wally! " I!!! Will/Ill! ”Nil/1'! ”W! l 93 10587 6324 7'AN ITATE UNIVERSITY LIBRARY THESIS This is to certify that the thesis entitled Do Juvenile Diversion Programs Widen the Net of Social Control? An Application of System—Level Evaluation Strategies presented by John A. Saul has been accepted towards fulfillment of the requirements for Master of Arts__degree in _Es¥r_bnlogy_ Major professor Date é" l/-,3/ 0-7639 OVERDUE FINES: 25¢ Per day per item ‘ \ .1 A .147; “F“.‘a. .‘ W: Place in book return to muove charge from circulatton records ‘ ‘ \ fluent” L w“ n” 4 53mg 8513 NM '2 V 57%! 3“; . , 93ml ,l . 5’00 [41% ' 3%8‘“ i . W£§fimmitgca 1952i 3 “Esq DO JUVENILE DIVERSION PROGRAMS WIDEN THE NET OF SOCIAL CONTROL? AN APPLICATION OF SYSTEM-LEVEL EVALUATION STRATEGIES By John A. Saul A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Psychology 1981 ABSTRACT DO JUVENILE DIVERSION PROGRAMS WIDEN THE NET OF SOCIAL CONTROL? AN APPLICATION OF SYSTEM-LEVEL EVALUATION STRATEGIES By John A. Saul Established as alternatives for youths in actual jeopardy of suffering the allegedly negative consequences of formal juvenile court handling, juvenile diversion pro- grams have instead been criticized for serving mostly those who would otherwise have been released. If this is true, diversion programs merely serve to "widen the nets" of the justice system. This research addresses the lack of con- clusive data on the issue by applying a multi-method system— level evaluation strategy to one ongoing diversion program. Four sources of data were analyzed, including control group dispositions, a random sample of archival court cases (using discriminant analysis), intake referee questionnaires, and system processing rates. The results show that most diver— sion cases would have otherwise been handled informally by the court, while only one-quarter to one-third were probably truly diverted from formal court involvement. Diversion's inherent conflicts are discussed as explanations for its inadvertent "net-widening" effects. ACKNOWLEDGMENTS The rocky road leading to the completion of this thesis had its ”falling rocks, bumps, and dangerous curves," and I am grateful to the many people who helped me along. My foremost thanks go to Dr. William Davidson for giving me the Opportunity to do this research, and for histfgresight in providing for a "systems—level" evaluation component in the project. His support, friendship, and tolerance for "mis- placed creativity” made this work possible. My other committee members, Dr. Robin Redner and Dr. Ralph Levine, also contributed important guidance and in- sights. I also thank my fine friends and fellow systems- group members, Diane Wresinski and Deb Bybee, who somehow let me get my hands on the choicest material, after patiently showing me the ropes. Randy Koch saved me hours of frus- tration by being a walking, talking juvenile justice liter- ature resource center. I also thank my friends in the EIEIO group for excusing my leap from faith into the harsh light of degree-oriented realities. Tremendous appreciation goes to the many data collectors and coders who worked on the research over the years. ii Sarah Lindsay endured my absurd procrastination and put together the final manuscript at break-neck speed, for which she deserves my utmost admiration. I also want to express my appreciation to my parents and family for their everlasting support through these many years of effort. Finally, I must mention my extreme gratefulness for the hot coffee and 3 a.m. breakfasts at Theio's Restaurant, without which I may have slept instead of worked. iii TABLE OF CONTENTS Page LIST OF TABLES ........................................ vi LIST OF FIGURES ..................... . ............... .. vii Chapter I. INTRODUCTION .................................... 1 Related Research.0.............000.0...0.....0.. 4 History and Rationales of Diversion........... 4 Critical Issues in Juvenile Diversion......... 7 Diversion as an Extension of Social Control... 9 Studies Reporting Data Related to Net- Widening................................... 13 Possible Explanations for Net-Widening........ 24 Objectives...................................... 28 II. METHOD.0...00... ....... 0.....000... ..... . ....... 32 setting..0..............0............0.0.. ..... . 32 Evaluation Design.... ...... ...... ...... .... ..... 33 Method 1——Analysis of Control Group DiSpOSitions. 0 . . . . . . . . . . . . . 0 . . . ............. . 33 Method 2——Discriminant Analysis of a Random Sample of Court Cases................. 34 Method 3——Analysis of Questionnaire Data Provided by Referees.................... 39 Method 4——Analysis of System Processing Rates... 40 Methods of Checking for Convergence of Results.. 41 III. RESULTS... .......... . ......................... .. 43 Method 1.... ...... .... ...... .. ........... ....... 45 Method 2........................................ 46 Method 3........................................ 51 General Convergence of Methods 1, 2, and 3...... 54 Method 4........................................ 56 Analysis of Status Offender Policy Change....... 63 Convergence Testing of Methods on Individual Cases........................................ 64 Method 1 vs. Method 2......................... 65 Method 1 vs. Method 3......................... 65 Method 2 vs. Method 3......................... 66 Method 1 vs. Method 2 vs. Method 3............ 67 iv Chapter IV. DISCUSSION....00.....0...00...000..0 ........... . Method 1............................ ......... ... Method 2........................................ Method 3........................................ Method 4........................................ Status Offender Policy Change................... Time Trends in ADP ”Formal" Cases............... Convergence of Results Across Methods........... Conclusions About Net—Widening and the ADP...... Comparison of Results With Other Studies........ Explanations for Net-Widening................... Prevention of Net-Widening...................... The Future of Diversion......................... Future Research................................. REFERENCE NOTES......00......0.......0....0........... REFERENCES ..... .0... ....... .....0.......0.....0...0... APPENDIX ....... .....0000 ........ .......00....0....0... LIST OF TABLES Table Page 1 Coding of Decision Sample Variables and ReliabilitieSeoooooooeeoooeooooooooeoeoeeeoeeo 36-38 2 Dispositions of ADP Control Group Cases Years I-V ...... .......0....0.0..... ..... 0.0... 46 3 Discriminant Analysis Summary Table............. 47—48 4 Discriminant Analysis Between—Group Statistics.. 49 5 Major Variables Entering Discriminant Functions.............0.............. ........ . 50 6 Discriminant Analysis Classification Results.... 51 7 Discriminant Analysis ADP Years I-VII... ...... .. 52 8 Alternative Dispositions for ADP Cases as Recommended by Intake Referees................ 53 9 General Convergence of Methods 1, 2, and 3...... 55 10 Analysis of Status Offender Policy Change....... 64 11 Classifications of Individual Cases by MethOdSIand 20.....0........0....... ....... . 66 12 Classifications of Individual Cases by MethOdS1and3......0...............0O. ..... . 67 13 Classifications of Individual Cases by MethOdszand 3.......0....................... 68 14 Summary Comparison of Methods 1, 2, and 3 ....... 69 vi LIST OF FIGURES Figure Page 1 Percent of Cases Classified as "Formal" over Time ....... .....................0. ....... 57 2 Percent of Cases Handled Formally by Year (Court Population Figures).................... 58 3 Percent of Cases Handled Formally by Year (All Cases in Sample)......................... 61 4 Percent of Cases Handled Formally by Year (Non-Status Cases Only)....................... 62 vii Chapter 1 Researchers concerned with social problems have recently become more aware that unintended negative consequences can result from social "reform" programs (Caplin & Nelson, 1973; Gaylin, Glasser, Marcus & Rothman, 1978; Klein, 1979; Rappa- port, 1981). Whether through faulty design or faulty imple- mentation, social programs may actually serve to exacerbate the problems they were supposed to solve. Concerns of this sort have been increasingly expressed about the impact of diversion programs on the juvenile justice system. Diversion programs were originally intended as positive alternatives to the juvenile justice system, a social institution whose own reformist ideals were said to have turned sour (U.S. President's Commission, 1967; In re Gault, 1967; James, 1969; Gold & Williams, 1969). However, the extent to which diver- sion programs actually achieve their goals must be evaluated before they can be considered as effective or true alterna- tives to traditional handling of youth in legal jeopardy. Hence, the purpose of the proposed research is to examine the major unintended consequence of juvenile diversion: the possibility that diversion programs, rather than diverting Youth away from the justice system, actually "widen the net" 0f justice system control over more youth. The potential "net-widening" effect of diversion pro- grams has attracted considerable attention in the juvenile justice literature. The issue is also referred to as "over- reach" (Klein, Teilmann, Styles, Lincoln & Labin, 1976) and the "acceleration of social control" (Blomberg, 1977). Each term refers to the process by which diversion programs result in an increase in the number of young people under the super- vision or control of justice—system agencies. Net-widening occurs when a youth who ordinarily would have been warned and released" by the police or the juvenile court is instead referred to a "diversion" program. As explained by Ruther- ford and McDermott (1976), "The result is that more juveniles come into contact with formal agencies of authority." This of course represents the exact opposite of diversion's original intent, which was to provide alternative services to youths who typically would have come under further court supervision. If the net—widening phenomenon is widespread, then diver- sion programs are not serving their intended type of client. This means that diversion has not actually been implemented (Klein, 1979), and thus cannot adequately be tested in its true form. Evaluation of diversion program effects on in— dividual-level outcomes such as recidivism becomes relatively meaningless because court-processed comparison cases would not be similar types of youths (Gibbons & Blake, 1976). In addition, diversion programs would lose their potential for cost-savings, because they would not be lessening the number of youths receiving police and court services (Rutherford & McDermott, 1976). Further, official "delinquency" could actually increase because greater numbers of youths would be monitored by justice-related agencies. Youths referred to diversion have been said to be more susceptible to the negative consequences of justice-system labeling, which may also lead to greater "delinquency" (Klein, Teilmann, Lincoln & Labin, 1979). Finally, diversion programming for those who would have been outright released may represent an abuse of individual rights due to the greater intrusiveness of the treatment (Howlett, 1973). Thus, the net-widening effects of diversion have poten- tially serious consequences. Despite the importance of this issue, there are few empirical studies evaluating diversion programs for evidence of net-widening. In contrast, there are many articles which deal with the issue on a conceptual basis only, presenting no original data (Bullington, Sprowls, Katkin & Phillips, 1978; Nejelski, 1976; Klapmuts, 1974; Blomberg, 1980). Pabon (1978) has characterized the avail- able research on net-widening as being "impressionistic." Others have also noted the paucity and inadequacy of the research to date (Blomberg, 1977; Gibbons & Blake, 1976; Rutherford & McDermott, 1976). This problem can at least partially be attributed to an over—emphasis on individual- level outcomes, occurring at the expense of system-level evaluation (Seidman, Note 1; Wresinski, Saul, Bybee, Davidson & Koch, Note 2). The present study assesses one ongoing juvenile diver— sion program for evidence of net-widening by developing and applying an appropriate system-level evaluation strategy. First, the historical background and theoretical rationale for juvenile diversion will be reviewed, followed by a more detailed discussion of the implications of the net-widening issue. Next, studies presenting data related to net-widening will be reviewed, with critiques of the evaluation strategies used to date. A summary of explanations for why and how net-widening occurs will then be presented. A comprehensive evaluation strategy will then be developed and applied to the analysis of the diversion program under study. Related Research The significance of the net-widening phenomenon can best be understood in the context of the original rationales for developing diversion programs. These rationales serve to underscore the ironic implications of diversion programs which unintentionally extend the range of justice system control. History and Rationales of Diversion Criticism of the juvenile justice system has occurred throughout its history, despite its own basis as an institu- tional alternative to save young people from treatment by the adult judicial system (Fox, 1970; Platt, 1974). One major criticism was that the juvenile court had become a large, overloaded and inefficient bureaucracy. Probation caseloads were said to be too high to permit effective treatment or even supervision. Costs of institutional treat- ment were becoming exorbitant. This situation encouraged the search for alternatives to court handling of youth. A second criticism focused on the paradoxical tendency for court processing to lead to more, rather than less, de- linquent behavior. A study by Gold and Williams (1969) showed that juvenile justice system handling of youth could actually result in more delinquency. Recidivism rates have been found to be as high as 50% to 85% (Scarpitti & Stephenson, 1968; Wolfgang, Figlio & Sellin, 1972; Jesness, 1975; Griffin & Griffin, 1978). These data have been interpreted as support ! ffor the labeling theory of delinquency, which states that V— _.._.-‘- $1.1... Xbourt involvement stigmatizes young people with the "delin- quent" label. This label may influence their self-concept, leading to an ongoing pattern of delinquent behavior, or the label may work to make law enforcement agents more vigilant in checking up on certain youths (Katkin, Hyman & Kramer, 1976; Klein et al., 1979). Combined with evidence that a sizeable number of young people are arrested at some time (Dinitz, 1969; Wolfgang et al., 1972), it became apparent that the traditional system was not adequately controlling or reducing delinquency. A third criticism dealt with abuses of the constitutional rights of juveniles coming into court contact. Particular concern was expressed for youths harbored in institutional facilities in which treatment was often inhumane (James, 1969; Richette, 1969). The parens patriae doctrine of the court allowed for "parentlike" treatment without the due process rights granted to most citizens. Three U.S. Supreme Court cases granted more rights to youth in legal jeopardy and re- jected the procedural informality which had characterized juvenile court practices (Kent vs. U.S., 1966; ln_£§ Gault, 1967; Ig_£§ Winship, 1970). All three criticisms provided impetus to the idea that juveniles would be better off if they could avoid traditional court processing. Backed by these rationales, the idea of diverting youth away from the justice system gained promin- ence. In 1967, the President's Commission on Law Enforcement and Administration of Justice recommended the following: (The) formal sanctioning system and pronounce- ment of delinquency should be used only as a last resort. In place of the formal system, dispositional alternatives to adjudication must be developed for dealing with juveniles... The primary objective noted by the U.S. Youth Development Delinquency Prevention Administration was to "divert youth away from the juvenile justice system" (Gemignani, 1973). In 1973, the National Advisory Commission on Criminal Justice Standards and Goals also embraced diversion: Every police agency, where permitted by law, immediately should divert from the criminal and juvenile justice system any individual who comes to the attention of the police, and for whom the purpose of the criminal or juvenile process would be inappropriate, or in whose case other resources would be more effective (p. 80). This recommendation was followed in 1974 by Congressional action stating that it was the ”declared policy of Congress ...to divert juveniles from the traditional juvenile justice system and to provide critically needed alternatives..."