A SYSTEMS MODEL APPROACH TO EVALUATING SOCIAL PROGRAMS: THE CASE OF YOUTH SERVICE BUREAUS Dissertation for the Degree of Ph. D. MICHIGAN STATE UNIVERSITY WILLIAM L SELKE 1977, LIBRARY Michigan State UHIVt Ity This is to certify that the thesis entitled A SYSTEMS MODEL APPROACH TO EVALUATING SOCIAL PROGRAMS: THE CASE OF YOUTH SERVICE BUREAUS presented by William L. Selke has been accepted towards fulfillment of the requirements for Ph.D. degreein SOCIAL SCIENCE 01.4. M M em Major professor 7 " 3 - '7 7 Date 0-7639 I mm; TIM III III III In IIIZIIIZIIQIII II 7% 175 afie- BIL, of _ “ “an ABSTRACT A SYSTEMS MODEL APPROACH TO EVALUATING SOCIAL PROGRAMS: THE CASE OF YOUTH SERVICE BUREAUS By William L. Selke The study which follows was an attempt to apply currently evolving systems research concepts to the evalua- tion of a group of delinquency~diversion programs. The primary purpose of the study was to clarify issues related to the implementation and effectiveness of Youth Service Bureaus. These programs are among the most popular innovations in the juvenile justice system today and are aimed at reducing delinquency by divert- ing youths from the formal processes of the justice system to alternative services outside the justice sys- tem, thus avoiding stigmatization. Three classes of variables were considered. The outcome measures used in the study were official crime rates based on law enforcement and juvenile court sta- tistics. Internal organizational variables including staff orientations to delinquency and program percep- tions by staff members were also addressed in the study. Finally, there was an examination of external William L. Selke environmental variables such as the degree of local sup- port, the location of the programs in the social struc- ture, and the socio-political conditions within a community. Findings from the study did not generally sup— port the hypothesis that Youth Service Bureaus would reduce the levels of delinquency in a community. With regard to crime reduction and diversion, results from time-series analyses of police and court data indicated only sporadic evidence that the programs had impacted on the justice system. There were no instances where a particular crime variable was consistently affected across sites, nor were there instances where a particu- lar site showed significant impact across the target variables. There were three general findings in the imple- mentation analysis section which shed light on the incon- sistent outcome results. First, it did not appear that Youth Service Bureau staffs held personal orientations to delinquency associated with the unique activities embodied in the original Bureau concept (i.e., resource development, coordination and service brokerage, systems modification, etc.). Second, it was concluded that the designs and functions of the various programs underwent such modifications during implementation that most of the core elements of the Youth Service Bureau concept William L. Selke were untested in this study. And third, it was found that program staff typically felt that they had been unprepared and ill—equipped for the complex task of implementing a planned social innovation aimed at affecting systems—level change. The implications of the study were discussed with regard to the program development, implementation, and evaluation phases of the social policy—making process. It was suggested throughout the study that the major problems in program evaluation were the unlimited num— ber of variations of program models and the absence of guidelines for introducing and maintaining program models. Recommendations were then put forth having to do with conceptual and operational refinements in program models, training seminars on the maintenance of program and evaluation design, and the incorporation of research concerns in the early stages of program development among others. A SYSTEMS MODEL APPROACH TO EVALUATING SOCIAL PROGRAMS: THE CASE OF YOUTH SERVICE BUREAUS By William L. Selke A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY College of Social Sciences 1977 The material in this project was partially pre- pared under Grant No. 77—N199-0095 from the Office of Criminal Justice Training and Educa- tion, Law Enforcement Assistance Administration, U.S. Department of Justice. Researchers under- taking such projects under Government sponsor- ship are encouraged to express freely their professional judgment. Therefore, points of View or opinions stated in this document do not necessarily represent the official position or policy of the U.S. Department of Justice. ii To my wife, Beth Montgomery, who was devoted in her support throughout this endeavor, and my parents, Virginia and Albert Selke, whose patience and understanding were appreciated. iii ACKNOWLEDGMENTS Dr. John McNamara and Dr. Ralph Lewis provided continual Opportunity for professional growth by demon- strating sensitivity to the realities of social pro- cesses and respect for the methods of social inquiry. Dr. Robert Trojanowicz was very helpful throughout my program and Dr. William Davidson, II was an invaluable coach during the dissertation phase of my studies. Special thanks to Dr. McNamara, my committee chairman, and each member of my committee. iv TABLE OF CONTENTS Page LIST OF TABLES . . . . . . . . . . . . . . . . . . vii LIST OF FIGURES . . . . . . . . . . . . . . . . . . ix Chapter I. INTRODUCTION 1 Purpose of the Study . . . . . . . . . . 1 Conceptual Framework . . . . . . . 3 Social Change and Experimentation . . . 4 Evaluation Research and Social Policy . 10 Organizational Theory . . . . . . . 19 Definition of the Systems Model . . . . . 23 II. THEORY, PRACTICE, AND EVALUATION IN DIVERSION . . . . . . . . . . . . . . . . 29 Delinquency Theory . . . . . . . . . 29 Diversion and Youth Service Bureaus . . . 40 Program Evaluations . . . . . . . 49 Summary Statistical EvaluatiOns . . . . 49 Narrative Descriptions . . . . . . . . 53 Experimental and Quasi-Experimental Studies 0 O O O O O O O O 0 0 6O Rationale for the Study . . . . . . . . . 71 III. METHODOLOGY . . . . . . . . . . . . . . . . 75 Research Questions . . . . . . . . . . 75 Research Design and Data Collection . . . 78 Impact Analysis and Time- Series Designs . . . . . . . . . . . . . . . 80 Crime Reduction . . . . . . . . . . . . 96 Diversion . . . . . . . . . . . . 104 Implementation Analysis . . . . . . . . 110 Organizational Factors . . . . . . . . 116 Environmental Assessment . . . . . . . 122 Limitations of the Study . . . . . . . . 124 Chapter Page IV. RESULTS . . . . . . . . . . . . . . . . . . 127 Time-Series Analyses . . . . . . . . . . 127 Crime Reduction . . . . . . . . . . . . 133 System Impact of Diversion . . . . . . 142 Implementation Analyses . . . . . . . . . 162 Delinquency Orientations of Project Staffs . . . . . . . . . . . 162 Program Perceptions . . . . . . . . . . 168 Environmental Influences . . . . . . . 173 V. CONCLUSIONS AND POLICY IMPLICATIONS . . . . 185 Crime Reduction and Diversion by Youth Service Bureaus . . . . . . . . . . . . 185 Program Implementation . . . . . . . . . 189 Implications for Policy . . . . . . . . . 195 Program Development . . . . . . . . . . 195 Implementation Guidelines . . . . . . . 198 Evaluation Strategies . . . . . . . . . 200 APPENDICES . . . . . . . . . . . . . . . . . . . . 206 A. PROJECT SUMMARIES . . . . . . . . . . . . . 207 B DEMOGRAPHIC CHARACTERISTICS . . . . . . . . 222 C. ANNUAL U.C.R. DATA-COLLECTION FORM . . . . 224 D. DELINQUENCY ORIENTATION SCALE . . . . . . . 226 E PROGRAM PERCEPTIONS SURVEY . . . . . . . . 230 F. ENVIRONMENTAL ASSESSMENT INTERVIEW GUIDE . 237 G. INSTRUMENT MODIFICATIONS ON THE PROGRAM PERCEPTIONS SURVEY . . . . . . . . . . . 241 BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . 243 vi Table 10. 11. 12. 13. LIST OF TABLES Summary of Study Components and Sites Time-Series Analysis Example . . . . . . . . Actual Burglary Offenses (Rates Per 1000 Total County Population, 1970) . . Actual Larceny Offenses (Rates Per 1000 Total County Population, 1970) . . Actual Vandalism Offenses (Rates Per 1000 Total County Population, 1970) . . . . Summary Statistics for Monthly Time-Series Analyses of Crime Reduction--Actual Burglary, Actual Larceny, and Actual Vandalism . . . . . . . . . . . . . . . . Total Delinquency Arrests (Rates Per 1000 Juveniles, Adjusted) . . . . . . . . . Delinquency Arrests for Burglary (Rates Per 1000 Juveniles, Adjusted) Delinquency Arrests for Larceny (Rates Per 1000 Juveniles, Adjusted) . . . . . Delinquency Arrests for Vandalism (Rates Per 1000 Juveniles, Adjusted) . . . Delinquency Arrests for Curfew/Loitering (Rates Per 1000 Juveniles, Adjusted) Delinquency Arrests for Runaway (Rates Per 1000 Juveniles, Adjusted) . . . Summary Statistics for Monthly Time-Series Analyses of Diversion--Total Delinquency Arrests, Delinquency Arrests for Burglary, and Delinquency Arrests for Larceny vii Page 97 128 135 136 137 139 144 146 147 148 151 153 154 Table 14. 15. 16. 17. 18. 19. 20. B1. Summary Statistics for Monthly Time-Series Analyses of Diversion--Delinquency Arrests for Vandalism, Delinquency Arrests for Curfew/Loitering, and Delinquency Arrests for Runaway . . . . . . . . . Police Referrals to Juvenile Court (Rates Per 1000 Juveniles, Adjusted) . . . . . Summary Statistics for Monthly Time-Series Analyses of Diversion--Part I and Part II Police Referrals to Juvenile Court, and Total Juvenile Court Petitions . . Internal Consistency Data for the Delinquency Orientation Scale . . . . . Mean Ratings on the Delinquency Scale by Site . . . . . . . . . . . . . . Internal Consistency Data for the Program Perception Survey . . . . . . . . . Mean Ratings on the Program Perceptions Survey by Site . . . . . . . . . . . . Population and Median Income for Youth Service Bureau and Comparison Counties viii Page 156 159 161 164 166 170" 172 223 Figure 1. 2. LIST OF FIGURES The Systems Model of Research Youth Service Bureau Evaluation Model ix Page 27 77 CHAPTER I INTRODUCTION Purpose of the Study The proliferation of social programs in recent years has provided a convenient testing ground for the practice of applied social research. But in moving from the isolated and highly controllable laboratory to the helter-skelter world of politics and policy making, social scientists have encountered a myriad of problems. One of the most perplexing of these has been the lack of consid- eration given social research findings by policy makers. The situation can be partially attributed to the advisable manner in which policy is formulated. But social scien— tists can also be cited for their hesitancy to expand the conceptual base of applied research so as to generate information more directly usable in the policy-making process. It is the purpose of this study to demonstrate the use of an expanded model of social research in the evalu- ation of a group of delinquency-prevention programs. The common practice in program evaluation has been to extract the articulated goals of a program and measure the extent to which these are realized. But in ignoring the implementation and metamorphosis of social innovations, the goal-outcome model of evaluation fails to provide information about the intermediary factors that influence the structure and functions of programs. The two major problems to be addressed in this study are the lack of information regarding the effectiveness of delinquency- diversion programs, and the inadequacies of traditional social research models. Because juveniles account for such a large propor- tion of official crime (over one-fourth of all arrests and nearly one-half of arrests for Index crimes),1 there has been much interest recently in developing innovative approaches to the delinquency problem. The programmatic focus of this research will be on one of the most popular of these approaches, diversion, as the concept is embodied in Youth Service Bureaus. These projects have been con- ceptualized as delinquency-prevention programs aimed at reducing delinquency by providing services at an early stage in the delinquent's career. The Bureaus are diver- sionary alternatives to the formal processes of juvenile courts. The study is aimed at accomplishing three things: (1) assessing the degree to which projects are successful in attaining stated goals, (2) examining the organiza- tional and environmental contexts to determine the effects 1U.S. Uniform Crime Reports (Washington, D.C.: Department of Justice, 1975), p. 188. of various factors on program development, and (3) clari- fying the relationships between program implementation and measures of effectiveness. In the remainder of this chapter, discussion will revolve around a number of diverse writings that have made important contributions to an expanded conceptual framework for applied social research. There will also be a presentation of the systems model to be used in this study, which incorporates many of the ideas included in the above writings. The following chapter contains a review of the historical development of the diversion concept and Youth Service Bureaus, as well as a critical review of past evaluation efforts in this area. Chapters III and IV are devoted to the application of the systems model evaluation. And in the final chapter, the discus- sion focuses on the implications of the study for policy making in the area of delinquency programming and social research. Conceptual Framework The framework within which this research was conceived borrowed from a wide range of influential writ- ings. While an exhaustive review of all relevant con— ceptual discussions is not feasible, certain bodies of the professional literature offer a wealth of information regarding the modification and expansion of classical research models for purposes of social research. Writers in the area of planned social change and experimentation have been instrumental in clarifying and legitimizing the role of social scientists in applied research. Discus- sions of program evaluation and policy research have similarly pointed to conceptual and methodological pre- requisites for a meaningful social research model. And the field of organizational theory has produced concepts of critical importance to social research which deal with understanding the structure and function of organizations‘ within the larger social context. Highlights from each of these areas are examined prior to defining the sys— tems model. Social Change and Experimentation The role of social scientists in the world of practical affairs has been a widely debated topic in the academic world for many years. Historically, some of the most influential writers have voiced strong reservations about an applied social science concerned with the solu- 1 tion of social problems. It has often been contended that engagement in science should be for the sake of 1See, for example, Max Weber, Max Weber: Essays in Sociology, ed. Hans Gerth and C. Wright Mills (New York: Oxford University Press, 1946); Robert Merton, Social Theory and Socia1_Structure (New York: Free Press, 1957); Talcott Parsons, The Social System (Glencoe, Illinois: The Free Press, 1951). science and that to judge the works of scientists on the basis of usefulness is to weaken the scientific approach. 0n the other side of this debate are those who believe that social scientists have an obligation to contribute their skills and understanding to the analysis of complex social problems. Writers like 0gburn, Etzioni and Etzioni, LaPiere, and Watson1 have emphasized the impor- tance of social science and applied research for improv- ing the information bases upon which public policy decisions are made. But while the debate may continue, social scientists in many disciplines are finding them- selves immersed in the complexities of social research and the need for expanded models of research is acutely perceived. In what has come to be labelled the interaction— ist perspective, Blumer posited a set of social processes that he felt should fall within the purview of social researchers—-the emergence and legitimization of social problems, the mobilization of action, the formation of an official plan of action, and the implementation of 1William F. 0gburn, Social Change, 2nd ed. (New York: Viking Press, 1950); Amitai Etzioni and Eva Etzioni, Social Change (New York: Basic Books, 1964); Richard T. LaPiere, Social Change (New York: McGraw- Hill, 1965); G. Watson, ed., Concepts for Social Change (Washington, D.C.: Institute for Applied Behavioral Sciences, 1967). the official plan.1 The influential writings of Gouldner have also suggested guidelines for a new model of social research.2 In discussing what he termed the "engineer— ing" model of research, Gouldner suggested that the model was inadequate because of an unquestioning acceptance of problem definitions and program formulations, and a naive expectation that research findings would be automatically utilized if the best canons of scientific research were followed.3 Davis called the traditional approach to research the "Los Alamos" model and described the under- lying rationale as follows: . . . Social policy should work like the development of atmoic weapons: the pure scientist tells the decision maker, "Hey, we have some pure research findings that suggest we can build an atomic bomb," and the decision maker says, "Well, I can't under- stand all this fancy stuff, but here's some money to try to make a bomb.” Davis faulted this model on three counts. First, social scientists seldom have entree to the levels where real lHerbert Blumer, "Social Problems as Collective Behavior," Social Problems 18 (1971): 298-306. 2A. Gouldner, "Theoretical Requirements of the Applied Social Sciences," American Sociolggical Review 22 (1957): 52-102; A. Gouldner, "Explorations in Applied Social Science," in Applied Sociology, ed. Alvin W. Gouldner and S. M. Miller (New York: The Free Press, 1965). 3Gouldner, "Explorations," p. 19. 4James A. Davis, "On the Remarkable Absence of Non-academic Implications in Academic Research," in Social Policy and Sociology, ed. N. J. Demerath, Otto Larsen, andearl F. Schuessler (New York: Academic Press, 1975), p. 40. decision making takes place. Second, policy makers probably know more about the realities of a social prob- lem in most areas than the social scientist. And third, policy decisions about a social problem are not made as simply as those regarding the production of military weapons. The issues raised by these three writers all suggest that there is a social as well as a scientific component in the applied research process. In the original works of Lippitt et al. on the dynamics of planned change, social relationships between the "client and the change-agent” were considered the critical factors in determining the degree to which research findings were utilized.1 Similarly, applied research has been conceptualized as a "complex of rela- tionships: between applied science and the scientific method, between the applied scientist and his subjects, organization and subjects turned clients."2 And the concept of "linkage” discussed by Havelock embodies these notions and stresses the importance of developing "reciprocal and collaborative” relationships with a variety of potential users.3 The common denominator in 1Lippitt et al., Dynamics of Planned Change (New York: Harcourt, Brace and World, 1958). 2Warren G. Bennis, Kenneth D. Benne, and Robert Chin, eds., The Planning of Change (New York: Holt, Rinehart and Winston, 1962), p. 629. 3Ronald G. Havelock, Planning for Innovation (Ann Arbor, Mich.: Institute for Social Research, 1971). all of these writings is the recognition that applied research requires a paradigm that suggests guidelines for combining social awareness and scientific rigor. The combination of experimental rigor and sensi- tivity to the social realities of applied research has been the distinguishing feature of the writings of Fairweather.1 His development and application of the "experimental social innovation" model illustrates a comprehensive approach to social research. In discuss- ing the definition and evaluation of new social subsys— tems (e.g., Youth Service Bureaus), Fairweather alluded to the range of factors that must be considered by the social researcher. He wrote, ”It is the functional rela- tionship between the outcome, participants, and social situation that the social innovative experimentalist uses to operationally define a social subsystem."2 Social research, then, needs to address the characteristics of those involved in the program and the nature of the con- text in which the program is implemented, as well as the question of goal attainment. The first two concerns have often been ignored in social research and the guide- lines for dealing with them (especially the second) are 1George W. Fairweather, Methods for Experimental Social Innovation (New York: John Wiley and Sons, 1967); George W. Fairweather, Social Psychology in Treating Mental Illness (New York: John Wiley and Sons, 1964). 2Fairweather, Methods, p. 77. currently vague. But writings such as those above have begun to delineate the crucial problem areas and posit alternative methods for dealing with them. Besides the conceptual and methodological issues raised in this section, writers in the area of social change and experimentation have called for broad shifts in political stance with regard to developing and evalu- ating social innovations. For example, Campbell has noted that because of the nature of the policy-making process, social programs are typically overadvocated and oversold, which places them in a position where the fail— ure to meet lofty goals is almost insured.1 In response to this condition, he called for a change in political posture "from the advocacy of a specific reform to the advocacy of the seriousness of the problem, and hence, to the advocacy of persistence in alternative reform 2 Riecken and Boruch efforts should the first one fail." also advocated a social experimentation process that begins with a planning phase, incorporates rigorous eval- uation, and uses the findings to accept, reject or modify the experimental intervention.3 1Donald T. Campbell, "Reforms as Experiments," American Psychologist 24 (April 1969): 409-29. 2 Ibid., p. 73. 3Henry W. Riecken and Robert F. Boruch, eds., Social Experimentation (New York: Academic Press, 1974), p. 14. 10 The literature on planned social change and experimentation has been very influential in stimulat— ing thought about the nature of an optimal paradigm for applied research. While such a paradigm has not yet crystallized in any concrete form, the writers referred to above have presented concepts and guidelines that are now beginning to affect the overall quality of social research. The discussion in the next section on evalua- tion research and its relationship to social policy has also had an impact on the thought and practices of social researchers. Evaluation Research and Social Policy In recent years, the most common type of applied social research has become program evaluation. The ever- increasing number of social programs coupled with the acute sensitivity to issues of efficiency and accounta- bility has resulted in a situation where social scien- tists are frequently called upon to apply their skills. With growing participation in the practice of evaluation research, it is not surprising that the topic has received greater attention in the literature. Since the seminal work of Suchman on the prin- ciples and types of evaluative research, there have been numerous program evaluators who have attempted to define the intricacies of their mission. Suchman himself 11 conceptualized evaluation research opposite "pure” research on the following continuum: Pure research aimed at the accumulation of knowledge -—Basic research with relevance for application-- Action research aimed at the process of application—— Engineering research dealing with the actual con- ditions of application--Evaluation research focused upon the administrative decisions following appli- cation. In moving from one end of this continuum to the other, there are basic differences in both the objectives and methods. The objective of pure research is the accumu— lation of knowledge and the generation of abstract theoretical generalizations while evaluation research is geared toward practical use and the production of con— crete guidelines for action in highly specified situations. The methods of evaluation research are not as specified and systematized as those of pure research, but the fol— lowing writings indicate that progress is being made. Because of the absence of parameters on the evolving evaluation research model, the ability to com- pare and contrast program evaluation is limited. Bern— stein and Freeman noted that there is a virtually endless series of designs, measurements, analyses, and outcome criteria utilized in the evaluation of any given program 1Edward A. Suchman, Evaluative Research (New York: Russell Sage, 1967), p. 89. 12 area.1 They recommended the creation of systematized evaluation-research modules that outline the basic requirements for each phase of the evaluation process. And Cain and Watts suggested the development of "repli— cability criteria" to be used as guideposts in the selec— tion of variables and the scope of evaluation efforts.2 In this manner, the range of possible variables would be limited to those over which policy makers have some control, and the types of programs subjected to evalua- tion would be restricted to those that can be replicated in other settings. While there is indeed justification for concern regarding the parameters of a comprehensive evaluation model, the field of evaluation research is awaiting guide- lines to a number of more immediate problems. Two of the most pressing are the maintenance of integrity in program and research designs and related questions revolving around the issue of validity in evaluation research. Probably the first major problem encountered by most evaluators is the absence of clarity and specificity about the goals, activities and expected outcomes of the lIlene N. Bernstein and Howard E. Freeman, Academic and Entrepreneurial Research (New York: Russell Sage, 1975), p. 140. ZGlen G. Cain and David Watts, "The Methodology of Evaluating Social Programs," in Evaluating Social Programs, ed. Peter H. Rossi and Walter Williams (New York: Seminar Press, 1972). 13 program. As was mentioned earlier, this situation can be partially attributed to the advisable manner in which policy is formulated. Suggestions have been forthcoming with regard to the clarification of program elements, but the effect of these on the reported evaluation studies is negligible at present. Weiss, for example, advocated the use of program models that would identify the intended processes of a program and the means and 1 And steps by which the program is intended to work. in a similar vein, Aronson and Sherwood suggested the use of impact models to explicate the primary target groups, expected nature of effects, and the extent to which effects must be present in order to be considered successful.2 An outlining procedure of the above nature is of critical importance to the evaluator lest, as often happens, the program takes on different operational characteristics and the research design is rendered useless. Perhaps the most important benefit of a program or impact model is the opportunity it provides for evaluators and administrators to sense changes that are occurring in the program and adjust evaluation designs accordingly. 1Carol H. Weiss, Evaluation Research (Englewood Cliffs, N.J.: Prentice—Hall, 1972), p. 50. 2Sidney H. Aronson and Clarence C. Sherwood, "Researcher Versus Practitioner: Problems in Social Action Research," Social Work 12 (1967): 89-96. 14 If social policy and program statements could be developed with a high degree of clarity and specificity, the next major obstaCle for the evaluator would be the maintenance of constancy in treatment interventions. When a program is implemented in a preexisting social structure, a variety of factors impinge on its evolu- tion, some of which are likely to be at odds with the original plans. This problem was illustrated by Short in the following presentation of a personal communica- tion from Lamar Empey regarding the well—known ”Provo Experiment": I have found staff members responding to emergent problems on ad hoc and individual bases, rather than turning to the theoretical guidelines of the pro- ject for solutions. There are always pressures to do this because . . . the problems of individuals often conflict with the problems of the organiza— tion and with what the theory says should be done. Therefore, faced with an ideological conflict, staff members reject the theory and do what they think is best for the individual. The result is that the theory often does not get a test.1 The manner in which the problem of a shifting program is handled has serious implications for the overall validity of evaluations, but there is little. agreement among writers as to the most beneficial and feasible solutions. On the one hand, Freeman and 1James F. Short, "The Natural History of an Applied Theory: Differential Opportunity and 'Mobiliza- tion for Youth,'” in Social Policy and Sociology, ed. N. J. Demerath, Otto Larsen and Karl F. Schuessler (New York: Academic Press, 1975), p. 204. 15 Sherwood contended that the evaluator has a responsi- bility to watch over the project like a "snarling watch— dog" in order to insure that program functions are undertaken precisely as prescribed.1 But Rivlin doubted the practicality of this approach as she wrote, "No good teacher or doctor or social worker will participate wholeheartedly in an experiment in which his every word is programmed and his freedom to adapt methods to cir— 2 Increasingly, cumstances is completely circumscribed." researchers are turning to the notion of process evalua- tion in response to this issue. While the concept does not necessarily increase the amount of control the researcher has over program elements nor guarantee the stability of program operations, it does, if carried out intensively, provide an ongoing, dynamic description of program development. What was traditionally known as "field" or "qualitative" research has been modified for use in program evaluation and is discussed currently under the headings of prOcess evaluation, operational research, or implementation analysis. While these concepts and other similar ones have unique properties in the various 1Howard E. Freeman and Clarence C. Sherwood, "Research in Large-Scale Intervention Programs," Journal of Social Issues 21 (1965): 11-28. 2A1ice M. Rivlin, Systematic Thinking for Social Action (Washington, D.C.: The Brookings Institution, 1971), pp. 115-16. 16 definitions assigned them, they do have similarities insofar as each is primarily concerned with the whats and whys of programs in contrast to the how wells. As Rossi noted, the primary purpose of qualitative, non- experimental research is to provide the decision maker with "information about the variety of forms the pro- grams are taking in individual projects, information which may be useful in setting up tighter evaluations at a later stage or in modifying policy to bring local practices in line with overall agency aims."1 An earlier distinction along these lines was that made by Scriven, who termed the above kind of evaluation "formative" as opposed to "summative" evaluations, which provide infor— mation at the conclusion of a program regarding the effectiveness (i.e., goal- or outcome-oriented evalua- tion).2 Formative evaluation was viewed by Scriven as a feedback mechanism that supplied developmental infor— mation throughout the life of a project to be used in making decisions about the future directions of the program. 1Peter H. Rossi, "Testing for Success and Failure in Social Action," in Evaluating Social Programs, ed. Peter H. Rossi and Walter Williams (New York: Seminar Press, 1972), p. 35. 2Michael Scriven, "The Methodology of Evaluation," in Perspectives on Curriculum Evaluation, ed. Ralph W. Tyler, Robert M. Gagne and Michael Scriven (Chicago: Rand McNally, 1967). 17 The concept of process evaluation should improve the usefulness of evaluation findings if the practice and development of procedures continues toward a compre- hensive model. But while process evaluation and its related concepts have been partially successful in addressing the first problem mentioned in this section-— the maintenance of program and research design-—the related validity problem remains a perplexing one. Short of a fully integrated evaluation component from the initial stages of program development, basic ques- tions regarding the overall validity of evaluation research have not been sufficiently answered. The most commendable efforts at improving the validity of evaluation research are those that have made use of the "stronger" quasi-experimental and experimen- tal designs. But most evaluation research has been of an ex post facto nature. Issues of validity in ex post facto research revolve around the fact that evaluators seldom have the opportunity to randomly select and assign participants and to manipulate independent variables (i.e., maintain experimental constancy) so that the likelihood of systematic biases is minimized. Kerlinger's warnings regarding the dangers of improper and erroneous interpretations of ex post facto research stem from the same kinds of problems and apply to most evaluation 18 research. He wrote that because of the "plausibility of many explanations of complex events . . . it is easy to accept the first and most obvious interpretation of an established relation, especially if we work without hypotheses to guide the investigation or proceed from the 1 And dependent variable to the independent variable." as Merton pointed out, post factum explanations do not lend themselves to nullifiability because they are so flexible and new interpretations can always be found to fit the facts.2 The earlier discourses on planned social change have been instrumental in legitimizing the participation of social scientists in the policy-making arena. Their emphasis on scientific rigor has begun to influence thinking about the potential role of social research in an increasingly planned society. Evaluation research writings have been most successful in supplying tenta- tive guidelines for researchers caught up in the prac— tice of ex post facto research. Another body of litera- ture that has had a positive impact on social research is that portion of organizational theory dealing with the relationships between organizations and their environments. The most important contribution of these lFred N. Kerlinger, Foundations of Behavioral Research, 2nd ed. (New York: Holt, Rinehart and Winston, 1973), p. 333. 2Merton, Social Theory. 19 writings has been the description of processes whereby organizational structures and functions are influenced in their interactions with the environment. Some of the more widely known of these discussions will be pre- sented before defining the systems model. Organizational Theory The literature in organizational theory is voluminous and most of the major works consider the environment to one degree or another. But there is a core of writers who have been particularly successful in demonstrating the importance of external environmen- tal factors in the analysis of organizations. Stinch- combe's discussion of organizational and environmental relationships has been singularly impressive in drawing 1 And the line of attention to this area of analysis. thought that has come to be known as the institutional school took these extra-organizational relations as the focal point for organizational analysis. Selznick, in particular, brought this topic to the foreground and his description of the manner in which the Tennessee Valley Authority (TVA) was ”co-opted" by local and national interest groups has become a classic in lArthur Stinchcombe, "Social Structure and Environment," in Handbook of Organizations, ed. James G. March (Chicago: Rand McNally, 1965). 20 organizational theory.1 The implications of his study were significant for understanding organizational "beha— vior" because of the way in which organizations were viewed as following the only feasible path in program development--that of adjusting to the environment and subsequently, changing in the process. This perspec- tive provided a framework for analyzing a number of perplexing organizational phenomena. One such phenomenon was the observed variation in patterns of organizational structures and managerial effectiveness. While there was a general recognition that different organizational structures were more or less successful in different settings, it remained for Burns and Stalker to make the connection between specific environmental factors and organizational control.2 They found that two Scottish factories had radically different managerial structures (highly structured vs. informal) and each had functioned with a high degree of success. But in the first instance, the environmental conditions were of a highly stable and patterned nature, while in the second there was much instability and fast- paced changes. And Perrow demonstrated the impact of 1Philip Selznick, TVA and the Grass Roots (Berkeley, Calif.: University of California Press, 1949). 2 T. Burns and G. M. Stalker, The Management of Innovation (London: Tavistok Institute, 1961). 21 another environmental factor, namely technology, on the power structure of hospitals.1 He found that as medical technologies became more complex, physicians replaced board trustees atop the power structure, but when tech- nologies began to require the use of outside consultants, the administrative group came to power. In each of these studies, the direct impact of external environmental factors on organizational structures is lucid. There have also been some illuminating discussions regarding the effects of the environment on the goals and activi- ties of organizations, as well as the structures. The environment has turned out to be an impor- tant factor in explaining the often-observed metamor- phosis that takes place in program objectives and functions. This phenomenon is not unexpected in social programs given the guidelines for operation are usually vague and general, but explanations for this critical process have been slow to evolve. Cyert and March coined the term "organizational drift” to refer to the directions and shifts in directions made by organiza- tions as a result of various member coalitions being formed and reformed within the environment.2 And lCharles Perrow, "Hospitals: Technology, Goals, and Structure," in Handbook of Organizations, ed. James G. March (Chicago: Rand McNally, 1965). 2Richard M. Cyert and James G. March, A Beha- vioral Theory of the Firm (Englewood Cliffs, N.J.: Prentice-Hall, 1963). 22 Thompson and McEwan outlined four basic approaches to interacting with a novel environment—-competition, bargaining, co-optation, and coalition.1 They suggested that these approaches are progressively more costly in terms of maintaining self-control and that new organi- zations are seldom in a position to compete or bargain. They must, therefore, relinquish a certain degree of autonomy in the search for support. Similarly, Thompson earlier discussed the costs to a new organization of creating what he called "domain consensus" or "a set of expectations for members of an organization and those with whom they interact about what the organization will and will not do."2 In carving out an operational domain, the organization has to modify and revise its original goals and objectives so as to bring them more in line with local perceptions. The literature is now becoming replete with descriptions of how organizational opera- tions can be radically altered by a wide range of exter- nal pressures, but this knowledge has not often been integrated in the evaluations of social programs. This 1James D. Thompson and William J. McEwan, "Organizational Goals and Environment: Goal Setting as an Interaction Process," in Complex Organizations and Their Environments, ed. Merlin B. Brinkerhoff and Philippe R. Kunz (Dubuque, Iowa: William C. Brown, 1972), p. 266. 2James D. Thompson, Organizations in Action (New York: McGraw-Hill, 1967), p. 29. 