University of Southern California Dissertations and Theses

Adoption and implementation of evidence-based tobacco use prevention curricula and programming

(USC Thesis Other)

Adoption and implementation of evidence-based tobacco use prevention curricula and programming

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content ADOPTION AND IMPLEMENTATION OF EVIDENCE-BASED TOBACCO USE PREVENTION CURRICULA AND PROGRAMMING by Silvana Nicolle Skara A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree of DOCTOR OF PHILOSOPHY (PREVENTIVE MEDICINE - HEALTH BEHAVIOR RESEARCH) August 2004 Copyright 2004 Silvana Nicolle Skara Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UMI Number: 3145288 INFORMATION TO USERS The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleed-through, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. ® UMI UMI Microform 3145288 Copyright 2004 by ProQuest Information and Learning Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code. ProQuest Information and Learning Company 300 North Zeeb Road P.O. Box 1346 Ann Arbor, Ml 48106-1346 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ACKNOWLEDGEMENTS ii This dissertation is the result of the collective efforts of a number of important and valued people who have directly or indirectly assisted and supported me during my doctoral studies and in this present endeavor. To all of these people, I owe my gratitude and thanks. Most of all, I would like to thank my supervisor and mentor, Dr. Luanne Rohrbach, for the many years of guidance, encouragement and patience which I have received. Without her knowledge and support, none of this was possible. I am fortunate to follow in her footsteps in advancing diffusion research in the area of substance use prevention programs for youth. I would like to thank Dr. Steve Sussman for providing inspiration, guidance and compassion throughout my graduate school experience, and, in particular, at times when I needed it most. His intelligence, creativity and his passion for research, truly inspire me. I am indebted to Drs. Mary Ann and Cappy Pentz for encouraging and facilitating my decision to pursue a career in research, starting back when I was completing my undergraduate studies and consistently throughout my MPH and doctoral studies. Both have been the needed source of optimism and support. I am grateful to all of the participating faculty and staff members at the University of Southern California who gave their time and provided their expertise. In particular, I thank Dr. Andy Johnson for providing leadership at the Institute for Health Promotion and Disease Prevention Research, and all of the other faculty Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. iii members who provided valuable training, knowledge and advice, including Drs. Chih-Ping Chou, Jennifer Unger, Clyde Dent and James Dwyer. I thank Mamy Barovich and Jolanda Lisath for all of their administrative help ever since I started graduate school. I also thank Dr. William McComas from the School of Education for agreeing to be on my dissertation committee, and for his interest in my research. I am very happy to thank my wonderful graduate school friends Drs. Cherilyn McMonigle, Michelle Weiner, Cheryl Nordstrom and Anamara Ritt-Olsen, as well as Darleen Schuster and Pam Elfenbaum (almost there!) for sharing the experience with me. I am grateful for their constant camaraderie and support. These friendships are very important to me. I am also grateful to all of the other USC friends who have provided assistance and encouragement, including Drs. Jenny Zogg, Terry Huang, Dennis Trinidad, Chao Yang Li, Amy Fan, Dongyun Yang, Sondos Islam and Joel Milam. I am grateful for the support from my longtime friends Sylvana Gusich, Araks Bagramyan, Jovie Bernardino, Staci Goodman, Don Smith and Samantha Simmons. Their constant faith and support is always greatly appreciated. I thank my entire family for their everlasting support. From the very beginning, they instilled in me the pride and confidence that I needed to accomplish every goal that I have ever set for myself. Finally, I would like to thank the Tobacco-Related Disease Research Program at the University of California for their financial support of my doctoral research (dissertation research award #11DT-0137). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. iv TABLE OF CONTENTS ACKNOWLEDGEMENTS........... ............ ii LIST OF TABLES.... ................. ..............................vii LIST OF FIGURES.............................................................. ....................................x ABSTRACT ........ xi INTRODUCTION..................... 1 Background .................................. 8 Literature Review. ...................... 13 Theoretical Frameworks Guiding Model Development................. ..14 Diffusion of Innovations Theory. .......14 Social Psychological Theories................................................19 Organizational Change Theories. ...................25 Empirical Research Results Guiding Model Development. ....... 29 Organizational Factors.................... ....30 Provider Factors ..................... . 3 8 Curriculum Factors ...................................... 3 9 Demographic Factors. .................. .40 Proposed Theoretical Model.................... 40 Hypotheses......................... .46 METHODS..... ...... 49 Overview. .............. 49 Independent Evaluation Overview, Design, and Sampling. ......49 Study One: Cross-sectional Study................ 55 Design and Sample... ................. ....55 Classroom Teacher Data. .................... 56 School Principal Data. ....... 57 Analytic Sample. ............................ .58 Measures.. ...... 58 Dependent Variables. ..................................... 59 Independent Variables..................... 66 Data Analysis Plan. .................... ..76 Data Analysis Overview. .................................. .76 Attrition Analyses. ................. ....77 Sample Characteristics ....77 Descriptive Analyses of Variables.. .......................... 78 Hypothesis Testing. .......... 78 Statistical Power. ............................... 81 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. V Study Two: Longitudinal Study ...... 81 Design and Sample ........................................81 Analytic Sample.. ........... 81 Measures... .......... 82 Dependent Variables. ...... 82 Independent Variables. ................... 84 Data Analysis Plan. ..................... .89 Data Analysis Overview ............ 89 Attrition Analyses ........................... .90 Sample Characteristics............................ .90 Descriptive Analyses of Variables.. ........................90 Hypothesis Testing. .................................91 Statistical Power. .......... ...91 RESULTS.... ...... .............................93 Overview. ...... 93 Study One: Cross-sectional Study................... ........................93 Attrition Analyses......................... .95 Sample Characteristics ..................... .95 School Data............... .........95 Teacher Data............................. 96 Principal Data. ..................... 99 Descriptive Analyses of Variables ...... ....101 Descriptive Analyses of Dependent Variables 101 Descriptive Analyses of Independent Variables .104 Correlations among All Analytic Variables..........109 Hypothesis Testing............................... 118 Adoption and Implementation of Evidence-based Tobacco Use Prevention Curricula. ...........119 Adoption and Implementation of Evidence-based Tobacco Use Prevention Programming .128 Study Two: Longitudinal Study.................. 138 Attrition Analyses.. .....................................138 Sample Characteristics. ...............................................140 School Data......................... ....140 T eacher Data .........................................141 Student Data. ............ .....144 Descriptive Analyses of Variables. ...... 145 Descriptive Analyses of Dependent V ariables .......145 Descriptive Analyses of Independent Variables 148 Correlations among All Analytic Variables. 152 Hypothesis Testing .................................. 160 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. vi DISCUSSION. ...... ....169 Overview. ...... 169 Study One; Cross-sectional Study............................................... 170 Summary of Results for Study O ne................................... 180 Study Two: Longitudinal Study. .......183 Summary of Results for Study Two.......... 186 Limitations ...... 187 Implications and Future Directions ......................... .....192 REFERENCES................. 209 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. vii LIST OF TABLES 1. Table 1. CDC Guidelines for School Health Programs to Prevent Tobacco Use and Addiction (1994). ...... ....11 2. Table 2. Outline of Proposed Factors Associated with the Adoption and Implementation of Evidence-based Tobacco Use Prevention Curricula and Programming in Schools.. ......................................... 42 3. Table 3. Study One: Prevalence of Use of Tobacco Use Prevention Curricula for Junior/Middle and High School Grades (n=75 Schools)....................... ..61 4. Table 4. Study One: Comparison of Characteristics of Schools Used in Analyses vs. Schools Not Used in Analyses ..... 95 5. Table 5. Study One: Demographic Characteristics of the Analytic School Sample (n=75 Schools). .................. ....96 6. Table 6. Study One: Demographic Characteristics of the Analytic Teacher Sample (n=75). ................................................. 98 7. Table 7. Study One: Demographic Characteristics of the Analytic Principal Sample (n=75). ....... 100 8. Table 8. Study One: Means (SD) and Percentages for Variables Used in Adoption and Implementation of Evidence-based Tobacco Use Prevention Programming (Continuous) Composite Index (n=51 Schools) ...... 103 9. Table 9. Study One: Means and Standard Deviations of All Analytic Correlates of Adoption and Implementation of Evidence-based Tobacco Use Prevention Curricula (n=75 Schools) ...... 106 10. Table 10. Study One: Intercorrelations of All Analytic Variables (n=75 Schools)................................................ ....Ill 11. Table 11. Study One: Univariate Logistic Regression Results of Factors Associated with the Adoption and Implementation of Evidence-based Tobacco Use Prevention Curricula in Schools. ..... ................ 120 12. Table 12. Study One: Multivariate Logistic Regression Results of Factors Associated with the Adoption and Implementation of Evidence-based Tobacco Use Prevention Curricula in Schools.................................. 123 13. Table 13. Study One: Multivariate Logistic Regression Results of Factors Associated with the Adoption and Implementation of Evidence-based Tobacco Use Prevention Curricula in Schools. .......................... ...125 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. v iii 14. Table 14. Study One: Multivariate Logistic Regression Results of Hypothesized Interactions Associated with the Adoption and Implementation of Evidence-based Tobacco Use Prevention Curricula in Schools............ 127 15. Table 15. Study One: Univariate Linear Regression Results of Factors Associated with the Adoption and Implementation of Evidence-based Tobacco Use Prevention Programming in Schools. ...... 130 16. Table 16. Study One: Multivariate Linear Regression Results of Factors Associated with the Adoption and Implementation of Evidence-based Tobacco Use Prevention Programming in Schools................... .....133 17. Table 17. Study One: Multivariate Linear Regression Results of Factors Associated with the Adoption and Implementation of Evidence-based Tobacco Use Prevention Programming in Schools .................... 