(42 U.S.C. S.5602 (b)). This high-level enthusiasm for diversion led to a pro— liferation of new programs throughout the country. The LEAA funded the formation of Youth Service Bureaus (YSB's) around the nation, along with other diversion programs. It was not long, however, before the need for evaluation of the impact of diversion programs was sensed (Palmer & Lewis, 1980). Several writers called for assessment of juvenile diversion, and noted the lack of research available on both individual and system effects (National Advisory Committee, 1973; Cressey & McDermott, 1973; Vorenberg & Vorenberg, 1973). As diver- sion programs became more closely scrutinized, criticisms began to emerge. Critical Issues in Juvenile Diversion Issues concerning juvenile diversion programs reflect many of the same problems discussed in relation to the tra- ditional juvenile justice system. As noted by Davidson, Koch and Snellman (Note 3), the diversion issue has now come "full circle." The same concerns about protection of individual rights, treatment effectiveness and avoidance of stigmatiza- tion raised about the juvenile court have recently been applied to diversion programs. The question of due process and youth rights arises -~vr ...—h -m—‘~-—.—» because of the informal, discretionary practices often in- volved in referring youth to diversion programs. Rights may be abused: (a) if participation in diversion programs is not voluntary, (b) if referral is disciminatory to certain types of youths or occurs without a legal finding of guilt, or (c) if a youth can be reinvolved with the court for in- adequate progress in the program ("double jeopardy"). Indi- vidual freedoms may be abused also by requiring youths to participate in programs that are potentially more intrusive than their behavior warrants (Howlett, 1973; Davidson et al., Note 3). The_second major criticism is that virtually nothing is” known about diversion program effectiveness in reducing de-I linquency. The lack of experimental programs providing con- trol group comparisons has hampered the evaluation of juvenile diversion. Without adequate evaluation, diversion is suscep- tible to demands for increased accountability. Justification of continued funding cannot be made on conceptual grounds alone. Questions of program effectiveness in reducing re- cidivism will remain until better evaluation results are produced (Binder, 1977; Rutherford & McDermott, 1976; Gibbons & Blake, 1976). Third, critics of diversion point out the problems associated with diverting juveniles to another service pro- “a- ...-..r gram, rather than diverting them away from all service ‘a systems (outright release). Diversion with services, which is the typical kind of_diversion, carries the potential for encapsulating the juvenile in another system of services and control. This Walternative encapsulation" may be just as deleterious and stigmatizing as court processing (Bullington etwal., 1978; Pabon, 1978; Klein et al., 1976). According to some researchers, ”true” diversion must avoid the mere transference of juveniles from one program to another. In effect, diversion to any program is contradictory to its goals; the true meaning of the concept only allows for diver- sion from the system (Bullington et al., 1978; Cressey & McDermott, 1974). Diversion as an Extension of Social Control The most fundamental issue in diversion is whether diver- sion programs accomplish their goal of diverting juveniles who would otherwise receive formal court processing. Nejelski (1973) has defined diversion as "the channeling of cases to non-court institutions or systems in instances where the cases would ordinarily have been processed by the juvenile court." If cases which would not have been formally processed are involved in diversion programs, two major consequences follow. One, the intended targets of diversion remain in the court system and suffer the allegedly negative results of court processing. Two, juveniles ordinarily released from the justice system become involved in some sort of ser— vice program, thus "widening the net" of social control. 10 It is important to note here that in the practice of "true” diversion the second consequence would not occur. That is, if diversion were operationalized as diversion with- out services, then even unintended targets of diversion would not be drawn into the supervision of any formal agency of authority. However, the point is made moot by the extreme rarity of ”true” diversion (Cressey & McDermott, 1974). Diversion programs (with services) thus have the potential for net—widening if they serve the wrong set of individuals. Several implications follow from the two major conse- quences of "diverting" the wrong set of juveniles. First, because the targeted set of juveniles are not diverted, there is little impact on the court. The number of youths receiving formal processing would not decrease, probation caseloads would remain the same, and no cost—savings would be achieved (Rutherford & McDermott, 1976). Life at the court would continue with "business-as-usual." The inclusion of juveniles not intended for diversion has considerably greater and more negative effects. At the system-level, diversion programs create new social agencies or enhance the resources and range of existing agencies. For the young person referred to this new service system, his or her participation in the diversion program represents a greater degree of system-penetration than would otherwise have been experienced. The diversion program creates rela- tively more social control, more supervision, and more chances of being reinvolved with the justice system. As Klein 11 (1979, p. 12) has explained, Diversion means to turn away from, and one cannot turn away from something toward which he was not heading in any case. Diversion programs must tar- get in on youngsters who otherwise would enter, or penetrate further into, the justice system. To the extent that "releasable" juveniles have been targeted for referral to programs, the idea of diversion has not been implemented in practice. Klein et a1. (1979) have explored the relationship between labeling theory and diversion. They found that young people often referred to diversion programs are also those who appear to be more susceptible to the effects of labeling. These are juveniles who, relatively speaking, ordinarily tend not to become further involved in the justice system, such as whites, females, higher status youth, and those with no prior involvement. They conclude that the labeling conse- quences of diversion program involvement for this type of client can lead to increased delinquency. In addition to extending social control over youths with minimal justice system contact, diversion programs may also widen the range of control over youth with no justice system contact. Blomberg (1977) studied a suburban California diversion program which focused on "family intervention" for delinquency prevention. The family focus brought "services" and supervision to siblings with no prior behavior problems. Other diversion programs have accepted referrals from 12 non-justice system sources such as schools, social service agencies, or families (Palmer & Lewis, 1980). These cases would not have ordinarily had any justice system contact, yet become involved in a "diversion” program. Net-widening also may increase the "extensiveness of control" over youngsters' lives (Klein et al., 1976). Any service program may impose some restrictions and controls on behavior. This may include behavioral contracting, required counseling sessions, curfews, and other restrictions. Thus control over clients can be achieved by mutual 'contract,‘ by suggestion and expectation, and by threat of reinvolvement with the police, probation, or court (Klein et al., 1976, p. 113). This is clearly a more intrusive level of treatment than need be experienced by a juvenile who, in the absence of a diver- sion program, would have been set unconditionally free. Net-widening also makes evaluation of diversion program effects on recidivism difficult. Even if some of the in- tended set of individuals are diverted, their potential im- provement will be disguised by the presence of unintended cases. Juveniles with few or no prior offenses cannot im- prove much; it is therefore very difficult to show results indicating reduced recidivism (Palmer & Lewis, 1980). The overall rate of delinquent behavior of diversion program cases may be lower due to the inclusion of "low-risk” youth, but there can be little chance of showing positive change over time. Further, since actual diversion has not been 13 implemented in such circumstances, neither can it be said to have been evaluated. Studies Reporting Data Related to Net-widening As stated earlier, the research related to the net- widening issue is "impressionistic" (Pabon, 1978). Few studies report original data, and those that do seldom pre— sent enough detail for critical evaluation of their method- ology. Some reports compare diversion program client char- acteristics with court-involved youth. Others simply report client characteristics without any comparison groups. Another method has been to interview or survey justice system officials responsible for diversion referrals. A few other articles analyze system processing rates after the establish- ment of diversion programs, to see if the number of formally- handled youth decreases. Each type of evaluation approach will be presented separately. Diversion program client characteristics. Klein (1979) has reviewed evidence from studies presenting demographic information on diversion project clients. He found few exceptions to the usual practice of taking "soft" offenders into diversion programs. The "soft” offender is contrasted with the profile of the "system-insertable" juvenile, who sets the standard by which diversion clients should be com— pared. ”System-insertable" youth, according to Klein, have the following characteristics: (a) they should be between 15-17 years of age, (b) proportionately more of them should be male (5 to 1), (c) they should have a prior record of 14 offenses with the current offense of at least medium serious- ness, and (d) there should be few status offenders in the group profile. Programs made up of non-system-insertable clients are engaged in net-widening. Two projects reviewed by Klein were found to be at least close to the standard. One program's clients had a mean age of about 16, with the modal age being 17. Fifty percent of the cases had prior offenses, with the current offenses usually involving formal arrest, and one-fourth of the clients were female. The percentage of females was a bit high when compared with Klein's standard (Klein, 1975). A second pro- ject with appropriate clientele was that reported by Berger, Lipsey, Dennison and Lange (1977). The project had only 20% females, 17% status offenders, and two-thirds of the cases were judged to be petitionable to the court. However, the average age was only 15, and only 45% of the cases had a prior record. However, most diversion programs reviewed by Klein con- sisted of generally inappropriate clienteles. A program reported by McAleenan (1976) had clients who average 13 years of age. Thirty percent were female, 77% had no prior record, and only 43% had been referred from justice system sources. In another program (Humphreys & Carter, 1976), the average age was 14, 35% were female, 43% had no priors, and only 41% were referred by the justice system. Berg and Shichor (Note 4) report a diversion program with 41% female clients, 73% of the clients aged 15 or under, and almost 92% with no prior 15 record. Bohnstedt et a1. (1975) studied other diversion programs averaging 39% female, with status offenders making up 51% of the clients. Klein (1979) concludes that programs of this sort are not meeting the standard diversion require- ments of handling ”system-insertable" clients. The diversion clients described in these reports clearly do not resemble a typical profile of the serious juvenile offender. Yet, mere reporting of the client characteristics does not give sufficient proof that net-widening is occurring. This is especially true if no information is given on type of offense, number of priors, or source of referral. Also, the absence of comparison cases makes interpretation difficult. The type of youth who may be "system-insertable” in one police or court jurisdiction may usually be released in another. The "standard” type of juvenile receiving further court processing may be a national composite bearing no particular resemblance to "system—insertable" youths in par- ticular counties. Interviews or surveys of justice system officials. Another method of assessing the tendency for diversion pro- grams to widen the net of control is to simply ask the rele- vant decision-makers. Justice system officials who refer juveniles to diversion programs have been questioned about their referral criteria in at least two studies (Klein et al., 1976; Palmer & Lewis, 1980). In the Klein et a1. study of 35 police department diversion programs, juvenile officers were interviewed about which types of youngsters were 16 preferred for referral. The officers preferred youngsters who had low-seriousness offenses, fewer prior offenses, were young in age, were unlikely to be rearrested, and who behaved cooperatively with the police. The authors conclude the following: This picture fits the profile of the youngster who heretofore would have been released out- right, for whom referral represents increased intervention rather than diversion £32m system intervention (Klein et al., 1976, p. 108). Police officers in this study were not putting the goals of diversion into practice. Palmer and Lewis (1980) evaluated California diversion programs, selecting 15 representative projects out of 74 under consideration. They gave justice agency officials a questionnaire containing various types of hypothetical case summaries. The officials were asked to indicate which cases they would choose to divert, warn and release, or petition for court processing. Those most likely to be diverted were characterized by low-seriousness offenses and a nonexistent or minor prior record. Those youths with a prior record typically had been only warned and released prior to the instant offense. Youths least likely to be picked were those with prior felony convictions or prior experience in a diver— sion program. These findings also indicate that diversion cases typically are dissimilar to those who would receive further court handling. 17 While interview or survey methods are a direct way of assessing decision-makers' beliefs about their diversion referral criteria, they do not necessarily assess their referral behavior. Officials may tend to respond with the answers they think are expected, rather than their actual attitudes. On the other hand, the data presented above does not show such a tendency. A second potential problem for survey methods is response rate. Palmer and Lewis (1980) report a 60% response rate to their case summary survey, which they conclude yields "moderately representative" results. This method of net-widening evaluation should be combined with analysis of actual dispositional comparisons to assure meaningful interpretation. System processing rates. Another type of study providing empirical evidence on the net-widening issue is the analysis of system processing rates. Programs which actually divert youths in imminent jeopardy of formal court processing should have the effect of reducing the overall rate of cases pene— trating the court system. Nejelski (1976) has offered this method as the primary strategy for the evaluation of diver- sion programs: A good test of the effects of diversion, therefore, is whether the number of juvenile court adjudi- cations is reduced as a result of new diversion programs (p. 397). This method has also been endorsed by Blomberg (1980) as an essential evaluation strategy for the assessment of net— widening. 