‘“ llllllll‘lhlllllllllllllllll llll‘llllllllllllllll . sagzr'mm 3 1293 10687 009 This is to certify that the thesis entitled AN EXAMINATION OF THE PROCESSES OF CHANGE: EXPLORING THE SUCCESS OF AN EFFECTIVE JUVENILE DIVERSION PROJECT presented by JULIA ANNE PARISIAN has been accepted towards fulfillment of the requirements for Major professor Date Jul! 29.1982 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution MSU LIBRARIES RETURNING MATERIALS: Place in book drop to remove this checkout from your record. FINES will be charged if book is returned after the date stamped below. Tyr 7w— AN EXAMINATION OF THE PROCESSES OF CHANGE: EXPLORING THE SUCCESS OF AN EFFECTIVE JUVENILE DIVERSION PROJECT By Julia Anne Parisian A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Psychology 1982 ABSTRACT AN EXAMINATION OF THE PROCESSES OF CHANGE: EXPLORING THE SUCCESS OF AN EFFECTIVE JUVENILE DIVERSION PROJECT By Julia Anne Parisian This study explored the relationships between the processes of change and delinquency outcomes in the context of a diversion program for juvenile offenders. One hundred and seven program participants were the subjects. Delinquency outcomes were based upon two distinct methods: Official data and self-report delinquency (SRD) rates. The process measures (predictor variables) included indices of intervention activities and general life domain conditions in the youths' lives. Separate discriminant analyses were performed on the two types of outcome groups. The results demon- strated SRD and offical delinquency to be distinct phenomena since the different discriminant functions included very different predictor vari- ables, depending on which measure defined the outcome groups. Also, the predictor variables provided a much poorer model for discriminating the official outcome groups than for the SRD groups. In addition, the results provided support for an "Action" intervention model with these diverted youths which included behavioral contracting and youth advocacy components. To Frances Kathryn Parisian and the memory of George Thomas Parisian ii ACKNOWLEDGEMENTS This effort would not have been possible without the support and encouragement of many people. First among them is Dr. William Davidson whose work has provided me (and many others) with the opportunity to be- come involved with the Adolescent Diversion Project in a way that has led to tremendous growth and learning. His support and encouragement through these years and this thesis have had a profound impact upon my education, for which I am very grateful. My thanks go also to Dr. Dozier w. Thornton who has provided me with guidance and inspiration through the "other half" of my education. He's always demonstrated more faith in my ability to complete this task than I had in myself. Dr. Robin Redner has been a good friend and role model throughout this endeavor, as well, and I appreciate all of the excellent suggestions and encouragement she has offered me. Throughout these past few months John Jeppesen has been putting up with my long hours and many moods, being an editor, sounding board and, most of all, a friend. His support has been my mainstay. And for their constant inspiration, friendship and encouragement, my thanks go also to Richard Alpert, Lynn Snellman and Patti Witte. Finally, this effort would not have been possible without the combined talents and seemingly endless energies of the ADP staff (past and present) through these years. Special thanks to Becky Mulholland, Craig Blakely, Debbie Bybee, Jim Emshoff, Marty Kushler, Cheryl Saylor, Sharon Dynak, Tina Mitchell, Jeana Chodakowski, John Saul and Bette Sue Rittman. Because of them, my involvement with the ADP has been fun, too. iii TABLE OF CONTENTS Contents LIST OF TABLES . . . . . . . . . . ................. LIST OF FIGURES. . . . . ...... . . . . . ..... . . . . . . INTRODUCTION . . ....... . . . . . . . . ..... Statement of Problem. . . . . . .......... Critique of the Juvenile Justice System . . . . . . . . . . Diversion ............ . . . . . . . . . ...... The Adolescent Diversion Project ...... . . . . The Utilization of Nonprofessionals as Change Agents . . Behavioral Contracting . . . . . . . . . . . . ....... Child Advocacy ....... . . . ...... Discussion . ........ . . . . . . . . ADP--The Outcome Research .......... . . . . . . . . Background ......... . . . . . . . ........ . . Outcome Results ............ . . . . . . ...... The Need to Examine Process Variables . . . . . . . . . . . . . . A Proposed "Hybrid Model" for the Task. . .......... ADP-~The Process Research. . .......... Review of Process Results Gained from Experimental Manipula- tion ....................... Review of Process Results Gained from Exploratory Components : Illinois Setting - Process Results. . ...... . Michigan Setting - The Intervention Interview Results . Michigan Setting - The Life Domain Interview Results. The Current Research ........ METHOD ................................ Subjects ........................ Youth Participants .............. . . ...... Student Participants . . . .................. Trainers and Supervisors ................... Procedures. . . . ........................ Recruitment of Students. . . . . ............... Referral of Youth ......... . . . ......... Training and Supervision ................... Intervention ................... Measures ............................. Archival Police and Court Data . . . . . ........ Process Measures . . . . ................... Self-Report Delinquency Measure . . .......... Life Domain Survey ..................... Intervention Survey . . . . . ............... iv Contents Page Data Modification Procedures for Process Measures ...... 78 Principal Components Analysis of Life Domain Scales. . . . 87 Life Domain - School .................. 88 Life Domain - Positive Change ............. 88 Life Domain — Peer Involvement ............. 88 Life Domain - Parental Control ..... . . ...... 88 Life Domain - Employment Motivation . . . ....... 9l Life Domain - Positive Home Involvement . . . . . . 9l Principal Components Analysis of Intervention Scales . . . 9l Intervention - Family .............. . . . 9l Intervention — School . . ............ . . . 94 Intervention - Job Seeking and Advocacy . ..... . 94 Intervention - Amount of Time and Parental Involvement. 94 Intervention - Recreation ............... 94 Intervention - Contracting ............ . . . 94 Computed Factor Scores . . . ...... . . . . . 94 Data Modification Procedures for. Outcome Measures ...... 95 Archival Outcome Measures. . . .............. 95 SRD Outcome Measures .................. . 95 RESULTS ............................. . 97 Analyses ........................... 97 Discriminant Function Results ................ lOO Archival Outcome Groups ........ . . . . . . . . . lOO Analysis with All Predictor Variables . . . ...... 100 Analysis with Only Process Variables. . . . . . . . . . lO3 Self-Report Delinquency Outcome Groups .......... lO3 DISCUSSION ............................ ll2 Results - Official Recidivism Groups ............. ll2 Important Predictors .............. . . ll2 Evaluation of Predictive and Explanatory Capabilities. . . ll7 Results - SRD Groups ..................... ll8 Important Predictors ................... ll8 Life Domain - Parental Control Factor ........ . ll8 Life Domain - Peer Involvement Factor ......... ll9 Life Domain - Employment Motivation Factor ....... l20 Intervention - Job Seeking and Advocacy Factor ..... l2l Intervention - Amount of Time and Parental Involvement Factor ....................... l22 Intervention - Family Factor .............. l23 C0nclusi0ns--Findings Discriminating the SRD Groups. . . . l23 Evaluation of Predictive and Explanatory Capabilities. . . l26 Comparison of the Two Discriminant Functions ...... . . . 127 Conclusions ......................... l28 APPENDIX A - Self-Report Delinquency Measure ........... l3l APPENDIX B - Recruitment Letter .................. l34 APPENDIX C - Participation Agreement ............... 135 V Contents Page APPENDIX D - Life Domain Survey .................. l36 APPENDIX E - Intervention Survey .................. l52 REFERENCES ............ . . . . . ............ l64 vi Table l4 l5 l6 l7 l8 l9 20 LIST OF TABLES Court Recidivism . Police Recidivism. . . . . . . . . . . ....... Incarceration. Relationships of the Multiple Contingency Model. . Distribution of ADP Participants by Referral Offense - Adolescent Diversion Project -- Fall, l976 through Septem- ber, l980. . . . . . . . . . . . . . . . . . . . . Breakdown of ADP Y0uth--Student Pairs by Condition and Year. Life Domain Scales . . . . . . . . . . ........ Life Domain Scale Correlations . . Inter-Interviewer Agreement - Years 2, 3, 4, 5 Life Domain. Intervention Scales. . . . . . . . . . ........... Correlation of Intervention Scales . . . . ....... Inter-Interviewer Agreement - Years 2, 3, 4, 5 Intervention. Means, Standard Deviations and Ranges of Life Domain Pre- dictor Variables . . . .................. Means, Standard Deviations and Ranges of Intervention Pre- dictor Variables . . . . . . . .............. Intercorrelations of Life Domain and Intervention Predictor Variables. . . . . . ..... . . . ......... Results of Factor Analysis of Life Domain Predictor Variables ....................... Communalities of Life Domain Predictors - Rotated Principal Components Solution - 6 Factors. . ............ Results of Factor Analysis of Intervention Predictor Variables ......................... Communalities of Intervention Predictors - Rotated Principal Components Solution ~ 6 Factors ............ Pooled Within-Groups Correlation Matrix of Predictor Variables. . . . . . . . . . . . . . . . . ........ vii 58 59 75 76 77 79 BI 82 84 85 86 89 90 92 93 99 Table Page 2T lbchival Success and Failure Groups - Means of Predictor Variables . ....................... lOl 22 Discriminant Analysis Summary Table - Archival Success and Failure Groups. . . ................... lOZ 23 Capability of the Discriminant Function to Predict Archival Outcome Groups. . . . ................ . . lO4 24 Discriminant Analysis Summary Table - Archival Success and Failure Groups ............... . ....... l05 25 Self-Report Delinquency Groups - Means of Predictor Variables ........................ 106 26 Discriminant Analysis Summary Table - Self-Report Delin- quency Groups ................... . . . lO7 27 Group Centroids and Between-Group Discriminant Ability of the Functions - Self-Report Delinquency Groups . . . . 109 28 Capability of the Discriminant Function to Predict Self- Report Delinquency Outcome Groups ........... . Ill 29 Comparison of Discriminant Function for Archival Outcome Groups with Functions for Self-Report Delinquency Outcome Groups ...................... ll3 viii Figure l LIST OF FIGURES Diagram of Measurement Administration ...... ix CHAPTER I INTRODUCTION Statement of Problem The problem of delinquency has been the focus of continuing concern and study in the country over the past few decades. Throughout the six- ties and seventies the incidence of delinquency grew at an alarming rate. Between l970 and l979, the number of violent crimes committed by persons under the age of 18 increased by 4l.3% (Crime in the United States, l980). Although the national rate of delinquency has decreased slightly during the past couple of years, this has most likely reflected the age shifts in the general population more than any meaningful changes in the behavior patterns of young people. In addition, the F.B.I. reports that even though on the national level the number of arrests for people under eighteen decreased by 6% from T970 to l980, this figure increased by 6.6% in the "East North Central" region of the country which includes, Illinois, Michi- gan, Ohio and Wisconsin (Crime in the United States, l981). During T980, young people accounted for 20.9% of all the arrests made in the U.S.--and for 40.2% of all arrests for property crimes. It is estimated that be- tween l7% to 33% of all males are arrested at least once before their eighteenth birthdays (Carter & Klein, T976; Davidson, Snellman, & Koch, l980). And yet even these figures greatly underestimate the rate of delin- quent acts, since so few of them are ever detected or observed and there~ fore do not appear in official statistics. Gold and Williams obtained l 2 self-report delinquency data from a random national sample of 847 young people in l967 and report that 88% of the youth confessed one or more chargeable delinquent acts in the prior three years. Only 3% of these offenses resulted in detection and contact with the police (Gold 8 Wil- liams, l969; Williams & Gold, I978). Clearly, delinquency continues to be a pressing problem in need of effective solutions. Unfortunately, the solutions have been elusive. The present discussion will briefly review the effectiveness of some of the treatment and rehabilitation efforts made by the traditional juve- nile justice system. Some of the major criticisms of this system will then be presented and the resulting movement toward alternatives, specific- ally toward diversion will be discussed. A brief review of the trends in diversion will follow. Then one particular diversion program (the Adoles- cent Diversion Project) will be discussed at length. The specific compon- ents of its intervention package will be reviewed. These include the use of nonprofessional change agents, and the utilization of behavioral con- tracting and of child advocacy as specific intervention tools. The out- come results of this program will then be presented. The need in program evaluation to examine and understand process variables as they impact upon outcome variables will be discussed generally as well as specifically in relationship to the Adolescent Diversion Project (ADP). The research model executed to evaluate the ADP allowed tremendous advancement of the understanding of this relationship between process and outcome variables in regard to the workings of this program. This research model will be presented and the process-related findings to date will be reviewed. The current study attempted to further examine the relationship between the variables which measure the ongoing processes of change throughout the 3 intervention period and the outcome variables of court recidivism rates, police recidivism rates, and measures of self-report delinquency. The second chapter presents the details of the study. Critique of Juvenile Justice System As delinquency has proliferated during the past two decades, so have criticisms of the juvenile justice system which has assumed the major responsibility of rehabilitating juvenile offenders. A major and inci- sive criticism has been that the system simply has not worked. In regard to its stated goals of rehabilitation and treatment of juvenile offenders, it has failed miserably (Krisberg 8 Austin, 1978). Recidivism rates are high at every step into the system. Wolfgang, Figlio, and Sellin (1972) report that, at the level of police arrests, 54% of all youths arrested for the first time will be arrested again; and 65% of those arrested twice will be arrested for a third time. The juvenile court does no better. Among youth probationers, 50% will be re-petitioned to court and the estimates of recidivism rates for residential institutions for juvenile offenders are between 50% to 85% (Davidson, et al., 1980). Recent reviews of the treatment efforts made by the juvenile justice system have reached discouraging conclusions. Riedel and Thornberry (1978) reviewed several evaluation studies which reported on hundreds of "treatment" approaches to rehabilitating juvenile offenders--inc1uding probation with variations in programming, imprisonment with variations in length, residential treatment programs and psychotherapy. Their conclu- sion was that "there is no systematic evidence that rehabilitation efforts have had significant impact on recidivism (p. 429). Romig (1978) arrived at a similar conclusion. He reviewed 170 controlled studies conducted in the area of delinquency rehabilitation between 1920 and 1976. He found 4 that, with yegy few exceptions, the treatment approaches utilized most frequently by the juvenile justice system are completely ineffective in rehabilitating (i.e., reducing the recidivism rates of) juvenile offen- ders. He finds that casework, the most frequently provided “treatment” for probationers (defined as including diagnosis, recommendations and direct services or counseling), is ggt effective for rehabilitation (p. 7). He finds similarly discouraging results with educational programs (which, while they may lead to improved academic performance, do not lead to a decrease in recidivism), most traditional vocational training and work programs, group counseling, the majority of individual psychotherapy approaches, community based residential programs and institutionaliza- tion. In short, he found that very little of what the juvenile justice system does with young people works to reduce future delinquency. [A second major criticism of the juvenile justice system has been that it operates inconsistently and unfairly in its handling of juvenile offenders)(Krisberg 8 Austin, 1978; Lemert, 1976). The juvenile court, since its inception, has maintained relative procedural informality in the name 0f providing individualized "treatment" to children. Conse- quently, historically there have been few legal protections or safeguards for the rights of the children in the court proceedings (Krisberg & Aus- tin, 1978, p. 3). During the mid-sixties the Supreme Court responded to this situation by handing down two major rulings (Kent vs. U.S. and the case of Gault) which stated that the constitutional rights of juveniles should be protected by establishing more formal proceedings within the juvenile court. (Some improvements have been realized since then, but concern about the biases and irregularities of the juvenile court remains> Krisberg and Austin state such concerns concisely in The Children of Ishmael (1978): 5 Laws governing juvenile justice agencies give license for widespread discretion. Police, probation, and court deci- sions are often made without significant parental or community input. Practices and procedures are so flexible that they allow gross distortions of cherished precepts of justice, including due process and equal protection of law (p. 570). (One of the most disturbing findings which signals the unfair workings of justice at the juvenile court level is that poor and nonwhite children consistently receive harsher disposition decisions at the hands of police and court personnel than do their wealthier, white counterparts) (Krisberg 8 Austin, 1978; Gold, 1966). These concerns have led many to entertain the notion that the pro- cessing of juvenile offenders through the juvenile justice system results in more malignant destructive consequences for the youths than in any benefits (Gold 8 Williams, 1969; Krisberg 8 Austin, 1978; Lemert, 1976; Schur, 1973). Gold and Williams' interview data from the National Youth Services suggests, in fact, that the apprehension of a youth by the po- lice for delinquent behavior leads to more subsequent delinquency (as measured by self-report) than if, in the same case, the youth were not apprehended (1969). They concluded: It appears, unfortunately, that what legal authorities now com- monly do upon apprehending a juvenile for his delinquent be- havior is worse than not apprehending him at all (p. 3). Lemert suggests that the labeling or "stigmatization" process that occurs when a youth becomes involved in the juvenile justice system accounts for the observed increases in delinquency. He describes this as "a process which assigns marks of moral inferiority to deviants; more simply it is a form of degradation which transforms identities and status for the worse (1976, p. 133). This, he maintains, simply creates more problems for the youth and his/her family. 6 Diversion These concerns and criticisms of the traditional juvenile justice system resulted in the current trend of "diversion" programs that direct youths away from normal processing through the system (Davidson 8 Kushler, 1979; Krisberg 8 Austin, 1978). The concept of diversion has been with us for over a century and, in fact, is what led to the creation of a separate juvenile court system in the first place (Palmer 8 Lewis, 1980). The current trend toward diversion was spurred by the 1967 Pres- ident's Commission on Law Enforcement and the Administration of Justice. Their report criticized the juvenile justice system's handling of juve- nile offenders and called for the widespread creation of alternatives in the form of diversion programs, Youth Service Bureaus, community-based treatment programs and the utilization of volunteers throughout the sys- tem (Carter 8 Klein, 1976; Davidson, et al., 1980; Palmer 8 Lewis, 1980). The call for diversion programs met with a widespread and rapid response over the next decade. In 1980, in Michigan alone, over 200 pro- grams identified "diversion" as at least one of their main purposes (Davidson, et al., 1980). Diversion's many supporters suggested that formal processing through the juvenile justice system has such deleter- ious effects that redirecting youthful offenders to alternative programs (or to nothing at all) was bound to be more beneficial (Lemert, 1976; Rappaport, 1977; Schur, 1973). Unfortunately, systematic evaluation efforts did not keep pace with program implementation and only recently have attempts been made to sort out the issue of the "effectiveness" of diversion projects (Palmer 8 Lewis, 1980). The findings and conclusions have been mixed. One of the major problems is'a definitional one (Klein, 1979). Diversion is usually defined negatively-~that is, as anything which removes offenders from formal handling within the justice system. 7 This, of course,1eaves tremendous latitude in terms of what diversion programs g9, There is also a great deal of variability between programs in the specific characteristics of the target populations and in the point of referral or diversion from the system. As an example, one so- called ”diversion" program might provide a few hours of counseling to a potential runaway. While another "diversion" program might provide community placements for repeat offenders who would otherwise be institu- tionalized in correctional facilities. Obviously, such diversity among programs renders the task of judging the effectiveness of "diversion” an extremely complex one. The problems which result from this situation are apparent in the recent reviews of diversion programs which are dis- cussed below. Klein (1979), in reporting on the effectiveness of diversion pro- grams to reduce recidivism rates and delinquency rates (as measured by self-report), notes the prevasive absence of evaluation efforts and the poor quality of evaluation in those rare cases when it is attempted. He reviews 13 diversion programs and reports that only three of these cite unequivocal positive findings in regard to the reduction of recidivism and delinquency. He finds that two more report negative findings in regard to these criteria and the remaining eight cite "equivocal" find- ings (e.g., a reduction of recidivism rates but no corresponding self- reported delinquency changes). Unfortunately, Klein's presentation does not include information about who the "diverted" youths were or what these programs entailed in terms of intervention. He does state that "almost none” of them employed random assignment to a control group. In another review, Davidson, Snellman, 8 Koch (1980), critiqued the reports of 50 evaluations of diversion programs. Of these 50, they 8 found only five that employed random assignment control groups. They limited their discussion to these five programs which included (1) a family focused intervention with diverted status offenders (Baron, Feeney, 8 Thornton, 1973), (2) a combination vocational, educational and counseling intervention package for youthful offenders (Quay 8 Love, 1977), (3) a program which referred diverted youths to local counseling services (Klein, 1979), (4) a broad-based intervention package which included relationship skills, behavioral contracting and advocacy for youth diverted from the police (Davidson, Seidman, Rappaport, Berck, Rapp, Rhodes, 8 Herring, 1977), and, (5) a program combining parent skills training, various environmental interventions and the assignment of an adult "role model/friend" to youths diverted from the police (Binder 8 Newkirk, 1977). Three of these evaluations which utilized treatment-as- usual control groups found significantly lower recidivism rates for diverted youths than for the control groups (Baron, et al., 1973; Quay 8 Love, 1977; Klein, 1979). Of those that utilized an outright release control group, two found lower recidivism rates for the group which received services than for the control group (Davidson, et al., 1977; Binder 8 Newkirk, 1977). On the other hand, Klein reports the lowest recidivism rates for the outright release control group, although, as reported above, the diverted-to-treatment group fared better than the treatment-as-usual group. Davidson, et al., concluded: First of all, it is obvious that the great majority of evalua- tions of diversion are of questionable credibility due to an inadequate design and/or other methodological deficiencies. In addition, the wide variability in the types of youths served, program types examined, and effectiveness criteria used makes definitive conclusions extremely difficult. The second major conclusion is that diversion has shown sufficient conceptual and empirical promise to warrant continued efforts in this area (p. V-9). 9 Romig (1978) also conducted a review of the literature of the effec- tiveness of "diversion" programs. He reports on eight studies of pro- grams which involved a total of 1,095 youth. All of the projects included randomly assigned control groups-~although it is not clear from his discussion if these were treatment-as-usual control groups or outright release groups. He reports that the results of these eight studies were uniformly negative. Seven of them resulted in no differences in the recidivism rates of the experimental and control groups and one (an intensive family counseling intervention administered to youths diverted from the court) resulted in greater recidivism rates for the experimental group than for the controls. It should be noted, however, that from Romig's description of these programs, only four of the eight were even attempting to divert youths who would otherwise be involved in the juvenile justice system away from that system. The other four were clearly "delinquency prevention" programs for "pre-delinquents." That Romig includes these as "diversion" projects again points out the defi- nitional ambiguity that plagues the literature in this area and makes the task of evaluation a difficult one. In discussing his findings, Romig reports that the overwhelming majority of diversion programs utilize individual counseling as the pri- mary intervention tactic. The second most frequently used “treatment" is referral to other agencies. Since these methods of treatment are consistently found to be ineffective with delinquent populations, Romig's conclusion about diversion is that there does not appear to be anything inherently effective about diversion programs independent of their treat- ment approaches. Diversion programs which utilize ineffective interven- tion tactics will be ineffective. Romig suggests,l§Diversion projects should be developed around teaching delivery strategies that will give 10 the youths the skills they need to succeed in school, work, and home. Such approaches have a greater likelihood of success than what has been continually tried in the pastf}(p. 121). Perhaps the most extensive recent evaluation of diversion programs was a three year study of eleven community-based diversion projects which was conducted by Palmer and Lewis (1980). The projects they examined served youths who were referred from either police departments or juvenile courts. The authors compared the recidivimnrates of the project client groups to those of "comparison” groups, two of which were random control groups; the rest were ”matched” comparison groups. Palmer and Lewis found that, for the six month period following the referral arrest, "clients had signifi» cantly lower recidivism rates than comparisons" (p. 87). For all programs combined, 25.4% of the clients had experienced one or more subsequent arrests and 30.7% of the comparisons had done so. When the influence of a youth's prior record on program success was examined, Palmer and Lewis discovered that, for youths with no arrests prior to the referral arrest, diversion programs were not effective. (The authors suggest that this is due to the fact that this population is not at risk anyway-~that is, they are not very likely to recidivate in any case. And yet, Palmer and Lewis found that they constitute the vast majority ofdiversion project clients) (In contrast to these low-risk youngsters, for persons with one arrest iprior to the referral arrest, there was a significant difference between the recidivism rates of the treatment group and the control group. This is the group for whom diversion projects appeared to be most promising) (For individuals with more than one priOr arrest, Palmer and Lewis found that "recidivism was reduced only in those few projects whose contacts with youth were relatively numerous (over 50) and that extended over several months"\(p. xxvi). (In contrast to these figures, it is striking that the l 11 average diversion client in these projects received a total of only 5.5 service hours over a period of six weeks.) When Palmer and Lewis report on the results of the 11 projects sep- arately, it becomes clear that the observed differences can be accounted for by the success rates of a few effective programs. The majority of the programs did not reduce recidivism rates. In fact, the results of three very effective programs accounted for nearly all of the variance between the client and comparison groups. Unfortunately, the brief program des~ criptions that Palmer and Lewis provide do not suggest hypotheses about what program variables might differentiate these three from the other, less effective programs. They conclude from their findings that, even though successful diversion projects were found to be somewhat rare, this should not be viewed as discouraging. Rather, they advocate that the implementa- tion of future diversion programs should include careful documentation of program elements "so that the successful ones can be replicated and the unsuccessful ones can be discarded" (p. 215). These reviews demonstrated that the results of the evaluation efforts of diversion to date are inconclusive. There seems to be agreement among these authors that the observed inconsistencies in the results of "diver- sion" studies stem from the variability between programs and/or the incon- sistencies in the applications of the concept of diversion (e.g., to in- appropriate populations at times). The current demand, then, is to exe- cute evaluation studies of diversion programs is such a way that specific program variables can be compared systematically. The need is to begin to specify the exact conditions under which diversion is effective. In the next section, one particular diversion project which has attempted to meet the challenge of this task will be described and discussed. 12 The Adolescent Diversion Project The Adolescent Diversion Project (ADP) provides services to juvenile offenders who are referred to the project by the Juvenile Court of Ingham County, Michigan in lieu of further court processing.