135 18. Table 18. Study One: Multivariate Linear Regression Results of Hypothesized Interactions Associated with the Adoption and Implementation of Evidence-based Tobacco Use Prevention Programming in Schools 137 19. Table 19. Study Two: Comparison of Characteristics of Schools Used in Analyses vs. Schools Not Used in Analyses. ......................... ....139 20. Table 20. Study Two: Demographic Characteristics of the Analytic School Sample (n=70 Schools) ...... ...141 21. Table 21. Study Two: Demographic Characteristics of the Analytic Teacher Sample (n=97 Teachers at 70 Schools).............................. 143 22. Table 22. Study Two: Demographic Characteristics of the Student Sample (n=6921 Students at 70 Schools) ............................ ..145 23. Table 23. Study Two: Means (SD) and Percentages of Adolescent Tobacco Use-related Program Outcome Variables, 1996 and 2000 (n=70 Schools).. ............. ....147 24. Table 24. Study Two: Means (SD) and Percentages for Variables Used in Implementation of CDC Guidelines Index (n-70 Schools). ............ 149 25. Table 25. Study Two: Means and Standard Deviations for Variables used in School Capacity to Implement Innovative Programming Index (n=70 Schools). ...... 151 26. Table 26. Study Two: Intercorrelations of All Analytic Variables (n=70 Schools).............. ..153 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ix 27. Table 27. Study Two: Effects of Implementation of Evidence-based Tobacco Use Prevention Programming Index on 2000 Adolescent Tobacco Use- related Program Mediators and Behavioral Outcomes. ........ 162 28. Table 28. Study Two: Effects of School Capacity to Implement Innovative Programming Index on 2000 Adolescent Tobacco Use-related Program Mediators and Behavioral Outcomes. ................. 165 29. Table 29. Study Two: Effects of Interaction Models (Implementation of Evidence-based Tobacco Use Prevention Programming Index X School Capacity to Implement Innovative Programming Index) on 2000 Adolescent Tobacco Use-related Program Mediators and Behavioral Outcomes...........167 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. X LIST OF FIGURES 1. Figure 1. Heuristic Model of Diffusion of Tobacco Use Prevention Curricula and Programming in Schools.. ........ 3 2. Figure 2. The Innovation-Decision Process Adapted from Rogers (1995, p. 163)...... .....15 3. Figure 3. Proposed Conceptual Model of the Adoption and Implementation of Evidence-based Tobacco Use Prevention Curricula and Programming in Schools ...... .........45 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. xi ABSTRACT Despite the availability of empirically-tested school-based tobacco use prevention curricula and new governmental policies mandating their use, a very small proportion of schools are implementing proven, effective tobacco use prevention programming. Thus, this dissertation investigated the process of diffusion of evidence-based tobacco use prevention programming through use of two studies that were comprised of two separate samples. Data for both studies were obtained from schools that participated in the large-scale Independent Evaluation study of the California Tobacco Control Program. Study 1 examined the adoption and implementation phases, by focusing on identification of variables (i.e., organizational, provider, curriculum and demographic) associated with the adoption and implementation of evidence-based tobacco use prevention curricula and programming. This first study utilized a cross- sectional sample comprised of both teachers and principals. Study 2 focused on the adoption and implementation phases, along with the program outcome phase, by examining the effect of an interaction between a newly-developed school capacity to implement innovative programming construct (derived from organizational-, provider- and curriculum-level variables) and the degree of implementation of evidence-based tobacco use prevention programming, on changes in adolescent tobacco use-related beliefs, attitudes, skills and behaviors. This second study utilized a longitudinal sample (at the school level) comprised of both teachers and students. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. xii Study l ’s results provided support for the primary hypothesis that variables from the organizational, provider, curriculum and demographic domains would be associated with the adoption and implementation of evidence-based tobacco use prevention curricula and programming. Organizational-level variables, particularly shared vision/goals, principal knowledge of effective tobacco use prevention programs, participatory decision-making, and adequate resources, showed the strongest associations. Study 2’s results provided minimal support for the main hypothesis that school capacity to implement innovative programming would have a significant moderating effect on the relationship between implementation of evidence-based tobacco use prevention programming and program outcomes. A significant moderating effect was found for decreases in students’ estimates of the proportion of peers who smoke monthly. The results of this study have important implications for diffusion of innovations in school settings, including governmental and school district-level policy decisions regarding tobacco use prevention education in our nation’s schools. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 INTRODUCTION It has been estimated that, for the last ten years, California has operated an annual $90 million budget for tobacco control, with one-third of the funds going towards school-based tobacco use prevention education (California Department of Health Services, 1998). The key component of the state’s tobacco control approach in school settings is the implementation of tobacco use prevention curricula delivered by classroom teachers to students. Numerous studies have shown that research-based tobacco use prevention curricula can be effective in producing consistent and significant reductions or delays in adolescent smoking (Botvin et al., 1995; Ellickson et al., 1993; Flay et al., 1985b; Pentz et al., 1989; Sussman et al., 1995). However, recent data indicate that a veiy small proportion of schools are implementing proven, effective tobacco use prevention programming (Centers for Disease Control, 1998; National Institute of Drag Abuse, 1997; Ringwalt et al., 2002; Rohrbach, Skara, Unger, et al., 2001; U.S. Department of Education, 2001). In fact, very little is known about the actual programs that are being implemented in our schools, and even less information exists on the diffusion of evidence-based tobacco use prevention curricula from researchers to the school districts, principals, teachers, and ultimately, the students. Purpose of the Present Investigation Despite the availability of empirically-tested school-based tobacco use prevention curricula and new governmental policies mandating their use, these validated preventive measures will not achieve maximum potential impact for Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 reducing smoking incidence and prevalence among youth unless they are widely disseminated and properly implemented. Yet, clearly there exists a gap in the communication channels between the scientific community and the provider community. Research advances have not been transmitted to schools in such a way that routine adoption of evidence-based curricula into comprehensive school-based programs can be assumed. To date, however, the diffusion process has not been adequately investigated, especially within the context of school-based drug use prevention programming. Thus, this dissertation proposal aimed to investigate the diffusion process through use of two studies (as presented in Figure 1). Study 1 examined the adoption and implementation phases, by focusing on correlates of the process that occurs within schools when a decision is made to adopt and implement a new program, specifically tobacco use prevention curricula and programming that are evidence- based. Study 2 examined the extent to which the association between implementation of evidence-based tobacco use prevention programming and changes in student program outcomes was moderated by the school’s capacity to implement innovative programming. For the purposes of this study, evidence-based tobacco use prevention programming is defined as comprehensive school-based programming practices that are based on research about the elements of effective programs, such as tobacco use policy, curriculum delivery, and teacher training. Please refer to the Guidelines for School Health Programs to Prevent Tobacco Use and Addiction, published by the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 Centers for Disease Control and Prevention (1994; see Table 1 on page 11 of this dissertation). Figure 1. Heuristic Model of Diffusion of Tobacco Use Prevention Curricula and Programming in Schools Dissemination Implementation by School Adoption by District and School Student Outcomes Organizational Characteristics Provider Characteristics Curriculum Characteristics Focus of Study 1 i i Focus o f Studv 2 The primary goal of this research was propose a socio-ecological conceptual model of the variables likely to influence the adoption and implementation processes; and examine the proposed model’s applicability to middle and high schools located across the state of California. This investigation builds upon diffusion research that has investigated various factors associated with the adoption and implementation of effective research-based substance use prevention curricula and programming. Although research in this specific area is relatively limited, results have indicated that some organizational, provider, curriculum and demographic Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 characteristics may be related to the selection and use of evidence-based curricula. Recent research conducted by Ennett et al. (2003) has shown that teachers who had received recent training in substance use prevention or were comfortable using interactive teaching methods in their delivery of programming were three to four times more likely to implement effective substance use prevention programming. Another study conducted by Hallfors and Godette (2002) indicated that school districts that had implemented the U.S. Department ofEducation’s “Principles of Effectiveness” policy were more likely to select a research-based program compared to school districts that had not implemented the policy. Rohrbach and colleagues’ (in press) research has provided cross-sectional evidence demonstrating that school districts that had consulted informational materials obtained from credible sources such as federal and state substance use prevention agencies (e.g., National Institute on Drug Abuse and Center for Substance Abuse Prevention) were significantly more likely to decide to adopt an evidence-based substance use prevention curriculum. Additional research conducted by Ringwalt et al. (2002) found that public middle schools were three times more likely to use effective programs compared to private schools. Further, middle schools that were located in the Northeast and Midwest, compared to the South and West, were also more likely to implement effective curricula. Other school and teacher background correlates included higher school poverty and a higher percentage of state teacher certification. Based on this information, the present study investigated new relationships that are unique to this study, but build upon previous research conducted by Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5 Rohrbach & Skara (in preparation; 2003) that has examined factors affecting the implementation of evidence-based tobacco use prevention practices in schools. Specifically, as part of Study 1 of this dissertation, this investigation utilized primary cross-sectional data to: (a) Examine additional theory-based organizational-, provider- and curriculum- and