18 Sarri and Bradley (1980) studied a South Australian equivalent of American diversion programs which operated through "juvenile aid panels." They found that the use of the panels was not associated with reduced referral to the juvenile court except briefly in 1973-74 and 1976-77. In the latter instance, the total number of youths referred by the police also declined (p. 51). Their data did not reveal whether youths referred to aid panels were in actual jeopardy of further court processing. They interpret the results as evidence of net-widening because of the large increases in numbers of juveniles receiving panel or court services after the initiation of the aid panels. Blomberg (1977) evaluated a "family intervention” diver- sion program in California for net-widening effects. He developed expected system processing rates from historical demographic data based on numbers of youths in court per total youth population. The percentage of "youth under con- trol"——those receiving formal, informal, or diversion control ——increased by 2.4% after the diversion program was started. Because of the family focus of the diversion program, more- over, additional youths were brought under supervision of some sort due to their status as siblings of offenders. Assuming one sibling per family, the percentage increase rose from 2.4% to 17% of youths under control. Assuming two 19 siblings per family, the increase in youths under supervision due to the diversion program's existence was 32%. Analysis of system processing rates has obvious advan- tages for diversion program evaluation, but also has some pitfalls. Official data are often not available, or are often unverifiable or unreliable. As Blomberg (1980) has noted, "the usefulness of this approach is dependent upon consistent data collection and reporting over time." Courts and police may not keep reliable statistics, and youth popu- lation figures may not be available at the county level. In addition, caution should be exercised when interpreting changes in system rates. These changes may be confounded by policy decisions and organizational adjustments unrelated to the diversion program's effects. Dispositions of comparison cases. The final method used to determine whether diversion results in net—widening in- volves the analysis of matched comparison or control cases. If juveniles similar to diversion clients further penetrate the justice system, then the program is probably serving the intended set of clients. If comparable cases are merely warned and released, then the program is probably extending the range of social control over more youths. A West Coast police diversion program was examined by Lincoln in an unpublished report cited by Gibbons and Blake (1976). Cases matched with diversion clients were studied as a control group, and it was found that nearly half of these were warned and released. According to Gibbons and 20 Blake, this ”suggests that the diversion group was made up of a considerable number of petty offenders." The evaluation dealt with the pilot phase of the program, which "diverted" a total of 30 youths. The Alternative Routes project in Orange County, Cali- fornia, also used matched comparison cases in its evaluation. However, the dispositions of matched cases were not reported. Instead, it is noted that 60% of the juveniles would have gone on to court processing. This information appears to have come from police estimates (Gibbons & Blake, 1976). However, it is also noted that only 87 of the 142 cases came from police referrals. Thus, only 60% of that 87 were actually in jeopardy of further court involvement. The result is that only 37% of their cases represented "at-risk" youth. Others in the diversion program were victims of net- widening, as they were referred by parents, schools, and community agencies without having justice system contact. In another study of police diversion, 120 diversion pro- gram clients were matched with a "comparable cohort” of 82 cases. While all 82 had petitions filed with the court, 20 were released and 30 received only informal probation (Pabon, 1978). Hence, about 61% of the diversion program clients were probably not in danger of deeper justice system pene- tration. One of the largest diversion studies, the California Youth Authority evaluation project, also used the matched- comparison method. As reported in Palmer and Lewis (1980), 21 matched comparison cases were developed for nine of the 15 programs under study. Various problems prevented the inclu— sion of the other six diversion programs. Further examination of the data reveals that 40% of the cases in the six programs dropped from the analysis came from non-justice system sources, compared with 28% in the nine sampled programs. The following results may therefore understate the entent of net-widening which occurred in the total sample of 15 diver- sion programs. Comparison cases were matched on age, sex, and ethnicity and had met diversion program referral eligibility criteria (although these criteria were unspecified). This matching procedure had "weaknesses," according to an earlier report (Bohnstedt, 1978), but these were not specified by the authors. Five dispositional alternatives were identified by the authors, and then classified according to whether or not they represented further justice system processing. Those youths referred to court intake by the police (31%), placed on informal probation by the court (7%), or petitioned for further court processing (13%) were considered as represen- tative of further justice system penetration. The authors conclude therefore that 51% of the diversion cases would have been further processed, and thus were actually "diverted." The remaining 49% include those comparison cases released by the police (14%), those dismissed at court intake (7%), and those referred from non-justice system sources (28%). These 49% were not diverted and their referral to diversion programs 22 "did involve an extension of society's services or controls" (Palmer & Lewis, 1980, p. 213). Some problems of interpretation arise from the study. The evaluation included both police and court diversion pro— grams, such that some overlap occurs in the dispositional categories. Cases diverted from court intake by the police (31%) were counted as proper examples of diversion from further justice system processing. However, many of these cases may have been subsequently dismissed at court intake, a category counted as improper for court—level diversion programs. If the goal of diversion centers on turning youth away from the juvenile 99333, then not all 31% of the suc- cessful police diversion cases should be counted as having been diverted from imminent danger of court processing. In addition, the 7% of youths placed on informal probation at the court do not represent further court processing to some researchers (Blomberg, 1977). It has been argued that diver— sion programs are more intrusive than informal probation, in which the court usually adjourns further action while waiting to see if the youth gets into more trouble. The number of youths in diversion programs who would have received further court processing is therefore subject to differential interpretation. Nonetheless, at least half of the diversion clients in the Palmer and Lewis study appear to be victims of more control and supervision than if they had not been referred to diversion. The interpretational problems associated with the above results are typical of 23 the matched-comparison case method of evaluating diversion programs. One problem is that comparison cases have to be drawn from a time period before the diversion program existed. Historical and/or policy changes in the interim may confound the results. Also, the matching process may be inexact. All cases cannot possibly have exact equivalents, and it is difficult to know which variables are important to match. Procedures which match cases on client demographic charac- teristics such as age, sex, and race may miss other important criteria such as offense seriousness, prior record, or type of previous adjudication. The comparisons may not be valid at all if the matching variables are not empirically—proven predictors of court processing. While all four of these evaluation methods may have their faults, the studies reviewed here appear to indicate converging evidence for the net-widening hypothesis. More research needs to be done using better evaluation methods. Yet, the tentative conclusion can be reached that most diversion programs studied to date have expanded the range of justice system control by "servicing" clients not actually in jeopardy of court processing. Gibbons and Blake (1976, p. 413) have cogently summarized the situation: Diversion, growing out of the sociologists' recommendations of 'radical nonintervention,‘ 'benign neglect,‘ or 'judicious nonintervention,‘ has become perverted in practice into a strategem that swells the population of acted-upon offenders. 24 Because of the serious implications this fact has for further diversion programming, it is necessary to next consider why net-widening occurs. Possible Explanations for Net-widening Although diversion programs have proliferated in recent years and have received high-level policy support and funding, there are inherent conflicts in the actual implementation of the diversion concept. The major conflict lies in the radi— cal implication that the traditional justice system has failed and may actually contribute to delinquency. Yet effec- tive implementation of diversion depends almost wholly on complete cooperation of the system so heavily criticized by diversion proponents. It is not far-fetched to assume that justice system agents might have their own ideas about how to best utilize the influx of resources and services that come with diversion program funding. One explanation offered for the net-widening effects of diversion is the ”penchant for control" on the part of jus- tice system agents (Klein, 1979). According to this view- point, the justice system not only seeks to maintain its usual level of control, but also to expand. Judges, no less than police, correctional personnel, and social agencies, develop strong preferences for control of client populations (Klein, 1979, p. 48). Police and other justice officials try to "keep the strings attached" when referring juveniles to diversion programs. 25 Other writers have also noted the pervasiveness of the social control functions of the justice system, and its tendency to extend such control (Blomberg, 1980; Pabon, 1978). According to Pabon (1978, p. 496), Current programs that have adopted new sponsorship and enlightened titles may merely be reincarnations of the traditional and costly institutions of control. Blomberg (1980) has cited historical evidence that social "reform" has traditionally served to "increase the network of social control and the population subject to control...". The operation of diversion programs can easily be coopted into a mechanism of extended social control by referring cases ordinarily released. This "penchant for control" need not be interpreted in fascist terms, however. It is partially motivated by a progressive philosophy of prevention, in which young people are seen as in need of treatment services. Several writers have commented on the transformation of diversion programs into "prevention" programs (Blomberg, 1980; Bullington et al., 1976; Klein, 1979; Sarri & Bradley, 1980). Because "preven— tion” programs depend on early identification and treatment, the extension of services to young, minor offenders is seen as a natural extension of professional "helping" roles. As Bullington et al. have explained, the parens patriae philo- sophy of the juvenile court need not be construed as incom- patible with diversion: 26 The extension of 'benign,‘ 'helping,’ community- based services to a larger population of young- sters is altogether compatible with the traditional parens patriae values of juvenile justice-namely that treatment for juveniles should be therapeutic and nonpunitive and that procedures should be in— formal and nonstigmatizing. Diversion programs fill the perceived need for services reaching the "at—risk” population of youngsters, who normally could not be "helped" due to lack of funding. Net-widening, with its capacity for increasing the number of juveniles likely to be detected by the court system for delinquent behavior, may be caused by a desire to provide services to youth and prevent increased delinquency. Besides the "penchant for control" and "penchant for prevention" explanations of net-widening, a third rationale might be termed the "penchant for self-perpetuation" of jus- tice system agencies. This organizational—maintenance per- spective derives from the common evidence that large bureau— cracies tend to perpetuate themselves. Diversion‘s goal of taking away clients from the court conflicts directly with the court's goal of maintaining caseloads, programs, and staff as continual justification for institutional funding. Rappaport, Seidman, and Davidson (1979) have pointed out that strict adherence by justice agencies to the diversion model would jeopardize their own existence: If (diversion) procedures were followed, the juvenile officers would find themselves with 27 less and less to do...The lack of large caseloads for each probation officer might eventually jeo- pardize their job security...(Justice) agencies would View themselves as moving with increasing speed along a pathway of self—annihilation (PP. 126-127). Since orgainzational "self-annihilation" is generally un- acceptable, justice agencies adapt diversion programs to their own goals and needs. Justice system organizations have managed to view diver— sion programs "as supplements to previous practice..., rather than as significant alternatives to previous practice" (Blom- berg, 1980). Service agencies will tend to increase the num- ber of youths under supervision when receiving additional resources, such as funding for diversion (Pabon, 1978). Blomberg's (1977) study of one diversion program revealed that court officials intentionally applied diversion program resources as a ”programmatic and service extension of infor- mal probation" rather than as a ”substitute for insertion into the formal juvenile court system" (p. 276). He explains further that: The net-widening brought on by diversion's 'dis- covery' of new clients is further influenced by the fact that justice agency funding is based upon numbers of client contacts or workload units (p. 584). Net-widening through diversion can thus enhance organizational viability. 28 To summarize, it is clear that a variety of forces influence the trend for diversion programs to widen the net of justice system control. Rutherford and McDermott (1976, p. 38) have summarized the explanations for net-widening this way: The conceptual confusion between prevention and diversion, the desire to offer services, and the general distrust that legal authorities have for nonprofessional helping agencies all lead to the development of Eggs programs for more juveniles within the existing systems of legal authority. There is always the danger, then, that diversion programs will serve to expand justice system control. Program moni- toring techniques and system-level evaluation strategies which can detect net-widening are therefore needed. More systematic research on the issue is necessary before net- widening's extent, causes, and solutions can be determined (Rutherford & McDermott, 1976; Seidman, Note 1). In the next section, the objectives and hypotheses of the present research will be outlined, followed by a presentation of the methodology. Objectives The present research attempts to fill the need for more conclusive empirical evidence on the net-widening phenomenon. By developing a more comprehensive evaluation strategy, the research also provides methods for improving the quality of future evaluations of diversion. Specifically, the study 29 applies a comprehensive multi-method evaluation strategy to the assessment of possible net-widening effects of one on- going juvenile diversion program. The diversion program under study is the Adolescent Diversion Program (ADP), which has been operating in a Mid- western urban area since late 1976. The program is a repli- cation and extension of a successful experimental diversion project conducted earlier in Illinois (see Davidson, Seidman, Rappaport, Berck, Rapp, Rhodes & Herring, 1977). Like the earlier program, the ADP was designed with the intention of avoiding net-widening. It utilizes an experimental design, where youths referred by the court are randomly assigned to experimental or control groups. The program provides an excellent setting for assessing the extent of net-widening in a well-intentioned and well-designed project. The multi-method evaluation approach tests four vari— ations of the net-widening assessment methods described earlier. Each presents a different type of evidence for the basic question under investigation: Is this diversion pro- gram serving the intended set of juvenile——those which would otherwise have received further court processing? The four methods, which will be discussed in more detail later, are the following: (1) Analysis of control group dispositions; (2) Discriminant analysis of a random sample of court cases; (3) Analysis of questionnaire data provided by intake referees; and (4) Analysis of system processing rates. These methods test different hypotheses, all related to the 30 basic question of net-widening effects. (1) If diversion is occurring as planned, ADP control group cases returned to intake referees for an alternative disposition will receive further court processing. (2) If diversion is occuring as planned, ADP cases should match-up better with cases receiving further court processing than with those dismissed at intake. This will be tested using discriminant analysis, which will compute a function based on variables which distinguish processed from dismissed cases drawn from a random sample of juvenile court files. The ADP cases will then be classified into either of the groups based on their particular characteristic pro— files on the discriminating variables. (3) If diversion is occurring as planned, intake referees will tend to choose further court processing when asked for their preferred alternative disposition for each diversion case. This will be assessed from a questionnaire filled out by referees at the time of each referral to the ADP. (4) If diversion is occurring as planned, and all other factors remain constant, the proportion of cases handled for- mally by the court should decrease after the initiation of the diversion program. System processing rates will be checked on a pre- and post-diversion program basis. These analyses will then be cross-checked in an effort to measure the convergence of evidence on the net-widening issue. The results should provide fairly conclusive evidence 31 of the extent of net-widening effects as a result of the ADP. The study will go on to discuss the findings in the context of how and why net-widening can occur as a result of diversion programming. Implications for future diversion efforts will also be discussed, with attention to the util— ization of monitoring and evaluation techniques for preven- tion or early detection of net-widening effects. Chapter 2 Method Setting The Adolescent Diversion Project (ADP) is operated as a court-level