23 situation will be addressed through the use of the systems model. Definition of the Systems Model Systems concepts in various forms have been explicated over the years by persons from many fields, but it is Bertalanffy who is generally credited with organizing these into a comprehensive theory.1 He sug- gested that the systems model was a "scientific revolu- tion" as defined by Kuhn2-—the appearance of new concep- tual schemes or "paradigms.” Bertalanffy noted that in all fields of science, "notions like wholeness, holistic, organismic, gestalt, etc. signify that, in the last resort, we must think in terms of systems of elements in mutual interaction.”3 This view is in contrast with the analytic, mechanistic, one-way causal paradigm of classical science. In moving from a classical to a systems model, the implications are far-reaching in all phases of inquiry. As Bertalanffy wrote, "compared to the ana- lytical nature of classical science with resolution into component elements and one-way or linear causality as the 1Ludwig von Bertalanffy, General System Theory (New York: George Braziller, 1968). 2T. S. Kuhn, The Structure of Scientific Revolu- tions (Chicago: University of Chicago Press, 1962). 3Bertalanffy, General System Theory, p. 45. 24 basic category, the investigation of organized wholes of many variables requires new categories of interaction, transaction, organization, teleology, etc."1 Certain of these new categories have already attracted much attention. For example, teleological advances can be seen in theories of criminology that have moved from narrow, one-factor (psychological or sociological) explanations of crime to a recognition of multiple causation. And also in the area of criminology, a cor- responding shift in the focus of research can be found in the changing emphasis from positivist to interaction- ist models. The positivist school focused on discover— ing and treating the singular "causes" of crime and delinquency while the interaction approach has begun to address a wider range of issues such as the creation of laws, the relationships between crime and crime control institutions,zuuithe interactions of rule-making, rule- enforcing, and rule-breaking activities.2 A parallel to the changing perspectives in criminology can be found in the organizational theory literature, where the tendency in recent years has been to View organizations as open systems. Katz and Kahn 1Ibid., p. xxii. 2Tony G. Poveda and Edward Schaffer, "Positivism and Interactionism: Two Traditions of Research and Crim- inology,” in Criminal Justice Research, ed. Emilio Viano (Lexington, Mass.: D. C. Heath, 1975), pp. 31-33. 25 reflected the open-systems model of organizations. They found fault with the Closed—system View because of its failure to recognize the interdependence of organi- zations and their environments, and the accompanying notion that system changes due to environmental factors should be treated as error variance.1 In contrast, the open-systems model maintains that organizational shifts caused by environmental pressures are not "sources of error variance, but are integrally related to the func— tioning of a social system. We cannot understand a system without a constant study of the forces that impinge upon it."2 The term "morphogenesis” was used by Buckley to describe the similar processes in cemplex system-environment exchanges that tend to elaborate or change a system's form, structure, or state.3 The task for this study is to relate the most pertinent of the sys- tems concepts in a comprehensive model that can be applied to the evaluation of the delinquency prevention programs. The underlying rationale for a systems approach in program evaluation was presented by Suchman in the following statement: 1Daniel Katz and Robert L. Kahn, The Social Psy— chology of Organizations (New York: John Wiley, 1966). 21bid., p. 27. 3Walter Buckley, Sociology and Modern Systems Theory (Englewood Cliffs, N.J.: Prentice-Hall, 1967), p. 58. 26 No event has a single cause and each event has multiple effects. All events are interrelated in a complex causal nexus open by nature and subject to rational intervention. No single factor is a neces- sary and sufficient cause of any other factor, and change in one part of the system may occur without necessitating a completely new equilibrium. Evaluations of success must be made in terms of conditional probabilities involving attacks upon causal factors which are only disposing, contribu- tory, or precipitating rather than determining. Figure 1 illustrates the "causal" sequence in which the social program is only one of many possible actions or events that may bring about (or deter) a desired effect. This study will concentrate on the first four stages of the sequence. The following characteristics outline the basic parameters of the systems model as it will be used here: 1. the meshing of process and goal-outcome eval- uation; 2. concern with different units and levels of analysis; and 3. a focus on maintenance outputs, as well as ”productivity." The systems model of evaluation, then, is an approach to examining the implementation and effectiveness of social programs (1) which recognizes the range of variables (2) that impinge upon survival and program activities. lSuchman, Evaluative Research, pp. 84-85. 27 .AsmmH .mwam HHmmmsm mmOOOSUOMGOO COHNQHOHpccca prow BUZV noacomom O>Hpmsam>m .ccsnosm .soncomon mo Hopes mEoummm 0291:.H madem moEOOpso mandfiuc> ucocsmamm AMOHQNHNNP Hwaoflu IcNHacwho .w.mv mucm>m wcflco>soucH m ENHMOMA mancflhc> pGOUGOQOUcH ovw 0Q .< Unmscm ”meadow AmnOpocw fiancee IGOHH>cm .w.cv mqofiuficcoomum 28 The specific form to be taken by the model in this study will be discussed in detail in Chapter III. The first chapter has been devoted to reviewing writings which contributed to the conceptual evolution of the systems model application in social research. An attempt has been made to delimit the parameters of the model and define the general characteristics it will take on in the study. Prior to discussing the methodology, however, writings on the theory, practice, and evaluation of diversion programs and Youth Service Bureaus will be examined in order to pull together what is known about the concepts and what is yet to be known. CHAPTER II THEORY, PRACTICE, AND EVALUATION IN DIVERSION Delinqpency Theory In Chapter I, highlights from various bodies of literature were surveyed because of their relevance to the research design of this study. The core of this chapter will be devoted to a critical review of some of the most widely recognized studies in the areas of diver- sion and Youth Service Bureaus. This review will pro— vide the basis for discussing the need and rationale for a systems model approach. Before looking at the studies, however, some introduction to delinquency theory is neces- sary because of the significant influence of these writ- ings on the evolution of diversion and Youth Service Bureau concepts. It is somewhat ironic that the diversion concept has become so popularized. The use of informal diversion has always been an integral part of the criminal justice process as a result of the individual discretionary powers of justice system officials. One factor that con— tributed to the current level of interest in diversion was increased concern about the informal practices in 29 30 the justice system.1 Because the informal dispositional patterns were often found to be discriminatory in nature, emphasis came to be on limiting the discretionary powers of justice system officials. Another very real factor that hastened the acceptance of diversion programs was the increasingly large number of persons coming to the attention of jus— tice system officials. As is the case with informal diversion, the primary purpose of formalized diversion is to lessen the burden on justice system institutions, especially the courts. The novel ingredient in formal diversion programs has been the provision of services as an alternative to the legal processes of the justice system. In addition to these factors, the growth and direction of delinquency and criminological theory played critical roles in suggesting the theoretical foundations upon which diversion programs might be built. The origin of today's most popular perspectives on crime and delin— quency can be found in Durkheim's works.2 It was he who first proposed that crime and deviance were not 1See, for example, Kenneth C. Davis, Discretionary Justice: A Preliminary Inquiry (Baton Rouge, Louisiana: Louisiana University Press, 1969); Jerome S. Skolnick, Justice Without Trial (New York: John Wiley and Sons, 1966). 2Emile Durkheim, The Rules of Sociological Method (New York: The Free Press of Glencoe, 1938). 31 necessarily pathological aberrations in behavior, but rather naturally occurring behaviors that served basic societal needs. From this viewpoint, crime was most readily explained by looking at the functions it served in maintaining a healthy, surviving entity.1 The func- tionalist perspective as it came to be known played a very major role in the evolution of present-day theory. It was not, however, until well into this century that Talcott Parsons introduced Durkheim's concepts into American circles.2 Durkheim's original notion of norm- lessness provided the stimulus for Merton's highly influ- ential writings on anomie.3 In developing the theory of anomie, Merton systematically examined the social factors underlying the appearance of normlessness and analyzed the relationships between societal structures and the manifestations of anomie such as poverty, mental illness, crime, and delinquency. The notion of social determinism became increasingly popular as it was continually demon- strated that social factors were, indeed, "related" to various forms of deviant behavior. It was the writings of the functionalists that set the stage for a basic shift in orientation from an individual, pathological View of 1Stuart H. Traub and Craig B. Little, eds., Theories of Deviance (Itasca, 111.: F. E. Peacock, 1975), p. xii. 2Talcott Parsons, The Structure of Social Action (Glencoe, Ill.: The Free Press, 1949)- 3Merton, Social Theory. 32 crime to one emphasizing the role of social structures and processes. The group most responsible for relating crime and delinquency to social factors was the Chicago School. Their writings were also aimed at debunking the patho- logical notions of deviance, but they went further than earlier writers in pointing to specific social struc— tures and processes related to the appearance of deviant behavior. In their ecological studies of Chicago, this group took issue with the original notion of normless- ness and illustrated that in high-crime areas there was a complex and well-understood system of values and norms. The fact that these differed from the mainstream of society was seen as resulting from a long history of discrimination and under—representation in the social institutions.1 The basic framework for delinquency theory was set with the functionalist recognition of the role of deviance in a society and the Chicago School's demon- stration of relationships between crime and delinquency and social characteristics. Most recent writings on the topic can be seen as elaborations of these basic concepts. For example, the many variations of cultural transmission theories all focused on determining the factors within the lower socioeconomic strata that 1Clifford Shaw and Henry McKay, Juvenile Delin- gyency and Urban Areas (Chicago: University of Chicago Press, 1942). 33 accounted for the disproportionate amount of crime and deviance. These theories took the existence of lower- class cultures as a given and energies were directed toward discovering how and why youths came to internalize delinquent subcultural value systems. Following the works of Cohen,1 most delinquency theories of the sub- cultural type began to view deviant behavior as a reac- tion to middle—Class values and statuses that were denied most working- and lower-class youths. In viewing delin- quency from this perspective, some of the emphasis was _shifted from the lower social classes to the middle—class values and institutions that served to frustrate aspiring lower-class youth and, in the process, create delinquency. One of the most systematic examinations of this process was opportunity theory as developed by Cloward and Ohlin. Besides addressing the "unavailability of legitimate means to 'success,'" they also pointed to the "availability of illegitimate means.”2 The theory of differential association was another highly influential factor in the evolution of delinquency theory.3 The popularity of this theory 1Albert H. Cohen, Delinquent‘Boys: The Culture of the Gang (New York: Free Press, 1955). 2Richard A. Cloward and Lloyd E. Ohlin, Delin- quency and Opportunity: A Theory of Delinquent Groups (New York: Free Press, 1970), p. 145. 3Edwin H. Sutherland, Principles of Criminology (New York: J. B. Lippincott, 1947). 34 stemmed from its explanation of both deviance and con- formity, its seemingly explicit and measurable hypotheses, and its emphasis on the interactional process as a unit of analysis. This social learning theory posited that delinquency was manifested on the basis of the frequency, intensity, and duration of associations with delinquency subcultural value systems. The writings of Sutherland, and later Sutherland and Cressey,1 spawned a great deal of research aimed at finding the number of contacts, the levels of intensity, and the length of duration required for the appearance of delinquent behavior. All of the above writings were important because they each in their own way contributed to the evolving sociological perspective on crime and delinquency. But there has been a great deal of criticism of social theories of delinquency from both sides. On the one hand, there are those who claim that by focusing on social factors and absolving the individual of responsi— bility for his or her behavior, delinquency is encouraged. On the other side of this issue are those who voice criti- cism at the class bias underlying most social theories and programs. Schur, for example, has criticized the "apparent assumption that delinquency is primarily a working-class phenomenon, and the analytic preoccupation 1Edwin H. Sutherland and Donald Cressey, Crimin- ology (New York: J. B. Lippincott, 1960). 35 with the delinquencies of working-class youth.”1 The social reform programs that have grown out of the social determinism idea have demonstrated similar kinds of biases in their attempts to modify the sociocultural contexts of lower-class youth. As far as diversion and Youth Service Bureau pro- grams are concerned, the most important theoretical writings are those that have attempted to move beyond the "provision of services" frame of mind. While many pro— grams have incorporated the notion that any kind of service is better than none at all, the theoretical jus- tifications for diversion did not necessarily imply that diversion from the justice system meant the provision of alternative kinds of treatment. In fact, the writings on interactionism and labeling theory that are most commonly quoted as the theoretical base for diversion have often looked on official intervention by any social agency with much skepticism. Schur's recent statements on radical noninter— vention typify this stance.2 His conclusion is that in many cases it is probably more beneficial to minimize the official attention given problem behaviors of youth, which are often only fleeting and likely to be exacerbated 1Edwin M. Schur, Radical Non—Intervention: Rethinking the Delinquency Problem (Englewood Cliffs, N.J.: Prentice-Hall, 1973), p. 92. 2 Ibid. 36 by formal sanctions. The basic contention of this view— point is that the traditional methods of treatment and rehabilitation have shown little in the way of effec- tiveness and have perhaps created delinquency by agitat- ing already sensitive conditions. The bases of these ideas can be seen in a brief review of interactionism and situational causation theories of delinquency. The interactionist, or situational, approach to understanding behavior has been traced to the writings of George Herbert Mead,1 and as Lofland pointed out, the approach contains three primary explanatory units-~the proximate, phenomenological, and processual.2 The notion of proximity merely suggests that the most important variables in the analysis of behavior are those that are spatio-temporally and conceptually close to the behaviors being studied. The phenomenology concept emphasizes the importance of considering the perceptions of situations and conditions by those who are subjected to study. The processual unit of explanation provides the basic model of causation in the interactionist approach and theories of delinquency have begun to incorporate this concept. 1See Herbert Blumer, "The Sociological Implica- tions of the Thought of George Herbert Mead," American Sociological Review 71 (March 1966): 535-47. 2John Lofland, Deviance and Identity (Englewood Cliffs, N.J.: Prentice-Hall, 1969), p. 297. 37 The emphasis on processual factors was evident first in the writings on the sociology of deviance. For example, Becker wrote, "One of the most crucial steps in the process of building a stable pattern of deviant behavior is likely to be the experience of being caught and publicly labeled as a deviant."1 This same line of thought had been presented by Lemert in his discussion of secondary deviance, which was seen as resulting from the formal processes of societal reactions to primary devi- ance.2 Erikson also contributed to the development of this perspective in suggesting that deviance was a prop- erty conferred ppon certain behaviors rather than an inherent quality, and that the critical variable in the study of deviance is the social audience, not the indi- vidual.3 In a widely quoted article, Garfinkel described the processes whereby deviant identities (and consequently, secondary deviance) were created and maintained through exposure to the "degradation ceremonies" inherent in the 1Howard S. Becker, Outsiders (New York: McGraw- Hill, 1963), p. 31. 2Edwin M. Lemert, Social Pathology (New York: McGraw-Hill, 1951). 3Kai T. Erikson, ”Notes on the Sociology of Deviance,” Social Problems 9 (Spring 1962): 307. 38 formal mechanisms of social control.1 Several recent studies have focused on the process by which persons are selected to receive formal sanctions or "treatment." Scheff and Culver found that the clinical diagnosis was the critical variable around which societal reactions were organized in mental health.2 Goffman documented the 3 and process of denunciation in mental hospitals, Hollingshead and Redlich demonstrated the relationship between social class and mental illness labels:4 In the area of law enforcement, Piliavin and Briar discovered that the demeanor of a youth was the most important variable in the decision by police to arrest or release. The role of organizational factors in the decision to arrest has also been explored.6 And the same kind of processual examination of the courts can be found in 1Garfinkel, "Conditions of Successful Degradation Ceremonies." 2T. J. Scheff and D. M. Culver, "The Societal Reaction to Deviance,” Social Problems 11 (Spring 1974): 401. 3Erving Goffman, Asylums (New York: Doubleday, 1961). 4A. B. Hollingshead and R. C. Redlich, Social Class and Mental Illness (New York: John Wiley, 1958). 5Irving Piliavin and Scott Briar, "Police Encoun- ters With Juveniles," American Journal of Sociology 70 (September 1964): 206-14. 6William J. Chambliss and John T. Liell, ”The Legal Process in the Community Setting," Crime and Delin- guency 12 (October 1966): 310-17. 39 Sudnow1 and Cameron.2 Each of these writings has had a significant impact on the development of delinquency theory as it came to be used in the conception of diver— sion programs. Directly in the area of delinquency theory, the writings that most clearly embody the notions from inter- actionism are those by Matza.3 In his theory of delin- quency and "drift," the basic contention was that delinquency was a pattern of behaviors that youth drifted into and out of at various time periods depending on the situation. Matza's primary concern was with the process that ensues following the detection of delinquent beha- vior. He was especially concerned with the stigmatization that accompanies contact and processing by the justice system. The potentially negative effect of being segre- gated and labeled as a deviant is the common thread running through all of the above interactionist discus- sions. Many of the theories and concepts contained in these writings are evident in the discussions of diversion and Youth Service Bureau programs. But, as was mentioned 1David Sudnow, "Normal Crimes: Sociological Fea- tures of the Penal Code in a Public Devender's Office,” Social Problems 12 (Winter 1965): 255-76. zMary Cameron, The Booster and the Snitch (New York: Free Press, 1964). 3David Matza, Becoming Deviant (Englewood Cliffs, N.J.: Prentice-Hall, 1969); Delinquency and Drift (New York: John Wiley, 1964). 4O earlier, an addendum to most diversion programs has been the provision of alternative services, which usually complicates testing the basic concepts of interactionism and labeling theory. Diversion and YOuth Service Bureaus While theoretical and conceptual discussions of diversion have taken place for some time, the implemen— tation of formal diversion programs has only come about in the last decade. In their analysis of the function- ing of the court system, the President's Crime Commission recommended the "early identification and diversion to other community resources of those offenders in need of treatment, for whom the full criminal disposition does not appear required."1 Much interest was generated over the notion of diversion through conferences, workshops, and involvement by state and federal agencies and by the end of the 1960's, diversion programs were springing up nationwide. There were a number of attempts to set down guidelines for the operation of diversion programs. For example, Gemignani wrote the following description of diversion for the Youth Development and Delinquency Prevention Administration: lPresident's Commission on Law Enforcement and the Administration of Justice, Task Fogpe Report: Juvenile Delinqpency and Youth Crime (Washington, D.C.: U.S. Government Printing Office, 1967), p. 332. 41 The strategy calls for the establishment, nation- wide, of youth services systems which will divert youth, insofar as possible, from the juvenile justice system by providing comprehensive, inte- grated, community-based programs designed to meet the needs of all youth, regardless of who they are or what their individual problems may be. But other guidelines which became available were simi- larly written in general terms and it is not surprising that program development in the area of diversion was disjointed. The widely varying structures and functions of diversion programs can be traced to the absence of defi- nitional clarity in the diversion concept. While there is consensus at a general conceptual level that certain cases do not warrant the sanctions of formal criminal processing, much disagreement exists regarding the opera- tional form that diversion programs should take. The term has been loosely equated with numerous criminal justice policies including the general policy of using the least restrictive alternative. It has also been used synonymously with several programmatic concepts such as referral to social services following decriminalization of certain offenses, pretrial release and deferred 1R. J. Gemignani, "Youth Service Systems: Diverting Youth From the Juvenile Justice System," Delinquency Prevention Reporter 8 (1972): 8. 42 prosecution programs, and referral of predelinquents to Youth Service Bureaus.l Closely related to the operational issues in diversion programming are questions regarding who should be diverted. In a widely recognized (although slightly heeded) article by Lemert, it was proposed that diver- sion programs should have as their primary goal the pre- emption of problems that would otherwise enter directly the justice system}; The practice of diversion, however, has often been such that an apparently high probability of entering the system is used as justification for the provision of preventive services via diversion. Critics have been quick to note that one of the unintended con- sequences of diversion programs may be the inclusion of larger numbers of persons in formal social control processes. This dilution of the diversion concept can be seen as symptomatic of the more basic problem regard- ing society's intolerance of differing cultures and lifestyles. Smith noted that "as long as the mainstream of America views deviation narrowly as evidence of pathology requiring some form of control, whether punitive or rehabilitative, diversion is likely to lNora Klamputs, Diversion From the Justice System (Hackensack, N.J.: National Council on Crime and Delin- quency, 1974), p. 3. 2Edwin M. Lemert, Instead of Court: Diversion in Juvenile Justice (Chevy Chase, Md.: National Institute of Mental Health, 1973). 43 remain largely a technique of enforcing conformity by alternative means."1 Besides the definitional and operational issues in diversion programs, there has recently been a great deal of controversy over the constitutionality of certain diversion practices. Much of this controversy revolves around the perception that constitutionally guaranteed safeguards may be bypassed as the scope of nonlegal mechanisms of social control is expanded under the guise of diversion. In reference to the Juvenile Conference Committees in New Jersey, Nejelski described an example of this problem: Instead of facing one judge, the juvenile faced nine. These committees were on occasion dealing with serious offenses—~aggravated homosexual attacks or repeated and serious burglaries. Juveniles were frequently put on probation, although there was absolutely no legal authority; in one county the committees were assessing fines. Another common practice was the ordering of psychiatric or psycho— logical tests and evaluations. All of this activ- ity was being conducted without any form of judicial review. In a similar vein, Harlow suggested that the use of civil processing in diversion programs is an attempt to subject persons whose behavior is held noncriminal 1Robert L. Smith, "Diversion: New Labe1--Old Practice," in New Approaches to the Diversion and Treat- ment of Juvenile Offenders (Washington, D.C.: L.E.A.A., 1973), p. 42. 2Paul Nejelski, ”Diversion of Juvenile Offenders in the Criminal Justice System," in New Approaches to the Diversion and Treatment of Juvenile Offenders (Washington, D.C.: L.E.A.A., 1973), p. 88. 44 to nonpenal measures that are similar to criminal sanctions.1 With the popularity and interest generated by the diversionary concept, it has expectedly fallen short of its overadvocated mission. Because of the rapid proliferation of diversion programs, policy makers and administrators were forced to chart the courses of programs on ad hoc bases, and it is not surprising that their activities took widely differing directions. The growth of the Youth Service Bureau concept has paralleled that of diversion and Youth Service Bureaus have become one of the most widespread types of diver- sion programming. Many of the same definitional prob- lems and operational issues mentioned above have come up in discussions of Youth Service Bureaus. But these projects have continued to flourish and are currently one of the most important subsystems of the juvenile justice system. The introduction of the Youth Service Bureau concept is usually credited to the President's Crime Commission,2 although very similar programs were in operation prior to that time (e.g., the Oakland County, Michigan, Youth Assistance Program predated the 1Eleanor Harlow, Diversion From the Criminal Justice System (Rockville, Md.: N.I.M.H., 1973), p. 8. 2President's Commission on Law Enforcement and the Administration of Justice, Task Force Report. 45 Commission by a decade). The Commission stated the pur— pose of Youth Service Bureaus should be "to provide and coordinate programs for young people."1 The concept was, thus, originally conceived by the Commission as an umbrella agency that would oversee and develop youth service programs in a community, as well as provide direct services where they were found to be lacking. But the National Council on Crime and Delinquency (N.C.C.D.) was another strong voice in the early develop- mental stages of Youth Service Bureaus, and they took issue with the President's Commission on two basic points. First, the recommendations of the N.C.C.D. suggested that the three interrelated functions of Youth Service Bureaus are service brokerage, systems modification, and resource development, which meant strengthening youth- serving agencies and stimulating the creation of missing elements, pg; providing direct services to fill the gaps}3 Second, the N.C.C.D. was much more adamant in emphasizing the noncoercive nature of Bureaus. In the growth of the Youth Service Bureau concept, it has been the model pre- sented by the President's Crime Commission that provided the primary framework for developing programs. lIbid., p. 69. ZSherwood Norman, The Youth Service Bureau (Paramus, N.J.: N.C.C.D., 1972), pp. 12—13. 46 While there has been a great deal of diversity in Youth Service Bureau programs, there is more agree- ment about what a Youth Service Bureau should look like than is the case with diversion programs in general. For example, it was pointed out above that the concept of diversion could be activated at a number of decision points through the justice system. Youth Service Bureaus, however, have been fairly consistent inasmuch as most of their referrals have been diverted prior to court processing. The vast majority of Youth Service Bureau referrals are from the schools, police, or the intake phase of the court process. It is also typical for Bureaus to concentrate their efforts on a specific target group of young, first-time, minor misdemeanant offenders. Nearly all Bureaus have spent much effort developing relationships with youth—serving agencies and the juvenile justice system, and have solicited community support from a wide range of sources. And structurally, there has been a lot of commonality with projects generally having a director, a casework super- visor or assistant director, and several caseworkers, counselors, outreach workers and/or case aides. Programmatically, Youth Service Bureaus have also developed similarly along certain general lines. In a national survey of Bureaus, the U.S. Department of Health, Education, and Welfare found that most of nearly 47 two hundred projects that responded to the survey were "primarily models for delivering direct services to chil- dren and youth."1 It appears from this report that innovative goals like systems modification, resource development and youth advocacy have been deemphasized or ignored in favor of more traditional approaches to solv- ing youth problems. The notion of service brokerage has continued to be a part of the Youth Service Bureau ter- minology, but it seems that this function is often quite similar to the past practices of referring out certain kinds of cases that can be more appropriately handled by other agencies. A final element that is conspicuous in its absence is the coordination and integration of community-wide youth services into a cohesive and encompassing system of youth-serving agencies. The above kinds of factors are among those that have prompted persons to express concerns over the potential use of Youth Service Bureaus as a catchall agency for law enforcement and school problems while creating the illu- sion of doing something constructive for children. The above statements have been intended to point out the origins and developments of diversion and Youth 1U.S. Department of Health, Education, and Welfare, The Challenge of Youth Service Bureaus (Washington, D.C.: Department of Health, Education, and Welfare, 1973), p. 26. 2F. W. Howlett, "Is the USB All It's Cracked Up to Be," Crime and Delinqpency 19 (1973): 491. 48 Service Bureau concepts and programs. Such discussions have tended to be interpreted negatively as a statement of the degree to which programs have failed to live up to their advanced billings. In this study, however, the assumption is that complex social programs are never implemented precisely according to a plan, and the critical role to be played by research and evaluation activities is that of demonstrating what factors-- individual, organizational, political, social, and environmental--impact on the evolution of programs. In terms of exportability, it is just as important to document the processes and results of interactions with the community as it is to assess the degree to which a program attained or did not attain its goals. In the following section, a number of studies representing the range of evaluative attempts are reviewed in order to illustrate (1) the empirical information cur- rently available on diversion and Youth Service Bureau programs, and (2) the refinement of research and evalua- tion techniques. The review is not exhaustive but rather focuses on studies that have had substantial impact on the operation and/or evaluation of Youth Service Bureaus. The studies focus on Youth Service Bureaus and similar diversion programs that have direct impli- cations for the evaluation of Youth Service Bureaus. An attempt is made to elucidate the shortcomings, as well as 49 the contributions, of this group of studies. The rationale for the systems model evaluation to be pre- sented at the end of the chapter draws heavily from statements and conclusions contained in the review. Program Evaluations Summary Statistical Evaluations Internal program evaluations have been roughly comparable nationwide because of the similarities in evaluation guidelines set by state criminal justice planning agencies. State and federal reporting require- ments usually call for summary statistics depicting the service activities of the project and the impact. In the state of Michigan, for example, Youth Service Bureau projects are required to report numbers of youth served by type of problem, referral source, age, school status, and justice system involvement. They also submit data regarding project activities including average length of stay in the program, numbers of cases referred out and terminated, average length of time for providing needed services, and cost figures per client for different services. In addition, impact data are collected on justice system variables such as the number of project youths who are arrested or referred to juvenile court, and nonjustice system factors like school improvement, 50 gaining employment, and reintegration with the family.1 Funded projects have complied with these reporting pro- cedures regularly on a yearly basis, but there is still little in the way of sound evidence that policy makers can utilize in formulating future goals and plans. The failure to generate systematic empirical information is not unique to Michigan nor Youth Service Bureaus, but is characteristic of states throughout the country and social programs of many different types. Part of this difficulty can be traced to the aforemen- tioned policy-making process, which results in vaguely defined and broad—aimed social programs. Funding agen- cies have seldom had the prerogatives and resources to spend sufficient time outlining what finite number of program variations will be considered, how each of these would be most beneficially implemented, and what process and impact-evaluation measures are most appropriate in each case. The failure of social scientists and agency personnel to develop comprehensive assessment procedures has also played a major role in obfuscating the meaning- fulness of social research findings. There were a number of suggestions discussed in the last chapter, but these have not been pulled together in a total evaluation model 1Michigan Comppehensive Law Enforcement and Criminal Justice Plan--1976 (Lansing, Michigan: Office of Criminal Justice Programs, 1976), pp. II-36, 37. 51 that is applicable across a range of programs and pro— vides information about the relative effectiveness of differing approaches to a problem. The results of the above factors are the lack of comparability between evaluation efforts and overstatements regarding the activities and impact of a project. A brief review of two project annual reports will illustrate the problem. The sources of these data will remain anonymous inasmuch as it could be misleading to single out these projects, which are, in actuality, among the most comprehensively evaluated in the state of Michigan. The first report addressed the stated goals of diverting youth from the juvenile justice sys— tem and reducing delinquency as the primary evaluation criterion. After noting a 28 percent decrease in the number of cases set for formal hearing and a 50 percent decrease in the number of youths placed on the juvenile court consent calendar for the year, it was concluded, that "the Bureau was very instrumental in bringing about the reductions." As far as reducing delinquency, the conclusion was that the Bureau was successful since 91 percent of the project youths were not arrested approximately six months subsequent to Y.S.B. interven- tion, 89 percent were not subsequently referred to juvenile court, and 93 percent were not adjudicated wards of the court. The second report contained figures 52 showing a total of 1,283 juvenile court petitions for the year prior to the project, 867 during the Bureau's first year, 923 during the second year, and concluded for the second year that ”the Youth Service Bureau has reduced the total number of petitions from the base year by 28.1 percent." Attention is also called to the fact that the Bureau was successful in diverting from the juvenile justice system 90 percent of the youths referred to it (i.e., only 174 of 1,805 referrals were eventually referred to juvenile court). Methodological constraints in social research almost always preclude conclusive statements as were made above, and the reality of gaining program support was undoubtedly the impetus for making such inferential leaps. But the fact remains that this type of evalua— tion activity only serves to complicate the task of assessing the concept of Youth Service Bureaus. Along with the external pressures on projects to ”look good," the following factors have devalued annual summary sta- tistical evaluations: l. The design and execution of evaluation by persons closely involved with and dedicated to the pro- gram ideals; 2. Failure to describe the intricacies of the intervention, and the implementation of plans; 53 3. Absence of attention to client selection and assignment processes, attrition, and follow-up proce- dures; 4. Lack of concern about the quality of evalua- tion data and absence of attempts to control for any number of highly plausible rival hypotheses (e.g., that statewide trends or changing population characteristics might have been responsible for observed changes). The narrative descriptions to be discussed in the next section have been successful to varying degrees in addressing the second factor. Little in the way of impact data is generated by these studies, but the des- criptions of program development contribute to under- standing the processual side of evaluation. The experi- mental and quasi—experimental studies to be examined following this group are those that have improved the quality of impact data by addressing the methodological concerns listed under three and four above. Some of the studies in this latter group have also begun to attempt integrating the processual and impact components of program evaluation. Narrative Descriptions Qualitative analyses of Youth Service Bureaus have been particularly useful in going beyond formal written statements about what a program was originally 54 to be and do. The effects of these program descriptions can be seen in the array of issues currently being dis— cussed regarding the implementation and operation of diversion projects. Narrative descriptions may focus on different aspects or phases of program development, but the overall effect is a clearer understanding of the translation of abstract theoretical concepts into spe- cific programmatic operations. The first attempt to describe Youth Service Bureaus on a national scale was the previously mentioned survey by the U.S. Department of Health, Education, and Welfare. 