1* . III. ; I. rm 1 v1, 1 1 1111111111311111112111{1111111111111 111.1 WI 111'11 1111111111111 11111 1111111“ "0‘ 1 II 1. ' 11111111111 11111111 11111 11111111111 1111 g Jab—— V— , “—1-:— £23;— fzf‘i 3 - ._ é"...— :__~:__'—‘__ 11l 111111111111 11‘ —~_ ”32* mm?“ _-—-—... t .0 - —-’ ’- __-————_.__ ;~‘.-.— “-2—- 111 1111111111111111111 11111111 1111111 11111 1 11111111'111 1' 11111 111 1111111111111 111111111 111111111111111111 . I I 1111111111 '1‘ 111111111111111111 1' 1 . _ ‘ 111111 111 111,111,111 13 1 . . ‘111'1'1111111111‘11111111111111111 1 1 ,111' '11 '1'1'1111,1'1.11;11"' 111111111 1111111 . - b ‘4 -— - . gW‘M-m’ A ~ . .- _ mToqr—I' .— M1“‘ _ 1.1'1 111. 111111111111 . ; 11111111 I 11 1 4' 111111111 .1111111'1111111111‘11 11111111111 111111111 111111111 "1111.111131-1‘1‘1 1111(1 11111111 11111131, 111111111131 ’ 1141111111111 11' II ' 111 . ,11“ 1.1111111 1111111111 11111 1 111111115!“ 111111111 11111111 I11 1111 111111 11111111111111111111111111111111 .' .. 1‘11 111111111111 11 111111,,‘1111111 1:11 . 1111 111111.11... 1'1 1.11;. 1111111111 1111111111 "111 1' 111 1‘1: '11 11 .,;;.~:1;.:= ; 111:1 1.111111 1 “ 11111 11111111111. ‘11111 .111 w— :b— “gm“. ’— ' ~.,., -E‘: :23: ..- _:——c~—:}:._L:‘:. ‘;1 _1 l _ ' m —... ‘— LT. ' —' a. . _<‘ - __:M‘: OCT—A - \ _ a «-— w‘s‘d.“ 1 11 v)"-— 111 1 3.3' 1.. 11 1111111111) 1111 1311111111 11111111 1111111111 11111111 ’ 1 11" 1 1111 111 1 1111111111 ~11 111111111 1111 1111 11111,, 1"1 111111111111 11 '1'1111‘1'11111'I 11,1 1!" '1 11111111111111111111111'1111111111 111111111 1 I11 1111111111111 ‘11 1' . 11111111111 1111111111 11111111111111 11:1 I 1 11... :, 11111111 111111 11 1.11111. 11111 1111111111 11111111 :1 1 11.11111'11 I '11. '1‘ 1' 1,1111” 11111 1) 11111111111111 1111,11111111111 111 1 ‘ THESI.‘ 'LIBRA ' Michigan State University 1 This is to certify that the thesis entitled Factors Affecting Youth Service Bureau Implementation presented by Rodney L . Witt has been accepted towards fulfillment of the requirements for M. S. degree in Criminal Justice MW Major professor Date October 28, 1977 0-7639 OVERDUE FINES: 25¢ per day per item RETURNING LIBRARY MATERIALS: Place in book return to remove charge from circulation records FACTORS AFFECTING YOUTH SERVICE BUREAU IMPLEMENTATION By Rodney Layne Nitt A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE School of Criminal Justice 1979 ABSTRACT FACTORS AFFECTING YOUTH SERVICE BUREAU IMPLEMENTATION By Rodney Layne Witt Problem. This study's purpose is to develop a form of evaluation which focuses upon the relationships between a social program and its environment. This type of analysis is classified as implementation evaluation, and the objective of the study is to determine if factors within a project's implementation process affect that project's outcomes. Method, A specific project, Youth Service Bureaus, was selected for evaluation. The evaluation is divided into two segments. The first segment deals with gathering qualitative information con- cerning the implementation process and perceived outcomes relative to those projects. The second segment defines factors found to influence implementation and perceived outcomes. Data was collected by a survey instrument and analyzed by path analysis. Findings. The analysis found statistically significant relationships between factors within the implementation process and perceptions of project outcomes. Those implementation factors influencing perceived outcomes were the degree of initial support and degree of participation in the project's planning and development. Dedicated to my wife Linda ii ACKNOWLEDGEMENTS As it must be with every thesis this effort could not have been accomplished without the aid, support, and comfort freely given by faculty, friends and family. For that reason I would eSpecially like to thank Ralph Lewis whohounded through my own inertia to prepare this document. Without his insights and advice this would have been an impossible task. My appreciation is also extended to those members of my thesis committee, Ken Christian, Robert Scott, and David Kalinich for making their time available. Most of all I would like to thank my wife Linda for her patience and understanding through what was often a very unsettling period. iii TABLE OF CONTENTS Page LIST OF TABLES . . . . . . . . . . . . . . . . vi LIST OF FIGURES . . . . . . . . . . . . . . . vii Chapter 1. THE PROBLEM . . . . . . . . . . . . . . l Need . 1 Purpose . 4 Theory . . . . . . . . . . . . . . . 4 Hypotheses . . . . . . . . . . . . . . 5 Overview . . . . . . . . . . . . 6 II. REVIEW OF THE LITERATURE . 9 Introduction . . . . . 9 Definition of the Implementation Process . . . . 9 Findings . . . . . . . . l4 Summary of Literature Findings . . . . . . . l7 Need for Additional Literature Review . . . . . l7 Systems Theory . . . . . . . . . . . . l8 Basic Assumptions . . . . . . . . . . . 25 Hypotheses . . . . . . . . . . . . . . 25 Summary . . . . . . . . . . . . . . . 26 III. METHODOLOGY . . . . . . . . . . . . . . 3l Introduction . . . . . . . . . . . . . 31 Model . . . . . . . . . . . . . . 32 Site Selection . . . . . . . . . . . . 32 Data Sources . . . . . . . . . . . . . 34 Interviews . . . . . . . . . . . . . . 34 Sample . . . . . . . . . . . . 37 Survey Questionnaire . . . . . . . . . . 44 Response Selections . . . . . . . . . . . 46 Operational Measures . . . . . . . . . . 47 Path Analysis . . . . . . . . . . . . . 48 Summary . . . . . . . . . . . . . . . 51 iv ANALYSIS . . . . . Introduction . . . . . . Aggregation of Data . . . Dependent and Independent Variable Correlations Dependent and Intervening Variable Correlations Intervening and Independent Variable Correlations : Selection of Variables . . Interpretation of Path Analysis Results Summary . V. SUMMARY AND CONCLUSIONS Summary . Findings . . Policy Implications . . Implications for Future Research APPENDICES A. SURVEY INSTRUMENT B. RESULTS OF PATH ANALYSIS FOR ALL VARIABLES ANALYZED I BIBLIOGRAPHY ll3 Table 3.1 3.2 4.l 4.2 4.3 LIST OF TABLES Questionnaire Distribution . Questionnaire Response Dependent and Independent Variable Correlations . Dependent and Intervening Variable Correlations . Intervening and Independent Variable Correlations vi Page 40 42 55 58 59 Figure (JON QWN hhhhhhbhbhbbhbhhww \OCDNOS O u—I 01 .10 .ll .12 .13 .14 .15 .16 LIST OF FIGURES Implementation Process Model . Straight Line Relationship Triangular Relationship Structure Structure Structure Structure Structure Structure Structure Structure Structure Structure Structure Structure Structure Structure Structure Structure #1 #l4 . #5 #IO . #15 . #6 #ll . #16 . #7 #12 . vii Page 33 50 50 62 64 66 67 68 70 71 72 73 74 75 77 78 79 81 82 CHAPTER I THE PROBLEM figggi Since the l930$ there has been an increasing willingness by government to intervene in social problems which were once seen as being areas of private concern. This increased commitment to the elimination of a wide range of social problems has resulted in a larger proportion of resources being devoted to their resolution. These resources are primarily utilized in the funding of what are known as social programming or planned changes. As with any effort which requires the expenditure of scarce resources there is a high level of interest as to whether or not these social programs are accomplishing their intended goals. This interest has led to the general acceptance of the inclusion of evaluation components within planned change programs. It is the purpose of these components to provide the desired information as to whether or not the programs are working, and why. With knowledge gained through experience and contributions from applied science, evaluations have moved from Simply determining how resources were spent, to more sophisticated efforts aimed at measur- ing the actual results. There are at least four reasons why projects may not meet the expectations which their developers have for them. The first I of these is programmatic over-expectation. That is, hopes of what the program will accomplish have been raised to levels which may be impossible to attain. One factor contributing to this is that those areas which are targeted for social interventions are, by their very definition, areas which are not resolvable by normal societal mechan- isms. Thus, the results of social programming efforts, particularly in terms of cost-benefit, are probably going to be lower than the dramatic changes often anticipated. The second closely allied reason, as defined by Campbell, is over-advocacy.1 Simply put, over-advocacy is defined as the promising of greater results than can actually be delivered. This may come about since, because we are dealing with scarce resources, it is only natural that those programs, whose advocates promise the greatest benefit for the least cost, will be funded. The result is the infla- tion of anticipated results in order to secure funding, or over- advocacy. The third reason why projects may fail to produce the results anticipated of them is because of a conceptual failure. Conceptual failure assumes that all projects, no matter how simple, are based on an underlying conceptual framework. This theoretical framework identifies the causal factors of the problem and points to the types of measures which might be employed to produce the desired outcomes. However, if the theoretical framework on which the project is based is incorrect or inappropriate, the desired effects may not be achieved, resulting in project failure. The fourth reason why projects may fail is referred to as implementation failure. Implementation is defined as: encompassing a process which includes the creation in a client-system of understaning of, and commitment to, a particular change which can solve problems, and devices whereby it can become an integral part of the client- systems operation.2 If the conditions suggested by this definition are not met the impact model upon which the project was based cannot be tested, because the project willnot becarried through as originally intended. The use of this approach to project analysis dictates that programs be viewed as being operationalized through specific organizations. An organization comes into being when: (1) there are persons able to communicate with each other (2) who are willing to contribute action (3) to accomplish a common purpose. The elements of an organization are. therefore (l) communication, (2) willingness to serve, and (3) common purpose. However, such an organization is also a component of an even larger organizational system or structure. As such, a complete evaluation cannot be undertaken without consideration of the program's relation- ships with the organizations composing its external environment. In pointing up the need to study the organization in relation to its environment, Etzioni specifies the area of interorganizational rela- tions as one meriting further investigation.4 However, until fairly recently social scientists have not concerned themselves with the relationships between a program and its implementing organization and those organizations in their external environment. This lack of interest is reflected in the evaluation literature by Evan who states: The relative neglect of interorganizational relationships is all the more surprising in view of the fact that all formal organizations are embedded in an environment of other organizations as well as in a complex of norms, values, and collectivities of the society at large. Inherent in the relationship between any formal organiza- tion and its environment is the fact that to some degree it is dependent upon its environment.5 Purpose This study was undertaken as part of a model evaluation project contracted by the Michigan Office of Criminal Justice Programs. Its purpose is to develop and demonstrate a form of evaluation which has as its focus the relationships between a specific type of program and its environment. This type of analysis may be classified as implementation evaluation, and its use will be demonstrated as applied to a specific type of delinquency prevention program, Youth Service Bureaus. The main objective of the study is to determine if factors present within the Youth Service Bureau implementation process have an effect upon the perceived project outcomes. Theory Implementation analysis represents a systemic model of program evaluation. In this type of analysis the formal organiza- tions, through which the programs are operationalized, are seen as being members of larger social and political systems. These systems are composed of other formal organizations which constitute the external environment for the program. Formal organizations are characterized by both an external environment and an internal environment (goals, division of labor, role expectations, etc.). The placement of an innovative project into an established system represents a possible disturbance for the internal environment \ >032 (mm Oz< <mo \ mU_.