diversion program accepting referrals from a county juvenile court. Youths are referred only at the pre- liminary hearing stage of the court process, where court intake referees conduct an ”inquiry" about the youth and the petitioned offense. Youngsters are only eligible for diversion if they have admitted guilt to a petitioned offense, are not guilty of severe injury-to-person crimes, are not currently on formal court probation, and would not have otherwise been dismissed by the court at intake. Participation in the ADP is voluntary, although very few youths (about 5%) refuse to participate. Because of the experimental design of the project, youths are randomly assigned to either the experimental or control group. Con- trol group youth are returned to the intake referee for an alternative disposition. The ADP has no control over this disposition. Experimental group youth are assigned to volunteers, who are typically college students engaged in a field-experience course at the large state university in town or the local community college. The volunteers are trained in behavioral contracting and advocacy skills, and 32 33 supervised by ADP staff members throughout their 18—week intervention with the juveniles. The research presented here covers the first seven groups of youths referred by the court for diversion. These groups will also be termed "Years" in this report. Referrals were taken in the Fall for one group or "Year,” and in the Spring for the next group or ”Year,” such that each group does not actually encompass an entire year. The seven groups or ”Years" actually comprise a three-year period from late-1976 to early-1980. The project has continued to operate since that time, but relevant data is not yet avail- able for the subsequent years. Evaluation Design Because none of the four evaluation methods used to date in assessing net—widening have produced conclusive results on their own, a more comprehensive multi-faceted approach is used in this research. Four separate methods are applied to the data in an attempt to produce converging evidence on the issue. This system-level multi-method approach results in a more conclusive evaluation than previous approaches. Sampling procedures, instrumentation, data collection, and statistical analyses will be discussed separately for each of the four methods. Then, procedures will be presented for checking the degree of convergence amongst the methods. Method 1——Analysis of Control Group Dispositions Sampling. There were 62 control group cases during the project's first seven groups of referrals. All of these 34 cases were analyzed for their eventual dispositions. Instrumentation. This analysis required no construc- tion of instruments. Data collection. Control group cases were checked for their dispositions by examining their court files for "pre- liminary orders." These are documents that state the court's decision concerning the filed petition (denied or authorized) and the handling of the youth (released, informal/consent probation, or care and custody——meaning that a formal hearing has been scheduled). The data were collected by ADP staff members and coded into the above—named dispositional cate- gories. Analysis. The analysis of the data consisted mainly of computation and comparison of the proportion of controls receiving each type of disposition. Changes over time were noted by checking for trend disparities over the years. Method 2——Discriminant Analysis of a Random Sample of Court m Sampling. Archival data from court files has been col- lected to evaluate court decision-making. A random sample of about 14% of the intake disposition decisions occurring from late-1976 to late-1979 was collected, to correspond with the period of the diversion program under examination. This procedure resulted in the sampling of 364 cases. All ADP cases were sampled for comparison purposes. These totalled 268 cases, although 20 of those had to be dropped because of missing files. 35 Instrumentation. A coding instrument was developed which included all or most of the pertinent information in case files which could have some bearing on the intake dis- position decision. The variables on which data was collected are listed in Table 1. Data collection. Data for each case came solely from the archival files kept at the court. Data collection was done by trained university undergraduates and ADP staff mem— bers. Staff members, who were graduate students, trained the coders in small groups at the court by going over example case files. The coders went through the files using a code- book which explained each variable and then coded the data to Opscan forms. Some earlier cases were coded onto other forms and then transferred later onto Opscans or were key— punched onto cards. Data collection took place intermittently over about a three-year period from 1977 to 1980, which en— tailed the use of many different coders. Sixteen different people coded cases during this time span. Reliability of coding was measured by percent agreement on each variable. Variables ranged from 56% to 100% in reli- ability, with about half the variables at 90% or better. Variables measuring parent's comments at the intake hearing had lower reliabilities because of the difficulty in rating these remarks. Reliabilities for the variables are listed along with the coding conventions on Table 1. Analysis. On the basis of a discriminant analysis, it was determined which variables combine to best distinguish 36 Table 1 Coding of Decision Sample Variables and Reliabilities* *Percent agreement in decimals Reliability .98 Sex .90 Living with (with whom .74 .77 .97 .91 .96 .94 1.00 .85 .63 .76 .70 the youth was living) Age # Siblings known to court (siblings with court petitions) # Days to hearing (days between subscribed and sworn date of petition and the preliminary hearing) Plea at hearing (plea entered by youth at pre- liminary hearing) Mother present (at pre- liminary hearing) Father present Police present Minor explains (situation at preliminary hearing) Parent's comments (rating of impact of comments at preliminary hearing) Parent blames other (parent blames someone or something other than the youth) Parent blames youth (parent blames youth for offense) 1=male; 2=female 1=both natural parents; 2=other family situation; 3=out of home 1-17 years O—9+ O—365+ 1=not guilty 2=guilty 1=no 2=yes same same same 1=negative 2=neutral 3=positive 1=yes 2=no 1=yes ‘2=no .72 .65 .67 .66 .94 .98 .56 37 Table 1 (cont'd.) Parent supports youth (in regard to the alleged offense) Parent wants help (parent wants help handling the youth) Par. on youth's behavior (parent's value judgments about the youth's behavior) Par. on youth's attitude (parent's value judgments on the youth's attitude) Seriousness of offense Committed alone or not (whether youth committed offense alone or with others) Recency of prior petition (how recent the most re- cent prior petition was) Parent present (whether either parent was present at preliminary hearing) Family as petitioner (family member filed the petition) School as petitioner Store as petitioner Police as petitioner Courtworker asypetitioner Parent comments or not (whether parent comments at preliminary hearing) Parent as complainant (parent as person who originally complained about the offense) 1=yes 2=no 1=yes 2=indifferent 3=no 1=positive 2=neutral 3=negative 1=positive 2=neutral 3=negative 1-27 (modified Sellin/ Wolfgang scale) 1=alone 2=with others 0=no priors 1=12+ months 2=6—12 months 3=1-6 months 4=less than 1 month 1=no 2=yes 1=no 2=yes 38 Table 1 (cont'd.) Victim as complainant same Police as complainant same .62 # Prior petitions (total 0—13 number of petitions prior to current petition) Cum. seriousness ofypriors 0-40+ (cumulative seriousness of .78 .96 .97 prior petitions) Most serious prior (seriousness of the most serious prior petition) Status offense or not (whether the alleged offense is a status offense) Person offense or not (whether the alleged offense is an injury to person offense) Disposition (disposition at the preliminary hearing) 0-27 (modified Sellin/ Wolfgang scale) 1=no 2=yes 1=no 2=yes 1=denied or dismissed 2=adjourned 3=formal hearing 1Reliability is for "Type of petitioner” 2Reliability is for "Type of complainant" 3Reliability is for "Type of offense" 39 cases released at intake from those which receive further court processing. The discriminating function which best predicted group membership for the random decision sample was then applied to ADP cases. The discriminant analysis computer program then classified ADP cases into one of the groups based on the data collected on distinguishing vari- ables. Diversion cases were then checked for the proportions that appeared similar to either the released-at-intake group or the further-processing group. The discriminant functions were derived from an analysis of cases randomly sampled from court files. An approximate 14% sample was taken, yielding 364 cases covering the three— year span of the project under study here. Cases which were on formal probation or in institutional placement at the time of the sampled preliminary hearing were dropped from the same ple. These cases were ineligible for referral to the diver- sion program, and therefore inappropriate for inclusion in the analysis. Also dropped were cases which had been referred to the diversion program at some point. These were also ineligible for ADP referral. The sample was reduced to 253 cases after excluding these types of cases. Method 3——Analysis of Questionnaire Data Provided by Referees Sampling. At the time of referral to the diversion pro- gram, the intake referee is asked to fill out a questionnaire about his impressions of the youth. This "Referee Referral Form" was not used for the first two groups referred to the ADP, so the total available cases number 208. Of these, 28 40 were missing or had never been filled out. All available questionnaires were used in the analysis. Forms were filled out regardless of whether the youth turned out to be a con- trol case or an experimental case. Instrumentation. The Referee Referral Form is included here in Appendix 1. It asks several types of questions about the intake referee's impressions of the youth, from his/her likelihood of success in the project to a rating of needed areas of improvement for the youth. The pertinent question for this research, however, is that listed as No. 3: "If this youth does not get into the Project what alternative will you recommend?" The referee is given the choices of dismissal, informal/consent probation, referral to another agency, or formal handling. Data collection. Referral forms are turned over to the ADP staff member at time of referral, or shortly thereafter. Data were coded according to the dispositional categories listed in the above item. Analysis. Responses to the item were totalled, and proportions of cases falling into each dispositional alter— native compared. Method 4——Analysis of System Processipg Rates Sampling. Total processing rates for all cases handled by the juvenile court were analyzed for the four-year period preceding the start of the diversion program and for the three-year period of its operation under examination here. All cases in the population handled by the court were included. 41 Data collection. Data were collected from official court records, which consist of monthly lists of petitions filed at the intake department. These are kept by the in— take supervisor and are marked with the preliminary hearing disposition. Analysis. Proportions of cases processed by the court in each dispositional category were computed both monthly and yearly. Chi-square tests were used to analyze global pre- vs. post-project proportions. Changes in organizational procedures or policy decisions were investigated as possible confounding variables. In addition, the results from a time-series analysis of the data will be presented. Using the Box and Jenkins (1976) approach, the analysis checks for shifts in trend or drift of the data, and tests for any significant change in the level of the dependent variable as a result of the inter- vention. The intervention impact segment of the analysis is based on computer analysis techniques derived by Glass, Willson and Gottman (1975). The time-series analysis inves- tigates correlational patterns in the data over time that may not be readily apparent by visual inspection of the graphed data. Methods of Checkingvfor Convergence of Results After the four types of analyses were complete, the results were compared across method. The first three methods each produce proportions of ADP cases classified into the various dispositional alternatives to diversion (release, 42 informal/consent probation, formal handling). These pro— portions were compared amongst the methods, and differences tested by chi—square analyses. In addition, results of the first three methods should be reflected in the system pro- cessing figures derived in method four. If ADP cases seem similar to those handled formally by the court, then the overall proportion of formal cases should show a corres- ponding decrease. All things being equal, this should show up in the system processing rates. Beyond the group comparisons across methods, classifi- cations of individual ADP cases were also compared across methods. Discriminant analysis classification of each case into its appropriate dispositional type was compared with the intake referee classification. For control group cases, there was the additional comparison on the actual alternative disposition ordered for each case. Chapter 3 Results The results of the analyses described previously will be presented separately for each given evaluation method. Along with the totals for classifications of cases into the dispositional categories, yearly figures for each method will also be presented. Next, the results of the comparisons of the various methods will be given. The general comparisons will be presented first, followed by the tests for conver- gence done on an individual case basis. Because of the importance of the dispositional cate— gories to this study, some explanation of dispositions is necessary. In collecting the data for this study, it was found that several different dispositions could result from the preliminary (intake) hearing. These are listed below: 1. Petition denied—~youth dismissed from court juris— diction. 2. Petition denied——youth referred to the Adolescent Diversion Project. 3. Petition authorized——youth placed on informal or consent probation with an intake referee as case supervisor. The case is reviewed in 30 to 90 days, at which time the youth may be dismissed from court jurisdiction, continued on informal probation, or 43 44 sent on to a formal court hearing if another petition has been filed. 4. Petitioned neither denied or authorized——hearing is adjourned for 30 days, usually so that the intake referee can observe, investigate, or monitor the youth. The case then has another preliminary hearing, at which time any of the other dispositions may be applied——dismissal, informal/consent pro- bation, or the scheduling of a formal hearing. 