‘ The ADP assigns each youth to work with an undergraduate student for eighteen weeks. The students and youths spend six to eight hours per week together working on mutually agreed upon goals. The students are enrolled in a three-term psychology course sequence at Michigan State University. During the weekly class meetings the students receive training and supervision in a specific and broad-based intervention model. The model stresses two main interven- tion tactics: behavioral contracting and youth advocacy. These are both implemented within the context of a deep, overriding commitment to deal with the youth's natural environment in any and all intervention strategies. (The ultimate goal of both strategies is for the youth to develop skills to improve his/her relationship with that environment so as to facilitate the potential exchange of positive resources within relationships and with- in the larger community in general? The intervention potentially addresses all major areas of the youth's life: home and family relationships, school, friends and free-time activities, and employment skills. The following discussion will first address the major components of the ADP's interven- tion package. A brief corresponding literature review will be presented for each of these and then the specifics of the approach within the ADP intervention package will be delineated. This procedure will be followed for (l) the utilization of nonprofessional change agents, (2) behavioral contracting, and (3) youth advocacy. The discussion will then turn to the presentation of the research components of the Adolescent Diversion Project and a review of the youth outcome results achieved to date. X'13 , / r / I The Utilization of Nonprofessionals a; Change Agents Nonprofessional (or "volunteer" involvement in the mental health and criminal justice fields became popular during the 1960's. Since that time, the utilization of nonprofessionals in various roles has become widespread. It is difficult to assess the extent of this trend, although it has been estimated that 50,000,000 Americans are involved in various volunteer activ- ities (Kelly, Snowden, 8 Munoz, 1977). Many different groups within society have been recruited to fill these nonprofessional helping roles. The lit- erature on nonprofessionals includes reference to indigenous community workers in low-income neighborhoods, parents, teachers, self—help groups like Alcoholics Anonymous and Parents Without Partners, college students, homemakers, police, and other non-mental health professionals (Mitchell, 1980; Rappaport, 1977). For the purposes of this discussion, the term "nonprofessional" will be used to include: Any individual who is recruited to provide mental health services without having completed the customary professional training in one of the traditional mental health disciplines. As such, the nonprofessional may be paid or unpaid, he may be a trained, prac- ticing professional in some other field (Zax 8 Specter, 1974, p. 369). The reasons cited in the literature for utilizing nonprofessional change agents in the mental health and criminal justice fields are many. Most often mentioned are the severe personpower shortages which result from overreliance on traditional approaches to solving human problems (Albee, 1970; Mitchell, 1980; Rappaport, 1977; Tharp 8 Wetzel, 1969). More specifically, traditional psychotherapy and counseling approaches require too many highly trained, inaccessible, expensive professionals to ever hope to meet the needs of this society. These services also tend to be best suited to (or at least most often rendered to) clients who are "young adult, intelligent, verbal, successful, white, (and) middle class" l4 (Rappaport, 1977, p. 8); in short, not those whose needs are greatest. The professionals have failed to reach large segments of the population (e.g., the poor and delinquents). In those instances when they have pro- vided traditional services to these groups the results have not been pro- mising (Rappaport, 1977, p. 374; Romig, 1978). These failures provided the greatest impetus to the utilization of nonprofessionals.ii1t is be— lieved that nonprofessionals provide a promising alternative to their pro- fessional counterparts because of their greater availability and accessi- bility, the relative dollar savings which result from their utilization, and most importantly, because empirical studies suggest that they can be very effective change agents in the human service fields (an issue which will be discussed at greater length below). These same factors which have resulted in the widespread utilization of nonprofessionals generally have also lead to a parallel development within the juvenile justice system. Schwartz, Jensen, and Mahoney (1977) report that there are 2,000 criminal justice volunteer programs in the United States, utilizing some 250,000 volunteers. These authors go on to advocate an even greater utili- zation of nonprofessionals within the juvenile justice system. Again, their primary assumption is that nonprofessionals are effective change agents when used in meaningful helping roles. Recent research suggests that this is a valid assumption. Durlak (1979) reviewed 42 studies which compared the effectiveness of professionals to nonprofessionals in the mental health fields. He reports that, in 28 studies no significant differences were found between the two groups in terms of positive client outcomes as measured by a variety of variables. In two studies, professionals were shown to be superior whereas in 12 studies superiority of the paraprofessionals' performance as helpers was demonstrated. All included studies involved multiple outcome ratings 15 and a follow-up measurement period. Durlak concludes that, "The central finding from these comparative studies is that the clinical outcomes that paraprofessionals achieve are equal to or significantly better than those obtained by professionals" (p. 84). In 1981 he reiterated his earlier findings and stated, "Research has failed to demonstrate significant dif- ferences in the outcomes obtained by paraprofessional and professional helpers" (Durlak, 1981, p. 569). Throughout the literature there are examples of studies or reviews of nonprofessional programs that find no clear or significant differences be- tween the performance of nonprofessionals and that of professionals and conclude that the utilization of nonprofessionals is unwarranted (Cook 8 Scioli, 1975; Dowell, 1978; Nietzel 8 Fisher, 1981). But the point has been made (Mitchell, 1980; Rappaport, 1977) that if the performance of pro- fessionals is not demonstrably superior to that of nonprofessionals, then the utilization of nonprofessionals should be strongly advocated on the basis of greater availability and cost effectiveness. One source of nonprofessionals that is frequently utilized because of its ready availability is the college campus. College students have been found to be effective change agents in mental health roles and they are particularly well-suited to the demand for several reasons. They are espe- cially willing and eager to become involved in practical "hands-on" exper- iences that will enhance their education and they are abundant in numbers (Heller 8 Monahan, 1977; Mitchell, 1980). The conclusions from the above discussion are: (1) that the utiliza— tion of nonprofessionals is a promising and constructive concept that may alleviate some of the strains on the fields of mental health and criminal justice, (2) nonprofessionals have been found to be effective change agents 16 in those fields and (3) college students may be particularly well suited to fill these roles. This, then provides the reasoning behind the Adolescent Diversion Project's utilization of nonprofessional change agents with delinquent youth. As previously stated, the ADP assigned each youth to work with a student who is enrolled in a three-term course which provides them with training and supervision in various intervention skills. One of the two major intervention techniques utilized is behavioral contracting. Behavioral Contracting (Behavioral interventions, several authors point out, are particularly well suited to implementation by nonprofessionals due to the relative ease with which they are taught and utilizedi(Albee, 1970; Heller 8 Monahan, 1977; Nietzel, Winett, MacDonald 8 Davidson, 1977; Tharp 8 Wetzel, 1969). This is in contrast to the more traditional "therapies" implied by the "illness model" of behavioral disorders, which are much more likely to require years of study to implement. Behavioral contracting is an intervention technique derived from social learning theory. This theory views individual behavior as a function of the environmental consequences for the behaviors (Bandura, 1969). From this perspective, since behavior results from environmental contingencies, the systematic re-organization of these contingencies will result in changes in behavior. The important environmental consequences are most often social ones--generating from habitual relationships with others (Tharp 8 Wetzel, 1969). Social learning theorists view anti«social or "abnormal" behaviors as learned behaviors--exactly like any other type of behavior. Heller and Monahan (1977) list four components of a typical behavioral intervention: (1) a specific definition of the desired change, (2) accurate observation and recording of the current behavior, (3) selection 17 of appropriate contingency alterations, and (4) the establishment and con- sistent maintenance of this new contingent relationship. Most behavioral change strategies, including "token economies" and behavioral contracting utilize a similar set of steps in achieving change. A review of the literature on behavioral interventions with delinquent youth reveals very optimistic and encouraging conclusions. The following quote from Nietzel, Winett, MacDonald, and Davidson (1977) is a representa- tive summary: I The research to date has indicated a pattern of positive results in terms of improvements in specific targeted behaviors. Addi- tionally, these positive results have shown some generalization to socially defined criteria such as recidivism, particularly when the interventions were accomplished in community settings (p. 81). Ross (1978), Rappaport (1977), and Heller and Monahan (1977) come to simi- lar conclusions. Davidson and Seidman (1974) reviewed the literature which appeared on behavioral intervention with delinquents between 1960 and June, 1973. Their discussion separates the studies into three cate- gories: those targeting educationally related behaviors as targets to change, those which target pro-social institutional behaviors, and those which utilize delinquent behaviors (e.g., recidivism rates) as the outcome criteria. Of the studies which focus on educational performance, the authors conclude that behavioral interventions attain positive changes in a variety of settings including residential institutions, schools and home- based interventions. They point out, however, that the issues of general- izability of the behavioral changes to other settings and durability of the obtained changes are not adequately addressed. In regard to the studies which focused on positive behaviors in institutional settings (like clean- ing one's room, saving money, being prompt to meals and meetings, etc.), the authors report that "Again, the effectiveness of behavioral principles has been demonstrated, yet the question of the lasting nature of effects 18 is largely unanswered" (p. 1002). Also, the generalizability of the changes to other settings is challenged. Finally, in regard to the third category of studies, the authors conclude, "The studies of interventions focusing on delinquent performances as criteria have also shown positive resu1t§'(p. 1005). The reviewed studies included programs that targeted stealing behaviors and aggressive behaviors in institutional settings. They also included studies which targeted a number of pro—social behaviors by utilizing "token economies," and behavioral contracting within families and then examined the reduction of delinquent behaviors as an outcome variable. In these instances, delinquent behavior was ”targeted“ only in an indirect sense. Some of these studies included follow-up criterion measurements. Of these, some demonstrate durability and others display a deterioration of obtained changes over time. So the issue of durability is again highlighted as one which needs greater empirical attention in the future. Davidson and Seidman (1974) conclude from their review that, "Overall, the behavioral intervention with delinquents have, to date, provided promising results" (pp. 1008-1009). But they also voice strong criticisms of the quality of the research conducted in this area. They cite as common problems, the absence of control groups, the absence of baseline or reversal data in A-B-A designs, the absence of multiple out- come measures and the lack of follow-up observations. The pervasiveness of these problems leads them to conclude, "This body of research cannot be classified as confirmatory in nature due to limitations of the research designs used and the concomitant failure to rule out alternative explana- tions"(p. 1008). Redner, Snellman and Davidson (1982) conducted a similar review as an update and complement to that of Davidson and Seidman. Their findings and conclusions were very similar to those of the earlier review. They » 19 find that most studies of behavioral interventions which focus on academic behaviors report positive results, although, "the outcomes are not uni- versally positive and are often of unknown strength" (p. 22). They find that studies geared toward changing in-program behaviors are consistently positive. They note that all such studies utilize token economies as the behavioral intervention tactic. They also note that, "attempts to modify and increase the pro—social performance of delinquent youth have met with consistent success” (p. 33). Examples of targeted behaviors in this cate- gory are pro-social comments, negotiation skills, work activity and appro- priate job behaviors. They further report that, ”the few studies targeting delinquent behaviors directly have mostly reported positive results." They note that very few systematic studies examining the impact of behav- ioral interventions on the reduction of delinquency are conducted.’ Only 34% of the reviewed studies actually utilized measurements of delinquency (either archival data or self-report) as criterion variables. Of these, half demonstrated reductions either across time or as compared to a con- trol group. The authors present this as a cautiously optimistic finding. These reviewers are also highly critical of the current state of the re- search in the area, noting many of the same serious weaknesses as the earlier review. They concur that these methodological problems lend a dubious light to the generally positive findings reported in the literature. In concluding their review, Redner, et alt (1982) call for an increased specification and systematic manipulation of the program variables, and the contexts within which they operate in futUre research efforts. They summarize: The results presented so far lead one to conclude that behavioral interventions are: somewhat successful in improving academic achievement; extremely successful in modifying behaviors related to program management; extremely successful in increasing pro- . social or reducing anti-social behaviors; and somewhat successful 20 in reducing delinquent activity. Within targeted behaviors, it is very difficult, however, to specify which intervention works best given a specific population, a specific setting, and a specific type of service provider (p. 38). It is clear from the above discussion that, even though serious methodological weaknesses are prevalent in the research efforts in this area, behavioral interventions have repeatedly been found to be construc- tive and successful intervention strategies with delinquent youth. This has been demonstrated, to varying degrees, across a wide variety of set- tings and specific intervention techniques. Similar conclusions led to the decision to include behavioral approaches as a primary intervention component in the Adolescent Diversion Project. The technique of behavioral modification had produced consis- tently positive results in both institutional and community settings. . . .This is not to say that the credibility of the positive findings was firm in terms of methodological consider- ations. . . .However, the behavioral approaches appeared to show sufficient promise and constituted a set of techniques readily usable by undergraduate students (Davidson, 1979, p. 22-23). Behavioral contracting, as proposed and described by Stuart (1973) is the specific behavioral technique utilized within the intervention package of the ADP. Stuart proposed four reasons for utilizing behavioral con- tracting in community-based interventions. First of all, it is demon- strated to be effective. Secondly, it requires little administration time. It also is a method that can effectively increase and extend family influ- ences over youthful behavior, and, finally, it can be administered success- fully by paraprofessionals (p. 335). In short, behavioral contracting is very well suited to the needs and goals of the ADP. Stuart suggests the implementation of a behavioral contract in family relationships when patterns of positive reciprocal exchange are absent or weak. In such situations, a contract provides a structure and schedule for such positive exchanges between family members. The process of 21 negotiation through which a contract is derived requires that family mem- bers specify, in positive terms, their role expectations and the benefits (and costs) which accrue by meeting or missing these. Stuart suggests that the communication and specification processes increase the probability that these expectations will be met and "contribute to interactional sta- bility" within the family (p. 337). élhese, then, are the goals of the ADP undergraduate change agents when they teach the youths and their families how to utilize the behavioral contracting model. The hope is to use the_ model as a way of initiating and maintaining more constructive communica- tion skills and the increased exchange of positives within family relation- shipsQB The undergraduates also utilize intervention tactics to facilitate the exchange of positive resources in other important areas of the youth's life. Child advocacy skills are most often utilized to accomplish these goals and this component of the ADP intervention package will be presented in the next section. Child Advocacy The advocacy approach stems from the environmental resources concep- tion of human behavior (Davidson 8 Rappaport, 1978). This perspective sup- ports and encourages diversity between peoples (as contrasted to the more traditional "Blaming the Victim" approach to human "problems," as described by Ryan (1976), which often pinpoints differences between groups of people as being the cause of whatever "problem" is currently being studied). (The environmental resources position maintains that §11_people have a right to equal access to society's resources-~social, material, educational, and psychological. If these resources are not provided to all members of society, then the society is seen as failing-«not the individuals who are experiencing the lack}(Davidson 8 Rappaport, 1978; Rappaport, 1977). This position is stated succinctly by Rappaport: 22 Diversity and cultural relativity are supported. . . .There is a respect for individual differences. At the same time, the economic policy is one of social responsibility and shared resources equitably distributed. There are no inferior persons or cultures and each has a right to be judged by its own stand- ards and to receive a fair share of the resources of the larger society (p. 24). Advocacy is suggested as the means of intervening on the individual's behalf (or group's behalf) in order to access or generate needed resources that are being withheld or denied. The target for change is the society, not the individual (Davidson 8 Rappaport, 1978; Rappaport, 1977). Child advocacy is this tactic applied on behalf of children (either individually or as a group), who are often the target of discriminatory practices be- cause of their age (Westman, 1979). lestman maintains that all children require advocates and that, most frequently, parents fulfill these roles. But, when they do not, others must do so (p. 165). Davidson and Rapp (1976) point out that "delinquents can be viewed in the same way as all other youths. Their areas of unmet need--which culminate most dramatically in their entry to the juvenile justice system-~happen to meet with severe social sanction“ (p. 227). Entry into the juvenile justice system, in turn, simply exacerbates the problem of resource inaccessibility. The goal of child advocacy efforts with this population would be to reverse this pro- cess and to generate and access resources for the youths. Advocacy became popular as a social intervention tactic in the 1960's during the civil rights movement (Westman, 1979). It proved to be a use- ful and effective approach to mobilizing resources for minority groups. The child advocacy movement received its primary impetus in 1969 when the report of The Joint Commission on the Mental Health of Children called for social advocacy efforts in addition to traditional health care approaches (Westman, 1979). This idea was reiterated in 1970 by the report of The White House Conference on Children (in Krisberg 8 Austin, 1978). This report stated the case for child advocacy more strongly: 23 Children, who are powerless and need a strong voice to repre- sent them as a minority group, are now without political clout in this country. Therefore, we recommend that top priority be given to quickly establish a child advocacy agency (p. 508). (:Westman lists three major tasks of the child advocate: “(1) to know every child, (2) to know what each child needs, and (3) to make sure that needed services are available";(p. 44). He suggests that these can be accomplished at the individual or group level. These tasks obviously incorporate a broad range of roles and functions. For that reason Westman cautions that, ”. . . advocacy is an interdisciplinary enterprise and can- not be complete without taking into account all facets of a child's life” (p. 193). From this discussion it appears that the definition of child advocacy is dangerously close to ”being all things to all children.” Certainly, as an intervention tactic, it suffers from conceptual ambiguity and the lack of a specific framework. This was the state of the art when the Adolescent Diversion Project was developed and implemented. Davidson (1979) writes: At the time that the planning of the Adolescent Project was in full swing (1972-1973) very little had been forthcoming about the specifics of an advocacy approach. It was essentially necessary to evolve an operational model for use in the latter years of the project. The model of child advocacy which was eventually implemented involved a sequential problem solving process aimed at activating, accessing, coordinating, etc. the necessary community resources to provide for the needs of the youth in the Project (p. 24). This model is fully articulated by Davidson and Rapp (1976) and by David- son and Rappaport (1978). (It involved four sequential stages: (1) assess- ment of the youth's unmet needs and of available potential resources to fulfill those needs, (2) the selection of the best specific strategy to maximize the possibility of obtaining the resource on behalf of the youth, (3) the implementation of the chosen strategy, and (4) careful monitoring 24 of the outcome of the advocacy effort. The fourth step might reveal the need for further, secondary advocacy efforts if the goal has not been obtained.) This is a "multiple strategy" model of child advocacy. Davidson and Rapp (1976) describe nine distinct advocacy strategies which result from the interaction of two continua. One describes the targeted level for change and is comprised of the individual, administrative or policy levels. The choice at this point might be determined by whether the advocate is interested in accessing a resource for one individual in one specific in- stance or effecting a more far-reaching change at the administrative or policy level. The second continuum describes the specific tactical approach the advocate uses to influence the persons who control the resource. Pos- sibilities here include positive, neutral (merely providing information to the individual), or negative, aversive tactics. The choice here is based upon a careful assessment of the critical individual in charge of the resources to determine what is most likely to be effective. These two con— tinua interact to yield nine possible strategies since the advocate will choose one type of strategy from both (e.g., positive approach at the individual level, neutral approach at the policy level). This multiple strategy model of child advocacy, as presented by David- son and Rapp (1976), is utilized by the undergraduate change agents in the ADP. Discussion Child advocacy, combined with behavioral contracting, as presented in the two previous sections, are the two major components of the ADP intervention package utilized with the diverted youths by the undergraduate change agents. 25 The model is applied in an individualized fashion with each youth. The information gained during an assessment phase is utilized to deter- mine, with the youth and his/her family, which specific goals should be set for the intervention. The undergraduate then utilizes either or both of the intervention strategies to best meet these goals. The entire inter- vention philosophy is one of focusing on positive aspects of the youth's life and developing strengths. This is an approach to human service efforts which is seen by many theorists in the field of community psychol- ogy as being an essential and most productive one (Albee, 1970; Heller 8 Monahan, 1977; Rappaport, 1977). Another major tenet of the ADP intervention philosophy is that in order for change tactics to be successful in any enduring sense, maximum involvement with the target's natural environment is essential. This is a necessary and central precept of both the environmental resources and the social learning perspectives of human behavior. This also is an inter- vention approach which has received substantial support in the literature-- particularly among behaviorists (Davidson 8 Wolfred, T976; Kazdin, 1978; Patterson, McNeal, Hawkins, 8 Phelps, 1970; Stuart, 1973). When combined with the emphasis upon the teaching of contracting and advocacy skills to the youths and their families, this should maximize the potential for the endurance of any benefits gained during the intervention period. This, then describes the basic components and philosophies of the Adolescent Diversion Project. One essential question that remains is whether or not and to what extent this is an effective intervention package. This will be discussed in the next section. ADP - The Outcome Research Background The Adolescent Diversion Project began in Champaign-Urbana, Illinois 26 in 1973 as a component of a larger research project which examined the effectiveness of nonprofessionals with four distinct target populations: elementary school children, institutionalized adult mental patients, elder- ly adults and adolescents in legal jeopardy. The intervention model which has been described above (hereinafter referred to as the “Action" model) was developed at that time. Undergraduate students from the University of Illinois worked on a one-to—one basis for eighteen weeks with youths who were diverted from the police department to the ADP in lieu of court peti- tioning. The results of this phase of operation (which will be discussed below) were very promising and indicated that the model warranted further examination (Davidson, Seidman, Rappaport, Berck, Rapp, Rhodes 8 Herring, 1977). The project was formally implemented in East Lansing, Michigan in 1977 for the purposes of replication and further research. Again, the Action intervention model was utilized among others. Michigan State Uni- versity undergraduates served as change agents and the youth, in this set- ting, were diverted from the Ingham County Probate Court. The project was funded under a five year grant by the National Institute of Mental Health. Although the scope of this research has been broad, with many different aspects, it has focused on two major areas of interest throughout this period: the impact of the intervention on the diverted youths and the impact of the intervention on the undergraduate change agents. This dis- cussion will primarily be restricted to the former issue. The major outcome measure of the project's effectiveness, of course, has been its impact on the youth's subsequent delinquency rates. A review of the literature highlights an important issue in regard to the measure- ment of delinquency. Historically, official archival records were utilized to measure "delinquency." Several authors in the field have criticized this practice (Blakely, Kushler, Parisian, 8 Davidson, 1980; Elliot 8 Voss, 27 1974; Fernandez, 1981; Gold, 1966; Gold 8 Williams, 1969). Fernandez lists three major concerns with official measures of delinquency. First of all, it has been argued that official delinquency rates are more a mea- sure of the behavior patterns and decisions of police and court staff than of delinquent behavior per se. For instance, Gold (1966) reports that, "boys who live in poorer parts of town . . . are four to five times more likely to appear in some official record than boys from wealthier sections who commit the same act" (p. 27). Clearly, then, police records are not an unbiased report of observed delinquent behaviors. The second major concern with official records of delinquency presented by Fernandez is that they represent only a very small subset of delinquent acts. Gold and Wil- liams (1969) compared self-reported delinquency data with official records for a national random sample of 847 young people and found that only 3% of all reported offenses resulted in official detection and contact with the police. Elliot and Voss (1974) reported that this rate was 5% for their sample of over 2,600 high school students in California. The propor- tion that appears in court records would be even smaller. Obviously, the inclusion of only official data in any study of "delinquency" will result in tremendous underestimation and misrepresentation of the phenomenon. This creates what Fernandez lists as the third major concern with official archival measures of delinquency. Many programs which are interested in curbing delinquency provide "preventative" services to low-risk populations. It is impossible to evaluate the effectiveness of such programs with such a low frequency occurrence as official detection and processing. Reliance on official measures to evaluate these programs could lead to unwarranted conclusions that a program is not effective. These dissatisfactions with official measures of delinquency resulted in the development of self- reported delinquency measures. This was seen as the solution to the problem 28 of distinguishing between official delinquency and law-violating behaviors (Gold 8 Williams, 1969, p. 4). It has been suggested that these two mea- sures of delinquency (official archival data and self-reported data) actu— ally measure different constructs, both of which are important and valid (Blakely, et al., 1980; Fernandez, 1981). Empirical data support this con- tention: the two measures of delinquency are consistently and significantly but weakly correlated with each other. The resulting consensus of opinion from the discussion of this issue in the literature is that both measures are useful and should be utilized in combination when evaluating programs which are attempting to affect “delinquency" rates (Blakely, et al., 1980; Elliot 8 Voss, 1974; Fernandez, 1981; Gold, 1966). The ADP has included both official archival data and self-reported measures of delinquency. Official measures of delinquency outcome that will be discussed here include the frequency and seriousness of court peti- tions, the frequency and seriousness of police contacts, and incarceration rates. A self-reported delinquency measure (see Appendix A) was also utilized. In addition, school records were accessed to include several measures of school performance, grades, attendance rates, and enrollment status. Outcome Results During the first phase of program implementation in Illinois, 37 youths were referred to the ADP by the local police department. (For a more detailed discussion of the results of this research, see Davidson, Seidman, Rappaport, Berck, Rapp, Rhodes, 8 Herring, 1977.) These youths were randomly assigned to the project or to an outright release control group. There were 25 project youths and 12 controls. The project youths received the Action intervention model described previously. The results were extremely favorable. The ADP had a dramatic impact on the reduction 29 of the number and seriousness of police contacts and the number of court petitions (Davidson, et al., 1977). The authors report that, during the intervention and a one year follow-up perkxh the project youths "had fewer contacts of less severity and fewer petitions filed than the control sub- jects" (p. 45). There were also promising trends in the school data. Al- though there were no between-group differences in school attendance or grades, it did appear that the project had impacted upon the youth's drop- out rate. "At termination, 71% of those in the project group remained in school, as compared to only 5 % of the control group” (p. 47). The second year of operation in Illinois replicated these positive findings (Rappa- port, 1977; Davidson, et al., 1977). One of the major research tasks in Michigan since 1977 has been to replicate the ADP intervention model to see if comparable recidivism effects are obtained (1) within a different setting, (2) with a larger N, and (3) in comparison to a treatment-as-usual control group. Although these questions were examined within the context of a much more extensive research design (as will become clear later), the results which bear on these questions will be presented first. Between the fall of 1976 and the spring of 1979, youths were referred to the ADP by the Ingham County Pro- bate Intake Department. The referred youths ranged from 9 to 17 years of age, with an average age of 14.21. They were 83% male and 74% white (Davidson 8 Redner, 1982). They had all been petitioned to court for the commission of an illegal act. The vast majority admitted to the alleged offense at the Intake Inquiry; those few who did not were not referred to the project until after formal pre-trial or trial proceedings had resulted in a determination that the youth had, in fact, committed the offense. The youths had an average of 1.6 prior court petitions (Davidson 8 Redner, 30 1982). Only 10.4% of them were petitioned to the court because of a status offense (Davidson 8 Redner, 1982). The court staff maintained that they did not refer youths who they would otherwise have warned and re- leased. Subsequent archival record searches suggested that this was, for the most part, true. After a youth was referred to the project an interview was scheduled for an ADP staff member, the youth and one or both of his/her parents. At the meeting, the project was explained to the youth and parent(s) and it was explained that participation was voluntary. Youths were then randomly assigned to a condition. During this two and one-half year time period, 60 subjects were randomly returned to the court intake worker for treatment-as—usual. These youths were the Control group. During this same time, 76 youths were randomly assigned to the ”Action" condition as described earlier. Tables 1-3 present the comparisons of these two groups on court recidivism rates, police recidivism rates and incarceration rates respectively. All of this data refers to a two and one-half year follow-up period subsequent to project referral. Table 1 shows that after two and one-half years, while 62% of the treatment-as-usual Control group were repetitioned to the court, only 38% of the Action group were. In other words, project participation had the effect of reducing court recidivism by 39% from what it would have been if this group received normal court processing. This is a significant dif- ference. The police data depicted in Table 2 demonstrate a similar find al— though the between group differences are not as great, nor are they sta- tistically significant. (Table 3 reveals a significant between group dif- ference in incarceration rates during the follow-up period. During this time, only 13% of the project youth are incarcerated for any length of time, whereas 28% of the Control group are. 31 Table 1 Court Recidivism Number of Youths by Condition Who Were and Were Not Repetitioned to the Court During a Two and One—Half Year Follow—up Period Group Repetitioned Not Repetitioned Total Action 29 (38%) 47 (62%) 76 Control 37 (62%) 23 (38%). 6O NOTE: x2 = 6.507 p < .02 The data presented in this table are from Davidson and Redner, 1982. 32 Table 2 Police Recidivism Number of Youths by Condition Who Did and Did Not Have Contact With the Police During a Two and One-Half Year Follow-Up Period Had Contact No Contact Group With the Police With the Police Total Action 41 (54%) 35 (46%) 76 Control 40 (67%) 20 (33%) 60 NOTE: X2 = 1.755 p < .20 The data presented in this table are from Davidson and Redner, 1982. 33 Table 3 Incarceration Number of Youths by Condition Who Were and Were Not Incarcerated at Some Time During a Two and One-Half Year Follow-up Period Group Incarcerated Not Incarcerated Total Action 10 (13%) 66 (87%)' 76 Control 17 (28%) 43 (72%) 6O NOTE: x2 = 3.946 p < .05 The data presented in this table are from Davidson and Redner, 1982. 