—.m_¢m._.Uuo PUqumn. \ / mU_.Pm_¢m._.U<¢<_.—U \ mU_Pm_mmPU.—._ZDESOU 34 effort; (2) representativeness of other projects in the state (length of operation, size of community served, etc.); and, (3) the availability and completeness of bureau records. On the basis of these criteria the six sites were Port Huron, Flint, East Detroit, Benton Harbor, Owosso, and White Cloud.2 Data Sources The research was started by an examination of the offical records of the state funding agency for the bureaus, the Michigan Office of Criminal Justice Programs. These records included the initial grant applications, correspondence between the bureaus and the funding agency, and their quarterly reports. From these documents many of the actors who were instrumental in the original planning and development were identified. The need for each bureau and its goals and objectives, as stated in the initial grant application, were noted. The quarterly reports yielded the activities undertaken by each program, and in some cases began to establish which agencies made up their input organization set. All documents were carefully examined for information on how implementation had progressed and for clues of any specific problems which might have arisen. Interviews As this was being done researchers assigned to other components of the model evaluation project were already in contact with the bureaus. Using information collected by them in the course of their work, combined with that which resulted from the records search, an 35 in-depth interview format was developed. This interview was then administered to project directors and staff, as well as to those relevant external environmental actors, such as police chiefs, school administrators and court officials, who had already been identified. The purpose of the interviews was to begin to compile more specific information about the interrelationships between the bureaus, their environment and outcomes. In all, thirty-five inter- views were conducted, ranging in duration from approximately one to three hours. While allowing latitude for the interviewees to move into areas not directlyassociated withthe implementation process, the interviews were structured in such a way as to cover the desired areas as thoroughly as possible. Several methods were utilized to ensure accurate and reliable interpretation of the interview data. Two interviewers participated in each interview, with at least one taking detailed notes of the conversation. As a supplement to the notes, tape recordings were made whenever feasible. In order to preclude the possibility that some interviewees might be hesitant to discuss sensitive areas while being recorded, all subjects were given an option of whether or not to be recorded. While there were no instances of any subject declining to be interviewed while being recorded, there were several cases where the interviewers decided not to use the recorder because they felt it could potentially inhibit the discussion and therefore the quality of information received. It was the opinion of the interviewers that in no instance where the recorder was used did it inhibit the interviewee. 36 After each day of interviews both sets of notes and the recordings were analyzed by the interviewers, and a composite transcription of the three was made. As a check against possible bias or misinterpretation on the part of the interviewers an independent member of the research team also analyzed both the tapes and notes. The two sets of analyses were then compared to one another and any differences in interpretation resolved. The results were then integrated into one permanent record. This record was then consulted before the next round of interviews, and any new questions or issues raised were incorporated into the structure of the remaining interviews. By following this procedure it allowed information derived from the initial interviews to be used in two ways. The first was that issues raised within-site by a subject were put into the interview structure for all remaining subjects. This technique served to make the resulting data less prone to individual perspectives or possible bias because of the constant crosschecking, verification, and synthesis of a number of viewpoints. The second use was the incorporation of issues raised at one site into the interview structure of other sites. The procedure aided in isolating those issues which were a result of purely local phenomenon and those larger issues cutting across sites. It was this latter information which was used in the construction of a survey instrument seeking quantitative data for the research questions. The format for the interview consisted of first obtaining the history of the bureau, which included such information as why was it 37 needed, who suggested it, were other alternatives considered and what were its original goals and objectives. The bureau's actual implementation was then probed, concentrating on such areas as what were the problems encountered during implementation, had the goals and objectives of the bureau changed from those originally intended, was the staff capable of handling the types of cases referred to. them, how were resources being allocated, etc. During the interviews each of the subjects were also asked for the names of other actors who were presently, or had been, involved with the particular bureau under examination at that time, in one of two capacities. Those were the initial planning and development of the bureau or its current operation. Involvement in current operations was defined as either serving in some advisory capacity to the bureau or providing inputs (clients) to it. Inter- views were then scheduled with as many as possible of those identified as being critical in either capacity. This procedure also contributed to the development of a broad and accurate picture of bureau implemen- tation. Sample Through information gained by the records review and the interviews, a sample population was selected. The data indicated that the overwhelming majority of those involved with either the plan- ning and development, advisory capacities, or provision of inputs for the bureaus were representatives of one of three institutions. These institutions were the probate courts, local schools, or local police. 38 Therefore, it is logical to assume that these organizations make up the input organization set for the bureaus. By virtue of the fact that these organizations controlled the inputs for the bureaus, the extent to which they agreed with and believed bureau goals were or were not being accomplished would be a major influence upon implemen- tation. For that reason, representatives of those organizations were targeted as the research population from which the research sample was drawn. The research sample itself actually consisted of two sub- samples: (1) all those major actors who had been identified as being directly involved in the bureau's planning, or who were critical to the bureau's operation by way of contributing a large proportion of their inputs; and, (2) those actors who used the bureaus sparingly or not at all. This latter group was identified by obtaining lists of all police agencies and schools within the bureau's jurisdiction and comparing them to records kept by the bureaus on the sources of their referrals. For the purposes of implementation evaluation, the latter group was especially critical because of their familiarity with, involvement in, and influence upon the bureaus. The initial research had established that the total numbers of actors composing the core group which had actively participated in the planning and development process was relatively small. Thus, a low return rate from this group would provide such a small data base as to make inferences from it potentially unreliable or impossible. To preclude that possibility 39 and to ensure a high return rate for the core group, the survey instruments were personally delivered and administered wherever possible. The distribution of the survey instrument to the remainder of the sample was by mail. This raised at least one other sampling problem. While it was possible to obtain the names of the school officials and police chiefs for those agencies within the areas serviced by the bureaus, there were no means available for securing the names of numbers of juvenile officers and school counselors who could potentially provide inputs. To overcome this, whenever a survey instrument was sent by name to the head of an agency, several extras were included and the chief or principal was requested to distribute them among the appropriate personnel. The results of this type of distribution method have several implications: (1) some questionnaires may never have been distributed by the head of an agency as was requested; (2) more questionnaires may have been distributed than there were applicable personnel to fill them out; or (3) fewer questionnaires may have been distriubted than there were appropriate personnel to fill them out. Since there is no way of calculating how many non-responses may have been due to the first two factors, the return rate given is the most conservative figure in that it is undoubtably low. Regarding the third factor, no requests for additional questionnaires were received, therefore, it is assumed that sufficient questionnaires were distributed. A detailed breakdown of the questionnaire distribution is shown in Table 1. It is divided by Site and by each of the 40 TABLE 3.1.--Questionnaire Distribution. School Police Courts % of Sample Port Huron Addressed 5 11 O 6.86 Unaddressed 6 3 O 3.86 Core 1 2 2 2.14 Flint Addressed 13 13 O 11.15 Unaddressed 24 13 2 16.73 Core 2 O 2 1.75 East Detroit Addressed 23 2 6 13.30 Unaddressed O O O - Core 3 2 1 2.57 Benton Harbor Addressed lO 3 l 6.00 Unaddressed 24 6 O 12.87 Core 2 2 l 2.14 Owosso Addressed 2 5 O 3.00 Unaddressed 8 18 O 11.15 Core 2 2 l 2.14 White Cloud Addressed 1 1 l 1.28 Unadressed O O O - Core 1 2.57 Other __1_ __:_ _-_ . 