5. Petition authorized——youth is scheduled to appear for a formal hearing, and a caseworker is assigned to investigate and make recommendations to the court. At the formal hearing, the youth may be dismissed from court jurisdiction, placed on formal probation, or placed in an institutional treatment facility. In practice, the intake referees rarely used option 3, in- formal or consent probation. Instead, cases were often adjourned for 30 days and placed under an intake worker's supervision or "observation." For analytical purposes, these two categories were considered identical. The dis- tinction appears to be only legal in nature, the youth in question experiencing the same degree of court supervision. The two categories of adjournment and informal probation were used differentially by intake referees, however. While only 4 of the 253 cases collected in the random sample of court files were given informal/consent probation, 23 of 60 45 ADP control group cases received this disposition. In addition, the referees who helped to develop the "Referee Referral Form" (see Appendix 1) listed "informal probation" as one of the options for alternative dispositions, but did not include ”30—day adjournment." Possible reasons for this situation will be considered in the discussion section. For now, however, it is safe to assume that the dispositions are functionally equivalent. Method 1——Analysis of Control Group Dispositions As shown in Table 2, dispositions received by the ADP control group could be found for 60 of the 62 cases in the first five project ”years." Only 15 of these, or 25.0%, were given formal handling by the court. The rest were split evenly between outright dismissal (36.7%), and infor- mal probation (38.3%). Given the small number of control group cases in individual years, it is difficult to inter- pret time trends in this data. However, it appears that the percentage of cases processed formally by the court increased somewhat over the years, rising to 43% in ADP Year V. The other categories fluctuated randomly over time. Combining the dismissed and informal probation groups into one category, and then analyzing a crosstabulation of time by disposition, it was found that the chi-square statistic was not significant. This indicates that the fifth group increase in the percent of youths sent to a formal hearing did not constitute a statistically significant change in the dispositional proportions over time. 46 Table 2 Dispositions of ADP Control Group Cases Years I—V ADP Informal Formal Year Dismissed Probation Handling Totals Missing I 3 5 2 10 0 30% 50% 20% II 6 2 1 9 0 67% 22% 11% III 3 9 3 15 1 20% 60% 20% IV 4 5 3 12 0 33% 42% 25% V 6 2 6 14 1 43% 14% 43% I-V 22 23 15 60 2 36.7% 38.3% 25.0% Method 2——Discriminant Analysis of a Random Sample of Court 93.822 This method was used to develop an empirically-based function which could predict the likely disposition of ADP cases had they p9; been diverted. Discriminant functions which best distinguished the dispositional groupings of cases were first derived, and then tested for significance. Applying the characteristics of each diversion project case to the discriminant functions, the analysis then classified each case into its most probable dispositional group member- ship. Table 3 lists the variables which entered the discri— minant functions formed to distinguish the three dispositional 47 Table 3 Discriminant Analysis Summary Table Standardized Canonical Disc. Function Coefficients Variable Wilks Entered Lambda Rao's V Function 1 Function 2 Parent's Comments .742 87.104 -.351 —.737 Recency of prior petition .625 145.905 .349 —.268 Plea at hearing .521 222.094 -.641 .103 Detention .481 257.694 .291 -.254 Living with .457 277.889 .210 —.503 Age .436 295.624 .143 .517 Parent wants help .423 311.607 -.231 —.091 Parent blames other .411 326.249 -.218 .116 Committed alone or not .399 341.439 -.129 .104 Police present .389 353.225 .213 .216 # Prior petitions .381 365.034 .273 -.006 Seriousness of offense .376 373.137 .240 .012 Status offense or not .368 384.298 .099 -.383 Victim as complainant .363 391.975 .165 -.210 Police as petetioner .357 400.592 -.334 -.035 Store as petetioner .350 410.491 -.227 -.379 Person offense or not .343 416.800 -.053 -.368 48 Table 3 (cont'd.) Father present .337 421.675 .011 -.386 Parent on youth's attitude .334 426.119 .143 —.301 groupings——Dismissed, Adjourned, Formal Handling. Nineteen of the 37 variables entered the equation, resulting in a Wilks Lambda statistic of .334. This indicates that about 67% (1-lambda) of the variance was accounted for by the functions. This high level of explained variance shows that the functions present a good model of the data. Table 4 illustrates the ability of the derived functions to adequately distinguish between cases in the three groups. The between-groups F statistics for each pair of groups are all significant beyond the .001 level. The placement of the group centroids on the derived canonical discriminant func- tions indicate that the first function best distinguishes the dismissed cases from the formal cases, while the second function polarizes the adjourned from the dismissed cases. In Table 5, the important variables in each function are presented, along with a comparison of the mean values for each group. Possible application of this information on significant distinguishing variables will be discussed in a later section. The definitions and coding schemes for each of the variables used in the analysis are presented in Table 1. Another measure of strength of the discriminant analysis is the accuracy of the group classifications. The original 49 Table 4 Discriminant Analysis Between Groups Statistics Between groups F statistics (df = 21, 230) Group 1-Dismissed 2—Adjourned Group 2.6173 2-Adjourned .0004 3-Formal 14.687 11.344 .0000 .0000 Canonical Discriminant Functions at Group Centroids Group Function 1 Function 2 Dismissed -1.45806 -.58215 Adjourned - .99763 .63974 Formal 1.20730 -.05743 cases are applied to the derived functions and then classi- fied into their most probable category. As noted in Table 6, the analysis correctly classified about 76% of the cases. The "adjourned" category had the highest error rate, with about 36% misclassified. The next step in the analysis consisted of predicting probable group membership of ADP cases based on the functions derived from the randomly sampled cases. If diversion was being carried out as planned, most or all of the cases should be predicted to be "formals." Table 7 shows that about 57% of the diversion cases would have most likely fallen into the "adjourned" category. Thus, had they not been referred to the diversion program, these cases would have probably been adjourned for about 30 days and then dismissed. Less than 10% of the adjourned cases typically receive formal 50 a we? wanna was» cw moanmfipw> no“ :ofluopfiuo :oflmsaocH oocmfihd> Hugo“ one wo xw.ww How oopcaooom H :oflpocsh bmo.l NN.N 5H. 00. em. Hm.H mw.H N®.¢H mb.H SON.H mm. hm.H mb.H 0®.H HNEHOW bb.H no. mH. vH. mv.H Ob.H HN.¢H mm.H >5. mm. mm.H mm.H AQQO HGOHOHHHQOU GOHQOGHXH .OMHQ HNOHGOCMU UQNHUHNUGMHW mwm.l "mesonucoo Qsonu as Amy .oczm mm.H Hom.a NH. wmm.| mm. msm.n om. mwm.| me.H owm.| m®.H mom.i mb.mH ham. >m.m ems.i m :oHpocsm wme.fiu "mefiocpeoo macho pa AHV .oesm no. wmm.u ma. mvm. sm.m Hmm.| m.H Heo.n oommfiEmHQ mucoflofiuwooo .ocsm .omam emueecaeaaHm H GOApoczm pcmcfiEHpomfiQ mcflhopqm moanmfihw> ache: m mande .mcowaocam m one an how oopzsoood .wmm. n moped; wxaflz Haweo>o .opoz .omHQ Haoflcocso onspfippm m.nwsom co vacuum you so omcommo compom poaofipfiaoa mm opowm you no owcommo mapsam unomonm nonpem spas wcfi>flq ow< mucoeeoo m.p:ohwm .omfia Hwoficocwo HoGOprpoQ mm ooHHom .cupoa uofiua mo mocooom mucoEEoo m.u:ohwm wcflumo: as «man moanmwpm> F1 51 Table 6 Discriminant Analysis Classification Results Predicted Group Membership Actual Group N of cases Dismissed Adjourned Formal Dismissed 60 39 15 6 (23.7%) (65.0%) (25.0%) (10.0%) Adjourned 66 12 42 12 (26.1%) (18.2%) (63.6%) (18.2%) Formal 127 6 10 111 (50.2%) (4.7%) (7.9%) (87.4%) Totals 253 57 67 129 (22.5%) (26.5%) (50.9%) Percent of grouped cases correctly classified: 75.9% court handling. About 27% of the ADP cases were classified as being similar to those which were sent on to a formal hearing. The remaining 16.5% fit the characteristics of cases immediately dismissed at the preliminary hearing. While the time trends for dismissed and adjourned cate- gories fluctuated over the years, the percentage classified under formal handling peaks at around 40% for the first two years and declines to around 20-25% over the remaining period. However, the chi-square statistic for this changing propor— tion of cases handled formally was not significant. Method 3——Analysis of Questionnaire Data Provided bijeferees In this method, intake referees indicated the most Probable alternative disposition they would recommend for a Youth being referred to the diversion program. This data, ‘taken from the "Referee Referral Form" (Appendix 1), was available for ADP years III-VII. The questionnaire had not 52 Table 7 Discriminant Analysis ADP Years I—VII ADP Dispositions Formal Year Dismissed Adjourned Handling I 1 16 13 3% 53% 43% II 6 11 10 22% 41% 37% III 11 40 16 16% 60% 24% IV 4 22 6 13% 69% 19% V 13 23 11 28% 49% 23% VI 1 8 3 8% 67% 25% VII 5 21 7 15% 64% 21% Total 41 141 66 I-VII 16.5% 56.9% 26.6% Total 30 27 67 32 47 12 33 248 100% 53 Table 8 Alternative Dispositions for ADP Cases as Recommended by Intake Referees Dispositions ADP Informal Formal Missing Grand Year Dismissed Probation Handling Totals Forms Totals III 7 48 14 69 4 73 10% 70% 20% IV 1 20 9 30 6 36 3% 67% 30% V 4 18 14 36 15 51 11% 50% 39% VI 2 3 6 11 3 14 18% 27% 55% VII 1 16 17 34 0 34 3% 47% 50% III- 15 105 60 180 28 208 VII 8.3% 58.3% 33.3% 100% (13%) of grand total been developed for use in the first two groups referred to the project. As shown in Table 8, 28 of the 208 cases had missing forms, or 13%. In addition, it was found that the forms were somewhat differentially missing between control group cases and experimental group cases. Of the controls, about 26% had missing forms, compared to about 10% of the experimentals. If diversion were occurring as planned, most of the ADP referrals should have been slated for a formal hearing as their alternative disposition. According to the intake referees, however, the most likely alternative disposition for cases referred to ADP was informal probation, which was chosen about 58% of the time. Formal handling was next, 54 with 33%, while the dismissal category accounted for only 8% of the cases. The percentage recommended for formal handling increased over the years, from 20% in Year III to 50% in Year VII. The percentage recommended for informal probation showed a somewhat corresponding decline over time. The trend towards higher proportions of formal cases versus all others was tested by a chi—square analysis, which was significant (K = 12.4, p < .01). General Convergence of Methods 1, 2, and 3 Before checking the analysis of system processing rates (method 4), the three methods which yielded dispositional classifications were compared. Table 9 presents a summary of the results for each method shown separately in earlier tables. Although there is some variation, the three methods produce very similar results. Diversion program referrals appear most similar to cases which would otherwise have received the "informal probation” or "adjourn for 30 days" disposition. In the discriminant analysis (method 2), 57% of the cases were classified in this category. Referees recommended ”informal probation" as the preferred alternative disposition for 58% of the total diversion referrals. While a lesser percentage, 38%, of control group cases actually were given this disposition, it was still the most frequently used category. The three methods resulted in very similar percentages of cases similar to those processed formally by the court. This category of ”formal handling” represents cases which 55 Table 9 General Convergence of Methods 1, 2, and 3 Dispositions Informal prob. Formal Method Dismissed or Adjourn. Handling Totals 1. Control 22 23 15 60 group dis— . . positions 36.7% 38.3% 25.0% 2. Discrimi- 41 141 66 248 nant analysis . classifications 16'5% 56'9A 26'6% 3. Referee's 15 105 60 180 alternative dis- 8 3% 58 37 33 37 position recom- ' ° ° ° ° mendation are in the most legal jeopardy, and which most deeply pene- trate the court system. As noted earlier, cases with the most likelihood of further court processing are the ideal intended target group for diversion. In method 1, the results show that 25% of the control group cases were sent on to formal court handling after their ADP referral. In method 2, the discriminant analysis found that about 27% of the ADP cases were similar to cases in the random court sample receiving formal handling. In method 3, intake referees recommended formal handling as their alternative disposition for 33% of the ADP referrals. VWhile the "informal probation/adjourn for 30 days" cate- gory represents a gray area for diversion program intentions, the "dismissal" category of cases was clearly to be avoided. However, almost 37% of the control group cases were dis— missed from the court after being referred to the diversion 56 program. The discriminant analysis indicated that only about 17% of the diversion referrals were similar to most dismissed cases. And only 8% of the diversion referrals were recommended by intake referees for dismissal if they did not get into the ADP experimental group. Possible ex- planations for these results will be considered in the Dis- cussion Section. The results from the three methods fluctuate consider— ably on a year-to—year basis, and there are few clear time trends in the data. The percentage of cases similar to those in the “formal handling” group, however, does exhibit some interesting trends over time. Figure 1 plots the "for- mal handling” percentage found from each method. Although the time periods are not strictly equivalent for the entire graph, note that the percentages in this group £i§g_over the years for control group cases and referees' recommended alternative dispositions. In contrast to this, the corres- ponding percentage of cases classified as "formals" by the discriminant analysis declines over the years. This result will also be discussed later. Method 4——Analysis of System Processing Rates To test the impact of the diversion program on the court system, the rates of dispositions before and after the program were compared. It was expected that if the program were serving the intended set of clients, the percentage of cases handled formally by the court would decrease over time. Figure 2 presents the court population rates on a yearly 57 60-7 R 50--» R '2‘ o D C m 40- S R m D o m m ‘3‘ 30-- R C D D D zo-u» c ‘ D E D 10-4 C I II III IV V VI VII ADP "YEARS" C = Control group dispositions D = Discriminant analysis classifications 'R = Referees' alternative dispositions Figure 1. Percent of cases classified as "formal" over time. PERCENT FORMAL 58 Pre-ADP 7o»- 50»- 404- 30-- # 4 l Post-ADP I J 1973 1974 1975 1976 52.5% 56.4% 60.5% 59.4% #F . 582 639 710 614 N ' 1108 TIE? 1173 1034 Totals: Pre-ADP 57% 2511/4371 1977 1978 1979 55.2% 57.1% 45.9% 3.21.42.81.21 883 715 939 Totals: Post-ADP 53% 1360/2565 Figure 2. Percent of cases handled formally by year (court population figures). 59 basis, including total cases processed and the number pro- cessed formally. The overall percentage of cases handled formally pre-project, 57%, was not significantly different from the post-project figure of 53%, using a chi-square test of significance. The seven-year random sample of court cases also showed non-significant differences for the pre— vs. post-comparison. Because of a change in court policy regarding status offen— ders, the sample was used to check changes across time while controlling for the policy shift. Shortly after the diver- sion program began, the court instituted a new policy of refusing to accept petitions on first-time status offenders. The pre- vs. post-comparison was then checked after removing status offenders from the sample. The court processed 45% of non-status offenders formally before the project began, and 50% afterwards. This is also a non-significant differ- ence, according to the chi-square test. In addition to this general comparison, a time—series analysis was done to investigate the possible impact of the project on court decision—making. For this analysis, monthly rates of cases handled formally by the court were needed. However, the monthly lists kept by the court staff on new petitions proved to be an unreliable source of data. Cases whose intake hearings were held in January would sometimes not show up on the lists until July. Because of this unre— liable record-keeping, figures were drawn instead from a monthly breakdown of the random sample of cases coded from 6O court files. The accuracy of the hearing dates in the sam- ple had been checked and confirmed, providing more confidence in the data despite the loss in monthly "n." Figures 3 and 4 present graphs of the data, for all cases in the sample, and for non-status cases, respectively. Both plots were best described by the time-series ARIMA (0,0,0) model-—essentially a "white-noise" model indicating little change in trend or drift. The (0,0,0) model indicates that there were no autoregressive or moving average para- meters needed to describe the time series, and that no dif- ferencing of the data was needed to produce stationarity. The data was tested using the Box—Jenkins (1976) time-series computer programs. The graph of the autocorrelation function (ACF) indicated that the series did not predict itself—— there were near-zero, non—significant correlations between all lags of the data. This was true for both the pre-ADP and post-ADP series of data. Because there were no inter— correlated time-lags in the data, the pre-ADP cases and post— ADP cases could be considered as independent samples and compared with a p-test of significance. For the analysis of all cases, the pre-ADP sample of cases handled formally had a mean of .56, with .12 SD. The post-ADP sample averaged .54, with .17 SD. The p-statistic, with 83 df, was non-significant. For non—status cases, the pre-ADP sample mean was .45, with .23 SD. The post-ADP sample averaged .50, with .19 SD. The p-test for this com- parison was also non-significant. 61 Pre-ADP Post-ADP 70-4- X X 60-6 x x x SO-L o m a 40-w z m o m 33 ao-I zo-L IO-w I I l I I I I ‘T T j 1 l l— l 1973 1974 1975 1976 1977 1978 1979 49.6% 55.3% 61.5% 61.1% 55.7% 55.1% 49.6% #F . 65 73 72 69 64 65 63 N ' 131 132 117 113 115 118 127 Totals: Pre-ADP Totals: Post-ADP 56.6% 53.3% 279/493 192/360 Figure 3. Percent of all cases handled formally by year (all cases in sample). 62 Pre-ADP Post-ADP 60-- .5 O l I X PERCENT FORMAL (o o I I 20-- 10-- 4. l I 1 1 1 n I I 77’ I If i 1973 1974 1975 1976 1977 1978 1979 38.2% 46.5% 54.4% 40.4% 50.7% 53.8% 48.2% .1522 _49 _l .2}. _3_9 _5_7. .53 N' 6 86 68 52 7 106 110 Totals: Pre-ADP Totals: Post-ADP 45.0% 50.9% 127/282 142/291 Figure 4. Percent of cases handled formally by year (non-status cases only). 