11m; following four influences were found to have been important in the development, organization, and primary service of Youth Service Bureaus: (l) the nature of the community; (2) the power base; (3) the funding sources; and (4) the orientation of staff—-the last being the single most important ingredient. It was concluded from this survey of 195 projects that the criti- cal element for success was a committed staff that is "aware of and sensitive to the power structure of the com- munity (and its effect on program).”1 More recently, Schuchter and Polk reported on a study of 45 planning agencies and 372 Youth Service IBureau projects that were contacted by phone for 1U.S. Department of Health, Education, and Welfare, (Challenge of Youth Service Bureaus, pp. 24-25. 55 interviews. In 17 locations, on—site Visits were made for a more thorough follow—up. One of the major findings of this study was that projects lose sight of the more innovative systems change and modification concepts because "a variety of institutional, community and other pressures push them into the delivery of direct services, overresponsiveness to justice system demands, and poten- tially coercive and stigmatizing practices as extensions of the justice system."1 They suggested that to under— stand better the metamorphosis of programs, it will be necessary to specify and measure two sets of interven- ing variables--program operations and intermediate goal attainment. In their words, "As the linking or bridg- ing variables between the program inputs and desired outcomes, these intermediate goal-attainment factors represent the theory of the program, while the program operation factors are the necessary conditions for the theory to work."2 A number of important issues regarding program operational procedures have been raised in other descriptive studies of Youth Service Bureaus and diver— sion programs. I A review by Klein of "several dozen" diversion ;programs pointed to unanswered questions about the lArnold Schucter and Kenneth Polk, Phase I figssessment of Youth Service Bureaus (Washington, D.C.: I;.E.A.A., 1975), p. 101. 2Ibid., p. 120. 56 location of diversion projects, diversion criteria, community tolerance, and funding instabilities, among others.1 The task for evaluators was seen as relating information on these types of factors to the impact measures of success or failure. But the problem lies in the fact that unresolved operational issues arise from the same value conflicts that also complicate the speci- fication of outcome criteria to measure success. For example, in his description of four juvenile diversion programs, Nejelski noted that nearly each proposed suc- cessful (or operational) criterion could be canceled out by a countervailing pressure. As he wrote, "Lack of treatment is favored by laissez—faire liberals and abhorred by interventions. A lack of procedure is decried by civil libertarians and applauded by champions of treat- ment.”2 These issues revolving around personal value sys- tems are only part of the problem underlying the explica- tion of program operations and success criteria. Another factor that comes up frequently is the complex and ambiguous nature of objectives and goals, which necessarily has a detrimental effect on implement- ing and evaluating programs. Klein, for example, noted 1Malcolm W. Klein, "Issues and Realities in jPolice Diversion Programs," Crime and Delinqpency 22 (October 1976): 421-27. 2Paul Nejelski, "Diversion: The Promise and the IDanger," Crime and Delinquency 22 (October 1976): 401. 57 three popular operational meanings for diversion: "true” diversion as the release of cases ordinarily slated for court petition, diversion as used synonymously with referral where larger numbers of offenders are not neceS- sarily taken out of the system, and diversion as it has been practiced for years by police departments in making informal referrals.1 Rutherford and McDermott recognized the same types of diversion practices and termed them "true diversion," "minimization of penetration," and ”screening.”2 A clear specification of goals, types of diversion to be undertaken, criteria for diverting, and operational procedures is the initial step in making diversion programs assessable and replicable. Besides shedding light on the definitional problems and lack of concrete operational guidelines, descriptive studies have been highly successful in demonstrating the importance of organizational and envi— ronmental contexts. On the basis of literature reviews and field research at 13 diversion projects, Rutherford and McDermott stated that a primary issue having to do with organizational milieu centers on the "regulations, rules, guidelines, and informal relationships that guide juvenile justice system personnel in their intra- and 1Klein, "Issues and Realities,” p. 426. 2Andrew Rutherford and Robert McDermott, Juvenile Diversion (Washington, D.C.: L.E.A.A., 1976), p. 3. 58 inter—agency interaction."1 They also cited the impor- tance of examining possible unanticipated consequences of diversion such as widening the net of social control mechanisms, more intense handling of nondiverted offend- ers, ignoring due process rights, and increasing the overall size of the system.2 The analysis of environ- mental and organizational factors is likely to consti- tute the only fruitful method for clarifying the omni- presence of unintended program consequences. The introduction of a social program into a pre- existing community structure will affect a broad range of power structures and interrelationships. It is highly unlikely that a new diversion program or Youth Service Bureau can or should myopically pursue a set of predeter- mined goals. There must be continual readjustments to and reassessments of the operational atmosphere, and it is unrealistic for researchers and policy makers to assume that a program will be implemented precisely according to a preconceived plan. This assumption, however, appears to have been commonplace as witnessed by the paucity of in—depth program descriptions that have gone beyond the formally stated outlines of projects. Cressey and McDermott concluded after detailed program 1Ibid., p. 5. 21bid. 59 analyses of three diversion and Youth Service Bureau programs that research must be addressed to "an under- standing of the organizational realities of bureaucratic professionals engaged in the dual process of implement— ing social ideals and establishing successful profes— sional careers."1 Unfortunately, the environmental and organizational realities of program implementation have been examined in a haphazard manner and guidelines for such processual research are scant. In this section, several issues have been raised in the narrative descriptions of Youth Service Bureaus and similar juvenile diversion programs. The major problems with the studies above are threefold. First, they seldom go beyond surface descriptions to point out specific guidelines for dealing with organizational, environmental, and social factors. Second, there is not usually a framework explicated from which issues are suggested for analysis and upon which a finite set of interrelated issues can be delineated. Third, and most important, is the fact that narrative descriptions are very seldom related to impact data and the implications drawn from these studies are typically based on the personal perceptions of the authors rather than on even 1Donald R. Cressey and Robert A. McDermott, Diversion From the Juvenile Justice System (Ann Arbor, Michigan: National Assessment of Juvenile Corrections, 1973), p. 60. 60 crude data. Writings like those just discussed have been most successful in generating healthy skepticism regard- ing the actual functions of diversion programs. They have, however, contributed little to understanding the impact of such programs on juveniles and the juvenile justice system. Experimental and Quasi- Experimental Studies The studies to be discussed in this section are those in which experimental and quasi-experimental designs as outlined by Campbell and Stanley1 have been utilized in the evaluation of Youth Service Bureau and diversion programs. In recent years, there have been laudable improvements in the design of program evalua— tions and the studies that follow are among those at the forefront in the growing field of evaluation research. What distinguishes these evaluations from those dis- cussed earlier is the integration of methodological controls that serve to enhance the meaningfulness of results. One of the earliest attempts at quasi—experimental research in the evaluation of Youth Service Bureaus was 1Donald T. Campbell and Julian C. Stanley, Experimental and Quasi-Experimental Designs for Research (Chicago: Rand McNally, 1966). 61 that of Duxbury. Included in this evaluation were seven Youth Service Bureau projects in the State of California. Systems-level data were collected for juvenile arrests, and the results ranged from a 42 percent reduction in juvenile arrests (Pacifica Police Department) to a 6 per- cent increase (San Diego Police Department) six months following the introduction of the Bureaus.1 In looking at probation referral statistics, Duxbury used non— equivalent control jurisdictions and again found varying results including one instance in which the San Fernando Youth Service Bureau district experienced a 40 percent decrease in the number of initial probation referrals, but a 34 percent decrease also occurred in the non—Y.S.B. comparison jurisdiction.2 She also collected individual- level data and using project youth as their own control group found that project clients had been arrested fewer times in the period six months after their program involvement as compared to the six months before.3 The major problem with studies of this sort is the ex post facto nature of the evaluation, which precludes the use of selection and assignment procedures that minimize the likelihood of biased samples. This concern:hsparticularly 1Elaine Duxbury, Evaluation of Youth Service Bureaus (Sacramento, Calif.: Department of the Youth Authority, 1973), p. 91. 2 Ibid., p. 93. 3Ibid., p. 97. 62 relevant in diversion programs in which the tendency is to divert the less serious and troublesome cases. Two other studies in which nonequivalent control groups were used demonstrate the potential shortcomings of this quasi-experimental design. Lincoln compared recidivism rates for youths in a West Coast juvenile diversion program with those of a "matched" control group selected from the police contacts 40 days prior to the initiation. She found that the control group had an average of only 1.1 subsequent offenses while the diverted group averaged 1.7 new offenses and more fre— quently had three or more new offenses.1 But since over half of the controls were warned and released, it could be argued that they were probably less serious than the referred project group and the project, thus, was faced with unfair odds in attempting to prevent future delin— quency. Similarly, Elliot and Blanchard used non- equivalent control youths chosen from probation case— loads in their study of a Youth Service Bureau and a similar juvenile diversion project;3 In this case, the selection of youths on probation would likely have biased the research in favor oftfluaproject youths since lSuzanne B. Lincoln, "Juvenile Referral and Recidivism,” p. 11. (Mimeographed.) 2Delbert S. Elliot and Fletcher Blanchard, "An Impact Study of Two Diversion Projects.” (Mimeographed.) 63 it could be expected that having penetrated further into the system, the control group would have more prior and subsequent offenses. Neither of the above studies found significant differences between control and project youths, but the potentially confounding effect of nonequivalent control groups is illustrated. In a recent review of nine of the "more ade- quately evaluated endeavors," Gibbons and Blake discussed two other widely quoted studies in which the control groups were selected in an ex post facto manner. For the evaluation of Project Crossroads (a diversionary program offering employment and counseling services), 191 first-time offenders in the project were compared with two control groups made up of those "routinely processed" youths who were "screened” prior to adjudi- cation and those who were "ultimately adjudicated." The l5-month follow-up check of police records indicated that 31 percent of the project group had been rearrested, 44 percent of the screened group, and 47 percent of the adjudicated cases.1 And questions regarding the com— parability of control groups is especially pertinent in the evaluation of Alternate Routes, a diversion pro- gram of the California Youth Authority that provides short—term individual, group, and family counseling. A 1Don C. Gibbons and Gerald F. Blake, "Evaluating the Impact of Juvenile Diversion Programs," Crime and Delinquency 22 (October 1976): 414. 64 comparison of 142 youths referred to the project in 1972 was made with 190 youths who were arrested in 1970 for similar offenses and the results showed only 6 percent of the project youths and 47 percent of the control group had subsequent petitions filed in juvenile court.1 The meaning of these results is debatable given the fact that nearly 40 percent of the project cases were not even police referrals but had come from parents, schools, and community agencies. The results are further compli- cated by the longer period of follow-up for control cases (up to three years for those arrested in early 1970). For most project cases, the follow-up period was less than 12 months, giving many of the controls up to thrice as much time in which to accumulate subsequent petitions. The most powerful evaluation designs are those in which the researcher is involved at the outset of the project and is capable of instituting random assign- ment (if not selection) procedures. In the Sacramento 601 Diversion Project, youth were assigned randomly to the diversion project four days of the week and the regular court intake unit the other three. The days were rotated each week so that the chances of inclusion in one group or the other were determined only by the 1Ibid. 65 day of the week on which the referral was made. After the program had been in operation for nine months, it was found that the project group had had court petitions filed on only 2.2 percent of the 803 cases, while the control group, which was handled in the traditional manner, had petitions filed on 21.3 percent of the 558 cases. As far as future arrests, a follow—up seven months after initial contact with the project disclosed that 35.0 percent of the project group were rebooked compared with 45.5 percent of the control group and 48.6 percent during a pre-project period. The conclu— sion reached was that the results provided "a powerful demonstration of the value of the diversion concept in combination with the use of family crisis counseling at 1 But other experimental the point of probation intake." studies of similar programs that have also utilized random assignment procedures have found less convincing results. Leidtke et al., for example, looked at the Portland (Maine) Youth Diversion Project, which offered services similar to those of Youth Service Bureaus-- counseling, advocacy, referral, employment counseling, etc. In a three—month follow-up of the diverted (experi- mental) and regularly processed court cases (control 1Roger Baron et al., ”Preventing Delinquency Through Diversion: The Sacramento County 601 Diversion Project," Federal Probation 37 (March 1973): 18. 66 group), no significant differences were found with regard to future delinquency rates, with 20 percent of the 40 controls and 18 percent of the 57 experimentals having been rearrested at least once.1 And in the Ado- lescent Diversion Project (Champaign-Urbana, Illinois), Davidson et al. found that the use of student volun— teers with diverted youths was successful in reducing recidivism whether a behavioral contracting or youth advocacy approach was taken. There were 16 of 25 experi- mental cases who had no police contacts in the one-year follow-up, but none of the 12 control youths fell into the same "success" category. Relatedly, the control group was found to have a higher group mean for police contacts, seriousness of contacts, and number of court petitions filed. But significant differences were not yielded for any of the groups on four questionnaires (social desirability scale, locus of control scale, social labeling scale, and self-report delinquency scale) nor school grade point average, although attendance figures were better for the project groups. 1K. Liedtke et al., Portland Youth Diversion Project (Portland, Maine: Office of Youth Diversion Services, 1974), p. 32. 2William S. Davidson et al., "The Diversion of Juvenile Delinquents: An Experimental Examination," p. 20. (Mimeographed.) 67 The Gibbons and Blake review mentioned earlier summarized another evaluation, which used randomization procedures and illustrated several of the difficulties encountered.1 In Klein's West Coast Study, four alter- native police dispositions were compared--counsel and release, nondetention petitioning, referral with pur- chase of services, and referral without purchase of services. In a six-month follow-up, youths randomly assigned to the latter three groups reported higher rearrest rates than those who were simply counseled and released. The counseled and released group reported as much delinquency involvement on a self-report delin- quency scale, leading Klein to conclude that "being rearrested is largely a function of visibility to the police rather than actual reinvolvement in misbehavior. The police simply do not 'see' as many repeaters among youths that are counseled and released.”2 Systems-level police arrest data were in essential agreement with the self-report materials. One of the first problems Klein faced was obtain- ing the sample. The design originally called for 3600 youths, 800 of whom would be randomly assigned to each group (200 per group). As it turned out, the groups 1Gibbons and Blake, "Evaluating the Impact of Juvenile Diversion Programs." 21bid., pp. 416-17. 68 contained only 81, 82, 88, and 55, respectively. The carefully structured random assignment procedure was also distorted because of numerous instances in which dispositions were meted out on the basis of perceived appropriateness rather than experimental design. This results in the same kinds of interpretation questions as were discussed with regard to the use of nonequivalent control groups. A final area in which Klein found obstacles was the collection of follow—up data. In several cases, youths had left the treatment program for one reason or another, and in some instances youths and families were uncooperative in completing the final phases of the program involving questionnaires regarding behavior. All of these factors cast serious doubt on the meaningfulness of this well-conceived study. The experimental and quasi-experimental studies, along with the narrative descriptions and summary sta- tistical evaluations discussed earlier, illustrate the variety of approaches and problems in evaluating diver— sion and Youth Service Bureau programs. Besides the conceptual and definitional issues presented at the beginning of this chapter, the review of specific pro- gram evaluations has pointed up a number of methodologi- cal concerns. Three particular areas require further consideration: (1) the integration and maintancne of strong evaluation designs, (2) monitoring and documenting 69 necessary program modifications, and (3) improving data— collection procedures and analyses. Using studies in this chapter as examples, some of the specific problems within each category can be illustrated. Much has already been said about the exclusion of evaluative considerations in the process of program formulation and implementation. But even when research— ers have a strong voice in all phases of program plan- ning and the evaluation design is fully integrated into the project, adherence to the research requirements is not guaranteed. In nearly all diversion programs, the selection of participants is severely restricted to the younger and less serious offenders. The problem of maintaining random assignment procedures was alluded to in Klein's study and is likely more prevalent than is reported. And where random assignment procedures are maintained, this is usually possible only for a short period of time, resulting in small sample sizes as was the case in the Liedtke and Davidson studies. A final consideration related to the maintenance of an evaluation design is attrition. In both the evaluation of Project Crossroads and Klein's West Coast study, a number of cases in each group were lost for different reasons at various phases of the programs, rendering the design powerless. It is premature to hope for conclusive 7O evaluation results until progress is made in explicating the importance of the integrity of evaluation designs. The issues having to do with monitoring and documenting programs go beyond the typical monitoring aimed at assuring funds are utilized in the manner speci- fied. Program documentation as implied in the process evaluation writings is a dynamic "case history" of the project rather than a static description of what a pro- gram was supposed to be. Studies such as those of Cressey and McDermott, Schucter and Polk, and Rutherford and McDermott have begun to set guidelines for demon— strating the processes that are used by new programs to manage their environment and garner the necessary commu- nity support. The literature is replete with examples of programs having "gone awry," which in most cases means nothing more than the program has taken a form not consistent with the preconceived notions of researchers and/or administrators. As procedures for documenting developmental histories are developed and utilized, it can be expected that many program shifts will come to be viewed positively as environmental management techniques rather than as "sources of error variance." In the area of data collection and analysis, one of the major obstacles to meaningful evaluations of diversion and Youth Service Bureaus is a lack of con- sistency in the selection of variables to be examined. 71 The reduction of delinquency (which is a primary goal of nearly all diversion programs) has been taken to mean a reduction in police contacts, arrests, court petitions, adjudications, school problems, family problems, or self— reported delinquency, to name but a few of the operational meanings of delinquency reduction. Similarly, there is a wide degree of variation in the duration of follow-up, with some studies such as Liedtke (three months), Duxbury (six months), and the Sacramento 601 Project (seven months) being extremely short. And with regard to analysis, the same kind of situation exists in which one study looks at the mean number of arrests for var- ious groups (e.g., Lincoln), another figures the per- centage of those in each group who have been rearrested (e.g., Alternate Routes), and yet another examines official police data on arrest rates (e.g., Elliot and Blanchard). Because of the incomparability of evalua- tion criteria and the absence of detailed operational histories of programs, there is little in the way of systematic information regarding the relative effec— tiveness of different types of programs for various groups of delinquents in different settings. Rationale for the Study The primary rationale for a systems model approach to evaluation is the broadening of information bases 72 upon which policy decisions are based. Regardless of the specific programmatic area, policy making can be improved as information is accumulated across a wider range of factors relevant to the process of social intervention. The systems model provides a framework within which to plan and organize the gathering of this kind of information. One major reason for the increasing popularity of systems models in research is the poor record of other approaches. While the systems model may be more difficult to deal with conceptually and methodologically, it offers a more encompassing perspective than the more traditional reductionistic approaches. The dearth of information about organizational and environmental influences on social programs bears witness to the nar- rowness of most traditional social research. A systems model approach requires the inclusion of these classes of variables in addition to commonly collected individ- ual variables. And the systems model is also more viable with respect to linking implementation and outcome fac- tors. Because of the emphasis placed on the mutual interactions of variables in the systems model, there is a greater potential for discerning the most important social influences on program development and outcome with this approach. 73 Support for the use of systems concepts can be found throughout the literature. For example, Coates and Miller described a systems model of evaluation as one that incorporates features of the goal—outcome model, but also includes for analysis the relationships and impacts of various intra- and extra-organizational linkages. It has also been suggested that evaluation designs employ measures that are sensitive to changes that might occur in 2 The individuals and relevant social institutions. National Advisory Commission on Criminal Justice Standards and Goals went so far as to state that "evaluation should focus mgrg on changes in institutional responses to youth problems than on behavioral changes in individual youth."3 Most evaluations have been limited to the collec- tion of individual information with organizational sources being tapped only rarely and the environment even less. The use of environmental data would have a significant impact on the quality of evaluation research especially 1Robert B. Coates and Alden D. Miller, "Evaluat- ing Large Scale Social Service Systems in Changing Environments," Journal of Research in Crime and Delin- guency 12 (1975): 93. 2Robert S. Weiss and Martin Rein, "The Evalua- tion of Broad-Aim Programs: A Cautionary Case and a Moral," Annals of the American Academy of Political and Social Science 385 (1969). 3National Advisory Commission on Criminal Justice Standards and Goals, Community Crime Prevention (Washing- ton, D.C.: U.S. Government Printing Office, 1973), p. 80. 74 if used as Hutcheson and Krause suggest as ”factors built into the analysis of trends and interactions rather than statically as in the case of categorizing people as 'social class I or V,' etc."1 Holland and Huntoon empha- sized the importance of environmental data for assessing the activities of a program that are aimed at maintaining internal operations and creating a viable position in the community.2 The study to be undertaken here attempts to incorporate these features and build upon the sounder research designs suggested by the studies discussed in this chapter. It is hoped that the study which follows will shed light on the functioning of the Youth Service Bureaus, as well as illustrate the potential utility of a systems approach to evaluation. 1Bellenden R. Hutcheson and Elliot A. Krause, "Systems Analysis and Mental Health Services," Community Mental Health Journal 5 (1969): 32. 2Winford E. Holland and Harrison H. Huntoon, "The Evaluation of Experimental Social Service Delivery Systems: An Organizational Effectiveness View," Community .Mental Health Journal 10 (1974): 41-51. CHAPTER III METHODOLOGY Research Questions This study is most appropriately defined as an -exploratory-correlational design. Because of the nature of relationships to be examined, the associated problems of measurement, and the lack of experimental control, no attempts can be made to impute causality. Rather, attempts will be made to analyze relationships between organizational and environmental factors and the indica- tors of success or failure for the Youth Service Bureau projects. For the overall evaluation design, the organi- zational and environmental factors will be viewed as the independent variables and the outcome measures as the dependent. Thus, the analytical strategy will be to examine observed variations in the impact measures in terms of the relationships to organizational and environ- mental findings. It should be noted that the explication of independent and dependent variables is somewhat arbi- trary since the status of any specific variable could change depending on the level of analysis (e.g., the orientation of staff would be the independent variable if examining client success rates, but the dependent if 75 76 discussing the effect of the environment on project operations). Youth Service Bureau programs provide a unique opportunity to apply current social research techniques because of the dual aims of affecting change in the juve— nile justice system and preventing individual cases of delinquent behavior. While the success of social programs has typically been assessed in terms of the individual client, the social subsystem through which individuals are processed constitutes an important intermediary focus for evaluation. A systems emphasis is particularly appropriate in this study because the programs are spe- cifically concerned with affecting the processes of organi- zations and agencies in the juvenile justice system. The individual effects of Youth Service Bureaus are partially dependent on.their degree of success in affecting change in the patterns of processing by the justice system (i.e., diversion). The underlying logic of the model to be used in this evaluation is presented in Figure 2. Major research questions addressed in the study are derived from this model. With regard to the analysis of project impact, two general research questions are suggested at the systems level related to ultimate and intermediate effects, respectively. These are: 1. How successful are the Youth Service Bureaus in reducing delinquency? 77 ax maps?» as: .3155 Ho 55.15509 unooopflppa 5 mega 58.58 180050 15% .58 55505 $095 05.5 155:0 5 mowqgo 1 \7 855: moofiopfim 1: 38an mo 8358 -ooospflpaa 5 momqmao QHGH§@#CH $502 pong 855$ 90 >535 830.5 c5 #5555 AI ._ 5mg 55m 55.8ch moonsomon m0 >Qmsdou< >55 0:9 H0 p:m=.050>5 28m no .3220 1 55 300.59 Samoan 503.5850 “Rum 0.55055 ”38.0me 55805 3:01.” H5550 .Hocos coflpcsfic>0 scoazm 005>H0m zusowln.m onsmfim 8056.0. 3550a 9030.5 .50 0.9de II. 85 955550 5880000508 95mm peg :5 95.5809 038mm 5095.: 185 5509:0508 50995 55:25.00 #00 no.5 m0 509 :0 msmcomcoo TI. 8050M 0551509 AR. mo 353353 #00 no.3 H0 :5 p5 5880 358555.. 550cm :5 Ho. EBUHE mummy; ASK—U; IHQZH mHmwg 78 2. How effective are the projects in affecting change in the policies of the juvenile jus— tice system regarding diversion? Related to the impact questions are those having to do with implementation. The general research questions revolving around the internal and external program fac- tors, respectively, are: 3. What is the relationship between organiza— tional characteristics and the ability to achieve stated goals? 4. What are the effects of environmental factors on the organization and its ability to func- tion effectively? It is in response to these questions that the study was designed. The specific research hypotheses will be dis- cussed within each component of the evaluation and, where appropriate, statistical hypotheses will be tested. Research Design and Data Collection Four interrelated components constitute the over- all design of the study with each addressing one of the general research questions above. The first component 'will involve the examination of official crime statistics and is aimed at determining the degree to which each pro- ject has been successful in reducing crime and delin- quency. The second involves a more detailed analysis of crime data, which will supplement the first component and ‘provide information regarding the intermediate goal of jprojects--diverting youths from the justice system. Both 79 of these components fall under the general rubric of impact analysis. The third component focuses on the internal organization of projects and will provide des- criptive information to be used in explaining observed variations in the impact measures. The final component deals with the external environment as it impinges upon the organization of projects and their ability of function effectively. The final two components constitute the implementation analysis. In the remainder of this chapter, each of the major components will be discussed in detail. Within each component, there will be a discussion of the methodology including rationale for the particular design, data-collection procedures, and analytical strategies. Discussions of data collection for each component will include descriptions of the samples of projects to be included in the analysis, specific variables to be exam- ined, and the procedures used in collecting the data. The discussions of analyses for each component will con- tain statements of research and/or statistical hypotheses and explanations of analytical techniques to be employed. The chapter will conclude with a discussion of the limi- tations of the design. 80 Impact Analysis and Time-Series Designs The two major components of the impact analysis section are aimed at answering questions related to the effectiveness of Youth Service Bureau projects in affect- ing change in the processes of the juvenile justice sys— tem. First, an assessment will be made of the degree to which projects have been successful in bringing about a reduction in crime, the ultimate effect (See Figure 2). Second, data will be examined in order to assess the extent to which projects have been able to accomplish the intermediate goal of diversion. The underlying logic of the impact analysis is that reductions in crime and delinquency rates (as measured by actual offenses) which might be attributable to the projects should follow in time earlier indications that diversion (as measured by delinquency arrests) has taken place since this is the primary intermediate goal in the program logic of the projects. Before addressing the specific designs of the crime reduction and diversion components, some discussion is necessary of the general research model to be used in this section, and the data that will be used. In the analyses of crime reduction and diversion, 'the time-series model will be employed. Time-series is a quasi-experimental design, which has been defined by Campbell and Stanley as follows: 81 . . attempts by a researcher to introduce something like experimental design into his scheduling of data collection procedures (e.g. the when andtfluawhom of measurement) even though he lacks full control over the scheduling of experimental stimuli (the when and the whom of exposure and the ability to randomize exposures) which makes a true experiment possible. A quasi-experimental approach is necessary here because of the ex post facto nature of the study wherein neither the selection of sites to receive Youth Service Bureau funding nor the exposure of youths to treatment at the specific projects meets the criteria for a true experi- mental design. The time-series designs are especially well-suited to measuring change in complex social systems (e.g. the justice system) where activity data are recorded on a regular basis. The quality of these data is, of course, of critical importance and will be discussed momentarily. Time-series designs are extensions of the classi- cal "pretest-posttest" designs in which one measurement is taken before and one after the intervention to deter- mine the extent to which there has been change. In time- series designs, measures are taken repeatedly before and after the intervention and observed changes following intervention can then be judged as "either the effect of the intervention or merely the progression of an evolving 1Campbell and Stanley, Experimental and Quasi- Experimental Designs for Research, p. 34. 82 and dynamic process unaffected by the intervention."1 The validity of these judgments depends on the extent to which controls are incorporated that rule out the rival plausible hypotheses that could be used to explain the 2 results. In its basic form, the single time-series design can be represented as follows: 01 02 03 X 04 05 06 where O signifies the repeated measurements or observa- tions and X denotes the treatment intervention. The more appropriate notation in this study, since the interven- tions or program are continuous, is the following: 01 02 03 X 04 X 05 X 06 where measurements are taken continually at the same intervals throughout the life of the program. In the case of the multiple time-series design where comparisons are to be made between two series of measurements, the proper notation is: 01 02 03 X 04 X 05 X 06 O1 O2 O3 O4 O5 06 lGene V. Glass, Victor L. Willson, and John M. Gottman, Design and Analysis of Time-Series Experiments (Boulder, Colorado: University of Colorado, 1972), p. 1. 2Campbell, "Reforms as Experiments." 83 where the second line represents the comparison series without the intervention. Examples of the use of time- series designs in social research can be found in the Ross, Campbell and Glass1 study of the effects of a breathalyser law on drunken driving in England, and the studies by Glass2 and Campbell and Ross3 on the effects of a speeding crackdown in Connecticut. A variety of intervention effects may result from the time-series designs.4 Some of these are: A. an abrupt change in level I B. an abrupt change in direction _________I C. A delayed change in level I 1Lawrence H. Ross, Donald T. Campbell, and Gene V. Glass, ”Determining the Social Effects of a Legal Reform,” American Behavioral Scientist 13 (1970): 493-509. 2Gene V. Glass, "Analysis of Data on the Connec— ticut Speeding Crackdown as a Time-Series Quasi— Experiment,” Law and Society Review 3 (August 1968): 55-76. 3Donald T. Campbell and H. Laurence Ross, "The Connecticut Crackdown on Speeding: Time-Series Data in Quasi-Experimental Analysis," Law and Society Review 3 (August 1968): 33-53. 4Glass, Willson, and Gottman, Design and Analysis of Time-Series Experiments, p. 46. 84 D. a delayed change in direction \ E. a temporary change in level I I F. a temporary change in direction ________I \\. G. a decaying change in level I.__-"/’ H. an accelerated change in direction _\ In both the crime reduction and diversion components, the I general analytical strategy will be to examine annual crime trends using a multiple—group time—series design and then to follow up with a more detailed look at monthly figures using a single—group time-series design where statistical tests of significance will be performed. As was mentioned above, the time-series approach is an expansion of the pretest-posttest design. The multiple time-series design is the similarly expanded version of the nonequivalent control group design. In both designs, the use of a "nontreatment" group is impor- tant because it provides a control for the rival 85 hypothesis that history (or the presence of a broad range of influences occurring at the same time as the interven- tion) "caused" observed changes. In the case of the Youth Service Bureaus, this means that the delinquency rates of comparison jurisdictions could be expected to have been influenced by the same historical factors that might be used to explain observed changes in project jurisdictions (e.g., if the size of the juvenile population has decreased statewide). The examination of annual data within the multiple time-series design will provide indi- cations of project successes in the areas of crime reduc- tion and diversion. Because of the problem of data instability over time (i.e., random fluctuations), the annual level data used in the multiple time-series design do not provide a sufficient number of data points to determine the signifi— cance of observed changes in post-intervention trends. The use of monthly data results in a much larger number of data points and enables such determinations to be made. Time and financial constraints made it necessary to limit collection of the much larger volume of monthly data to project sites only. Thus, the single—group time-series design will be employed for the monthly level analyses of crime reduction and diversion. 