demographic-level constructs that are hypothesized to be associated with the adoption and implementation of evidence-based tobacco use prevention curricula and programming; and (b) examine the interactive relationships between organizational-, provider, and curriculum-level factors hypothesized to influence the adoption and implementation of evidence-based tobacco prevention curricula and programming by schools. The second goal of this research was to examine the extent to which the effects of implementation of evidence-based tobacco use prevention programming on subsequent changes in adolescent tobacco-related outcomes was moderated by organizational-, provider- and curriculum-level constructs, or a composite index of the school’s capacity to implement innovative programming. Thus, this study extends previous research conducted by Rohrbach and colleagues (Rohrbach, de Calice et al., 2001, 2002; Rohrbach, Unger et al., 2002) that has shown a positive relationship between the implementation of the effective elements of tobacco use prevention programming, as specified by guidelines published by the Centers for Disease Control and Prevention (CDC; Centers for Disease Control and Prevention, 1994), and changes in adolescent tobacco use knowledge, beliefs, attitudes, skills and behaviors. Study results obtained by Rohrbach and colleagues indicated that higher Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 levels of implementation of effective program elements predicted significant improvements in students’ beliefs about the negative consequences of tobacco use, as well as their tobacco-related knowledge. The prevalence of 30-day cigarette smoking was also found to be marginally significant (p<06). These results provided the basis for hypothesized relationships in the present investigation. Specifically, as part of Study 2 of this dissertation, this paper utilized secondary longitudinal data that were collected from 1996 to 2000 as part of the Tobacco Use Prevention Education (TUPE) component of the Independent Evaluation (BE) of the California Tobacco Control Program (Independent Evaluation Consortium, 1998, 2001, 2003). The present study contributes new knowledge to this particular area of investigation because it was designed to: (c) Examine the importance of the interactive effect between a newly-developed school capacity to implement innovative programming construct (derived from organizational-, provider- and curriculum-level variables) and the degree of implementation of evidence-based tobacco use prevention programming, with subsequent changes in adolescent tobacco use-related knowledge, attitudes and behaviors. Significance of this Study Given the size of investment in school-based tobacco use prevention education across the nation, the evidence that specific curricula can reduce adolescent tobacco use, and the anticipated increase in demand by governmental agencies for the use of evidence-based curricula, it is necessary to increase our understanding of the factors associated with adoption and implementation of Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 evidence-based programs. The results of this investigation may be useful to various stakeholders including: (1) governmental policymakers involved in drug use prevention research and practice; (2) school personnel who occupy leadership positions, such as administrators and principals; and (3) curriculum developers and distributors. In particular, the findings from this study should be useful in informing policymakers regarding the need and the extent to which policies and guidelines need to be developed and implemented to control and increase the use of effective tobacco prevention programs across all school districts. School leaders with decision-making power may also gain insight from the current description of faculty adoption and implementation patterns of the programs currently used in their schools. Results may suggest whether there are major factors that are amenable to change that may promote or limit the selection and use of effective programs by principals and teachers. This information may also be usefiil to school personnel who are interested in further developing their skills and knowledge about teaching and integrating innovative technology in their schools. Program developers may use information about the correlates of program adoption and implementation to develop more appealing programming content features, materials, and teaching techniques, including possible changes in the training and education needed by teachers to implement programs that are effective. Program distributors may also learn whether more appropriate packaging and marketing techniques, such as highlighting specific Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8 advantages of the new program, are needed to increase the diffusion of promising prevention curricula. BACKGROUND Tobacco use remains the single major preventable cause of death and disease in the United States. According to the United States Department of Health and Human Services (2000), tobacco smoke is responsible for more than 400,000 deaths each year, resulting in more than $50 billion in direct medical costs. Despite the adverse health outcomes and enormous costs associated with smoking, nearly one- quarter of adult Americans—an estimated fifty million people—continue to smoke cigarettes. In addition, it has been estimated that more than 3,000 young people, most of them children and adolescents, begin smoking each day. Research indicates that at least seventy percent of these current smokers want to quit smoking, and each year about half of them make a serious quit attempt, but only a very small percentage (approximately five percent) succeed (U.S. Department of Health and Human Services, 2000). In an effort to reduce the continuing toll of tobacco use by discouraging smoking initiation and promoting smoking cessation, California was the first state to establish a tobacco control program funded by an increased tax. In 1988, Proposition 99 passed and raised the tax by 25 cents on each pack of cigarettes sold in the state. The California Tobacco Control Program (CTCP) involves the media, health departments, community-based organizations and schools as part of a statewide effort designed to change the social and legal environments so that tobacco use is Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 9 considered unacceptable to children and young adults. It has been estimated that for the last ten years California has operated an annual $90 million budget for tobacco control, with one-third of the funds going towards school-based tobacco use prevention education (California Department of Health Services, 1998). Schools have served as a key component and primary setting for these tobacco use prevention education efforts because the vast majority of smokers begin using tobacco before the age of 18 (United States Department of Health and Human Services, 1989). Thus, the California Department of Education established the Tobacco Use Prevention Education (TUPE) Program to provide funding and assistance to schools and districts for the specific purpose of reducing tobacco use through utilization of proven, research-tested prevention practices delivered by classroom teachers to students. Numerous individual studies have shown that school-based curricula that are evidence-based have been effective in producing consistent and significant reductions or delays in adolescent smoking (Botvin et al., 1995; Ellickson et al., 1993; Flay et al., 1985b; Pentz et al., 1989, Sussman et al., 1995). The most widespread educational prevention approaches implemented through the school system are delivered universally (i.e., prevention programming delivered to an entire population regardless of risk status) and based on curricula designed to counteract the psychosocial influences that promote tobacco use initiation. The two major psychosocial approaches that have been adopted by schools are the social influences approach (Evans, 1976) and the more comprehensive personal and social skills Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 10 enhancement (e.g., Life Skills Training, Project Towards No Tobacco Use) strategy (Botvin et al., 1980; Sussman et al., 1995). Social influences programs are designed to increase the awareness of social influences promoting drug use, alter norms regarding the prevalence and acceptability of drug use, and build drug resistance skills. Skills enhancement programs incorporate aspects of the social influence approach and also include general self-management and social competence skills. Literature reviews (Flay, 1985a; Hansen, 1992; Perry & Kelder, 1992) and meta-analyses (Bruvold, 1993; Tobler & Stratton, 1997; Tobler et al., 2000) of these social influences programs have indicated short-term (under 24 months) reductions in the rate of initiation of tobacco use generally ranging from 30% to 50% or more in students exposed to social influences programs compared to control students. A few targeted programs, implemented with older teens have found over 25% prevalence reductions in tobacco use lasting up to two years post-program (Sussman et al., 2003). These programs take more of a skills enhancement approach and also incorporate motivation enhancement material to effect changes in personal attitudes that may impede skills development and behavior change. In addition, in both universal and targeted programs, effective educational strategies are those that are implemented interactively rather than with a didactic lecture-style (Tobler et al., 2000). Although much less evidence exists for the long-term follow-up success of these tobacco use interventions, current empirical data indicate reductions ranging from approximately 10% to 15%—lasting for up to 15 years after programming (Skara & Sussman, 2003). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 11 Currently, many such empirically-validated programs are available to schools, and have been endorsed and promoted for use through federal and state policies. For example, in California, schools that receive Tobacco Use Prevention Education (TUPE) funds from the state’s tobacco control program are required to design their prevention programming based on the Guidelines for School Health Programs to Prevent Tobacco Use and Addiction, published by the Centers for Disease Control and Prevention (1994; California Department of Education, 2000). These guidelines are based on research about the elements of effective programs, and specify the principles as presented in Table 1. Table I. CPC Guidelines for School Health Programs to Prevent Tobacco Use and Addiction (1994) 1. Develop and enforce a school policy on tobacco use. 2. Provide instruction about the short- and long-term negative physiologic and social consequences of tobacco use, social influences on tobacco use, peer norms regarding tobacco use, and refusal skills. 3. Provide tobacco-use instruction in kindergarten through 12th grade; this instruction should be especially intensive in junior high/middle school and should be reinforced in high school. 4. Provide program-specific training for teachers. 5. Involve parents or families in support of school-based tobacco use prevention programs. 6 . Support cessation efforts among students and all staff who use tobacco. 