42 TOTAL 128 85 18 100% Total of Sample Addressed 41.59% Total Unaddressed 44.61% Total of Sample Core 13.31% Other .42% 41 institutions making up the input organization set. In addition, the total number of questionnaires distributed to the core subsample and to the remaining subsample are presented. The noncore subsample is divided into the number of questionnsires which were addressed directly to an actor by name and those which were addressed to be distributed by the head of an agency. In all, a total of 233 question- naires were distributed by the model evaluation staff. The response rate for the questionnaire is shown in Table 2. It is divided by Site and by core and noncore subsample for each site. A total of 80 questionnaires were returned for a return rate of 34.33%. This figure includes those questionnaires which were administered personally. While the return rate for the core group was very satisfactory, the issue of sampling bias for the sample as a whole must be confronted. Bias may be defined as, "any influence, condition, or set of conditions, which, singly, or together, cause distortion or aberration of the data from those which may have been obtained under the conditions of pure chance; furthermore, bias is any influence which may have disturbed the randomicity by which the choice of a sample population was selected."3 In this study, sample selection was done in a purposive, as opposed to random, manner. All actors who could be identified as being instrumental in the planning and development of the bureaus were included in the sample. The same is true of those actors who potentially could, or did, provide inputs for the bureaus. Therefore, the question of sample bias due to the selection procedures may be TABLE 3.2--Questionnaire Response. 42 School Police Court % of Return Port Huron General 2 2 O 5.00 Core 1 2 l 5.00 Flint General 5 8 2 18.75 Core 2 O 2 5.00 East Detroit General 8 O 1 11.25 Core 3 2 l 7.50 Benton Harbor General 14 1 1 20.00 Core 2 O 1 3.75 Owosso General 4 1 O 6.25 Core 2 2 l 6.25 White Cloud General 0 l l 2.50 Gore 2 2 2 7 50 Other 1_ ___ ___ ___ 1.25 TOTAL 1 45 21 13 100.00% Average Total Response Rate 34.33% Average Total Core Response Rate 96.7 % Average Response Rate By Agency School 35.15% Courts 72.22% Police 24.70% Total % of Responses By Agency School 56.25% Courts 16.25% Police 26.25% 43 raised only if all the actors who fit the conditions for either of the sample subgroups were not identified, and as a result did not have an opportunity to respond to the questionnaire. Due to the extensive document research and interview schedule, it is unlikely that a sufficient number of actors, relevant to the bureaus, remained unidentified. For that reason it is also unlikely that the results of the questionnaire have been biased because of improper sample selection. However, there still remains a question of bias that may have resulted from differences in those who returned the question- naire and those who did not. To deal with the question of systematic bias caused by response or nonresponse patterns, an assumption must be made. This assumption is that the greater the relevance of the bureaus to the potential respondent the more likely the return of the questionnaire. Thus, those who have no strong feelings toward the bureaus, either positively or negatively, are least likely to reply. Consequently their absence will not skew the sample results in either direction. Extending this assumption, those who have very strong feelings, either positively or negatively, will be the most likely to respond. If such is true, the sample results will reflect a true picture of the population. Unfortunately it is impossible to discount the presence of some degree of bias in the study results. It is felt that, based upon those who did respond, if bias is present, it is weighted in the direction of those who view the bureaus in a positive manner. This tendency was allowed for in the data interpretation. 44 Examination of the return percentages by Site and respondent class (police, courts and schools) Show some variation. In all instances where the return rate was lower than average there were a large percentage of questionnaires sent which were not addressed to individuals, but were to be distributed by agency heads. Therefore, it is reasonable to assume that the low response rates were a result of overestimating the number of potential respondents, or a failure on the part of agency heads to distribute them, and not because of a resistance in responding to the questionnaire. Surveygguestionnaire The purpose of the survey instrument was to obtain quantita- tive data on issues of implementation identified by the qualitative portion of the study. Because the needs of the model evaluation project demanded covering a wider range of issues than could be examined by this study, a limited number of issues were selected for an intensive evaluation at this time.4 Here, those issues under evaluation are questions of relationships between a desire for alternatives, desire for outside alternatives, degree of initial support, and the degree of participation in the bureau's planning and development to perceptions of project outcomes. Stated more precisely, this study will test the hypotheses that a desire for alternatives, degree of initial support, and participation in planning and development, as the independent variables, will have a positive relationship to the dependent variables of perceptions of project success, and that this 45 relationship is not the result of intermediate ultimate effects, or perceptions of internal environmental success. The null hypotheses would then be that there is no positive relationship between the independent and dependent variables, and variation in the dependent variables is the result of the intermediate ultimate effects. The independent variables were obtained from respondent answers to the following questions: Prior to the Y.S.B. there were definitely sufficient outside agencies available to the police for juvenile referral. Prior to the Y.S.B. the police were highly desirous of additional alternatives to the Options of warn and release to the court for juveniles. Relevant representatives of the police were in favor of creating a Y.S.B. Relevant representatives of the police actively participated in the planning and development of the Y.S.B. These items were then repeated in a form applicable to both the courts and schools.5 The initial stage of the study established that there were intermediate ultimate effects (intervening variables) of the project's internal environment which were recognized by and important to rele- vant external environmental actors. These items, selected to measure perceptions of intermediate ultimate outcomes, appeared in the questionnaire as: Juveniles from my agency referred to the Y.S.B. always receive service within a reasonable period of time. I have no problem contacting the appropriate Y.S.B. staff people whenever I need information and/or service. 46 People from the Y.S.B. display a facilitative (helpful, cooperative) attitude toward your organization. The individual Y.S.B. personnel were equipped to handle the type of cases you referred to them from the beginning. The training provided by the Y.S.B. for its personnel has improved their ability to deal with the type of cases you refer to them. 1 am definitely satisfied with the feedback the Y.S.B. provides me. From the goals and objectives common to most Y.S.B.s a related set of outcomes were selected as the dependent variables. These were formulated as: The Y.S.B. has been responsible for a reduction in recidivism among juveniles. Officials (police and courts) have used the Y.S.B. to divert juveniles from the court. Since the Y.S.B. began operations there has been a reduction in the number of status offenses formally processed and petitioned to the court. Since the Y.S.B. began operations there has been a reduction in the number of non-status offenses formally processed and petitioned to the court. The Y.S.B. has changed the type of juvenile petitioned to the court.* Response Selections The response selections for all of the variables under con- sideration were based upon a Likert Scale whose divisions were: totally agree, strongly agree, agree, disagree, strongly disagree, *Appendix A contains the independent variables numbered as items 1 through 12, dependent variables as items 80-84 and the inter- mediate ultimate effects as items 70-74 and 76. 47 and totally disagree. There are several limitations associated with employing Likert Scales for this research. Briefly, there are four types of scalar categories for the measurement of data: nominal, ordinal, interval, and ratio. These categories are arranged in a hierarchy by degree of refinement, each possessing aspects which determine the inferences that may be made from them. The Likert Scale is ordinal in nature. Ordinal scales indicate classes or categories and rank order. 'They do not possess 6 The use of equal units of measurement or an absolute zero point. the non-standardized Likert Scale means that: (1) groups may be ordered with respect to the degree to which they possess a certain characteristic, and yet we may not be able to say exactly how much of that characteristic they possess; (2) we cannot make direct comparisons between scores derived from a Likert Scale because the Likert Scale may tap into different segments of the underlying variable. Operational Measures With those limitations in mind, the statistical techniques for use in hypothesis testing were selected. Given the manner in which the hypotheses were structured, several different stages of data manipulation were necessary for deriving the information upon which the hypotheses could be accepted or rejected. The first of these was to establish if there were, in fact, positive relationships between the dependent and independent variables. For that purpose 48 the variables were subjected to a Pearson r test. The purpose of this test was to provide a measure of association indicating the strength and direction of any relationship present. On the basis of the results from this operation it was possible to determine if the study would fail to reject any of the null hypotheses. This was true in those cases where no significant relationship was found between a dependent and independent variable. However, from the results of the Pearson r alone, it was not possible to determine if the variation of the dependent variable was uniquely attributable to the independent variable or if it might be the result of intervening variables, defined here as intermediate ultimate effects. This was accomplished by the use of what is known as path analysis. Path Analysis Path analysis is defined as "primarily a method of decompos- ing and interpreting relationships among a set of variables . . . ."7 One of the major assumptions upon which path analysis is based is that a weak causal relationship among the variables is known. To establish the presence of such a relationship between the intermediate ultimate effects and both the dependent and independent variables a Pearson r test was again used. In that step, as with other tests of relationship used in this study, the significance level was set at .05. Those variables which were found to have no significant relationships to others of the study were discarded. The remaining variables, therefore, satisfy the requirement for path analysis, 49 that of a weak causal relationship. However, given the result of the Pearson r test alone, no determination can be made as to which form, of two possible structures, the relationships take. The two structures possible are illustrated in Figures 3.2 and 3.3. In each of the examples given the intervening variable Y is dependent in relation to the independent variable X, and independent in relation to the dependent variable Z. In Figure 3.2 the variable subsystem of Y and Z is partly due to the causal dependence of Z on Y and partly due to their sharing of a common cause, independent variable X.8 By the use of path analysis the exact configuration of the structure can be determined and the unique contribution of each independent variable in relation to the dependent variable can be ascertained. If the results of the analysis Show that there is a measure of association which runs from X directly to Z, and from X to Y and then to Z, the structure will be as appears in Figure 3.3. If the path of association runs entirely from X through Y to Z the structure will correspond to Figure 3.2. While there are several measures derived from path analysis, those which will be used by this study are the nonstandardized regression coefficients, the standard error of B, and the standard- ized regression coefficients. The nonstandardized regression coeffi- cient is the slope of the regression line and indicates the magnitude of the expected change in Y for one unit change in X. The standard- ized regression coefficient (beta weight) indicates the same slope of the regression line except that both X and Y are expressed in 50 FIGURE 3.2.