20-7- right to be represented to waive his or her 600; 20-7- by an attorney in any right to counsel, he or 740. delinquency she must explicitly do proceedings. A juvenile so. However, the must be informed of juvenile must consult that right when he or with an attorney at she is brought to least once prior to a court. If the juvenile detention hearing. or his/or her parent or guardian are unable to employ counsel, counsel will be appointed. Texas Tex Fam. Code A juvenile has a right A juvenile cannot waive Ann. sec. to be represented by an the right to counsel in 51.09, 51.10. attorney at every stage the following of a delinquency proceedings: a hearing proceeding. If the to consider transfer to juvenile's parent or criminal court; an guardian is financially adjudication or unable to pay for an disposition hearing; a attorney, the court hearing prior to will appoint one. In commitment to the Texas addition, the court may Youth Commission; or a appoint an attorney in hearing required any case in which it because it is a deems representation proceeding regarding a necessary to protect child with mental the interests of the illness, retardation, juvenile. disease or defect. In all other circumstances, a juvenile may waive counsel if the waiver is made by the juvenile and his or her attorney; they are both informed of and understand the right and the possible consequences of waiving it; the waiver is voluntary; and the waiver is made in writing or in court proceedings that are recorded. Utah Utah Code Ann. There is a statutory The statute does not sec. 78-3a- right for a child to address waiver of 35. have an attorney at counsel. However, Utah every stage of the officials told us that delinquency juveniles can waive the proceedings. Upon a right to counsel. request for an attorney, or upon the court's own motion, in addition to a finding that the juvenile is indigent, the court shall appoint counsel. -------------------------------------------------------------------------------- (See figure in printed edition.)Appendix V SURVEY OF PROSECUTORS ========================================================== Appendix IV (See figure in printed edition.) (See figure in printed edition.) (See figure in printed edition.) (See figure in printed edition.) (See figure in printed edition.) (See figure in printed edition.) (See figure in printed edition.) (See figure in printed edition.) (See figure in printed edition.) (See figure in printed edition.) (See figure in printed edition.) (See figure in printed edition.) MAJOR CONTRIBUTORS TO THIS REPORT ========================================================== Appendix VI GENERAL GOVERNMENT DIVISION, WASHINGTON, D.C. James M. Blume, Assistant Director, Administration of Justice Issues Barry Jay Seltser, Assistant Director Thomas L. Davies, Senior Evaluator William J. Sabol, Senior Social Science Analyst Barbara A. Stolz, Senior Social Science Analyst Joanne M. Parker, Senior Social Science Analyst Bonita J. Vines, Computer Specialist Brenda I. Rabinowitz, Evaluator Maria D. Strudwick, Evaluator Pamela V. Williams, Communications Analyst Michelle D. Wiggins, Information Processing Assistant OFFICE OF THE GENERAL COUNSEL, WASHINGTON, D.C. Jan B. Montgomery, Assistant General Counsel Rachel A. Ramsey, Legal Intern Matthew L. Robey, Legal Intern ATLANTA REGIONAL OFFICE Frankie L. Fulton, Regional Management Representative Cynthia C. Teddleton, Evaluator-in-Charge Veronica O. Mayhand, Senior Evaluator