7. Assess the tobacco use prevention program at regular intervals. However, it appears that these “proven” practices and/or programs still have not been widely adopted by schools in California or across the nation. Instead, a large majority of schools are implementing tobacco prevention programs that are Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 12 aggressively marketed and relatively easy to implement, but have not shown evidence of effectiveness and have not been property evaluated (U.S. Department of Health and Human Services, 2000). For example, data from the 2000 evaluation of the Tobacco Use Prevention Education (TUPE) Program in California’s schools indicate that the majority of school districts were using programs that have not been identified as “exemplary” or “promising” by expert panels, such as Here's Looking at You, 2000 (60%); materials developed by the American Cancer Society (43%), American Heart Association (27%), and American Lung Association (27%); and locally developed materials (58%). Programs that have been endorsed by the experts were less common, such as Project ALERT (26%), Life Skills Training (15%), and Quest: Skills fo r Adolescence (18%) (Rohrbach, Skara, Unger, et al., 2001; California Department of Education, 2000). The widespread use of ineffective or unevaluated programming may account, in part, for the lack of decline in smoking prevalence among youth in California during the early and mid-1990’s. Despite the resources currently expended on empirically-validated school-based tobacco use prevention programs, these programs will not be successful without accompanying research on dissemination, adoption and implementation (Elias, 1997). The effective diffusion of evidence-based tobacco use prevention curricula and programming from prevention scientists to the classrooms is essential to improve adolescent tobacco use outcomes. Yet, clearly there exists a gap in the communication channels between the scientific community and the provider community. Research advances have not been transmitted to Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 13 schools in such a way that routine integration of evidence-based curricula into comprehensive school-based programs can be assumed. An understanding of a key piece of this problem, the process by which research findings are disseminated, adopted and implemented in classrooms, has not been adequately investigated, especially within the context of school-based drug prevention programming (Pentz, 2003; Pentz et al., 1990). In order for innovative educational programs to be diffused successfully and achieve maximum impact, the factors that influence the adoption and implementation process need to be well understood. These factors will provide information about what characteristics are essential to adoption, implementation and institutionalization of innovations. LITERATURE REVIEW Although the literature related to innovations introduced into organizations is large and diverse, efforts towards the scientific study of diffusion have largely relied on the theoretical perspectives related to social change (i.e., social behavior and its change within organizations). This body of literature has evolved over the last 100 years from different disciplines, each pursuing their own variables of interest within specific organizational contexts, in an attempt to understand the factors that inhibit and facilitate the process of planned or deliberate organizational change. Consequently, fragmented findings from various studies utilizing differing social science perspectives have not yet coalesced into a unified and comprehensive body of tested propositions that can be applied to program adoption and implementation across various real world settings. Nonetheless, taken together, the numerous Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 14 propositions, models and findings from these disparate studies can serve as the basis of a new conceptual model that may offer an understanding of the processes involved in the adoption and implementation of innovations in schools. Several perspectives have been identified that are most relevant to understanding change within organizations. Specifically, theories of diffusion, social psychology, and organizational development are fundamental to our understanding of how schools improve and systems change. This section will review each perspective and its related findings (in terms of their application to schools), and examine the potential contribution of each to an integrated conceptual model of adoption and implementation. The first two groups of perspectives on change in organizations tend to view the adoption process through the eyes of individual adopters—the people who will be affected by the change. The remaining perspective, organizational change, shifts the focus from the individual user within an adopting organization to a broader approach that includes both user and organizational viewpoints—both of which are intimately linked within a comprehensive social system. Theoretical Frameworks Guiding Model Development Diffusion of Innovations Theory Everett Rogers’ Diffusion of Innovations theory has been the definitive source used to explain and predict the diffusion of innovations (Rogers, 1983, 1995), and thus, served as a starting point in building a theoretical framework. Over the last five decades, the diffusion paradigm has been applied to a number of diverse fields, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 15 such as sociology, anthropology, geography, marketing, communication, public health and education. This approach is based on well-established theories in sociology, psychology, and mass communications, and it offers a concise and easily understood approach to a potential adopter’s acceptance of innovations. The diffusion framework is concerned with the communication of an innovation through certain channels among members of a social system over time. Focus was placed on the theoretical dimension most relevant to this paper, namely, the innovation- decision process. Rogers’ model of the innovation-decision process is depicted in Figure 2 (See Appendix A). The innovation-decision process is “the process through which an individual (or other decision-making unit) passes from first knowledge of an innovation, to forming an attitude toward the innovation, to a decision to adopt or reject, to implementation of the new idea, and to confirmation of this decision” (Rogers, 1995, p. 163). Communication channels that provide information and feedback from one stage to the others link all these stages together. The major factors that influence the diffusion process include how information about the innovation is communicated, the innovation itself, and the nature of the social system into which the innovation is being introduced (Rogers, 1995). The model delineates a five-stage process from the perspective of those who adopt and implement an innovation. According to this model, potential users (individuals or other decision-making units, hereafter referred to as “an individual”) move through five stages in their decision to adopt or reject an innovation. First, the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 16 individual is exposed to an innovation’s existence and obtains some knowledge and understanding of how it functions (Stage 1, knowledge). Types of knowledge include awareness about the innovation, how to use the innovation properly, and how the innovation works. Prior characteristics and conditions of the individual, including factors such as previous practice, perceived needs or problems, innovativeness and norms of the social system, will affect the outcome of the knowledge stage. Innovativeness is the degree to which an individual is relatively earlier in adopting new ideas than others in a social system. Innovativeness is often affected by the characteristics of the individual, which include the socioeconomic characteristics, personality variables and communication behavior. At the next stage (Stage 2, persuasion), the individual gathers further information and develops a favorable or unfavorable attitude or opinion of the innovation. The outcome of this stage will be affected by the perceived characteristics of the innovation, such as (a) the relative advantages of the innovation over existing programs, (b) its compatibility with existing values, experiences and needs of the potential adopter (c) the complexity of implementation, (d) its trialability, i.e., the possibility of testing the innovation on a limited basis, and (e) its observability, i.e., the degree to which the results of the innovation are visible to others. Innovations that are perceived to have less complexity but greater relative advantage, compatibility, trialability and observability, will be adopted more rapidly. Persuasion is also influenced by information from interpersonal networks and peers or competitors whose subjective opinion is most convincing. When someone who is Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 17 similar to the individual shares a positive evaluation of the innovation, the individual is more motivated to adopt it. Eventually, the individual makes a decision to adopt or reject the innovation (Stage 3, decision) through a series of deliberations. Adoption is a decision to make M l use of the innovation as the best course of action available. Active rejection means considering and trying the innovation out on a limited basis, and deciding not to adopt. Passive rejection, also called non-adoption, consists of never really considering the use of the innovation. Implementation occurs when the individual actually carries out the decision and uses the innovation to produce any benefits intended (Stage 4, implementation). Until this stage the process has been a mental exercise. This stage may continue for a lengthy period of time until the innovation finally loses its distinctive quality as a new idea. Re-invention, the degree to which an innovation is modified by the user, can also occur at this stage. The experience gained from the implementation stage, along with other information that flows from various communication channels help the individual to confirm or change the decision during the subsequent confirmation stage. In this stage (Stage 5, confirmation), the individual seeks reinforcement of an innovation-decision already made, changes from adoption to discontinuance, or changes from rejection to later adoption if exposed to conflicting messages about the innovation. Further, each stage of the innovation-decision process is a potential rejection point. Although this model has frequently been cited in the educational literature as a useful framework for explaining the successful implementation of innovations in Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 18 schools (Eicholz & Rogers, 1964; Miles, 1964; Owens, 1970), it is considered to be of limited utility mostly due to its focus on individual-level factors while largely ignoring the organizational context inherent in schools. For example, one of the assumptions of Rogers’ model is that during any of the five stages, the individual user is free to decide whether the innovation shall be tried, continued or rejected. This assumption may not apply to educational innovations that are mandated by administrators for use by teachers. Thus, the individual teacher may not be in a position to move through the stages as posited by the model. Moreover, the adoption of an innovation by administrators does not necessarily ensure that it will ever be implemented or institutionalized at the classroom level. Similarly, the framework posits that during the initial knowledge stage, the prior characteristics or conditions (e.g., innovativeness) of the individual adopter (or aggregates of individuals) will affect the outcome of this knowledge stage. Again, individual-level factors are emphasized; whereas, the importance of organizational-level factors (such as norms of the social system) are minimized. The framework, however, is useful in helping to explain the importance of the innovation itself. As highlighted by the model, the perceived merits of the innovation play an important role in the persuasion stage of the innovation-decision process. Hallfors and Godette (2002) applied diffusion theory to identify the determinants of adoption and implementation of research-based substance use prevention programs in a sample of 104 school districts in 12 states. The authors hypothesized that districts that reported higher levels of perceived relative advantage, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 19 higher levels of perceived compatibility, and lower levels of perceived complexity were more likely to select and use a research-based program. Overall results supported all three hypotheses suggesting that these constructs are potential important determinants of the diffusion process. A second example of how diffusion of innovation theory constructs can be applied to research concerning the adoption and implementation of effective research-based prevention programs in schools has been illustrated in the research conducted by Parcel and colleagues (1995). The study examined three diffusion theory constructs, relative advantage, compatibility and complexity, believed to be associated with the adoption of an effective tobacco prevention program in schools throughout Texas. Results indicated that both teachers’ and administrators’ perceptions of the relative advantage of the curriculum (as compared to their current practices) was a strong significant predictor of adoption of the program. Social Psychological Theories A second set of theories that is relevant to planned social change within schools is based on the social psychological literature. Both the theory of planned behavior and social cognitive theory are useful frameworks for conceptualizing how people’s attitudes, expectations and beliefs interact with social systems to create certain behaviors and processes. Fundamental to both perspectives is the question of what motivates individuals’ behavior. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 20 Theory of Planned Behavior Ajzen’s (1988) theory of planned behavior is an extension of the theory of reasoned action (Fishbein and Ajzen, 1975), which postulates that individual performance (or non-performance) of a given behavior is primarily determined by an individual’s intention to perform that behavior. Intentions are indicators of how hard a person is willing to try to achieve behavioral goals. The theory of reasoned action includes two major factors that determine intention, one personal (attitude), and one social (subjective norm). An attitude indicates a person’s beliefs about the outcomes of the behavior, and the negative or positive evaluation of performing or not performing a behavior. Subjective norms can be described as normative beliefs that are based on the likelihood that other people would approve or disapprove of the behavior, as well as the person's motivation to comply with the opinions of others. A limitation of the theory of reasoned action is its assumption that an individual has complete volitional control over his/her own behavior. The extended theory of planned behavior, however, takes into consideration that people do not have complete control over their behaviors by adding a third construct, perceived behavioral control, which is believed to be a critical aspect of behavior change processes. The concept of perceived behavioral control is closely related to Bandura’s (1977) concept of self-efficacy. Perceived behavioral control indicates beliefs about opportunities, resources, and skills necessary to perform a behavior, and how much influence these beliefs have on performing the behavior. All three constructs, attitude, subjective norm and perceived behavioral control, can Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 21 exert independent influences on the intention to perform a given behavior (Ajzen, 1991). Perceived behavioral control can also directly affect behavior, independent of attitude and subjective norm. Thus, according to the theoiy of planned behavior, people will tend to adopt a behavior if they perceive it as having a positive value for themselves, believe that important social referents (e.g., friends, colleagues) think they should perform the behavior, and believe that they possess the capability to perform the activity to attain desired outcomes. The theory of planned behavior is often applied to the study of individual health behaviors, but it has also been successfully applied to program diffusion in schools. Researchers in the Netherlands utilized the theory of planned behavior to investigate the determinants of teachers’ decision making about using AIDS education curricula (Paulussen et al., 1994, 1995). Guided by this theoretical basis, the investigators examined the predictive validity of attitudes (e.g., outcome beliefs, perceived student interest, perceived instrumentality), subjective norms, and self- efficacy on the dissemination, adoption and implementation of prevention curricula in secondary schools. Results indicated that collegial interaction was significantly associated with teachers’ awareness knowledge. That is, those teachers who named colleagues as an information source, had discussed materials with them, and knew that colleagues were using the curricula were more likely to indicate awareness knowledge of the AIDS curricula. Further, both adoption and implementation were associated with perceived instrumentality, subjective norms, perceived colleague behavior, and teachers’ moral opinion on sexuality. Perceived instrumentality was Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 22 defined as the teachers’ perceptions of students’ enthusiasm for the curriculum, and whether the innovation was easy to use, provided clear procedural instruction, and met their planning concerns in terms of time required to prepare for classroom instruction. Social Cognitive Theory Bandura’s (1986) social cognitive theory, earlier named social learning theory, posits that human behavior change is affected by personal factors (i.e., values, knowledge, attitudes, perceptions), environmental influences (i.e., barriers, reinforcers, and environmental stimuli), and behavior (i.e., effects on and consequences of behaviors). These factors operate interactively as determinants of each other in a process called reciprocal determinism. This concept means that change is bi-directional and that behavior and the environment are reciprocal systems such that the person can shape the environment as well as environment shaping the person. According to this theory, human motivation and action are regulated by forethought. Beliefs about one’s ability to enact a specific behavior (self-efficacy) and beliefs about outcomes likely to occur if one performs the behavior (outcome expectancies) influence the types of behavior one chooses to engage in, the level of perseverance, and the amount of effort expended to enact the behavior. The more highly valued (expectancies) the expected outcome, the more likely the person will perform the behavior to yield the outcome. Furthermore, in order for change to occur, factual information must be provided to the individual about the behavior change (knowledge), the individual must believe that the behavior is acceptable to Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 23 their peer group, and they must have the skills to perform the behavior (behavioral capability). The concept of self-efficacy was considered by Bandura to be the single most important personal factor that determines one’s effort to change behavior. Self- efficacy reflects the degree of confidence an individual has in their ability to perform a behavior, not necessarily his/her true capabilities. Thus, Bandura’s (1997) work has focused on self-efficacy expectations and the manipulation of self-efficacy expectations, as he believed that outcome expectations were dependent on self- efficacy expectations—especially when outcomes were based on performance competence. For example, a person may believe that a behavior will produce a certain outcome but fail to act on that outcome if they do not believe they are capable of successfully performing it. In this situation, self-efficacy expectations are thought to have both a direct effect on the behavior, and an indirect effect on the behavior through outcome expectations. Social cognitive theory has been applied to diffusion research in schools. For example, the Smart Choices Diffusion Project intervention utilized social cognitive theory methods to influence the diffusion of a tobacco prevention programs to Texas schools (Parcel, Eriksen, et al, 1989; Parcel, Taylor, et al. 1989; Parcel, et al., 1995). The project addressed hypothesized social cognitive theory determinants (e.g., knowledge and awareness, outcome expectations, expectancies, reinforcement of adoption, behavioral capability, and self-efficacy for implementation) believed to be associated with four phases of diffusion in schools, dissemination, adoption, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 24 implementation, and maintenance. Social learning theory methods utilized to facilitate behavior change included direct modeling, guided mastery (i.e., practice of the behavior in a simulated situation), self-application of acquired skills, incentives, and social contracting (a public commitment to the behavior change goal). The study evaluated the effectiveness of the intervention strategies used to accomplish each phase of diffusion, and provided information on the factors associated with the specific outcomes (increasing knowledge and awareness, influencing school districts to adopt the program, accomplishing implementation, and maintaining use of the program). At the dissemination phase, it was found that an intervention that trained local opinion leaders to show videotapes modeling program adoption had no effect on the teachers’ and principals’ readiness to adopt the program (Brink et al., 1995). At the adoption stage, the study found that distribution of a newsletter, which summarized program effectiveness data and modeled adoption by successful school districts, resulted in higher adoption rates in the intervention school districts (Parcel, O’Hara-Tompkins et al, 1995). At the implementation stage, the investigators found that face-to-face teacher training workshops resulted in great levels of program implementation; however, the workshop had no effect on overall completeness and fidelity of program implementation (Basen-Engquist et al. 1994). Finally, at the maintenance stage, it was determined that distribution of a newsletter and the use of incentives such as teacher recognition and feedback on performance did not have an affect on maintenance of the program (Parcel, O’Hara-Tompkins et al, 1995). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 25 Organizational Change Theories Within the organizational literature there are a number of orientations and approaches that attempt to explain how and why organizations change. At present, one of the most important limitations in school improvement research has been a lack of understanding of schools as organizations. In order to increase our understanding of how to facilitate change in schools, it is necessary to see schools from the viewpoint of different perspectives. Two approaches to organizational behavior that that are applicable to health education practice are stage theory of organizational change and organizational development. Stage Theory of Organizational Change As conceptualized by Beyer and Trice (1978), stage theory of organizational change helps to explain how organizations plan and implement new goals, programs, technologies, and ideas. Organizations are believed to pass through a series of steps or stages as they change, with each stage requiring a unique set of actors, variables, circumstances and strategies that facilitate an innovation’s level of development and use. Movement through the stages can be forward, backward, or abandoned at any time during the process. The seven stages to this theory are: (1) Detection of unsatisfied demands on the system wherein a problem is uncovered; (2) Search for possible responses wherein alternative solutions are compared; (3) Evaluation of alternatives wherein potential solutions are compared; (4) Decision to adopt a course of action wherein a strategy is adopted from a number of alternatives; (5) Initiation of action within the system, which requires policy changes and resources for Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 26 implementation; (6) Implementation of change wherein the innovation is actually put into practice and usually requires resources and changes in organization members’ work behaviors and relationships; and (7) Institutionalization of change wherein organizational structures and processes (e.g., plans, job descriptions, and budgets) are changed in order for the intervention to become part of routine organizational operations. Within this model, different actors play leading roles at different stages of the organizational change. For example, decisions to adopt and institutionalize are typically political, therefore administrators take the leading roles during these stages. During implementation, however, focus is on the people who need to make changes in their practice. Thus, the determinants that are key to the innovation’s progress will change at different stages. Practical information about the characteristics of the innovation may be needed when leaders are ready to decide on a course of action, whereas training, technical assistance, and reinforcement may be most suitable at the implementation stage. To facilitate successful diffusion, an innovation’s current stage of development must be correctly assessed so that strategies to promote change can be matched to various points in the process of change. Organizational Development Theory Organizational development is an improvement strategy that applies behavioral sciences to improve organizational or system effectiveness. In the late 50’s and early 1960’s, the theory evolved from insights from group dynamics and from the theory and practice of planned change, which revealed that organizational Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 27 structures, processes, and culture influence work behavior and motivation. Organizational development programs typically identify problems that impede an organization’s functioning and then develop an overall plan or strategy that includes a series of activities, each designed to achieve an outcome that moves the organization towards its goals. The two major aims of organizational development programs are to improve the functioning and performance of individuals, teams, and the total organization, and to teach organization members how to continuously improve their own functioning. Thus, the targets of organizational development interventions are humans and their social processes. Efforts for change include a spectrum of people problems and work system problems such as: poor morale, interpersonal conflict, intergroup conflict, low productivity, poor performance, inappropriate organization structure, unclear or inappropriate goals, and inadequate response to environmental demands. Strategies involve diagnosis, action planning, interventions, and evaluations. Organizational change theories have also been applied to diffusion research in schools. Goodman and colleagues (Goodman, et al., 1991, 1997; Smith et al., 1995; Steckler et al ., 1992) incorporated elements of both stage theory of organizational change and organizational development theory to explore how to improve the diffusion of a tobacco prevention curriculum to twenty-two school districts in North Carolina. The principles of stage theory were used to inform the intervention, and organizational development theory was the basis for the strategies utilized during different stages of diffusion. For example, at the awareness stage, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 28 informational presentations and on-site meetings were conducted with principals to increase awareness (dissemination) of three empirically-based smoking prevention curricula. The site visits were found to be successful in increasing administrators’ awareness and concerns regarding tobacco prevention. At the adoption stage, intervention strategies were used to address three identified barriers to diffusion of the particular innovation: the project was sponsored by outsiders (i.e., research team), school personnel had limited knowledge about the proposed curricula and its requirements for instruction, and tobacco prevention was not a priority in the school districts (and perhaps was considered risky in a “tobacco state” such as North Carolina). One adoption-level intervention technique consisted of process consultations designed to help introduce schools to the curriculum and to clarify the training and commitment necessary for implementation. The consultations, however, had no effect on the districts’ adoption decisions (Goodman et al., 1992). To promote program implementation, teachers were provided training and ongoing on-site technical assistance. This strategy resulted in greater quantity of program lessons being implemented by trained teachers compared to teachers who did not participate in training (McCormick, et al., 1995; Smith et al., 1995) At the institutionalization stage, the strategy of process consultation focused on the political skills of those individuals who championed the program. Participants were given skills instruction on how to build coalitions of programs advocates so that the program could become entrenched within the school system. This strategy resulted in comparably low levels of maintenance in intervention and comparison Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 29 districts (McCormick, et a!., 1995). In addition to tailoring strategies to fit specific stages, the project intervened at different organizational levels. That is, at the awareness stage, a conference was directed at both administrators and teachers. At the adoption stage, the process consultation again included both administrators and teachers, whereas implementation was primarily focused on teachers. At the institutionalization stage, process consultations were re-focused once again on administrators. Thus, this intervention followed stage theory’s posit of utilizing different strategies and different stages, which proved to be beneficial in moving the innovation from one stage of development to the next. Empirical Research Results Gaiding Model Development Numerous research studies, guided by the tenets of differing social science perspectives, have provided empirical evidence of the various characteristics that are associated with successful change and improvement within schools. Although many of these studies have investigated factors specific to the prediction of adoption and implementation of innovative school-based programs, some factors have only been investigated in terms of other types of successful school change and improvement, such as providing professional development to staff or increasing student test scores. However, it is believed that the same factors and processes that determine various changes within schools will determine the successful adoption and implementation of prevention programs. The factors may best be classified into the following four groups: organizational, provider, curriculum, and demographics. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 30 Organizational Factors Organizational-level characteristics are key to promoting, change within schools because successful change is likely to occur if efforts are focused on the organization as a whole, not just at the level of the individuals working within the organization. Numerous research studies have suggested that successful adoption and implementation of innovative school programming requires that schools alter their goals, values, beliefs, structures, systems, and strategies (Stoll & Fink, 1996; Sammons, et al. 1995; Berends et al., 2002; Fullan, 1992; Palestini, 2000; Hall & Hord, 1987). Principal characteristics. One factor that appears to be critical to increased implementation and adoption of a program is role of the “gatekeeper” of educational innovations. The principal is typically the gatekeeper for adoption and continued use of new practices and programs in a school (Berman & MacLauglin, 1978; Fullan, 1991). Since the early 1970s Michael Fullan and other major figures in the field of educational change and improvement have provided considerable evidence suggesting that the school principal is central and most closely associated with leading and supporting change (Fullan, 1991, 1992; Hall & Hord, 1987; Hord, 1992; Miles, 1983). Numerous studies have shown that the knowledge, attitudes, and behaviors of the gatekeeper significantly facilitate or inhibit adoption and implementation. For example, Hall and colleagues’ (1980) investigation of innovations in the teaching of science in 60 elementary schools provided evidence that “the single most important hypothesis emanating from these data is that the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 31 degree of implementation of the innovation is different in different schools because of the actions and concerns of principals” (p.26). However, from this and other research (Hall & Hord, 1987) it was concluded that principal influence by itself does not directly account for school improvement; instead, principals facilitate the process of change. For example, the principal’s leadership may translate into the schools’ ability to obtain sufficient resources that, in turn, support their teachers’ efforts to implement an innovation (Berends et al., 2002). According to Louis and Miles (1990), an innovation will be successfully implemented if the principal is both effective in leadership and management. Leadership involves articulating a clear vision and/or mission, giving direction, and providing inspiration for the school as an organization. Management relates to designing and carrying out plans, negotiating resources, and working effectively with people. Although leadership skills and abilities have been studied more often in the literature, management functions are equally essential; thus, an effective principal should possess both sets of characteristics. Leadership variables that have been found to be positively related to innovation adoption include the leader’s attitude towards the innovation (Fullan, 1992; Hall & Hord, 1987; Sergiovanni, 1992; Vaill, 1998), clearly articulated vision (Donaldson, Jr., 2001; Glickman, 1993), high standards for and performance expectations from teachers (Sammons et al., 1995; Sergiovanni, 1992, 1996), power and respect (Block, 1987), and innovativeness or entrepreneurial spirit, especially in terms of technology adoption (Fullan & Miles, 1992; Hall & Hord, 1987). Management variables include planning for interventions (Pallestini, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 32 2000; Stoll & Fink, 1996), collecting and using information to keep aware of how the change effort is progressing (Fullan & Miles, 1992; Pallestini, 2000; Stoll & Fink, 1996), providing active ongoing support to teachers and programming in- service (Berman & McLauglin, 1977), managing conflict (Donaldson, Jr., 2001), and working well with teachers and school staff (Fullan & Hargreaves, 1998; Glickman, 1993; Sizer, 1986). A recent study (Kam et al., 2003) examined the role of principal leadership in the implementation of an empirically-validated school-based delinquency prevention program. Very interestingly, results indicated that the degree of principal leadership (i.e., provides support in general for the program, sees program as part of the central mission of the schools, speaks positively about the program with staff, collaborates with program coordinator and technical assistance