--Straight Line Relationship. / \ FIGURE 3. 3.-—Triangular Relationship. 51 standardized units. It is the standardized regression coefficient which will be used by this study to determine the strength of the relationships between the variables. The standard error of B is used in measuring the standard deviation of the sampling variability of B. As a rough test of significance the standardized regression coefficient should be at least twice as large as the standard error of B. However, for the purposes of this study significance will be determined by evaluation of the F ratio. Summar In summary, the research design was divided into two segments. The first segment consisted of a records review and series of inter- views for the purpose of securing information on the implementation process at six selected Youth Service Bureau sites. The second seg- ment was a survey instrument directed at relevant actors in the Youth Service Bureau's external environment. The purpose of this instrument was to obtain quantitative data about critical factors in the implementation process which had been discovered during the first portion of the research study. The analysis of the resulting data was done by means of the Pearson r and path analysis. The purpose of these statistical operations was to determine if factors in the bureau's implementation were related to perceptions of bureau outcomes, and if so, to ascertain if the relationships were direct and unique or the result of intervening variables. The analysis of the results of the statistical operations will be displayed in Chapter IV. CHAPTER III FOOTNOTES 1Berman and McLaughlin, Federal Programs, p. 19. 2While most of the bureaus were countywide in their operating area, they will, for this study, be identified by the town in which their offices were located. 3Paul D. Leedy, Practical Research: Planning and Design, (New York, New York, Macmillan Publishing Company, Inc., 1975), p. 107. 4The complete questionnaire as utilized by the model evaluation project is contained in Appendix A. 5N. M. Downie and R. W. Heath, Basic Statistical Methods, (New York, New York: Harper & Row, 1965), p. 14. 6Jae-On Kim and Frank J. Kahout, "Special Tapics in General Linear Models," in Statistical Package for the Social Sciences, ed. Norman H. Nie et a1. (McGraw-Hill, Inc., 1975), p. 383. 7 Ibid., p. 385. 52 CHAPTER IV ANALYSIS Introduction The purpose of this chapter is to present the analysis of the data generated by this study. Specific variations of the follow- ing hypotheses will be tested. H]: There is a positive relationship between a high degree of desire for alternatives and high percep- tions of project success. H2: There is a positive relationship between a high degree of desire for outside alternatives and high perceptions of project success. H3: There is a positive relationship between a high degree of initial support and high perceptions of project success. H4: There is a positive relationship between a high degree of participation in the planning and develop- ment of the project and high perceptions of project success. These hypotheses will be tested in such a manner as to determine that the relationships set forth are direct and not the result of intermediate ultimate effects, such as internal environmental SUCCESS . Aggregation of Data As has been detailed in the preceding chapter, there were several statistical procedures which were applied to the raw data before it was in a form suitable for determining and testing each of 53 54 the specific hypotheses. The first of these steps was the applica- tion of a Pearson r test to the dependent and independent variables to determine which were related. However in the survey instrument, the independent variables (desire for alternatives, desire for outside alternatives, degree of initial support, and participation in the planning and development) were presented in a series of twelve items. Each of the independent variables was presented three times. One for each of the organiza- tions making up the bureau's input organization set, the school's, police and court's. For example: Prior to the Y.S.B. there were definitely sufficient out- side alternatives available to the police for juvenile referral. Prior to the Y.S.B. there were definitely sufficient out- side alternatives available to the schools for juvenile referral. Prior to the Y.S.B. there were definitely sufficient out- side alternatives to formal disposition for the court. For the purposes of analysis, all items which were identical, except for being agency specific, were combined into one measure. By doing this the responses of court personnel on a court specific question were added to school personnel responses on the identical school specific question, and then both were added to the police response. The procedure reduced the twelve items to four separate measures where the questions for each specific agency type were only used for representatives of that type of agency. The combined measures are denoted in the analysis as NEED (corresponding to a desire for alternatives), DESIRE, (a desire for outside alternatives), 55 FAVOR (initial support), and PARTIC (participation in the planning and development). Dependent and Independent Variable Correlations The four combined independent variables were then correlated with the dependent variables, which were: VAR081 VAR082 VAR083 VAR084 VAR085 The Y.S.B. has been responsible for a reduction in recidivism among juveniles. Officals (police and court) have used the Y.S.B. to divert juveniles from the court. Since the Y.S.B. began operations there has been a reduction in the number of status offenders formally processed and petitioned to the court. Since the Y.S.B. began operations there has been a reduction in the number of non-status offenders formally processed and petitioned to the court. The Y.S.B. has changed the type of juvenile petitioned to the court. The variables were then subjected to the Pearson r test; the results are presented in Table 4.1 TABLE 4.1.--Dependent and Independent Variable Correlations. VAR081 VAR082 VAR083 VAR084 VAR085 NEED -.0833 .0329 -.0725 -.0031 .0033 S = .2600 S = .3930 S = .2910 S = .4910 S = .4900 DESIRE .0807 .2298 .0636 .0198 .1669 S = .2660 S = .0280 S = .3150 S = .4400 S = .0970 FAVOR .3846 .1810 .2307 .2360 .1027 S = .0030 S = .0670 S = .0380 S = .0340 S = .2140 PARTIC .5741 .2308 .5528 .3775 .4573 S = .0010 S = .0270 S = .0010 S = .0010 S = .0010 56 From the results generated by the first step, and displayed in Table 4.1, it is impossible to accept any of the research hypothe- ses. This is because, even though there may be a significant correla- tion between variables, it is not known if the correlation is a result of the independent variable or the result of an intervening variable, intermediate ultimate effects. However, on the basis of the Pearson r test alone, it can be stated that a number of independ— ent variables seem to have no relationship to a number of dependent variables, because of a lack of significant correlation between 2 them. For that reason the study has failed to reject the following null hypotheses: H1 : There is no positive relationship between a high degree of desire for alternatives and high perceptions of the Y.S.B. being responsible for a reduction of recidivism among juveniles. H2a: There is no positive relationship between a high degree of desire for outside alternatives and high perceptions of the Y.S.B. being responsible for a reduction of recidivism among juveniles. 2C: There is no positive relationship between a high degree of desire for outside alternatives and high perceptions of there having been a reduction in the number of status offenders formally processed and petitioned to the court since the Y.S.B. began operations. 2d: There is no positive relationship between a high degree of desire for outside alternatives and high perceptions of there having been a reduction in the number of non- status offenders formally processed and petitioned to the court since the Y.S.B. began operations. 2e: There is no positive relationship between a high degree of desire for outside alternatives and high perceptions of the Y.S.B. having changed the type of juvenile petitioned to court. 3p: 3e: 57 There is no positive relationship between a high degree of initial support and high perceptions of officials (police and court) having used the Y.S.B. to divert juveniles from the court. There is no positive relationship between a high degree of initial support and high perceptions of the Y.S.B. havin93changed the type of juvenile petitioned to the court. Dependent and Intervening Variable Correlatiahs Having once established which dependent and independent variables were related, the second step in the data analysis was to determine which intermediate ultimate effects (intervening variables) were related to the dependent variables. The specific intermediate ultimate effects to be tested were: VAR070 VAR071 VAR072 VAR073 VAR074 VAR076 Juveniles from my agency referred to the Y.S.B. always receive service within a reasonable period of time. I have no problem contacting the appropriate Y.S.B. staff people whenever I need information and/or service from them. People from the Y.S.B. display a facilitative (help- ful, cooperative) attitude toward your organization. The individual Y.S.B. personnel were equipped to handle the types of cases you referred to them from the beginning. The training provided by the Y.S.B. for its personnel has improved their ability to deal with the types of cases you refer to them. I am definitely satisfied with the feedback the Y.S.B. provides me. The results of the Pearson r test for this step are displayed in Table 4.2. TABLE 4.2.--Dependent and Intervening Variable Correlations. VARO81 VAR082 VAR083 VAR084 VAR085 VAR070 .3130 .4001 .2934 .2698 .2707 .0080 .0010 .0130 .0210 .0190 VAR071 .4382 .3528 .3243 .3065 .3344 .0010 .0010 .0060 .0080 .0040 VAR072 .6106 .3790 .4394 .2026 .5298 .0010 .0010 .0010 .0620 .0010 VAR073 .3331 .4280 .2727 .1821 .3451 .0060 .0010 .0220 .0920 .0040 VAR074 .4752 .5045 .3999 .1930 .3573 .0010 .0010 .0020 .0850 .0040 VAR076 .5576 .3782 .4865 .4364 .3756 .0010 .0010 .0010 .0010 .0020 Intervening and Independent variable Correlations In order to complete the causal relationship structure to be examined by path analysis it was necessary to establish if there were relationships present between the independent variables and the intermediate ultimate effects (intervening variables). This was again done by the use of the Pearson r test, and the results are contained in Table 4.3. Selection of Variables By rejecting those pairs of variables whose correlation coefficients were not significant at the .05 level it is now possible to determine which combinations of dependent, independent, and inter- vening variables may be arranged together in a structure, for the 59 opoo. u omoo. u opoo. n m opoo. u m opoo. u m opoo. u m mnme. wemm. omcm. mmmv. mmmm. mmme. o~km vuom<> mnom<> Numm<> Fmom<> ouom<> .mcowpmpmccou mpnmvcm> ucmccmamuca use mcvco>cmchiu.m.e m4m_o_oo.. E c3328.. a .5» 0.9333; .m.m.