86 Discussing the problem of data instability and statistical analyses of time—series data, Campbell wrote, The plausibility of the hypothesis that instability accounts for the effect can be judged by Visual inspection of the graphed figures or by qualitative discussion, but in addition it is this one threat to validity which can be evaluated by tests of signifi- cance. In the multiple time—series analysis, a decrease in the rates of actual criminal offenses or delinquency arrests in the absence of similar decreases in comparison juris- dictions would provide initial evidence that the Youth Service Bureaus were successful in reducing crime and diverting youths from the justice system. If these decreases are found to be statistically significant in the single-group time-series analysis of monthly data, the evidence is stronger yet that the projects were effective. The statistical model upon which the analysis of nuanthly time-series data is based was developed by Box zuad Tiao.2 The objective of a statistical analysis is to separate out the true effect of an intervention on a tixne-series from random shocks and determine whether the irfl:roduction of the intervention decreased, increased, 1Campbell, "Reforms as Experiments," p. 117. , ZGeorge E. P. Box and George C. Tiao, "A Change '1“ ‘Zhe Level of Non—Stationary Time-Series," Biometrics '52 ( 1965) : 181-92. 87 or did not affect the variable on which time—series data were collected. According to Glass, . . . the statistical analysis answers the question of whether the observations following the enactment of a law (or introduction of a program) are simply a continuation of the time-series of the preenactment observations or whether they have shifted up or down from the general level of the preenactment time- series. Thus, the basic function of the statistical analysis of monthly data is to determine the general level and slope of the time-series data before the intervention in order that comparisons can be made with the level and slope of the post-intervention data. Because of the data insta— bility problem, various mathematical properties of the time-series data must be examined to determine which model the data fit so that the appropriate time-series analysis can be performed. The model identification process involves several complex mathematical concepts and func- tions and will be discussed only briefly here. Thorough ctiscussions of this procedure can be found in Box and Tiao2 and Glass, Willson, and Gottman.3 There are three properties of the data which are innportant in the model identification process. The order 1Glass, ”Connecticut Speeding Crackdown," p. 66. 2Box and Tiao, "A Change in the Level of Non- St at ionary Time-Series . " 3Glass, Willson, and Gottman, Design and Analysis Mime-Series Experiments, pp. 97—101. 88 of differencing is the first and indicates the number of times differencing (subtracting each observation from the one following it) must be carried out to reduce trends in the preintervention data to a constant and stationary level. While the time-series analysis to be used here allows for any of four orders of differencing to be used, social science data will seldom require more than a first order of differencing which removes linear trends (second order removes quadratic trends, third order cubic trends, and fourth order quartic trends). The second and third parameters included in the model identification process are the orders of autoregres- sion and moving average. Time-series analysis is based on a multiple regression model and the order of auto— regression and moving averages correspond to and function as beta weights in a multiple regression equation. Both the autoregressive and the moving average functions are reflated to issues of interdependence and instability of nusasurements in a time-series. The autoregressive process gauges the extent to which a given data point is affected tn; the measurementS‘preceding it (e.g., what effect the Chalinquency arrest rate for burglary in January has on the Efiune rate for February). The moving average process zrttempts ‘UD take into account the effects of past random Shocks to the time-series on current observations (e.g., what effect past changes in pOpulation make-up have on current observations). The overall purpose of the model identification process is to identify and correct for data instabilities which complicate the calculations of general levels of the time-series prior to and follow- ing the introduction of an intervention. These adjusted, general levels provide the basis upon which determinations of pre-post changes are made. Model identification is accomplished using the computer program CORREL at the Michigan State University Computer Center.1 The CORREL program computes autocor- relation and partial autocorrelation coefficients which can be used to determine orders of differencing, auto— regression, and moving average. The order of differenc- ing is determined by determining the lowest order (with zero being the lowest) in which all but the lag 1 auto- correlation (the correlation between two successive data points) do not differ significantly from zero. The deter- mination of autoregressive and moving average orders is made by examining the autocorrelation and partial auto- correlation coefficients for the order of differencing 1For a detailed discussion of the content and use of this program, see Lynn D. Miller, Time-Series Analysis (East Lansing, Michigan: Criminal Justice Systems Center, Michigan State University, 1976). 90 which has already been determined. Miller presents a table that summarizes the identification of these last two parameters.1 The model identification process is a complex one requiring an understanding of trigonometric functions and higher level mathematics. To follow this procedure for each of the variables at all sites would be impossible. It was, therefore, necessary to determine the model that most closely fit crime statistics in general and use this model for each specific crime variable. This was accomp- lished by examining the correlograms and autocorrelations for a limited number of randomly chosen crime variables. A review of the literature on crime statistics was also undertaken to shed light on this question. In both instances, indications were that the seasonal model (0,1,1) was the most appropriate.2 The actual time-series analysis is accomplished using the TSX computer program, which is available at the Michigan State University Computer Center.3 Four parameters of a time-series data set are tested in TSX. 1Ibid., p. 43. ZSee, for example, Michigan Uniform Crime Report-— 1975 (Lansing, Michigan: Department of State Police, 1976), pp. 16-17. 3For a detailed description of the program, see Miller, Time-Series Analysis. 91 First, the general level of the pre-intervention data is calculated. With crime statistics, this phase of the analysis is not particularly useful since the level is tested for difference from zero and nearly all sites experience crime rates significantly above zero. Second, the change in level is computed by subtracting the last pre-intervention data point from the first post- intervention point. Both of these figures are adjusted to minimize the impact of seasonal fluctuations and the influence of earlier random shocks to the time-series. The t-statistic tests the difference between the observed level of change and no change which would be expected if the projects had no effect. Thus, where there is a negative change in level accompanied by a sufficiently large t-statistic (-l.67) it would be concluded that the project under scrutiny has had a significant impact on the particular variable. Third, the drift (or slope) of the pre-intervention trend line is calculated and tested for its difference from zero. This parameter is not directly relevant in this study except for its use in the computation of change in drift. The fourth and final parameter tested in the TSX program is the change in drift which is defined as the slope of the pre- intervention line minus the slope of the post-intervention line. A negative change in drift indicates that a rising 92 crime rate was retarded or a falling rate hastened while a positive change in drift means that a rising crime rate was hastened or a falling rate retarded following the intervention. Where a significant negative change in drift is observed, it will be viewed as evidence in sup- port of the effectiveness of projects. There is reason to expect that perhaps a project would not have a signifi- cant impact on the overall level of a specific crime vari- able although the rate at which the variable had been increasing would show a marked decline. In order to determine the significance of the t-statistics, degrees of freedom are computed by sub- tracting the number of parameters tested (four) from the total number of data points in the time-series (usually 56). Using a one—tailed test since the hypothesized direction of change is negative, examination of a t-table will indicate that with 50 degrees of freedom it is necessary for the t-statistic to exceed -l.67 for sig- nificance at the .05 level (i.e., five times in a hundred one would mistakenly accept a false hypothesis). The num- ber of pre- and post-intervention points will vary from Project to project since most projects started at different times and the monthly data were collected for only one time Period--January, 1972, through August, 1976. Before 93 detailing the methodologies for each component of the impact analysis section, a brief discussion of official crime statistics is in order because of the controversial nature of the data. In both components of the impact analysis section, the primary data source will be the Uniform Crime Reports compiled each year by the Michigan State Police in con- junction with the U.S. Federal Bureau of Investigation. These official police data have been the focal point of a great deal of criticism, but much of the concern revolves around the misinterpretation and misrepresentation of the information. For example, Chambliss and Nagawasa con- cluded that official crime statistics were useless as indicators of "actual deviance."1 From a naturalistic perspective, however, the extent of "actual deviance" is infinite and the vast proportion of it is tolerated and absorbed by individuals and the community as it must be (informal social control mechanisms). Therefore, the small proportion of all illegally defined behavior which exceeds the tolerance levels and comes to the attention of authorities is correctly the primary concern of those interested in formal mechanisms of social control such as the justice system. For those interested in discovering lWilliam J. Cambliss and Richard H. Nagawasa, "On the Validity of Official Statistics," gournal of Research in Crime and Delinquency 6 (1969): 71-77. 94 the "true" picture of crime, official statistics are obviously inadequate. The complexity of social data, coupled with the emotionalism surrounding the crime problem, has also resulted in the manipulation and distortion of official crime data for self—serving and politically motivated ends. An example is the U.S. Uniform Crime Report pub— lished each year by the F.B.I. It has been charged that the F.B.I. generates the maximum amount of terror from these reports by presenting only the upward side of crime charts, using "crime clocks" to show a progressively shorter time period between the commission of crimes without correcting for the large growths in population, and compiling an index of "serious crime" in which crimes like ”joyriding" and "entering a building without permis- sion" are given equal weight with crimes of murder and rape.1 Theoretically, the crimes of violence about which there is the greatest personal concern could come to an abrupt end, while an increase in the incidence of stealing wheelcovers (larceny over $50) could produce an alarming rise in the index of serious crime. The same kind of manipulation can be seen in the use of one set of data by law enforcement, court, or correctional agencies to 1National Institute of Mental Health, Criminal Statistics (Washington, D.C.: Department of Health, Education and Welfare, 1973). 95 demonstrate the effectiveness of an existing program and the use of another set to illustrate the need for a new program. Besides the potential misuses and misinterpreta— tions of crime data, concern has been expressed regarding the reliability and comparability across time and juris- dictions.1 Because of ambiguous definitions and individ- ual officer discretion, there is skepticism that crime- recording procedures vary so widely that they are meaning- less. However, as Wheeler pointed out, variability in the manner of responding to different crimes in different jurisdictions is an inherent characteristic of the jus- tice process and a legitimate area for investigation. Relatedly, Skogan suggested that the pressures to over- or under-report and record crime are likely to be distrib- uted across time and jurisdictions in a random fashion such that relative comparisons are justified although the "true" levels of crime may be obfuscated in all 1D. J. Black, "Production of Crime Rates," American Sociological Review 39 (1970): 733-48; T.<3.Smith, "Crime Statistics--Can They Be Trusted," American Criminal Law Review 11 (1973): 1045-86; J. E. Price, "A Test of the Accuracy of Crime Statistics," Social_§roblems 14 (1966): 214-22; Thorsten Sellin and Wolfgang, The Measurement of Delinquency_(New York: John Wiley and Sons, 1964). ZStanton Wheeler, "Criminal Statistics: A Refor- mulation of the Problem,” Journal of Criminal Law, Crimin- ology and Police Science 58 (1967): 317-24. 96 jurisdictions.1 It is on the basis of this reasoning that the official crime data are employed in this study. Crime Reduction The total sample of Youth Service Bureau projects to be included in the overall evaluation design is 13 (see Table 1). These are all of the projects receiving funds during 1976 under the Youth Service Bureau element in the state plan of the Michigan Office of Criminal I Justice Programs. In Appendix A, program descriptions of each project can be found. These were abstracted from the original grant applications submitted to the Office of Criminal Justice Programs by each subgrantee. For both components in this impact analysis sec- tion, seven of the sites can be included (Calhoun, Berrien, Genesee, Van Buren, St. Clair, Newaygo, and St. Joseph Counties). One project ceased operations during the study (Kalamazoo), another is a multi-county project not amenable to time-series analysis (Alpena), and two others began too late for the collection of post-intervention data points (Shiawassee and Grand Traverse). One of the programs began at the time from which data were col- lected, resulting in an absence of pre-intervention 1W. G. Skogan, "The Validity of Official Crime Statistics: An Empirical Investigation," Social Science Quarterly 55 (1974): 25-38. 97 NNXNNX NNNNNN NNNNMNN NNNMMMN 3:000 gag 080803 58:00 0:03.: #:0de 3:000 00.0229 0:05 3:0 08022:... 5:000 0:00.30 00: 0:00 3:500 000003030 00095 3:160 9500:: ”30.300 003:. ammspo ngmmow .pm 0:05m 00.59 mu::00 ow>m$z 0:30 03:? >038 $.30 . pm :onsm 9:00 3:500 8:5 :«> 30: 30: 3:500 8:05.00 00:0 0303 3:300 0000:00 05E 3:000 :0H90m pong 89:00 0:0.“ ”500:0: p:0:00000< 50:00:: 3308:0925: :oH Hm 0:020 5:000:38 0:300: 3:3 ”BEE 0030:: :oH p: ”308395 00: 0H5 no: :9.” 0.030 :0 :0.“ 926% 00030:: 90005 mmp:m .:wH000 :0H000H0>0 Hamh0>o m0 >A088301|.H 0Hnwe 98 (East Detroit), and one project is a part of a larger youth-serving system which has functioned for 20 years (Oakland). In order to utilize the multiple time-series design, it was necessary to select a comparison juris- diction for each of the seven sites. Since all of the projects have a county—wide focus, the unit of analysis in this component is the county. Thus, "matched" non- Youth Service Bureau comparison counties were selected on the basis of geographical location, total population (1970), and median family income (1969) as reported in the Michigan Statistical Abstract (1974).1 Comparison counties were chosen which were the nearest match to the Youth Service Bureau counties located in the same geo- graphic region of the state. These demographic data are contained for each of the '7 pairs of counties in Appendix B. While the Youth Service Bureau counties and the comparison counties can by no stretch of the imagi- nation be viewed as equivalent, the comparisons serve as an important checkpoint for examining the long-term delinquency trends in the Youth Service Bureau areas. Official police data were collected for the 14 counties from the Uniform Crime Reports compiled annually by the Michigan State Police. Using the form in lMichigan Statistical Abstract--l975, 10th ed. (East Lansing, Michigan: Michigan State University, 1975). 99 Appendix C, annual data were coded from computer print- outs maintained by the State Police which summarize the activities of the law enforcement agencies in each juris- diction. Six years annual data were collected (1971-1976.) The monthly data to be used in the second step of the crime reduction analysis were also collected from the Michigan State Police. These data were put on tapes and trans— ferred to the Michigan State University Computer Center, where files were created upon which the statistical analy- ses were performed. Fifty-six months of data were col— lected representing the period from January, 1972, until August, 1976. The Michigan U.C.R. data are submitted by 498 of the 510 law enforcement agencies in the state, and it is estimated that the data represent 99.5 percent of all the officially recognized crime in Michigan.1 The variables to be analyzed in this component are those most directly pertinent to the question of crime reduction-~actual or founded offenses. These are to be distinguished from the arrest data which will be used in the next component addressing the issue of diver- sion. The specific offenses which will be examined are burglary, larceny, and vandalism. These three offenses were selected for analysis because they are the most common offenses among delinquents and along with the lMichigan Uniform Crime Report, 1975, p. 4. 100 status offenses (those acts which are criminal only when committed by a juvenile) constitute over 60 percent of all juvenile arrests in Michigan.1 They also represent the offenses most common among Youth Service Bureau cli— ents. Status offenses in the U.C.R. data (runaway and curfew/loitering) are not included in the actual-founded section although arrest data are collected for the two and will be used in the next component. Actual or founded crime is that proportion of all reported crime in which a determination was made that a crime had, in fact, been committed. The arrest data to be used in the analysis of diversion are a subset of actual crime wherein suspects were formally arrested. It is a premise of this impact analysis section that the actual offense data represent the most accurate estimate of crime levels and should be used in analyzing the question of crime reduction. Furthermore, the arrest data are viewed as being the best indicator avail- able of law enforcement activities and therefore, the appropriate level of data to use in examining the diver- sion question. Since nothing is known of the offenders responsible for all actual crime, no age breakdown is possible, but the assumption can be made that a signifi- cant reduction in delinquency rates would show up in the 11bid., p. 27. 101 overall actual crime picture. In order to facilitate inspection of the annual data, all figures have been calculated into rates per 1,000 total county population (based on 1970 Census). To summarize, the first of a two-step analytical procedure for answering questions regarding the success of Youth Service Bureau projects in reducing crime will be a multiple time-series analy- sis of annual data on three actual crime categories for seven project counties and their comparison counties during the time period for 1971-1976. The research hypothesis in this component of the study is that the Youth Service Bureau counties will experience a reduction in the level of delinquency fol- lowing the introduction of the Youth Service Bureau pro— jects. Each of the variables just discussed was selected for inclusion in the analysis because of its relevance to this hypothesis. For each variable, the analytical procedure will be the same. Since the use of annual data does not provide enough data points for a statis- tical analysis of post—intervention change, the multiple time-series data will be visually inspected to determine those instances where it appears that changes in the trend line have occurred. Inspection of the data will be focused on ascertaining whether or not there has been a change in the level and/or slope of the trend lines. As was previously mentioned, a change in level 102 suggests that the jurisdiction went from an increasing delinquency rate to a decreasing one or vice versa while a change in slope implies that a rising or falling delinquency rate was either hastened or hindered after the introduction of the Youth Service Bureau. To examine the data for these changes, it is suggested that the first calendar year in which a project has functioned for six months be used as the intervention point (Calhoun, 1972; Berrien, 1973; Genesee, 1973; Van Buren, 1974; St. Clair, 1975; St. Joseph, 1976). The six- month period allows for the time it takes a project to become fully operational, and delinquent behavior which has been prevented through contact with the projects should be reflected in the official crime statistics within this period given the rapidity with which delin- quent careers typically develop. The second step in the analysis of crime reduc- tion involves the use of a single group time-series design which provides a more in-depth look at the statis- tical significance of observed changes in the annual multiple time-series data. When concomitant variation is observed between the introduction of a Youth Service Bureau and changes in a trend line, the analyses in this phase can be used to verify the significance of an observed change. Each of the variables included at the 103 annual level will also be analyzed in this second step since annual level data could conceal significant monthly shifts. It should be noted the analyses of monthly data do not improve the power of the design for generating causal statements. They will, however, indicate whether observed changes in variables are sta- tistically significant. The same seven Youth Service Bureau jurisdictions will be included in the statistical analyses of monthly data on the same three actual crime variables. Because of the problem of aggregating all reporting jurisdic- tions in a county, the central city in which the project is located will be the unit of analysis (i.e., Battle Creek, Benton Harbor, Flint, Paw Paw, Port Huron, White Cloud, and Three Rivers). This is not considered trouble— some because the projects focus their energies in the central cities. Since the primary concern in this analy- sis is determining whether observed changes in the levels of trend lines are statistically significant, it is not necessary to include the comparison jurisdictions. For each of the three actual crime variables at the seven sites, t-tests and significance levels will be presented for the changes in level and drift as outlined earlier. Since the monthly data were collected for one time period (1/72-8/76), and projects began at various times, the number of pre- and post-intervention data 104 points for each site will be different. Allowing six months after initial funding for start-up time and the recycling of clients, pre- and post-intervention points for each site are as follows: Battle Creek (2-73)-l3 and 43; Benton Harbor (l-74)-24 and 32; Flint (l-74)-24 and 32; Paw Paw (l-75)-36 and 20; Port Huron (7—75)-42 and 14; White Cloud (l-76)-48 and 8; and Three Rivers (l-76)—48 and 8. In Battle Creek, it was necessary to set the intervention point at seven months after initial funding so that the required 13 pre-intervention data points were present for the seasonal adjustments. Neither White Cloud nor Three Rivers data could be adjusted for seasonal fluctuations because of insuffi- cient post-intervention points. The interpretation of results from the monthly analysis in this component, as well as the next, will follow the steps outlined earlier in the chapter. Diversion In the first component where the focus was on crime reduction, the most appropriate type of U.C.R. data was the actual or founded offenses which are the best indicators of crime levels. The focus on diversion in this component suggests the use of arrest data since they represent the best indicator of police processes and diversion is aimed at affecting police processes. 105 The same seven Youth Service Bureau counties and their comparison counties will be included in the annual level analysis of diversion, and the same seven Youth Service Bureau central cities in the statistical analy- sis of monthly data. The variables to be included in the annual multiple time—series analysis of diversion are those which should reflect the effects of diversion on the juvenile justice process. The variables for which data will be analyzed provide a comprehensive picture of the delinquency situation in each jurisdiction, and they were selected particularly because they represent areas in which Youth Service Bureaus focus their energies. Total delinquency arrests (under 17) will be examined to provide an overview of the delinquency situation. Then the five most common delinquent offenses--burglary, larceny, vandalism, runaway, and curfew/loitering--will be examined. These are hypothesized to be the most sen- sitive measures of diversion activities not only because they are the largest offense categories, but also because the Youth Service Bureaus concentrate on the same types of offenses. Finally, data on police ggfgg— gals to juvenile court will be analyzed to determine whether or not the projects are having an impact further into the justice system. 106 For all of the above delinquency arrest data, the figures represent the arrests of persons under the age of 17 (the statutory definition of a juvenile in the state of Michigan). As in the last component, all figures have been transformed into rates to facilitate inspection of the data by correcting for population dif- ferences between counties. Since this component focuses on delinquency arrests rather than overall extent of actual crime as was the case in the first component, rates were calculated per 1,000 juveniles (ages 7-16) in the county rather than per 1,000 total population. Also, adjustments for the changing size of the juvenile population at risk were made by using the number of youths between 7 and 16 as the base for 1970, the number between 6 and 15 (from the same 1970 Census) for the 1971 rates, 5 and 14 for the 1972, and so on. The research hypothesis in this component is that if Youth Service Bureaus have been successful in divert— ing youths from the juvenile justice system, delinquency arrest rates will decrease. And similar to the annual multiple time-series analysis of crime reduction, inspec- tion of the data will be aimed at determining whether changes of trend lines in either level or slope are apparent. Although it would be expected that the effects of diversion should show up earlier than the effects of crime reduction since it is the intermediate goal 107 preceding crime reduction, this distinction does not affect the year specified as the intervention point in the last section. Even if we only require three months of time in operation during a calendar year to qualify that year as the intervention point (rather than six as in the crime reduction component), the year does not change for any of the sites. The monthly time—series analysis of diversion will be carried out in the same follow—up manner as it was in the crime reduction component. Each of the seven variables discussed above will be analyzed for each of the seven Youth Service Bureau central cities. In addition, data were collected from four major projects (Berrien, Genesee, St. Clair, and Macomb Counties) on total number of court referrals. This provides an oppor- tunity to examine the impact of projects on overall juvenile court activities. As before, the monthly data in this component represent the central city except for the total juvenile court petitions which are county—wide in scope. The number of data points (months) for total delinquency petitions varied and will be presented along with the results. Intervention points for this variable will be the same as for the other variables in this component (see below). The hypothesis to be tested for each site on each variable is that there will be a significant 108 reduction in the levels and/or slopes of delinquency trends following the initiation of diversion activities by the Youth Service Bureau projects. Because diversion is being viewed as an intermediate goal preceding the accomplishment of the overall goal of crime reduction, the intervention point for the monthly analysis in this component was set at three months after initial funding began. This means that the pre- and post-intervention points for the sites are the following: Battle Creek (2-73)—13 and 43; Benton Harbor (lO-73)-21 and 35; Flint (lO-73)—21 and 35; Paw Paw (lO-74)-33 and 23; Port Huron (4-75)—39 and 17; White Cloud (10-75)-45 and 11; and Three Rivers (10-75)-45 and 11. As was the case in the last component, the intervention point for Battle Creek was moved back (from three to seven months here) so that seasonal adjustments could be made. Also, the removal of seasonal fluctuations could not be carried out for White Cloud and Three Rivers data because of too few post-intervention points. The analysis and interpretation of the monthly arrest data will follow the same lines as that of the monthly actual offense data in the crime reduction com- ponent. The t-statistics and confidence levels for changes in level and drift will be examined to determine Whether changes in arrest patterns following the intro- duction of a project are significant. And,1x>reiterate, 109 a significant decrease in the level of a variable will be interpreted to mean a reduction in the absolute level of the particular variable following the intervention while a significant decrease in drift or slope will be taken to mean that the rate of increase in the variable was retarded. It is possible that the absolute level of a variable may be unaffected if it is high to begin with, but that the rate at which it was increasing is signifi- cantly lessened. The following summarizes the methodology for the two components in the impact analysis section: I. Crime Reduction A. Annual Multiple Time-Series (1971-1976)--rates per 1,000 total county population 1. Variables a. actual burglary b. actual larceny 0. actual vandalism Sites-Comparison Sites (intervention points): (Calhoun—Jackson (1972), Berrien-Muskegon (1973), Genesee-Saginaw (1973), Van Buren- Allegan (1974), St. Clair-Lapeer (1975), Newaygo—Mecosta (1976), St. Joseph—Branch (1976) B. Monthly Single Time-Series (1/72-8/76) 1. 2. Variables--same as I—A-l Sites (intervention points--six months after funding): Battle Creek (2-73), Benton Harbor (1-74), Flint, 1-74), Paw Paw (1—75), Port Huron (7-75), White Cloud (1-76), Three Rivers (l-76) II. Diversion A. Annual Multiple Time—Series (1971-1976)--rates per 1,000 juvenile population under the age of 17, 1. adjusted by year Variables a. total delinquency arrests b. delinquency for burglary 110 delinquency for larceny delinquency for vandalism delinquency for runaway . delinquency for curfew/loitering g. referrals to juvenile court 2. Sites-Comparison Sites (intervention points) —-same as I—A-2 B. Monthly Single Time-Series (l/72-8/76)—- juveniles under the age of 17 l. Variables-~same as II-A-l 2. Sites (intervention points--three months after funding): Battle Creek (2-73), Benton Harbor (10-73), Flint (IO-73), Paw Paw (IO-74), Port Huron (4-75), White Cloud (10-75), Three Rivers (10-75) HJCDQ-O Implementation Analysis As was mentioned earlier, the overall analytical procedure for the study will be to move from the impact analysis of ultimate and intermediate effects to the examination of organizational and environmental factors that may have influenced the success or failure of a project in attaining stated goals (refer back to Figure 2). As far as the overall study is concerned, the variables in the implementation analysis section will be viewed as the independent variables and attempts will be made to relate these findings to those resulting from the impact analysis. Implementation analysis is that portion of the previously defined systems model of evaluation that focuses attention on the processual aspects of program evaluation. It is aimed at answering questions about the political and social interactions necessary in the 111 introduction of a new social program. Implementation analysis has been defined by Williams and Elmore as follows: Scrutiny of (l) the preliminary policy specifica- tions to determine their clarity, precision and reasonableness; and (2) staff, organizational, and managerial capabilities and implementation strate- gies to determine the degree to which the proposed policy alternative can be specified and implemented in its bureaucratic/political setting. The primary rationale for addressing issues of imple- mentation is that without information about how programs go about their daily business, it is impossible to repli- cate programs found to be successful, and it is next to impossible to determine the barriers responsible for the unsuccessful programs. The implementation process has begun to attract a great deal of attention partially because of the increasing evidence that social programs often fail to achieve the expected goals. One reason for the failure of many social programs is of a political nature and has to do with the fact that programs usually must be oversold to attract funding which results in the speci- fication of lofty and unattainable goals. A second pos- sible reason for the failure of a program to achieve its goals is that the theory upon which the program was based may be inadequate. Finally, a program may be 1Walter Williams and Richard F. Elmore, Social ProgramImplementation (New York: Academic Press, 1976), p. 270. 112 unsuccessful because it was not implemented to the extent necessary for processes to be activated which were to bring about change. All of these factors are inter- related and it is hoped that the implementation analysis will provide clues to the range of factors which facili- tate and impede the successful introduction of a new program. The lack of attention given implementation in program evaluations can be traced to the lack of under- standing and guidelines in this area. As Williams and Elmore suggested, implementation analysis involves the study of dynamics, which is the weakest area method- ologically in the social sciences.1 They went on to point out that the questions addressed by implementation analy- sis fall into the "uncharted research terrain": Implementation analysis must ask whether the organi- zation can do what is desired in technical terms, whether it can function well in a bureaucratic sense (which involves micro—organizational issues), and whether it can Operate successfully in its larger environment (macro-organizational/political issues). Thus, while the importance of the implementation pro- cess is now being recognized, designs and data-collection procedures are not nearly as specified nor systematized as they are in the analysis of outcomes or effects. One of the few studies which has focused spe— (lifically on the implementation process is that of ‘ 1Ibid., p. 271. 2Ibid. 113 Pressman and Wildavsky.1 They examined the implementa- tion of a federal Economic Development Administration (E.D.A.) employment program in Oakland, California. Their conclusion was that the multi-million-dollar pro- gram failed to achieve its original goals because of: (l) the federal grandeur in formulating idealistic pro- grams, goals, and principles; (2) the overestimation by federal policy-makers of the level of local support which can and will be generated; and (3) the separation of policy from implementation with the latter being viewed as "technical questions that would resolve themselves if the initial agreements were negotiated and commitments were made."2 The primary contribution of the Pressman and Wildavsky study was its translation of conceptual implementation concerns into research and evaluation activities. It did not, however, point out specific methodological procedures for addressing questions about implementation. The suggested course of action was stated in more general terms of making implementation problems part of the initial policy-formulation process, and bringing into closer correspondence means and ends by making each partially dependent on the other.3 1Jeffrey L. Pressman and Aaron Wildavsky, Imple— mentation (Berkeley, Calif.: University of California Press, 1973). 2Ibid., pp. 142—43. 31bid., p. 143. 114 Another study noteworthy because of its empha- sis on the implementation process is the analysis of planned educational change by Gross, Giacquinta, and Bernstein.1 The primary impetus for the study was skepticism by the authors that traditional theories of planned change were too narrow in focusing on individual organizational members' initial resistance to change. It was felt that the results of the study demonstrated the need for a new theoretical formulation of planned change which would "take into account organizational variables that could influence the implementation phase of the process . . . a theory that would take into account the complex, dynamic nature of the process involved in successfully implementing organization inno— vations."2 Reference is also made to an article by Greiner,3 in which the importance of historical and developmental factors is emphasized with regard to the study of implementation. Special emphasis was placed on the developing relationship between an organization and its environment. 1Neal Gross et al., Implementing_Organizational Innovations (New York: Basic Books, 1972). 2 Ibid., p. 9. 3Larry E. Greiner, "Antecedents of Planned Organi- zational Change," Journal of Applied Behavioral Sciences 3 (1967): 51-85. 115 The writings which have just been discussed are at the forefront in the area of implementation analysis. But it is obvious that only the most basic conceptual and methodological concerns have been addressed. Research design and evaluation procedures to be used in the two major components of the implementation-analysis section of this study represent an initial attempt to look more closely at organizational and environmental factors that influence the introduction of a social program. The variables to be included in this section represent those that are hypothesized to be critical in the implementation of a Youth Service Bureau. The organi- zational factors to be examined revolve around staff per- ceptions and orientations. Studies reviewed in the last chapter suggested this area was an important determi- nant of the basic nature of a project, and it will be hypothesized in this study that certain perceptual frameworks would be more highly characteristic of suc- cesful programs. Environmental assessment will be aimed at discovering how various projects dealt with a number of social and political issues that appear to be common in all social innovations. It is hoped that the organi- zational and environmental components of the study will generate guidelines for improving the implementation and evaluation of Youth Service Bureau programs. 