teams, etc.) did not influence the success of the program dissemination on student behaviors. Further, high teacher implementation quality also d id not influence intervention effectiveness. Thus, neither high principal leadership nor high implementation quality by itself predicted program effectiveness. However, it was found that the intervention was effective in schools with both high principal leadership and high quality of program implementation by teachers. Specifically, when principal support was low, high quality of implementation in the classroom did not ensure that the intervention would be effective. Similarly, Smith and colleagues (1993) conducted research specific to the implementation of school-based tobacco use prevention curricula and found that Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 33 administrative support was a key determinant of program implementation. Results indicated that teachers in districts with supportive administrators were significantly more likely to implement the curriculum. Further, these teachers were more likely to implement a greater percentage of tobacco use prevention curriculum activities than teachers without administrative-level support. Additionally, it was found that those teachers who were better prepared to implement the curriculum (due to the effects of training) were more likely to implement the curriculum than teachers who were less prepared. However, the effect of being better prepared did not significantly influence completeness or extent of implementation. Thus, the authors concluded that regardless of how prepared or dedicated a teacher may be to implement programming, administrative and structural mechanisms (including the provision of sufficient resources) must support classroom activities to achieve successful implementation. These findings point to the very important need for studies to investigate the interrelationships among multiple predictors of implementation, or else risk that significant existing relationships may go undetected (Greenberg et al., 2000). Accordingly, because considerable evidence exists indicating that the school principal is central and most closely associated with leading and supporting change in schools (Fullan, 1991, 1992; Hall & Hord, 1987; Hord, 1992; Miles, 1983), factors surrounding the principal need to be better understood especially in combination with provider characteristics that are known or believed to influence implementation of school-based prevention programs. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 34 In addition to principal leadership and management, the principal’s personal commitment is also important to facilitating change. Administrators who are committed can be expected to support programs and work harder to make them succeed. Commitment may be defined as the principal’s interest, involvement, or feeling of being personally responsible for the execution and the outcome of change efforts. Studies indicate that giving high priority to the project (Donaldson, Jr., 2001; Fullan, 1992; Senge, 1999), having a strong personal belief in the cause (Glickman, 1993; Sergiovanni, 1992, 1996; Vaill, 1998), and feeling personal accountability for the success of the program (Barth, 1988; Senge, 1999; Vaill, 1998) are associated with successful school change. Rohrbach and colleagues (1993) found that their intervention designed to increase principals’ commitment to encourage and monitor program implementation in their schools resulted in greater rates of implementation of an evidence-based substance abuse prevention program. Knowledge about the innovation is another factor that contributes to the adoption of new educational programs. Research indicates that knowledge about an innovation is a key criterion for decision-making (Fullan, 1992; Paulussen et al., 1995). People who have sufficient valid information are in a better position to make well informed choices, and thus, are expected to make decisions that will advance their organization toward its goals. Further, making informed choices has another beneficial outcome; that is, people feel involved and more inclined to support decisions they participated in making, especially if they feel that their decision was based on a good appraisal of the situation. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 35 School Structure and Processes. Characteristics of the school structure and processes are also likely to influence the adoption and implementation of schoolwide change efforts. For example, organizational form (hierarchical versus network), number and types of organizational layers, communication structures and decision making channels determine the flexibility of an organization, and therefore, its ability to adapt to change. The degree of decentralization of power and authority in schools is viewed as an important structural factor that influences an organization’s capacity to change. The traditional command and control model of organizational structure is not suitable for school reform efforts because top-down management reinforces fear, distrust and internal competition, and reduces collaboration and cooperation. It may lead to compliance, but commitment is required for lasting change. Power sharing generates increased empowerment, and subsequent motivation and interest in school improvement among teachers (Fullan, 1992). Accordingly, studies have found that schools that have management structures that are decentralized, such as site based management systems, are more likely to implement educational programming (Louis & Miles, 1990). The importance of subordinate participation in decision making regarding innovations has also been given great importance in the school improvement literature. Shared decision making per se is not the significant feature, rather, the feeling of control that is created through joint participation. Studies have found that schools that engage in participatory decision-making have higher decision quality, increased ownership, commitment and satisfaction, and reduced resistance to change, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 36 and thus, will have the greatest impact on the degree to which an innovation is successfully adopted and implemented (Bemd, 1992; Boyd, 1992; Stoll & Fink, 1996). Organizational structure also refers to the communication infrastructure. A communication network that is comprehensive, flexible, and allows for a free flow of communication in all directions is more likely to be able to initiate and sustain change because employees will be clear about the organization’s direction. Organizations must clearly communicate the expectations, commitments, needs, and elements necessary to make the change effective. Multiple channels, including informal lines of communication are necessary because change tends to ignore the proper formal channels and established bureaucratic lines (Palestini, 2000). Within schools, clear communication especially among administrators and teachers has been identified as being a critical factor in the selection and implementation of innovative programming (Fullan, 2001; Palestini, 2000). The provision of adequate resources is also a necessary antecedent of successful adoption and implementation of educational programming. Material supports, such as funds to pay for program materials or training workshops, have been routinely documented in the school change literature to facilitate implementation of programming (Fullan, 1992; Huberman & Miles, 1984; Milstein, 1993; Yin & White, 1984). Specifically, schools that have administrators who provide assistance in the form of support personnel (coordinator or consultants) or Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 37 release time to attend program trainings have also proved to be more likely to continuously implement programming (Fullan, 1992; Huberman & Miles, 1984). School Culture. Considerable research indicates that schools have been extremely resistant to significant organizational change, in large part due to deeply embedded cultures that impede reform efforts (Bolman & Deal, 1993; Conley, 1997; Cunningham & Gresso, 1993; Stoll & Fink, 1996). Oftentimes, reform efforts have focused on changing the formal organizational structures of the school, while neglecting its informal and cultural dimension. It is believed that systemic change will not occur by introducing some specific modification, policy, or regulation. Rather, transformational change will be most successful if it is made at a deeper level, which would entail changing the culture of the school. School culture resides in the members of the organization, and it develops and evolves over time. An organization’s culture is difficult to discern because it is elusive and not clearly visible. Most simply, it is the basic beliefs and assumptions, which operate unconsciously, that the administrators, teachers, and support staff hold about each other, the school, the students, and the community. Research on school culture indicates that a whole-school vision is an important factor in the change process (Conley, 1997; Fullan, 1993; Hord, 1992; Nanus, 1992). Vision can be defined as the shared values and beliefs of a group of people. Vision helps schools to define their own direction and to develop a shared attitude that reflects a belief that the school as a whole is in charge of change. Other factors related to successful adoption and implementation of programming include the degree of the school’s innovativeness or Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 38 willingness to initiate and accept change (Nias et al., 1989; Stoll & Fink, 1996), respect for one another (Louis et al, 1995; Stoll & Fink, 1996), and a unified focus on students or student learning (Little, 1990; Louis et al., 1995). When attempting change, more successful schools have also established a warm climate, which includes higher levels of trust and openness between administrators, teachers and other staff members, and higher levels of collaboration among all school members (Joyce & Murphy, 1990; Little, 1990; Louis et al., 1995). Provider Factors A school’s readiness for change also depends to a large extent on the individual teachers. Teachers’ personal commitment has also been found to be a factor that determines to a great extent whether a school has the capacity to adapt to change. Like administrators, teachers who are committed can be expected to support programs and to work harder to make them succeed. Research indicates that teachers who have a strong personal belief in the cause (Fullan, 1992; Glickman, 1993; Miles, 1983; Vaill, 1998), and feel personal accountability for the success of the program (Donaldson, Jr., 2001; Louis et al., 1995; Vaill, 1998) are more likely to implement school improvements. Teachers’ knowledge about an innovation is another determinant of adoption of educational programming (Paulussen et al., 1995; Steckler, et al., 1992). Regarding types of informational sources used to influence program selection decisions, a recent study conducted by Rohrbach et al. (in press) found that schools were significantly more likely to decide to adopt an evidence- based substance use prevention curriculum if they consulted informational materials Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 39 obtained from credible sources such as federal and state substance use prevention agencies (e.g., National Institute on Drug Abuse and Center for Substance Abuse Prevention). Other provider characteristics that were significantly related to teacher- reported levels of program acceptance or implementation include teacher perceptions of students and their readiness to learn, greater program acceptance by teachers, good overall teaching skills, and teacher self-efficacy to deliver the program (Gingiss et al., 1994a, 1994b; Guskey, 1988; Parcel et al., 1995; Rohrbach