> . ; mgauuagumii . — .3 mango—n. 2m . .Qm 223. Sam /\ 2e . .mfi Ream . 3am “39:30.. momma yo monk. Evans 3 323360 _occomcoa .m.m.> mo> .g mgauuagumil .N .: mmDnu. n. as . .mfi :2..- Sam /\ 3c . whosm . .mfi 30m 2.82 3 3:33 :2: Uo>oLQE_ mm: _occomion Le .m.m.> 2: .3 9:59; mo> 9: Dom: 2.3050 mm. . .mfi 23o . 2mm /\ U.._.m of 3 octave 3:52... 25 . .mfi ~35 . Sam 67 «.58 «5 59¢ 3:533 “.52“. 3 .m.m.> 2.: com: 33050 6* mean—03.25115 .2 meOE 8m . .mfi e33 . Sam A _ U_._.m_ 2: :5: piece 52m. Bum .m ; mgzuoagum .mé mmDUE 68 m3. .mfi 33.. 30m $38 05 80..» 3.282: “.520 3 .m.m.> A U_._.m 9: 23. .m5 Swan. 30m xumnvmmu ct? cozuma—me—mm 325. Sam 69 H4b: There is no positive relationship between a high degree of part1c1pat1on 1n the planning and develop- ment of the project and high perceptions of officials (police and court) having used the Y.S.B. to divert juveniles from the court. The next three structures all share the common dependent variable VAR083, since the Y.S.B. began operations there has been a reduction in the number of status offenses formally processed and petitioned to the court. They are structures #4, #9, and #14 as illustrated in the proceeding diagrams. In all three cases the independent variable PARTIC is significantly related to the depend- ent variable VAR083, while controlling for the effects of the inter- vening variables VAR070, VAR072, and VAR076. The path coefficients between the dependent and independent variables are all significant at the .000 level. PARTIC is also significantly related to all three intervening variables. In structures #9 and #14 the inter- vening variables, VAR072 and VAR076, are significantly related to the dependent variable VAR083. However, in structure #4 the inter- vening variable is not significantly related to the dependent variable. On the basis of the results of these three structures the study may reject the following null hypothesis: H4c: There is no positive relationship between a high degree of participation in the planning and develop- ment of the project and high perceptions of there having been a reduction in the number of status offenses formally processed and petitioned to the court since the Y.S.B. began operations. The next three structures to be examined are #5, #10, and #15 which share the common dependent variable of perceptions of 7O 2:8 or: 8 8:253 Ucm commouoLQ 3.258 momcoto magnum E cozozoom .1 ocauuabmiié .: mMDUE e8 . .9m :22. Sam A U_._.m._om o>_ooo.. m>m5m .m.m.> of 3 00:39. 3:552. e8 . .mfi 33mm . Bum 71 {sou 05 2 8:253 Ocm commenced 33:58 momcoto magnum c. actuate”. .mu ogauusgumiiK .: mmDUT... o8. .mfi :33. Sam A mmo . :w—cm. .mfi 30m ocaztm 0233.33 at tam—am? .m.m.> 2.: Eat cacao BEE so . 6rd. :8... . Sam 72 #58 05 3 3:258 ncm 3mmwuoLQ 3.2.28 mmmcoto magnum c_ c2333”. /\ .3; wgauUSLumlué... MMDUE 8° . .Em 33.. . 3am Urey—(n— 8: . .9m ”88 . Sam .m.m.> 2: E .8233. 53.88 53 cozuemsmm NS. .mfi 25m . Sam 73 2:8 3 «5:23qu Ucm 60308.5 momcwto maumumucoc c. cozuavam .5“ mgsuuabmund... mmDUE o8 . .m.m 3......” . 3mm /\ o2. .m.m 88.. Sam 9:: 03233.. a E 8.3mm 0280.. m>m.s_m .m.m.> 3 3.3%.. 3.252... 0.5.}. $8. .mfi :33. Sum 74 £38 3 U0co_:u0a new 0030095 .0030qu magnumicoc 5 9.0.6330”. .c 2* Osauusbmnl .o— .3 MEAD—u. i... .m.m 3:... . 500 A =8. .mfi 3...... . Sam 0U33um 0235.03 0 >0_n_m_v .m.m.> Eat 0300.... 0.5.5. 8.. . .m_m 88... Sam 75 0.500 3 00.8352. 0.5 00000095 00020000 0:030:20: 5 5003001 .m 5 mgauusxumnu. — p .3 mmamv—u— =8 . .90 :02. 300 % /\ _. 3... .90 028 . 30m. .m.0.> 2: >0. .8255 0.03080 5.; 8:800:00 0.52“. .8. .90 28... Sam 76 there having been a reduction in the number of non-status offenses formally processed and petitioned to the court since the Y.S.B. began operations. In all three structures the independent variable PARTIC is significantly related to the dependent variable VAR084. Both intervening variables VAR070 and VAR072 are not significantly related to the dependent variable. However, VAR076 has a significant coefficient with VAR084 of .33256. The relation- ship between VARO76 and VAR084 appears to be Stronger than the relationship between PARTIC and VAR084 suggesting that the interven- ing variable has an equal or greater effect upon the dependent variable than does the independent variable. On the basis of these three structures the study may reject the following hypothesis: H4d: There is no positive relationship between a high degree of part1c1pation 1n the planning and develop- ment of the project and high perceptions of there having been a reduction in the number of non-status offenses formally processed and petitioned to the court since the Y.S.B. began operations. The next three structures which will be examined are #6, #11, and #16 which all share the common dependent variable of percep- tions of the Y.S.B. having changed the type of juvenile petitioned to the court. In all three structures the independent variable PARTIC is significantly related to the dependent variable VAR085. PARTIC is also significantly related to all intervening variables. Two of the intervening variables, VAR070 and VAR076, have no significant relationships with the dependent variable. However, VAR072 is significantly related to VAR085. Thus, in Structure #11 77 «.500 0.: 8 00:00:00. 22.02: 00 095 05 009.050 00... .m.m.> 2:. .3. whauuagumll .NF .3 mango—u. 8.. . .90 83... Sam /\ fl _ mam . .90 02.: . 300 0E: h.0 00200 03000000.. 0 E 00330 02000.. 0>03_0 .m.m.> 05 00. 00.2000. 0020.024 0.55 8.. . .90 82...... . Sum 78 0.500 05 3 00:05:00 05.02.: 00 00.5 05 .8055 00.. .m.0.> 0.? . F g 0.5.533qu :2 .3 mum—DD."— .. S . .90 .5 5... Sam /\ 8.. . .90 02:... Sum 003330 0200:2000 0 >0_00_0 .m.m.> 05 80.: 0.0000 0.52.. 8.. . .90 30:. . 300 79 {:00 05 00 00:03:00 0:002: 00 003 05 .8055 00.. .m.0.> of. .m ; unnuuagumll .3 F .3 mmDUE 08. .90 mmmom. 0.5m /\ ”___—1 03 . .90 MNQNN. 30m .m.m.> 05 >0 00030.0 0.03000. 5.; 5:080:00 0.52.. s... .90 3333 . 000m 80 the intervening variable seems to influence the dependent variable to the same extent as does the dependent variable. On the basis of these three structures the study has been able to reject the following null hypohtesis: H4e: There is no positive relationship between a high degree of participation in the planning and develop- ment of the project and high perceptions of the Y.S.B. having changed the type of juvenile petitioned to the court. The final structures to be examined are #7, and #l2 which share the common dependent variable of perceptions of the Y.S.B being responsible for a reduction in recidivism among juveniles. In both cases the independent variable PARTIC and the dependent variable VAR081 have a significant relationship. Also, in both structures, it may be said that the path coefficients between the independent and intervening variables and the dependent variable are of approximately the same magnitude. This indicates that the independent and the intermediate ultimate effects are almost equally important in their influence upon perceptions of outcomes. 0n the basis of the results from these two structures the study may reject the following null hypothesis: H4a: There is no positive relationship between a high degree of participation in the planning and develop- ment of the project and high perceptions of the Y.S.B. having been responsible for a reduction in recidivism among juveniles. Examination of the overall results finds several trends indicated by the data. First of these is that the results of the 81 00:c0>:_ 9.080 503.030.. 0. c0:0300.. 0 .00 0.0.0.5000; .m.m.> /\ .20 finneggumllfip .3 mmDnu—n— 8.. . .90 830 . 300 8o. .90 2.3.. . Bum 0.03.00 o>...3...u£ 0 >200... .m.0.> 0... so... 0.08.. 0:02.". 8.. . .90 Sam. 0.00 .N :0 wgauusbmun .2 .3 mmDUE 82 8... .90 hmmg. 000m 00::0>:_ 0:050 Em_>_0_00.. K :_ :000300: 0 ..00 / 0.0.03.0 0.0.0:0000.. . m .m. > S... .90 .m.0.> 2.. .3 00039... S... .90 $30. 300 53.080 5.; 5:800:00 3...... 500 83 Pearson r test show no significant correlations between the independ- ent variables of a desire for alternatives and a desire for outside alternatives among members of the bureaus' input organization set to any of the dependent variables.5 It did find that the degree of initial support (FAVOR) and participation in the planning and develop- ment of the project (PARTIC) both correlated significantly with all of the dependent variables. when these relationships were tested by the use of path analysis, which controlled for the effects of specified intervening variables, it was found that the relationship between FAVOR and VAR081 was not significant but was the result of the effects of VAR073 and VAR074. The relationship between PARTIC and VAR082 was not significant when controlling for VAR070, VAR072, or VAR076. However, even when controlling for the effects of the intervening variables, PARTIC does have significant relationships with dependent variables VAROBl, VAR083, VAR084, and VAR085. From these results it is reasonable to assume that the degree of participation in the planning and development of the project does influence the manner in which specific project outcomes are perceived. The study also found that FAVOR and PARTIC also had signifi- cant relationships to all intervening variables with which they were tested. This suggests that these early activities may also influence perceptions of intermediate ultimate outcomes for the projects. Finally it was found that all of the intermediate ultimate effects had significant relationships with each of the dependent variables tested. This was true even when controlling for the 84 effects of the independent variables. It is therefore reasonable to assume, that the perceptions of internal environmental outcomes are related to the manner in which the project outcomes are perceived. Summary The purpose of this chapter was to determine, by the use of path analysis, if the independent variables had significant relation- ships with the dependent variables when the effects of the interven- ing variables were controlled for. One of the requirements of path analysis is that there be a weak causal relationship between the variables. In order to satisfy this requirement the variables under consideration were tested by the use of the Pearson r. Only those variables which were significantly correlated were to be used in the path analysis. However, the number which proved to be so related was too large to be analyzed with the amount of resources available. For that reason the combinations tested were limited by selecting only those variations where the correlation coefficient between the dependent and independent variables were greater than .3000, and where the coefficient between the dependent and intervening variables was greater than .4000. The results of the path analysis found that the following null hypotheses could not be rejected: H3 : There is no positive relationship between a high degree of initial support and high perceptions of the Y.S.B. being responsible for a reduction in recidivism among juveniles. 85 H4b: There is no positive relationship between a high degree of participation in the planning and development of the project and high perceptions of officials (police and court) having used the Y.S.B. to divert juveniles from the court. On the basis of the results of the path analysis the study was able to accept the following hypotheses: H4c: There is a positive relationship between a high degree of participation in the planning and development of the project and high perceptions of there having been a reduction in the number of status offenses formally processed and petitioned to the court since the Y.S.B. began operations. H4d: There is a positive relationship between a high degree of participation in the planning and development of the project and high perceptions of there having been a reduction in the number of non-status offenses formally processed and petitioned to the court since the Y.S.B. began operations. H4e‘ There is a positive relationship between a high degree of participation in the planning and development of the project and high perceptions of the Y.S.B. having changed the type of juvenile petitioned to the court. 