34 The comparisons of the two groups' self-report delinquency data reveal no differences either during the project intervention period or one or two years after project termination. The two groups also did not differ on any of the school variables. The results from this research, then, are mixed. The court recidivism results obtained in Illinois were strongly replicated. Furthemore, strong group differences are maintained for at least two years after project termination. Also, the differences between the two groups' incarceration rates are striking and lend further support to the conclusion that this is an effective intervention package that has an enduring impact upon the lives of the youths who are participants. But the absence of between group differences on the other outcome measures casts some doubt on this opti- mistic conclusion. It is particularly perplexing to note such differences in the number of court petitions and, yet, not to achieve group differences in self-report delinquency rates nor in the frequency of police contacts. These latter measures are purported to be more comprehensive indices of illegal behavior, since they are both less "distant" from the actual act than court records would be expected to be (Elliot 8 Voss, 1974, p. 81). At the least, these results highlight (l) the importance of including mul- tiple outcome criteria in program evaluation research and (2) the extremely complex nature of "change" as it relates to social phenomena like ”delin- quency" in open social systems. A great deal more information than that presented here would be necessary to unravel and to understand the observed results. The Need to Examine Process Variables Research results such as those just presented on the effectiveness of the ADP point up the importance of developing and utilizing adequate re- search designs and measures to tap the underlying processes of change and 35 the relationship of these processes to specific outcomes (Davidson, Redner 8 Saul, in press). This need is frequently discussed and delineated in the literature, particularly in relationship to the research conducted in the field of psychotherapy (Durlak, 1979; Gomes-Schwartz, Hadley 8 Srupp, 1978; Meltzoff 8 Kornreich, 1970; Strupp, 1978). Among these authors there is unanimous dissatisfaction with the traditional mode of conducting psycho— therapy research. Meltzoff 8 Kornreich (1970) report that relatively few attempts are made to specify and verify the therapeutic variables which relate to Specific outcomes (p. 203). Strupp (1978) and Gomes-Schwartz, et a1. (1978) both call for a movement away from the general question of whether or not psychotherapy is effective, and toward a greater specifica- tion of the issues involved. Gomes-Schwartz, et a1. state: Simple group comparisons do not answer questions about how therapy works--what elements of technique and what patient and therapist characteristics facilitate therapeutic move- ment or, conversely, contribute to negative change (p. 436). Durlak (1979) notes this same complaint about the research comparing the effectiveness of professional change agents with that of nonprofessionals: Probably the most serious weakness in comparative research lies in the failure to examine the factors that account for paraprofessionals' effectiveness. Investigators have failed to relate specific intervention techniques to specific client changes (p. 78). Even when investigators have examined the relative impacts of various types of interventions, Meltzoff and Kornreich (1970) note that the typical approach has been to label or describe the various theoretical orientations underlying the particular interventions and then to assume within group homogeneity of intervention. These authors point out that this is an erroneous assumption and that such research cannot further the task of understanding how particular interventions are helpful with particular tar- get populations under specific conditions. 36 The research in psychotherapy has been categorized by Gottman and Markman (1978) into three types-~each of which attempts to answer a dif— ferent question: Is psychotherapy effective? What kind of therapy or therapeutic system is the most effective? Or, what therapeutic processes lead to the most change? The last of these is generally referred to as the "process study." The methodology employed in such research might pro- vide guidelines for attempting to sort out the underlying processes opera- tive within the ADP. Unfortunately, such process studies in psychotherapy seldom utilize objective outcome criteria-~a fact which limits their use— fulness in determining the relative impact of specific intervention compon~ ents. In discussing the problem, Meltzoff and Kornreich (1970) noted that most researchers in the area of psychotherapy were apparently so committed, a priori, to some particular school of therapy that it never occurred to them to test the theory by utilizing objective outcome criteria--rather they would explore the conditions which seemed to maximize the occurrence of some therapist variable like, "unconditional positive regard" or “depth of interpretation,‘ which was assumed to be helpful. Still one might agree that the methodology employed by the psycho- therapy process researchers could, nevertheless, be instructive to the current task. Schlien's (1968) presentation of several process studies suggests that the most common methodology involves a content analysis of the in—session behaviors of the therapist, the client or both. Gottman and Markman's more recent review (1978) also reflects this fact. Many such studies focus on the relationship of highly specific, operationalized interpersonal behaviors--like, the therapist's verbal reinforcement behav- iors, physical gestures, or the duration of speech and silence--to the dependent variables. Unfortunately, the dependent variables most often are 37 also within-session behaviors and not, as mentioned earlier, more general- ized objective outcome criteria. But, at any rate, the focus of the research tends to be limited to the examination of minute aspects of the interaction between the therapist and client in the hopes of specifying some of these as more or less helpful in achieving certain outcomes. This perspective stems from a belief which is stated succinctly by Meltzoff and Kornreich (1970). ”The proof of the therapeutic pudding lies in what actually happens between the participants” (p. 439). Although this is certainly an essential aspect of understanding the processes which lead to behavior changes, the social learning and the environmental resources theories of human behavior would disagree vehe- mently with such a limited focus of inquiry. The person's environment-- outside of any therapeutic interactions-—would be seen as having a tremen- dous impact upon any observed or hoped for changes. From this perspective, the exclusive examination of in-session variables or variables related directly to the intervention activities of the change agent is seen as too limited. Such a restricted focus suggests a simplistic understanding of human behavior and change. Another, more serious, obstacle to utilizing the techniques suggested from this source, stems from the fact that the intervention activities employed by the traditional psychotherapy model are of a fundamentally different (and mggh_more restricted) scope than those employed within the ADP "Action" intervention package. In the latter, the "intervention“ includes a wide variety of activities which occur in a wide variety of settings. For example, an undergraduate change agent who works within the ADP model might (among other things) spend recreational time with the youth, visit school personnel on the youth's behalf, role— play job interviewing situations, visit a neighborhood youth center to gather information about available activities, and engage in family 38 negotiation sessions to develop a contract. These activities obviously occur throughout the community. As such, they are not amenable to record- ing procedures or observation in the way that more traditional interven- tions are. And the psychotherapy research reported in the literature offers little in the way of suggestions for measuring such a diverse array of intervention components. In speaking to this problem, Tharp and Wetzel (1969) state: From an experimental point of view, behavior modification in the natural environment is difficult because of problems in the maintenance of control and collection of data. The functioning source of change, should change be accomplished, is difficult to pinpoint and isolate (p. 37). For reasons described above, then, the techniques employed in therapy "process” studies have limited value for gaining a better understanding of the changes which underlie the observed results of the ADP intervention package. In discussing these problems and the complexities involved in explor- ing the specific processes which underlie changes in human behavior as they relate to therapeutic interventions, several authorsgyg make constructive proposals about possible research tactics. Paul (1967) suggests: In all its complexity, the question towards which all outcome research should ultimately be directed is the following: What treatment, by whom, is most effective for this individual with that specific problem, and under which specific set of circumstances? (p. 111). He maintains that no single study will yield such a set of answers. Instead, he advocates sequential studies with systematic variations in the process and intervention variables and careful monitoring of the differen- tial results. He also advocates that each study extensively measure and describe each of the important variables (emphasized in his quote). He thus foresees an accumulative process of arriving at the answers to the process question that he poses. 39 Several authors offer an additional suggestion: that of repeated measurement of multiple outcome and process variables throughout the inter- vention period (Gottman 8 Markman, 1978; Orlinsky 8 Howard, 1978; Tharp 8 Wetzel, 1969). Gottman and Markman (1978) argue for the use of.such designs: Their most useful contribution is that they offer a unique perspective on the assessment of interventions . . . . Interventions with clients, institutions, communities, and societies do not merely have an "effect" but an "effect pattern” across time . . . . The time series designs pro- vide a methodology appropriate to the complexity of the effects of interventions in human systems. A Proposed "Hybrid” Model for the Task Davidson, Redner and Saul (in press) note the inadequacy of current research paradigms for examining the impact of social programs and propose a hybrid model which they suggest is more able to meet the complex demands of the task. Their proposed model (of which the ADP research, presented below, is an example) includes both of the suggestions presented above. That is, it includes a sequential experimental component for hypothesis confirmation purposes and it utilizes multiple measurement points to add a longitudinal perspective. In addition, it also advocates multiple mea- surement tools rather than relying on a single outcome measure. The hybrid model that is proposed is particularly well suited to examine the relationship of on-going process variables to specific outcomes. It includes "exploratory components" devised particularly for such a task. The authors propose that these exploratory components would be designed especially to match the needs of a given research setting. They might include a battery of measures including interviews with several sources, observations, questionnaires and archival data, the focus of which would be geared specifically toward the goals of the given program. These would be administered at several points in time, allowing the monitoring of the 4O patterns of change over time of a number of variables, as well as their interrelationships. Davidson, et a1. advocate that these measures be designed to include maximum input from all of the project participants (target or client population, change agents, and staff). They suggest that this procedure could provide (1) manipulation checks of the experi- mental component of the model, (2) important longitudinal information about the relationships between the intervention activities and the observed out- comes and (3) information about other intervening variables which might be operating upon the participants and influencing outcomes. Since this model demands that measures be tailored to suit the particular needs of each research setting, "measurement development will necessarily become part of research efforts occurring simultaneously with intervention develop— ment, initiation and evaluation" (Davidson, et al., in press). The next section will describe how this "hybrid model" of program evaluation research was utilized to help to explore the relationship between process and outcome variables in the Adolescent Diversion Project. ADP--The Process Research The hybrid model proposed by Davidson, Redner and Saul (in press), as mentioned earlier, consists of two major components. The first of these is the experimental, or confirmatory component--which, through systematic manipulation of process variables and repetition over time, yields con- trolled tests of certain hypothesized relationships between process and outcome variables. The second component is the exploratory component described above. Its purpose is to provide a maximum amount of informa- tion about the entire context of the program participants' lives (includ- ing, but not limited to, the intervention activities). The result will be to obtain information about the change patterns demonstrated by multiple variables over time and thus to generate hypotheses about possible 41 relationships between process and outcome variables. The discussion in this section will present the results of this type of research conducted at the ADP. First, a description and the results of the experimental com- ponents of the model (as they relate to process variables) will be pre- sented. Review of Process Results Gained From Experimental Manipulation During the second year of operation in Illinois, in order to gain a better understanding of the impact of the "Action" intervention model, the model was separated into its two main components: behavioral contracting and child advocacy (Davidson, Seidman, Rappaport, Berck, Rapp, Rhodes, 8 Herring, 1977). In this phase there were two experimental groups--a con- tracting group and an advocacy group--and the outright release control group. Thirty-six youths were diverted from the police and randomly assigned to one of the three conditions. As mentioned earlier, the comparison be- tween the controls and the experimentals taken as one group replicated the findings of the first year (fewer court petitions and fewer police peti- tions for less serious offenses among the experimental groups). But there were pp significant differences between the two experimental groups on any outcome measures. Davidson, et a1. conclude that both intervention com- ponents are effective and advocate the utilization of the combination pro- vided with the Action model for maximum flexibility and individualized interventions. The next experimental manipulation designed to explore the specific important aspects of the intervention model involved the implementation of a "Relationship" focused intervention model which provided the same amount of training and supervision for the undergraduates as the Action model-- but the intervention was based upon the relationship therapy principles and 42 procedures developed by Carl Rogers (Kantrowitz, 1979). Another condition included in this stage of the research was a "Low-Intensity" training and supervision condition. This condition was devised to test more commonly utilized volunteer-intervention models for adolescents-~like those employed by the Big Brother/Big Sisters program. This group received much less training than the other groups and only monthly (as compared to weekly) supervision meetings with a focus on natural skills. The results of this phase of the ADP research are discussed fully by Kantrowitz (1979). Both the Action and Relationship conditions achieved significant reductions in the youths' court recidivism rates, and similar but smaller and nonsignifi~ cant trends in police recidivism. But the Low-Intensity groups did not demonstrate these effects. They were statistically indistinguishable from the treatment-as-usual control group on all outcome criteria. These effects have recently been demonstrated to have endured through a two and one-half year follow-up period (Davidson 8 Redner, 1982). Interview data with the youths further revealed that all three groups of undergraduate change agents spent similar amounts of time with their youths--so the observed differences are pp: due to simple differences in the "amounts" of intervention provided to the groups. Kantrowitz concludes: It appears that highly specific contents of training/supervision, combined with close, careful monitoring of cases, done in small classes and on a weekly basis, are crucial components of a suc- cessful nonprofessional program working with delinquent youth (p. 194). These data suggest that the actual contents of the intervention model are less important than the fact that a specific and comprehensive model is provided within the context of intense and frequent training and super- vision sessions. Kantrowitz also finds, from measures administered to the students, that those involved with the Action model reported greater satis- faction with their youths and with the project in general than did those in the Relationship condition. 43 The next phase of the ADP research was designed to explore the impor- tance of providing a wide range of services to the youths with an interven- tion designed for implementation within a broad range of areas of the youth's life (like the Action model). This was contrasted with an interven- tion which was limited to dealing with the youth's family from the perspec- tive of the social learning model (the "Family" model). Emshoff (T980) and Blakely, Kushler, Emshoff, Parisian, and Davidson (1980) present the results of this research. After a one and a half year follow-up period, the Family and Action groups combined had significantly fewer court peti- tions than did the Control group. The two groups did not differ signifi- cantly from one another on any of the delinquency outcome variables (David- son 8 Redner, 1982). Emshoff noted differences in school variables, how- ever. The Action group maintained steady rates of drop-out throughout the intervention period whereas the Family and Control groups' drop-out rates increased during this time. The same trend was noted in grade point aver- age and attendance: the other groups demonstrated a deterioration not pre- sent within the Action group. Emshoff concluded that the data demonstrated a slight superiority of the Action condition-—particular1y vis-a-vis the school variables. In the next year of operation, the ADP compared the Action and Control conditions to a condition which was devised to examine the importance of diversion away from the court system to the Action intervention package (Blakely, 1981). The students in this condition received training that was identical in content and structure to those in the Action condition, but the weekly supervision of cases was conducted by the court intake staff. Blakely reports that, by the end of the intervention period, only 18% of the "Action" youths had recidivated to either police or court, whereas this figure was 50% for the "Court" group. This difference was not significant, 44 however. But, Davidson and Redner (1982) report that, after one and a half years, 75% of the Court group had had further contact with the police, whereas only 39% of the Action group has. This is a significant differ- ence. Blakely concludes that the diversion model is essential to the effectiveness of the project. Even though both groups of undergraduates were trained to carry out the same intervention model, the students super- vised by the court staff were found (as measured by questionnaires adminis- tered to the students) to lose their commitment to the concept of diver- sion, and to become more punitive in their approaches to intervening with delinquents than were the Action students. These results have shed some light on the issue of which intervention processes or components are most important in accounting for the reported success of the ADP. The following conclusions can be made: (1) both the behavioral contracting and the child advocacy components of the model are useful and effective (Davidson, et al., 1977); (2) intense supervision and training in a specific and detailed intervention model are essential (Kan- trowitz, 1979); (3) a broad based intervention approach is slightly more effective than one limited to interventions within the family domain (Emshoff, 1980); and, (4) diversion from the court is an essential aspect of the ADP intervention package, and independence from the juvenile justice system should be maintained (Blakely, 1981). This presentation of this sequence of studies conducted on the ADP reveals how a series of experi- ments with repeated manipulations of process variables can help to unravel the question of the specific relationships between process variables and outcome criteria. This demonstrates the confirmatory component of the hybrid research paradigm proposed by Davidson, Redner and Saul (in press). The next section will describe and present the findings of the exploratory component of that model as it was executed within the ADP research setting. 45 Review of Process Results Gained From Exploratory Components As stated previously, Davidson et al. (in press) advocate the inclu- sion of exploratory components in the assessment process when examining the effectiveness of social interventions. These are designed specifically to explore the context of the particular program in question to achieve a clearer picture of the processes of change which underlie any observed effects in the outcome criteria. One of the primary measures developed and utilized for this purpose vis-a-vis the youth-related outcome effects of the Adolescent Diversion Project is the "process interview." This was first utilized during the last phase of research conducted at Illinois. In discussing the rationale for this interview, Davidson (1979) states: It was felt essential to not only replicate the findings of the second year but to begin attempts at analyzing the processes which might be leading to the observed results. This led to the develooment of an intricate set of longitudinal, explora- tory process interviewing procedures . . . (p. 56). He later goes further in explaining the function and purpose of these interviews: One of the major foci of this research was to gain a detailed monitoring and understanding of the ongoing process of events in the lives of the youth, the components of the intervention approaches, and the salient features of the training and supervision sessions (p. 67). It was believed to be extremely important to obtain this detailed infor- mation from the perspective of the program participants (youths, parents, undergraduate change agents, and superVisors). When first implemented, the interviews were conducted at three points throughout the intervention period with each of four sources: the youth, a parent, the student, and the student's supervisor. The interviews covered all important aspects of the youth's life (school, friends, free time activities, home life, and involvement with the legal system) and the intervention process in detail. The interview information was coded into items which were combined into 46 scales via the rational/empirical method (Jackson, 1970; Selltiz, Wright- man 8 Cook, 1976). The measurement development procedures utilized will be discussed at greater length in Chapter II. The Illinois Setting--Process Results The major findings of the analysis of the relationships between these scales, the three conditions, and a success/failure criteria (wherein suc- cess is defined as the absence of further police or court contacts during project involvement and attendance at school of at least a weekly average of two days) are presented in Table 4 (reprinted from Davidson, 1979). Davidson summarizes these results: For all conditions it was apparent that the success-failure criteria was closely related to what has been described as socially acceptable or sanctioned role involvement . . The youth who end up in further trouble with the police and completely uninvolved in school attendance are characterized by low levels of involvement at home, with the school system, and with the employment market (p. 87). The Michigan Setting--Intervention Interview Results The process interviews were also utilized in the research carried on at the Michigan site. Here the interviews were conducted at four points: at project referral and six, 12 and 18 weeks later. This last time coincided with project termination. The interviews were conducted with the youth, a parent and the volunteer in each case. There were two major components of the interview: (1) a "Life Domain Interview" which sought specific information about the youth's life in many areas, includ- ing school, home, and free time activities, and (2) an "Intervention Inter- view" which focused solely upon the activities directly involved in the intervention. This latter component, obviously, was only conducted with the five different experimental groups (Action, Relationship, Low Intensity, Family and Court). The results discussed in this section include the com- bined data from all of the youths who participated in the project between .Ammmpv cemuw>eo Soc; cmpcweqmm .muozr .EO wucm>mecw FOOLUm .cowp:m>cmu:w Poozom Poewcwe use mm Loo cuzoa :o nmmzuom 47 Foewcwa nco xpwsow oz .o cowpcm>empcp xpwEMJ .o .umxopqu .umeFQEm zuzox “on cu mcwuasmuu< .m sync» pom on mcwuaewuu< .m .Empmzm mowumsn .Emamxm mowumsn mpwcm>zn ou ocwvcogmmm .q mecm>zn op mcwucoamwm .v mes—wed .pmoos .Poooe mcwpoecucou oceuowuwcw mewpooeucoo mcwuowumcw epnzocp powuwcH .m anzocp PowuwcH .m .mcowmcosro .mcowmcmewo .mcompcmewo omcegu co cowbecowcmpmo .N mmcogu co coeuoeowcmpmo .N mmcozu co cowpocowewpmo .N .moFOL nm>ocaao . .mmFoL om>ocaoo .mcwxmmm new :w vm>Fo>cH .P >Fwowu0m cw nw>_o>:w:: .P xprowoom cw uw>~o>cwcz ._ .mucoseomcma .Ewpmxm Foonum on» Foozom m.;g:oz mzu cw cw mmmcmno co mewxcoz .m mmmcozo co mcwxcoz .m .mocmwew .omcm waEmm ogy m.;u:o> we“ cpwz mcwxeoz .e cw mmmcmgo co mcwxeoz .q .Fmooe .—mooe mmmousm xueoo>ue mo cowbemuwcH .m mewuueeucou mewuowuwcH .m .mcowmcmewc .mcowmcwswo .mcowmcmewo seemed :0 seepweeom .N emeeeo ee sowpweeem .N emceeo :0 seepeeeom .N .Poogom new .mmpoc um>ocaao .mmpog um>ocaao x_w5od cw vm>_o>cH ._ xrpowuom cw om>Po>cH ._ xppomoom cw nw>Fo>cH .p _ocucou soeoo>e< ep_eo mcwuuogpcoo Foeow>mzmm m me>11mumo mwocw—PH Aoc:__ee\mmmuo:m x cowuwvcoo x mmcoom mroom mmmuocov —mooz xucmmcwpcou mFawupzz on» we mawsmcowumpmm so m—nmh 48 the spring of 1977 and the spring of 1979. The information yielded by the intervention interview provided an excellent “manipulation check" of the various intervention conditions. There were, accordingly, significant differences found in the intervention interview data between the various conditions. These demonstrated model adherence in most instances and have been discussed at length previously (Blakely, 1981; Emshoff, 1980; Kantrowitz, 1979). The data were analyzed using five (condition) by two (success vs. failure) by three (time period) analyses of variance. A youth is included in the "failure" group if any contact with the police or court occurred subsequent to project referral. Only the results which differen- tiate between the success and failure groups will be discussed here. These analyses included the data from the youth interviews only. (All the infor- mation in this and the following section was obtained from Davidson and Redner, 1982.) Three intervention scales yielded significant results involving the success/failure variable: 1. Amount of Time. This scale measured the amount of time that the volunteer spent working with or on behalf of his/her youth. The results showed that the volunteers who worked with youths who recontacted the police or court at some time during the intervention or one year afterwards spent increasingly less time with their youths throughout the intervention period (time x success/failure F = 4.34, p < .01). These analyses, of course, do not allow a determination of the direction of causality, but this probably reflects the volunteer's reaction to the youth's reinvolve- ment in the juvenile justice system. 2. School Intervention-~System Focus. This scale measured the extent that intervention activities focused on improving the school situation by attempting changes at the school (e.g., changing classes or teachers). 49 This type of intervention activity happened more within the "failure" group than the ”success" group. This effect is strongest when the success/ failure (S/F) classification is based upon the intervention period (F = 4.81, p < .03); only a marginal difference is obtained when S/F includes the one year follow-up period (F = 3.31, p < .07). 3. Legal System Involvement. This scale measured the extent to which the volunteer became involved with the juvenile justice system on behalf of the youth. Again, and obviously, the "failure“ group received more of this type of intervention than the "success” group. This effect was apparent and strong when S/F was measured during the intervention period (F = 10.96, p < .001) and when it included the one year follow-up S/F information (F = 9.83, p < .002). It is also informative to note some of the intervention scales which do not demonstrate S/F group differences. For instance, none of the scales which measure the extent that the intervention focuses on the home domain demonstrated between group differences in regard to the success/failure variable. Similarly, the two scales which measured the amount of behavioral contracting activity and the amount of advocacy activity did not show any such effects. These results from the intervention interview suggest that the youths who become re-involved with the juvenile justice system receive a very different intervention package than those who do not. The youths who become re-involved receive a decreasing amount of intervention time from their volunteers and this time is relatively more often spent in dealing with the school system and the juvenile justice system. It is obvious that the latter involvement is elicited by the occurrence of a crisis (i.e., the volunteer only becomes involved in this system as a result of the youth having gotten into trouble). One might surmise that the between 50 group differences on the School Intervention scale can be similarly under- stood. That is, it is possible that the "failure" group volunteers more frequently become involved with the school system in response to crises (e.g., suspension or failing grades). The next section will discuss the results from the life domain component of the interview which demonstrated significant differences between the success and failure groups. The Michigan Setting-—Life Domain Interview Results. The data from the life domain interview were analyzed using a five (condition) by two (success vs. failure) by four (time period) analysis of variance. Five life domain scales demonstrated significant effects with regard to success/failure. 1. School Involvement. This scale measures the extent to which the youth is involved in school (how often he/she goes, whether or not he/she likes school or does well in it, etc.). The analyses here reveal that youths who recontact the juvenile justice system are less involved in school than are those youths who do not. This is true whether S/F is based upon the actual intervention period, a one year, or two year follow-up period (F = 5.21, p = .02; F = 11.53, p = .001; and F = 7.43, p = .03, respectively). This replicates earlier findings reported by Blakely, et a1. (1980) and Davidson, et a1. (1977). 2. Positive School Change. This scale measures the extent of change in the school domain in such variables as attendance, grades, classroom behavior, homework completion, etc. This scale demonstrates an interaction effect with time and S/F (F = 3.78, p = .01). The "failure" group demon- strates more positive change over time. Since this scale measures change, this is somewhat difficult to interpret. Since the previously discussed School Involvement scale demonstrated that this group was less involved with school than the "success" group, this result may reflect the fact 51 that these youths have more frequent problems within the school domain and therefore there is more room for improvement and change in this area. This effect is only apparent when S/F is measured during the intervention period-~not when measured during one or two year follow-up periods. 3. Positive Home Change. This scale measures the extent of change in the home domain. It includes items such as, "What change has occurred in terms of the parent and youth arguing?” This scale demonstrates a con- dition by success/failure effect (F = 2.70, p = .02). Among the Action, Relationship and Family conditions the success group demonstrates more positive changes than the failure group, whereas for the Court and Control conditions, the failure group has higher means. The Low Intensity successes and failures do not differ. These condition differences may reflect the more proactive, strength—building approach of the Action, Family and Relationship change agents, whereas those in the Court group may have been more likely to work toward improvement in this area after a crisis (like re-involvement in the juvenile justice system). The same might be true in the Control group. 4. Parental Control. This scale measures the extent that the par- ents attempt to control, by negative means, the behaviors of the youth. It includes several items which focus on the frequency of arguments over various concerns (chores, friends, the use of the phone, etc.). This scale demonstrates a strong main effect for the S/F variable when it is based on one year and two year follow-up periods (F = 6.23, p = .01; F = 8.16, p = .005, respectively). It does not, however, yield significant differ- ences between the S/F groups when they are determined on the basis of the "during-intervention" time period. This suggests that the more parents attempt to control the youth (which reflects dissatisfaction with the 52 youth's behavior), the more likely it is that the youth will get into legal trouble in the future. 