116 Organizational‘Factors In this component, S3 of the 13 sites will be included (see Table l). Kalamazoo County was not included because the project was no longer in operation when the data were collected, and Calhoun County was not among the original group of funded programs during the initiation of the study (time—series data were col- lected on the project because it was informally regarded as the prototype Youth Service Bureau in the state). Two instruments were used for data collection--the Delinquency Orientation Scale and the Program Percep— tions Survey. This information was requested from all project staff members who were involved in the adminis- tration and/or service delivery aspects of the program (i.e., directors, supervisors, caseworkers and youth workers, casework aides, and student interns). In sites where environmental assessment interviews were performed (the next component), these instruments were delivered to project directors for circulation among staff members. Copies of the instruments were mailed to the other sites with an accompanying letter of explanation. The return rate for the Delinquency Orientation Scale was very high, with 51 of the 59 distributed completed and used in the analysis. On the Delinquency Orientation Scale (Appen- dix D), four major conceptual frameworks for viewing 117 delinquent problems are included. These orientations are based on a classification of reactions to delin- quency by Schur, and are briefly the following: 1. the get-tough antipermissive approach-~an insistence that wrongdoers must be dealt with sternly and that misconduct ”will not be tol- erated,” the ”good guys vs. bad guys" theme; 2. the individual treatment approach—-emphasizes the distinctive characteristics of individual offenders and the modification of individual attitudes and behaviors; 3. the liberal reform approach--emphasizes the socio-cultural aspects of deviance and the improvement of community programs and insti— tutions; 4. the nonintervention approach-~recognizes the widespread and temporary nature of most "mis- conduct” and seeks to delimit the application of formal sanctions. Schur pointed out that individuals will rarely exhibit a pure form of one of these orientations, but that they are models around which persons organize their responses because "each pattern is grounded in certain core assumptions and basic outlooks that in turn imply a whole complex of interrelated preferences." The Delinquency Orientation Scale was developed by creating statements felt to represent the position suggested by each approach on five issues--causes of crime and delinquency, most appr0priate responses, role of the juvenile court, approach to prevention, and the use of diversion. Two statements were formulated for lSchur, Radical Non—Intervention, pp. 19-23. 2Ibid., p. 22. 118 each of the four approaches on the issues of causation (antipermissive-—numbers 24 and 10, treatment-~31 and 6, reform-—l4 and 8, and nonintervention--9 and 23); response (16 and 30, 25 and 11, 29 and 18, and 4 and 28, respectively); and the role of juvenile court (5 and 32, l and 17, 2 and 22, and 19 and 7). One statement for each approach was included on the issues of prevention (20, 15, 12, and 27, respectively) and diversion (3, 21, 26, and 13). A rating was obtained for each of the 32 statements using a Likert-type rating system from one (strongly agree) to six (strongly disagree). The first step in the analysis of these data will be to examine the internal consistency of the instru- ment. Using subprogram Reliability of the Statistical Package for Social Sciences (SPSS) computer program, Cronbach's alpha will be computed for each of the four subscales to determine the extent to which variance in subscale scores is accounted for by common variance with the subscale items.1 Then the correlation of each item (statement) with the total subscale score will be exam- ined to see if the item is most appropriately placed in the subscale. Finally, the intercorrelations of the four subscales will be analyzed to test the discriminant 1Norman H. Nie et al., Statistical Package for the Social Sciences, 2nd ed. (New York: McGraw-Hill, 1975). 119 validity of the instrument (or its success in tapping into distinct response patterns). Following instrument validation procedures, the data will be used to examine the dominant orientations of staffs at each project. The number of respondents in each site, of course, varied and these figures will be presented along with mean project ratings on each of the four subscales. Mean project ratings will be calculated by adding the ratings of each staff member on all eight statements in each subscale and dividing by the total number of items rated in the subscale. Given the rating system used (1 = strongly agree and 6 = strongly disagree), the lower the mean rating of a project on a subscale the more that subscale is characteristic of project staff orientations. While there are no standardized norms to relate the mean project ratings to, they can be viewed relative to the other Youth Service Bureau projects to determine if there are significant differences among the projects with regard to dominant delinquency orientation(s). Where this kind of variation is observed, the impact measures of effec- tiveness will be reviewed to see if they appear to be related to any particular response patterns on the Delin- quency Orientation Scale. The second part of the organizational component will focus on staff perceptions of several important internal operational variables. The same 9 projects 120 were included and data were collected in the same manner as with the Delinquency Scale. Of the 64 Program Per- ceptions Surveys that were distributed, 57 were returned and are included in the analysis. The Project Perceptions Survey (Appendix E) is a modified version of an instrument developed by Moos to assess the organizational environment of cor- rectional programs?‘ It contains the following nine sub- scales: Relationship dimensions 1. Involvement--measures the degree of partici- pation by clients in the ongoing operations of the project; 2. Support--measures the level of support given clients by project staff; 3. Expressiveness--measures the extent to which open expression of feeling is encouraged; Treatment dimensions 4. Autonomy-—measures the extent to which clients are encouraged to take part in plan— ning and leadership activities; 5. Practical Orientation--measures the degree to which clients are prepared for leaving the program; 6. Personal Problem Orientation—-measures the amount of concentration on understanding personal problems and feelings; Systems maintenance dimensions 7. Order and Organization-—measures how impor— tant order and organization are in the pro- gram; 8. C1arity-—measures the explicitness of program rules and procedures; and 9. Staff control--measures the extent to which regulations are used to control clients.2 1Rudolf H. Moos, Evaluatingygorrectional and Communitnyettings (New York: John Wiley, 1975). 2Ibid., p. 41. 121 The basic purpose for which this instrument was used was to obtain a comprehensive outline of the operational nature of the projects. The format of the Program Perceptions Survey was true-false, and the scoring of the items is specified by Moos.1 Basically, item responses which indicated positive perceptions (true for statements characteris- tic of the subscale and false for those not) were scored as one, and those which indicated negative per- ceptions as zero. Thus, the higher the mean project score on a subscale, the more characteristic that vari- able is of the project (mean subscale scores could range from .00 if each respondent answered each subscale item in the negative direction to 1.0 if each respondent answered each subscale item in the positive direction). The same validation procedures will be performed with this instrument as were described for the Delinquency , Scale. And similarly, mean project scores on each of the subscales of the Program Perceptions Survey will be examined and related back to impact measures (where they are available) to determine if any of the subscales vary concomitantly with impact results. 1Ibid., Appendix A. 122 Environmental Assessment The final component of the study involves an attempt to examine a range of social, historical, and political variables in the environmental contexts of each project. Seven projects were included in this component (see Table 1). These were chosen because the time-consuming nature of data collection precluded the inclusion of all sites and these seven projects were felt to be representative of the entire sample-~they represent large and small projects, old and new, urban and rural, accepted and rejected, and variations on the Youth Service Bureau concept. The primary rationale for the environmental assessment is the need to understand the influence of extra-organizational factors on the development and success of projects. Using the Environmental Assessment Guide (Appen— dix F), data were gathered in structured, open-ended interviews with a range of persons at each site. Those interviewed included project directors and staff, as well as a number of individuals external to the pro- jects who had had contacts with the programs in an adVisory and/or utilization capacity and were familiar with the evolution of the program. The external persons represented law enforcement agencies and the courts, schools, social service agencies, planning units, and related programs. Thirty—four persons were interviewed 123 and over 75 hours of tapes were reviewed in preparation for writing the descriptive analyses of environmental factors. Interviews and discussions of this material were organized around certain conceptual areas of interest. First, there was a focus on the involvement of and sup— port by individuals and organizations in the community. Second, questions were presented regarding the formal and informal positions of projects in the community social structure, and the external relationships necessary to implement the program. Third, energies were directed toward examining the perceptions of project staffs and external others with regard to the degree of clarity and consensus on program goals and objectives. Finally, attempts were made to clarify the historical factors which may have facilitated, hindered, and/or modified the directions and operations of projects. In reviewing the taped interviews with these topical areas in mind, certain environmental issues were found to be common among several projects while others were unique to the situation of a particular project. The results of the environmental assessment will consist of a descriptive summary of each of these environmental issue areas, as well as a discussion of the techniques and effectiveness of approaches to these issues by project staffs. It is hoped that this final component 124 of the study will provide clues to understanding per- ceived differences in success on the outcome variables. Limitations of the Study The major limitation of the study is the fact that it was carried out in an ex post facto manner. All program evaluations which are designed and executed following the initiation of a program are weak insofar as the ability to demonstrate causal relationships is concerned. While social science research is not ever likely to produce conclusive causal statements where human behavior is involved, it can be expected to gen— erate empirically supported statements of relationships useful in the formulation of social policy. The criti- cal factor, however, remains the failure to utilize principles of social research in a manner which permits the generation of meaningful results. At a minimum, program evaluation needs to be included in original program formulations, and resources need to be allocated to insure the integrity of research designs is maintained. Another shortcoming of the study has to do with the all-encompassing nature of the general systems model and the inherent problem of not being able to examine all of1fluepossible interrelationships suggested by a systems approach. The selection of official crime data as the primary outcome measure, for example, was made 125 at the expense of having to ignore other potentially important systems-level measures (e.g., school and social service data). Also, the inclusion of organiza- tional and environmental factors was limited to those hypothesized to be most influential in the evolution of a program. Other organizational factors (e.g., director leadership qualities, the recruiting and training of personnel, etc.) and environmental issues (e.g., economic conditions, civic concern) would no doubt warrant close scrutiny in a follow-up to this study. This basic prob- lem in systems research is most adequately addressed by using sound research designs that generate meaningful results, and compiling evaluation results in a systematic fashion so that variations in evaluation models are exploited for broadening the information base upon which policy decisions are made. Finally, the time factor is nearly always a con- sideration in social research. Many social programs would not be expected to have significant impact for years given the extent and complexity of social problems like poverty andcufinmaand delinquency. It is unfortu- nate that most social research misses the earliest phases of program development where the general frame- works for program structures and functions are con- ceived. 126 In this study, evaluation activities were car- ried out during a relatively short time period in the overall life spans of1flueprojects. Efforts were made to gather historical and developmental information, and there were attempts to gauge the probabilities for con- tinued operations with local funds. But in order to thoroughly understand the intricacies of implementation and the ultimate impact of programs on a social problem, it is mandatory that comprehensive evaluation designs be conceived and executed in numbers of similar pro- grams throughout the existence of each project. While these limitations are troublesome in most social research today, they do not preclude the genera- tion of information useful for planning purposes. This is particularly true when addressing an area like imple- mentation, about which there is little in the way of published knowledge. At a minimum, it is hoped that the study will demonstrate one possible approach to obtaining comprehensive and useful information about the relation- ship between implementation and overall program effec- tiveness. CHAPTER IV RESULTS Time-Series Analyses The examination of Youth Service Bureau impact focused on crime reduction as measured by actual offense data, and diversion as measured by delinquency arrest data. To recapitulate, diversion has been viewed as an intermediate goal of the projects related to affecting the processes of police decision-making and disposi- tional patterns. Crime reduction is a less direct, long-range goal which has been suggested as a measure of the ultimate effects of projects reflecting their successes in preventing delinquent behavior. Both com- ponents were addressed in this study using annual multiple-group and monthly single-group time-series designs and official crime statistics. The form taken by the results of statistical time-series analyses of monthly data can be seen in the example in Table 2. Initial information is presented for the number of pre— and post-intervention points (N1 and N2), as well as model identification data including the orders of autoregression, differencing, and moving 127 omo.u mmo.- pawn 9 ago pmflpa mmm. mmm. pms.:u :00.¢mn vos.o: smm.mv: «0:.mHs: mm. uppm e u:::: uppm e 0:0 Hm>mq papm a Hm>mq Ha> an: apps: ”000:9 :0\0:0 Hz: m0 003H0> 00:0H:0> :0H00 8:8“:HE How 0:: 0wa:00u m:HBOHHom 0:9 mwa 50H mbH mNH wma mmH whH 50H 00H me HbH wma mmH OHM HNH mNH NbH me 00H NAN 0wH HvN mmH me mbH me mud wwfl woa HNH mma wad “was: :0wp:0>:00:wlpmom 128 woa mvH mvH mNH 00H omH mmH va MHH mm 0mH wvm Hm mvfi NmH 05H mmH HmH va b0H mmH “0:00 :owp:0>:0u:H|0:m H 0: 0H0>0 H0:00:00 Ho :pw:0q 0H 0wwp0>0 w:H>oa 00 :00:0 0: w:wo:0:0wmfiv m0 :00:0 0: :0H000Hw0:ou:0 00 :00:0 Or-IHN ”noflppofimfipaopfl proz 0: u :wfl00p man: you 80000nm :0 000:000 .mm H Nz .HN n Hz 0HHH>:HH80 :w 0000::: >0:0:U:HH0Q H0909 H 00:8:2 80Hnoum .0H080x0 0H0>H0:0 00Hh00|0awenl.m 0Hnafi 129 average, and the length of the seasonal cycle. Then the raw data for the pre- and post-intervention periods are listed (January, 1972, is the first data point). The bottom line of the table contains the statistics which will be examined. Data on level, level change, drift, drift change, and the corresponding t-statistics are presented for the theta value where there is a mini- mum error variance around the regression 1ine. The basic function of the time-series analysis is to predict a post-intervention trend line based on adjusted pre- intervention data (adjusted for seasonal fluctuations and the moving average process) and compare the observed level to zero and the observed slope to the predicted. Looking at the time-series analysis example in Table 2, it can be seen that the actual number of delinquency arrests in January, 1972, was 139 while the adjusted time-series estimate for initial level is 149.6. The t—statistic associated with the figure for- initial level is for a t-test to determine whether the estimate of initial level is significantly different from zero. In general, the t-values resulting from this analysis of crime data would be expected to be large and statistically significant since almost every jurisdiction has a crime rate somewhat higher than zero. Therefore, the initial level of a variable and its 130 associated t-value provide little in the way of useful information and will not receive detailed consideration. The figure for change in level represents the estimated change in the level of the variable (number of delinquency arrests) at the time specified as the point of project intervention. Because of the adjust- ments that have been made to account for systematic instability in the data series, this change may be interpreted as the effect due to the intervention pro- ject. As indicated above, we have utilized a point three months after the formal funding (July, 1973, for this example) as the intervention point to test for their effects on the diversion of delinquents. The actual figure presented for change in level is calcu- lated on the difference between estimates of the last pre—intervention data observation and the first post— intervention data observation (post-intervention minus pre-intervention). For example, Table 2 shows a -24.7 for the change in level of the data on total delinquency arrests. This represents a decrease of 24.7 arrests between the pre— and post-intervention periods. This figure is based on the difference between the estimated number of arrests for September and October of 1973, the intervention point. According to the raw data presented in Table 2, the figures for these two months were 168 and 138, 131 respectively, which means that without the time-series adjustments the change in level would have been -30. Using a one-tailed test, the t-statistic for change in level (-l.7) is significant at the .05 level (with 49 degrees of freedom), indicating that there was a sig- nificant decrease in total delinquency arrests follow- ing the intervention. It should be reiterated that change in level is calculated using only the last pre- intervention and first post—intervention adjusted data points. Therefore, moving the intervention point back or up one month could have a significant effect on the level change statistics. Given the general upward trends of crime sta- tistics, it is perhaps the data for slope in trend lines that provide the most sensitive measures of project impact. The measure of slope refers to the general slope of the pre-intervention time-series line while change in slope is the difference between the predicted slope of the post-intervention series and the actual slope of the trend line following the inter- vention. Looking back at Table 2, it can be seen that in the example there was a slight positive slope before the intervention (.39) and a very slight decrease in the slope of the trend line following the intervention (-.04). These data indicate that the positive rate of increasing delinquency arrests before the intervention 132 was slightly retarded in the post-intervention period. Neither of the t-statistics--.34 for slope and -.03 for change in slope—-approaches the .05 level of signifi- cance. Thus, the conclusion for this time-series analysis example with regard to slope would be that slightly increasing rates of delinquency arrests prior to the initiation of the project were not significantly affected by the activities of the project. The statistical values at the bottom of the table will be summarized in table form for each of the sites by variable. For the three actual offense vari- ables in the crime reduction component, values for the change in level and slope and their associated t-statistics will be presented for the seven project jurisdictions. Changes in slope provide the most sen- sitive measures for assessing project impact at the systems level because of the problem regarding the effects of choosing an intervention point on the measure of level change. The same form of data presentation will be followed for the nine arrest and court petition variables in the diversion component. In both compo— nents, summaries of the monthly time-series analyses will be discussed in conjunction with the annual crime rate data for the seven Youth Service Bureau juris- dictions and their comparison sites. Following the presentation and discussion of the time-series analyses 133 of project impacts, results from the organizational and environmental components of the implementation section will be presented. Crime Reduction As has been stated, crime reduction is con- sidered the ultimate outcome measure at the systems level and actual offense data from official police sta- tistics have been used in addressing this research question. The actual offense data are the most accu— rate measure of crime and delinquency levels since they are the subset of all reported offenses that are inves- tigated and determined to have in actuality been com— mitted (in other words, founded complaints). While many minor criminal and delinquent offenses are not reported and many reported offenses are not founded, these data are the most consistently and comprehensively recorded information regarding crime levels across all jurisdictions. Annual data for project and comparison jurisdictions have been transformed into rates per 1000 total county population (1970 Census) and are presented for the six-year period from 1971 through 1976. A heavy black line is used in these tables to indicate the last pre—intervention year and the first post- intervention as determined by the formula spelled out in the last chapter (i.e., six months in operation 134 during a calendar year equals first post-intervention year). And as discussed above, results from the monthly time-series analyses will be presented to supplement the annual crime rate data. The first variable in the actual offense cate- gory was burglary. Rates of actual burglary offenses for the six-year period at the project and comparison sites are presented in Table 3, with the dark lines indicating the appropriate intervention points. Only Calhoun County showed a decrease in actual burglary rate from the last pre-intervention year to the first post year while the comparison county did not, and the general trend for the intervention year (1972) was in an increasing direction. Newaygo and St. Joseph Coun— ties experienced post-intervention decreases, but these occurred in a year (1976) when all sites including the comparison counties also showed decreases. Tables 4 and 5 contain the same type of infor— mation for rates of actual larceny and actual vandalism, respectively. The same overall trends can be seen for both offenses, with a consistent increase through 1974 and decreases beginning to show up occasionally in 1975 and consistently in 1976. Calhoun and Newaygo Counties show a post-intervention decrease in rates of actual larceny offenses although the comparison county for Newaygo also had a decrease. With regard to rates of 135 Table 3.--Actual burglary offenses (rates per 1000 total county population, 1970). Year counties 1971 1972 1973 1974 1975 1976 Calhoun 14.92 14.18 15.70 20.14 16.17 15.15 Jackson 13.12 14.55 16.19 19.50 19.61 17.03 Berrien 14.02 17.36 17.84 22.02 18.73 17.86 Muskegon 16.36 12.23 13.89 19.77 20.77 16.03 Genesee 13.75 14.66 17.97 22.89 21.93 19.68 Saginaw 22.78 19.69 19.61 25.90 23.16 18.96 Van Buren 12.35 11.25 16.38 20.42 20.76 18.18 Allegan 8.90 9.76 9.39 15.13 13.35 10.81 St. Clair 12.57 13.80 16.08 19.07 19.72 15.96 Lapeer 5.63 7.03 8.35 11.76 8.52 7.86 Newaygo 10.57 10.97 10.82 14.18 15.33 10.75 Mecosta 6.68 9.86 8.68 12.04 11.11 4.70 St. Joseph 10.70 10.99 11.52 12.79 11.01 9.92 Branch 7.86 11.85 11.42 11.16 9.08 7.23 136 Table 4.--Actua1 larceny offenses (rates per 1000 total county population, 1970). Year countles 1971 1972 1973 1974 1975 1976 Calhoun 29.82 28.49 31.58 38.65 36.31 39.34 Jackson 24.80 27.82 26.85 29.47 29.26 32.24 Berrien 32.60 33.29 37.03 41.21 45.89 43.17 Muskegon 31.79 29.59 32.63 36.22 39.64 40.60 Genesee 31.55 29.03 33.65 43.31 48.08 46.07 Saginaw 37.17 38.32 38.34 48.51 44.21 43.88 Van Buren 19.69 17.43 21.75 27.70 34.79 32.31 Allegan 13.31 12.05 15.40 20.91 21.33 17.65 St. Clair 23.99 24.65 29.64 35.66 41.81 36.05 Lapeer 11.06 11.04 13.20 18.26 17.21 16.84 Newaygo 12.97 13.29 14.54 15.00 20.29 14.61 Mecosta 33.97 33.87 34.58 43.01 48.23 41.48 St. Joseph 24.41 20.78 25.09 31.80 27.79 29.29 Branch 21.16 25.32 28.41 26.51 20.31 19.71 137 Table 5.-—Actua1 vandalism offenses (rates per 1000 total county population, 1970). Year counties 1971 1972 1973 1974 1975 1976 Calhoun 5.88 6.48 6.63 9.33 11.25 10.12 Jackson 8.72 9.66 7.04 9.96 11.31 11.38 Berrien 15.27 15.91 18.76 21.74 22.72 21.80 Muskegon 13.08 12.67 13.96 18.24 22.11 19.23 Genesee 8.23 8.60 11.62 15.44 16.10 18.16 Saginaw 2.80 3.48 3.50 5.65 9.22 8.11 Van Buren 5.68 5.36 7.28 9.27 10.56 13.32 Allegan 3.91 5.39 5.80 9.13 9.37 8.28 St. Clair 9.81 11.90 13.36 17.95 20.54 17.32 Lapeer 1.95 2.52 3.63 5.75 5.84 4.70 Newaygo 3.18 5.57 4.79 4.18 2.93 2.79 Mecosta 6.47 7.86 9.65 13.90 14.93 11.22 St. Joseph 13.59 13.53 14.26 19.71 18.34 18.65 Branch 8.02 8.47 5.65 6.28 4.83 5.49 138 actual vandalism, Newaygo County was the only project jurisdiction to experience a post-intervention decrease but its comparison county again showed a similar decrease. Thus, while Newaygo County had post-intervention decreases for the three crime reduction variables, it should be noted that the overall trend was decreasing that year (1976) and the comparison county had similar decreases for each variable. Calhoun County, however, had post-intervention decreases for two of the three variables (burglary and larceny) in a year when the overall rates were increasing (1972) and its comparison county was also increasing. The apparent success in crime reduction for Calhoun County may have to do with the fact that it has the oldest Youth Service Bureau project and perhaps, several years are required in order to impact on the overall levels of actual crime vari— ables. It should also be mentioned that while Calhoun County had a decrease in actual burglary and larceny offenses in the first post-intervention year, this was not the case in the following years where the county experienced changes (increases and decreases) similar to the overall trends. The results of the monthly time-series analyses of the three crime reduction variables are summarized in Table 6. It should be reiterated that the statis- tical time—series analysis of monthly data involves Table 6.--Summary statistics for monthly time-series analyses of crime reduction--actual burglary, 139 actual larceny, and actual vandalism. Change Change Pre P051: df in level t in Drift t Actual Burglary Battle Greek 13 42 52 3.161 .157 5.468 1.041 Benton Harbor 24 31 51 25.940 2.779 -2.286 -2.871* Flint 24 29 49 92.237 3. 436 -5.594 —2.405* Paw Paw 36 20 52 .395 .249 .243 .691 Port Huron 42 13 51 4. 807 .653 -1. 883 -2.281* White Cloud 48 8 52 - .935 -1.156 .106 .701 Three Rivers 48 8 52 1.072 .417 .273 .570 Actual Larceny Battle creek 13_ 42 52 -2. 666 .822 .194 .185 Benton Harbor 24 31 51.678 - .435 .085 .636 Flint 24 29 49 8.071 1.496 L 407 3. 010 Paw Paw 36 20 52 .385 .831 - .030 - .283 Port Huron 42 13 51 - .905 - .523 .151.730 White Cloud 48 8 52 .267 - .781 .080 1. 258 Three Rivers 48 8 52 .003 .004 - .041 - .363 Actual vandalrsn Battle Creek 13 42 52 ...a Benton Harbor 24 31 52 - 6.087 - .872 - .900 -1.511 Flint 24 29 49 17.374 .290 2.298 .129 Paw'PaW' 36 20 52 - 3.623 -2.512* .110 .841 Port Huron 42 13 51 1.598 .142 -2.090 -1.548 White Cloud 48 8 52 .226 - .338 - .069 - .552 Three Rivers 48 8 52 -l9.688 -3.510* 3.571 3.417 4: 'Significant at the -1.67) with 50 degrees of freedom using a one-tailed test. .05 level (t must be less than aData in these sites could not be analyzed for this variable because of the large number of months in which no offenses were founded or reported. 140 making data adjustments to remove the effects of sea- sonal fluctuations and overall time-series trends due to past shocks to the system (i.e., unemployment, urbanization, etc.). The change in level represents the difference between the last adjusted pre-intervention data point and the first adjusted post-intervention point . Change in drift measures the difference between the slope of the predicted post-intervention line and the slope of the observed post-intervention line. Using a one-tailed t-test (since the hypothesized direction of change is negative) and the usual .05 confidence level, a t-Value must be less than -l.67 in order to be considered significant. All such t-values are noted in Table 6. It can be seen in Table 6 that all Youth Service Bureau jurisdictions except one experienced increases in the level of actual burglary offenses following the interventions. There were, however, three sites (Benton Harbor, Flint, and Port Huron) that showed significant decreases in slope following the initiation of pro- jects. In these three sites, the level of actual bur- glary did not decrease although the significant changes in slope indicate that increasing rates of actual bur- glary offenses were significantly retarded in each locale. The results of time-series analyses on actual 141 larceny and actual vandalism offenses are also con- tained in Table 6. None of the seven sites was found to have experienced significant decreases in either the level or slope of the larceny time-series. With regard to actual vandalism, the analyses indicated that Paw Paw and Three Rivers had significant decreases in level after the Youth Service Bureaus began. These sites did not experience similar decreases in drift, which would have increased our confidence in the positive nature of these findings. Benton Harbor and Port Huron experi- enced reductions in post-intervention slope that approached significance at the .05 level. The hypothesis that Youth Service Bureau juris- dictions would experience decreased crime rates follow- ing intervention was not supported by the results of the time-series analyses. From the annual data, it appeared that Calhoun and Newaygo Counties may have had post-intervention decreases, but these were not veri- fied in the statistical analyses of monthly data. Some significant changes were observed for level and drift of the three actual offense variables, although they were not consistent across either site or variable. For the three actual crime variables, there were only two instances where a significant decrease in level was found (vandalism in Paw Paw and Three Rivers) and three 142 where a significant decrease in slope was observed (burglary in Benton Harbor, Flint, and Port Huron). In order to constitute support for the hypothesized reduction in crime, time-series results would need to show some degree of consistency across site and/or variable. System Impact of Diversion The time-series analyses of arrest data were aimed at generating information regarding the effec- tiveness of Youth Service Bureaus in accomplishing the intermediate systems-level goal of diversion. Because diversion is considered an intermediate step in the attainment of crime reduction goals, the arrest vari- ables in this component are those that should be most sensitive to the initial impacts of projects on the justice system. Arrest data and the court petition data included in the latter part of this component are the most accurate indicators of statewide juvenile jus- tice system activities. Insofar as the Youth Service Bureau projects are geared to affecting change in the processes of the juvenile justice system by advocating diversion, this component provides the most direct test of project impacts at the systems level. Annual delinquency arrest rates for the Youth Service Bureau and comparison jurisdictions are presented 143 in Table 7. These rates per 1000 are based on the number of juveniles ages 7 through 16 in the county (as opposed to total population), and have been adjusted on a year—to-year basis to take into account the decreasing size of the at-risk juvenile population over the six years. The presence of the same overall trend characteristic of the actual offense data can be discerned in Table 7 for this first arrest variable. Total delinquency arrest rates were generally on the rise until 1975, when they began to decrease and con- tinued this trend through 1976. The only Youth Service Bureau jurisdictions which experienced post-intervention decreases for total delinquency arrest rates were Newaygo and St. Joseph Counties, but the two comparison counties showed similar decreases and the general arrest rate for delinquency was declining in the post- intervention year for both sites (1976). Thus, it does not appear that the projects were successful in having a perceptible impact on the total delinquency problem in the communities as measured by rates of arrest. It could be that projects have been successful in encour- aging diversion although the police continue to make the formal arrest prior to diverting a juvenile. In this case, the system impact of diversion would be expected to show up in the rates of juvenile court petitions, which will be examined momentarily. It may 144 Table 7.-—Total delinquency arrests (rates per 1000 juve- niles, adjusted). Year 0°“ntles 1971 1972 1973 1974 1975 1976 Calhoun 27.30 50.83 53.53 53.42 51.54 52.24 Jackson 30.37 33.50 34.87 36.73 41.53 36.49 Berrien 49.90 52.73 57.52 65.67 60.39 53.99 Muskegon 48.15 52.51 54.06 62.47 61.15 56.12 Genesee 28.46 30.17 34.54 39.62 35.06 30.17 Saginaw 54.25 70.66 70.66 64.09 36.91 45.42 Van Buren 37.14 44.19 42.64 52.20 39.03 52.16 Allegan 24.96 27.81 29.64 29.37 23.62 27.35 St. Clair 60.98 58.54 65.39 40.12 75.33 63.67 Lapeer 10.94 14.77 23.66 26.93 23.24 21.66 Newaygo 27.79 29.65 28.56 38.80 31.30 21.45 Mecosta 8.80 7.88 8.13 23.11 39.30 35.97 St. Joseph 41.81 40.14 49.50 59.54 64.29 52.43 Branch 13.93 18.84 15.83 13.00 11.67 11.37 145 also be that the impact of diversion activities is limited to specific areas of concentration, which is why the arrest rates for certain offenses have been included in the analysis. Rates of delinquency arrests for burglary over the six-year period are presented in Table 8. The same general trend is present in the data for the five specific offenses (i.e., increases through 1974 and decreases in 1975 and 1976) and post-intervention changes in Youth Service Bureau jurisdiction must be interpreted with this in mind. It can be seen from Table 8 that four of the Youth Service Bureau counties-- Berrien, St. Clair, Newaygo, and St. Joseph-~experienced post-intervention decreases in the rates of burglary arrests. In each of these instances, the decreases were found in the absence of similar decreases in the comparison counties although the post-intervention years for the last three counties were ones in which the overall rate of burglary was on the decline (1975, 1976). Table 9 contains the delinquency arrest rates for larceny. St. Clair, Newaygo, and St. Joseph Coun- ties registered decreases in rates of larceny for the post-intervention years, but in each case the comparison counties showed decreases in larceny rates of a similar magnitude. The data in Table 10 for vandalism arrest rates indicate that Berrien and St. Joseph Counties 146 Table 8.--Delinquency arrests for burglary (rates per 1000 juveniles, adjusted). Year C°unties 1971 1972 1973 1974 1975 1976 Calhoun 2.46 4.44 4.90 6.08 3.75 5.23 Jackson 3.94 4.62 3.64 4.78 5.57 4.92 Berrien 6.17 7.85 6.28 8.40 6.27 4.73 Muskegon 7.65 6.49 7.94 7.95 7.63 4.87 Genesee 3.62 3.60 4.11 5.29 4.61 4.44 Saginaw 3.86 4.68 6.07 5.13 2.21 3.47 Van Buren 7.24 8.37 5.80 8.78 8.06 7.87 Allegan 2.92 5.72 3.46 4.18 2.19 3.87 St. Clair 7.12 4.11 7.08 7.64 7.12 7.40 Lapeer 1.26 2.88 .97 2.28 2.55 1.65 Newaygo 4.59 3.26 4.37 5.84 8.48 3.34 Mecosta 80 2.42 1.24 3.36 1.68 3.54 St. Joseph 2.60 1.78 4.86 6.49 5.84 5.79 Branch 2.54 2.03 1.85 1.91 .86 1.14 147 Table 9.--De1inquency arrests for larceny (rates per 1000 juveniles, adjusted). Year countles 1971 1972 1973 1974 1975 1976 Calhoun 5.80 14.44 14.61 16.87 17.00 13.62 Jackson 6.74 6.83 6.78 6.28 9.63 7.34 Berrien 11.61 12.51 14.75 18.90 20.65 16.84 Muskegon 10.65 11.84 13.26 12.29 12.04 12.23 Genesee 7.22 6.70 8.11 9.87 9.65 7.12 Saginaw 7.11 15.43 15.40 15.86 11.67 10.66 Van Buren 3.54 8.60 6.94 9.71 12.61 9.33 Allegan 8.58 5.84 6.32 8.74 7.08 6.49 St. Clair 11.31 9.70 9.85 16.35 14.61 10.82 Lapeer 2.73 3.25 7.69 7.89 6.42 7.56 Newaygo 6.11 8.79 5.83 10.19 6.17 5.56 Mecosta 2.60 1.01 1.03 13.44' 19.28 16.22 St. Joseph 12.52 8.98 9.44 14.91 17.21 11.88 Branch 3.98 6.99 5.31 4.17 5.16 4.30 148 TablelfiL-De1inquencyarrests for vandalism (rates per 1000 juveniles, adjusted). Year C°untles 1971 1972 1973 1974 1975 1976 Calhoun 1.56 3.08 2.48 2.73 3.51 2.85 Jackson 2.80 3.67 2.62 2.80 3.41 1.57 Berrien 3.12 3.13 2.39 5.06 4.58 3.30 Muskegon 3.62 4.84 5.27 5.61 5.04 3.83 Genesee .93 1.00 1.16 1.10 1.12 1.50 Saginaw 1.46 1.13 1.73 1.72 1.13 1.60 Van Buren 3.40 3.29 2.90 2.95 2.55 3.65 Allegan 1.25 2.62 2.85 2.98 2.96 .52 St. Clair 7.01 6.87 7.36 6.99 8.76 5.76 Lapeer 1.18 74 1.34 2.20 1.24 55 Newaygo 2.50 2.13 .87 2.70 2.93 1.43 Mecosta 6O .40 .83 .21 3.98 1.04 St. Joseph 5.75 3.84 4.58 6.78 7.71 5.39 Branch .99 1.02 1.39 .12 .00 .25 149 have post-intervention decreases in vandalism rates, while both respective comparison counties showed increasing rates of vandalism in the same year. Also, Berrien County registered the decrease in a year when the overall vandalism arrest rate was on the rise (1973). Newaygo County again registered a decrease in vandalism rates following the introduction of the project. To summarize, the annual data presented on delinquency arrest rates (total, burglary, larceny, and vandalism) provide little support for the hypothe- sized post-intervention decreases in arrest rates for Youth Service Bureau jurisdictions--evidence that the projects made an impact on the justice system by encour- aging the use of diversionary alternatives. For example, Newaygo County showed a post-intervention decrease in rates for all four variables. It should be noted, however, that the post-intervention year for Newaygo was 1976, a year in which most jurisdictions were experiencing decreases in delinquent arrest rates (its comparison county had similar decreases for three of the four variables). St. Joseph County had the same post-intervention year and the data should be viewed with the same caution as was just suggested. St. Joseph also showed post-intervention decreases for the four variables and on two of these the comparison county had an increase (burglary and vandalism). St. Clair 150 County experienced decreases for burglary and larceny while the comparison county showed an increase in bur- glary rates. The post-intervention year for St. Clair County was 1975, another year in which overall delin- quency rates were on the decline. Berrien County also experienced decreases for two variables (burglary and vandalism). And the post-intervention decreases in Berrien County came in a year when the general trend for delinquency rates was on the rise (1973). The comparison county had increased rates for burglary and vandalism the same year. The significance of these apparent decreases will be examined shortly in the discussion of results from the statistical analyses of monthly data. The final two arrest variables to be examined are the two status offenses recorded in the U.C.R. data--curfew/loitering and runaway. Other status offenses such as truancy and incorrigibility would have been important variables to analyze since most of the projects focus a great deal of energy on the status offender, but these are not presently included in the U.C.R. reporting system. Table 11 contains juvenile arrest rates for curfew/loitering, and it can be seen that Van Buren, Newaygo, and St. Joseph Counties showed decreases in curfew/loitering rates for the post— intervention year. In all three instances, however, 151 Table 11.