4a: There is a positive relationship between a high degree of participation in the planning and development of the project and high perceptions of the Y.S.B. being responsible for a reduction in recidivism among juveniles. The implications of these findings for policy formulation will be discussed in the following chapter. CHAPTER IV FOOTNOTES 1In the survey instrument, scales for the various items ran in two directions. Thus, a numerical value of 6 on one item might indicate the same as a numerical value of l on another item. In such instances the reverse order was brought into agreement before the responses were added together. 2This statement cannot definitely be said to be true in all possible cases because of the possibility of a supressor variable. 3The hypotheses notation has been standardized throughout this study. A notation of H] refers to a hypotheses containing the independent variable NEED and all possible variations of it with the dependent variables. Following this, H2 is DESIRE, H3 FAVOR, H4 PARTIC. The letter in the subscript refers to the specific dependent variable. These are denoted by the letter a through e, which corresponds to the dependent variables VAR08l through VAR085. 4For the complete results of all combinations which were tested see Appendix B. 5Except between DESIRE and VAR082. 86 CHAPTER V SUMMARY AND CONCLUSIONS Summary The purpose of this study was to develop and demonstrate the use of implementation evaluation as applied to planned change pro- grams. For the study, implementation was defined as: encompassing a process which includes the creation in a client-system of understanding of, and commitment to, a particular change which can solve problems, and devices whereby it can become an integral part of the client- systems operation;1 Implementation evaluation is based upon the assumption that if the conditions suggested above are not met then the impact model for the project cannot be tested, because the project will not be carried out as originally intended. The use of this approach to project analysis dictates that programs be viewed as being Operational- ized through specific organizations which are components of even larger organizational structures or systems. As such, a complete evaluation cannot be undertaken without considering the relationships, and their effects, which have developed between a planned innovation and those organizations composing its external environment. In reviewing the literature it was found that very few studies made the implementation process their basic unit of analysis. As Williams states, 87 88 I am certain that no bibliographic effort, at least in the social policy areas, would refute the point that little research has been carried out either on implementation of social policiei, programs, or on the implementation process agency . . The literature review for this study found only four works dealing with implementation. For that reason consideration was given to two additional bodies of literature: systems theory and interorgani- zational theory. From these three sources, a set of assumptions were generated upon which the hypotheses for this study were based. These assumptions were: 1. Planned innovations may be viewed as open systems. As such they are characterized by both an internal and external environment. Open systems are also characterized by their constant interaction with and dependency upon their external environment for a supply of inputs and consumption of outputs. The external environment of a planned innovation is composed of formal organizations which may also be viewed as Open systems. They too are characterized by an internal environment and an external environment in which the innovation takes its place. Planned innovations represent potential changes in the external environment of existing organizations. These changes in the external environment of existing organiza- tions may have consequences for those organizations internal environment. The degree and type of change required by the innovation, for those organizations, will determine the degree of support or opposition for the innovation. This support or opposition may be expressed by the degree to which the innovation's survival resources are provided. Given this dependency upon other organizations for the provision of survival resources, the innovation's external environment may influence its goals and act1v1t1es. 89 7. The degree to which planned innovations are implemented, as intended, will be influenced by the support/opposition they receive from their external environment. 8. The degree of implementation will influence the effects achieved by, and the potential for institutionalization of, an innovation. This study was undertaken as one component of a model evalua- tion program whose overall objective was the evaluation of a specific delinquency prevention program, Youth Service Bureaus. In that capacity the goal of this study was to define and evaluate factors within the Youth Service Bureau's implementation process which affected both their implementation and perceived outcomes. Much of the theory developed during the review of the litera- ture was concerned with the effects of support or opposition exhibited toward the innovation by organizations in its input organizational set. Of primary interest to the study was the potential for this support or opposition to influence perceptions of the innovation's internal environmental outcomes, overall outcomes, and potential for institutionalization. The initial stage of the research design found that this support or Opposition not only influenced the actual implementation and outcomes, but also seemed to influence perceptions of outcomes. The second stage of the research design was therefore devoted to determining if activities carried on during the initial stages of implementation directly influenced the manner in which the project outcomes were perceived. However, there was also evidence which indicated that perceptions of internal environmental outcomes also influenced actual environ- mental outcomes. For that reason path analysis was chosen to 90 analyze the hypothesized relationships to determine the unique contributions of each variable to the perceptions of project out- comes. One of the requirements of path analysis is a weak causal relationship between the variables in question. This requirement was met by determining a correlational relationship, by using a Pearson r test among the variables under consideration. While a causal relationship may not be directly inferred from a significant correlational relationship, it is reasonable to assume that in most cases there would not be a causal relationship and not a correlational relationship. For that reason the Pearson r test was used as a selection device for the variables to be examined by path analysis. The initial determination of the variables to be tested by the Pearson r test was done on the basis of the results of an exten- sive records review of all documents related to the conception and implementation of the bureaus. Additionally, a series of structured interviews were conducted with relevant external environmental actors from the bureaus input organization set. From those activities applicable examples of initial support, intermediate ultimate effects, and project outcomes were selected. A survey instrument was then developed and administered in order to obtain quantitative data on the perceptions of the actors in question. Findings The results of the correlational tests performed on the survey data found: 91 There was no positive relationship between a high degree of desire for alternatives and any of the selected variables measuring perceptions of project success. That there was no positive relationship between a high degree of desire for outside alternatives and any of the selected variables measuring perceptions of project success except perceptions of officials (police and court) using the Y.S.B. to divert juveniles from the court. That there were no positive relationships between the degree of initial support and high perceptions of officials (police and court) using the Y.S.B. to divert juveniles from the court, and the Y.S.B. having changed the type of juvenile petitioned to the court. Once the presence of correlational relationships had been established, the next step was to determine the magnitude and struc- ture of the proposed causal relationships. The instrument used for that purpose was path analysis. Since sufficient resources were not available to test all variations, specific combinations were selected on the basis of criterion established earlier in this study. The results of the path analysis were: 4. There was a positive relationship between the degree of initial support and perceptions of the intermediate ultimate effects of people from the Y.S.B. displaying a facilitative (helpful, cooperative) attitude toward referring organizations and satisfaction with the feed- back provided by the Y.S.B. There was a positive relationship between a high degree of participation in the planning and development of the project and high perceptions of the intermediate ultimate effects of people from the Y.S.B. displaying a facilita- tive (helpful, c00perative) attitude toward referring organizations, and between satisfaction with the feedback provided by the Y.S.B., and also between perceptions of juveniles referred to the Y.S.B. always receiving service within a reasonable period of time. 92 6. Controlling for the effects of the intervening variables of intermediate ultimate effects there was no positive relationship between the degree of participation in the planning and development of the project and high percep- tions of officials (police and court) using the Y.S.B. to divert juveniles from the court. 7. Controlling for the effects of the intervening variables intermediate ultimate effects there was a positive relationship between the degree of participation in the planning and development of the project and perceptions of the following project outcomes; there being a reduce tion in the number of both status and non-status offenses formally processed and petitioned to the court since the Y.S.B. began operation; and the Y.S.B. having changed the type of juvenile petitioned to the court. 8. Controlling for the effects of the independent variables there was a positive relationship between all intervening variables and all dependent variables tested. Policy Implications These findings contain several implications for successful implementation and institutionalization of innovative projects in general and Youth Service Bureaus in particular. It is obvious that the higher the perceptions of the success of project outcomes, among relevant actors, the higher the support which will be generated for the institutionalization of the project. This study has concen- trated upon several factors which seem to be related to high percep- tions of project outcomes. However, for the studies findings to be of any practical use the variables dealt with must be amenable to manipulation in order to produce the desired results. For that reason several implementation strategies which might impact upon the variables in the desired fashion will be discussed with each of the policy implications. 