5. Leggl Involvement. This scale measures the extent of the youth's involvement in the legal system. The results replicate those found in the corresponding intervention interview scale-~the “failure” group is (obviously) more involved with these activities than the ”success” group F = 22.73, p = .0005). These results from the life domain interview data demonstrate impor- tant differences between the youth who are successful in remaining unin- volved with the juvenile justice system and those who become re-involved. The "successful" youth are more involved in school and apparently have more positive relationships with their parents (since the "failures'" higher scores on the Parental Control scale can be viewed as reflecting parental dissatisfaction with the youths' behaviors). Ohis is consistent with the earlier finding by Davidson (1979) that the youths who were suc- cessful in the Illinois project tended to be more involved in socially acceptable roles! More generally, the results demonstrate the importance of including the "exploratory component" of the hybrid model as proposed by Davidson, Redner, and Saul (in press). The inclusion of the process interview allows a careful examination of the relationship between contextual/environmental variables, the ongoing processes of the intervention and the observed changes in the criterion or outcome variables. Unfortunately, the analyses reported above do not allow a comparative examination of the relative importance of the various process and contextual variables in accounting for the variance of the outcome measures. This is the focus of this pro- posed study. 53 The Current Research The preceding review suggests that the relationships between process variables and the outcome results achieved by social intervention programs must be examined if we are to achieve appropriate applications of specific interventions to specific human problems. We have seen that this approach to evaluation research is widely advocated in the human service fields (Davidson, Redner, 8 Saul, in press; Durlak, 1979; Gottman 8 Markman, 1978; Meltzoff 8 Kornreich, 1970; Orlinsky 8 Howard, 1978; Paul, 1967) but seldom applied. This has been due in part to a lack of appropriate methodology to meet the scope of the task. The “hybrid model" proposed by Davidson, Redner, and Saul (in press) was suggested as a research paradigm which ppglg_meet this need. We have seen how the combined applications of the con- firmatory and exploratory components of this model, in the evaluation of the Adolescent Diversion Project,have been extremely productive in this regard. This effort to "untangle" the impact on outcome of a wide array of intervention-process variables and other contextual variables operative in the lives of the project youths is especially important in regard to the ADP. This is because of the perplexing outcome results which have been reported to date (Davidson, et al., 1977; Davidson 8 Redner, 1982; Blakely, 1981; Emshoff, 1980; Kantrowitz, 1979). The project has been demonstrated to be very effective in reducing "delinquency," as measured by certain criteria (court recidivism and incarceration); but not equally effective in reducing "delinquency" as measured by other criteria (self-reported delinquency and police recidivism). These results demonstrate the impor- tance of including multiple outcome measures in the evaluation of social intervention programs. Specifically, they highlight the importance of 54 including self-reported delinquency (SRD) measures and support the conten- tion that SRD and archival records of formal contacts with the juvenile justice system in fact measure two different constructs or aspects of the phenomenon of "delinquency“ (Fernandez, 1981). These apparently contradictory results also point out the complexity of the problem of achieving (and measuring) changes in human behavior as a result of an intervention package applied within an open social system. ' The careful monitoring of process and contextual variables is essential if we ever hope to understand the complexities of the intricate patterns which lead to change. The process interviews, which were devised especially for such an exploratory task within the ADP setting, have provided an excellent tool for achieving this understanding. Previous univariate examinations of the process interview data have resulted in substantial gains in information about the patterns of change which are associated with the observed outcome results. But these analyses do not allow an examination of the relative importance of these variables as compared with one another in accounting for the obtained variance. A multivariate examination was required for this next step. The current study conducted such an examination. It explored the com- bined influences of a number of process variables (as predictors) on two different outcome or criterion variables. The process variables were derived from the process interviews described above. These variables included the youths' scores on the three life domain scales which measure various aspects of the youth's relationship to his/her family, two scales which measure his/her relationship to school, one which focuses on job seeking activities, two which focus on peer relationships, and ten of the 55 intervention scales which measure different components of the students' intervention activities throughout the youth's involvement with the program. These eighteen scales were utilized as predictor variables in two dis- criminant function analyses, each examining a different outcome or cri- terion variable. The first of these utilized as the criterion variables the occurrence of court or police recidivism during a one year follow-up period to categorize the youths as ”successes” or "failures.“ The second utilized self-reported delinquency rates during this same period. In this case the youths were assigned to one of three categories determined by the magnitude of the SRD score (Low, Medium, and High scores). These analyses addressed the following research questions: 1) Which process variables appear to be most strongly associated with the occurrence of official recidivism during a one year follow-up period? And, similarly, which ones appear to be strongly associated with the level of SRD rates? 2) To what extent can a combination of these variables account for the observed differentiation on these two outcome criteria? (i.e, To what extent can you predict delinquency outcomes one year after project termi- nation based upon the knowledge of these process measures?) 3) 00 different combinations of the process variables account for official delinquency outcome and for SRD outcome? How do the solutions for these two discriminant functions compare? The following chapter will describe the study in greater detail. CHAPTER II METHOD Subjects Youth participants. The sample for this research consisted of 107 youths who participated in the ADP between April, 1977 and June, 1979. These 107 youths were among the 133 who were randomly assigned to one of the five experimental conditions discussed in Chapter I which were under study within the context of this research project during that time (Action, Relationship, Low Intensity, Family and Court). Twenty-one youths were eliminated from the total sample due to the inability to conduct follow-up measures with them one year after the completion of their project involve- ment. These follow-up interviews were not possible either because the youth could not be located or they refused to be interviewed. The comple- tion rates of these follow-up measures for the five conditions were nearly identical (Davidson 8 Redner, 1982). Another five youths were eliminated because of missing interview data from the period during their involvement with the project. These youths were also nearly equally distributed be- tween conditions. Throughout this time, the Adolescent Diversion Project operated under a federal grant from the National Institute of Mental Health. It diverted youths from the Ingham County Juvenile Court-~The Intake Division. All youths referred to the project had been petitioned to the court for a vio- lation of the juvenile code.1 The nature of the offense varied greatly. Table 5 presents the frequency distribution for all youths by referral 56 57 offense. This table includes all youths referred to the project between the Fall of 1976 and the Spring of 1980. The youths who were referred during this time had an average of 1.6 prior court petitions (including the referral petition) and 1.4 prior arrests (Davidson 8 Redner, 1982). The project was operated in conjunction with the Department of Psy— chology at Michigan State University in East Lansing, Michigan, where the undergraduates who acted as service providers to the diverted youths were enrolled. The average age of the 107 youths who were included in this study, during their interventions, was 14.25 with a range from 9 to 17 years. Seventy-eight and one-half percent of the youths were Caucasian. Only 18.7% were female. The youths represented a wide range of socio-economic levels, but the majority were from lower and lower-middle income homes. Student participants. Each youth worked for 18 weeks with an under- graduate change agent who was enrolled in a three-term course sequence in the Psychology Department. The sequence began either in Fall or Spring Term and ran consecutively through the next two academic terms. Each stu- dent enrolled in a four credit course for each of the three terms that he/ she was involved with the project. The students were primarily social science majors. Most were in their junior year. The students were ran- domly assigned to condition during Years 111, IV, and V (see Table 6). Table 6 provides a breakdown of the 107 youth-student pairs who will be included in this study by condition and year. Trainers and supervisors. The students were trained in the interven- tion model dictated by their particular condition in small groups of six to eight students each (with the exception of the Low Intensity group, which was supervised less intensively as will be described below). Each 58 Table 5 Distribution of ADP Participants by Referral Offense Adolescent Diversion Project —- Fall, 1976 through September, 1980 Type of Offense .5299; B 8 E (breaking and entering) 6O malicious destruction of property 14 RAW (runaway) 8 incorrigibility/curfew (pg: school incorrigibility). 7 larceny 83 possession of a weapon 1 assault without weapon 7 assault with weapon (including armed robbery) 4 possession of controlled substance (other than 9 alcohol) arson 1 UDAA (car theft)/joy riding 11 false alarm 4 eluding/resisting arrest 1 possession of stolen property ' l4 forgery 3 truancy (including school incorrigibility) ll bomb threat 1 leaving the scene 3 reckless or felonious driving, DUIL 2 reckless use of firearm 2 criminal sexual conduct (CSC)--4th degree 1 obstructing justice __41 Total 248 Note: This information was compiled by Davidson and Redner, 1982. Pct. 24. 5 33. moo monoobmoomcsm A r: a: on a: n: e» a. n) gs as as e» $- —J O O O 59 .xoapm meg“ cw wouDFUCw mam: on: gene» monapocw anou mmcpe Noe am e_ es up WPeoee F_ P_ -- -- -- peace mp _F m -- -- s_w2ea Fm -- --. _~ -- seemeoeeu so; NF -- -- N_ -- geeseoeoepoe es A a PP N_ eo_oo< .mammmm .Naemfi .meeeem dense .__eu “weep «me_eem ~Nem_ .Ppee cowoweeoo -mea_ .__eav -mem_ .meeeemv -eemy .__eev -Nmm_ .meweemv > me> >H me> H: me> : (80> ieeo> eee_ee_oeeeoo so mewea oeooaom--eo=o> ao< eo ezoexeoem o anmh 60 small group in the Action, Relationship and Family group was trained and supervised by a pair of psychology graduate students. The Court group was trained by one advanced psychology graduate student and a Court Intake Referee. They were subsequently supervised by the Court Referee alone. . The graduate students who were responsible for the training and super- vision of the undergraduates were each simultaneously responsible for con- ducting some specific research related tasks on the project. They were rotated through the various conditions over time as much as possible so that no one supervisor would be exclusively associated with one of the conditions--however, they were not randomly assigned, nor was it possible to completely cross supervisors with conditions since none of the condi- tions except for Action was implemented often enough to achieve this. The organizational structure of the ADP was modeled after the "Educa— tional Pyramid" model suggested by Rappaport (1977). This model advocated a hierarchical structure wherein the target population (in this instance, the diverted youths) constituted the bottom tier. They were served exclu- sively by the undergraduate change agents who made up the next level in the pyramid. The students were Supervised by graduate students, who also coordinated the research efforts. They were, in turn, supervised by the Project Director who oversaw the entire organization. - Procedures Recruitment of students. Each Winter and Spring Term, course announce- ments were sent out to over 1,000 social science majors informing them of the availability of this course sequence beginning the following Spring or Fall Term. (Fall Term students were recruited during the prior Spring Term, since many of them left town over the summer.) The announcement (see Appendix 8) provided a general description of the course experience and instructed interested students to call for more information. Those 61 who called were told to attend a recruitment meeting where they were pro- vided with more detailed information about the content and requirements of the course (by the Project Director). Students who were still interested in pursuing enrollment in the course were told to come to a subsequent meeting where they would be required to fill out a number of demographic measures, attitude questionnaires, and Jackson's Peronality Research Form (Jackson, 1968). These measures were utilized in conjunction with the larger research project and are not directly related to the current study. From the pool of students who had attended both meetings and completed all of the measures, the necessary number of students for the following ”year" of the project were chosen.' An equal number of males and females were selected randomly. An additional group was chosen to make up a mailing list, should some of those selected to begin the class change their minds. The final set of students was randomly assigned to condition. Additional students, who had been rejected through the above described selection pro- cess, were given the option of enrolling in a simultaneous course sequence. This course would be involved with administering the process interviews which were conducted four times with each diverted youth, a parent, a peer (in some instances), and--for experimental youth-~a student. Some other students were recruited by the academic counselors in the Psychology and Criminal Justice Departments for this interviewing course sequence. Approximately 10 to 15 students were enrolled in this course during each "year" of the project. Referral of youth. The youths were referred to the project from the Intake Department of the Ingahm County Juvenile Court after a preliminary hearing had been held in conjunction with the charge for which they had been petitioned. If the youth admitted committing the offense at this point, the ADP was one of the referral options available to the Intake Referee. 62 The referees had agreed not to refer youths who they would otherwise warn and release. If, after having heard a brief description of the project from the referee the youths and parents were interested in pursuing the possibility of becoming involved, the referee would schedule an appointment with one of the project staff. This meeting with an ADP staff person took place at the court. The youth and parents were given a detailed description of what project involvement would entail. They were told that involvement was on a volun- tary basis; and if they decided to participate, the project would last for 18 weeks and during that time they would spend six to eight hours per week working with a student. It was explained that the intervention was an individualized one, so that everyone's experiences were unique, but a general description of the kinds of activities that might be included in their work with the student was provided. They were also told that not every youth would be assigned to work with a student--that only three-fourths of the youths would be and that a random lottery system would determine which youths would work with a student. They were told that, if they agreed to participate, they would be expected to be a part of the evaluation procedures whether or not they were assigned to a student. It was explained that these evaluation procedures entailed (1) re- leasing access to school records, (2) releasing access to all police and court records, and (3) participating in a series of interviews. The youths would be paid $5.00 for each interview. It was explained that the project would want to interview the youth, parent and one of the youth's friends (to be nominated by him/her). The interviewer, they were told, would inquire about many different areas of the youth's life, including 63 home, school, freetime activities, and intervention activities with the student to whom they were assigned. The voluntary nature of participation was again stressed. They were told that if either of them did not wish to participate they would go back to the referee for another disposition decision. This was also the proce- dure if they happened not to be assigned to a volunteer after deciding to participate with the ADP. If, at that point, they made a commitment, the participation agreement was filled out and signed (see Appendix C). Then the youth was randomly assigned to condition by drawing an envelope that contained the condition the youth would be assigned to. The envelopes were stratified by referee, the sex of the youth and race. The youths who had been assigned to the Control condition were told to expect to hear from the interviewer and were returned to their Intake Referee. Youths who had been assigned to one of the experimental conditions were told that they would be contacted by an interviewer and their assigned student within a day or two. The case information was then forwarded to the supervisors in charge of that condition and they assigned a student to the youth. Assignment was random with two exceptions; students without cars were assigned to youths who lived close to bus routes and girls were always assigned to female students. The students were told to contact their youth as soon as possible to set up an initial meeting time. Training and supervision. For Action, Relationship, Family and Court conditions, classes met in small groups of six students with two supervisors on a weekly basis. Each class session lasted for 2% hours. The students were supplied with training manuals designed for their particular condi- tion (Court and Action groups used the same manual). The training por- tion of the classes lasted for eight weeks. The manuals for each condition 64 were divided into eight sections. The material in each was to be mastered for the corresponding class meeting. In addition to the manual, outside readings were assigned each week. These were provided to the students on a "reserved" basis by the University Library. These courses were taught according to a "mastery" model. The beginning of each class meeting was spent answering any questions the students had about the material assigned for the week. After this discussion session, the students responded to three short essay questions on the material. These were then corrected immediately and returned to the students. If the students had not demon- strated an 85% mastery level on each question, they were required to re- write the answers and turn them in again within the next few days. The correct answers to the questions were discussed during class. Next, the group responded to oral questions covering the same material which were similarly evaluated. Again, if the students did not demonstrate mastery, they wrote out their answers and resubmitted them. From the fourth week on, these didactic activities were supplemented with role-plays of impor- tant types of intervention interactions (e.g., contract negotiation ses- sions within the Action, Family and Court groups). The emphasis on mas- tery and practice were intended to maximize the probability that the students were, indeed, learning the intervention models. The content of each of the training components of these four conditions were: 1. Aptipp. The Action group was trained to utilize behavioral con- tracting and child advocacy skills. Both of these intervention approaches were discussed in Chapter I. They were expected to utilize these tech- niques to aid the youth in dealing with a wide range of situations in their lives (school, home, freetime, employment and the legal domain, if necessary). They were trained in (a) assessment techniques, (b) the imple- mentation of intervention strategies, (c) careful monitoring and "trouble 65 shooting" skills, (d) secondary intervention strategies to incorporate the information gained from the monitoring activities, and (e) termination strategies which shifted the major responsibilities for carrying on with these activities to the youth and his/her family. 2. Relationship. This group's training was based upon the inter- personal theories of human behavior put forth by Harry Stack Sullivan and Carl Rogers (Kantrowitz, 1979). The intervention strategies which were taught focused upon establishing meaningful, authentic relationships be- tween the students and youths. This was expected to impact upon the youths' self-concepts in constructive ways which would diminish the likeli- hood of future delinquent behaviors. Students were taught to utilize accurate empathy skills, a problem-solving model, feedback skills and other communication and relationship skills to facilitate the development of a close, warm relationship between the student and youth. 3. fgpily. This condition was devised to test the importance of the broad focus of intervention of the Action model against a scope of inter- vention limited to the family domain (Emshoff, 1980). Training in this case was similar to the Action group's training in every respect except that it focused solely on interventions within the family system. The role of behavioral contracting, accordingly, assumed greater emphasis in the training of this group. As in the Action group, the students were taught assessment, intervention, monitoring, "trouble shooting," and term- ination strategies to utilize in their interventions. 4. Eggpp, The Court group received exactly the same training as did the Action group (Blakely, 1981). The difference between the two groups was that the Court group's case supervision was carried out by a Court Intake Referee (as mentioned previously). He met with the students in a 66 group every week after training to review and to plan the individual case interventions. Supervision of cases for the Action, Relationship and Family groups was also carried out in the small groups on a weekly basis-—but by the ADP supervisor pairs who had also been responsible for the training of the groups. The format of these supervision groups allowed each student to present the goals, activities and accomplishments of the preceding week and to discuss any unresolved issues or questions with the group. They developed and presented their plans for the next week as a result of this process. The goal was to create an active problem-solving orientation within the groups and to foster group input into the individual cases. This supervision format allowed a close monitoring of the cases and pro- vided the students with the resources of their fellow students' experi- ences. In addition to the weekly group meetings, students were encouraged to call their supervisors at any time if problems or questions arose. In contrast to these intensive training and supervision models, the fifth condition--Low Intensity-~was managed quite differently (Kantrowitz, 1979). These students received training sessions which presented very gen- eral information about the volunteering role with juvenile offenders. The focus of the training and intervention was on utilizing the students' natural helping skills for the benefit of their youths. This would pri- marily include such things as being the youth's friend, providing a good role model and sharing in mutually enjoyable activities. The students met with their supervisiors on a monthly basis for case supervision. This condition was devised to reproduce the circumstances which are more common within programs that serve young people with nonprofessional "helpers." This model could then be compared to the more intensive training and super- vision models presented above. 67 Intervention. The intervention activities of each group were deter- mined by the content and format of the different training and supervision conditions. Manipulation checks revealed that the students in each condi- tion (with the exceptions noted earlier regarding the Court group) adhered to the models which they had been taught (Kantrowitz, 1979; Emshoff, 1980; Blakely, 1981). All students were expected to spend between six to eight hours per week working with their youths. The intervention period lasted for 18 weeks in all instances. This excluded any academic breaks observed by the University. All students were expected to carry out their inter- vention activities within the community--the youth's natural environment. Although the extent to which this occurred would obviously be influenced by the training models. For instance, only the Action and Court groups would be expected to become directly involved with the youths' schools. But the intensity of the intervention, as measured by the amount of time spent with the youth on a weekly basis, and the length of the intervention period, were the same across all conditions (Davidson 8 Redner, 1982). Measures The measures which were utilized for the purposes of this study include: 1. An indication of the occurrence of police and court contacts sub- sequent to referral to the ADP up until one year after the termination of project involvement. This measure was based upon archival record searches. 2. A measure of self-reported delinquent behaviors. 3. Detailed information about significant areas of the youth's life-- school, home, friends, freetime activities and employment. This informa- tion was based upon the process interviews conducted with the youth. This will subsequently be referred to as the Life Domain Survey. 4. Detailed information about the extent and nature of the interven- tion activities that the students carried out with or on behalf of the youth. 68 This information was also based upon the youth interviews. This will sub- sequently be called the Intervention Survey. A depiction of the measurementschedule is presented in Figure 1. Each of the measures will be discussed separately. Archival police and court data. The information about subsequent police and court contacts was collected semi-annuallyrthroughout this time (Spring of 1977 through the Spring of 1979) and for several years after- wards. This length of time increased the probability that the records concerning any contacts with the police or any petitions to the court would be in the files and accessible to the ADP research staff at some time during this period. The ADP staff's experience with these records suggested that it often took months for records to reach the files and that, even then, files would often disappear for unexplained reasons and variable lengths of time. For these reasons, repeated, frequent collections were deemed necessary to maximize the probability of a complete data set. The archival data were gathered by graduate research assistants, with the help of under- graduate coders. The police data were gathered from the County Sheriff's Department, all major police stations in the area, the local branches of the State Police and several smaller township police departments. Several variables were collected from the police records, but the only one of concern to the present study is the number of police contacts. The court data were all collected from the local County Juvenile Court. Again, several court variables were collected, but only the number of peti- tions subsequent to project referral is of interest for this study. Process measures. The Self-Report Delinquency measure (SRD), and the Life Domain and Intervention Surveys were all administered to the youths by the interviewers who were enrolled in the research course sequence cowumuumwafiav< ucmfiouammoz mo Emuwwwn .H ouswam mo>u=m acauao>uoucH .ucH mo>u=m nausea owaa an hocozccwaon uuoaom maom mam .AcoHuooHHoo «not uov aofiumuumwcfiva ucoaousmwoa «0 mafia uoon mommaamam x .oHanHm>w mw wumv noun? you moduoa mwacwwm abouu< AIIIIIIIIV v L 'y kl t x .8511 i. x Jodi . x JS 9 ,o r. “V.“ VWAT .r Aw x an x 9A x Ga an 1. ii 11 iv A 1w J“ 95 J“ an x 58 “mom mono» osevx x x mummzflwzwfiww - 3-2 uauum .2-ng umom your one a o a u c o > H o u a H w a w u a a mum o3003.o 1— 70 that ran concurrently with the intervention sequence. This course involved a three-term commitment from the students as did the other sequence. It was taught by ADP graduate student research assistants. The first term of this course involved training the students in interviewing techniques, familiarizing them with the ADP, teaching them coding and data management skills and training them to administer several paper and pencil measures, one of which was the SRD. (The others were not pertinent to this study.) This portion of the course relied heavily upon role-playing and practice interviews combined with frequent, specific feedback from the supervisors. The interviewing technique which was taught was an open-ended one. The interviewers had a specific set of items to code from the information gained during the interviews but they obtained this information by asking open-ended questions, followed by increasingly specific probes--rather than administering a set of items. It was felt that this interview tech- nique would yield higher quality data than more traditional methods (Davidson, 1979). The interviews were all audio-recorded to maximize the accuracy and retention of the information when the interview was coded later. The Life Domain and Intervention Surveys will be discussed at greater length in later sections. At the end of the training segment of the course, before being assigned to interview a current ADP case, the interviewers conducted a complete practice interview anwi measure administration with previous interviewers or ADP staff members. They recorded the entire interview, coded all of the information and then reviewed the entire procedure with a supervisor. They were then assigned cases. Each case consisted of four interviews with the youth and four with the parent (at the point of referral; and six, twelve, and 18 weeks later). It also involved three interviews with the youth's volunteer (at six, 71 twelve, and 18 weeks). One interviewer was responsible for all of these interviews for each case assigned to her/him. Once assigned, the inter- viewer contacted the youth and parent and scheduled appointments for the interview. The interviews were generally conducted at the youth's home in a place where privacy could be assured. All sources were interviewed separately. The interviews generally lasted for one to one and a half hours. The interviewers assured the sources of the confidentiality of the interview and the anonymous handling of the information. The youths were paid $5.00 to complete each interview. Every interviewer was responsible for four to six cases, one of which was an interrater case to which two interviewers were assigned. One of the interviewers in the pair conducted the interview, the other observed and tape-recorded the procedure and both coded the information independently. This allowed the assessment of inter- rater reliability of the Life Domain and Intervention Survey data. Throughout the last two terms of the course sequence, interviewers met regularly with supervisors to discuss any problems which might have developed with specific cases, to talk about coding questions, listen to tapes of interviews, check over data sets as they were completed and handed in, and continue to work on improving interviewing techniques. Each of the components of this process interview which were pertinent to this study will be discussed below. These include the SRD, Life Domain Survey and Intervention Survey. Self-Report Delinquency measure. The SRD measure which was utilized is included in Appendix A. The items inquire about a wide range of delin- quent behaviors that are frequently occurring. Very serious crimes that are infrequently committed are not included. The measure also includes several positive items to discourage response—set biases. The youth re- sponded to each item twice: first with the frequency of the behavior during 72 the past six weeks, and secondly with the frequency during the past year. The coded frequencies were, "never,“ "once," "twice," and "more than twice." The SRD was administered by the interviewer immediately following the open- ended interview. The interviewer read each item and recorded the youth's responses to eliminate error variance due to differences in reading skills. The measure was administered at the four time periods throughout the intervention period and also one year after project termination (by a dif- ferent interviewer, but following a similar open-ended interview). Blakely, et a1. (1980) reported that this measure generates one major scale. The rational subscales were found to be less reliable. These authors demonstrated the reliability of the total scale via internal con- sistency analyses and test-retest methods. Concurrent validity of the total scale was assessed by examining the relationship between it and official outcome measures (police and court records). The authors report correlations in the mid to upper 20's and conclude that the total scale "demonstrated strong validity properties." Life Domain Survey. The process interview was developed to provide a source of information about the impact of the intervention model on various areas of the youth's life (school, home, freetime, etc.). Specific information about these areas, combined with information about the nature and extent of the intervention activities could also provide information about contextual variables that might be affecting project outcomes. In short, these interviews were seen as tools which could be used to begin to examine the complex processes which underlie ”change" in human behavior within social contexts. This interview was devised specifically to meet these needs as they related to the ADP (Davidson, et al., in press). Davidson (1979) described the process of developing the interview, which took place during the second year of research at the Champaign-Urbana 73 setting. Initially, a handful of interviewers conducted the first process interviews with the youth participants and one of their parents three times throughout the course of their involvement. The general areas of interest were predetermined rationally. The item set was developed through a number of steps. First, an extensive list of potential items were generated rationally, based upon the first few interviews. Then, audio- tapes of interviews were analyzed to generate new items, ascertain appropri- ate response categories for the items and to check the "ratability" of them. Following this process, all of the interviews were coded with this new item set. Then the task of scale development began. It was accomp- lished according to the methods outlined by Jackson (1970). First, items were rejected that did not demonstrate adequate differentiation properties. All items with over 90% endorsement in one response category were eliminated. The second step was to group items into scales rationally. Then these scales were examined for internal consistency and convergent/discriminant validity. The scales were modified wherever possible to maximize these properties. Finally, a multi-trait/multi-method analysis was conducted to further examine the scales' reliability and validity properties. The codebooks which resulted are included in the appendices. The Life Domain Survey (Appendix 0) included 132 items which made up eleven scales. The Intervention Survey (Appendix E) had 146 items which composed thirteen scales. The Intervention Survey will be discussed further in the next section. Information concerning the eleven Life Domain scales in their final form are presented in Tables 7, 8, and 9 which are reprinted by permis- sion of Davidson and Redner (1982). Table 7 presents the content descrip- tion of each scale, an example of a scale item and the internal consistency alphas for each scale. The alphas ranged from .59 to .90. Table 8 74 presents the validity and reliability information obtained from the multi- trait/multi—method analysis. The results demonstrated the desired pattern of higher mono-trait/hetero-method correlations than mono-method/hetero- trait correlations on all eleven scales. Table 9 presents the inter- interviewer reliability data. In all instances the between-interviewer correlations were higher than the within—interviewer (between other scales) correlations. But the interrater correlations for some of the scales were lower than is desirable. They ranged from .19 to .81 with an average of .45. The correlations for three scales were particularly low: Parental Knowledge of Friends, Parental Knowledge of School and Legal System Involve- ment. The interrater correlations for the remaining eight scales ranged from .41 to .81 with an average of .53--a more acceptable range. These three scales were, therefore, excluded from the analyses conducted in this study. The Life Domain Surveys were conducted with the youths and a parent at four times during project involvement, as described earlier (at 0, 6, 12 and 18 weeks). Each interview focused exclusively on the past six weeks. Thus, the first interview provided information about the six weeks prior to project involvement and the second, third, and fourth provided information concerning the eighteen week period that coincided with the ADP intervention. During the years of interest, the completion rates of these interviews was 9 %, and was nearly identical across conditions. Only the data from the youth source were used for the analyses in this study. Intervention Survey. The scales of the Intervention Survey were created in an identical manner to that used for the Life Domain Survey. Information concerning the thirteen Intervention scales are presented in Tables 10, 11, and 12 (reprinted by permission of Davidson and Redner, 1982). II. III. IV. VI. VII. VIII. IX. XI. 75 Table 7 Life Domain Scales POSITIVE INVOLVEMENT IN THE HOME. The degree to which the youth spends time at home and is involved in activities with his parents and siblings, e.g., "How often does youth play indoor activities with parents?" and "How often does the youth eat dinner at home?