--Delinquency arrests for curfew/loitering (rates per 1000 juveniles, adjusted). Year countles 1971 1972 1973 1974 1975 1976 Calhoun 1.72 3.78 4.71 2.93 2.99 2.07 Jackson .81 1.43 2.03 1.23 1.16 1.31 Berrien 3.75 3.10 3.30 3.94 2.98 .03 Muskegon 2.24 1.75 1.60 2.79 1.03 1.43 Genesee 1.07 .99 1.28 1.39 .85 1.11 Saginaw 1.03 54 .85 2.39 1.60 19 Van Buren 1.55 .90 3.28 1.55 1.68 1.22 Allegan 30 .61 .93 .57 .84 .07 St. Clair 5.12 5.14 5.47 5.38 3.03 1.82 Lapeer 59 .59 1.57 76 .39 .39 Newaygo 56 1.99 .29 1.50 1.09 .00 Mecosta .00 .00 .00 .00 .21 .12 St. Joseph 2.50 2.99 3.53 2.32 3.07 1.10 Branch .99 .45 69 36 .49 .00 152 the comparison counties had similar decreases. A similar situation is found in the data in Table 12 for delinquency arrest rates for runaway. Van Buren, Newaygo, and St. Joseph Counties had post-intervention decreases in runaway rates, but the respective compari- son counties also registered decreases in the same year. Thus, there is no strong evidence to support the hypothesis that Youth Service Bureau jurisdictions would experience decreased rates of arrests for status offenses as a result of diversion activities by the projects. It is possible, however, that subtle changes in arrest rates would not show up in gross annual fig- ures, which is part of the reason for including the same variables in the monthly analyses. The results of the monthly time-series analyses of total delinquency arrests, burglary arrests, and lar- ceny arrests are presented in Table 13. Looking down the t-values for change in level, it can be seen that none of the apparent post-intervention decreases observed in the annual data are supported by signifi- cant findings in the analyses of monthly data. There are, however, two Youth Service Bureau sites where sig- nificant decreases in slope were found. Paw Paw experienced a significant post—intervention decrease in slope for total delinquency arrests and Three Rivers had a significant decrease in slope for delinquent 153 Table 12.-—De1inquency arrests for runaway (rates per 1000 juveniles, adjusted). Year countles 1971 1972 1973 1974 1975 1976 Calhoun 6.46 10.34 10.19 8.91 9.50 6.73 Jackson 5.12 4.46 7.09 8.73 9.72 6.00 Berrien 6.17 7.24 8.77 6.53 5.63 4.17 Muskegon 7.53 10.14 8.22 13.83 15.39 14.72 Genesee 6.96 8.98 9.14 11.44 10.22 7.99 Saginaw 6.36 5.90 11.08 17.16 8.86 10.30 Van Buren 6.28 6.65 8.09 6.37 7.90 5.52 Allegan 4.11 4.20 4.34 3.54 3.73 2.62 St. Clair 11.38 10.67 12.79 14.06 15.19 13.05 Lapeer 1.70 2.59 2.69 5.08 6.04 1.42 Newaygo 7.23 3.97 4.08 7.19 4.78 1.11 Mecosta 1.20 .61 2.90 2.10 6.08 2.50 St. Joseph 4.45 3.84 6.01 6.68 8.31 6.49 Branch 1.88 3.38 1.85 3.10 1.84 .51 154 Table 13.—-Summary statistics for monthly time-series analyses of diversion-~total delinquency arrests, delin- quency arrests for burglary, and delinquency arrests for larceny. Pre Post df Ihzunge t Change t in Level. in Drift Tbtal Delin- quency Arrests Battle Creek 13 42 52 2.670 .227 -.895 - .634 Benton Harbor 21 34 51 —7.784 -.721 2.801 1.199 Flint 21 32 49 24.661 1.734 -.038 - .030 Paw Paw 33 23 52 -1.056 —.877 -.287 —2.763* Port Huron 39 16 51 -4.136 -.401 -.O47 - .045 White Cloud 45 11 52 1.334 .612. —.155 - .499 Three Rivers 45 ll 52 -1.539 -.260 -.493 - .615 Delinquency Arrests for 91113123 Battle Greek 13 42 52 .090 .037 -.221 - .753 Benton Harbor 21 34 51 -l.656 — .824 .088 .495 Flint 21 32 49 12.404 3.579 .020 .065 PawPaw 33 23 52 ...a Pert Huron 39 16 51 -4.934 -1.358 .647 1.754 White Cloud 45 11 52 ... ... ... ... Three Rivers 45 11 52 -1.870 - .790 .450 1.406 Delinquency Arrests for larceny Battle Greek 13 42 52 - 2.364 - .386 -.633 - .858 Benton Harbor 21 34 51 - 1.601 - .495 .170 .593 Flint 21 32 49 -10.088 -1.l38 2.065 1.049 PaW'PaW' 33 23 52 ... ... ... ... Port Huron 39 16 51 - 2.988 — .755 -.361 — .899 White Cloud 45 11 52 ... ... ... ... Three Rivers 45 ll 52 .759 .337 -.584 -1.918* *Significant at the .05 level (t must be less than -1.67) with 50 degrees of freedom using a one-tailed test. aData in these sites could not be analyzed for this variable because of the large number of months in which no arrests were made or reported. 155 larceny arrests. Table 14 contains results from the monthly time—series analyses of the other three arrest variables-—vandalism, curfew/loitering, and runaway. It can be seen that Battle Creek showed significant decreases in post-intervention level for both of the status offenses (i.e., curfew/loitering and runaway), and a significant decrease in slope for curfew/loitering. Flint experienced a negative change in slope for curfew/ loitering arrests and Port Huron had a negative change in level for runaway arrests, both of which approached significance at the .05 level. In summarizing the annual and monthly data for the six arrest variables, several things can be said about the impacts of projects at the systems level which might result from the encouragement of diversion. Newaygo County (White Cloud) and St. Joseph County (Three Rivers) showed post-intervention decreases for all six arrest variables in annual data. But only the change in larceny arrests for Three Rivers was supported by the monthly data with a significant decrease in slope (the five specific offenses were not subjected to analysis for White Cloud because of insufficient data). St. Clair County (Port Huron) had post- intervention decreases in rates of burglary, larceny, and curfew/loitering at the annual level of analysis. While these observations were not supported by 156 Table 14.-~Summary statistics for monthly time—series analyses of diversion-~delinquency arrests for vandalism, delinquency arrests for curfew/loitering, arrests for runaway. and delinquency Gmnge Gmnge Pre P0513 df in Level t in Drift ‘3 Delinquency Arrests for vandalrsn Battle Creek 13 42 52 ...a Benton Harbor 21 34 51 2.338 2.451 .109 1.283 Flint 21 32 49 .438 .332 -.008 - .067 Paw Paw 33 23 52 ... ... ... ... Port Enron 39 16 51 8.152 2.388 -.348 ~1.003 White Cloud 45 11 52 ... ... ... ... Three Rivers 45 11 52 .494 .246 -.316 —1.162 Delinquency Arrests for Curfew/Loitering Battle Creek 13 42 52 3.497 -l.688* -.792 -3.177* Benton Harbor 21 34 51 2.107 1.034 .041 .229 Flint 21 32 49 4.913 2.530 -.272 -l.575 PawPaw 33 23 52 Pert Huron 39 16 51 —2.292 - .977 —.l79 — .750 White Cloud 45 11 52 ... ... ... . . Three Rivers 45 11 52 -1.134 - .833 .094 .513 Delinquency Arrests for Runaway Battle Creek 13 42 52 -4.021 -l.880* -.076 - .294 Benton Harbor 21 34 51 - .386 - .405 —.128 -1.217 Flint 21 32 49 -4.714 -1.227 .131 .381 PawPaw 33 23 52 Pert Huron 39 16 51 -4.445 —l.522 .391 1.317 White Cloud 45 11 52 ... ... ... ... Three Rivers 45 11 52 -1.085 - .758 .069 .317 *Significant at the .05 level (t must be less than -l.67) with 50 degrees of freedom using a one-tailed test. aData in these sites could not be analyzed for this variable because of the large number of months in which no arrests were made or reported. 157 statistically significant results from monthly analyses, it could be seen that related decreases in the monthly data followed the same general pattern with some approaching significance (e.g., change in level for burglary and runaway). Berrien County (Benton Harbor) experienced annual post-intervention decreases for two of the six variables (burglary and vandalism), although neither of these was verified in the statistical analy- ses of monthly data. Van Buren County (Paw Paw) showed post-intervention decreases for the status offenses-- curfew/loitering and runaway. Statistical analyses of monthly data for these two variables could not be carried out for Paw Paw because of insufficient data, although the analysis was performed on total delin- quency arrests for Paw Paw and a significant post- intervention decrease in slope was found. Neither Genesee County (Flint) nor Calhoun County (Battle Creek) showed post—intervention decreases for any of the arrest variables at the annual level. However, in the analyses of monthly figures, Battle Creek was found to have sta- tistically significant post-intervention decreases in the level of curfew/loitering and runaway arrests. Similarly, Flint was found to have a significant decrease in slope for curfew/loitering arrests, which did not show up in the annual data. 158 The evidence discussed thus far does not lend support to the hypothesis that diversion activities of the projects would result in decreased arrest rates in the respective communities. Some of the jurisdictions did appear to have experienced decreases on certain variables, but no jurisdictions showed consistent decreases across arrest variables and no variable was consistently affected across jurisdictions. As was mentioned earlier, it may be that projects are suc— cessful in encouraging the use of diversionary alter— natives but that law enforcement officials continue to invoke formal arrest sanctions prior to diverting. The following data on juvenile court referrals were included in the analysis to address this possibility. Table 15 contains the rates of police referrals to juvenile court for the six-year period. Berrien, St. Clair, Newaygo, and St. Joseph Counties had decreases in court referral rates in the post-intervention years although the latter two counties experienced the decreases in a year when the general trend for juve— nile court referrals was downward (1974) and the com- parison counties also showed decreases. In Berrien and St. Clair Counties, the post-intervention decreases in juvenile court referral rates were found in the absence of similar decreases in the comparison. None of these decreases was found to be statistically 159 Table 15.——Police referrals to juvenile court (rates per 1000 juveniles, adjusted). Year 0°“ntles 1971 1972 1973 1974 1975 1976 Calhoun 11.67 17.13 18.25 17.65 22.49 20.55 Jackson 16.40 19.68 17.20 23.46 28.27 22.46 Berrien 34.40 35.37 31.30 29.26 20.41 21.12 Muskegon 33.36 30.76 36.60 34.73 36.85 34.68 Genesee 16.28 16.49 17.51 20.12 19.78 18.79 Saginaw 16.24 19.43 17.66 17.54 10.72 10.51 Van Buren 21.71 30.95 24.71 28.66 30.73 32.37 Allegan 14.59 16.73 17.10 14.88 13.14 17.84 St. Clair 8.53 10.32 9.99 7.89 7.84 11.26 Lapeer 5.25 7.68 12.69 10.32 12.15 9.29 Newaygo 14.04 17.87 16.32 16.48 6.94 11.44 Mecosta 12.40 8.89 6.00 22.25 25.37 24.53 St. Joseph 9.27 15.25 24.32 29.91 35.80 31.36 Branch 8.95 15.57 13.05 10.50 8.35 8.34 160 significant in the analyses of monthly data, which are summarized in Table 16. Police referrals to court were broken down into Part I and Part II offenses in the statistical analyses of monthly data in order to sensi- tize the analyses to more subtle changes that might have been occurring. It can be seen in Table 16 that only the t-Value for the change in level for Part II referrals in Three Rivers even approached the .05 level of significance. But it should be mentioned that suf- ficient data for statistical analysis were not avail- able for several sites, making it impossible to check these sites for significant changes in police refer— rals to court. At the bottom of Table 16, results are presented for the statistical analysis of the final measure of court activity--total delinquency petitions. This measure represents the overall activity of the juvenile court insofar as it includes petitions from all sources (i.e., parents, schools, and other agencies, as well as the police). Only St. Clair County (Port Huron) experienced a statistically significant post- intervention decrease in level of total delinquency petitions substantiating the observed decrease in annual- level data on police referrals to court in St. Clair County. The examination of arrest and court petition data did not reveal consistent data in support of the 161 Table l6.—-Summary statistics for monthly time-series analyses of diversion--Part I and Part 11 police referrals to juvenile court, and total juvenile court petitions. Pre Post df Change t Change t inlxwel inIhift lkut Ilkflice Referrals to aunt Battle Greek 13 42 52 -.040 —.008 .074 .046 Benton Harbor 21 34 51 ...a ... ... ... Flint 21 32 49 16.009 2.460 .336 .576 PawPaw 33 23 52 PCrt Huron 39 16 51 ... ... ... ... White Cloud 45 11 52 ... ... ... ... Three Rivers 45 ll 52 .293 .090 -.329 -.745 Pwnlelkflice Referrals to Court Battle Creek 13 42 52 .877 .489 .258 .996 Benton Harbor 21 34 51 ... ... ... ... Flint 21 32 49 8.598 1.299 -.422 -.712 PaW'Paw 33 23 52 ... ... ... ... Pert Huron 39 16 51 .239 .107 .002 .200 White Cloud 45 11 52 ... ... ... . Three Rivers 45 ll 52 —4.588 -l.462 .152 .358 Tbtal .Delinquency Petitions Berrien Co. 33 39 68 26.202 1.929 3.489 .962 Genesee Co. 13 27 41 .850 .032 .499 .073 Macomb Cb. 43 63 104 - 3.453 - .562 .852 2.105 St. Clair Co. 27 21 44 -12.789 -l.746* 1.054 .503 *Significant at the .05 level (t must be less than -1.67) with 50 degrees of freedom using a one-tailed test. aData in these sites could not be analyzed for this variable because of the large number of months in which no arrests were made or reported. 162 general research hypothesis that Bureau jurisdictions would experience decreases in arrest and court petition trends following the initiation of diversion activities by the projects. Some of the sites experienced signifi- cant post-intervention decreases when the analyses focused on specific offense categories, and some sites had decreases on two or three of the eight arrest and court petition variables. But because of the large number of analyses carried out, a certain number of significant findings could be expected to have occurred by chance, thus making it necessary to have required consistent findings in order to claim support for the hypothesized system impact of diversion activities by the Youth Service Bureaus. The analyses to be dis- cussed next in the implementation section were designed to shed light on the differential impacts of projects and to clarify the factors that might have facilitated or hindered the projects in achieving systems-level goals. Implementation Analyses Delinquency Orientations of Project Staffs One of the most consistently mentioned factors in discussions of program operations and effectiveness has been staff orientation. Interest in this particu- lar organizational factor has stemmed from the belief 163 that operational guidelines are usually open to a cer— tain degree of interpretation so that the actual activi- ties of a program oftentimes mirror the personal orien- tations of staff members. It is for this reason that the Delinquency Orientation Scale was developed and utilized iniflmastudy. As was pointed out earlier, the overall framework of the study was to View organiza— tional and environmental factors as independent vari- ables that influence the activities and consequently, the outcomes of projects (the dependent variables). Initial analyses of the Delinquency Orientation Scale were focused on determining the psychometric properties of the instrument. Table 17 contains data pertaining to the internal consistency of the instru— ment. The initial alpha level for each original scale is shown, along with the final alpha after scale modi- fication (the antipermissive scale does not have an adjusted alpha since no modifications were made). Item frequencies were examined to determine if any item had insufficient variance to be included in further analyses (the criterion used was, at least, 10% variance). None of the items was deleted on this basis since no item was completed in the same manner by 90% of the respon- dents. Item-scale correlations are also presented in Table 17. Five items were deleted because they cor- related negatively with their scale, were not critical 164 Table 17.--Interna1 consistency data for the Delinquency Orientation Scale. Antipermissive Scale Alpha = .82339 N of cases = 46 Treatment Scale Alpha = .31221 N of cases = 46 Scale Scale Item Correlation Item Correlation V03 .481 V01 .030 V05 .582 Vll .274 V10 .475 V17 .140 V16 .787 V21 .103 V20 .610 V25 .130 V24 .424 V31 .033 V30 .601 V32 .439 Scale Scale Scale Correlation Scale Correlation SC2 .298 SC3 .025 SC3 -.070 SC4 —.130 SC4 .201 Reform Scale Alpha = .49188 N of cases = 46 Nonintervention Scale Alpha = .63443 N of cases = 46 Scale Scale Item Correlation Item Correlation V08 .192 V04 .223 V12 .443 V07 .417 V14 .211 V09 .483 V18 .313 V13 .457 V22 .014 V23 .191 V29 .246 V27 .178 V28 .250 Scale Scale Correlation SC4 .276 165 in the rational construction of their scale, and were not appropriate for inclusion in any of the other scales (variables 02, 0.6, 15, 19, and 26). The final alpha levels of the antipermissive and nonintervention scales fall into the range of acceptability (greater than .6) although the alpha levels of the treatment and reform scale fall short of the desired level. Finally, the scale—scale correlations are presented and it can be seen that the intercorrelations among scales range from -.130 to .298, indicating that the four scales are orthogonal, or tapping into independent conceptual dimensions. Using the modified scales, mean ratings were calculated for each of the nine sites that responded to the survey. These mean project ratings for each of the scales by site are presented in Table 18, along with the grand means for each scale. It can be seen that the Youth Service Bureau staffs, in general, indi- cated a fairly high degree of agreement with the state- ments representing individual treatment and social reform orientations (grand means equal 2.62 and 2.46, respectively). The grand means for the antipermissive and nonintervention scales were 3.81 and 3.95, indi- cating that there was a general disagreement with state- ments characteristic of these two orientations. Mean project ratings ranged from 2.21 to 5.14 and in every 166 .oouwdmfip mawqonpm u oo.® can monws hawGOMHm u oo.H d mo.m oe.m me.m Hw.m modes ensue sm.m om.m Ho.m ew.m on «soofls eo.m sm.m mm.m mH.e Aev onso>ese shone so.m mm.m mm.m 66.6 AHHV oooaxoo 3H.m so.m mo.m mH.o Ame oomnoseesm oH.e om.m mm.m mm.e Ame neooos oo.¢ oo.m mo.m ms.m Amv soonoo .em sm.m Hm.m oo.m ss.m Ase sHoHo .em no.3 os.m sm.m wo.m Amv oossm 53> mm.m Hm.m mo.m s~m.m A¢v oonosou ooHuco>Hopqflsoz ShOmmm unoSusone o>HmmHEHoQfiun¢ Ase moeam .opflm an onom moqosoawfimn may no mwafipsn adozll.wH manna 167 site, the treatment and reform scales received the highest ratings although neither of the two consis- tently received the highest. Similarly, the anti- permissive and nonintervention scales received somewhat lower ratings in every site although neither was con- sistently rated lowest. It appears that the Youth Service Bureau staffs combine individual treatment and social reform concepts in forming their dominant orien- tations to the problem of delinquency while rejecting (albeit, not strongly) the antipermissiVe and non— intervention notions. If certain activities and procedures are asso- ciated with these orientations, as was hypothesized in Chapter III, this analysis should be taken one step further to examine the relationships between dominant delinquency orientations and the actual Operations of programs as determined through discussions with project and related personnel. The overall high ratings given the treatment scale are consistent with the basic nature of most of the programs that place primary emphasis on individual treatment activities (e.g., casework and counseling, or direct services). Similarly, there is a degree of consistency between the low ratings given the nonintervention scale and the absence of "true" diversion activities associated with the nonintervention orientation (i.e., diversion out of the system without 168 the provision of alternative services). There is a puzzling incongruence, however, between the generally high ratings on the social reform scale and the lack of emphasis by most projects on directly affecting change in social institutions such as the schools and courts. The potential conflict between low ratings on the anti- permissive scale and cooperation with justice system officials who may not share this perspective also has important implications for the implementation of pro— grams. Both of these issues will be discussed at greater length in the final section of this chapter. Program Perceptions The Program Perceptions Survey was included in the study as an attempt to examine the Youth Service Bureau programs along certain organizational dimen- sions. The instrument contains three treatment scales (autonomy, practical orientation, and personal problem orientation), three relationship scales (involvement, support, and expressiveness), and three systems main- tenance scales (order and organization, clarity, and staff control). Brief descriptions of each scale can be reviewed in Chapter III. This phase of the study was intended to be exploratory in nature, as was the delinquency orienta- tion analysis, and the primary goal is similar inasmuch 169 as the analysis is focused on describing the organi- zational "characters" of the projects. Initial analyses were again aimed at examining the internal consistency of the instrument. First, it was necessary to delete a total of 15 items that were completed similarly by over 90% of the respondents (52 or more of the 57 respondents). Internal consistency analyses were then carried out and 17 other items were deleted because of their low or negative correlations with their scales (see Appendix G). Even after making the above revi- sions, the alpha levels for most of the scales remained low, as can be seen in Table 19. Alpha levels for the involvement, autonomy, and personal problem orientation scales were the only ones to exceed .6. Furthermore, it can be seen in Table 19 that several of the scale intercorrelations (ll of 36) are significant at the .001 level, indicating that there is a high degree of inter- dependence among scales and that they are not neces- sarily measuring separate dimensions. The psychometric analyses could not be carried further because of time constraints and a small sample size so that the meaning and validity of the Program Perceptions Survey data are open to question. With the shortcomings of this instrument in mind, mean ratings on each scale were calculated by site 170 .Ho>m~ H00. one as ucmowmficmwm zfiamowumflumwm who wcowumeuhoohoucfi ofimom omens. moo. owe. «me. mmo.u 022.- va.- wofi.n 842.- Amwm.v Honesoo wesem *mnm. one. mam. mom. sane. *mse. smae. Aswo.v spanseo 062. 4208. 6mm. won. some. amen. Ammo.v noose moo. mum. mwm. «on. coo. Adam.v soaeoesoaao $386 Hosomaoo mwm. 88m. same. mom. Ammo.v sosososoaao 3258.5 4666. 4086. ammo. Acme.v aeosooss 488. ...mmo. $va whosofiwmosea amen. Abom.v opossum :8; 58.5395 HWWWMMO 432.230 .820 :EHMMWMmWMWWMWCQ :MWWMWWMMMO 35:83» mmocmimmopg phonesm 2954,3495 OH 8m .>o>p=m mcoHuQmonmm EahMOpm mzu AOm «can accopwflmcoo A238 6238 HossoosH--.oH edema 171 and these data are presented in Table 20. The mean ratings have been multiplied by 10 in order to facili- tate inspection of the data. This means that the possible range for the mean ratings is 0 to 10 (0 would indicate that the dimension measured by a particular scale was not viewed as characteristic of the project, while 10 would indicate that it is highly characteristic). With regard to the three relationship dimensions, the scales measur- ing involvement of clients in program operations and ’support given clients by staff members generally received mid-range ratings while the expressiveness scale was rated among the highest overall. This is consistent with the ratings on the treatment dimensions since the expressiveness scale had to do with the open expression of feelings and the treatment dimension receiving the highest rating was personal problem orientation, which is concerned with examining and understanding personal problems and feelings. The personal problem orientation scale was consistently among the highest rated scales, while the autonomy scale, which assessed the degree to which initiative and leadership were encouraged, and the practical orientation scale, which assessed the extent to which clients were prepared for terminating from the program, consistently received ratings that were among the lowest. The systems maintenance scales measuring the importance of organization in the programs (order) 1372 .oflumfiumuomuwzo magma: $6 62 366 66:» 6626 6666666666 OH 6:6 cowu6mflcmuno uwmnp 80 oHumHuouowhano ma :ofimcmeflo x 666 6 eoa6 66 66666 was» :6 66:68 manfimmoq man .6666 one 80 6066666668 6666666666 06 86680 :6 OH 69 UCHQHHHSE mama 6:608 666 wonfiwa 6666 36«> oos 666 666656 eoo6oao page usaoaoaosa 6 6663 16.2 x 666 62 oo A66. 86.6 66.8 66.8 86.8 68.6 66.6 66.8 68.6 68.6 6666: 66666 86.6 66.6 66.6 26.8 26.6 66.6 66.6 66.6 68.6 A66 6:66H6 66.6 66.6 66.6 8H.6 66.6 66.6 66.6 68.6 66.6 A66 onao>688 66666 66.6 86.8 86.6 66.8 66.6 66.6 68.6 66.6 86.6 A666 6666666 66.6 66.8 66.6 66.8 66.8 66.8 66.6 86.6 66.8 A66 6666636666 86.6 66.8 68.8 66.6 66.6 86.8 66.8 66.8 68.6 A66 65666: 66.6 66.6 66.6 66.6 66.6 6. 66.6 86.6 66.6 A66 soomow .66 66.6 66.6 66.8 66.6 86.6 86.6 86.6 86.6 66.6 A66 66666 .66 66.6 66.8 66.6 86.6 66.6 66.6 66.6 66.8 66.8 A66 666:6 66> 86.6 66.6 66.6 66.6 66.6 66.6 86.6 66.6 668.6 A66 6666666 HHMMMO 36.6.68 666.5 Sfimfim NMWWMLOCQ Emma“. WWMMMO 6506063.. 6666966666966 9.869% 66686395 2: 86m mfimom .Illlll .6666 up mw>psm mcofluomouwd EwngpQ map so mwcfiumh 666511.06 mHn6B 173 and the explicitness of project activities (clarity) generally received high ratings in all sites. The §EE££ control scale, which measured the extent to which staff members use rules and regulations to control clients, was consistently rated lowest in all sites except one. Thus, it appears that the Youth Service Bureau staffs at the various projects had similar perceptions of the internal organizational characteristics of the programs. While this organizational component of the study concentrated on factors hypothesized to affect the imple- mentation of programs, delinquency orientations and pro- gram perceptions probably have a less direct impact on the form taken by the program than do the interactions of project staff with external persons and organiza— tions. The implementation strategies used by projects can be viewed as having direct and critical impact on the eventual substance of a program, as well as its ability to effectively pursue social change goals. In the final component of the study, attention will be focused on the approaches taken in the various projects to initiating and maintaining programs within the pre- existing community social structures. Environmental Influences The information regarding implementation strate— gies and the effects of the environment on program 174 development was obtained through structured interviews with a number of persons at each of the sites included in this component (see Table 1 in Chapter III). Each of the topical areas to be discussed in this component was raised with all persons interviewed so that a range of opinions could be examined. As expected, most of the projects faced a number of similar obstacles during the implementation phase. The primary goal in the follow- ing discussion is to explicate the various procedures used by the projects to overcome implementation barriers and the perceived effectiveness of different approaches. Perhaps the most common problem facing new social programs is the garnering of public support. All of the Youth Service Bureau programs were relatively small programs within the communities and many of their activities were highly dependent on other agencies and institutions. In almost every case, initial project activities were aimed at developing the relationships and arrangements necessary for the accomplishment of project goals. One of the most common methods of gen- erating support was the involvement of critical persons in the program. Certain individuals and organizations were involved in the program in a policy capacity (e.g., as a member of an advisory body) while others were encouraged to become involved with the program as a user of services. Most of the projects were successful in 175 generating these initial bases of support without which they would have had little opportunity to succeed. How- ever, the elicitation of a support base is not generally without implications for the evolution of a program. In many instances, persons in the community close to the delinquency problem were involved in the initial formulation of programs. Juvenile court judges and other court personnel were particularly active in the origination of Youth Service Bureaus. There are a number of obvious advantages to having court personnel (especially a judge) closely involved in a program from the outset, as well as law enforcement representatives and school authorities. The inclusion of these indi- viduals in the activities of the project provides a certain degree of credibility in the community and opens channels of communication which are critical links if the programs are to develop diversionary procedures and alternatives. But in every site where it had been neces- sary to gain support through encouraging active partici- pation in the program by external actors, the projects were expectedly affected by these interactions. Often- times the results appeared to be significant modifica- tions in the basic program concept. At one site, at least (several were facing a similar situation), it became necessary for the project to institute formal procedures for applying legal sanctions in cases where 176 clients did not conform to program expectations. This was done in response to criticism by school and police officials who felt that the project should take a ”harder” approach to dealing with referred cases. Faced with the potential loss of referrals, it is not sur- prising that projects must consider changes that may alter the basic program model. There are several possible approaches to the potential conflict between engendering support by encour- aging critical actors to participate in the program and maintaining the integrity of the program design. One of the most common (although, perhaps, least effective) is that of making a concentrated effort in the early phases of the project to educate a broad range of indi- viduals throughout the community as to the logic, goals, and activities of the project. All too often, the ambiguous nature of a program is maintained under the guise of flexibility in order to facilitate ad hoc adap- tations to changing social and political whims. This is not to say that a certain degree of flexibility is not desirable for meeting changing needs in the commu- nity. But if the program model is to be evaluated and compared with other alternatives to dealing with a prob- lem, basic characteristics that distinguish the program from others must remain intact. 177 Another approach to soliciting community par- ticipation without sacrificing program design is to recruit influential persons who are in basic agreement with the goals and methods of the program. While this is obviously a more difficult endeavor, if carried out successfully the individuals involved can lend credi- bility and stability to the new program, as well as serve as spokesmen for the basic ideals of the program. In some of the sites, it was possible to generate the necessary level of support without formally involving community influentials in program operations. This was accomplished through far-reaching and continual efforts to familiarize the entire community with the nature, objectives, and activities of the program. A final approach to this problem seen in the projects was to concentrate on developing and nurturing one or two crit— ical relationships (e.g., with a school system), which would have the potential to maintain the project when "soft" monies were no longer available. Two other issues investigated during project assessment interviews have to do with the location of projects in the community social structures and the degree of clarity and consensus regarding the role of the project in the system of human services. Both of these issues may have significant impact on the develop- mental paths taken by projects and the resultant 178 abilities to effectively pursue stated goals. As for the location of the project, a similar kind of situation exists as was just discussed with regard to the rela- tionship between program participation and support. There are the obvious advantages to being closely aligned with the juvenile court or some other stable institution in the community. But serious questions arise regarding the degree to which innovative activities can be stimu- lated within the framework of a traditional, well- entrenched institution. There were indications that projects related to the court, for instance, were often perceived as an auxiliary of the court even though con— centrated efforts may have been made to repudiate this image. The location of the project is likely to influ— ence the activities of a program, as well as its image. As was seen in the last component, for example, several project staffs indicated a fairly high level of agree- ment with the social reform orientation to delinquency although there were few examples of these kinds of activities seen in the project assessment sites. It appeared that certain projects had contemplated activi- ties aimed at affecting systems-level reform (e.g., the modification of school expulsion policies, the nonlegal handling of delinquent status offenders, etc.), but more often than not these efforts were sidetracked. Part 179 of the problem undoubtedly had to do with the complexity of these kinds of endeavors and the relative paucity of procedural guidelines. But in many cases, the major obstacle to innovative social reform was the absence of 'understanding and/or agreement with this type of activity. The problem appeared to be exacerbated when a project was formally or informally aligned with an established agency. The one project in the study that pursued institutional change goals as a primary activity encountered a myriad of problems and was unsuccessful in garnering enough support to complete the initial three-year grant period. Discussions with persons involved in this project indicated that the most trouble- some elements had to do with the nature of the program rather than its location in the social structure (the project was a relatively independent operation within the community mental health organization). Throughout the interviews, there was a great deal of discussion regarding the nature of programs in terms of goal clarity and consensus and the influence of these factors on the implementation and success of the various projects. There was some discussion earlier about the manner in which social policy is formulated, and the evolution of program designs from the generally broad and abstract guidelines. This situation is often decried as one of the most basic problems in social 180 programs. It can be argued, however, that such open- ness is necessary because it allows adjustments to be made in program designs that are more in line with local needs. A critical element generally missing in the process of formulating and implementing social programs is a systematization and delineation of the acceptable range of program models. The tradeoff is between allow- ing for responsiveness to local needs and maintaining those characteristics of a program that are integral parts of the underlying conceptual framework. It was felt in several sites that the basic Youth Service Bureau concept had been diluted due to a variety of external pressures. An example of this was alluded to earlier in talking about the element of coercion being built into certain programs. Other facets of the Youth Service Bureau concept have undergone similar transformations including the previously mentioned absence of institu- tional change activities, the de-emphasis of resource development and coordination of services, the downplay— ing of youth advocacy functions, and the perceptible shifts from the original notion of diversion out of the justice system to one emphasizing the prevention of entrance into the system (i.e., the targeting of young, first-time, minor offenders). The goal clarity issue is important because it is so intricately tied to project activities and the level 181 of success achieved. Without some firm sense of the range of acceptable goals and activities, the feasi- bility and measurability of the specific subgoals, and the criteria for success, it is impossible to meaning- fully test the precepts of a program concept. At the various sites, it was found that while the general goals of diversion and crime reduction were consistently present there was a wide variety of subgoals and activi- ties. The subgoals derived from the two general goals reflected the move away from systems change toward an individual effect orientation (e.g., instead of specify- ing measurable subgoals in terms of systems variables such as crime statistics or school dropout rates, there was a tendency to interpret the major goals in terms of individual project youths resulting in subgoals like decreasing the arrests of project clients by 10%). There was a commonality throughout the projects insofar as most concentrated on activities that were individual— oriented direct services such as counseling and case— work. But under the rubric of direct services, there were activities ranging from job counseling, tutoring, and testing to individual therapy, group discussions, and family counseling to name but a few. The effect of this situation is to render the evaluation of a program concept ambiguous because of the lack of comparability across sites—-the concept is tested disjointedly in 182 various forms in unique settings rather than sys- tematically as one program model under different commu- nity conditions viewed as intervening variables. The issues regarding goal clarity, feasibility, and measurability are matters which can be directly addressed at various stages of the policy-making pro— cess. Issues having to do with goal consensus, however, are not as accessible for manipulation in the policy- making process and are likely to be dealt with at the project level. Interviews at all sites revealed that there was a lack of consensus regarding the roles to be played by the new projects, which is to be expected when a social innovation affects established practices and relationships. Particularly evident was the dis- agreement among police and school officials with pro- grams advocating the expanded use of nonlegal approaches to youth problems. In most cases, it appeared that police and school personnel who dealt with the problems most directly generally felt that the most effective approach was the use of more and stricter legal sanc- tions. And officials of the juvenile courts were often found to be in disagreement with project guidelines regarding who is to be diverted, at what point in the system, and in what manner (true diversion or minimiza- tion of penetration with alternative services). The primary approach to these problems has been to educate 183 the community as to the rationale, form, and goals of the program. While certain public relations activities are undertaken at the various sites (e.g., speaking engagements, use of the media, etc.), little systematic planning is done to clarify the overall role of a project and sensitize the staff to the problems of infusing the basic notions of the program throughout the community. The final area addressed in the interviews was the presence or absence of socio-historical factors that hindered or facilitated the implementation of projects. It is obviously not possible to summarize the many unique events that arose in the different communities. But there were such factors that appeared to be wide- spread enough that numerous projects had experienced them. For example, a long—term growth in the size of the juvenile population has begun to level off in this decade and part of the rationale for Youth Service Bureaus (i.e., alleviating the overburden of cases in the justice system) has been brought into question. Increased concern over the legal rights of juveniles has perhaps worked in favor of the Bureaus since many agencies would likely prefer making use of an informal "helping" program rather than a depersonalized, legal- istic justice system in all but the most troublesome cases. Moves toward limiting the scope of the justice 184 system through decriminalization have also probably worked in favor of the projects as they came to be viewed as alternatives to the justice system for deal- ing nonlegally with certain classes of problems (e.g., the status offenses). Another widespread phenomenon that has undoubtedly influenced the perceptions of Youth Service Bureaus is the general disillusionment with new social programs. Numerous programs have come and gone without having highly visible effects on the problem area to which they were addressed, and it appeared through the interviews that many persons were skeptical about the prospects of a successful Youth Service Bureau even before it started. There is, of course, an endless list of such phenomena that could have affected the implementation of the Youth Service Bureau projects. The purpose here was not to survey the range of environmental influences, but rather to illustrate that uncontrollable external variables are at work and should be considered within the particular situation during the development, implementation, and evaluation of new social programs. CHAPTER V CONCLUSIONS AND POLICY IMPLICATIONS Crime Reduction and Diversion by Youth Service Bureaus With regard to the systems-level impact of Youth Service Bureaus, the outcome analyses in this study lead to conclusions not dissimilar to the growing body of evaluation research literature which has found the effect of "little effect," to use Weiss' phrase.1 In both the crime reduction and diversion components, results from time-series analyses of police and court data indicated only sporadic evidence of project impact which could have been expected to occur by chance given the large number of analyses carried out. There were no instances where a particular variable in the crime reduction or diversion components was consistently affected across sites, nor were there any instances where a particular site showed significant impact across the target vari- ables. Several factors can be cited as possible explanations for the general lack of positive findings in evaluation research. First, there is the problem of 1Weiss, Evaluation Research, p. 126. 