63 Analysis of Status Offender Policy Change Because of the possible confounding effects of the status offender policy change, additional analyses were done to investigate these potential effects on court decision— making. In the pre—diversion program sample of cases (1/‘73 through 11/‘76), status offenders made up 42.8% of the total cases. A large majority of these cases (72%) were processed formally by the court. The change in policy came around August of 1977, after which the percentage of status offen— ders in the sample dropped to 14.4% (8/‘76 through 1/‘80). For the first eight months of the ADP's existence, however, status offenders still comprised about 40% of the sample. The proportion of these cases which were handled formally remained at about 64-65% throughout the post—ADP period. The eight-month period at the beginning of the diver- sion program before the policy shift involves the first two ADP "years.” Thus, while these 57 cases were being referred to the project, status offenders still made up about 40% of the court's caseload at intake. This fact is reflected in the make-up of ADP referrals; 32% were status offenders in these first two ADP "years," compared to 10% for the fol- lowing five ”years." Table 10 summarizes these figures. As noted in Table 7 and Figure 1 above, the percentage of cases classified in the ”formal handling" group by the discriminant analysis declined after the first two project "years" or groups. This is in accordance with the drop in status offenders, most of whom were handled formally. In 64 Table 10 Analysis of Status Offender Policy Change Pre-ADP Post-ADP Post-ADP (1/‘73——11/'76) (12/‘76——7/'77) (8/‘77—-1/'80) #Status/#Total #Status/#Total #Status/#Total 211/493 25/62 43/298 42.8% 40.3% 14.4% Status sent Formal Status sent Formal Status sent Formal 152/211 16/25 28/43 72.0% 64.0% 65.0% Status in ADP Status in ADP 18/57 19/191 31.6% 9.9% Status classified as ”Formal” by Discri- 12/18 6/19 minant Analysis: 66.7% 31.6% the first two ADP years, 67% (12/18) of the status offenders referred to the project were classified in the formal hand- ling category by the discriminant analysis. However, in addition to the sharp decline in status offenders referred in the the next five project groups, only 32% (6/19) were classified as "formal" cases. Most of the status offense cases referred to the ADP after the policy change must have therefore had characteristics which distinguished them from the typical set of status offenders. Conveggence Testing of Methods on Individual Cases In addition to the general comparisons of results pre- sented earlier, an analysis of the first three methods was done for classification differences based on individual cases. Each method's classifications were compared with the individual cases for which data was available from the other 65 methods. Control group cases occurred only in Years I to V of the project, and the Referee Referral Form (method 3) was used only in Years III to VII. The discriminant analysis covered all cases from Years I to VII. Method 1 vs. method 2——control group vs. discriminant analysis. The discriminant analysis classifications of cases agreed with the actual dispositions of controls on 30 of the available 58 cases. This was a 52% rate of correct classifi- cation. The chi-square test of the 3 x 3 table showed that the methods were significantly associated (K = 11.76, df = 4, p < .019). The gamma coefficient, indicating the strength of the relationship was .437. Of the disagreements between methods, 19 involved a more serious disposition predicted by the discriminant analysis than was actually given to the control group case. Only 8 of the 22 dismissed control group cases were correctly pre- dicted by method 2, with 11 of these misclassified as "ad- journed for 30 days/informal probation." The classification comparisons are summarized in Table 11. Method 1 vs. method 3——control group vs. referee's recommendations. These methods had comparable cases only for Years III to V, resulting in only 29 cases. They agreed on 18 of the 29 cases, for a 62% agreement rate. Method 3 also had trouble agreeing with the number of control group cases which were dismissed. While the referees recommended dismissal for only 2 of these cases, 9 were actually dis- missed. Most of these had been recommended for "informal 66 Table 11 Classifications of Individual Cases by Methods 1 and 2 Method 1 Method 2——Discriminant Analysis Control Group Formal Dispositions Dismissed Adjourned Handling Totals Dismissed 8 11 3 22 (Col%)/(Row%) 57%/36% 39%/50% 19%/14% 37.9% Informal Prob. 3 14 5 22 21%/14% 50%/64% 31%/23% 37.9% Formal Handling 3 3 8 14 21%/21% 11%/21% 50%/57% 24.1% Totals 14 28 16 58 24.1% 48.3% 27.6% probation" by the intake referees in filling out the referral form. While 21 of the control group cases were recommended for informal probation on the referral form, 12 of these actually received that disposition. Six of the others were dismissed, and 3 were sent on to formal handling. The chi- square test was significant (K = 11.99, df = 4, p < .017). and the gamma coefficient was .735. Table 12 summarizes the results of this comparison. Method 2 vs. method 3——discriminant analysis vs. referees' forms. These methods had comparable cases for ADP Years III-VII, for a total of 165 cases. There were 28 missing referral forms. The methods agreed on 98 of the 165 cases, for a rate of about 59%. Most of the disagreements, 48 or 67, were in cases where the referee recommended a more serious disposition than that which fit the discriminant analysis classification of the case. For example, the dis- criminant analysis classified 32 cases as "dismissed" or 67 Table 12 Classifications of Individual Cases by Methods 1 and 3 Method 1 Method 3——Referees' Recommendations Control Group Informal Formal Dispositions Dismissed Probation Handling Totals Dismissed 2 6 1 9 (Col%)/(Row%) 100%/22% 29%/67% 17%/11% (31%) Informal Prob. 0 12 1 13 57%/92% 17%/7% (44.8%) Formal Handling 0 3 4 7 14%/43% 67%/57% (24.1%) Totals 2 21 6 29 (9.5%) (72.4%) (20.7%) "adjourn for 30 days," when the referees' alternative recom— mendation had actually been "formal handling." On the other hand, 23 cases predicted as "dismissed" by the discriminant analysis were recommended for "informal probation" or "for— mal handling" by the referees. The chi-square test for this crosstabulation, which is shown in Table 13, was significant (§.= 23.26, df = 4, p < .0001), and the gamma coefficient was .473. Method 1 vs. method 2 vs. method 3. In the final test of convergence, the three methods were compared on the 29 cases for which each had data. These are cases from ADP Years III to V. The methods agreed perfectly on 10 of these, or about 35%. There was partial agreement (two of the methods agreed) on another 17 of the 29, or 59%. Only 2 of the cases had completely different dispositions for each method. Because of the three-dimensional matrix required to present the results in detail, only a summary table will be 68 Table 13 Classifications of Individual Cases by Methods 2 and 3 Method 3 Method 2——Discriminant Analysis Referees' Formal Form Dismissed Adjourn Handling Totals Dismissed 4 7 1 12 (Col%)/(Row%) 15%/33% 7%/58% 3%/8% 7.2% Informal Probation 16 71 11 98 59%/16% 69%/72% 31%/11% 59.4% Formal Handling 7 25 23 55 26%/13% 24%/46% 66%/42% 33.3% Totals 27 103 35 165 16.4% 62.4% 21.2% presented here (Table 14). This table again shows consider— able similarity in the dispositional categorization of diver- wion program cases. This is especially true for the impor- tant ”formal handling" category, where the three methods differ at most by only about 3%. while methods 1 and 2 each show nine cases in the It should be noted that "dismissed" category, these are not the same nine individual cases. In actuality, there are four cases common to the two methods in this category. While the three methods do not show exact agreement of classification of individual cases, the sub- category totals are fairly consistent. all of the results will be summarized and discussed. In the next section, 69 Table 14 Summary Comparison of Methods 1, 2, and 3 Dispositions Methods Dismissed Inf/Adjourn Formal Totals 1. Control Group 9 13 7 29 Dispositions 31% 44.8% 24.1% 2. Discriminant 9 14 6 29 Analysis 31% 48.3% 20.7% 3. Referees' 2 21 6 29 Recommendations 6.9% 72.4% 20.7% Chapter 4 Discussion In this section, the results will be summarized and dis— cussed in relation to the hypothetical "net-widening" effects of diversion programs. The methods will be compared to assess the relative validity of the findings across method, with attention to the results of the testing for convergence. The findings will then be discussed in terms of conclusions that can be drawn about the diversion program under study and its goals. After comparing the results with those of other studies, a more general discussion of probable causes of the net-widening phenomenon will follow. The section will close with suggested ways to avoid net-widening, some thoughts on the future of diversion programs, and recommen- dations for future research on this topic. Method 1 The analysis of control group dispositions showed that only a quarter of these control cases are sent on to formal handling by the court. About 37% are outright dismissed from court jurisdiction, which indicates that these cases were in no jeopardy of deeper court system involVement. While another 38% received informal probation, this typically involves little court intervention. According to these results, assuming generalizability to all ADP cases, a 70 71 substantial number of ADP referrals were clear instances of "widening the net" of the justice system. If 37% of all project youth would otherwise have been released from the court, their "diversion" program involvement represents more system intrusion than they need have experienced. The 38% receiving informal probation are not clear examples of truly diverted cases, since there was some, albeit minimal, court penetration. However, they were diverted from some amount of court supervision and investigation. Are the control group dispositions valid as reference points for the typical diversion program referral case? Perhaps, but there are a few threats to the validity of the findings from this method. For one thing, the number of cases is very small, about one-quarter of the total. In addition, there were no control group cases in ADP groups VI and VII. If types of cases differed between years, then control group cases would not be reflective of these later years. Also, and probably most importantly, there is a confounding effect stemming from the cases being "rejected" by the ADP and sent back to the referee for another dis- position. The referee may give a lighter disposition to the youth who is not accepted into the diversion program service component. In fact, as seen in Table 9, control group cases were dismissed in higher proportion than that predicted by either the discriminant analysis or the referees' question- naires. And in the analysis of individual case classifica- tion convergence between methods, control group dispositions 72 tended to be less serious than "predicted" by the other methods. Method 2 The classification of ADP cases by the discriminant analysis showed that most project referrals were similar to court cases which were given informal probation or adjourned for 30 days. In contrast to control group dispositions, only about 17% of the cases were classified as similar to dismissed cases. About the same amount, 27%, fit the char— acteristics of "formal" cases. Comparison of these results with the typical composite picture of court cases aids in the interpretation of the findings. In Table 6, it was noted that 50% of the random sample of court cases received formal handling. Almost half of this proportion of ADP cases would have been processed formally, however, according to the discriminant analysis. While 26% of the court sample cases were adjourned for 30 days or more, about 57% of the ADP cases would have likely received this disposition. Since nearly all adjourned cases are eventually dismissed, a substantial majority of ADP cases would never have had any "formal" court involvement. These findings can be interpreted differently depending on one's perspective. If one of the goals of diversion is to serve only those cases which would not otherwise have been outright released by the court, the Adolescent Diversion Project can be said to have succeeded in having about 17% of their cases be of this type. However, if the goal of 73 diversion is to serve only those cases which would other— wise have penetrated the court system to the point of having a formal hearing, then the ADP failed. Only about 27% of the ADP referrals would otherwise have received formal hand- ling, according to the discriminant analysis. Was discriminant analysis a valid method for distin- guishing the dispositional groupings of cases and classifying ADP youths accordingly? All of the evidence seems to support an affirmative answer. The variables entering the functions accounted for about 67% of the variance (Table 3). The between—groups 3 statistics were all significant at the p < .001 level, indicating that the three types of disposi- tions involved highly distinguishable types of cases (Table 4). In addition, the discriminant functions correctly classified 76% of the cases from which they were derived (Table 6). The variables which distinguished the three categories are shown on Table 5. "Formal" cases were more likely to have a ”not-guilty" plea by the youth at the hearing than either of the dismissed or adjourned cases. Formal cases were also more likely to have had negative comments made by the youth's parents at the hearing, but adjourned cases had more negative parent comments than dismissed cases. Formal cases were much more likely to have had a prior court record than either dismissed or adjourned cases. Adjourned cases were more likely to have been petitioned by the police than the other cases, however. Formals and adjourned cases tended 74 to be about a year older than dismissed cases (K = 14.6 vs. 13.7). Adjourned cases more typically involved a family situation where the youth was living with the two natural parents, while formal cases were the least likely to show this pattern. Diversion program cases usually fit the characteristics of the adjourned cases, as indicated by the discriminant analysis classifications. In Table 5, the ADP cases are compared with the others on the important distinguishing variables. A few discrepancies appear that prevent all of the ADP cases from being classified as similar to the typical adjourned case. In brief, ADP cases seem similar to adjourned cases except that they have the most negative parent's comments, are slightly more likely to have had a prior court record, are more similar to formal cases in their living situation, and occupy a middle ground in terms of age. Method 3 In the analysis of referee questionnaire data, ADP cases once more were viewed as similar to the informal probation (or “adjourned”) category of cases. Referees recommended this as the recommended alternative disposition for 58% of their ADP referrals. In contrast to the other methods, a somewhat higher percentage, 33%, were found by the referees to be fit for formal handling. Only about 8% were recom— mended for dismissal if they were not accepted into the diversion program. This percentage was much lower than the 75 17% classified as similar to dismissed cases by the discri- minant analysis and the 37% dismissed by the intake referees after becoming ADP control group cases. The Referee Referral Form questionnaire data is sub- ject to at least one threat to validity, however. The fact that these forms are reviewed by diversion project staff may produce some demand characteristics in the referees' responses. The referees may answer the questionnaire with responses they think would be desired or expected by the ADP staff. For example, knowing that the ADP staff did not want referrals of cases who would otherwise be immediately released by the court, the referees would be unlikely to recommend "dismissal” as their preferred alternative dis- position, even if that was actually their preference. When compared both generally and for individual cases, the ques- tionnaire data did produce more "serious" dispositional classifications of ADP cases than the other methods. Another problem with the Referral Form data is the use of the "infor- mal probation" category, which was found to take the form of "adjourned for 30 days” in actual practice. It is not clear whether the referees had something different in mind for ADP cases than the procedures usually followed for a 30-day adjournment. The referees may have planned a more intensive investigation or supervision than normally used. Method 4 If the ADP had diverted only cases which otherwise would have received formal handling, the overall rate of cases 76 handled formally by the court would have been expected to have decreased after the initiation of the project. As indicated by the other analysis methods, however, the per- centage of "formal" cases in the program was probably no more than 25—33%. With these results, there should not be much of an impact on the overall system processing rates of the court. The pre-post change in percentage of cases formally processed was indeed minimal. The court population rates of formally handled cases dropped from 57% to 53%, a non- significant difference. Analysis of the random sample of archival court data produced the same finding, with a 56% to 54% pre-post change. Controlling for the sharp decline in status offenders in the post-ADP period, the analysis showed that the percentage of non-status cases which were formally processed actually increased from 45% to 50%. This is a non-significant difference, however. The time-series analysis also revealed no significant patterns of differences between the pre/post ADP time periods. The impact on the overall court figures would have most likely not been large even if all of the ADP cases were drawn from those in jeopardy of formal processing. Total ADP cases for the time period under study were 268, which was only about 10% of the court's total caseload at intake. Given the findings that only about a quarter of these were "for- mals," the lack of impact on the court's formal processing rates is not surprising. 