“ oi=.78 PARENTAL CONTROL. The degree to which the parents make attempts, positive or negative, to control the actions or conduct of the youth, e.g., ”How much do the parents argue with the youth about his/her friends?" and "How often do the parents talk to the youth about changing?“ a = .72 POSITIVE CHANGE IN THE HOME. This scale was drawn from common com- plaints from parents and youth in the interviews and the content of those issues were included as change items, e.g., "How much has the youth's spending time at home in the evenings changed?" a = .79 INVOLVEMENT IN THE SCHOOL. The degree to which the youth went to school, did well in school, liked school, etc., e.g., ”How often does the youth attend school?“ and "How often do the teachers hassle the youth? a = .90 POSITIVE CHANGE IN SCHOOL. E.g., "What change has occurred in school attendance?" o = .85 ACTIVITIES WITH PEERS. The degree to which the youth engaged in activities with his peer group or friends, e.g., "How often does the youth spend time with friends on the weekends?“ o = .75 POSITIVE CHANGE IN FREETIME ACTIVITIES. E.g., "What change has occurred in the youth spending freetime constructively?" o = .59 PARENTAL KNOWLEDGE OF FRIENDS. The extent to which parents have knowledge of the youth's friends and his activities with them, e.g., "To what extent to parent(s) know who the youth's friends are?" a = .79 PARENTAL KNOWLEDGE OF SCHOOL. The extent to which parents have knowledge of the youth‘s school performance, e.g., "How many of the youth's teachers do the parents know of?" a = .63 EMPLOYMENT MOTIVATION. The extent to which the youth desires and initiates action to obtain a job, e.g., "How often does the youth actively seek employment?" 6 = .69 LEGAL SYSTEM INVOLVEMENT (high score means low involvement) The extent to which the youth has contacts with the police or juvenile court, e.g., "How often has the youth had contacts with the police lately?" a = .71 76 om. mm. mo. mm. mm. mm. mw. um. cm. cu. om. OH NO. ©F. ow mo. N_. 0» mo. mm. 1 0“ mo. mm. 0» mo. cm. 0» 00.1 q—. 0p <0. cm. i on we. mm. 1 0» moo.1 «F. o» No.1 RN. oe No. m_. .wmmmm .mmmz pcmema pwmeuiocmpmx\cozumeio:oz .Nmm_ .eoeeoe eee comee>ea seed ”meoz on _o.. oF. mm. ucmsm>~o>cH EmHmAm Foam; eo _o.- N_. _m. eo_ee>_ooz oeoesOPQEU on No.- efi. om. Foeeom\omeo_3oe¥ Feoeoeea on mo. em. ox. meeoPLE\omeopzoe¥ _eeeoeee on me. mp. Fe. oeeoeoeu\oeeeeo o>wommoe on Fo.- ON. ee. meeoa gee; mowow>weo< on No. op. me. _oo;om e_ omeeeo o>_oemoa - oo oo.. _N. me. _ooeom e? peoEe>_o>eH on _o.- ~_. a“. meal e_ emceeo o>woemoa co co. Fm. oe. Foeoeoo Feeeoeea 0» ~00. m_. _w. acmem>po>cH Mflmummwm .mmmmm .11111 .mmwm “edema ense> not: Logo» A.m_eeeaeee so_eepe>. ..o.Po cozpmeioempwz\uwmcplocoz meowpopmecou mFoom cpoeoo ween w mpnwk 77 Table 9 Inter-Interviewer Agreement Years 2, 3, 4, 5 Life Domain Within Interviewer Between Correlations Interviewers Range Scale Correlation Lg_ Hi_ Mean Positive Involvement in the Home .81 .05 -.57 .26 Positive Change in School .41 -.02 .25 .10 Activities with Peers .47 -.05 .33 .19 Parental Control .51 -.OO .42 .21 *Parental Knowledge of Friends .19 .Ol .38 .15 *Parental Knowledge of School .29 .01 .34 .19 Positive Change in the Home .54 .03 .34 .17 Involvement in School .44 .05 .40 .17 Positive Change in Freetime Activities .48 .02 .55 .23 Employment Motivation .56 .07 .52 .24 *Legal System Involvement .27 -.Ol .38 .18 (high score means low involve- ____ ____ ment) .45 _ .19 NOTE: These scales were excluded from analyses in the current study. 78 Table 10 presents a description of each scale, an example item for each, and the internal consistency alphas for each. The alphas ranged from .58 to .93. Table 11 presents the results of the multi-trait/multi-method analyses. Again, the desired pattern of higher mono-trait/hetero—method correlation predominated. Table 12 presents the interrater data. The overall picture of these inter-interviewer correlations was superior to that of the Life Domain Survey scales. These correlations ranged from .22 to .87 with a mean of .61. Three scales performed questionably in these analyses-~both in terms of demonstrating relatively low interrater correla~ tions and in the fact that these correlations were closely approximated or exceeded by the within interviewer correlation ranges. These scales were: Positive Involvement, Peer Involvement, and Legal System Involvement. These were excluded from the analyses in this study. Interrater correla- tions for the remaining ten scales ranged from .47 to .87 with an average of .70. The Intervention Surveys were conducted with each project youth, a parent and the volunteer three times-—at six, 12 and 18 weeks. The focus of the interview was on the past six weeks. Combined, then, the three interviews provided inclusive information about the entire intervention period. Again, as with the Life Domain Survey, only the youth data was used for this study. Data Modification Procedures for Process Measures As discussed in the preceding section, for the purposes of the current study, eight Life Domain Survey scales were utilized. These were: Positive Involvement in the Home Positive Change in School Activities with Peers Parental Control Positive Change in the Home Involvement in School Positive Change in Freetime Activities Employment Motivation 79 Table 10 Intervention Scales I. AMOUNT OF TIME. Frequency and amount of contact, e.g., "How much time does the volunteer spend working on the case?" a = .76 II. POSITIVE INVOLVEMENT. The extent to which the youth and the assigned volunteer get along and the lack of problems involved in the intervention process, e.g., "To what extent does the youth like the volunteer?" a = .87 III. PARENTAL INVOLVEMENT. The extent to which parent(s) are included in the intervention and the extent of a relationship built up be- tween the parent(s) and the volunteer, e.g., "How often does the volunteer talk with the parent(s)?" a = .81 IV. SCHOOL INTERVENTION: YOUTH FOCUS. Extent of the intervention focusing on school behavior of the youth, e.g., "To what extent is the volunteer trying to get the target to go to school more?" a=.87 V. SCHOOL INTERVENTION: SYSTEM FOCUS. Extent of the intervention focusing on bringing improvements to the school environment by focusing on school staff, e.g., "To what extent is the volunteer working on curriculum changes?" a = .78 VI. JOB SEEKING ACTIVITIES. Extent to which the intervention focused on getting the youth employment, e.g., "How much has the volunteer instructed the youth in job seeking?" a = .85 VII. FAMILY INTERVENTION: YOUTH FOCUS. Extent to which the intervention focused upon changing the youth within the family context, e.g., "How often does the volunteer talk to the youth about home?'o = .87 VIII. FAMILY INTERVENTION: PARENT FOCUS. Extent to which the interven- tion focused upon changing the parents' behavior in the family, e.g., "To what extent is the home intervention focused on improving the parents' household rules?" a = .78 IX. RECREATIONAL ACTIVITY. Amount of recreation involved in the time spent with youth by volunteer, e.g., "How often do the volunteer and youth do athletic activities together." a = .66 X. PEER INVOLVEMENT. Extent to which friends of the youth are involved in the intervention, e.g., "How often do the youth's friends spend time with the volunteer and youth?” a = .58 XI. LEGAL SYSTEM INTERVENTION. Extent to which the volunteer became *Tnvolved in the juvenile justice system for the youth, e.g., "Has volunteer assisted in negotiating a court disposition?" o = .83 XII. XIII. 80 Table 10 (continued) Intervention Scales ADVOCACY ACTIVITIES. Extent to which the volunteer intervened on behalf of the youth to gain needed resources, e.g., "To what extent does the volunteer take action to generate new resources (e.g., employment, new club) for the target?" a = .76 CONTRACTING ACTIVITIES. Extent to which the volunteer utilized behavioral contracting as an option in the intervention, e.g., "To what extent has the volunteer been instructing the youth and significant others in the methods of contracting?" o = .93 81 mm. op mo. Nm. mo. 0p mo. mm. mm. 0p no. mm. we. no. Fm. .m— mm. op 00.1 mm. om. op mo. mm. Rm. op mo.- mm. mm. pm. om. .Np mm. op po. op. mm. op Fo.1 o—. mm. op NO. OF. mm. Fm. mm. .Fp pm. op moo.- mp. om. op po.- op. mo. op No. mp. pm. om. mm. .o_ oo. op mp. mm. mm. op mo. om. oo. op co. .m. ow. mm. mm. .m om. op oo. mm. mm. op poo. om. ow. op pp. mm. mm. mm. om. .m pm. op No. mm. mm. op oooo.1 om. ms. op mo. pm. ox. ms. cm. .5 pm. op mo. pm. om. op poo. om. mp. op pp. om. mu. mm. ow. .0 mm. op moo.1 mm. mm. op 00. em. oo. op po.1 om. mm. pm. om. . .m Km. op 00. mm. mm. op po.1 mm. mo. op op. om. me. No. mm. .o mm. op mo. om. om. op mo.1 mm. mm. op moo.1 mm. pm. ow. mm. .m pa. op po. om. mm. op mo. om. oo. op 00. mm. mm. mm. mm. .N am. op po.1 pm. mm. op oo. mm. om. op oo.- mm. No. on. om. ._ a p82 madam awed. mam an mg g peoeea 111111111. .111111 11111. opp; opp: opp: Loopoopo> pcoeoo :poo> poocoo spoo> spoo> ppoep1ocopoz\oo;poe-oooz Aempocomopo proppo>= .o.pv oogpoeiocopo:\ppoepiocoz mopoom coppoo>copcp wo ooppopoecou FF open» 82 Table 12 Inter-Interviewer Agreement Years 2, 3, 4, 5 Intervention Within Interviewer Between Correlations Interviewers Rppgg Scale Correlation Lp_ Hj_ Mean Amount of Time .47 -.02 .46 .25 *Positive Involvement .22 ' .00 .36 .18 Parental Involvement .61 .06 .62 .34 *Peer Involvement .34 .01 .52 .26 Recreational Activities .57 -.02 .42 .21 Fpmily Intervention: Parental .75 .05 .67 .39 ocus F232;: Intervention: Youth .64 .16 .67 .42 Spgpfil Intervention: System .87 _.05 .61 .28 School Intervention: Youth .79 .04 .63 .33 Focus Job Seeking Activities .83 .03 .26 .13 *Legal System Involvement .40 .03 .38 .15 Contracting Activities .77 .12 .62 .36 Advocacy Activities ,70 .03 .56 ,3; .61 .28 *NOTE: These scales were excluded from analyses in the current study. 83 For each of these, each youth's scale scores from the first, second, third and fourth interviews were summed across time to yield one general measure for each scale. This combined score, it will be remembered, con- tained information about ore-intervention levels and also contained infor- mation about the during intervention scores which would indicate any improvements or detriments realized during this period. The combined scale scores yielded rough, overall indications of both sources of varia- tion. Table 13 presents the means and standard deviations of the resulting eight Life Domain scale scores. Ten Intervention Survey scales were also utilized in this study, as discussed previously. They were: Amount of Time Parental Involvement Recreational Activities Family Intervention: Parental Focus Family Intervention: Youth Focus School Intervention: System Focus School Intervention: Youth Focus Job Seeking Activities Contracting Activities Advocacy Activities The scores for these ten scales were also summed across time. Each youth's scale scores from the six, 12, and 18 week interviews were combined to provide an overall measure of the intervention activities which were carried out throughout the youth's involvement with the ADP in each area. Table 14 describes the resulting ten Intervention scale scores. The research questions presented at the end of Chapter I were to be addressed via discriminant analyses utilizing the contents of both the Life Domain and Intervention scales as predictor variables. The intercorre- lations of these scales were therefore of interest. Table 15 presents the intercorrelations of these 18 scale scores. As can be seen from this table, many of these variables were highly correlated with one another. 84 Table 13 Means, Standard Deviations and Ranges of Life Domain Predictor Variables £2919. Positive Home Involvement Positive School Change Activities of Peers Parental Control Positive Change at Home Involvement in School Positive Change — Freetime Activities Employment Motivation |>d 10. 12. 13. 12. 13. 13. ll. 19 .84 .76 ,31 .79 .29 .26 .05 b-b-b-b-D-D-D-b Possible Range .00-20. .00-20. .00—20. .00-20. .00-20. .00—20. .00-20. .00-20. 00 OO 00 00 00 OO 00 00 85 Table 14 Means, Standard Deviations and Ranges of Intervention Predictor Variables 3991.9. Amount of Time Parental Involvement Recreational Activities Family Intervention Parental Focus Youth Focus Family Intervention School Intervention System Focus School Intervention Youth Focus Job Seeking Activities Contracting Activities Advocacy Activities l>ppo< soeoo>o< mo. poop moppp>ppo< oeppoeepeoo po.- pomp moppp>ppo< oepxoom new mp. “App maooe spec» r 1 coppcm>gopcp poocom mp. Aopv wooed sopmxm a 1 coppco>gopcp Foosom No.- ”opp oeooe ep=o> 1 coppco>eopcp AFFEoJ mo.1 Aopv moood popcocoo 1 coppoo>gopcp appEou om. pmpv moppp>ppo< peeoppeeeome «8* PF. ANHV pcoeo>Po>cp popcoeoo mm. Appo oepp to peeoe< *8 mp.- Amoov coppesppoz peoesep85o .1 N . peony moppp>ppo< «_ weppomcu 1 mmcocu o>pppmoo mwp Aooov pooeom ep peoeo>po>ep No.- “wasp wee: pa ooeeeo o>pppmoe oo.- peony poopeeo pepeeeea mm.- Amoov meooa co moppp>ppo< mm“ Amado ooeeeo pooeom o>pppmoa Apoov peoeo>po>ep oaox o>pppmoo poo mopoopce> Lopopomco coppom>copcp one apogee opp; mo mcoppopogeougopcp m» m—noh 87 Fifty-seven percent of the correlations between the Life Domain scale scores were significant (p < .05). All 28 intercorrelations of these Life Domain scales ranged from -.23 to .89. The average of the absolute values of these correlations (which measure the average amount of association--either positive or negative-~between Life Domain scales scores) was .22. The intercorrelations among the Intervention scale scores were even more prevalent. Thirty-four out of the 45 Intervention scale inter- correlations, or 75% of them, were significant (p < .05). They ranged from -.13 to .83 with an average of the absolute values of these intercorre- lations of .32. The portion of Table 15 which presents the correlations between Life Domain and Intervention scale scores reveals that 45% of these correlations were significant. They ranged from -.25 to .60 and the average association between Life Domain and Intervention scale scores was .17 (again, using the absolute values). In order to reduce the problem which the intercorrelation of these variables would present in the discriminant function analyses, the Life Domain scales and the Intervention scales were both subjected to factor analyses. While a combined factor analysis approach could have been uti- lized, it was thought that the separate analyses of Life Domain and Inter- vention variables would address the most serious problems of intercorrela- tions while yielding the most conceptually concise factors. The Life Domain and Intervention analyses will be discussed separately below. Principal Components Analysis of Life Domain Scales The eight Life Domain scales were analyzed by means of a principal components analysis using the varimax rotation method to achieve orthogonal factors (Nie, Hull, Jenkins, Steinbrenner 8 Bent, 1975). The results are presented in Tables 16 and 17. Table 16 reveals that adherence to the convention of utilizing the number of factors which result in eigenvalues 88 of 1.0 or greater would have led to a decision to create three factors. However, the six factor solution was preferable due to the greater con- ceptual simplicity of the resulting factors. The six factor solution accounted for 93.8% of the total variance of the eight Life Domain scale scores. Examination of the communalities in Table 17 reveals that a large proportion of the variance of each of the eight variables was included in the six factor solution. Thus, the solution accomplished the creation of six orthogonal factors which represented theoretically meaningful com- ponents while retaining the majority of the information contained in the original eight scale scores. The six resulting factors are presented below. Life Domain - School. The variable loadings on this factor reveal that it primarily included information from the two Life Domain scales which measured the youth's school involvement: Positive School Change and Involvement in School. This factor, then, provided a general indication of the quality and quantity of the youth's school related activities. Life Domain - Positive Change. This component included, primarily, the two Life Domain scales which measured change: Positive Change in the Home and Positive Change in Freetime Activities. It, therefore, provided a general indication of the degree of change over the period from six weeks prior to ADP involvement to the end of their intervention 24 weeks later. Life Domain - Peer Involvement. This factor primarily represented the degree to which the youth was involved with peers as measured by the Life Domain scale Activities with Peers. It should be noted that the scale which measured Positive Change in Freetime Activities was negatively loaded on this factor. Life Domain - Parental Control. An examination of the factor loadings 'on this component reveal that it primarily included the Parental Control 89 m.mm o.om m.mm F.No “.mm m.Pm com oopcoooo< mocopco> eo u o>ppopoEoo . . . . . . sow ompcoouu< e N p a N op e ep N _N m pm ooeepee> pepop to a Nam. omN. ope. om_.p emo.p me.N oepe>eoopm cowpopom oopepoeez Lee moppmppepm . . . . . . Low ompcoooo< N mm m mm o NN m mm 0 ea o NN ooeepee> co p o>ppepee=o . . . . . . com oopczooo< m N_ m Np m Np m ep N pN N NN eoeepee> pepop to p pmo.p eoo.p eNo._ omp._ emo.p NNo._ ospe>eoopm . coppopom eopepoe Lop moppmppepm mNNoo.- pmmom. onmo. moeop. emooo. NoNeo.- coppe>ppoz peeesopeEM . . . .1111.. .11111. . moppp>ppo< oeppoooe mmmmp oNomp moNo_ onmo 1 Nmepo omNmp ep ooeego o>pppmoa NoNeN. omeo.- oomop.- ommpN.- pmmmo.1 NNoom. pooeom ep peoeo>po>ep Nopeo.- mNmeo. oNepo. ooemo. aNpom. NepNo. moo: pa ooeeeo o>pppmoa NoaNo.- Noooo. mNNNm. _oemp. emmmo. oemNo.- poepeeo Pepeeeea ooeoo.- meNmN. eeopm. Nempm. Nmeop. mppop.- stoma eo moppp>ppo< Nmpeo. mNomo.- mmmeo. mmNmo.- NNoop. poomm. omeezo pooeom ospppmoe 11.111 :9: NpNoa. NNmmo.- oNpmo.- moooo.- oONoo.- ONNNP. -o>pe>ep ego: ospwpmoa pcoem>Po>cp coppo>ppoz poepcou poms mmcogo poonom mmpnopgo> oEo: o>pppmoo pcoexoposm popcopoo 1o>Po>:H Loom m>pppmoo 1cposoo mwpp -epeeoo oepo -epeEOQ oepo -epeeOo apps -epeeoo oops -epeeoo apps mopoopco> Lopopomeo cpoeoo ompp mo upmzpoc< Lopood mo oppommm op opoop 90 Table 17 Communalities of Life Domain Predictors Rotated Principal Components Solution 6 Factors .3919 Positive Home Involvement Positive School Change Activities with Peers Parental Control Positive Change at Home Involvement in School Positive Change - Freetime Activities Employment Motivation .99197 .94799 .87232 .98327 .93689 .92898 .86639 .97748 91 scale. This scale, it will be remembered, measured the frequency of the parents' negative attempts to change or control the youth (i e., by argu- ing or utilizing punishments). Life Domain - Employment Motivation. This component represented, predominantly, the Employment Motivation scale from the Life Domain Sur- vey. It indicated the degree to which the youth wanted and was seeking employment. Life Domain - Positive Home Involvement. As before, this factor repre- sented primarily one Life Domain scale--Positive Home Involvement. This scale included items which measured the extent and type of activities that the youth engaged in with his/her parents and siblings and information about the amount of time the youth spent at home. Principal Components Analysis of Intervention Scales 'The ten Intervention scales were also subjected to a principal com- ponents analysis with varimax rotation. The results of this analysis are presented in Tables 18 and 19. Again, observance of the 1.0 or greater eigenvalue convention for the determination of the number of factors would have yielded three factors. But the six factor solution was again prefer- able due to greater conceptual simplicity. This solution accounted for 90.3% of the total variance of these ten scales. The communalities pre- sented in Table 19, again, reveal that the majority of the variance of each of the variables was included in this six factor solution. The six resulting Intervention factors are presented below. Intervention - Family. An examination of the variable loadings on this component reveals that it primarily reflected the two family inter- vention scales: Family Intervention-Parent Focus and Family Intervention - Youth Focus. It provided, then, an overall indication of the degree to 2 9 m.om m.mm m.m~ m.o~ N.wm m.mm n.v m.m ¢.m m.NP ¢.w_ w.mm moo. an. Pom. mmm.~ mmw.P Nmm.m m.om N.om m.mo o.mn o.o¢ «.mm m.o~ m.- m.m~ m.v~ o.mp ¢.m~ mmo.~ mo~.F mom.~ mm¢.P mmn.— mmm.m «comm. mommp. momma. momma. nmmmm. «ammo. Nommw. Pmpop. mowmp. mpmmo.1 mmmvp. memm. NPNNP.1 umNP.1 ommmp. mmomm. ommmo. mnopp. vmmmo.1 muoom. mmomm. mmmoo. ommvm. ovomp. ommmp. mmom0.1 ONNN—. mummF. pooom. mammo. mmmop. moooo. mfimuo. mmmo—. mommo. mmmmm. wwomm. wmoxo.1 momma. Novmm. omwmo. momma. oommo. ommom. mmmom. mwmuo. mummo. ommmo.1 mvomp. Pmmpo. ommmm. mmmmo. pwmom. oopmv. Nmmmo. muonm. ommmm. ampwp. awnpm. Nopmc.1 .Imcppoocpcou coppooeoom pooEo>Po>cp moooo>o< oco poozom anEom 1coppco>copop 1coppco>eopcp _op:mcoo oco ocpxoom goo 1coppco>copcp 1coppco>eopop we?» mo pcooe< 1coppco>cmpcp mopoopeo> Lopopooeo coppcm>gopcp mo mpmxpoc< gopuou mo oppommm 1coppco>eopcp mp mpam» com mooopso> yo a com oocopgo> oopopoecz Low com mocopgo> we a Low mocopgo> oopopom cop oopcoooo< w>ppopoe=u oopcoouo< popo» mo a oopo>cmmpm coppepom moppmppepm oopcoooo< o>ppoposou ompcoooo< popok mo N oopo>ooopm coppepom moppmppepm xoooo>o< moppooepcou moppp>ppo< Nepeoom new mooou spoo> coppco>cmpop poogom moood Eopmxm co wucm>LmucH Pogum wooed spoo> coppco>gopcp appsod moood popcopoo coppcm>copcp APpEom moppp>ppo< pocoppooeomm pcoem>po>cp popcmcoo weph mo pcooe< mopoopgo> 93 Table 19 Communalities of Intervention Predictors Rotated Principal Components Solution 6 Factors Amount of Time .86062 Parental Involvement .85971 Recreational Activities .96015 Family Intervention - Parental Focus - .89695 Family Intervention - Youth Focus .90407 School Intervention - System Focus .90343 School Intervention - Youth Focus .89285 Job Seeking Activities .92099 Contracting .97245 Advocacy .86031 94 which the intervention activities were focused upon improving the family situation. Intervention - School. This factor predominantly included the infor— mation combined in the two school-related Intervention scales-~School Intervention - System Focus and School Intervention - Youth Focus. It provided a general measure of the degree to which the intervention activi— ties focused upon improving the youth's school situation. Intervention - Job Seeking and Advocacy. This component reflected, for the most part, the Job Seeking Activities and the Advocacy Intervention scales. Since much activity which was coded in the Advocacy scale related to seeking employment for the youth, this factor may be thought of pri- marily as an indication of the degree to which the intervention activities were focused upon seeking the youth employment. Intervention - Amount of Time and Parental Involvement. As the name indicates, this factor represented primarily the Amount of Time and the Parental Involvement Intervention scales. Since both of these scales were significantly correlated with the majority of the other Intervention scales, and they were intercorrelated at a level of .50 (see Table 15), this factor can be interpreted as an overall indication of the intensity of the intervention that the youth received during project involvement. Intervention - Recreation. This factor indicated the degree to which the intervention included recreational activities. It primarily reflected the corresponding Recreation scale from the Intervention Survey. Intervention - Contracting. This component, as the name implies, reflected the Contracting scale from the Intervention Survey. Computer Factor Scores Standardized factor scores were computed for each youth for these six Life Domain and six Intervention factors. These twelve factor scores 95 provided the predictor variables for the discriminant analyses which will be discussed in Chapter III. Data Modification Procedures for Outcome Measures The discriminant function analyses, utilizing the twelve process factors described in the previous section were performed to examine the differences between youth outcome groups as defined in two different ways: The first method based group membership upon the archival police and court data. The second method defined groups according to self—report delin- quency scores. Each of these is discussed separately below. Archival Outcome Measures The archival police and court data indicated the frequency of police and court contacts for the time period from the point of project referral until one year after project termination. For the purposes of this study, this information was reduced to a simple dichotomous variable indicating either the absence or presence of contact with the juvenile justice system during this period of time. By one year after the project termination, 50 of the youths included in this study had pp$_recontacted the legal authorities, while the remain- ing 57 had had at least one such contact during this time. Self-Report Delinquency Outcome Measure . The Self-Report Delinquency (SRD) variable which was utilized in the present study was computed as a two-part sum. The first component in- cluded the youth's reported delinquency rates for the "past six weeks" from the interviews conducted at 6, 12, and 18 weeks into the intervention. These three, combined, would yield a measurement of delinquent behaviors during the entire intervention period. The second component consisted of the frequency of the youth's delinquent behavior during the year following project termination as measured by the youth's responses to the SRD 96 instrument at the one-year follow-up interview. This overall SRD score, then, provided a measurement of delinquent behaviors from the point of project referral to one year after project termination. This was the same period for which the archival police and court contact data were collected. The resulting SRD variable for the 107 youths included in this study, ranged from 0 to 143 with a mean of 48.98 and a standard deviation of 30.11. For the purposes of this study, this SRD variable was used to categorize the 107 youths into three equal groups. The Low SRO group included those youth with scores from O to 35. The second group, with moderate SRD rates, included scores from 36 to 54 and the High SRO group included scores over 54. Two separate discriminant function analyses were performed. The first utilized the two criterion groups created from the archival outcome data. The second utilized the three SRO-defined groups discussed above. The results of these analyses will be presented in the next chapter. CHAPTER III RESULTS Analyses The research questions set forth at the end of Chapter I were addressed by means of a series of discriminant analyses (Nie, et al., 1975). The Wilks' method of stepwise inclusion of predictor variables was utilized. This method bases the inclusion decisions on the greatest decrease in Wilks' Lambda. A variable was considered for inclusion into the discriminant function only if its partial multivariate F ratio was larger than 1.0 (Nie, et al., 1975, p .448). The first set of analyses explored the differences between the youth outcome groups whose membership was determined on the basis of archival data, as discussed in the previous chapter. Group I included youths who did not have any contacts with either the police or court from the point of referral to the ADP through a one-year follow-up period (a period of approximately 1% years in all). This group is subsequently referred to as the success group. Group II included all youths who had one or more con- tacts with the police or court during this time period. They are subse- quently referred to as the failure group. The next analysis explored the differences between three outcome groups that were determined on the basis of the youths' Self-Report Delin- quency (SRD) scores, as discussed previously. Group I included the third of the youths who reported the lowest delinquency rates during the 15 year 97 98 period of interest (the low SRD group). Group II consists of the middle third (the medium SRD group) and Group III consists of the highest scorers (the high SRD group). For both phases of the analysis, the six Life Domain factors and six Intervention factors described in Chapter II were used as predictor vari- ables. These were the "process-related“ variables. In addition, the youth's age at the time of project referral was included as a predictor since delinquency rates have been reported to vary systematically with age (Williams 8 Gold, 1978). Condition was also included since it has been found to be an important variable in determining archival outcomes in the ADP (Blakely, 1981; Davidson, et al., 1977; Kantrowitz, 1979). Conditions were transformed into a dummy variable--using the Low Intensity condition as the reference category (Nie, et al., 1975, p. 374). As discussed in Chapter II, the Life Domain factors were orthogonally related to one another-~as were the Intervention factors. However, the correlations of these two groups of factors across group were of interest-- as were the correlations of age and condition with the process-related pre- dictor variables. Table 20 presents the correlation matrix of all the predictor variables. Although a few of these correlations were higher than was desirable, the decision was made to maintain the conceptual clarity of the existing set of predictor variables without further modification or reduction. The results should be interpreted accordingly. In particular, one should not necessarily assume by a variable's absence in a function, that it is not an important discriminator if there are variables in the function that are highly correlated with it. With this caution in mind, the discussion will now turn to the results of the discriminant function analyses. The analyses relating to the two different types of criterion groups will be discussed separately in the next section. 99 o~.1 mp. «0.1 op. om< pN. NF.1 om. k. op.1 oH op. no.1 op. WW.1 .ONol 3.. 00. mp oo. mp.- mm. Np.- mm. oo. pN.- po. mp.- mN.- mN. mp.- No.1 Np.- mo.- mo.- MN.- mp.- mo. oo.- No. op.- _wm. op. pp. No.- No. Np. pN.- mp. pp.- op. owe- .om. op. x r .3. .11 oo. oo. oo. oo. eo. op.- mp.- oo. oo. oo. oo. ON. Np.- op.- po. po.- No.1 pp. po.- mo.- po. po. oo. .mm. op. we p0,- mo. po.- mN. Np.- pp.- .wm. pr 00. Fo.1 Po. op mp NH PH mop mop cop mo.1 00.1 No.1 mo. m—. .<. o_.1 mo.1 op. mo; mo.1 mp. No.1 o_. oo. op. mm. .1 ”o. po.1 Po. 00. N04 po. v gar mo. v an mN. pop peeoo mN.- pep pppEee mo. pap opemeoppapom se.- p