185 186 overadvocacy wherein programs are set up for failure because of the lofty and unrealistic goals which are espoused. The Youth Service Bureaus examined in this study were relatively small and minor components of the total youth-serving systems in the communities, yet they were expected to impact upon major social prob- lems which had evolved over many years. Closely related to this problem is the fact evaluation is usually centered on new programs which have not been integrated into the community social structure, making it even less likely that the cooperation and support necessary for attaining stated goals could be generated. There is also a critical time factor involved in most program evaluations. The Youth Service Bureaus in this study, for example, had been in operation for an aver- age of about two years and part of this time had been spent becoming functional. It is very unlikely that even the most well-conceived and implemented program could impact upon a broad-spectrum problem such as crime and delinquency in such a short period. Another response to the lack of positive find- ings at the systems level might be that the projects have actually focused their energies on direct ser- vices and significant impact should be expected at the individual level. But the argument was made earlier that in the case of the Bureaus, success at the systems 187 level (i.e., in diverting youths from the justice system) was a prerequisite for meaningfully assessing impact at the individual level. This is the case because unless the projects were, in fact, working with youths who had been diverted from the justice system (which was not indicated), the question of indi- vidual effects is a moot one. And since the Bureaus do serve a large number of youths in the absence of evidence that these clients are diverted from the jus— tice system, it may be concluded (as it has been often in past evaluations) that an unanticipated consequence of diversionary programs is to include a wider range of youths in the expanded formal social control system. Another unanticipated consequence of diversion programs, in general, might be to increase the severity with which those who are processed in the justice sys- tem are dealt. Operating under the assumption that only the "tough kids” are processed through the system, there may be a tendency for justice system officials to become stricter with what is essentially the same sample of youth problems. Finally, the criticism has been made generally of diversion programs that they are piecemeal programs which may have the effect of hindering attempts at bringing about more basic changes in the justice system. For example, many Youth Service Bureaus have made it known that they are ready and 188 willing to provide "needed” services to status offenders if these behaviors are removed from the jurisdiction of the justice system through decriminalization. If this were to happen, the basic goal of decriminalization will have been defeated (the basic goal being the removal of a behavior from the sphere of formal social control for regulation through informal mechanisms of control like the family). If benefits are to accrue from program evalua- tions, they must be in the area of clarifying why so many programs appear to be unsuccessful in achieving their goals and objectives. While the untrained observer can likely come up with a range of possible explana- tions similar to those discussed above, what is needed is an explication of program-specific factors assessable through the policy-making process which might facili- tate the functioning of social programs. The implemen— tation analysis components of this study were geared toward this end. By focusing on a number of organiza— tional and environmental variables it was hoped that information would be generated which would help clarify the internal and external processes affecting the suc- cess or failure of projects. After examining these independent, or intervening, variables, there was to be a discussion of how these factors might be manipulated 189 through the policy process. Such is the plan of the remainder of this chapter. Program Implementation It was originally planned that the implementa- tion results would be related back to observed varia— tions in success from the impact analyses. However, since there was little variation in success rates with all of the projects having minimal impact on the justice system, the implementation data took on a different meaning. Rather than clarifying the influences of organi— zational and environmental factors on specific programs, the implementation analyses become important in the explanation of the general lack of success of the Youth Service Bureau concept in affecting systems-level change. The environmental assessment interviews pointed to a number of problems which undoubtedly explain, in part, the overall failure of the Youth Service Bureau projects to affect systems-level change. Most of these problems including the lack of local support, inadequate resources, and position in the social structure were troublesome at each project site. And the absence of significantly more or less successful projects precluded the possibility of pulling out specific implementation factors related to project effectiveness. While the projects showed little variation as far as overall 190 systems impact, there were differences in the one or two variables which were affected at the various sites. However, there were no implementation data to support the notion that there was any logic to the project impact patterns where one site would experience decreases in actual burglary and runaway arrest rates, another in larceny and vandalism arrest rates, and yet another in actual vandalism and juvenile court petition rates. 'The status offenses and court petitions were the only outcome variables verbalized as target variables, and they were not affected to a greater extent than any of the other crime variables. Overall, the interviews with external persons (police, court, school officials, etc.) resulted in few data regarding program activities which might have been useful in explaining specific program outcomes. A wealth of information was generated with regard to per- ceptions of and attitudes toward the general Youth Ser— vice Bureau model. But in only a very few instances were external persons familiar enough with the programs to discuss the details of what the expected outcomes of a particular program might be. Thus, while the projects did appear successful in educating the commu- nity as to the existence and general nature of the projects, there was little understanding of program inner workings within the group of persons interviewed. 191 The major obstacle to collecting meaningful implementation data for this study was the ex post facto nature of the research. In order to gather the most useful implementation data, it is necessary for project staff to have continually documented program development. But as was pointed out in a quote earlier in the study (p. 14), day-to-day problems are typically worked out on an individual, ad hoc basis resulting in undocumented program shifts which in and of themselves are minor but which cumulatively can alter the basic nature of the pro- gram. This was found to be the case at the Youth Service Bureau sites and, consequently, documentation of program development was sketchy with little information regard- ing subtle alterations of the original program concept. It has only been recently that the importance of the implementation process in social programs has been recognized. As was discussed earlier, the parameters and guidelines for implementation analysis are not as thoroughly conceptualized as those for the assessment of outcome. Because the measurement of implementation fac— tors is so much less sophisticated, the results of the implementation analysis are not as concrete as those in the outcome analysis section. But it is believed that the implementation results provide usable information 192 regarding the selected factors, and suggest further paths of analysis. The implementation analysis was limited by time and financial constraints to a fairly small number of variables, and there are unquestionably other impor- tant organizational and environmental factors which are not included here. But from the implementation analyses which were carried out in this study, three general conclusions can be drawn. First, it did not appear that Youth Service Bureau staffs held personal orientations to delinquency which are associated with the unique activities embodied in the original Youth Service Bureau concept (i.e., resource development, coordination and service brokerage, systems modifica- tion, etc.). In general, the project staffs appeared on both survey instruments to have strong inclinations toward viewing delinquency in the traditional manner as primarily an individual problem requiring the use of commonly accepted treatment approaches such as coun- seling. Second, it can be concluded that the designs and functions of the various projects underwent enough modification during implementation that most of the core elements of the Youth Service Bureau concept were 193 untested in this study. What was tested was the effec- tiveness of traditional direct services with identified pre-delinquents outside the confines of the juvenile justice system. The third conclusion to be drawn from the imple- mentation analysis is that the projects were unprepared and ill-equipped for the complex task of implementing a program aimed at affecting systems-level change. In most instances, it was felt that time, training, and resources were oftentimes insufficient given the range of activities required by the projects. These types of findings with regard to imple- mentation are not unique to the Youth Service Bureaus and have undoubtedly been discovered in other program evaluations which looked at the processual side of evaluation. This is especially likely in programs where attention is directed toward innovative approaches such as systems change. Typically in such programs, there are few theoretical or practical guidelines and the programs tend to evolve on an ad hoc basis often result- ing in the dilution of the original program concept. Perhaps the most important contribution to be made by those emphasizing the implementation aspect of evalua- tion is the documentation of program evolution. More will be said about this topic momentarily when the implications of the study are discussed. 194 Another important benefit of implementation analyses is a clearer understanding of goal—outcome assessment. If evaluation results are to become an integral part of the policy-making process, more is needed than mere statements that a program has been successful or unsuccessful in attaining its stated goals and objectives. At this stage when most social programs appear somewhat less than successful in achiev- ing their broad-scale aims, knowledge about implemen- tation can be used as feedbaCk in on-going planning and development activities. As this kind of information accumulates and is used to improve the functioning of programs, implementation data could then be utilized to enhance the replicability of programs found to be suc- cessful. The results from analyses of implementation factors often have direct implications for the various phases of the policy-making process. While impact analyses provide information for basic policy decisions, the validity of the information is questionable if it is not supported by a comprehensive review of imple- mentation procedures. Throughout the policy—making process there are opportunities to address the kinds of issues arising out of implementation analyses. To conclude this study, some of the implications of the 195 findings for the policy—making process will be dis- cussed. Implications for Policy Two of the major problems alluded to throughout this study were: (1) the need for a delimitation of an acceptable range of explicit program models which can be used to operationalize a concept; and (2) the absence of guidelines for the introduction and maintenance of program models. The first problem is one requiring attention early in the program development phase. The second becomes more critical after the program models are delineated and the implementation phase has begun. And both of these problems are inextricably linked to the evaluation of program concepts and models. In the following sections, the policy implications from this study will be discussed specifically with regard to the program development, implementation, and evaluation phases of social programming. Program Development One of the most noticeable shortcomings in program development is the apparent lack of attention given the theoretical underpinnings of program concepts. This has become more and more apparent from the evalua— tions of Youth Service Bureaus nationWide. Even where program models have been specified, they usually have 196 not been systematically derived as planned variations of the program concept. In the case of Youth Service Bureaus, there have been three general program models specified--service brokerage (coordination and referral), resource development, and systems modification. There is some question as to the congruence of the first two models with labeling theories underlying the Youth Service Bureau concept. Both of these approaches imply the continued use of formal social control processes outside of the justice system and are in contrast with labeling theory notions about the avoidance of stigma associated with processing by formal social control agencies (i.e., being labeled as a "Y.S.B. kid" may have a stigma similar to that accompanying the label "delin- quent"). The systems modification model is probably the closest conceptually to the labeling theory notions underlying diversion programs, but this model is rarely found in operation. The general trend has been for Youth Service Bureaus to move in the direction of more familiar direct service models. Policy implications in the area of program development revolve around gaining the degree of control over program evolution necessary to prevent the dilution of the program concept. This ideally would involve several interrelated steps. First, detailed study of theoretical and philosophical writings pertinent 197 to the subject area would need to be undertaken. There would also need to be a thorough review of empirical and qualitative research studies of past and existing programs with characteristics similar to those of the proposed program. Then, on the basis of this informa- tion, a limited number of operationalizable program models would be explicated. And for each of the speci- fied models, parameters would be set regarding target groups, ranges of activities, expected impact areas, and criteria for success. Finally, program models would be piloted in a limited number of experimental sites and final refinements of program models would be completed prior to large-scale implementation. Following the systematic derivation and pilot- ing of operational program models, the issue then becomes one of maintaining program designs during wide— range implementation. It has already been suggested that the maintenance of program design does not neces- sarily go against notions that programs must be flexible enough to adjust to varying community needs and desires. The critical factor is the recognition of which program elements are unalterable for the testing of the program concept and which can be adjusted (and by how much) without affecting the essential nature of the concept. These are concerns which are most effectively addressed during the implementation of program models. 198 Implementation Guidelines As in the program development phase, initial implementation activities would involve a detailed review of relevant theories and studies. In the case of Youth Service Bureaus, this would have meant the intensive examination of theories on planned social change and studies of planned change approaches to implementing innovations. Through this process general strategies for implementation could be formulated and guidelines for dealing with specific factors clarified in the literature set forth. It is, however, during the actual implementation of projects that the most useful kinds of information become available. What has been missing in the area of social programming has been a feedback mechanism for solicit- ing, codifying, and generating information about the day-to-day activities and issues arising during imple- mentation. The results of this study and those of studies in many program areas indicate that there is a general insensitivity to the importance and complexity of implementation. Perhaps the most direct implication of this for policy in the area of implementation is the suggested emphasis on gathering, submitting, and con- stantly reviewing implementation data. Just as guide- lines are set out with regard to outcome data (e.g., in diversion programs, the suggested focus on rearrest 199 rates or court petitions), directions could be set for the collection and use of implementation data. In this study alone, several implementation issues sur- faced about which little has been said in policy and program guidelines--1ocating in the community social structure, training staff in the theories and methods of specific program models, educating individuals and agencies in the community as to the program goals and activities, creating a base of support, and becoming integrated with on-going community activities. Along with detailed study of the programmatic area and feedback mechanisms for both outcome and imple- mentation data, other policy-related matters could also improve the implementation and, subsequently, the effectiveness of social programs. One such matter involves the inclusion of project applicants in inten- sive training and educational seminars during the grant application period to thoroughly familiarize them with program requirements and expectations. The effective- ness of the pre-funding orientations would be highly dependent on the degree of success during the program development phase in spelling out clearly the parameters and expectations for various program models. The dual aims of orientation seminars would be to heighten awareness of program theories, models, and goals and 200 to dissuade those who are not in agreement with basic program assumptions from further pursuing projects. Another policy—related matter which could have a significant influence on the quality of social pro- grams is the collection of data from relevant actors in the communities regarding their perceptions of and attitudes toward the projects. This type of informa— tion could be used in assessing the projects' successes with regard to informing the community of the nature and goals of programs. But more importantly, it might be used to determine the general receptiveness to the new program and to indicate where there are likely to be problems during implementation. Finally, with an increased sensitivity to the complexities of implemen- tation, it would be hoped that funding patterns could be modified to reflect the awareness of implementation concerns, thus alleviating the major implementation barrier or instability. Policy implications discussed so far with regard to program development and implemen- tation would also have a significant impact on program evaluations and the use of evaluation results in the policy-making process. Evaluation Strategies Numerous evaluation problems result from the failures during program development and implementation. 201 When program models are not clearly specified and adherence to program guidelines is loose, the evaluator is faced with the problem of having to decipher a defi- nition of the program and continually redefine the evaluation design as characteristics and perceptions of the program change. Part of this problem is insoluble given the lesser degree of control over variables in social research as compared with the laboratory (e.g., it will never be possible to manipulate economic condi- tions the way noise levels are dealt with in a labora- tory experiment). But the policy-level activities discussed thus far would provide a major boost to the improvement of social research. There are also direct policy implications for the development of evaluation strategies which could have measurable impacts on the quality of evaluation research findings. One of the major implications of this and other ex post facto studies has to do with the failure to integrate evaluation concerns with social and political considerations in the earliest stages of program development. If the policy-making process were to include considerations of expected program outcomes and how they would be measured, it is likely that the aims and objectives of programs would be stated in more realistic terms (instead of such lofty terms as "ending poverty” or "stopping crime"). 202 Even under ideal conditions of program devel- opment and implementation, certain basic policy deci— sions regarding evaluation would be required. First, just as it was suggested that parameters be set for the range of acceptable program models, parameters need to be set for the range of acceptable evaluation designs. For example, it is not productive to collect a large amount of outcome data if comparison data are not also gathered. And if no data are collected prior to the introduction of the program, it becomes impossible to get an accurate measurement of pre-post change on any variable. Along with compendia on acceptable evaluation designs, specifications of relevant outcome criteria are required. The list of outcome criteria should follow directly from explicit program models origi- nated during program development activities. Besides being relevant to program goals and activities, outcome criteria must meet the requirements of measurability and replicability. If an outcome variable cannot be accurately measured using current research techniques and if these measurements cannot also be made in a wide range of situations, the variable is not likely to generate information useful for making policy decisions about the validity of a program concept. 203 With evaluation designs and outcome criteria specified, attention then needs to be directed to detailing how measurements are to be taken on each variable. The Youth Service Bureaus, for example, have placed a high degree of emphasis on preventing delinquency and, assumably, affecting delinquency rates. But there is little consensus on what measures best reflect "delinquency." The term can be taken to mean anything from the number of policy contacts with juveniles to the number of juveniles in correctional facilities. With an unlimited variety of meanings and measurements of delinquency, little can be done in the way of cross-program comparisons. Policy decisions are required to insure a degree of comparability along outcome measures across the range of projects within a particular program model, at least. There are also needs for policy directives regarding the analysis and presentation of data. In reviewing various types of data collected by the pro- jects in this study, it was noticed that widely varying analytical techniques were used making the comparisons of program outcomes difficult. Some projects examined client outcomes by looking at the total number of rearrests, others focused on the percentage of clients rearrested, others looked only at recidivism for those clients who completed the program, and still others 204 checked recidivism only during the period of partici- pation in the program with no follow-up. With all the unique twists to data analysis and presentation for the various outcome measures (the same situation existed in the analyses of court variables, school-related measures, etc.), it was next to impossible to make comparisons across programs. Also, statistical tests for signifi— cance were rarely used (not even simple t—tests) and the use of even nonequivalent control groups was almost nonexistent, making the meaning of the findings highly suspect. Finally, it has been strongly suggested that implementation issues must be given greater emphasis in all phases of policy making and program evaluation. More attention is required in the formulation and main- tenance of programmatic approaches. While it is doubt- ful that the "experimental treatment" in social research can ever be controlled the way it is in the laboratory, implementation analysis can provide dynamic program descriptions which would serve as documentation of pro- gram change and metamorphosis. This study has attempted to demonstrate the use of a systems model approach in the evaluation of the Youth Service Bureau concept. It has illustrated the complexities of the systems model and has highlighted the difficult problems encountered in program evaluation. 205 The whole field of social research is in an embryonic stage where major advancements are needed. Throughout this discussion, it has been suggested that refinements are called for in both the social policy-making process and the fields of applied social research and program evaluation. It is unlikely, however, that such refine- ments will be made unless cooperative working relation- ships are fostered between policy makers and researchers so that the agenda and activities of policy making and social research are intertwined. In order to most effectively address the deep-rooted and complex social problems of today, it will be necessary to integrate the measurement and validity concerns of research with the financial and political concerns of policy making. APPENDI CBS 206 HrabaH wNHOLomK‘lmUlvbWND-l APPENDIX A PROJECT SUMMARIES Berrien County Genesee County Calhoun County Van Buren County St. Clair County Newaygo County St. Joseph County Macomb County Shiawassee County Oakland County Grand Traverse County Alpena County Kalamazoo County 207 APPENDIX A PROJECT SUMMARIES Berrien County (Benton Harbor) The Berrien County Youth Service Bureau began operation on July 1, 1973. The implementing agency was the Juvenile Division of the Berrien County Probate Court. One of the recommendations of the John Howard Association in 1971 was to provide a community-based, diversion alternative for youth, especially status offend- ers. In addition, Berrien County had the second highest juvenile crime rate in Michigan and suffered from social economic problems--unemployment, racial conflicts, low academic achievement, etc. Project objectives included the significant reduc- tion of the number of official arrests, school suspensions or expulsions, and court petitions involving youth referred to Youth Service Bureaus. Program activities included a juvenile information exchange for police agencies and schools, short-term coun- seling, screening of all police complaints to determine appropriate action, referral services to community agen- cies, consultation to parents and agencies, follow-up evaluations of youth with high potential for recidivism and a volunteer program. 208 209 The staff of nine included a director, case super- visor, five youth counselors, and two secretaries. Genesee County (Flinp) The Genesee County Youth Service Bureau began operation July 1, 1973. The implementing agency was Flint Community Schools and Director of Youth Projects was named project director. In 1972, the Genesee County Juvenile Delinquency Planning Unit determined a need for (1) diverting youth from official adjudication and (2) coordinating community youth service agencies. Genesee County was designated as an LEAA high crime area and Flint was one of ten Michigan Crime Impact Cities. While probability of arrest was low, 68 percent of youth arrested were referred to probate court. Also noted were lack of secondary prevention ser- vices, lack of coordinated community programming, high rate of school suspensions, and lack of information, resources. Project objectives included: (1) diverting first and second offenders and nonassaultive offenders; (2) reducing number of accepted court petitions, Youth Service Bureau participant arrests and school suspensions (or expulsions; (3) mobilizing community resources to pro- «vide needed youth services; (4) reducing delinquency in (Genesee County; (5) reducing size of probate court case— ]xaads; (6) referring 25 percent of its referrals to 210 existing community agencies; and (7) documenting youth programming needs and developing appropriate services. Services provided were short-term counseling to youth and parents, service brokerage, resource develop- ment, and systems modification. Staff included a director, two community service coordinators, two youth workers, a program evaluator, and a secretary. Calhoun County (Battle Creek) The Calhoun County Youth Services Bureau began operation January 1, 1972. The Calhoun County Juvenile Court was the implementing agency and the Director of Court Services was named project director. A growing rate of delinquency in the county pre- cipitated the formation of the Youth Service Bureau. Over half the juveniles arrested were warned and released with no services provided. Furthermore, police agencies had no standardized referral procedures, schools were hesitant to refer truant and incorrigible youth to court, and coordination of referral to social agencies was min- imal. The goals of the Youth Services Bureau aimed at ;providing services to previously unserviced juveniles, at increasing resource development, and at decreasing probate czourt caseloads. Its objectives were: reduce the number 211 and/or quality of official arrests, school suspensions or expulsions, and accepted court petitions of youth partici- pating in the Bureau program. The Bureau provided a Juvenile Information Exchange Service for police agencies, coordinated referrals to com- munity agencies, provided short-term counseling to youth referred by schools and police, consulted and advised parents and professionals working with involved youth, screened all police complaints on first-offense juveniles, conducted follow-up evaluations of "high-risk" youth, referred to juvenile court when necessary, and planned to implement a volunteer program in the future. The Bureau operated as a branch of the Calhoun County Juvenile Court. An advisory council comprised of representatives of school, police, court, and agency per- sonnel provided on—going planning, training, consultation, and evaluation of the Youth Service Bureau. Bureau staff included a director, assistant direc- tor, coordinator of volunteer services, a senior youth counselor, three youth counselors, and two secretaries. Van Buren County (Paw Paw) The Van Buren County Volunteer Court Friends was funded beginning July 1, 1974. The project director was the Probate Judge and the implementing agency was the Van Buren County Probate Court. 212 The project was developed to divert youth from the juvenile justice system. A rural county, 17 percent of the Van Buren County population received some type of public assistance and 25 percent of county families had yearly incomes below $3,000. The Youth Service Bureau provided referral and supportive counseling services for predelinquent and delinquent youth, utilizing the assistance of volunteers. At least half of the status offenders referred to the court were to be referred to Volunteer Court Friends. The staff consisted of a project coordinator, counselor, field workers, and secretary. St. Clair County (Port Huron) The St. Clair County Youth Service Bureau began operation in January, 1975. The implementing agency was the St. Clair County Probate Court. The Youth Service Bureau was formed because of an absence of appropriate referral sources for predelinquent and delinquent youth and their families. (Approximately 80 percent of youth arrested were warned and released.) The project accepted referrals from police, court, and schools. Its objectives included a significant reduc- tion in juvenile arrests, school suspensions, and refer- rals to juvenile court. It acted as a central referral source to youth service agencies, had a county-wide youth 213 information system on youth apprehended by the police, and provided referral and counseling services for county youth. The staff included a director, assistant director, five youth counselors, and two secretaries. Newaygo County (White Cloud) The Newaygo County Youth Service Bureau began fund— ing on July 1, 1975. The implementing agency was the Newaygo County Probate Court. The Youth Service Bureau was formed in Newaygo County to provide needed alternative services to the pro- bate court for juvenile offenders and their families. It was the practice of the probate court to reject petitions relating to school problems or minor juvenile offenses. Objectives of the Youth Service Bureau included reducing juvenile arrests by 10 percent; school suspen- sions, expulsions, and drop-outs by 10 percent; and cases coming under juvenile court jurisdiction by 15 percent. The project intended to provide referral, screen- ing and counseling services, and to initiate needed new services. The Bureau concentrated on services to the family unit . The Bureau accepted referrals from police, school, cujurt, and parents. It was governed by a policy board with citizen and agency representation. 214 Staff included a director-counselor, one youth counselor, and a secretary. St. Joseph County (Three Rivers) The St. Joseph County Youth Service Bureau was funded beginning July 1, 1975. The implementing agency was the St. Joseph County Probate Court and the Court Director was named project director. The Bureau was developed to fill a gap in services, to provide dispositional alternatives for police, and to aid the schools in handling behavioral problems (since the juvenile court did not accept school petitions). The Project provided diversion services for youth and their families, offering counseling and making refer- rals to appropriate community agencies. Objectives included: (a) reducing recidivism among youth referred to the Youth Service Bureau by 5 percent, (b) reducing school suspensions and expulsions of youth referred to the Youth Service Bureau by 5 percent, and (c) reducing rate of petitioning to juvenile court of Youth Service Bureau participants by 10 percent. Project personnel included a director, three counselors, and a secretary. .3'1' 233.”; a I . “3‘ 1 215 Macomb County (East Detroit) The Youth Services Center began operation in June, 1971. The City of East Detroit was the implementing agency. During 1969 and 1970, the community perceived an increase in delinquent behavior and attributed it, in part, to increased drug usage. The Protective Services Commis- sion, established by the East Detroit City Council, took the initiative of planning the Youth Service Center. The program was aimed at prevention, rather than rehabilita- tion, and drug usage was a primary target. Originally the Center followed a crisis-intervention model. Goals of the Bureau have been expanded to include diverting youth from the juvenile justice system, strength- ening family life and parent-child relationships, involving tfluacommunityin.providing for the needs of youth, and help- ing youth experiencing behavioral problems at school or in the community. The Bureau provided individual and group counsel- ing for youth, referral and information services, crisis intervention, youth advocacy, family counseling, drug etucation, etc. It received referrals from court, police, jparents, and youth themselves. The project is currently funded by the City of Imast Detroit in cooperation with the East Detroit School system. It employs one director and one youth counselor. 216 Shiawassee County (Owosso) The Shiawassee County Youth Service Bureau began operation October 1, 1975. The Shiawassee County Probate Court was the implementing agency and an employee of the Juvenile Court was named project director. The Youth Service Bureau was developed to address the need for alternative resources for the large numbers of cases referred to juvenile court. Services for county delinquent youth were extremely limited. The project sought to reduce the number of court petitions, the sizecxf probation caseloads, and the rates of school suspensions and expulsions. Program and activities were aimed at providing casework and counseling services to status offenders, first offenders, and predelinquents. The program was oriented toward decentralized services and outreach in order to service outlying areas of this rural county. Bureau employees included a director, three case- workers, and a secretary. Oakland County (Pontiac) The Youth Service Bureau portion of the Oakland County Youth Assistant Program began January 1, 1974. The Director of Youth Assistance was project director. The Youth Assistance Program was well-established and extensive, employing 22 professional social workers and 217 utilizing over 1,000 volunteers. In 1972, 4,000 youth received casework services and 6,000 additional youth participated in Y.A.P.