77 Status Offender Policy Change A bigger impact on court dispositions was produced by the change in policy towards status offenders. The court's decision to not accept petitions on most status offenders dropped their percentage of the caseload from about 40% to 14% (Table 10). As explained eariler, this affected the make—up of ADP referrals. Status offenders made up 32% of ADP cases in the first two project groups, but only 10% in the next five groups. Since most status offenders were typically sent to a formal hearing, the first two groups of ADP cases have a higher percentage of "formal”—type cases than later years (Table 7). The drop in status offenders lowered the court's total intake caseload in 1977 and 1978, although it rose again in 1979. The number of non-status cases in this period which were handled formally increased somewhat, which could be construed as a response by the court to try to maintain formal probation caseload levels. However, this interpre- tation may be overstated, given the fact that the increase in non-status cases handled formally was a non-significant 5%, while the total intake caseload fell by 24% after 1976. The slight change in the handling of non-status cases clearly was not enough to make up for the drop in total caseload. Time Trends in ADP "Formal" Cases While the drop in ADP cases classified as "formal" by the discriminant analysis can be partially explained by the status offender policy shift, it is not clear why this 78 category increased over time in control group dispositions and referees' recommendations (Figure 1). For control group dispositions, this increase may be due to random events. As shown in Table 2, the numbers involved are very small. While the percentage of cases sent to a formal hearing increased from 20% to 43% over the five—group period of ADP referrals, the number of cases increased only from two to six. It may be that the intake referees were responding to feedback from ADP staff that more referrals should come from "formal" cases. This interpretation is fairly speculative, however. For the referee questionnaire data, the percentage of cases which had formal handling recommended as the preferred alternative disposition rose from 20% in ADP Year III to 50% in ADP Year VII. This change does not seem to involve the status offender issue, as the data does not show differential responses for status offenders on the Referral Form. It may be that the social desirability response set mentioned earlier became stronger over time. That is, intake referees may have wanted to appear to be complying with ADP staff expectations by more frequently checking the "formal hand- ling” category. It should also be noted that in ADP Years VI and VII there were no control group cases. In other words, the referee could say that the preferred alternative disposition for the case was formal handling, without having to worry about the possibility of having to follow through on the statement. In these two groups, the percentages checked as ”formal” were 55% and 50%, in contrast to the 79 20-39% range for years in which there was a control group. Convergence of Results Across Methods As noted earlier, the methods produced very consistent findings. Classification of diversion project cases into dispositional types yielded very similar proportions under methods 1, 2, and 3. Additionally, the lack of impact indi- cated in method 4 followed logically from the findings of the other methods. Classification of individual cases dif- fered somewhat between methods, yet the rates of agreement were higher than of disagreement. The trend for disagree- ments was that of more ”serious" dispositions recommended by referees, followed by discriminant analysis classifi- cations, and with the least serious dispositions experienced by control group cases. The results were so similar, how- ever, that the conclusion must be reached that there was great convergence across method. In reporting the results of each method, possible validity problems have been mentioned which may have some- what distorted a few of the findings. The similarity of the findings across the four methods should produce more con- fidence in the general results. If one of the methods were to be chosen as the most valid, it would seem to be the discriminant analysis classification of cases (method 2). The data used in this method were collected unobtrusively, producing no interaction with the actual referral behavior of the intake referees. This method had the soundest em- pirical support, as the classifications were based on 80 functions derived from cases with known dispositions. The three types of dispositional category groupings were statis- tically proven to be distinguishable, and the error rate of classification prediction was relatively small. In contrast, the analysis of control group dispositions may have been biased by the effect on referees of knowing that the case had been "rejected" by the project. In addition, the number of control group cases was not large enough to insure good generalizability of the results. In method 3, the analysis of referee questionnaire data, the findings were subject to invalidity produced by the possible social desirability response bias of the referees. Also, the absence of a control group in two of the ADP groups for which the form was used could have affected the responses. In method 4, the level of analysis was perhaps too gross to reveal specifics about the kinds of cases referred to the diversion project. While it was clear that the project did not produce a large change in system processing rates, the precise reasons for the lack of impact could not be ascer- tained from this method. Conclusions about Net-widening and the ADP Were the goals of diversion carried out during the first three years of the Adolescent Diversion Project? Did the project have the effect of widening the net of social con- trol by involving youths who would not otherwise have pene- trated the court system? Using the results of the discri- minant analysis (method 2) as a guide, it can be concluded 81 that the project did not divert substantial numbers of cases which were in jeopardy of formal handling and therefore pos- sible formal probation or institutionalization. Only about one-fourth of the ADP cases would have otherwise been for- mally processed. In addition, it was clear from method 4 that the project did not have an impact on the proportion of cases handled formally by the court. The ideal diversion program would produce a reduction in the court's formal case- load (Nejelski, 1976). In these terms, there is support for the notion that the diversion program led to ”net-widening.” Cases which would have received little or no court supervision and treatment were involved in diversion program services. On the other hand, only about 17% of these cases would have been immediately dismissed by the court. These represent clear instances of "net-widening." Most of the diversion referrals (57%) were "diverted” from the state of limbo known as the 30-day adjournment. This can sometimes involve supervision and monitoring of behavior by an intake referee or a student intern. It can sometimes involve some rules of informal probation that the youth must follow. It can also involve merely a rescheduling of the intake hearing with no contact with court officials in the interim. Immediate referral to the diversion program probably lessened the amount of court contact that these cases would have received had they been on informal probation or 30-day adjournment. The Adolescent Diversion Project was certainly a more 82 intrusive experience, however, because of the one-to-one contact with a volunteer 6-8 hours per week for 18 weeks. It may be that the types of cases given 30-day adjourn- ments or informal probation are cases in which the court does not quite know how to handle. The youth's problems seem too serious to dismiss the case, yet not serious enough to justify formal handling. Intake referees attempt to pro- vide informal "services" or monitoring to fill the gap between outright dismissal and possible formal probation. The option of diversion program referral thus appears desir- able for these cases since it provides services to the youth without engaging the resources of the court probation staff. This may explain why so many of the ADP cases fit the char- acteristics of "adjourned” cases. Thus, the ADP did not achieve the objective of diverting mostly cases in jeopardy of formal court processing or the goal of reducing the court's formal probation caseload. 0n the other hand, all but 17% of the project's cases would have received some further court involvement had they not been diverted. The evaluation of the project's success is therefore dependent on the degree of success expected, as well as on one's definition of diversion. If diversion means alternative service provision away from the court for young people in need of help, then perhaps the ADP fulfilled this model of intervention. But if diversion means removing from court jurisdiction those who are in distinct jeopardy of traditional formal court handling, then the ADP cannot be 83 said to have completely fulfilled the terms of this model. Comparison of Results with Other Studies The findings of this study are fairly similar to other evaluations of diversion programs. Most of the diversion literature involves police-based diversion programs, however, which are not strictly comparable to the court-based program examined here. Cases in this study represent deeper system penetration than most cases involved in police diversion programs. Cases which would have been dismissed at intake by the court in this study do not count as truly "diverted" cases, whereas cases in a police diversion program which would otherwise have been referred to the court and then dismissed at court intake do count among the truly "diverted." This fact should be kept in mind while comparing the findings. Only one study reviewed in Chapter 1 reported the dis- positions of control group cases. In that study, involving a police diversion program, nearly half of the cases were merely warned and released. For ADP control group cases, 37% were warned and released (dismissed). The preliminary results of the Alternative Routes project in California showed that only about 37% of the cases were in jeopardy of going on to court involvement by police referral. In a study cited by Pabon, about 24% of cases matched with the diversion program cases were released, while about 37% received informal probation. Another 39% were handled for- mally by the court. The California Youth Authority evaluation of nine diversion programs concluded that 51% of the cases 84 were diverted from some degree of further justice system penetration. However, only about 13% were likely to have received formal court handling, and another 7% would have gotten informal probation. These results are confounded by type of program, as police and court diversion programs were not analyzed separately, as noted in Chapter 1. Two studies of system processing rates concluded, like the present study, that the diversion program did not reduce the percentage of cases handled formally (Blomberg, 1977; Sarri & Bradley, 1980). There were no studies which used a referee questionnaire similar to the one used here, so no comparison data is available for this method. In terms of client characteristics, ADP cases appear to be less serious than the "system-insertable" standard com— posite reported by Klein (1979). ADP cases were younger (average age = 14.2), had somewhat less serious offenses, and were rarely likely to have had prior court petitions (2.2% had prior petitions). However, the types of cases receiving formal handling in the court under study were less serious than Klein's national composite. Fewer were male, more were younger, and more had less serious offenses. Thus, diverted cases in the ADP should not be expected to look quite the same as "system-insertable" cases in other juris- dictions. In general, the ADP evaluation yields results consistent with the diversion literature. If anything, the ADP cases had a higher percentage of the "truly diverted" than other 85 programs. Yet it does appear to be true that diversion pro- grams, including the ADP, involve youths in their programs who would not otherwise have been in jeopardy of further justice system involvement. The inclusion of these cases produces a widening net of social control. By failing to reach their targeted group of cases, diversion programs fail to reduce the number of young people subject to formal court "treatment" or supervision. Explanations for Net-widening How do diversion programs end up serving the wrong set of clients? The reasons seem to be tied closely to the or— ganizational goals and resources of the court system. In the Adolescent Diversion Project, and in other programs, a large percentage of the cases appear similar to those which would otherwise have been adjourned for 30 days or put on informal probation. As mentioned earlier, this type of case usually receives only minimal "treatment” by the court. They are perceived by court personnel as needing some sort of services, yet do not warrant formal probation services. Diversion programs with services provide a ready alternative for these cases. Services are provided, but with little court involvement. The court then has an outlet for these types of cases, and fills its quotas of diversion referrals with them, leaving less room for the "formal" types of cases. The court feels it has helped the youths by providing ser- vices, and can justify the referrals by the prevention philosophy of service provision. 86 Under this philOSOphy, it is better to be "safe than sorry”——not unlike the situations in which criminal justice officials are involved in predictions of dangerousness. There is less risk involved in predicting a "false positive” ——predicting someone to be violent who turns out not to be—— than in predicting a "false negative," where a violent per- son is sent free (Monahan, 1981). Thus, the relative levels of risk in prediction contribute to the practice of providing services or treatment in cases where there is some uncer- tainty. By ”widening the net” of service over more people, the agency can maximize its chances of "catching" the "true positives"——the criminals who will go on to commit violent acts. Hence, the court tries to expand the number of young- sters receiving some sort of services rather than reduce it. Diverting a youth who would otherwise receive formal pro- bation services could thus be construed as a wasted oppor- tunity to provide diversion program services to a youth who would otherwise have to be dismissed. The "penchant for prevention" philosophy alluded to in Chapter 1 can therefore contribute significantly to net-widening. Another explanation is related to this, but centers more directly on the resources available to the court. In this sense, the court has a usually fixed amount of resources that it can devote to service—provision or "treatment." Diversion programs often have been initiated with external funds and can thus be considered as supplemental to the 87 regular court programs. Court officials thus may view diversion programs as opportunities for additional service provision, rather than as alternatives to their traditional services. The motivation for expanding court service pro- vision to more youth through referrals to diversion pro- grams comes from the court's perception that there are more youth needing services that it can typically provide. The court's "penchant for self-perpetuation," typical of any bureaucratic organization, prevents it from dismantling its formal probation caseload by diverting these cases. Instead, the court maintains the same level of formally-handled cases, and expands the number of cases receiving services by refer— ring less-serious cases to diversion. Prevention of Net-widening Currently, diversion programs depend on the cooperation of court officials to receive referrals. The control over the types of cases to be referred rests almost totally in the hands of court officials. Diversion advocates wishing to prevent net-widening must either gain more control over the referral process or persuade court officials to refer the "right" kind of cases. The organizational forces in the court operating to produce net-widening appear too strong to be overcome by mere persuasion, however. Diversion advocates need to make administrative agreements with court officials that are directed by a higher agency or legislature, and which give less discretionary power to court referees. 