pmpcp .wm. pmpv coppoopooa1coppoo>gopop it. pepv Np. pomEo>po>cp popcogoo oco map» mo pcooe<1coppcm>popcp pmpv zoooo>o< oco mo.1 ocpxoom ooo1coppco>popcp mm. ANHV poonom1coppco>pmpcp k. mo.1 pppv appEou1coppco>popcp . poapv pooEo>po>cp po osoz o>pppmoo1cposoo ompp o.- pmoov eoppe>ppoz p pooEXopoEu1cpoeoo map; No. peony poppeoo popcopoo1cpoeoo oppp po. pmoop peoe 1m>po>cp pooo1cpoeoo.oppp po.- pNoov emceeo o>pppmoo 1 :poeoo ompp ppopv poosom1cposoa mmpp pop mopooppo> popopooco po xpppoz coppopoppou mooopw1cpgpp3 oopooo om opnoh 100 Discriminant Function Results Archival Outcome Groups Analysis with all predictor variables. Tables 21 and 22 present the results of the discriminant function analysis of the differences between the successes and failures as defined by the archival data. Table 21 presents the two groups' means on all predictor variables. Table 22 pre- sents a summary of the discriminant function which resulted from the analy- sis. Six of the predictor variables entered the equation, resulting in a Wilks Lambda statistic of .857. This demonstrated that only a little over 14% of the variance between the criterion groups was accounted for by the function. This suggested that these predictors provided a relatively poor model for the data. The discriminant function coefficients reflect the relative importance of the predictors to the differentiation of the two groups. The Action condition was the most important predictor. Examination of the group means in Table 21 reveals that the success group was more likely to be in the Action condition than was the failure group. The next most important predictor variable was the Intervention-Family factor. Successes were less likely than failures to report interventions in the family domain. Another important predictor was the Life Domain-School factor. The group means reveal that the success group reported more positive involvement in school than the failure group. Intervention in the school area was also demonstrated to be an important variable in the function. Here again, the successes were less likely than the failures to report this type of intervention activity. Another measure of the performance of a discriminant function is its ability to accurately predict the group membership of the original cases from which the function was derived. Table 23 reveals that this function 101 Table 21 Archival Success and Failure Groups Means of Predictor Variables GROUPS Variables Success (1) Failure (2) Life Domain—School .152 -.134 Life Domain-Positive Change .095 -.083 Life Domain-Peer Involvement -.054 .047 Life Domain-Parental Control -.143 .125 Life Domain-Employment Motivation .070 -.O62 Life Domain-Positive Home Involvement -.O69 .061 Intervention-Family -.182 .160 Intervention-School -.O73 .064 Intervention-Job Seeking and Advocacy .079 -.O69 '"SSTZ§2§1°1332132§..31 1.... 6"" ”107 -094 Intervention-Recreation -.026 .023 Intervention—Contracting -.002 .002 Age 14.400 14.088 Action Condition .520 .316 Relationship Condition .140 i .088 Family Condition .120 .228 Court Condition .060 .140 102 Table 22 Discriminant Analysis Sommary Table Archival Success and Failure Groups Standardized Canonical Wilks Discriminant Function Variable Entered Lambda Significance Coefficients Action Condition .957 .032 .631 Intervention-Family .922 .015 -.559 Life Domain-School .902 .014 1 .553 Intervention-School .881 .011 -.470 Life Domain-Change .869 .013 .363 Intervention-Amount of Time and Parental Involvement ’857 ‘015 -'328 Canonical Wilks Chi- Eigenvalue Correlation Lambda Sguare _QE Significance .16736 .3786350 .8566356 15.784 6 .0150 Group Centroids Group 1 - No Juvenile Justice System Con- .433 tacts Group 2 — One or More Juvenile Justice -.380 System Contacts 103 achieved the correct classification in 69% of the cases. It performed about equally with respect to the success and failure groups. Analysis with only process variables. The fact that the Action con- dition turned out to be the major discriminator in the preceding function led to a second discriminant function analysis which included only the process variables. This was done to ascertain the impact of some process variables on the archival criterion groups that the presence of the Action condition predictor could be masking due to problems of intercorrelation (see Table 20). Table 24 presents the results of the second discriminant function analysis performed on the two archival outcome groups. As before, six predictor variables entered the equation, resulting in a Wilks Lambda of .895. This function accounted for about 10% of the variance, which reflec- ted a loss of about 4% due to the exclusion of the Action condition varia- ble. This function did not achieve a .05 level of significance. A compari- son of the discriminant function coefficients in Tables 22 and 24 reveals that, except for slight ordering effects, the second analysis did not lead to any substantially different conclusions about the differentiation of the two archival outcome groups on the basis of the process variables. Fur- thermore, the exclusion of the condition variable resulted in an even poorer "fit" for the data. Self—Report Delinquency Outcome Group Tables 25 and 26 present the results of the discriminant function analysis of the three outcome groups as defined by SRD scores. Table 25 presents the group means of the predictor variables and Table 26 displays the discriminant function summary table._ The resulting solution consisted of two equations with ten predictor variables. Wilks' Lambda statistic for the functions was .502 (i.e., the functions account for 50% of the variance 104 Table 23 Capability of the Discriminant Function to Predict Archival Outcome Groups Actual Group Membership No further contact with juvenile justice system One or more contacts with juvenile justice system Percent of Cases Correctly Classified - 69.16% Predicted Group Membership 1 2 1 34 16 1 (68.0%) (32%) 2 17 40 (29.8%) (70.2%) 51 56 50 57 105 Table 24 Discriminant Analysis Summary Table Archival Success and Failure Groups (Including Only Life Domain and Intervention Predictor Variables) Standardized Canonical Wilks Discriminant Function Variable Entered Lambda Significance Coefficients Intervention—Family .971 .077 —.602 Life Domain-School .954 .084 .715 Life Domain-Change .938 .084 ' .515 Intervention-Amount of Time and Parental Involvement '920 ’07] 7'486 Intervention-School .904 .067 -.449 Intervention-Recreation .895 .079 -.343 Canonical Wilks Chi- Eigenvalue Correlation Lambda Sguare .QF Significance .11756 .3243376 .8948052 11.337 6 .0785 Group Centroids Group 1 - N0 Juvenile Justice System Con- .363 tacts Group 2 - One or More Juvenile Justice -.318 System Contacts 106 Table 25 Self-Report Delinquency Groups Means of Predictor Variables Variables Life Domain-School Life Domain-Positive Change Life Domain-Peer Involvement Life Domain-Parental Control Life Domain-Employment Motivation Life Domain-Positive Home Involve— ment Intervention-Family Intervention—School Intervention-Job Seeking and Advocacy Intervention-Amount of Time and Parental Involvement Intervention-Recreation Intervention—Contracting Age Action Relationship Family Court GROUPS Low(l) Medium(2) .117 -.052 -.151 .204 -.511 '.O77 -.387 -.O92 -.319 .031 .042 .138 -.117 .134 -.257 .100 —.l78 .198 .212 -.313 .129 .146 .236 .183 13.806 14.429 .361 .457 .111 .086 .222 .200 .111 .143 14 .176 .013 .159 .014 .091 .270 .414 .472 .417 .139 .111 .056 107 mmo. m me.mp 000. ON mom.wo .mpm .mm opoomm. 1pzu moo.1 mmm. mmm.1 ofim. mpm.1 Nmm.1 mmm.1 mmm. moo.1 pom. mmm.1 «mm. mp5. mop. Npo. ~w©.1 mwm. mmw.1 Pom.1 ovm.1 N :oppocod p ooppocod mpcopopmmoou coppocod poocpeppompa poopcocou omNpoLoocopm oemommo. p oNNNm. mNppNom. o mopeo. oooEop ooppocom coppopoppou moppz eopp< peopeeeeo ooo. Now. ooo. mpm. ooo. on. ooo. pem. ooo. mmm. ooo. pom. ooo. mpo. ooo. mmo. ooo. opp. ooo. eeo. ooeeopppeopm menses moppz moooeo Aooooooppoo ppooom1ppom oo.m~ oFFNF. N mm.om Nmoon. F oocoppo> oopo>cmmpm coppocod we & om< moppooppco01coppco>popcp omcocu o>pppmoo1cpoeoo oppp poms 1o>po>cp mEo: o>pppmoo1cpoeoo oppp xppEodiooppco>popcp zoooo>o< ooo mcpxoom ooo1coppco>popcp poo2o>po>cp popcopoo oo< ospp1coppoo>popcp coppo>ppoz pcoexopo5m1cpoeoo opp; pocpcou popcogoo1cpoeoo opp; pooeo>po>cp cooo1cpoeoo opp; oopopcm opoopeo> opoop xeoEE:m mpmxpoc< poocpeppompo om mpnoh 108 in SRD scores). This suggested that these data were reasonany well described by this set of predictor variables--un1ike the archival data. The significance level of the solution was less than .001. The first function accounted for 80% of the explained variance, while the second one accounted for about 20%. The second was only marginally significant (p = .073). Table 27 describes the function's ability to dis- criminate between the three groups. The between group F's which described the differences between the high SRD group and the other two were both significant at less than .001. The between group F which reflected the difference between the low and medium SRD groups was significant at less than .05. Table 27 also displays the two functions as they are represented by the group centroids. The first function-~which accounted for the majority of the variance--appeared to discriminate between all three groups very well. While the second function performed less well. There were three predictor variables that were heavily loaded in the first function. The first of these was the Life Domain-Parental Control factor. Table 25 reveals that the low SRD group scored low on this factor; the medium SRD group, moderately; and the high SRD group reported the highest level of Parental Control. (This factor, as discussed previously, primarily measured the frequency with which parents engage in arguments or punishments to control their children.) The second most important pre- dictor variable in this function was the Life Domain-Peer Involvement factor. Again, the scores on this factor co-varied with the SRD scores, with the low SRD group reporting low Peer Involvement and so on. The third important discriminator was the Life Domain-Employment Motivation factor. As with the previous two predictors, the youths reporting the lowest SRD scores reported the lowest desire for employment, the high scorers 109 Table 27 Group Centroids and Between-Group Discriminant Ability of the Functions - Self-Report Delinquency Groups Gpppp Group Centroids Function 1 Function 2 Low SRD (l) .940 .335 Medium SRD (2) .124 “—.582 High SRD (3) -1.06l .231 Between Group F-Statistics and Significance Levels Group 1 Group 2 Group 2 2.4404 .0123 Group 3 6.5994 3.3471 .0000 .0009 110 reported the highest, and the middle group fell in between. These rela- tionships will be discussed at greater length in the next chapter. Table 28 presents the capability of these disciminant functions to accurately predict the group membership of this sample. Sixty-four per- cent of the cases were predicted accurately. The functions predicted the high and low groups very well (75% and 72% accurately) but performed poorly with the middle group (54% were classified incorrectly). These figures reflect the fact that there was a great deal of overlap between these three groups, even though the group centroids were well differentiated from one another. These results will be discussed at greater length in the next chapter. 111 Table 28 Capability of the Discriminant Function to Predict Self-Report Delinquency Outcome Groups Actual Group Membership Predicted Group Membership 1 2 3 Low SRD 1 26 7 ' 3 36 (72.2%) (19.4%) (8.3%) Medium SRD 2 10 16 9 35 (28.6%) (45.7%) (25.7%) High SRD 3 4 5 27 36 (11.1%) (13.9%) (75.0%) 4O 28 39 Percent of cases correctly classified - 64.49% CHAPTER IV DISCUSSION Three research questions were posed in Chapter I concerning the rela- tionships between delinquency and several process variables which were operative within the lives of the ADP participants while they were involved in the intervention. These specific questions were: 1) Which process variables were more strongly associated with offi- cial recidivism and which with self-report delinquency? 2) To what extent did a combination of these process variables, derived from a discriminant function, explain the variance of these two measures of delinquency outcome? How successfully did they predict delin— quency outcomes for the same population? 3) How did the function derived for the archival groups compare with those derived for the SRD groups? How dissimilar were they? The two separate discriminant functions which were reported in Chapter III have been compiled in Table 29 for comparison. Each of the three re- search questions will be discussed below. The first two will be discussed-- first in regard to the official outcome groups and then in regard to the Self-Report Delinquency groups. Lastly, the third question will be addressed. Results-~Official Recidivism Groups Important Predictors The discriminant function analysis of the two officially determined groups led to a characterization of the "success" group as: 112 oooo. ompo. mochmecmmm 113 mNppNOm. ommoomo. messed mxppg coppocod www.mp Nmm.om 11111 opp: oopopoomm< oocop -La> =eoepepoxm= p6 p omo. map. pppu1 oep. ooo. eopppoeoo peeoo ppp. oom. mmm. mmm. omp. oopppocoo A—pEMJ amp. owe. ppp. mmo. oop. copppocoo opcmcoppopom Npe. Nme. pom. ope. ONm. p wepmo. copppoeoo eoppo< Npe.ep mNe.ep oON.mp moo.ep ooe.ep a eommo.- p epNNN. om< ape.- mop. omN.. Noo. Noo.- o pONNN.- m mprN. oeppoeepeeo-.p oNN.- esp. oNp. NNo. oNo.- coppeoeoox-.p :mEm; o>: c :me poo. mpm.- NpN. eoo. pop.- p emopp. op mpmep. o NNNNN.- wee oewp p“ wewoe<-ep .- . .- .- . .- . poeoo>o< epo map opp moo mpo e ooomm e peeNm eee mepxoom eon-.p amp. oop. NmN.- eoo. Mpo.- e oeope.- pooeom-.p mpo.- amp. ppp.- cop. Nop.- N omooe.- m oepom. N ponm.- pppEed-.p .1 . . . .1 .1 . pomso>po>cp opp Nmp Neo poo moo m Noopm o pNoNN oeo: o>pppmea-.o.o . . .1 .1 . . .1 copp ooN pmo mpm Noo epo op poppo m poooo -e>ppoz peoEpopeEW-.o. ope. Nao.- Nom.- mNp. Nep.- N NoeNN. p Nmmmo.- peepeeo pepeopee-.o.o Nme. ppo. ppm.- peo. emo.- m Nmoem.- N eomep.- peoEd>po>ep eooa-.o.o Neo.- eoN. pmp.- meo.- Nmo. o moopm.- o NONMN.- m NNNoN. omeeeo o>pppmea-.o.o Noo.- Nmo.- Npp. emp.- Nmp. m eonm. pooeom-.o.o 1pmv pr «pp pr ppv Ne pp pp opeepee> saw: Um: 301— mmpzzpmm mwmmmouzm W :Owuuczm W cowuuczu W corpoczm % a HUM 0mm W mam W4 _.Q>_.:u..p< u U U 5 Ru 6 mzpcop< pop coppocod poocpeppompo do coopLoQEou mm w—nMH 114 1) relatively more frequently involved with the Action condition than with any of the other experimental conditions. 2) less likely than failures to be involved with interventions focused on the family domain (Intervention-Family). 3) more positively involved in school (Life Domain-School). 4) less likely to be involved with interventions in the school domain (Intervention-School). Although mmaother predictor variables entered into the function (Life Domain-Positive Change and Intervention-Amount of Time and Parental Involve- ment) they each accounted for only about 1% of additional variance and were, therefore, not considered to exercise important discriminating abili- ties. An examination of the discriminant coefficients in Table 29 reveals that the Action condition was found to be the single most important pre- dictor of success-group membership of all the predictor variables. As will be remembered, when this variable was removed from the equation, a signifi- cant solution was not achieved. This finding, of course, was not unexpected. The relative effectiveness of the Action Condition in achieving success (as defined by the absence of contacts with the police or court) for youthful offenders was well documented previously (Blakely, 1981; Davidson, et al., 1977; Davidson 8 Redner, 1982; Kantrowitz, 1979) as discussed in Chapter I. The current results re-emphasize this finding in a multivariate sense. The three other important predictor variables (Intervention-Family, Life Domain-School, and Intervention-School) which discriminated between the two groups had also been examined previously in univariate analyses. These findings, some of which were presented in Chapter I, are of interest to the current discussion. The second most important discriminating vari- able in this equation, the Intervention-Family factor, measured the extent 115 to which the intervention focused on changing the home domain. The group means revealed that the success group reported less of this type of inter- vention activity. However, earlier analyses reported by Davidson and Redner (1982) place this finding in a clearer perspective. These research- ers reported strong condition effects with the Family Intervention-Parental Focus scale. Specifically, the Court condition was found to be the only condition which reflected a difference between the success and failure groups similar to that found in the study (i.e., the failure groups reporting relatively mp:p_family-focused intervention activities than the success group). With the Family Intervention-Youth Focus scale, Davidson and Red- ner report that the Court, Low Intensity and Relationship conditions all reflected similar differences between the success and failure groups (al- though nonsignificant). In light of these earlier findings, it is likely that the group differences found in the present analysis do not reflect a pattern that is common to all conditions equally. It is also likely that the observed differences between the success and failure groups on this factor reflect an increase in family intervention activities subsequent to a youth's reinvolvement with the juvenile justice system. The current analyses, of course, do not permit a closer examination of the causal underpinnings of the observed association. The third most heavily weighted predictor in the discriminant function was the Life Domain-School Involvement factor. This demonstrated a greater involvement in school on the part of the success group-~another replication of earlier findings concerning the ADP (Blakely, 1981; Davidson, 1979; Davidson, et al., 1977; Davidson 8 Redner, 1982), and also more generally (Elliot 8 Voss, 1974). Elliot and Voss found a weak relationship between academic achievement and subsequent delinquency (as measured, however, by SRD). They also reported that normlessness in the school domain (as C) ('1’ 116 measured by dislike of school, reported frequency of "skipping," getting into trouble with school officials, etc.) was a predictor of subsequent delinquency (Elliot 8 Voss, 1974, p. 151). Their data demonstrated, how- ever, that the causal connection between school involvement and delin- quency was a reciprocal one—-i.e., delinquency contributed to alienation with- in the school domain and predicted future academic failure as much or more than it was a consequence of these phenomena (p. 135). This again points out that the causal relationship underlying the observed association be— tween school involvement and official delinquency in the current study can- not be assumed. It is possible that involvement in the juvenile justice system leads to disengagement in the school domain. Subsequent studies will have to explore this possibility. The last meaningful discriminating variable in this function was the factor which measures school intervention activities. Again, as with intervention in the family domain, the success group reported fewer such activities. This also replicated earlier findings reported by Davidson and Redner(l982). This relationship may reflect the fact that more of the success cases were positively involved in the school domain and were, therefore, not in need of intervention efforts in this area. Alternatively, this relationship could, again, reflect an increase in intervention activi- ties in this domain subsequent to and reacting to further involvement in the juvenile justice system. These results, considered together, lead to several conclusions. First of all, the Action condition was again demonstrated to be an effec- tive approach to helping youthful offenders avoid future involvement with the juvenile justice system. The current study did not lead to a clear understanding of this efficacy. None of the intervention components which might have been expected to account for this difference performed as 117 important group discriminators in this analysis (e.g., contracting or advocacy). Secondly, the school area appeared to have an important bearing on official delinquency. Although these results did not lead to a clear understanding of this relationship, they suggest, at the least, that change agents in other settings who work on behalf of youthful offenders might be well advised to attend to the youth's relationship to school. Finally, these results suggest that the decision to combine all the ADP experimental condi- tions into one discriminant analysis may have resulted in misleading conclu- sions and that a condition specific approach might have been preferable in achieving a clear understanding of the relationships between intervention pro— cesses and outcome. Evaluation of Predictive and Explanatory Capabilities The discriminant function derived from these predictor variables did not perform well in regard to the archival outcome groups. Only 14% of the observed variance in official outcomes was explained by this combination of variables. In addition, the function did not predict group membership very accurately. These two facts indicate that whether or not a youth was contacted by the police or court was primarily determined by factors that were not adequately reflected in these predictor variables. Perhaps since official outcome can be viewed as largely a function of official behaviors (by police and other agents of social control), this is not surprising, since these predictor variables measured youth behaviors, primarily. The Self-Reported measure of delinquency, which may be thought of as being more directly related to the youth's behaviors, was, in fact, more profitably explored via the discriminant function analysis--as will be discussed in the next section. 118 Results - SRD Groups Important Predictors The discriminant function analysis of the three Self-Report Delin- quency groups highlighted the following important group differences: Life Domain—Parental Control factor. The low SRD group members were much less likely to report negative coercive interactions with their parents than the middle group--who in turn was less likely than the high scoring group to report such interactions. The low SRD group reported fewer argu- ments, and fewer incidents of their parents utilizing aversive controls. In short, these youths reported that their parents were less dissatisfied with them and that there was less friction in these relationships. Davidson and Redner (1982) reported the same relationships in regard to the archival success and failure groups--as reported in Chapter I. In their report it appeared that this relationship became stronger in a predic- tive sense-—that is, the group differences on the Parental Control scale (which measures the parent-youth relationship during project involvement) were larger when the groups were determined by two-year follow-up archival outcome than when they were based on one-year follow-up data. And the rela- tionship did not appear when based upon official recidivism gppipg the inter- vention period. This finding and that of the current study strongly sug- gest that this variable was an extremely important factor in the development of delinquency among these youth. These results strongly support Richard Stuart's (1973) position that, "the family plays a critical role in the etiology of delinquency when certain dysfunctional family interaction pat- terns coexists with a paucity of opportunities for acceptable performance in the community" (p. 335). Stuart reports both that families of delin- quents are more likely to exhibit coercive interaction patterns and that they are less likely to utilize positive controls to influence their children. 119 He suggests that behavioral contracting can be a useful method of revers- ing this pattern and extending the constructive, positive influences of the family. In regard to this, it is interesting to note, in Table 29, that the SRD group» means f0r the Intervention-Contracting factor provided further support for this contention. (Although this factor entered the function as a significant predictor, it accounted for only a little over % of the variance between SRD groups and therefore cannot be said to have been an important predictor.) The relationship between parental dissatisfaction and SRD which was demonstrated by these data is consistent with findings of other researchers. Conger (1976) found that the incidence of parents' reinforcing behaviors and the frequency of positive communication wenenegatively related to de- linquency. He further reports that as the incidence of parental punish- ment increases, this inverse relationship between parental reinforcement and delinquency disappears. Conger concludes that parental response behaviors--whether rewarding or punishing--are important determinants of the bond between the youth and parent. He views this bond as an important potential protection against delinquent involvement. Of Conger's findings vis-a-vis this important parent/child relationship, Austin (1977) states, "mutual acceptance is associated with least delinquency and mutual rejec- tion with most delinquency" (p. 112). The findings of the current study were consistent with this assertion. Life Domain-Peer Involvement factor. The low SRD group also reported much less involvement with their peers and more positive change in the freetime domain than did the middle or high scoring groups. This finding also replicated those reported by other investigators in the area of delinquency. Elliot and Voss (1974), for instance, reported that among 120 their sample of youths, as measured by self-report, "commitment to peers has a strong, direct effect on delinquent behavior" (p. 165). Conger (1976) examined the question of peer involvement as it related to delin- quency in a large sample of white male urban junior and high school stu- dents. He found no overall relationship between attachment to peers and self-reported delinquency rates in his sample. He suggested that this could be understood in light of social learning theory. Social learning theory would predict that peer involvement would lead to increased delin- quency only if a youth's peer group was reinforcing delinquent behaviors. Involvement with a conventional or pro-social peer group would be expected to lead to decreased delinquency. In a large, mixed sample of youths, one would expect to find no generalized relationship between peer involve- ment and delinquency. Conger would maintain that the results of the cur- rent study suggest that the "peer groups" of these youths are attending to and reinforcing delinquent behaviors and that the higher level of peer involvement of the high SRD group would therefore be expected. Conger further suggests that the influence of delinquent peers (of the rewards they offer for delinquent behaviors) will be greatest in a youth's life in the absence of sources of reinforcement for pro-social behavior. He predicts that, "delinquency will be greatest where attach- ments to conventional environments are weak and attachments to deviant environments are strong." The high SRD groups reported negative relation- ships with their parents (Life Domain-Parental Control), low involvement at home (Life Domain-Positive Home Involvement) and high peer involvement, as measured by this Peer Involvement factor. In combination all provide strong support for Conger's position. Life Domain-Employment Motivation factor. The low SRD group was also less likely to report high motivation for employment than the middle group 121 (who scored moderately on this factor) or the high group (who reported high motivation for employment). This effect might, in part, be a reflection of the fact that the low SRD group tended to be somewhat younger than the other two groups, since Table 20 indicates that age and employment motivation were correlated. This effect may also stem from the fact that the low SRD group was relatively more positively involved at home and school than were the other two groups. .Perhaps if these other important, traditional youth roles became more dissatisfying or aversive (as they appeared to be for the other two groups) then employment was seen as a more important goal-~a kind of substitute arena in which to prove oneself successful. Intervention--Job Advocacy factor. Closely related to the above pre- dictor, the low SRD group reported the lowest frequency of intervention focused on obtaining employment and other "advocacy" interventions. The middle SRD group reported the highest incidence of these types of inter- ventions and the high SRO group mean fell in between the other two groups. As stated above, the low SRD group did not appear to want employment, so their report of little intervention in this regard was not surprising. But both the middle and high SRO groups did report high desire for employ- ment--particularly the high SRO group. The relatively low score on this factor for that group was therefore perplexing. These data did not provide any direct explanation of this observed group difference. It is possible that these high SRD youths were involved with crises in other areas of- their lives which became the focus of the intervention activities, but that is not clear from this study. It is, perhaps, equally as likely that the relationships between these youths and the students with whom they worked were less positive than those in the other groups and that the students were less likely to take proactive steps toward goals which were 122 desired by their youths. Further investigation into this area is neces- sary to gain a clearer understanding of these results. Intervention-Amount of Time and Parental Involvement factor. Table 29 reveals that this predictor, while not heavily weighted in the first function--was the most important predictor in the second function which discriminated the SRD groups. Although this second function was not very successful in discriminating the three SRD groups, this predictor entered the overall discriminant function procedure at the fourth step and accounted for an additional 4% of the observed differences between the three groups. The group means revealed that the low SRD group reported the highest level of involvement with their student change agents, as measured by this factor. The middle SRD group reported the lowest group mean here and the high SRD group fell in the middle. The high score of the low SRD group might have been due to the assigned students finding it relatively easier and more satisfying to work with this group. They may have been a more rewarding and responsive group to work with since they tended to be more positively involved at home (which would have made it easier to involve their parents in the intervention) and at school. They also were not engaging in delinquent behaviors as much as the other groups and this fact perhaps resulted in relatively greater rewards for the students who worked with this group. The situation which sometimes faced those students in the other groups was that they became aware that their youth was continuing to engage in delinquent behaviors and they felt less "successful" or effective themselves as a result. The difference between the medium SRD and the high SRD group was more puzzling--particularly the very low score of the medium SRD group. It may be that the high SRD group provided more apparent, clear-cut issues 123 for the students to work on than did the middle group. Furtherinformation would be necessary to obtain a clearer understanding of these group dif- ferences. Intervention-Family_factor. The group means on this factor revealed that the low SRD group reported the least amount of intervention in the family domain, the medium SRD group reported the highest amount, and the high SRD group fell in the middle. This may reflect that the low SRO group was not in need of interventions in this area and therefore received less. The medium SRD group, however, reported high involvement in the home (see the Life Domain-Home Involvement factor means) and a moderate amount of stress in their relationships with their parents (see Life Domain- Parental Control). Under these circumstances, there would have been an increased need for interventions with the family. The high SRD group was relatively uninvolved in the home domain and their relationships with their parents were very stressed. Under these circumstances it would have been relatively difficult to get the youth and parent engaged in family intervention activities. Conclusions-~Findings Discriminating the SRD Groups The remaining four predictor variables which entered the discriminant function accounted for approximately 1% of the variance each. They, there- fore, cannot be viewed as important discriminating variables in regard to the SRD groups. The overall picture of the high SRD group presented by the group means of the predictor variables was not a new one. This group reported nega- tive relationships with their parents governed primarily by coercive con- trols. They were highly involved with their peer group and relatively less involved at home. They were not positively involved in school. They wanted employment but were most frequently not employed. Davidson (1979) 124 reported similar findings in regard to the “failure" youths who partici- pated in the ADP in Illinois. In that instance, the group was determined by police and court recidivism and poor school attendance levels. Elliot and Voss (1979) also reported similar relationships between many of these “life domain” variables and delinquent behavior in their sample of 2,617 metropolitan high school students. It is a picture which is consistent with the predictions of Conger's (1976) synthesis of social control and social learning theories of delinquent behavior. The youths who reported high levels of delinquency in this study also reported relatively weak attachments to conventional roles and environments (school, home) and they also reported strong attachments to their peers. Conger would suggest that these peers were very likely to be highly delinquent themselves and that the youths' commitment to their peers were most probably very strong as well. These results are similarly consistent with the related theory of "differential opportunity" put forth by Cloward and Ohlin (1978) which united the concept of anomie (differential access to legitimate means of obtaining desired ends) with "delinquent subculture" theories which sug- gested the idea of differential access to illegitimate means of obtaining desired ends as well. This theory would also predict the pattern of "life domain" components reported by the high SRD group in this study. The fact that the Intervention predictors played a relatively minor role in these discriminant function solutions is somewhat discouraging. However, it is not surprising, since the previous research on the ADP did not demonstrate condition effects in regard to Self-Reported Delinquency rates. The Life Domain predictors did provide strong support for the rationale behind the inclusion of the behavioral contracting approaches utilized in the Action model. They also provided support for advocacy 125 interventions ahned at achieving desired goals in the employment and school areas. Although the importance of the peer involvement factor in discriminating the SRD groups might be viewed as an indication that peer relationships should become a focus of intervention efforts, the most productive approach in this area is more likely to be an indirect one. Even in an intervention model which is as deeply embedded in the youth's natural environment as is the ADP, peer group relationships are difficult and relatively inaccessible targets for change. Conger's (1976) findings, however, suggest an alternative, indirect strategy for addressing the concern about the influence of delinquent peers. He maintains that delin- quent peers are primarily influential with youths who are not receiving