—sponsored activities. They saw a need for better integrated and coordinated services, child management training for parents, technical assistance to field workers via local committees, and a central intake process. Objectives included: preventing arrest and school suspension and expulsion of project youth, preventing project youth from coming under court jurisdiction, and identifying and coordinating public and private financial resources aimed at prevention and control of delinquency on the county or regional level. Activities were to establish a central intake process, to provide service within 48 hours of referral, to coordinate existing youth services, to establish a countywide advisory board, and to provide child manage- ment training classes for parents. The project staff included three field super- visors, one intake worker, one program development and community organization specialist, one child management training specialist, and two secretaries. 218 Grand Traverse County (Traverse City) The Grand Traverse Youth Service Bureau began April 1976. The implementing agency was the Grand Traverse County Probate Court. The project director was the Coordinator of Volunteer Services for the Grand Traverse County Probate Court. Services available to youth (e.g., school social workers) in the county were minimal. The Bureau, therefore, was to provide services to children 7-17 who were identi- fied as behavioral problems and to provide a springboard for community development. Goals aimed toward prevention of delinquency by early identification and immediate attention, better uti- lization of existing resources, relief for the overbur- dened court, and reduction of taxpayer costs (by using volunteers and reducing delinquency). Objectives were to reduce the number of official arrests and school suspensions or expulsions involving Bureau youth. Also listed were the reduction in number of institutional placements, the diversion of youth from the juvenile justice system, and the development of new resources. Activities indicated that the Bureau was to act as a service broker, as a catalyst toward improved agency cooperation, as a provider of services to all areas of the county, as a clearinghouse of police, and as an 219 "employer" of volunteers. Referrals were accepted from children, parents, and agencies. Staff included a director-coordinator, a senior counselor, two junior counselors, and one secretary-clerk. Alpena County (Alpena) The Alpena County Youth Service Bureau began June 1, 1974. The implementing agency was the Youth Service Bureau of Northeast Michigan and the director of same was named project director. Alpena County is a large rural area with few ser- vices provided to delinquent and predelinquent youth. A disproportionately high number of arrested youth (70-80 percent in county vs. statewide average of 43 percent) were referred to probate court. The purpose of the project was to determine exist- ing youth services, make a needs assessment of unmet ser- vice needs, and to develop a five-year comprehensive plan to coordinate and implement needed services in the Alpena County area. Project activities included the completion of the above five-year plan, and the development of diversion services for youth--referra1 to existing community agen- (gies for 90 percent of referred youth and provision of (direct short-term counseling services for the remaining 10 percent. 220 Project personnel included a director, three counselors, and a secretary. Kalamazoo County (Kalamazoo) The Kalamazoo Youth Service System was funded beginning July 1, 1974. Kalamazoo County Community Mental Health was the implementing agency. The Y.S.S. sought to reduce delinquency by effect- ing systems modification through intervention in the schools. At the outset, it provided no direct services to youth. The goals of the Youth Service System included the following: (1) increase communication between schools and youth-serving agencies, (2) provide information on ser— vices available to youth, (3) improve Coordination of ser- vices to youth, and (4) develop new and modify existing programs involving youth and/or agencies. Objectives were to prevent youth from dropping out of school and from being referred to juvenile court for school-related problems, to reduce the number of suspen— sions of target youth, to provide appropriate referrals, to identify needs, and to develop new programs. Activities included the development of an advisory committee to outline referral procedures, a community resource directory, educational resource teams for service delivery in the schools, inservice training for school and flu : raging-nuagi . 221 police personnel, and career development for potential drop-outs. The staff included a project director, program specialist, senior consultant, five outreach consultants, one executive secretary, and two clerk-typists. This program ceased operations after two years (1976) because of the unavailability of local funding. APPENDIX B DEMOGRAPHIC CHARACTERISTICS 222 APPENDIX B DEMOGRAPHIC CHARACTERISTICS Table B1.—-Popu1ation and median income for Youth Service Bureau and comparison counties. Counties Population Median Family Income Berrien 163,875 $6,145 Muskegon 157,426 6,048 Genesee 445,589 6,340 Saginaw 219,743 5,983 Calhoun 141,963 6,376 Jackson 143,274 6,421 Van Buren 56,173 5,196 Allegan 66,575 5,532 St. Clair 120,175 5,546 Lapeer 52,361 5,282 Newaygo 27,992 4,583 Mecosta 27,992 4,322 St. Joseph 47,392 5,626 Branch 37,906 5,449 Source: Michigan Statistical Abstract (East Lansing, Michigan: School of Business Administration, Michigan State University, 1975). 223 APPENDIX C ANNUAL U.C.R. DATA—COLLECTION FORM 224 APPENDIX C ANNUAL U.C.R. DATA-COLLECTION FORM Site Code Coder Year of Data Date of Coding ***************************Iklk Reported Burglary Reported Stolen Property Actual Burglary Actual Stolen Property Cleared Burglary Cleared Stolen Property Reported Vandalism Actual Vandalism Cleared Vandalism Reported Larceny Actual Larceny Cleared Larceny Reported Part 2 Actual Part 2 Cleared Part 2 Reported Part 1 Actual Part 1 Cleared Part 1 Delinquency Arrests Adult Arrests Burglary Larceny Part 1 Stolen Property Vandalism Part 2 Runaway Curfew Grand Total Burglary Arrests Referred to Juvenile Court Larceny Arrests Referred to Juvenile Court Part 1 Arrests Referred to Juvenile Court Stolen Property Referred to Juvenile Court Vandalism Referred to Juvenile Court Part 2 Referred to Juvenile Court Grand Total Referred to Juvenile Court 225 1.1—.59.: o I. A 1; ._,_.. APPENDIX D DELINQUENCY ORIENTATION SCALE 226 APPENDIX D DELINQUENCY ORIENTATION SCALE Date Position City ' Degrees Length of Employment at Youth Service Bureau The following statements represent a wide range of opinions regarding the causes and treatment of delinquency, as well as the role of the juvenile justice system. Please indicate the extent to which you agree or disagree with each by circling the appropriate number. The scale is as follows: Strongly agree Agree Partially agree Partially disagree Disagree Strongly disagree aim-puma o o o o o o l. A major advantage of the juvenile court is the ability to informally determine the best approach to rehabilitation .................. 1 2 3 4 5 6 2. With well-trained personnel and small caseloads the juvenile court can offer quality services and legal safeguards are unnecessary ........... l 2 3 4 5 6 3. The diversion of delinquents from the juvenile system is not likely to reduce delinquency because of the "soft" approach usually taken in these programs ....................... 1 2 3 4 5 6 4. Treatment for crimes other than the most serious is best carried out on a voluntary basis ....... T 2 3 4 5 6 5. The juvenile court is generally too lenient with delinquents ..................... T 2 3 4 5 6 6. Those individual emotional and psychological factors underlying crime are not generally under the control of the person .............. l 2 3 4 5 6 7. It is of primary importance that the juvenile court limit its activities to only the serious offenders ...................... l 2 3 4 5 6 227' 10. ll. 12. 13. 14. 15. 16. 17. 18. 19. 20. .228 Societal factors like racism and poverty are the critical variables underlying crime and delinquency. To search for the cause of crime is fruitless since everybody at times is criminal but only certain persons happen to come to the attention of officials .................... The most important cause of crime can be found in the person themself ............... Smaller caseloads and more intense individual therapy are the keys to reducing crime ....... The prevention of delinquency is best accomp- lished by providing economic and social programs for those groups involved in criminal activity . . . Juveniles are best served if they are diverted totally from the social service system and not officially handled by any agency .......... The most important causes of crime are to be found outside of the individual and not under their control .................... The most feasible way to prevent delinquency is the early identification of pre-delinquents and the provision of services to this group ..... The best way to reduce crime and delinquency is to ensure that the potential punishment always outweighs the benefits derived from committing a crime ................. The introduction of legal safeguards into the juvenile justice process is likely to hinder its effectiveness .................. T In order to reduce delinquency there must be changes made in the educational and social institutions which serve youth ............ l The juvenile court procedure should contain all the legal protections afforded by adults . . . . To prevent crime it is necessary to make it known that swift and sure punishment will result . . l l l l l l l l >U O‘U 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 225) Diverting youth from the juvenile justice system is important since it allows for the provision of services for younger and less serious offenders . It is important for the juvenile court to become more involved in the social and familial aspects >41) of delinquency .................. . 1 Understanding how laws are conceived, passed and enforced is more important than studying the causes of crime ................... An individual chooses by "free will“ to commit a crime ....................... The most beneficial approach to the crime and delinquency problem is to improve the quality and quantity of counseling and casework services . . The diversion concept is most important because it allows a youth to avoid the stigma of formal processing and still offers an avenue for deliver- ing services .................... The best way to prevent crime and delinquency is to bring about broad changes in the economic structures of society so as to lessen the discrim- ination of inequality ................ Where treatment is required it is necessary to clearly define expectations and specify the length of time to be involved ............ In treating delinquency it is most important to develop a broad range of coordinated programs in the community above and beyond counseling ...... Stricter and longer sentences would go a long way toward reducing criminal behavior ........ The major cause of crime and delinquency can usually be traced to emotional and psychological factors ....................... Juvenile courts have gone too far in their attempts to help juveniles ............. l l l l l l l l l APPENDIX E PROGRAM PERCEPTIONS SURVEY 230 APPENDIX E PROGRAM PERCEPTIONS SURVEY Instructions This questionnaire is designed to get your impres- sions about the Youth Service Bureau. The questions are not designed to find out if the Bureau if "good" or "bad," but rather are focused on what kind of a program it is, what kinds of things go on in the program, what it's like working at the Bureau, what it's like working with the kids, and so on. The questionnaire includes 86 statements in "true- false" format. If the statement is characteristic of the Youth Service Bureau, you should circle the "T." If the statement is not characteristic of the Bureau, you should circle the ”F." The questions cover a wide variety of areas including the kids, the program, kinds of services, etc. Each question should be read carefully before responding. Remember the point is not to make your pro- gram look good or bad since there are no right or wrong answers. Rather please respond as accurately and hon— estly as possible. It is very important to represent your program as it really is. 231 232 Youth Service Bureau Environment Staff Form 10. ll. 12. 13. 14. 16. 17. 18. The kids are proud of this program. Staff have very little time to encourage the kids. The youth are encouraged to show their feel- ings. The staff act on the kids' suggestions. There is very little emphasis on making plans for getting out of the program. The clients are expected to share their per- sonal problems with the staff. The staff make sure that the YSB is always neat. Staff sometimes argue with each other. Once an appointment schedule is arranged for a client he/she must follow it. The youth we get in the YSB really try to improve and get better. The staff are interested in following up the kids once they terminate. Our clients tend to hide their feelings. The kids are expected to take initiative in this program. The kids are encouraged to plan for the future. The kids rarely talk about their personal problems. The offices are often messy. If the staff's approach to a client is changed the staff always tells him/her why. The kids may criticize staff members to their face. “:57“ g T" d y: 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 233 The kids in this program care about each other. The staff help new kids get acquainted with the YSB and its approach. The staff and clients say how they feel about each other. The staff give kids very little responsibil- ity for their improvement. The clients are encouraged to learn new ways of doing things. Personal problems are openly talked about. The conference room usually looks a little messy. When kids first come to the YSB someone explains how the YSB operates. The kids will be terminated from this program if they don't obey the rules. There is very little group spirit in this program. The more mature kids in this program often work with the younger kids. People say what they really think around here. The clients have a say about what goes on here. There is very little emphasis on what the kids will be doing after they terminate with the YSB. Discussions in this program emphasize under- standing personal problems. This is a very well organized program. Staff are always changing their minds here. All decisions about the program are made by the staff and not by the kids. The kids put a lot of energy into what they do in the YSB. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 56. 234 The kids rarely help each other. The kids say anything they want to say to the staff. The staff discourages criticism from the kids. Staff care more about how the kids feel than about their day-to-day problems. Staff are mainly interested in learning about the kids' feelings. Things are sometimes very disorganized around here. Staff tell the kids when they're doing well. The staff very rarely punishes kids by detaining them. The program has very few social activities for the kids. Staff go out of their way to help the kids. The kids are careful about what they say when the staff are around. Staff encourage the clients to initiate their own activities. This YSB emphasizes training for new kinds of jobs. The clients are rarely asked personal ques- tions by the staff. Many of the kids look messy. If a youth breaks a rule of the YSB he knows what will happen to him/her. Staff don't order the kids around. Very few things around here ever get people excited. Staff are involved in the youth's activities in the community. F 57 F 58 F 59. F 60. F 61. F 62 F 63 F 64 F 65 F 66 F 67 F 68 F 69 F 70. F 71. F 72 F 73 F 74 F 75 235 When the kids disagree with the staff they keep it to themselves. Staff rarely give in to client pressure. The kids in this program are expected to work toward their goals. The staff discourage talking about sex roles. Sessions with the kids are carefully planned. The kids are always changing their minds about what they want. If a client argues he/she will get into trouble with the staff. Discussions are pretty interesting in this program. Counselors have very little time to encour- age clients. It is hard to tell how the kids are feeling in this program. The kids in this program are encouraged to be independent. The new treatment approaches are often tried in this program. Staff try to help the kids understand them- selves. The staff sometimes miss their appointments with clients. The kids never know when a counselor will ask to see them. The staff regularly check up on each youth. The youth don't do anything for themselves unless the staff ask them to. Staff encourage group activities among the youth. In this program staff think it is a healthy thing to argue. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 236 There is no client input to this program. The kids must make special plans before ter- minating with the program. The clients hardly ever discuss their sexual lives. The staff set an example for neatness and orderliness. The clients never know when they will be terminated. The clients can call staff by their first names. This is a friendly program. The staff knows what the kids need. There is very little emphasis on making the kids more practical. The kids are rarely kept waiting when they have appointments with the staff. The kids know when counselors will want to see them. APPENDIX F ENVIRONMENTAL ASSESSMENT INTERVIEW GUIDE 237 8:8o8m oofiraom :80» 8 .8 :0: 8:5 800 o:: .8.“ 30: 98:: m: 9.888 A>pfl8v 088 :02: 30: . 8:8o8m oofifiom :80» may :0 mmooosm on: 88:88:88 no empapfiafloa: 98: 0: 8on8 30: :08 8.883 ofi. o8 8:3 . 85.885 8 .8m coo: 855800 o:p 98098 08 com: m08m8o888:0 >88 93 o.8 8:3 . 8:883: oofifiom :80» 8 m0 8o3 o:w omomao .8 8858 0:3 mo8:ow8 :8 $8532.68 o:p :2.” own—8:0 >8 :oo: 893 m8: . A880: s:a\8saav 8:88:58 mofl>pom specs on: yo mofipfl>flpoa U8 w:H:8HQ 8:898 o:p :H oo>8>5 ow 0:3 . 88.8mm o0Eom. :80w 8 mo 82% o:p 80:98 39:88: 0:3 8:088NH8who n8 magfifiofi o:p o8 0:3 . 8:088”: :8o8m oofirmom :psow 80m gm? .w HO GOfiufimho 063. .HOH 98:: 83 98:98 apfi8v 088 :02: 30: . wood/88.30 :8 8808 8on8: .8 p:o:o>ofl:08 o5 c830: moofiwonm 88:8 .8 389888.44 8:803 88880:: so» was» A.opo .spflflaso :88:8 ”8:088:00 08:00o .8808 #808303 8.883 9:. was 8:3 883938 was Us.” med. 23 3. . 8A .08 .ooofifiom mo :03 $088858: 3895.3 mm» 8 now poo: o:p 80:98 0: wow: moflpmfimo888:0 >38 o:p ohoa 8:3 . 88o8m ooflPHom :88» 8 now 8o“: 8:880 o:p oomoaqo .8 898998 Aoo8:o.w8 U8 m8sew>8fiv 0:3 . «mm» on» no: wafleqafi: Haaflmflno on: 8: 839:5 W83 Amo8:ow8 6:8 magvfifiofiv 0:3 . wflsmhooamv o8fiw80 mm» o:p .8m 88:: 8:880 9: 88 $30888 8:8 £888,830 .883 fig . 8.888.: :8o8m o0H>pom :80» ohm snowmfim mewmmbm MOH>mmm mBDON MQHDO BmH>mmezH Ezmsmmmmm¢ Azm h NHQmem< 238 239 wm58o8m o0:>.8m :80» o:: »: 88:88 8:08 85:80:00 o:: :: 8.88:0 »8 :oo: o.8:: 98:: 8588 o0:>8m :80» o:: :0 mo: :8: :08 :8 8808 o:: oN: ::.8:.:9 88888 .8::0 0: 30: 8:09:89 o:: :0 o::: o:: .8>0 888:0 880w 80:85 0: 88:88 mo:::.8:.:9 o:: 98: 888 o0:>.8m. :80» o:: :0 mo>:8o.n:0 :8 m80w ::omo.:9 o:: :::3 8.88 8888 8:8 0: 88o: 8:3 0:. 888m o0:>.8m :80» o:: :0 mo>: 8o .30 :8 8808 .v: .m: .N: .H: :o::o::: o:: :: 898:0 >8 :ooo. o.8:: o>8m . o: 88:888 o.83 85:88.0 :8 m80w omo:: 883 :095 708 $085888: :0 8.8580 o:: w:::.80 I:00 w:0: 8:583 8:08 85:98:00 o:: 83 8:3 88>: :8 Bo :8 880m omo:: oN:::.8:.:9 mo:0:ow8 .8:8 :8 30: 8.88:9 8:880 o:: 8. mo>: 8o .30 :8 888 80:85 o:: 0: 88:88 o83 mo:::.8:.:9 8:3 . 88588 $8: :83 888 mo8:o.w8 .8::0 :8 8: 8:3 0:. Em: 9: .8: $8: 855 mo>: 8o .30 :8 m8ow :o::o::: o:: o83 8:3 moi :8 .30 8:8 888 858o8m o0:>.8m :80» o:: :0 8:88: 98:5: o:: .8: m:0o9mo.:9 ::owo.:9 o:: o8 8:3 . 8:0: 8.89000 .8: o::8o88 o:: o8: 30: 5898:: o0:>.8m :80» 8:: moon 8888 .880 A38 :3 :23 fig . 8:89: 50» :803 888:0 8:3 . :800 50» :: 8885.88 888 F838: 8:88 8:9: as: a: :88 8:58 8:0: 8:: :0 :0::800: o:: :: wow8:0 »:8 :oo: o.8:: o>8m . $888.: 58o8m o0:>.8m :80» 809 888:8 was 88:8 8 9: :88 89:89 mm» o:: :0 w::::5: o88: .8: m8o9 180:9 o:: o.83 8:3 .:o:m:::8:mo .83 :: 8:3 . 88: 8.89000 98:5: .8: 388888 o>8: mm» o:: :8 20:88:88 .880 :8 :3 8:, 88s . 830:8 8: m8 88::88o .888 o0:>.8w 3:8 .888: 8:89 . 803 8888 888 >858: 88.8: 8:9» on: a: 8:88: 8 8 mm: on: was 82: . W888: 588 o0:>.8m :80» o.:9 .v: .H: .0: 240 @339» .5393. suds damn mwfioqmmd 9930 £033 5 SEE: may no Um: 5.8.55 moflfimm £50» 93 mo monmpmfixw map mm: #895 £055 30m . wwfimmmooam #3033. 55.3% Cu m3 Humanoid ad mm msawhfi moifimm 5:0.» wfimd 939;: 9.58 mi: learn. ”Bond 33 3 ”Emmmpm mmfloqmwm .828 06 30m . wmmUSHHpud mmmnw 2H mmwcaso ”283353 had :89 393 9a: wgoo mi: Im>s.n map 80.5 2959» 8930.5 wfifimfic mo .36.“ map #509.» 33 zfipcmmwhg wmwocmmd .895 06 so: wmmvdfippm mmmfi 3 mqumso ”5833me has 88, 8.93 02mm «30: 550% 8E8 530% a go «m3 93 ”Bond H03 mmflocmwm .850 On 30: ~33 98 fig pmmn 338 >9: 8 888mm 633 wmmflonmwm umnpo ow 8.38 Edam 5593mm 8?me £8» 93 mo Egg 8 :3 son 98 8:5 Bow .ON .OH M: 3959» .5389 “6352 mmflocmwm .830 £033 5 .8588 map 885:3 3:03 mm» a mo 88me 05 9898 :9» EU :02: 30m wmfimmmoonm H3035“. 3 m5 .553? Sm. mm mm» 93 wfim: 9:35 9.58 wflqwkfin mo 83 9: 20% x55 868% uofio v.8 ”Ens . 3:500 magma—n. map 88% 39:9» Samoan wfitQfiv mo meH 9: 38a xfifi 888w“ 8:8 Eu fins E a we .83 on» Egon mmfiocmwm .828 no mcsfippm map was was; panama p8 .33 wfi 5:3 838%.. .850 Cu Swamp 3 Edam m”: was How .nonm mmw map #898 50% 36 £058 30m mcow ”58me ~43ng mmmfipfi>flpom :dmnsm wofi>hmm manor ow mmflocmwm Hmnpo hp cm>ww whoaqsm may ca mwawno >qd :mmn mnmnp mam ~3de mofirnmm 530% 93 mo Amfihmmfi magma mficfiofiv m2 #330.“ mfi 5 9533 an mwfloqmwm H930 CU mmpwmv p.33 09 38.899 m3 wfipquwHQ—fi 2H 58% 8Emm 38> 93 E 83:”? Amflmflfipm c3 pamfiflfiv 83333 93 :H mmwSEO EB 58, 985 gum 23me 38.5mm moflfimm €50» awom wuss wmsmmhsm moflfimm A350» new 833mm 83.3305 mg .5 fits 8% mmfiocmwm .8590 .36 $836 333 09 wmmflpgwpnud m8~3 mo nofiomamm map 3 . ma mfiwmflow aha mmfioammd .585 63 gamma p.33 09 38 no.3 samba mofifimm £50» 93 mo cofludpcmcwamfi H5505 93 now 6303mm. mums .2 Amflwmpafim aofiamfimpfiv 83H>38 323 m3 ufifi 90¢ @8ng 523m moEmm 5:0» 95 .NN .ON .mH .wH .NH .QH .mH APPENDIX G INSTRUMENT MODIFICATIONS ON THE PROGRAM PERCEPTIONS SURVEY 241 APPENDIX G INSTRUMENT MODIFICATIONS ON THE PROGRAM Scale Involvement Support Expressiveness Autonomy Practical Orientation Personal Problem Orientation Order/Organization Clarity Staff Control PERCEPTIONS SURVEY Items Deleted on the Basis of Frequencies Items Deleted on the Basis of Algha Levels 82 02,47 03 22 14,23,84 33,69 79,85 44 54,63 242 10,64 29,83 12,30,57 13,40,49 68 60 43,70 08 18,27 B I BL I OGRAPHY 243 BIBLIOGRAPHY Allport, Gordon. Becoming. New Haven, Connecticut: Yale University Press, 1955. Alwin, Duane F., and Sullivan, Michael J. "Issues of Design and Analysis in Evaluation Research." In Validity Issues in Evaluation Research. Edited by Ilene N. Bernstein. Beverly Hills, California: Sage Publications, 1976. Aronson, Sidney H., and Sherwood, Clarence C. "Researcher Versus Practitioner: Problems in Social Action Research." Social Work 12 (1967): 89-96. Baron, Roger et al. "Preventing Delinquency Through Diversion: The Sacramento County 601 Diversion Project.” Federal Probation 37 (March 1973): 13-19. Becker, Howard S. Outsiders. New York: McGraw—Hill, 1963. Bennis, Warren G.; Benne, Kenneth D.; and Chin, Robert, eds. The Planning*of Change. New York: Holt, Rinehart and Winston, 1962. Bertalanffy, Ludwig von. General System Theory. New York: George Braziller, 1968. Bernstein, Ilene N., and Freeman, Howard E. Academic and Entrepreneurial Research. New York: Russell Sage, 1975. Black, D. J. "Production of Crime Rates." American Sociological Review 39 (1970): 733-48. Blumer, Herbert. "Social Problems as Collective Behavior." Social Problems 18 (1971): 298-306. "The Sociological Implications of the Thought of George Herbert Mead." American Sociological Review 71 (March 1966): 535-47. Boorkman, David. Community-Based Corrections iniDes Moines. Washington, D.C.: L.E.A.A., 1976. 244 245 Box, George E. P., and Tiao, George C. "A Change in the Level of Non-Stationary Time—Series." Biometrics 52 (1965): 181-92. Buckley, Walter. Sociology and Modern Systems Theogy. Englewood Cliffs, N.J.: Prentice—Hall, 1967. Burns, T., and Stalker, G. M. The Management of Inno— vation. London: Tavistok Institute, 1961. Cain, Glen G., and Watts, David. "The Method— ology of Evaluating Social Programs." In Evaluating Social Programs. Edited by Peter H. Rossi and Walter Williams. New York: Seminar Press, 1972. Cameron, Mary. The Booster and the Switch. New York: Free Press, 1964. Campbell, Donald T. "Reforms as Experiments." American Psychologist 24 (April 1969): 409-29. , and Erlebacher, Albert. "How Regression Artifacts in Quasi-Experimental Evaluations Can Mistakenly Make Compensatory Education Look Harmful." In Handbook of Evaluation Research, Vol. 1. Edited by Marcia Guttenberg and Elmer L. Struening. Beverly Hills, California: Sage Publications, 1975. Campbell, Donald T., and Ross, H. Laurence. "The Connecticut Crackdown on Speeding: Time-Series Data in Quasi-Experimental Analysis." Law and Society Review 3 (August 1968): 33-53. Campbell, Donald T., and Stanley, Julian C. Experimental and Quasi-Experimental Designs for Research. Chicago: Rand McNally, 1963. Caro, Francis, ed. Readings in Evaluative Research. New York: Russell Sage, 1971. Carter, Robert, and Gilbert, John. "Alternate Routes: An Evaluation." Sacramento: California Youth Authority, 1973. Cavan, Ruth Shonle, ed. Readings in Juvenile Delinquency. 2nd ed. Philadelphia: J. B. Lippincott, 1969. Chambliss, William J., and Liell, John T. "The Legal Process in the Community Setting." Crime and Delinquency 12 (October 1966): 310—17. 246 Chambliss, William J., and Nagawasa, Richard H. "On the Validity of Official Statistics.” Journal of Research in Crime and Delinquency 6 (1969): 71-77. Cloward, Richard A., and Ohlin, Lloyd E. Delinquency and Opportunity: A Theory of Delinquent Groups. New York: Free Press, 1970. Coates, Robert B., and Miller, Alden D. "Evaluating Large Scale Social Service Systems in Changing Environments." Journal of Research in Crime and Delinquency 12 (1975): 92-106. Cohen, Albert H. Delinquent Boys: The Culture of the Gang. New York: Free Press, 1955. Cressey, Donald R., and McDermott, Robert A. Diversion From the Juvenile Justice System. Ann Arbor, Michigan: National Assessment of Juvenile Corrections, 1973. Cyert, Richard M., and March, James G. A Behavioral Theory of the Firm. Englewood Cliffs, N.J.: Prentice-Hall, 1963. Davidson, William S.; Rappaport, Julian; Seidman, Edward; Berck, Philip L.; and Herring, Jacob. "The Diversion of Juvenile Delinquents: An Experi- mental Examination." (Mimeographed.) Davis, James A. "On the Remarkable Absence of Non- academic Implications in Academic Research." In Social Policy and Sociology. Edited by N. J. Demerath, Otto Larsen, and Karl F. Schuessler. New York: Academic Press, 1975. Davis, Kenneth C. Discretionary Justice: A Preliminary Inquiry. Baton Rouge, La.: Louisiana Univer— sity Press, 1969. Durkheim, Emile. The Rules of Sociological Method. New York: The Free Press of Glencoe, 1938. . Suicide. New York: Free Press, 1951. Duxbury, Elaine. Evaluation of Youth Service Bureaus. Sacramento, Calif.: Department of the Youth Authority, 1973. 247 Elliot, Delbert S. National Evaluation of Youth Service Systems. Boulder, Colorado: Behavioral Research and Evaluation Corporation, 1974. , and Blanchard, Fletcher. "An Impact Study of Two Diversion Projects." (Mimeographed.) Empey, LaMar T. A Model for the Evaluation of Programs in Juvenile Justice. Washington: Law Enforce- ment Assistance Administration, 1977. Erikson, Kai T. "Notes on the Sociology of Deviance." Social Problems 9 (Spring 1962): 307—14. Etzioni, Amitai, and Etzioni, Eva. Social Change. New York: Basic Books, 1964. Fairweather, George W. Methods for Egperimental Social Innovation. New York: John Wiley, 1967. . Social Psychology in Treating Mental Illness. New York: John Wiley and Sons, 1964. Fishman, Robert. An Evaluation of the Effect on Criminal Recidivism of New York City Projects Providing Rehabilitation and Diversion. New York: New York Criminal Justice Coordinating Council, 1975. Freeman, Howard E., and Sherwood, Clarence C. "Research in Large-Scale Intervention Programs." Journal of Social Issues 21 (1965): 11-28. . Social Research and Social Policy. Englewood Cliffs, N.J.: Prentice-Hall, 1970. Garfinkel, Harold. "Conditions of Successful Degradation Ceremonies." American Journal of Sociology 61 (March 1956): 42-44. Gemignani, IR. J} ”Youth Service Systems: Diverting Youth From the Juvenile Justice System." Delinquency Prevention Reporter 8 (1972): 1-12. Gibbons, Don C., and Blake, Gerald F. "Evaluating the Impact of Juvenile Diversion Programs." Crime and Delinquency 22 (October 1976): 411-20. Gibbs, Jack P. ”Conceptions of Deviant Behavior: The Old and the New." Pacific Sociological Review 9 (1966): 9-14. 248 Glass, Gene V. "Analysis of Data on the Connecticut Speeding Crackdown as a Time-Series Quasi- Experiment." Law and Society Review 3 (August 1968): 55-76. ; Willson, Victor L.; and Gottman, John M. Design and Analysis of Time—Series Experiments. Boulder, Colorado: University of Colorado, 1972. Goffman, Erving. Asylums. New York: Doubleday, 1961. Gouldner, A. "Theoretical Requirements of the Applied Social Sciences.” American Sociological Review 22 (1957): 52-102. "Explorations in Applied Social Science." In Applied Sociology. Edited by Alvin W. Gouldner and S. M. Miller. New York: The Free Press, 1965. Greiner, Larry E. "Antecedents of Planned Organizational Change." Journal of Applied Behavioral Sciences 3 (1967): 51-85. Gross, Neal et a1. Implementing Organizational Innova- tions. New York: Basic Books, 1972. Hagen, John L. "The Labeling Perspective, the Delin- quent and the Police: A Review of the Literature." Canadian Journal of Criminology and Corrections 14 (1972): 150—65. Harlow, Eleanor. Diversion From the Criminal Justice System. Rockville, Md.: N.I.M.H., 1973. Havelock, Ronald G. Planning for Innovation. Ann Arbor, Michigan: Institute for Social Research, 1971. Holland, Winford E., and Huntoon, Harrison H. "The Evaluation of Experimental Social Service Delivery Systems: An Organizational Effective- ness View." Community Mental Health Journal 10 (1974): 41—51. Hollingshead, A. B., and Redlick, R. C. Socigl Class and Mental Illness. New York: John Wiley, 1958. Howlett, F. W. ”Is the USB All It's Cracked Up to Be." Crime and Delinquency 19 (1973): 485-92. 249 Hutcheson, Bellenden R., and Krause, Elliot A. "Systems Analysis and Mental Health Services." Community Mental Health Journal 5 (1969): 29-45. Katz, Daniel, and Kahn, Robert L. The Social Psychology of Organizations. New York: John Wiley, 1966. Kerlinger, Fred N. Foundations of Behavioral Research. 2nd ed. New York: Holt, Rinehart and Winston, 1973. Kitsouse, John L. "Societal Reaction to Deviant Beha- vior." Social Problems 9 (Spring 1962): 247-56. Klamputs, Nora. Diversion From the Justice System. Hackensack, N.J.: N.C.C.D., 1974. Klein, Malcolm W. "Issues and Realities in Police Diversion Programs." Crime and Delinquency 22 (October 1976): 421-27. . "Alternative Dispositions for Juvenile Offenders: An Assessment of the . . . Juvenile Referral and Resource Development Program." "Mimeographed.) Kobrin, Solomon, and Lubeck, Steven G. "Problems in the Evaluation of Crime Control Policy." In Public Policy Evaluation, Vol. 2. Edited by Kenneth M. Dolbear. Beverly Hills, California: Sage Publications, 1975. Kuhn, T. S. The Structure of Scientific Revolutions. Chicago: University of Chicago Press, 1962. Kvaraceus, William C., and Miller, Walter B. Delinquent Behavior, Culture and the Individual. Washington: National Educational Association, 1959. La Piere, Richard T. Social Change. New York: McGraw- Hill, 1965. Leiber, Leon. Project Crossroads: A Final chort to the Manpower Administration. Washington, D.C.: Department of Labor, 1971. Lemert, Edwin M. Instead of Court: Diversion in Juvenile Justice. Chevy Chase, Md.: N.I.M.N., 1973. . Social Pathology. New York: McGraw—Hill, 1951. Liedtke, K., et a1. Portland Youth Diversion Project. Portland, Maine: Office of Youth Diversion Services, 1974. Lincoln, Suzanne B. "Juvenile Referral and Recidivism." (Mimeographed.) Lippitt et a1. Dynamics of Planned Change. New York: Harcourt, Brace and World, 1958. Lofland, John. Deviance and Identity. Englewood Cliffs, N.J.: Prentice-Hall, 1969. Matza, David. Becoming Deviant. Englewood Cliffs, N.J.: Prentice-Hall, 1969. . Delinquency and Drift. New York: John Wiley, 1964. Merton, Robert. Social Theory and Social Structure. New York: Free Press, 1957. Michigan Comprehensive Law Enforcement and Criminal Justice Plan--1976. Lansing, Michigan: Office of Criminal Justice Programs, 1976. Michigan Statistical Abstract--1975. 10th ed. East Lansing, Michigan: Michigan State University, 1975. Michigan Uniform Crime Report--1975. Lansing, Michigan: Department of State Police, 1976. Moore, Eugene A. "Youth Service Bureaus." Journal of the American Judicature Society 52 (June-July, 1968-69). Moos, Rudolf H. Evaluating Correctional and Community Settings. New York: John Wiley, 1975. National Advisory Commission on Criminal Justice Standards and Goals. Community Crime Prevention. Washington, D.C.: U.S. Government Printing Office, 1973. National Institute of Mental Health. Criminal Statis- tics. Washington, D.C.: Department of Health, Education and Welfare, 1973. 251 Nejelski, Paul. "Diversion of Juvenile Offenders in the Criminal Justice System." In New Approaches to the Diversion and Treatment of Juvenile Offenders. Washington, D.C.: L.E.A.A., 1973. . "Diversion: The Promise and the Danger." Crime and Delinquency 22 (October 1976): 393-410. Nie, Norman H. et a1. Statistical Package for the Social Sciences. 2nd ed. New York: McGraw-Hill, 1975. Norman, Sherwood. The Youth Service Bureau. Paramus, N.J.: N.C.C.D., 1972. 0gburn, William F. Social Change. 2nd ed. New York: Viking Press, 1950. ‘ Parsons, Talcott. The Social System. Glencoe, 111.: The Free Press, 1951. . The Structure of Social Action. Glencoe, Ill.: The Free Press, 1949. Perrow, Charles. "Hospitals: Technology, Goals, and Structure." In Handbook of Organizations. Edited by James G. March. Chicago: Rand McNally, 1965. Piliavin, Irving, and Briar, Scott. "Police Encounters With Juveniles." American Journal of Sociology 70 (September 1964): 206-14. Poveda, Tony G., and Schaffer, Edward. "Positivism and Interactionism: Two Traditions of Research and Criminology." In Criminal Justice Research. Edited by Emilio Viano. Lexington, Mass.: D. C. Heath, 1975. President's Commission on Law Enforcement and the Admin- istration of Justice. Task Force chort: Juvenile Delinquency and Youth Crime. Washington, D.C.: U.S. Government Printing Office, 1967. Pressman, Jeffrey L., and Wildovsky, Aaron. Implemen- tation. Berkeley, Calif.: University of California Press, 1973. Price, J. E. ”A Test of the Accuracy of Crime Statis- tics." Social Problems 14 (1966): 214-22. 252 Quinney, Richard. Critique of Legal Order. Boston: Little, Brown and Company, 1974. . The Social Reality of Crime. Boston: Little, Brown and Company, 1970. Reckless, Walter C. The Crime Problem. New York: Appleton—Century-Crofts, 1961. Riecken, Henry W., and Boruch, Robert F., eds. Social Experimentation. New York: Academic Press, 1974. Rivlin, Alice M. Systematic Thinking for Social Action. Washington, D.C.: The Brookings Institution, 1971. Rosenheim, Margaret K. ”Youth Service Bureaus: A Concept in Search of Definition." Juvenile Court Judges' Journal 20 (Summer 1969): 69-74. Ross, H. Lawrence; Campbell, Donald T.; and Glass, Gene V. "Determining the Social Effects of a Legal Reform.” American Behavioral Scientist 13 (1970): 493-509. Rossi, Peter H. ”Testing for Success and Failure in Social Action." In Evaluating Social Programs. Edited by Peter H. Rossi and Walter Williams. New York: Seminar Press, 1972. Rutherford, Andrew, and McDermott, Robert. Juvenile Diversion. Washington, D.C.: L.E.A.A., 1976. Scheff, T. J., and Culver, D. M. "The Societal Reaction to Deviance.” Social Problems 11 (Spring 1974): 401-13. Schucter, Arnold, and Polk, Kenneth. Phase I Assessment of Youth Service Bureaus. Washington, D.C.: L.E.A.A., 1975. Schur, Edwin M. Crimes Without Victims. Englewood Cliffs, N.J.: Prentice-Hall, 1965. . Radical Non-Intervention: Rethinking the Delinquency Problem. Englewood Cliffs, N.J.: Prentice-Hall, 1973. Scriven, Michael. "The Methodology of Evaluation." In Perspectives on Curriculum Evaluation. Edited by Ralph W. Tyler, Robert M. Gagne and Michael Scriven. Chicago: Rand McNally, 1967. 253 Sellin, Thorsten, and Wolfgang. The Measurement of Delinquency. New York: John Wiley and Sons, 1964. Selznick, Philip. TVA and the Grass Roots. Berkeley, Calif.: University of California Press, 1949. Shaw, Clifford, and McKay, Henry. Juvenile Delinquency and Urban Areas. Chicago: University of Chicago Press, 1942. Short, James F. "The Natural History of an Applied Theory: Differential Opportunity and 'Mobili- zation for Youth.'" In Social Policy and Sociology. Edited by N. J. Demerath, Otto Larsen and Karl F. Schuessler. New York: Academic Press, 1975. Sjoberg, Gideon. ”Politics, Ethics and Evaluation Research." In Handbook of Evaluation Research, Vol. 2. Edited by Marcia Guttenberg and Elmer L. Struening. Beverly Hills, Calif.: Sage Publi— cations, 1976. Skogan, W. G. "The Validity of Official Crime Statis- tics: An Empirical Investigation." Social Science Quarterly 55 (1974): 25-38. Skolnick, Jerome S. Justice Without Trial. New York: John Wiley and Sons, 1966. Smith, Robert L. "Diversion: New Labe1--Old Practice.” In New Approaches to the Diversion and Treatment of Juvenile Offenders. Washington, D.C.: L.E.A.A., 1973. Smith. T. G. ”Crime Statistics—-Can They Be Trusted." American Criminal Law Review 11 (1973): 1045—86. Stinchcombe, Arthur. "Social Structure and Environment." In Handbook of Organizations. Edited by James G. March. Chicago: Rand McNally, 1965. Suchman, Edward A. Evaluative Research. New York: Russell Sage, 1967. Sudnow, David. "Normal Crimes: Sociological Features of the Penal Code in a Public Defender's Office." Social Problems 12 (Winter 1965): 255—76. 254 Sutherland, Edwin H. Principles of Criminology. New York: J. B. Lippincott, 1947. , and Cressey, Donald. Criminology. New York: J. B. Lippincott, 1960. Thompson, James D. Organizations in Action. New York: McGraw-Hill, 1967. , and McEwan, William J. "Organizational Goals and Environment: Goal Setting as an Interaction Process." In Complex Organizations and Their Environments. Edited by Merlin B. Brinkerhoff and Phillip R. Kunz. Dubuque, Iowa: William C. Brown, 1972. Traub, Stuart H., and Little, Craig B., eds. Theories of Deviance. Itasca, 111.: F. E. Peacock, 1975. Tripodi, T., et a1. Social Program Evaluation. Itasca, 111.: F. E. Peacock, 1971. Trojanowicz, Robert C. Juvenile Delinquency: Concepts and Control. Englewood Cliffs, N.J.: Prentice- Hall, 1973. U.S. Department of Health, Education and Welfare. ‘de Challenge of Youth Service Bureaus. Washington, D.C.: Department of Health, Education and Welfare, 1973. U.S. Uniform Crime Reports. Washington, D.C.: Department of Justice, 1971. Watson, G., ed. Concepts for Social Change. Washington, D.C.: Institute for Applied Behavioral Sciences, 1967. Weber, Max. Max Weber: Escays in Sociology. Edited by Hans Gerth and C. Wright Mills. New York: Oxford University Press, 1946. Weiss, Carol H. Evaluation Research. Englewood Cliffs, N.J.: Prentice-Hall, 1972. Weiss, Robert S., and Rein, Martin. "The Evaluation of Broad—Aim Programs: A Cautionary Case and a Moral.” Annals of the American Academy of Political and Social Science 385 (1969). 255 Wheeler, Stanton. "Criminal Statistics: A Reformulation of the Problem." Journal of Criminal Law, Criminology and Police Science 58 (1967): 317-24. Wholey, Joseph et a1. Federal Evaluation Policy. Washington, D.C.: The Urban Institute, 1970. Williams, Walter. Social Policy Research and Analysis. New York: American Elsevier, 1971. , and Elmore, Richard F. Social Program Imple- mentation. New York: Academic Press, 1976. HICH G m111111111711 3