93 The first finding which has direct policy implications is that both the degree of initial support and the degree of participa- tion in the planning and development of the project was positively related with the perceptions of internal environmental outcomes. It was also found that the internal environmental outcomes signifi- cantly affected perceptions of the projects outcomes. For that reason an effective implementation plan should include provisions for developing the latter two areas. There are several methods which seem to effectively increase the degree of initial support. These would include identi- fication of those actors critical to the project's success prior to the time that it actually begins operations. Once this is done representatives of the project may meet with those actors personally, explain the project's goals, and try to overcome or reduce any opposition. This approach would also facilitate the identification of potential conflicts between established organizations and the project. When done at an early date arrangement or modifications would be easier to incorporate into the proposed project than after it was once established. Initial support could also be increased by educating the public (potential consumers of the projects services) by means of such activities as a media campaign, utilizing newspapers, radio or television, and speeches to local organizations. This in effect creates a constituency for the project which is one consideration when it is reviewed for potential institutionalization. 94 The degree of participation in the planning and development of the project is a variable which is also open to manipulation. Such practices as design of the pr0ject by use of a planning committee increases the opportunity for the inclusion of relevant actors. Use of this procedure tends to co-opt actors who might otherwise oppose the project and gives them a stake in its successful operation. However, care must be exercised in the use of this procedure to ensure that it is not manipulative for the purpose of gaining support and that it does not contribute to modification of the project solely for the sake of serving established interests. If a planning committee or a similar such group is not feasible, individual actors should be contacted on an informal basis to advise them of planning activities and to solicit their advice and suggestions. This procedure has the advantage of including, in the planning process, a large number of actors and their opinions without the disadvantages posed by a formal committee. The second important finding of the study which has implica- tions f0r policy development is that perceptions of internal environ- mental outcomes influence perceptions of project outcomes. The study examined three of these intermediate ultimate effects and found they all had significant relationships with all project outcomes tested. The three, perceptions of juveniles referred to the Y.S.B. always receiving service within a reasonable period of time, people from the Y.S.B. displaying a facilitative (helpful, cooperative) attitude toward referring organizations, and satisfaction with the 95 feedback provided by the Y.S.B., are all amenable to manipulation. One of the primary methods of ensuring a facilitative attitude on the part of project staff toward referring organizations is a comprehensive indoctrination of the staff on the importance of such an attitude. Another technique might be periodic meetings between project staff and members of the input organization staff to bring out and resolve any problems which might devel0p. A third would be to periodically assign members of the staff to referring organiza- tions for a short period of time. This would allow them to see first hand the types of problems other agencies must cope with and at the same time permit them to develop personal contacts with the other organization's staff. During the onsite interviews with actors relevant to the Y.S.B. the researchers for this study found one of the most common complaints to be a lack of feedback from the Y.S.B.s on cases referred to them. Since this seemed to be such an important factor it is recommended that Youth Service Bureaus should make it a routine matter of policy to provide feedback to the relevant actors in cases referred to them. If possible this should be done by some form of personal communication, rather than by mail, thereby increas- ing the opportunities for the identification of potential problems which might be developing. Implications for Future Research There are several areas for future research which are logical extensions of this study. In establishing the presence, and 96 magnitude, of causal relationships between the independent and dependent variables, only one intervening variable at a time was controlled for. Yet, the results indicated that there were a cluster of intervening variables which had significant relationships with the dependent variables. One avenue of future research would be to test whether those independent variables found by this study to be significantly related to the dependent variables would remain when the effects of several appropriate intervening Variables were considered at one time. The results of this study indicate that only a small grouping of the factors which influence perceptions of project outcomes were identified. Future research would be of value in identifying additional implementational factors in this area. Such areas as clarity of goal definition, agreement with goals, and questions of resource allocation, all seem to be variables which might be explored. A final area which might be examined is the relationships of the independent variables identified by this study. For instance, is a desire for alternatives causally related to the degree of initial support and is the degree of initial support related to the degree of participation in the planning and devel0pment of the project? As the answers to these and other questions are established the practice of implementation will become a valuable and precise tool in ensuring project success. CHAPTER V FOOTNOTES 1Bennis, Changing Organizations, p. 175. 2Williams and Elmore, Social Program Implementation, p. 4. 97 APPENDICES 98 APPENDIX A SURVEY INSTRUMENT 99 Please circle the appr0priate number indicating the extent of your agreement/disagreement with the following statements. Totally Agree Strongly Agree Agree Disagree Strongly Disagree Totally Disagree 0501-9de l. Prior to the Y.S.B. there were definitely l 2 3 4 5 6 sufficient outside agencies available to the police for juvenile referral. 2. Prior to the Y.S.B. the police were highly l 2 3 4 5 6 desirous of additional alternatives to the options of warning and release or referral to the court for juveniles. 3. Relevant representatives of the police l 2 3 4 5 6 definitely were in favor of creating a Y.S.B. 4. Relevant representatives of the police l 2 3 4 5 6 actively participated in the planning and develOpment of the Y.S.B. 5. Prior to the Y.S.B. there were definitely l 2 3 4 5 6 sufficient outside alternatives available to the schools for juvenile referral. 6. Prior to the Y.S.B. the schools definitely l 2 3 4 5 6 had sufficient internal alternatives for non-academic school related problems of juveniles. 7. Relevant representatives of the schools 1 2 3 4 5 6 definitely were in favor of creating a Y.S.B. 8. Relevant representatives of the schools 1 2 3 4 5 6 actively participated in the development and planning of the Y.S.B. 9. Prior to the Y.S.B. there were definitely l 2 3 4 5 6 sufficient alternatives to formal disposi- tion for the court. TOO 101 10. Prior to the Y.S.B. the court was highly l 2 3 4 5 6 desirous of additional alternatives to formal disposition for juveniles. ll. Relevant representatives of the court l 2 3 4 5 6 definitely were in favor of creating a Y.S.B. 12. Relevant representatives of the court l 2 3 4 5 6 actively participated in the planning and development of the Y.S.B. l3. Relevant representatives of other youth 1 2 3 4 5 6 service agencies actively participated in the planning and development of the Y.S.B. l4. Relevant representatives of other youth l 2 3 4 5 6 service agencies definitely were in favor of creating a Y.S.B. 15. After considering a variety of alternative l 2 3 4 5 6 plans the Y.S.B was selected as the preferred option. l6. As proposed the Y.S.B. filled a definite l 2 3 4 5 6 gap in the youth development and delin— quency prevention services of the community. The Y.S.B. was established to accomplish specific goals. Please rate each of the following goals as to the priority placed upon it during the planning and development of your Y.S.B. Indicate your choice by circling the corresponding number. l - High Priority 2 - Medium Priority 3 - Low Priority 4 - Not a Goal l7a. Diversion of status offenders from the court 1 2 3 4 b. Diversion of misdemeanants from the court l 2 3 4 c. Diversion of first offenders from the court l 2 3 4 d. Direct service/treatment l 2 3 4 102 Help modify the existing juvenile justice Provide service brokerage and referral for problem youth Provide focal point for the advocacy of youth and their problems in the community While implementing the Y.S.B. project, which of the issues below were problems which had to be overcome. Please rate each item on scale provided. #wNfl Not a Factor Highly Difficult to Overcome Moderately Difficult to Overcome Little Difficulty in Overcoming Goals not sufficiently defined Techniques to accomplish goals complicated or unclear Unrealistic goals Police Resistance Court Resistance School Resistance Resistance from relevant political office holders Community not sufficiently attuned to juvenile problems Communications problems between Y.S.B. and agencies which refer clients to it Lack of trust between Y.S.B. and agencies which refer clients to it Insufficiently trained personnel Insufficient resources N NNNNN w wwwww .b h-b-b-D-b 103 18m. Lack of technical assistance from D.C.J. l 2 3 4 Please list those factors which you believe have helped the Y.S.B. in terms of its getting established and operational. Then rate each factor on the scale provided. 1 - Highly Important 2 - Moderately Important 3 - Little Importance 19a 123 b 123 c 123 Below is a list of various Y.S.B. goals. Please rate each of them according to the priority you feel the Y.S.B. is currently placing on them. l - High Priority 2 - Medium Priority 3 - Low Priority 4 - Not a Goal 20a. Diversion of status offenders from the court I 2 3 4 b. Diversion of misdemeanants from the court l 2 3 4 c. Diversion of first offenders from the court 1 2 3 4 d. Direct service/treatment l 2 3 4 e. To help modify the existing juvenile justice l 2 3 4 system f. Provide service brokerage and referral for l 2 3 4 problem youth 9. Provide a focal point for the advocacy to l 2 3 4 youth and their problems in the community 104 Below is a list of Y.S.B. services. Please consider your Y.S.B.s operations and then indicate the percentage of their resources you feel they actually allocate to each activity. Your Y.S.B. may not attempt all of the activities listed so please provide answers only for those they are actively attempting. For example if you feel 60% of their resources are devoted to counseling and 40% to service coordination write those figures after your choices. The total should equal lOO%. Indicate % here 2la. Service coordination between agencies b. Family Counseling c. Individual short term counseling d. Individual long term counseling e. Psychotherapy type counseling f. Maintenance of a police contact file 9. Providing jobs for juveniles h. Referral of juveniles to appropriate outside agencies Using the space provided below write in those activites from the previous question which you have indicated your Y.S.B. is allocating a percentage of their resources to. Then, by circling the appro- priate answer, rate how well you feel they are doing with each activity. (write in only those you marked in the previous question). Very Hell Moderately Well Well Bad Moderately Bad Very Bad mm-wa-J 105 22c. l 2 3 4 5 6 d. l 2 3 4 5 6 e. l 2 3 4 5 6 f. l 2 3 4 5 6 g. l 2 3 4 5 6 h l 2 3 4 5 6 Below is a list of services offered by various Y.S.B.s. Please indicate what percentage of their resources you feel your Y.S.B. should be allocating to each one. Indicate only those you feel they should attempt. The total must equal lOO%. Indicate % here 23a. Service coordination between agencies b. Family Counseling c. Individual short term counseling d. Individual long term counseling e. Intensive psychotherapy type counseling f. Maintenance of a police contact file 9. Providing jobs for juveniles h. Referral to appropriate outside agencies i. Other (please specify) Since the Y.S.B. has been in operation your cooperative interactions with them have increased or decreased. (Circle one of the below that is most near applicable.) 