adequate reinforcements from more conventional interpersonal situations (specifically the parents). He suggests that a more positive parent/child relationship can serve as an important protection against deleterious involvement with delinquent peers. The suggestion, then, is that since youths who are more involved with and committed to peers generally are so because they receive a greater amount of positive attention from these peers than they do from their parents, the most productive solution is to increase the amount of reinforcement available to the youth from the parent(s). This approach focuses on strengthening the positive bond be- tween the youth and the parent, thereby increasing the importance of that relationship to the youth. The resulting more positive parent/child involvement (or home involvement) has repeatedly been shown to be negative- ly related to delinquency (Conger, 1976; Davidson, 1979; Elliot 8 Voss, 1979). The relatively minor role that intervention activities played in discriminating the SRD groups could be due to a number of factors. First 126 of all, it is possible that the merging of these conditions masked some of the intervention relationships that might exist in some of the condi- tions. The Action group, in particular, might yield stronger intervention/ outcome relationships in a discriminant function analysis when viewed alone. Another possibility is that the Life Domain and Intervention inter- views conducted with the youths did not measure the intervention activi- ties with enough precision to explore the relationships between these activities and SRD outcomes (even if such relationships did exist). This was of particular concern in regard to the advocacy portion of the inter- vention. Perhaps the volunteer interviews would have provided a more accurate or complete measure of the intervention predictor variables. Another likely possibility is that the interventions supplied by the ADP to the court referred youth--although very "intensive" in comparison to most traditional intervention approaches--was simply not an important enough influence in the youth's life to affect the desired changes in de- linquent behaviors. Finally, it is possible that the intervention packages offered by the ADP were "off target" in some important ways and simply did not achieve changes in areas of the youths' lives that would need to be changed in order to decrease delinquent behaviors. Future studies-— specifically of the Action condition-~would be necessary to begin to unravel this puzzle. One profitable approach would be to carefully examine the interventions that youths who gig demonstrate decreases in SRD rates received and attempt to differentiate these interventions from those of the other youth. Evaluation of Predictive and Explanatory Capabilities The discriminant functions for the SRD groups provided a relatively good model to account for SRO variance, accounting for 50% of the observed 127 variance. An important caution must be born in mind in regard to these results, however. Both the SRD data and the predictor variables were ob- tained through interviews. Under these circumstances, method variance would be expected to inflate the "variance accounted for" by these func- tions. Even in light of this caution though, these results can be said to indicate that the ten predictors included in the equations were impor- tant factors in accounting for the delinquency rates of these youths as measured by SRD. It is also true, however, that 50% of the variance in SRO group membership was pp; accounted for by this combination of predic- tor variables. So clearly, several other variables which are important in understanding the etiology of delinquency were pp; included in this set (for instance, the level of delinquent activity among a youth's peer or family group might be such variables). The predictive ability of these discriminant functions--although superior to chance--was not very accurate and yielded a particularly high percentage of "misses" in regard to the middle group. So these groups were not clearly differentiated from one another by the functions. In short, these derived functions were more productive in an analytical, exploratory sense than in a predictive one. Comparison of the Two Discriminant Functions An examination of Table 29 yields a very clear-cut answer to the third research question of whether or not the derived discriminant func- tions for the SRD groups and the archival groups were similar. The answer is that they were not at all similar. In fact--and almost unbelievably, given such a relatively small number of predictors--of the top four pre- dictors of each function there was absolutely no overlap (i.e., no common, important predictors). This provides clear evidence that "delinquency" as measured by archival data and "delinquency" as measured by SRD are two 128 very distinct phenomena. These are clearly pp§_two ways of measuring the same construct, but, in fact, measure separate constructs--both of which are valid and important to the overall phenomenon of "delinquency." This is the position put forth by Blakely, et al. (1980) and Fernandez (1981). It receives strong confirmation from these results. The current study does not lead to any clear understanding of the observed differences between these two indices of delinquency. It does appear that the discriminant function which described the officially de- fined groups included several factors which measured strong environmental reactions to a youth's involvement in the juvenile justice system. Specific- ally, the observed group differences on the Life Domain—School factor, and the two Intervention factors which discriminated between the two groups (Family and School) could all be interpreted as differential responses to a youth's re-entry into the juvenile justice system. It is not clear whether or not this was the case since all of these measures were collapsed across time. A time series analysis of these data would lead to a clearer understanding of this matter. Conclusions This study attempted to explore the relationship between the processes of change and delinquency outcomes in the context of the Adolescent Diver- sion Project, a diversion program for juvenile offenders. Delinquency outcomes were measured by two distinct methods: Archival data concerning police and court contacts and a Self-Report Delinquency measure. The pro- cess measures included indices of intervention activities and general life domain conditions operative in the youths' lives. The discriminant analyses performed led to several major conclusions. First of all, self-report delinquency and official delinquency were demonstrated to be two distinct constructs by the fact that the derived 129 functions which described the various groups in the two instances had virtually nothing in common. The process variables that discriminated groups of SRD respondents did not discriminate between the officially de- fined ”success" and "failure" groups and vice versa. Secondly, these process variables which were utilized as predictors provided a very poor model for the explanation of official delinquency. They accounted for very little of the difference between outcome groups. Self—Reported Delinquency was more adequately explained by this particular set of predictor variables—-although there was certainly a great deal of unexplained variance (50%) even here. A comparison of the predictors which were determined to be important with respect to the official outcome groups and those with respect to the SRD groups leads to a perplexing situation. The successful official out- come group was, in part, predicted by participation in the "Action" inter- vention condition--the SRD groups were not related to condition. Yet the Life Domain predictors in the SRD functions provided support for the specific intervention components of the Action model (e.g., the importance of the Life Domain-Parental Control factor provides strong support for the inclusion of behavioral contracting). But the Action model did not account for any of the variance in group membership among the SRD groups. This leads to the inevitable question of how to account for the Action model's impact on official recidivism when official delinquency was not demonstrated to be related to the specific intervention components as measured by the Intervention Survey. This study does not provide the information necessary to answer this question, although some of the results suggest that impor- tant steps toward achieving a clearer picture would be to examine the Action condition youths separately and to examine the data utilizing a time series approach rather than collapsing across time as was done in this study. 130 Two final, and more general conclusions can be made on the basis of the results of this study. First, the observed distinctions between official and self—reported delinquency emphasize the need to include pppp meaSures to evaluate social programs related to delinquent--or pre- delinquent--populations. Both measures of delinquency are valid and important ones. Delinquency programs, such as the ADP, should attempt to tailor their intervention efforts to maximize a decrease in pppp_types of delinquency. Secondly, this study clearly demonstrates the importance of conducting research in the manner prescribed by the Hybrid Model proposed by Davidson, Redner, and Saul (in press) of which the ADP research has been an example as described in Chapter I. The relationships between pro- cess variables and outcomes in social interventions are extremely compli- cated. The task of sorting out and explaining these relationships is an essential and extremely difficult one--particu1ar1y when the interventions are carried out in the context of the natural environment. An appropriately complex approach to measurement such as that proposed by Davidson, et a1. is essential if such explanantions and understandings are ever to be achieved. APPENDIX A Self Report Delinquency Measure 131 SELF REPORT DELINOUEI‘TCY (SPOT HOW OFTEN HAVE YOU: HON OFTEN HAS YOUR SOT/DAUGHTER: IN THE LAST HOW OFTEN HAS YOUR YOUTH: 6 WEEKS XEAB 2085 2086 2087 r)! (JO! L ’0 2089 2090 2091 SKIPPED CLASS WHEN YOU/HE/SHE WAS IN SCHOOL? GONE ONTO SOMEONE’S LAND WHEN THEY DIDN’T WANT YOU/HIM/HER TO BE THERE. 0R WITHOUT THEIR PER'115510N? GONE INTO A HOUSE 0R BUILDING WHEN YOU/HE/SHE WASN’T SUPPOSED TO BE THERE? PLAYED ON A SCHOOL ATHLETIC TEAN? THREATENED To HURT SO‘~'1EONE? BEEN TOLD TO BRING YOUR/HIS/HER PARENTS TO SCHOOL FOR SOMETHING YOU/HE/SHE DID WRONG? DAMAGED 0R MESSED UP SOMETHING NOT BELONGING TO You/HIM/HER? HURT SOMEONE BADLY ENOUGH FOR HIM/HER To NEED BANDAGES OR A DOCTOR? GOTTEN ON THE HONOR ROLL FOR GOOD GRADES IN SCHOOL? TAKEN SOME PART OF A CAR 0R SOME GASOLINE? HIT A MEMBER OF YOUR/HIS/HER FAMILY? (IN ANGER) HAS NOT BEEN ALLOWED TO Go TO SCHOOL UNTIL THE SUPERINTENDANT 0R PRINCIPAL TOLD YOU/HIM/HER THAT You/HFJSHE COULD Go AGAIN? (BEEN SUSPavDED) 2097 2098 2100 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 132 IN THE LAST 6 WEEKS m TAKEN SOMETHING NOT BELINGING TO YOU/HIM/HER WORTH LESS THAN $2.00? EARNED SOME MONEY AT A JOB? DRUNK BEER 0R LIOUOR? (INCLUDES SIPS) RUN AWAY FROM HOME? SKIPPED A FULL DAY OF SCHOOL? BEEN SENT To THE SCHOOL PRINCIPAL’S OFFICE FOR BAD BEHAVIOR IN CLASS? CARRIED A GUN OR A KNIFE? WORKED ON A SCHOOL NEWSPAPER 0R YEARBOOK? TAKEN SOMETHING NOT BELONGING TO YOU/HIM/HER WORTH OVER $50.00? DONE SOMETHING AROUND THE HOUSE OR FOR THE FAMILY THAT REALLY PLEASED YOUR/HIs/HER FAMILY? SET FIRE To SOMEONE ELSE’S PROPERTY? USED 0R THREATENED TO USE A WEAPON TO GET SOMETHING FROM A PERSON? TAKEN SOMETHING FROM A STORE WITHOUT PAYING FOR IT? (REGARDLESS 0F PRICE) SMOKED WITHOUT YOUR PARENTS/YOUR/HIS PARENTS KNOWING ABOUT IT OR WITHOUT PERMISSION? (REGULAR CIGS.) WORKED FREE FOR A CHARITY ORGANIZATION? TAKEN A CAR WITHOUT THE OWNER’S PERMISSION? (INCLUDES JOYRIDING) XEAB 2115 2116 2117 2118 2119 T33 SMOKED MARIJUANA? TAKEN SOMETHING FROM A PERSON BY FORCE? (MAY OR MAY NOT USE A WEAPON) IN THE LAST 61EEI§S IEAB BEATEN UP ON SOMEBODY OR FOUGHT SOMEONE (PHYSICALLY)? TAKEN DRUGS OR PILLS) OTHER THAN MARIJUANA? BOUGHT OR GOTTEN SOMETHING THAT WAS STOLEN BY SOMEONE ELSE? BRO’EN INTO A PLACE AND STOLEN SOMETHING? TAKEN THEINGS WORTH LESS THAN $50.00? (OVER *W***M**************H* 2.03) APPENDIX B Recruitment Letter . . 34 IIICEIIGALI STATE lfi‘IIVERSITY - EAST LANSING 1 Department of Psychology - Snyder Hall May 14. 1979 COURS E AZ‘INOUl-I CEIME'I‘IT Dear Student; We would like to inform you about a sequence of courses available to under- graduates beginning in Fall, 1979. The course involves a three*term commitment. The course will be offered under Psychology 371 (Fall, 1979). Psychology 372 (Winter, 1980), and Psychology 373 (Spring, 1980) for 4 credits each term. The course will provide you the opportunity to work on a one-to-one basis with an adolescent from the local community. These youths have been referred by the juvenile court. The course will provide you with initial training in working with adolescents and ongoing supervision of your actual involvement and progress. The course will be a unique combination of information, intervention techniques, and actual community experience. This course will be a valuable experience for anyone considering graduate work and/or a career in the human service fields. .g;§5§§_ggg§; This course requires that you be enrolled the entire three terms. Due to the typical overwhelming response to this course, we are limited to the number of people we can process. Therefore, the first 300 people who call the office number (355-1814) starting Wednesday, May 23 through Tuesday, May 29 will be the only people invited to the initial meeting. If you or any of your friends would be interested in this project, please call 355-1814 between 8:00 a.m. to 12:00 p.m. and 1:00 p.m. to 5:00 p.m. no earlier than May 23 for information concerning the first meeting. Sincerely, /v ‘william S. Davidson II, Ph.D. Associate Professor of Psychology iDirector, Adolescent Diversion Project 195Dtrjm APPENDIX C Participation Agreement T35 USU ADOLESCENT DIVERSION PROJECT Participation Agreement In the Fall of 1976, Michigan State University and the Ingham County Probate Court started a program for youth in this community. The idea for the program is to provide an alternative to the juvenile court for local youth. This project involves several things. First, some of the youth in the program will be working with volun- teers The volunteers will be working with kids on a lot of different things depend- ing on the kinds of changes which need to be made for particular kids. This might involve working things out between you and your parents, working things out at school, helping you find a job, things to do in your free time and so on. Second, not all of the kids who'decide to be in this program will be working with a volunteer. About 3/4 oftfluakids will be working with a volunteer between six and eight hours per week on the kinds of things described above. The rest of the kids will be re— turned to the intake worker. whether or not you will be working with a volunteer will be decided by lottery and in no way reflects what we think of you. When you complete the program on / / and if you've had no further trouble with the court, you will have no further court responsibility. If you would like to-talk to a la"yer about this decision ’ttorney Daniel Rankine will_be willing to meet with vou without charge. ‘ This form is for you to indicate your intention to participate in the program. It‘s important for you to know exactly what you're agreeing to. 1. We hereby indicate that we are participating in this project voluntarily and understand that we have the right to withdraw if we see fit. 2. We understand that this project is being evaluated and that our input will be an important part of that evaluation. We, therefore, agree to provide honest and accurate information to the project staff. 3. We agree to be interviewed by the project staff on six: occasions. Once within two days, once in six weeks, once in twelve weeks, once in eighteen weeks, cme and two years from now. We understand that this information as all information, ”111 be kept confidential. We also understand that both and his/her. parents will participate in each interview. will be paid $5.00 for each interview. Finally, we understand that one of '5 friends will also participate in the interview. 4. We agree to work with a volunteer should one be assigned. We understand that this may involve six to eights hours per week. We understand that the volunteer will be able to work around our schedule. 5. We hereby give permission for the project staff to examine and record the police and court records for . It is also understood that this information will be handled confidentially and anonymously. This permiss- ion includes access to juvenile as well as adult records should they occur. This permission applies to all records from 192_ through 190_. Youth Date Parent Project Staff Court Staff APPENDIX D Life Domain Survey 1015. 1016. 1017. 1018. 1019. 1020. 1021. 1022. 1023. T36 Life Domain Survey (All questiomsrefer to the last six weeks.) How often does youth spend time with parents in athletics? (last 6 weeks) 1 2 . 3 4 5 Never Once Twice 3 or 4 More than times 4 times How often do the youth and parent(s) go to movies together? 1 2 3 4 5 Never Once Twice 3 or 4 More than , times 4 times ,— How often do the youth and parent(s) go camping/Iishing/hunting, etc.? I 1 2 3 4 3 Never Once Twice 3 or 4 More than times 4 times How often does youth visit relatives with parents? 1 2 3 4 5 Never Once Twice 3 or 4 More than times 4 times How often does the parent(s) instruct the youth in some skill/activity? l 2 3 4 5 Never Once Twice 3 or 4 More than times 4 times How often does the youth participate in purchased activities with parents? 1 2 3 4 5 Never Once Twice 3 or 4 More than times 4 times How often do the parent(s) talk with the youth about day-to-day things? 1 2 3 4 5 Never Once a Once a More than Daily month week once/week How often does the youth spend time with siblings in athletics? (no siblings) l 2 3 4 5 Never Once Twice 3-4 times More than 4 times How often does the youth spend time with siblings going to movies? (no siblings) l 2 3 4 5 Never Once Twice 3—4 times .More than 4 times T37 . r F.' ~s)1 (All questions refer to the 13st olh VCtKS-) 1024. How often does the youth spend time with siblings camping/fishing/hunting, etc.? * (no siblings) 1 2 3 4 5 Never Once' Twice 3-4 times More than 4 times 1025. How often does the youth spend time with siblings going out of town? * (no siblings) 1 2 3 4 5 Never . Once Twice 3—4 times More than ' 4 times 1026. How often does the youth spend time with siblings at indoor activities (TV)? * (no siblings) 1 2 3 4 5 Never Once 2-6 times More than Daily once/week 1027. How often does the source say the youth and siblings "hang around" together? * (no siblings) 1 2 3 4 5 Never Once 2-6 times More than Daily once/week 1028. How much is expected of youth in terms of household responsibilities? 1 2 3 4 5 Nothing Very little Some A fair A lot amount 1029. How often does the youth complete his/her household responsibilities? * (no responsibilities) 1 2 3 4 5 Never Seldom Half the Most of All the time the time time 1030. How often does the youth spend evenings at home? 1 2 3 4 5 Never Less than More than Almost Daily once a week once a week everyday 1031. How often does the youth engage in other spontaneous activities with his/her parent(s)? l 2 3 4 5 Never Once 2-6 times More than Daily once a week 1032. 1033. 1034. 1035. 1036. 1037. 1038. 1039. 138 (All questions refer to the last six weeks.) How often does the youth engage in spontaneous activities with his/her siblings? (activities not covered by other items) (no siblings) 1 2 3 4 5 Never Once 2-6 times More than Daily once a week How often does the youth engage in purchased activities with siblings (no siblings) l ,. 2 3 4 5 Never Once Twice 3—4 times More than 4 times How often does the youth sleep at home at night? 1 2 3 4 5 Never Less than More than Almost daily Daily once a week once a week How often does the youth sleep at a relative's home? 1 2 3 4 5 Never Once 2-6 times More than Daily once/week How much has the youth's neatness around the house improved? I 1 2 3 4 5 Much worse Somewhat No change Somewhat Much better worse better How much has the frequency of the youth's talking with parents changed? 1 2 3 4 5 Much less Somewhat No change More Much more less How has the youth's and parent's "getting along" changed? 1 2 3 4 5 Much worse Somewhat No change Somewhat Much better worse better How has the youth's performance of household responsibilities changed? 1 2 3 4 5 Much less Somewhat No change Somewhat Much often less often more often more often 1040. 1041. 1043. 1044. 1045. 1046. 1047. 1048. .vv (All questions refer to the last six weeks.) How much has the youth's spending time at home (evenings) changed? 1 2 3 4 5 Much less Somewhat No change Somewhat Much more 1ess~ more What change has occurred in terms of the parent's lessening restrictions on the youth and/or allowing the youth to do things outside the home? 1 2 3 4 5 Much worse Somewhat No change Somewhat Much better , worse better What change has occurred in terms of the youth's moodiness? 1 2 3 4 5 Much worse Somewhat No change Somewhat Much better worse better What change has occurred in terms of the parent and youth arguing? l 2 3 4 5 Much more Somewhat more No change Somewhat Much less frequent less frequent What is the view of the change which has occurred in the home area? 1 2 3 4 5 Completely Somewhat No change Somewhat Completely dissatisfied dissatisfied satisfied satisfied What change has occurred in the parent(s) hassling the youth about school? 1 2 3 4 5 Much more Somewhat more No change Somewhat Much less frequent less frequent How often do parents use punishment to control the youth? 1 2 3 4 5 Never Not very Sometimes Most of All the often the time time How often do the parent(s) talk to the youth about changing? 1 2 3 4 5 Never Not very Sometimes Most of All the often the time time Does source say that the youth lies to the parents? 1 5 No Yes 1049. 1050. 1051. 1052. 1053. 1054. 1055. 1056. 1057. 140 (All questions refer to the last six weeks.) Do the parent(s) suspect the youth of illegal activity? 1 No How much do 1 Never How much do 1 Never How much do 1 Never How much do (no phone) 1 Never How much do 1 Never How much do 1 Never How much do 1 Never How necessary is change in the home domain? 1 Unimportant 5 Yes the parent(s) and the youth argue about where the youth is going? 2 3 4 5 Once a week More than Almost daily Daily or less once a week the parent(s) hassle youth about the way the youth looks? 2 3 4 5 Once a week More than Almost daily Daily or less once a week parent(s) and youth argue about chores? 2 3 4 5 Once a week More than Almost daily Daily or less once a week the parent(s) and the youth argue about the use of the phone? 2 3 4 5 Once a week More than Almost daily Daily or less once a week the parent(s) and youth argue about the youth's friends? 2 3 4 5 Once a week More than Almost daily Daily or less ,once a week the parent(s) and youth argue about curfew? 2 3 4 5 Once a week More than Almost daily Daily or less once a week the parent(s) and youth argue in general? 2 3 4 5 Once a week More than Almost daily More than or less once a week or daily once a day (according to source) 2 3 4 5 Partially Relevant Central Crucial important 1058. 1060. 1061. 1062. 1063. 1064. 1065. 141 All questions refer to the last six weeks.) To what extent do the parent(s) hassle the youth about school? 1 2 3 4 5 Never Once a week More than Almost daily Daily or less once a week How often do the parent(s) intervene with youth's peers? l 2 3 4 5 More than Once a week 2—6 times Once Never once a week To what extent do the parent(s) know the specific things the youth does with friends? - 1 2 3 4 5 Not at all Vaguely Specifically Vaguely Specifically mention one mention one mention mention thing thing several several things things To what extent do parent(s) know who youth's friends are? l 2 3 4 5 Not at all knows small knows half knows most knows all proportion of them of them of them of them To what extent do the parent(s) know specific things youth does in free time? 1 2 3 4 5 Not at all Vaguely Specifically Vaguely Specifically mention mention mention mention one thing one thing several several things things To what extent do the parent(s)know what classes the youth takes? (not in school last 6 weeks) 1 2 3 4 5 Not at all Mention Mention Mention Mention one several most almost all How many of the youth's teachers do the parents know of? (not in school last 6 weeks) 1 2 3 4 5 None at all Mention one Mention Mention Mention several most almost all To what extent do the parent(s) know specific things that the youth does in school? (not in school last 6 weeks) 1 2 3 4 5 Not at all Vaguely Specifically Vaguely Specifically mention one mention one mention mention several thing thing several 1066. 1067. 1068. 1069. 1070. 1071. 1072. 1073. 142 (All questions refer to the last six weeks.) How often does the school inform the parents when the youth is absent? (not in school last 6 weeks) 1 2 _ 3 4 5 Never Not very Sometimes Most of All the often the time time To what extent have the parents taken action to change the youth's school situation? 1 2 3 4 5 Not at all Has talked Planning Taking Making very about it specific action intense action . effort To what extent have the parent(s) been successful (in making changes)? (no attempt above) 1 2 3 4 5 Not at all Very little Some impact Somewhat Completely impact satisfied satisfied To what extent has the change resulted in improved attendance? (no attempted change or not in school last 6 weeks) 1 2 3 4 5 Much worse Somewhat No change Somewhat Much better worse better To what extent has the change resulted in improved academic performance? (no attempted change or not in school last 6 weeks) 1 2 3 4 5 Much worse Somewhat No change Somewhat Much better worse better Is the youth currently enrolled in school? 1 5 No Yes Has5the youth been. suspended/expelled from schoolin last 6 weeks? 1 5 No Yes Is the youth good at particular classes? (not in school last 6 weeks) 1 2 3 4 5 No classes One class Some classes Most classes All classes 1074. 1075. 1076. 1077. 1078. 1079. 1080. 1081. 1082. 1083. 1084. I7Y\J (All questions refer to the last six weeks.) How many classes does the youth know particular things which are going on? (not in school last 6 weeks) 1 2 3 4 5 No classes One class Some classes Most classes All classes How often does the youth attend school? (not in school last 6 weeks) 1 2 3 4 5 Never Once/week More than Almost Daily (or the . or less once a week everyday required ° amount) How often does the youth talk to teachers outside'of class? (not in school last 6 weeks) 1 2 3 4 5 Never Less than Once/month Once/week More than once/month once/week How often does the youth talk to the principal at school? (not in school last 6 weeks) 1 2 3 4 5 Never Less than Once/month Once/week More than once/month once/week How many academic classes is the youth good at? (not in school last 6 weeks) 1 2 , 3 4 5 No classes One class Some classes Most classes All classes How many activity classes is the youth good at? (not in school last 6 weeks) 1 2 3 4 5 No classes One class Some classes Most classes All classes To what extent is the youth good at P.E.? (not in school last 6 weeks) 1 2 3 4 5 Not at all Very little O.K. Pretty good Very good Source Source Card Card 1085. 1086. 1087. 1088. 1089. 1090. 1091. 1092. 144' (All questions refer to the last six weeks.) How often does the youth participate in extracurricular activities? (not in school last 6 weeks) 1 2 3 4 5 Never Less than Once a More than Almost every once a week week once a week day How often does the youth skip any classes (on days when youth is in school)? (not in school last 6 weeks) 1 2 3 4 5 Everyday. Almost Once a Less than Never ' everyday week once a week To what extent does the youth want to switch schools? 1 2 3 4 5 A great Definitely Would Has Not an deal an issue like to mentioned issue How often does the youth get in trouble with the teachers? (not in school last 6 weeks) 1 2 3 4 5 More than Once 3 Once a Less than Never once a week week month once a month To what extent does the youth get a fair deal at school? (not in school last 6 weeks) 1 2 3 4 5 Never Not very Sometimes Most of All the often the time time To what extent does the youth get passing grades in school? (not in school last 6 weeks) 1 2 3 4 5 No classes One class Some classes Most classes All classes How often does the youth do homework in school? (not in school last 6 weeks) 1 2 3 4 5 Never (or no Less than Once a More than Almost homework) once a week week once a week every day How often does the parent(s) help youth with homework? (no school last 6 weeks) 1 2 3 4 5 Never Once a Once a More than Daily month week once/week 1093. 1094. 1095. 1096. 1097. 1098. 1099. 1100. 145 (All questions refer to the last six weeks.) How often does the youth do homework outside the school? (not in school last 6 weeks) 1 2 3 4 Never Less than Once a More than once a week week once a week How many classes does the youth like? (not in school last 6 weeks) 1 2 3 4 No classes One class Some classes Most classes 0 How many classes does the youth dislike? (not in school last 6 weeks) 1 2 3 4 All classes Most classes Some classes One class How many teachers does the youth like? (not in school last 6 weeks) 1 2 3 4 None Mentions one Mentions more Mentions one as likai than one as as really liked liked How many teachers does the youth dislike? (not in school last 6 weeks) 1 2 3 4 All Most all Some One 5 Almost every day 5 All classes 5 No classes 5 Mentions more than one as really liked 5 None To what extent does the youth like activity classes? (art, shop, etc.) (not in school last 6 weeks) 1 2 3 4 Not at all Dislikes Says is Likes O.K. To what extent does the youth like P.E.? (not in school last 6 weeks) 1 2 3 4 Not at Dislikes Says is Likes all O.K. To what extent does the youth dislike the rules at school? (not in school last 6 weeks) 1 2 3 4 A major reason Definitely Some A little for not dissatisfied dissatisfaction dissatisfaction liking school ' 5 Mentioned as a favorite 5 Mentions as a favorite class 5 Not an issue —- 146 (All questions refer to the last six weeks.) 1101. To what extent does the youth like the school administrators? * (not in school last 6 weeks) 1 2 3 4 5 Intense Dislike Not an Likes them Really dislike issue likes them 1102. To what extent does the youth care about school? 1 2 3 4 5 Not at all Very little Somewhat Concerned Very concerned 1103. To what extent is the youth concerned about finishing school? 1 2 3 4' 5 Not at all Very little Somewhat Concerned Very concerned 1104. In general, what is the youth's attitude toward school? I 1 2 3 4 5 Very negative Negative Neutral Positive Very positive 1105. How often do the administrators hassle the youth? * (not in school last 6 weeks) 1 2 3 4 5 More than Once a Once a Less than Never once a week week month once a month 1106. How often do the teachers hassle the youth? * (not in school last 6 weeks) 1 2 3 4 5 More than Once a Once a Less than Never once a week week month once a month 1107. What change has occurred in school attendance? * (not in school last 6 weeks) 1 2 3 4 5 Much worse Somewhat No change Somewhat Much better worse better 1108. What change has occurred in academic performance? * (not in school last 6 weeks) 1 2 3 4 5 Much worse Somewhat No change Somewhat Much better worse better 1109. What change has occurred in cflassroom behavior? * (not in school last 6 weeks) 1 2 3 4 5 Much worse Somewhat No change Somewhat Much better worse better gm 1110. 1111. 1112. 1113. 1114. 1115. 1116. 1117. 147 (All questions refer to the last six weeks.) What (not in school Much worse What 1 Much worse Wha t (not in school 1 Much worse , change has occurred in the last 6 weeks) 2 Somewhat worse change has occurred in the 2 Somewhat worse change has occurred in the last 6 weeks) 2 Somewhat worse What is the view of any changes 1 Completely dissatisfied How often does (not in school 1 Never How often does (not in school 1 Daily How often does 1 Never How often does 1 Never 2 Somewhat dissatisfied youth spend time last 6 weeks) 2 2 times a week youth completing homework? 3 4 5 No change Somewhat Much better better youth's attitude towards school? 3 4 5 No change Somewhat Much better better youth's relationship to teachers? 3 4 5 No change Somewhat Much better better which have occurred in the school area? 3 4 5 No change Somewhat Completely satisfied satisfied with friends during school time? youth skip school with friends?’ last 6 weeks) 2 More than once a week youth spend time with friends on weekends? 2 One weekend per month or less 3 4 5 Almost Few times Major part everyday a day of the day 3 4 5 2-6 times Once Never 3 4 5 A part of A part of Most of almost every every every weekend weekend weekend youth participate in purchased activities with friends? 2 Once 3 4 S 2-6 times More than Daily once a week —_M’ 148 WW (All questions refer to the last six weeks.) 1118. How often does youth participate in other spontaneous activities with friends? 1 2 3 4 5 Never Once a week More than Almost Daily or less once a week every day 1119. How often does youth spend time with friends in the afternoons? 1 2 3 4 5 Never Once a week More than Almost every Every or less once a week afternoon afternoon 1120. How often does youth spend time with friends evenings? 1 2 3 4 ’ 5 Never Once 2~6 times More than one Every evening/week evening 1121. How often does youth drink with friends? 1 2 3 4 5 Never Once' 2-6 times More than Daily once a week 1122. How often does youth go to parties with friends? 1 2 3 4 5 Never Once Twice 3-4 times More than 4 times 1123. How often does youth smoke dope with friends? 1 2 3 4 5 Never Not very Sometimes Most of All the often the time time 1124. How often does youth spend time at a friend's home? 1 2 3 4 5 Never Once 2—6 times More than Daily once a week 1125. How many close friends does youth associate with? l 2 3 4 5 None (no One Two Three More than particular set three of friends) 1126. How much has the frequency of the youth's spending time at home (during the day) changed? 1 2 3 4 5 Much less Somewhat No change Somewhat Much more frequent less more frequent warm-um“~ 1127. 1128. 1129. 1130. 1131. 1132. 1133. 1134. 149 (All questions refer to the last six weeks.) What change has occurred with respect to youth spending free time constructively? 1 2 3 4 Much worse Somewhat No change Somewhat worse better 5 Much better What change has occurred in youth's spending time with the peers that the youth usually gets in trouble with? 1 2 3 4 A lot : A little No change A little more more time less time 5 A lot less time What is the view of the change which has occurred in the friends-—free time area? 1 2 3 4 Completely Somewhat No change Somewhat dissatisfied dissatisfied satisfied How often does youth spend time hanging around alone? V l 2 3 4 All the Most of Sometimes Not very time the time often How often does the youth actively seek employment? 