88 These agreements should provide strict referral guide- lines based on empirically—derived profiles of the kinds of cases to be eligible for diversion referral. In brief, the profiles would be developed from an archival study of court decision-making, focusing on the characteristics of cases sent on to formal handling. Specific ranges could be given for key variables predicting formal handling. For example, the guidelines might specify that all referred individuals must have had at least one prior court petition, and must be 15 or over. Exceptions could be spelled out, such as first time offenders who have more serious crimes or more serious family problems. At the time of each referral, the diversion program staff person would check the case with the guidelines and determine the appropriateness of the referral. Continual monitoring would thus be built into the referral process, and diversion program staff would be free to reject a case if it was not appropriate. The initial agreement would of course have to include the approximate number of referrals to be made each year to insure receiving enough cases. Another method of changing the way the court uses diver— sion programs would be to set up diversion programs without services. The court would then not be able to View the "program" in terms of supplemental service provision. The rarity of this kind of "true” diversion is testimony to the truth of the proposition that courts wish to employ diver- sion program services for their own purposes. 89 The Future of Diversion Can diversion programs succeed in reaching their tar— geted group of clients? Are diversion programs as now operated worthwhile? The future success of diversion would seem to depend on radical changes in the way diversion pro- grams are established. Essentially, diversion advocates must overcome a tradition of wishful thinking and naiveté that has led to simplistic approaches to juvenile justice reform. Despite high—level policy support and federal funding, the fate of diversion programming has been left in the hands of hundreds of local—level police and court officials. Program development has neglected to consider the private agendas and organizational allegiances dic— tating much of these officials' behavior. Lacking the clout to gain control over the referral process, diversion planners have had to sit back and watch as local law enforce- ment agencies manipulate the programs to meet their own needs and goals. Without some amount of leverage over the actions of court or police officials, diversion cannot be expected to succeed. Why expect law enforcement and justice system workers to cooperate fully with a program that is basically antagonistic to their traditional operations? Diversion programs were developed from the rationale that traditional court and police handling of delinquents was ineffective and harmful, and should be reduced or even destroyed. The stakes are too high to expect acquiescence to the reformist goals 90 of diversion——jobs, funding, status, and organizational sur- vival lie in the balance. Diversion programs as now operated are of ambiguous value. As seen in this study and others, the risk of widening the net of social control over an expanded range of youths is high. The benefits to these youths of diversion services may not outweigh the added risk of reinvolvement in the justice system. This question brings other evaluative criteria into the examination of diversion programs. For example, does the program tend to bring the youths into more contact with law enforcement agents? Does the program "encapsulate" the youths into an alternate service system—— trading the "delinquency" label for the "socially handi- capped” label? Does the program provide skills and oppor- tunities for the youths to help avoid future legal trouble? Some diversion programs, such as the Adolescent Diversion Project, are specifically geared towards avoiding justice system involvement and providing empowerment skills to the youths. In these terms, the program could not be said to be harmful to its clients, and in fact may be quite bene— ficial. However, by allowing referral of cases not targeted for diversion, the program may still fail to help the very people for which it was created. In addition, it will fail to produce radical change in the court system by failing to remove any substantial number of cases from traditional formal court handling. 91 At this point, the future of diversion looks bleak. External funding, traditionally from the federal government, is drying up, leaving diversion programs even more at the mercy of local justice system agencies. Those agencies which do decide to pick up the tab for continuation of diver- sion programming will likely demand even more control over the programs. This would likely reach the point where "diversion" programs become just another component of traditional court services. Wheel I | '. l The need for ”true” diversion still exists. The rationale for the concept still has solid supporting evidencei showing that traditional court handling of delinquents is f I not effective. And it can still be said that diversion has i not yet been actually implemented in practice (Klein, 1979). 3 Yet these facts are inconsequential if control over the diversion referral process remains in the hands of justice system officials. Diversion programs lacking this control are doomed to repeat the mistakes of the past. Future Research Given the importance of organizational variables to court decision—making and diversion referrals, they warrant close investigation. Future studies should collect and examine data on agency budgets, probation caseloads, number of staff, number of spots available in detention facilities, availability of institutional placements, etc. These vari- ables could be correlated with general court processing rates and types of diversion referrals. 92 Future research should also test the success of alter- native diversion program procedures. Programs with typical referral procedures could be compared experimentally with programs that used the empirically-developed guideline system of referrals described above in the section on pre- vention of net-widening. Also, diversion with services could be compared with diversion without services, if the necessary administrative agreements could be established. One of the major purposes of this study has been to develop and employ better evaluative strategies for testing the types of cases referred to diversion programs. Future research should proceed from this development to insure adequate program evaluation. The multi-method evaluation system utilized here has proven useful and is recommended for future studies. Explicit evaluation of the organiza- tional and system—level impacts of diversion will help us to initiate social interventions that achieve their goals. 93 REFERENCE NOTES Seidman, E. The route from the successful experiment to policy formation: Falling rocks, bumps and dangerous curves. In R. Roesch (Chair), Panel of Evaluation and Policy. American Society of Criminology, San Francisco, Nov. 1980. Wresinski, M. D., Saul, J. A., Bybee, D. I., Davidson, W. S., & Koch, J. R. Prevention of juvenile delin— gpency: Evaluating the systems impact of diversion. Paper presented at the 88th Annual Convention of the American Psychological Association, Montreal, Canada, 1980. Davidson, W. S., Koch, J. R., & Snellman, L. Current status of diversion research: Implications for polipy and programming. Paper presented at the meeting of the American Society of Criminology, San Francisco, Nov. 1980. Berg, D., & Shichor, D. Methodological and theoretical issues in juvenile diversion: Implications for evalu- ation. Paper presented at the National Conference on Criminal Justice Evaluation, Washington, D.C., Feb. 1977. L I ST OF REFERENCES 94 LIST or REFERENCES Berger, D., Lipsey, M. W., Dennison, L. B., & Lange, J. M. The effectiveness of the Sheriff's Department juvenile diversion projects in Southeast Los Angeles Coun_y. Claremont, CA: Claremont Graduate School, 1977. Binder, A. Diversion and the justice system: Evaluating the results. In A. W. Cohn (Ed.), Criminal justice plan- ningyand development. Beverly Hills: Sage Publications, 1977. Blomberg, T. Diversion and accelerated social control. Journal of Criminal Law and Criminology, 1977, 68 (2), 274-282. Blomberg, T. Widening the net: An anomaly in the evaluation of diversion programs. In M. Klein & K. Teilmann (Eds.), Handbook of criminal justice evaluation. Beverly Hills: Sage Publications, 1980, 572-592. Bohnstedt, M. Answers to three questions about juvenile diversion. Journal of Research in Crime and Delinquency, 1978, lg (1), 109—123. Bohnstedt, M., et al. The evaluation of juvenile diversion programs. Sacramento: California Youth Authority, 1975. Box, G. E. P., & Jenkins, G. M. Time-series analysis: Fore- castipg and control. San Francisco: Holden-Day, 1976. Bullington, B., Sprowls, J., Katkin, D., & Phillips, M. A critique of diversionary juvenile justice. Crime and Delinquency, 1978, 23 (1), 59-71. Caplan, N., & Nelson, S. D. On being useful: The nature and consequences of psychological research on social problems. American Psychologist, 1973, 28, 199-211. Cressey, D. R., & McDermott, R. A. Diversion from the juvenile justice system. Ann Arbor, MI: National Assessment of Juvenile Corrections, University of Michigan, 1973. 95 Davidson, W. S., Seidman, E., Rappaport, J., Berck, P., Rapp, N., Rhodes, W., & Herring, J. Diversion program for juvenile offenders. Social Work Research and Abstracts, 1977, 13, 313—338. Dinitz, S. Emerging issues in delinquency. In The threat of crime in AmericaJ the 1967—68 E. Paul dePont lecture on crime, delinquency and corrections. Newark: University of Delaware, 1969. Fox, S. J. Juvenile justice reform: An historical perspec- tive. Stanford Law Review, 1970, 22, 1187. Gaylin, W., Glasser, I., Marcus, S., & Rothman, D. Doing good: The limits of benevolence. New York: Pantheon, 1978. Gemignani. Diversion of juvenile offenders from the juvenile justice system, in Criminal Justice Monogpaph: New approaches to diversion and treatment of juvenile offen- ders 39, 1973, 8. Gibbons, D. C., & Blake, G. F. Evaluating the impact of juvenile diversion programs. Crime and Delinquency, 1976, 22 (4), 411-420. Glass, G. V., Willson, V. L., & Gottman, J. M. Design and analysis of time-series experiments. Boulder: Colorado Associated University Press, 1975. Gold, M., & Williams, J. R. The effects of getting caught: Apprehension of the juvenile offender as a cause of sub— sequent delinquencies. Prospectus: A Journal of Law Reform, 1969, 2 (1), 1-12. Griffin, B. S., & Griffin, C. T. Juvenile delinquency in perspective. New York: Harper & Row, 1978. Howlett, F. W. Is the YSB all it's cracked up to be? Crime and Delinquency, 1973, 22 (4), 485-492. Humphreys, L., & Carrier, J. M., Second annual evaluation report: Pomona Valley Juvenile Diversion Project. Claremont, CA: Pitzer College, 1976. In re Gault, 387 U.S. 1 (1967). In re Winship, 397 U.S. 358 (1970). James, H. Children in trouble: ‘A national scandal. New York: Cristian Science Monitor Publishers, 1969. 96 Jesness, C. F. Comparative effect of two institutional treatment programs for delinquents. Child Care Quar- terly, 1975, l (2), 72-79. Katkin, D., Hyman, D., & Kramer, J. Delinquency and the juvenile justice system. North Scituate, MA: Duxbury Press, 1976. Klapmuts, N. Diversion from the justice system. Crime and Delinquency, 1974, 2, 108-131. Klein, M. Deinstitutionalization and diversion of juvenile offenders: A litany of impediments. In N. Morris & M. Tonry (Eds.), Crime and justice: An annual review of research, Vol. 1. Chicago: The University of Chicago Press, 1979. Klein, M., Teilmann, K. S., Styles, J. A., Lincoln, S. B., & Labin-Rosensweig, S. The explosion in police diver- sion programs: Evaluating the structural dimensions of a social fad. In M. Klein (Ed.), The juvenile justice system, Beverly Hills: Sage Publications, 1976, 101-119. Klein, M., Teilmann, K. S., Lincoln, S. B., & Labin, S. Diversion as operationalization of labeling theopy. Final report on Project #MH26147 to the Center for Studies in Crime and Delinquency, National Institute of Mental Health. Los Angeles: Social Science Research Institute, Univer- sity of Southern California, 1979. McAleenan, M. The West San Gabriel Valley Juvenile Diver- sion Project: Annual Report, Second Year. Los Angeles: Occidental College, 1976. Monahan, J. Predicting violent behavior: An assessment of clinical techniques. Beverly Hills: Sage Publications, 1981. National Advisory Commission on Criminal Justice Standards and Goals. Diversion. In its Police. Washington, D.C., U.S. Government Printing Office, 1973. Nejelski, P. Diversion of juvenile offenders in the criminal justice system. In Criminal Justice Monograph: New approachesypo diversion and treatment of juVenile offen— ders 39, 1973, 83. Nejelski, P. Diversion: The promise and the danger. Crime and Delinquengy, 1976, 22 (4), 393-410. Pabon, E. Changes in juvenile justice: Evolution or reform. Social Work, 1978, 22 (6), 492-497. 97 Palmer, T., & Lewis, R. An evaluation of juvenile diversion. Cambridge, MA: Oelgeschlager, Gunn, & Hain, Publishers, 1980. Platt, A. The triumph of benevolence: The origins of the juvenile justice system in the United States. In R. Quinney (Ed.), Criminal justice in America. Boston: Little, Brown, 1974. Rappaport, J. In praise of paradox: A social policy of empowerment over prevention. American Journal of Community Psychology, 1981, 2 (1). Rappaport, J., Seidman, E., & Davidson, W. S. Demonstration research and manifest versus true adoption: The natural history of a research project to divert adolescents from the legal system. In R. F. Munoz, L. R. Snowden & J. G. Kelly (Eds.), Social and psychological research in com- munity settings: Designing and conducting programs for social and personal well-being. San Francisco: Jessey- Bass, 1979, 101-144. Richette, L. A. The throwaway children. Philadelphia: Lippincott, 1969. Rutherford, A., & McDermott, R. Juvenile diversion. Wash- ington, D.C.: National Institute of Law Enforcement and Criminal Justice, LEAA, 1976. Sarri, R., & Bradley, P. W. Juvenile aid panels: An alter- native to juvenile court processing in South Australia. Crime and Delinquency, 1980, 2g (1), 42-62. Scarpitti, F. R., & Stephenson, R. N. A study of probation effectiveness. Journal of Criminal Law, Criminology, and Police Science, 1968, 22, 361—369. U.S. President's Commission on Law Enforcement and the Admin- istration of Justice. The challenges of crime in a free society. Washington, D.C.: U.S. Government Printing Office, 1967. Vorenberg, E. W., & Vorenberg, J. Early diversion from the criminal justice system: Practice in search of a theory. In L. E. Ohlin (Ed.), Prisioners in America. Englewood Cliffs, N.J.: Prentice Hall, 1975, 151—183. Wolfgang, M., Figlio, R., & Sellin, T. Delinquency in a birth cohort. Chicago: University of Chicago Press, 1972. APPEND IX 98 APPENDIX 1 Referee Referral Form Michigan State University Adolescent Diversion Project Youth's Name Referee Length of Formal Court Involvement Date Please complete the following ratings on this youth. 1. Compared to other youth you have referred to the Project, how serious is his/her history of delinquency? (Circle appropriate rating.) / / / / / / / 1 2 3 4 5 6 7 Extremely Average Extremely Minor Serious 2. Compared to other youth you have referred to the Project, how serious is the current allggation? (Circle appro- priate rating.) / / / / / / / 1 2 3 4 5 6 7 Extremely Average Extremely Minor Serious 3. If this youth does not get into the Project, what alter- native will you recommend? (Check alternative recom- mendation.) Referral to: Dismissal Equal Ground Informal/Consent Probation CHM Formal Hearing - 99 4. How would you rate this youth's likelihood of success in the Project? (Circle appropriate rating.) / / / / / / / 1 2 3 4 5 6 7 Extremely Average Extremely Doubtful Likely 5. How would you rate this youth's chances of being formally petitioned to court again within the next year? (Circle appropriate rating.) / / / / / / / 1 2 3 4 5 6 7 Extremely Average Extremely Likely Unlikely 6. Compared to other youth you've referred to this Project, how easy will he/she be to get along with? (Circle appropriate rating.) / / / / / / / 1 2 3 4 5 6 7 Extremely Average Extremely Difficult Easy 7. Compared to other families you've referred to this Project, how easy will they be to get along with? (Circle appropriate rating.) / / / / / / / 1 2 3 4 5 6 7 Extremely Average Extremely Difficult Easy 8. Compared to other youth you've referred to this Project, how suitable is this Project for this youth's needs? (Circle appropriate rating.) / / / / / / / 1 2 3 4 5 6 7 Extremely Average Extremely Unsuitable Suitable 9. Taking this youth and his/her situation into consider- ation, what would you consider a successful outcome for this case? (+ = relevant, - = irrelevant, please mark each.) Neatness around home Improved classroom improved behavior Personal appearance Completing homework improved improved Talking to parents improved Getting along with parents improved Doing household chores improved Spending more time at home Parents lessen restrictions Youth's moodiness improved Arguing with parents improved Improved school attendance Improved academic performance 100 Improved attitude toward school Participation in extra- curricular activities increased Improved relationship to teachers Constructive use of free time increased Formed new relationships Hanging around with better crowd Getting a job No further police contacts No further court petitions Alleged Offense