24. Have continually increased since the establishment of the Y.S.B. 106 24. Initially increased but have now leveled off at the highest level Initally increased but have recently begun to decrease Initially increased but dropped off rapidly Never increased beyond occasional contact Never had operational contacts The Y.S.B. was completely Operational within how many months of grant initiation. (Circle the one that is most nearly accurate.) 25. Immediately upon grant initiation 1 - 2 months 3 - 4 months 5 - 6 months 7 months to 1 year Over 1 year Never has become fully operational Please circle the appropriate number indicating the extent of your agreement/disagreement with the following statements. Totally Agree Strongly Agree Agree Disagree Strongly Disagree Totally Disagree mmwa-d llllll 26. Juveniles in my agency referred to the 1 2 3 4 5 6 Y.S.B. always receive service within a reasonable period of time. 107 27. I have no problem contacting the appropri- l 2 3 4 5 6 ate Y.S.B. staff people whenever I need infbrmation and/or service from them. 28. Pe0ple from the Y.S.B. display a facili- 1 2 3 4 5 6 tative (helpful, cooperative) attitude toward your organization. 29. The individual Y.S.B. personnel were 1 2 3 4 5 6 equipped to handle the type of cases you referred to them from the beginning. 30. The training provided by the Y.S.B. for l 2 3 4 5 6 its personnel has improved their ability to deal with the type of cases you refer to them. 31. Through thecourt the Y.S.B. should have 1 2 3 4 5 6 the authority to impose sanctions upon those juveniles that do not cooperate with it. 32. I am definitely satisfied with the 1 2 3 4 5 6 feedback the Y.S.B. provides me. In the sapce provided below please list how the feedback provided by the Y.S.B. could be improved. Please circle the appropriate number indicating the extent of your agreement/disagreement with the following statements. 108 Totally Agree Strongly Agree Agree Disagree Strongly Disagree Totally Disagree mm-th—I I I l I I 34. The Y.S.B. director is well known and l 2 3 4 5 6 respected among those people relevant to the bureau. 35. The Y.S.B. Operations have definitely . 1 2 3 4 5 6 adhered to the grant proposal. 36. The Y.S.B. has definitely been able to 1 2 3 4 5 6 change to meet ongoing needs. 37. There are definitely goal/grant related 1 2 3 4 5 6 criteria for where the Y.S.B. is placed administratively. 38. The Y.S.B. has been responsible for a 1 2 3 4 5 6 reduction in recidivism among juveniles. 39. Officials (police and court) have used 1 2 3 4 5 6 the Y.S.B. to divert juveniles from the court. 40. Since the Y.S.B. began operations there 1 2 3 4 5 6 has been a reduction in the number of status offenses formally processed and petitioned to the court. 41. Since the Y.S.B. began operations there has 1 2 3 4 5 6 been a reduction in the number of non-status offenses formally processed and petitioned to the court. 42. The Y.S.B. has changed the type of juvenile 1 2 3 4 5 6 petitioned to the court. 43. The age of juveniles petitioned to the court has increased, decreased, stayed the same, since the Y.S.B. began operations. 109 44. The number of minor offenses petitioned to the court has increased, decreased, stayed the same, since the Y.S.B. began operations. 45. The number of major offenses petitioned to the court has increased, decreased, stayed the same, since the Y.S.B. began. Please circle the appropriate number indicating the extent of your 9 agreement/disagreement with the following statements. Totally Agree » Strongly Agree Agree ' Disagree 1 Strongly Disagree 1 Totally Disagree mm-wa—J IIIIII 46. There is definitely strong support within 1 2 3 4 5 6 your agency for continuing the Y.S.B. 47. Obtaining local financing was/is the 1 2 3 4 5 6 major blockage in continuing the Y.S.B. once its grant expires. 48. If necessary I feel money from my agency 1 2 3 4 5 6 would be well spent in support of the Y.S.B. 49. Y.S.B. administrators made/have every 1 2 3 4 5 6 effort to insure continued funding after its grant expires. 50. The Y.S.B. has definitely made my job 1 2 3 4 5 6 easier. APPENDIX B RESULTS OF PATH ANALYSIS FOR ALL VARIABLES ANALYZED 110 111 No. Variable B Std. Error B Sig. Beta 1. FAVOR to VAR073 .39319345 .11318414 .001 .4273102 FAVOR to VAR081 .72682984E-01 .12023700 .548 .0862978 VAR073 to VAR081 .27124504 .13069969 .043 .2962730 2. FAVOR to VAR074 .56000000 .11303181 .000 .5816766 FAVOR to VAR081 .9573505lE-Ol .13328680 .476 .1127131 VAR074 to VAR081 .47707790 .13844611 .001 .5407531 3. PARTIC to VAR070 .22317819 .69757614E-Ol .002 .3764298 PARTIC to VAR082 .92178924E-01 .67174399E-Ol .175 .1710995 VAR070 to VAR082 .30596838 .11330160 .009 .3367142 4. PARTIC to VAR070 .22047136 .77923383E-Ol .006 .3564482 PARTIC to VAR083 .40391918 .10066756 .000 .4903499 VAR070 to VAR083 .15798497 .16275478 .336 .1186269 5. PARTIC to VAR070 .22047136 .77923383E-01 .006 .3564482 PARTIC to VAR084 .21194341 .94983283E-Ol .030 .2994462 VAR070 to VAR084 .18659852 .15356470 .230 .1630658 6. PARTIC to VAR070 .20711801 .75856266E-01 .008 .3400931 PARTIC to VAR085 .37537677 .91515559E-Ol .000 .4920388 VAR070 to VAR085 .12947935 .15027089 .393 .1033600 7. PARTIC to VAR072 .27438411 .85240145E—Ol .002 .3922007 PARTIC to VAR081 .27611996 .73090075E-01 .000 .3879651 VAR072 to VAR08l .46641760 .10447390 .000 .4584800 8. PARTIC to VAR072 .23398026 .70715294E-01 .002 .3796714 PARTIC to VAR082 .42489538E-Ol .63556195E-Ol .506 .0833026 VAR072 to VAR082 .28747946 .10313037 .007 .3473398 9. PARTIC to VAR072 .25584168 .80539776E-Ol .002 .3907425 PARTIC to VAR083 .37182929 .95157184E-Ol .000 .4544188 VAR072 to VAR083 .32678300 .14533190 .029 .2618435 112 No. Variable B Std. Error B Sig. Beta 10. PARTIC to VAR072 .25694596 .77284435E-Ol .002 .4030181 PARTIC to VAR084 .22000170 .87192447E-01 .014 .3418663 VAR072 to VAR084 .65451024E-01 .13676080 .634 .0648431 11. PARTIC to VAR072 .27309237 .78937295E-01 .001 .4106707 PARTIC to VAR085 .20115142 .83072193E-Ol .019 .2818422 VAR072 to VAR085 .44439489 .12492224 .001 .4140644 12. PARTIC to VAR076 .44727808 .11929715 .000 .4416815 PARTIC to VAR081 .31257895 .79975905E-01 .000 .4253777 VAR076 to VAR081 .26825825 .78975294E-01 .001 .3696889 13. PARTIC to VAR076 .32952319 .10057583 .002 .3740217 PARTIC to VAR082 .66150201E-01 .6l702751E-01 .288 .1315719 VAR076 to VAR082 .18775826 .70035038E-Ol .009 .3290185 14. PARTIC to VAR076 .36279286 .11301628 .002 .3912866 PARTIC to VAR083 .37301705 .87984277E-01 .000 .4679739 VAR076 to VAR083 .26082065 .94894572E-Ol .008 .3033882 15. PARTIC to VAR076 .36247335 .1083211 .001 .4022667 PARTIC to VAR084 .16993731 .82672001E-Ol .044 .2581745 VAR076 to VAR084 .24293765 .91747957E-01 .010 .3325688 16. PARTIC to VAR076 .37386570 .11067946 .001 .4054494 PARTIC to VAR085 .26445534 .89822247E-01 .001 .3683570 VAR076 to VAR085 .17614130 .97410308E-01 .076 .2262335 BIBLIOGRAPHY 113 BIBLIOGRAPHY Bernard, Chester I. "Formal Organization." In Social Work Administration: A Resource Book. Edited by Henry A. Schatz. Council On SociaTPWork Education, 1970. Bennis, William G. Changing Organizations. New York: McGraw Hill Co., 1966. Berman, Paul and McLaughlin, Milbrey W. Federal Programs Supporting Educational Change: A Model of Educational Change. Santa Monica, California: The Rand Corp., 1975. Bertalanffy, Ludwig Von. General Systems Theory. New York: George Braziller Co., 1968. Campbell, Donald T. "Reforms as Experiments." American Psychologist, April, 1969. Chiru Robert. "The Utility of Systems Models and Development Models For Practitioners." In Management Systems. Edited by Peter D. Schoderbeck. New York: John Wiley and Sons, Inc., 1971. Churchman, C. West. The Systems Approach. New York: Dell Publishing Co., 1968. Clark, B. R. "Interorganizational Patterns in Education." Administrative Science Quarterly 10 (1965). Downie, N. M. and Heath, R. W. Basic Statistical Methods. New York: Harper and Row, Inc., 1965. Etzioni, Amitai. “New Directions in the Study of Organizations and Society." Social Research 27 (l960):223-28. Etzioni, Amitai. Modern Organizations. Englewood Cliffs, New Jersey: Prentice Hall, Inc., 1964. Evan, William M. "The Organization Set: Toward a Theory of Interorganizational Relations." In Approaches to Organiza- tional Design. Edited by James D. Thompson. Pittsburgh, Pennsylvania: University of Pittsburgh Press, 1966. Gross, Bertram. The Management of Organizations. New York: Glencoe Free Pressjil969. 114 115 Gross, Neal; Giacquinta, Joseph B.; and Berstein, Marilyn. Implementing Organizational Innovations. New York: Basic Books, Inc., 19 1 Katz, Daniel and Kahn, Robert L. The Social Psychology of Organiza- tions. New York: John Wiley and Sons, Inc., 1966. Kim, Jae-On and Kahout, Frank J. "Special Topics in General Linear Models." In Statistical Package for the Social Sciences. Edited by Norman H. Nie. McGraw Hill, Inc., 1975. Lawrence, Paul and Lorsch, Jay W. Organization and Environment. Homewood, Illinois: Robert D. Irwin, Inc., 1969. Leedy, Paul. Practical Research: Planning and Design. New York: Macmillan Publishing Co., Inc., 1975. Levine, $01 and White, Paul E. "Exchange as a Conceptual Framework for the Study of Interorganizational Relationships." Administrative Sciepce Quarterly 5 (1960). Lewis, Ralph G. and Green, Jack. "Environmental Effects on Project Implementation." Paper presented to the National Conference on Criminal Justice Evaluation. Washington, 22 October 1975. Marrett, Bageley. "0n the Specification of Interorganizational Relationships." Sociologypand Social Research 61 (1971). Negandhi, Anant R., ed. Interorganization Theory. Kent, Ohio: Kent State University Press, 1975. Pressman, Jeffrey and Wildavsky, Aaron. Implementation. Los Angeles: University of California Press, 1973. Rogers, Everett M. Diffusion of Innovations. New York: Glencoe Free Press, 1962. Starbuck, William. "Organizational Growth and Development." In Handbook on Organizations. Edited by James March. Chicago: Rand McNally, 1965. Thompson, James 0. "Organizational Goals and Environment." American Sociological Review 23 (1968). Van de Ven, Andrew H.; Emmett, Dennis C.; and Keonig, Richard. "Frameworks for Interorganizational Analysis." In Inter- organization Theory. Edited by Anant Negandhi. Kent, Ohio: Kent State University Press, 1975. 116 Williams, Walter and Elmore, Richard. Social Program Implementation. New York: Academic Press, 1976. Wolman, Harold. "Organization Theory and Community Agencies." Public Administration Review (January-February 1972).