1 2 3 4 Never Once Twice Three or four times Has the youth had a job in the last 6 weeks? 1 3 No Self employed How much time does youth spend at the job? (no job last 6 weeks) 1 2 1-4 hrs. 3 (1—4) 4 None Less than Once a 4a-10 hrs./wk. week or less 0 a k (eég¥; odggl week) To what extent is the youth taking action to get a job? 1 2 3 4 Not at all Has talked Planning Taking about it specific action action 5 Completely satisfied Never 5 More than four times 5 Yes Formally has job 5 More than 10 hrs./wk. 5 Making very intense effort. Now employed 1136. 1137. 1138. 1139. 1140. 1141. 1142. IUV (All questions refer to the last six weeks.) To what extent does the youth want a job? (including current job if already employed) (i.e., getting a job or a new job is an issue) 1 2 . 3 4 5 Against Indifferent Somewhat Definitely A great deal the idea for the desired (wants a job or idea a new job) To what extent do the youth's parent(s) want him or her to have a job? (including current job if already employed) 1 . 2 3 4 5 Against. Indifferent Somewhat Definitely A great the idea for the desired deal idea Has the youth acquired a job recently? (last 6 weeks) 1 5 No Yes What kind of relationship does the youth and his/her supervisor have? (no job last 6 weeks or no supervisor) 1 2 3 4 5 Strong Negative Neutral Positive Strong negative positive What kind of relationship does the youth have with her/his peers at work? (no job last 6 weeks or no peers at work, i.e., works alone) 1 2 3 4 5 Strong Negative Neutral Positive Strong negative positive How often does youth talk to his/her supervisor? (no job last 6 weeks or no supervisor) l 2 3 4 5 Never Not very Sometimes Most of All the often the time time Does the supervisor say good things about the youth? (no job last 6 weeks or no supervisor) 1 5 No Yes What change has occurred in the frequency of the youth's illegal activities? 1 2 3 4 5 Much more Somewhat No change Somehwat Much less more less 151 (All questions refer to the last six weeks.) 1143. What change has occurred in the frequency of the youth's contact with the police? 1 2 . g 3 4 5 Much more Somewhat No change Somewhat Much less more less 1144. Is there a pending legal case? 1 3 5 Two or more cases One case None 1145. Is the yduth on probation? l 5 Yes No 1146. How often has youth had contacts with police recently? (6 weeks) 1 2 3 4 5 More than 3-4 times Twice Once Never 4 times 1147. What is the nature of the police disposition in comparison to what would be expected? * (none last 6 weeks) 1 3 5 Unfavorable No different Favorable than expected 1148. What was the court disposition of any petition which had been filed? 1 2 3 4 5 Institution- Probation Diversion to Case No alized or us or to dismissed petitions removed from other social home agencies 1149. Has the youth been detained? 1 5 Yes No 1150. What is the extent of the interviewer's confidence in the information contained in this section? 1 2 3 4 5 Completely Somewhat Somewhat Confident Completely uncertain uncertain confident confident (702) (90%) (992) Ft‘m APPENDIX E Intervention Survey 1155. 1156. 1157. 1158. 1159. 1160. 1161. 1162. 1163. 1164. 1165. How often does 1 Once/Month or less 152 Intervention Survey the volunteer have contact with the youth (not including phone). 2 3 4 5 2—4 times 5—7 times Twice a week More than a Month a Month twice/week How much time does the volunteer spend working on the case? 1 One hour or less/week To whao extent youth?- 1 Never 2 3 4 5 Two or three Four or five Six or seven Eight or more hours/week hours/week hours/week hours/week does the volunteer engage in spontaneous activities with the 2 3 4 ° 5 Once 3-5 times Once a week More than once a week What is the frequency of phone contact between the volunteer and youth? 1 Once a month or less How often does 1 Once a month or less How often does 1 Once a month or less To what extent 1 Not at all To what extent 1 Intensely dislikes To what extent 1 Intensely dislikes To what extent I Started off very poorly To what extent 1 Never 2 3 4 5 2-4 times 5-7 times Twice a week More than a month a month twice a week the volunteer call the target? 2 3 4 5 2~4 times 5-7 times Twice a week More than a month a month twice a week the target cell the volunteer? 2 3 4 5 2—4 times 5-7 times Twice a week More than a month a month twice a week does the youth like what the volunteer is doing in general? 2 3 4 5 Barely men— Seems to Is very Say it's the best tions positive like it pleased thing that's happened things does the youth like the volunteer? 2 3 4 5 Dislikes Okay Likes Intensely likes does the volunteer like the youth? 2 3 4 5 Dislikes Okay Likes Intensely likes has the volunteer had an early success with the case? 2 3 4 5 Had trouble It started Got things Had an early getting okay going quickly major success started does the youth share personal things with the volunteer? 2 3 > 4 5 Not very Sometimes Most of the All the time often time 153 Complaints 1166. 1167. 1168. 1170. 1171. 1...: 1"“ \1 IO 1173. 1174. 1175. 1176. 1177. 1178. 1179. 11 " Volunteer feels role is undefined 1 5 Yes No ‘ Volunteer has no car 1 5 Yes No Volunteer can't find things youth likes to do 1 5 Yes : No Volunteer has problems finding focus for approach] 1 5 Yes No Target doesn't show up for contacts 1 5 Yes No Volunteer doesn't have enough time 1 5 Yes No Volunteer is angry at target 1 5 Yes No Parents wanted volunteer of same sex 1 5 Yes No Parents wonder about the purpose of the program 1 5 Yes No Target is too busy to see volunteer 1 5 Yes No Target finds the volunteer aversive (gross) I 5 Yes No Target says program takes too much time 1 5 Yes No Volunteer is frustrated with lack of progress 1 5 Yes No Target has no enthusiasm l 5 Yes No Volunteer feels that program has nothing to offer target 1181. 1182. 1183. 1185. 1186. 1187. 1188. 1189. 1190. 1191. 1192. 154 Parent is uncooperative? 1 2 Yes No How often does the parent call the volunteer? l 2 3 4 5 Never Less than Once a 2-4 times More than once a month month a month once a week How often does the volunteer call the parents? 1 2 3 4 5 Less than Once a 2-4 times More than Almost once a month month a month once a week daily How often do the parents do things to assist the volunteer and youth getting together? ' l 2 3 4 ' 5 Never Not very Sometimes Most of All the often the time time To what extent do the parent(s) like what the volunteer is doing? 1 2 3 4 5 Not at Barely men- Seems to Are very Say it's the all tions positive like it pleased best thing that's things happened To what extent does the parent like the volunteer? l 2 3 4 5 Intensely Dislikes Okay Likes Intensely dislikes likes To what extent does the volunteer like the parent(s)? l 2 3 4 5 Intensely Dislikes Okay Likes Intensely dislikes likes To what extent has a friendship developed between the volunteer and the parent(s)? l 2 3 4 5 Never talk to Talk very Talk most Talk all Developed a each other seldom of the time of the time strong friendship To what extent do the parents want this volunteer workinq with the youth? 1 2 3 4 5 Strongly Mildly Doesn‘t Mildly in Strongly in opposed opposed care favor favor To what extent does the source mention the volunteer spontaneously? (leave blank for volunteer) 1 2 3 4 5 Not at all Once In two areas Several times In most areas How often does the volunteer talk with the parents? 1 2 3 4 5 Never Not very Sometimes Most of All the often the time time How often does the volunteer talk with the parent(s) about day-to-day things? 1 2 3 4 5 Never Not very Sometimes Most of All the often the tire time 1193. 1194. 1196. 1197. 1198. 1200. 1201. 1202. 1203. 155 To what extent do the parent(s) play a role in the school intervention? (no neg. of school change) 1 2 3 4 5 Never Not very Sometimes Most of All the time often the time How often do the youth's friends 3 end time with the volunteer and outh? . P y 1 2 3 4 5 Never Not very Sometimes Most of All the time often the time How often does the volunteer monitor the youth's performance in school? (not in.school last 6 weeks) 1 ‘ 2 3 4 5 Never Less than Once a month 2—4 times More than 4 once a month a month times a month How often does the volunteer involve the youth's friends in recreation? 1 2 3 4 5 Never Not very Sometimes Most of All the time often the time How often does the volunteer talk to the youth's friends independently? 1 2 3 4 5 Never Not very Sometimes Most of All the time often the time To what extent has the volunteer specified changes relevant to the friends? 1 2 3 4 5 Not at all Very little Somewhat Considerably A major focus How often does the volunteer involve the youth's peers in things? 1 2 3 4 5 Never Not very Sometimes Most of All the time often the time To what extent do the volunteer and youth have a mutual recreational interest? 1 2 3 4 5 Not at all Have some- Have one Have two Have several thing to (to do (to do (to do talk about together) together) together) How often do the volunteer and youth do purchased activities together? 1 2 3 4 5 Never Less than Once a month 2-4 times More than once a month a month 4 times a month How often do the volunteer and youth do athletic activities together? 1 2 3 4 5 Never Less than Once a month 2-4 times More than 4 once a month a month times a month What is the extent of the interviewer's confidence in the information contained in this section? 1 2 3 4 5 Completely Somewhat Somewhat Confident Completely uncertain uncertain confident confident (70%) (90%) (992) 1204. 1205. 1206. *4 iv C: \J 1209. 1210. 1211. 1212. 1213. 156 How often does the volunteer talk with the parent(s) alone about home problems? 1 2 3 4 5 Never Less than Once a 2—4 times More than 4 once a month month a month times a month How often does the volunteer talk to the youth about home? 1 2 3 4 5 Never Not very Sometimes Most of All the time often the time To what extent is the intervention focused on the home? 1 2 3 4 5 Not at all A very Part of, but One of the The major ° minor part not a major major foci focus focus To what extent is the home intervention focused on the youth doing household chores? 1 2 3 4 5 Not at all A very Part of, but One of the The major minor part not a major major foci focus focus To what extent is the home intervention focused on improving the youth's attitude? 1 2 3 4 5 Not at all A very Part of, but One of the The major minor part not a major major foci focus focus To what extent is the intervention focused on providing the parents information about the youth's comings and goings? 1 2 3 4 5 Not at all A very Part of, but One of the The major minor part not a major major foci focus focus How often does the volunteer talk with the parent and youth about home problems? 1 2 3 4 5 Never Less than Once a 2-4 times More than 4 once a month month a month times a month How often has the volunteer mediated a family disagreement? 1 2 3 4 5 Never Once Twice Three times More than 3 times To what extent does the volunteer involve the parent(s) in the planning of the intervention? 1 2 3 4 5 Never Not very Sometimes Most of All the time often the time How much of the time do the youth's siblings spend with the youth and volunteer? (no siblings) l 2 3 4 5 Never Less than Once a month 2-4 times More than 4 once a month a month times a month 1217. 1218. 2015. 2016. 2017. 2018. 157 To what extent is thelxmw intervention focused on improving the parent's household rules? 1 - Not at all 2 3 A very minor p rt focus Part of, but not a major 4 5 One of the The major major foci focus To what extent is the home intervention focused on having the parent(s) treat the youth more positively? 1 Not at all 2 3 4 5 A very Part of, but One of the The major minor part not a major major foci focus focus To what extent is the intervention focused on getting the parent(s) and youth to talk more? 1 Not at all 2 3 A very minor part focus Part of, but not a major 4 ; 5 One of the The major major foci focus To what extent is the intervention focused on improving the youth's school performance? 1 Not at all 1 Never 2 3 4 5 A very Part of, but One of the The major minor part not a major major foci focus focus How much do the volunteer and youth talk about school? 2 3 4 5 Less than Once a 2-4 times More than 4 once a month month a month times a month How often does 1 Never How often does (not in school 1 Never To what extent 1 Not at all To what extent 1 Not at all To what extent (not in school 1 Not at all To what extent 1 Not at all the volunteer talk to parent(s) about school? 2 3 Once Twice the volunteer help the youth with homework? last 6 weeks) 2 3 Less than Once a once a month month has the volunteer specified 2 3 Very little Some is the volunteer working on 2 3 Very little Somewhat is the volunteer working on last 6 weeks) 2 3 Very little Somewhat is the volunteer working on 2 3 Very little Somewhat 4 5 3—4 times More than 4 times 4 5 2—4 times More than 4 a month times a month school as a major change area? 4 5 Considerably A major focus the youth changing schools? 4 5 Considerably A major focus school—related activity changes? 4 5 Considerably A major focus changing the youth's classroom behavior? 4 5 Considerably A major focus 2019. 2020. 2021. 2028. 2029. 2030. 2031. 158 Has the volunteer set up a contract with the youth's school staff? (not in school last 6 weeks) 1 5 No Yes To what extent is the volunteer trying to get the target to go to school more? 1 2 3 4 5 Not at all Very little Somewhat Considerably A major focus How often does the volunteer monitor the changes the school is supposed to make? (not in school last 6 weeks) 1 2 3 4 5 Never Less than Once a 2-4 times More than 4 once a month month a month times a month To what extent is the volunteer trying to get the youth to do more homework? (not in school last 6 weeks) 1 2 3 4 5 Not at all Very little Somewhat Considerably A major focus How often does the volunteer visit the school? 1 2 3 ' 4 5 Never Once Twice 3-4 times More than 4 times How often has the volunteer talked to the school administrators? l 2 3 4 5 Never Once Twice 3-4 times More than 4 times How often has the volunteer talked to the school counselors? 1 2 3 4 5 Never Once Twice 3-4 times More than 4 times To what extent is the volunteer working on curriculum changes? 1 2 3 4 5 Not at all Very little Somewhat Considerably A major focus To what extent does the volunteer involve the youth in the negotiation of school changes? (no neg. of school change) 1 2 3 4 5 Never Not very Sometimes Most of All the time often the time How often has the volunteer talked to the teachers? 1 2 3 4 5 Never Once Twice 3-4 times More than 4 times How much has the volunteer talked with the youth about a job? 1 2 3 4 . 5 Not at all Very little Some Considerably A major focus To what extent has the volunteer identified the job area as a focus of change? 1 2 3 4 5 Not at all Very little Some Considerably A major focus How much has the volunteer instructed the youth in job seeking? l 2 3 4 5 Not at all Very little Some Considerably A major focus to O to to 2033. 2034. 2035. 2036. 2037. 2038. 2039. 2040. 2041. 2042. 2043. 2044. 2045. 159 How often does the volunteer take the youth for job interviews? 1 2 . 3 4 5 Never Once Twice 3-4 times More than 4 times To what extent has the volunteer contacted local youth employment resources? 1 2 3 4 5 Not at all Talks about Made contact Made contact Actively using it. with one with several/ several resource using one Has the volunteer gotten the youth a job? (Youth already has’job - or clearly does not want a job) 1 5 No Yes Has the volunteer gotten the youth into a work-study program? (Youth already in W.S. Program or has no interest to be in one) 1 5 ' No Yes To what extent does the parent(s) play a role in the legal intervention? (no legal issue) 1 2 3 4 5 Not at all Very little Some Considerably A major focus How often does the volunteer talk to the police? 1 2 3 4 5 Never Once Twice 3-4 times More than 4 times How often does the volunteer talk to probation/D88 staff? 1 2 3 4 5 Never Once Twice 3-4 times More than 4 times How often does the volunteer talk to a lawyer? I 2 3 4 5 Never Once Twice 3-4 times More than 4 times Has the volunteer assisted in getting the youth a lawyer? I 5 No Yes Has the volunteer attended a hearing? 1 5 No Yes Has the volunteer assisted in negotiating a court disposition? 1 5 No Yes How often has the volunteer visited the youth in detention? 1 2 3 4 5 Never Not very Sometimes Most of All the time often the time Has the volunteer aided in getting the youth released from detention? 1 5 No Yes How often do the volunteer and youth talk about legal problems? 1 2 3 4 5 Never Once Twice 3-4 times More than 4 times 2046. 2048. 2050. 2051. 2052. 2053. 2054. 160 How often does the volunteer talk to the parent(s) about legal problems? 1 2 3 4 5 Never Once Twice 3—4 times More than 4 times To what extent has the volunteer specified interpersonal contingencies which need alteration? 1 2 3 4 5 Mentions Mentions only Mentions one Mentions two Mentions several nothing general things specific specific changes change changes To whafi extent hastfluzvolunteer specified a contract between the youth and the significant others in his/her life? 1 2 3 4 5 5 Mentions Mentions the Talks of Specified the Specified nothing idea in wanting to people the people passing do a contract and behaviors To what extent has a contract been used? 1 2 3 4 . 5 Mentions Has specific Has talked Has 3 Has had a contract nothing plan with both contract in operation parties written To what extent has the volunteer set up a monitoring system for the contract? 1 2 3 4 5 None Mentions plans Has a Has started Has started mentioned to do so specific using one using one and knows system ready of each party's to implement compliance To what extent has the volunteer involved the youth and the relevant significant(s) in the contract negotiations? 1 2 3 4 5 Completely Minimally Somewhat Involved in Involved in uninvolved involved involved most everything (or no contract) To what extent has the volunteer been instructing the youth and significant others in the methods of contracting? 1 2 3 4 5 Not at all Very little Somewhat They have Very actively talked about involves youth in it a fair instruction and amount practice To what extent does the youth like the contracting approach? (Kid knows nothing of approach) 1 2 3 4 5 Intense Dislike . Neutral Likes Likes dislike intensely To what extent is the parent following through with the contract? (no contract) ‘ 1 2 3 4 5 Not at all Very little Somewhat Most of Complete the time followthrough 2056. 2057. 2059. 2060. 2061. 2062. 2063. To what extent is the y uth following through the contract? (no contract) 1 2 3 4 5 Not at all Very little Somewhat Most of Complete the time followthrough To what extent do the parent(s) like the contracting approach? (source knows nothing of approach) 1 2 3 4 5 Intense Dislike Neutral Likes Likes dislike intensely To what extent has the volunteer specified changes the youth would like made in his/her environment? 1 2 3 4 5 Mentions Mentions only Mentions one Mentions two Mentions several nothing general specific specific changes things change changes To what extent has the volunteer specifi-d courses of action to facilitate change? 1 2 3 4 5 Mentions Mentions only Mentions one Mentions two Mentions several nothing vague actions Specific specific actions action actions To what extent has the volunteer specified individual targets for change? 1 2 3 4 5 Source Source Source Source Source mentions mentions only mentions one mentions two mentions nothing categories specific specific several of individuals person individuals To what extent has the volunteer (with or without the kid) taken specific action to initiate change? 1 2 3 4 5 Source Source Source Source Source mentions mentions only mentions one mentions two mentions nothing vague action specific specific several action actions (Refer to Item 2060) To what extent has youth alone taken specific action to initiate change? 1 2 3 4 5 Source Source Source Source Source mentions mentions only mentions one mentions two mentions nothing vague action specific specific several action actions actions To what extent has the volunteer followed up on change areas? (Monitoring the change attempts) I 2 3 4 5 Mentions Mentions only Followed up Followed up Followed up on nothing general things on one on two several changes specific specific change changes To what extent did the volunteer involve the youth in the planning and action which has been accomplished? 1 2 3 4 5 Completely Minimally Somewhat Involved Involved in uninvolved involved involved in most everything 2064. 2066. 2067. 2069. 2071. 2072. 162 To what extent has the volunteer been instructing the youth and/or parent(s) in advocacy? 1 2 3 4 5 Not at all Very little Somewhat They have Very actively talked about involved in it a fair instruction amount and practice To what extent does the youth like the advocacy approach? (never used approach) 1 2 3 4 5 Intense Dislike Neutral Like Like intensely dislike To what extent does the volunteer take action to generate new resources (e.g., employment, new club) for the target? 1 2 3 4 5 Never Talks about Has a Has taken Substantial it specific plan initial progress action To what extent is the volunteer using outside resources (e.g., Ed. programs employment programs)? 1 2 3 4 5 Not at all Talks about Made contact Made contact Actively using it with one with several/ several resource using one To what extent do the parent(s) like the approach? (never used approach) 1 2 3 4 5 Intense Dislike Neutral Likes Likes dislike intensely To what extent does the volunteer talk about himself/herself and his/her feelings to the youth? 1 2 3 4 5 Never Not very Sometimes Most of All the time often the time How much effort does the volunteer put into trying to understand the youth's feelings, beliefs, values, etc? 1 2 3 4 5 None Very little Some Considerable Major amount To what extent is the volunteer trying to get the parent(s) to advocate for the youth? 1 2 3 4 5 Not at all Very little Somewhat Considerably A major focus To what extent does the source view the "home domain" as an important focus for change? 1 2 3 4 5 Not important Somewhat ) Important Very A major focus important important 2073. 2074. 2076. 2077. 2078. 2079. 2080. To what extent does the source focus for change? 1 2 Not important Somewhat important To what extent does the source important focus for change? 1 2 Not important Somewhat important .0 To what extent does the source focus for change? 1 2 Not important Somewhat important To what extent does the source focus for change? 1 2 Not important Somewhat important To what extent does the source focus for change? 1 2 Not important Somewhat important View the "friend domain" as an important 3 4 5 Important Very Major focus important view the youth's freetime activities as an 3 4 Important Very important 5 A major focus view the "school domain” 3 4 5 Important Very A major focus important View the "job domain" as an important 3 4 Important Very important 5 A major focus View the "legal domain" as an important 3 4 5 Important Very A major focus important Has the youth mentioned his involvement in the ADDP to any court personnel he/she has come in contact with? (no contact with court personnel) 1 5 Yes No Has the youth mentioned his involvement in the ADD? to any police personnel he/she has come in contact with? (no contact with police personnel) 1 5 Yes No Has mentioning his/her involvement in the ADD? kept the youth out of further trouble? (no mention of the project to other source) 1 5 Yes No REFERENCES Albee, G.W. Conceptual models and manpower requirements in psychology. In Cook, P.E. (Ed.), Community psychology and community mental *1 health. San Francisco: Holden-Day, 1970. Bandura,‘A. Principles of behavior modification. New York: Holt, Rine- hart & Winston, 1969. Baron, R., Feeney, R., & Thornton, W. Preventing delinquency through diversion; The Sacramento County 601 Diversion Project. Federal Probation, 1973, 32, 13-18. Binder, A., & Newkirk, M. A program to extend police service capability. Crime Prevention Review, 1977, 4, 26-32. Blakely, C.H. The diversion of juvenile delinquents: A first step toward the dissemination of a successful innovation. Unpublished doctoral dissertation. Michigan State University, 1981. Blakely, C.H., Kushler, M., Emshoff, J., Parisian, J., & Davidson, W. An experimental examination of behavioral contracting with delinquent youth. Paper presented at the annual meeting of the American Psychological Association, Montreal, 1980. Blakely, C., Kushler, M., Parisian, J., & Davidson, W. Self-reported delinquency as an evaluation measure: Comparative reliability and validity of alternative weighting schemes. Criminal Justice and Behavior, 1980, 2(4), 369-386. Carter, Robert M., & Klein, Malcolm W. (Eds.) Back on the street: The diversion of juvenile offenders. Englewood Cliffs, NJ: Pren- tice-Hall, Inc., 1976. *2,3-- Cook, Thomas, J., & Scioli, Frank, P. The effectiveness of volunteer programs in courts and corrections: An evaluation of policy related research. Chicago: University of Illinois, 1975. Crime in the U.S.: FBI Uniform Crime Reports, 1979. Washington, DC: U.S. Department of Justice, 1980. Crime in the U.S.: FBI Uniform Crime Reports, 1980. Washington, DC: U.S. Department of Justice, 1981. Davidson, W.S. Adolescents in legal jeopardy. In Seidman, E., Rappaport, J., Davidson, W.S., & Linney, J. (Eds.), Changing human service systems: Interventions with children, adolescents, adults and the elderly. Unpublished manuscript, 1979. L * see last page 164 165 Davidson, W.S., & Kushler, M.G. Social science contribution to crime and delinguency control: Alternatives to watching the pendulum. Unpublished report prepared under contract to the National Science Foundation, 1979. Davidson, W.S., & Rapp, C.A. Child advocacy in the justice system. Social Work, May 1976, 21(3), 225-232. Davidson, W.S., & Rappaport, J. Advocacy and community psychology. In G. Weber & G. McCall (Eds.), Social scientists as advocates. Bever- ly Hills: Sage Publications, 1978. 0 Davidson, W.S., & Redner, R. Personal conmunication. Spring, 1982. Davidson, W.S., Redner, R., & Saul, J. Modes of measuring social and com- munity change. In Seidman, E. (Ed.), Handbook of social and community intervention. Beverly Hills: Sage Publication, in press. Davidson, W.S., & Seidman, E. Studies of behavior modification and juve- nile delinquency: A review, methodological critique, and social perspective. Psychological Bulletin, 1974, 31(12), 998-1011. Davidson, W., Seidman, E., Rappaport, J., Berck, P., Rapp, N., Rhodes, W., & Herring, J. Diversion program for juvenile offenders. Social Work Research and Abstracts, 1977, 13(2), 40—49. Davidson, W.S., Snellman, L., & Koch, R. Diversion in Michigan. Michigan Department of Social Services: Mimeo, 1980. Davidson, W.S., & Wolfred, T.R. Evaluation of a community based behavior modification program for the prevention of delinquency: The failure of success. Community Mental Health, 1976. Dowell, David A. Volunteers in probation: A research note on evaluation. Journal of Criminal Justice, 1978, 6, 357-361. Durlak, J.A. Comparative effectiveness of paraprofessionals and profes- sional helpers. Psychological Bulletin, 1979, 86) 80-92. Durlak, J.A. Evaluating comparative studies of paraprofessional and pro- fessional helpers: A reply to Nietzel and Fisher. Psychological Bulletin, 1981, 89, 566-569. Elliot, D.S., & Voss, H.L. Delinquengy and dropout. Lexington, MA: Lex- ington Books, 1974. Emshoff, J.G. The diversion of juvenile delinquents: A comparison of intervention with the family and intervention with alT7life systems. UnpubliShed doctoral dissertation. Michigan State University, 1980. Fernandez, Maria Isabel. Self-report delinquency and official archival records: A comparison. Unpublished master’s thesis, Michigan State University, 1981. 166 Gold, Martin. Undetected delinquent behavior. Journal of Research in Crime and Delinquency, 1966, 12, 27-46. Gold, M., & Mann, 0. Delinquency as defense. American Journal of Ortho- psychiatny, 1972, 32(3). Gold, M., & William, J. The effect of "getting caught": Apprehension of the juvenile offender as a cause of subsequent delinquencies. Prospectus: A Journal of Law Reform, 1969, 2, 1-12. Comes—Schwartz, B., Hadley, S., & Strupp, J. Individual psychotherapy . and behavior therapy. In Rosenzweig, M.R. & Porter, L. W. (Eds.), Annual Review of Psychology (Vol. 29). Palo Alto, CA: Annual Review, Inc., 1978. - Gottman, J., & Markman, H. Experimental designs in psychotherapy research. In Garfield, S. & Bergin, A. (Eds.), Handbook of psychotherapy and behavior change (2nd ed.). New York: John Wiley & Sons, 1978. Heller, K., & Monahan, J. Psychology and community change. Homewood, IL: *4 5__ The Dorsey Press, 1977. Kantrowitz, R. Training professionals to work with delinquents: Differen- tial impact of varying training/supervision/intervention strate- gi s. Unpublished doctoral dissertation, Michigan State Uni- versity, 1979. Kazdin, A. The application of operant techniques in treatment, rehabilita- tion and education. In Garfield, S., & Bergin, A. (Eds.), Hand- bookoftmychotheragyand behavior change (2nd ed.). New York: John Wiley & Sons, 1978. Kelly, J.G., Snowden, L.R., & Munoz, R.F. Social and community interven- tions. Annual Review of Psychology. 1977, 28, 323—361. Klein, M.W. Deinstitutionalization and diversion of juvenile offenders: A litany of impediments. In Morris, M. & Tonry, M. (Eds.), Crime and Justice, 1978. Chicago: University of Chicago Press, 1979. Krisberg, B., & Austin, J. (Eds.) The children of Ishmael. Palo Alto, CA: Mayfield Publishing Co., 1978. Lemert, E.M. Instead of court: Diversion in juvenile justice. In Carter, R.M. & Klein, M.W. (Eds.), Back on the streets: The diversion of juvenile offenders. Englewood Cliffs, NJ: Prentice Hall, Inc., 1976. Meltzoff, J., & Kornreich, M. Research in psychotherapy. New York: Atherton Press, 1970. Messinger, Sheldon. Confinement in the community: A selective assessment of Paul Lerman's community treatment and social control: A critical analysis of juvenile correctional policy. In Krisberg, 8., & Austin, J. (Eds.), The children of Ishmael. Palo Alto, CA: Mayfield Publishing Company, 1978. *see last page 167 Mitchell, C.M. Nonprofessionals working with delinquent youth: An experimental comparison of university, community college and community nonprofessionals. Unpublished doctoral dissertation, Michigan State University, 1980. *6-- Nietzel, M.T., & Fisher, S.G. Effectiveness of professional and parapro- fessional helpers: A comment on Durlak. Psychological Bulletin, l981, B9, 555-565. Nietzel, M.T., Winett, R.A., MacDonald, M.L., & Davidson, W.S. Behavioral approaches to community psychology. New York: Pergamon Press, 1977. Orlinsky, D., & Howard, K. The relation of process to outcome in psycho- therapy. In Garfield, S., & Bergin, A. (Eds.), Handbook of psychotherapy and behavior change (2nd ed.). New York: John Wiley & Sons, 1978. Palmer, T., & Lewis, R. An evaluation oquuvenile diversion. Cambridge, MA: Oelgeschlager, Gunn, & Hain Publishers, Inc., 1980. Patterson, G., McNeal, S., Hawkins, N., & Phelps, R. Reprogramming the social environment. In Ulrich, R., Stachnik, T., & Mabry, J. (Eds.), Control of human behavior (Vol. 21. Glenview, IL: Scott, Foresman & Co., l970. Paul, G.L. Strategy of outcome research in psychotherapy. Journal of Consulting Psychology, 1967, g1, l09—118. Quay, H.C., & Love, C.T. The effects of a juvenile diversion program on rearrests. Criminal Justice and Behavior, 1977, AKA), 377-395. Rappaport, J. Community psychology: Values, research and action. New York: Holt, Rinehart & Winston, l977. Redner, R., Snellman, L., & Davidson, W.S. A review of behavioral methods in the treatment of delinquency. In Morris, R. & Kratochwill, T. (Eds.), Practice of therapy with children: A textbook of methods. New York: Pergamon Press, in press. Riedel, M., & Thornberry, T. The effectiveness of correctional programs: An assessment of the field. In Krisberg, B. & Austin, J. (Eds.), The children of Ishmael. Palo Alto, CA: Mayfield Publishing Co., 1978. Romig, D.A. Justice for our children. Lexington, MA: Lexington Books, T978. Ross, A.D. Behavior therapy with children. In Garfield, S. & Bergin, A. (Eds.), Handbook ofypsychotherapy and behavior change: An empirical analysis. New York: John Wiley and Sons, 1978. Ryan, W. Blaming the victim (2nd edition). New York: Vintage Books, 1976. *see last page *7__ Schur, E. Schwartz, Shlien, J. StrUppa H. 168 Radical nonintervention. Englewood Cliffs, NJ: Prentice-Hall, Inc., 1973. I., Jensen, D., & Mahoney, M. Volunteers in juvenile justice. National Institute of Law Enforcement and Criminal Justice. LEAA U.S. Dept. of Justice. Washington, DC: U.S. Government Printing Office, 1977. M. (Ed.) Research in psychotherapy. Washington, DC: American Psychological Association, 1968. Psychotherapy research and practice: An overview. In Garfield, ‘S. 8 Bergin, A. (Eds.), Handbook of psychotherapy and behavior Stuart, R. _qhapge (2nd edition). New York: John Wiley & Sons, 1978. B. Behavioral contracting within the families of delinquents. In Stumphauser, J.S. (Ed.), Behavioral therapy with delinquents. Springfield, IL: Charles C. Thomas, 1973. Tharp, R., & Wetzel, R. Behavior modification in the natural environment. New York: Academic Press, 1969. Westman, J.C. Child advocacy: New professional roles for helping families. The White Williams, Wolfgang, New York: The Free Press, l979. House Conference on Children. Emergence of identity. In Krisberg, B. & Austin, J. (Eds.), The children of Ishmael. Palo Alto, CA: Mayfield Publishing Co., 1978. J.R., & Gold, M. From delinquent behavior to official delinquency. In Krisberg, B. & Austin, J. (Eds.), The children of Ishmael. Palo Alto, CA: Mayfield Publishing Co., 1978. M., Figlio, R., & Sellin, T. Delinquency in a birth cohort. Chicago: University of Chicago Press, 1972. *see last page ADDITIONS TO REFERENCES Austin, R. Social learning and social control: A comment on Conger. Criminology, 1977, 16(1), 111-116. 2Cloward,~R., & Ohlin, L. Illegitimate means, differential opportunity and delinquent subcultures. In Krisberg, B., and Austin, J. (Eds.), The children of Ishmael. Palo Alto, CA: Mayfield Publishing, Co., l978. 3Conger, R. Social control and social learning models of delinquent behavior: A synthesis. Criminology, 1976, 1311), 17-40. 4Jackson, D.N. The personality research form. Palo Alto: Research Psychologists Press, 1968. C JJackson, D.N. A sequential system for personality scale development. In C. Spielberger (Ed.), Current topics in clincial and community psychology (Vol. 2). New York: Academic Press, 1970. 6Nie, N., Hull, C., Jenkins, J., Steinbrenner, K., & Bent, D. Statistical package for the social sciences (2nd Ed.). New York: McGraw- Hill Book Co., 1975. 7Seltiz, C., Wrightman, L.S., & Cook, S.W. Research methods in social relations. New York: Holt, Rinehart & Winston, 1976. 169