et al., 1993). Curriculum Factors The importance of the curriculum itself should not be underestimated when studying adoption and implementation of innovative curricula. In order to put the innovation into practice in classrooms, implementers must have a clear understanding of the innovation’s technical requirements, as well its evidence of success. Curriculum characteristics shown to be associated with higher levels of adoption and implementation include utilization of teaching methods that are familiar to the provider, types of instructional methods used (e.g., didactic or role playing), ease of use, and clarity of curriculum’s procedural instruction (Gingiss et al., 1994a; Huberman & Miles, 1984; Parcel et al., 1995; Rohrbach et al., 1993; Smith et al., 1993). The chances for successful adoption and implementation are also greater if the teacher perceives a relative advantage of the innovation over existing programs (Hallfors & Godette, 2001; Parcel et al., 1995), or has been provided with information regarding the beneficial effects of the innovation (Fullan, 1992). For Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 40 example, if the impact of a prevention intervention is uncertain in terms of improved student outcomes, teachers are less likely to start or continue use of the innovation. Demographic Factors School demographic factors are also significantly related to program implementation, including socioeconomic status, minority composition, school level (elementary, middle, and high school), and school size. In a longitudinal school reform study conducted by RAND, implementation of school improvement efforts was positively associated with high levels of poverty among students, and higher percentages of minority students (Berends et al., 2002). Implementation levels were also higher in schools that were elementary schools versus middle or high schools (Bodilly, 1998) and smaller in size (Berends et al., 2001; Bodilly, 1998). This may be largely due to the hypothesis that secondary schools and larger schools are more complex organizations and are likely to resist organizational change (Lee & Smith, 1995, 1997). Other studies have also found that schools that are located in urban (versus suburban and rural) areas and have higher percentages of minority students are significantly more likely to implement evidence-based tobacco prevention practices (Rohrbach, de Calice et al., 2001, 2002; Rohrbach & Skara, in preparation; 2003; Rohrbach, Unger et al., 2002). PROPOSED THEORETICAL MODEL Theoretical and empirical research on understanding change within schools indicates that a combination of factors occurring at multiple levels may substantively influence the process by which schools adopt and implement prevention Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 41 programming. Emphasis should not be placed on understanding individual factors separately; instead, successful adoption and implementation should be viewed as resulting from a whole system of factors with mutually contingent interrelationships occurring over time. Thus, adoption and implementation must be considered as a research problem involving the whole organizational system, and should be studied from a socio-ecological perspective. Table 2 provides an outline of the factors selected for the proposed conceptual model. Organizational-level characteristics that affect adoption and implementation include principal characteristics (including principal leadership, commitment and knowledge), the organizational structure and processes of the school, and school culture. Provider-level characteristics include individual teachers’ knowledge, attitudes, and behaviors. Curriculum-level variables focus on the nature of the innovation itself. Demographic variables include exogenous variables related to the school composition. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 42 Table 2. Outline of Proposed Factors Associated with the Adoption and Implementation of Evidence-based Tobacco Use Prevention Curricula and Programming in Schools Organizational-level Components Principal Leadership of tobacco use prevention education Commitment to tobacco use prevention education Knowledge of effective tobacco use prevention programs Structure & Processes Management (decentralized) Participatory decision-making Communication Adequate resource (training, materials, time) School Culture Shared vision/goals Innovativeness Climate Collaboration Provider-level Components Commitment to tobacco use prevention education Knowledge of effective tobacco use prevention programs Perceived mandate to use effective tobacco use prevention curricula Expectations and values regarding tobacco use prevention programming Self-efficacy to implement tobacco use prevention programming Curriculum-level Components Provider perception that curriculum has been demonstrated or endorsed as being effective Utilized teaching methods familiar to provider, minimal training required Compatibility of interactive instructional methods Demographic-level Components Population density (rural, suburban or urban) School size (number of students enrolled) Socioeconomic status (percentage of students eligible to receive free or reduced cost lunch) Ethnic minority (percentage of non-white students) School type (high or junior/middle school) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 43 The framework portrayed in Figure 3 is an attempt to capture the complex system of variables that are necessary for successful adoption and implementation of evidence-based tobacco use prevention curricula. Organizations are complex and layered social systems, composed of members, roles, exchanges, cultures and resources. Thus, organizational change can best be promoted by working at multiple levels within the organization. Adoption and implementation of programming is more likely to occur when there is clear consistent leadership, communication, and commitment from the principal, both initially at adoption and throughout implementation. The principal should also have a clear understanding of what to do and change in order to put the innovation into practice. The innovation’s effectiveness should be known to the principal in order to improve the chances that it will be adopted and stay implemented. Structures and processes of the school should support organizational effectiveness through sharing of power, authority, and decision-making among the members who will be involved in facilitating and implementing the programming. The school structure should also allow for the provision of assistance in the form of adequate resources for training, release time and materials throughout the change effort. Successful adoption and implementation will also depend on the school culture. Members involved in the change effort must have a shared vision of their schools’ goals, and feel that they are working within a supportive environment with a warm climate that fosters acceptance of change (i.e., innovativeness). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 44 For adoption and implementation to succeed, the providers of the programming should also be committed to delivering the educational innovation to their students, have sufficient understanding of the full potential of the particular technology, perceive a mandate from the school district to use effective curricula, hold high expectations that it will be successful in changing student behaviors, as well as possess the self-efficacy to implement the programming. The nature of the innovation itself is another important factor in the diffusion process. In particular, the implementer of the innovation should perceive the curriculum as being evidence- based, and requiring teaching methods that are already familiar to the provider. Further, the provider should possess skills that are compatible with a curriculum that requires interactive instructional methods (i.e., discussion, activities, role playing) to deliver the curriculum. The socio-ecological perspective used here allows for simultaneous consideration of factors at each level, involving a whole set of concurrent processes in interaction as a whole system. Although the factors are presented as separate components, and were tested with cross-sectional data, the model is intended to portray dynamic processes with mutually contingent interrelationships acting over time. The framework of adoption and implementation developed here was not meant to be tested as a single equation model; rather, for purposes of this study, the framework has been developed to integrate and summarize factors derived from the theoretical and empirical literature, to test the preliminary hypotheses described herein, and serve as a stimulus for future empirical research. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 45 Figure 3. Proposed Conceptual Model of the Adoption and Implementation of Evidence-based Tobacco Use Prevention Curricula and Programming in Schools Organizational Principal Characteristics Structure & Processes School Culture Provider Adoption and Implementation of Evidence-based Tobacco Use Prevention Curricula and Programming Demographics Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 46 HYPOTHESES Study One: Cross-sectional Study Overview of Hypotheses It was hypothesized that the adoption and implementation of evidence-based tobacco use prevention curricula and programming would be positively associated with each individual variable listed in Table 2, and with each level of factors (i.e., organizational, provider, curriculum, and demographic). More specifically, it was believed that variance in the adoption and implementation of evidence-based tobacco use prevention curricula and programming would be best explained by the organizational context in which individual implementers work. Of the variables at the organizational level, principal leadership was thought to be the strongest correlate of whether an evidence-based curriculum and program would be adopted and implemented. However, characteristics of the organizational context, the provider, or the curriculum itself will not provide as adequate an explanation of program adoption as will the combination of all three levels of variables. Thus, in addition to testing for the associations between each variable and program adoption, independent of others, interactions between correlates at different levels were also examined. In particular, this investigation explored theoretically- and empirically- driven hypotheses, including a two-way interaction between principal leadership of tobacco use prevention education and provider self-efficacy to implement tobacco use prevention education, as well as a two-way interaction between provider self- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 47 efficacy to implement tobacco use prevention education and provider’s perception that the curriculum was demonstrated or endorsed as being effective. Specific Hypotheses Adoption and implementation of evidence-based tobacco use prevention curricula and programming would be positively associated with: 1. Organizational-level (Principal Characteristics): Schools that have principals who possess strong leadership skills, commitment and knowledge of tobacco use prevention programming 2. Organizational-level (Structure & Processes): Schools that have an organizational structure which supports organizational effectiveness, including decentralized management, participatory decision making, good communication, and provision of adequate resources 3. Organizational-level (School Culture): Schools that have a culture that is adaptive, yet consistent in planning its common goals, within a context that promotes collaboration, honest communications and a positive caring attitude as its ideal 4. Provider-level: Schools that have providers who possess strong commitment, attitudes, skills and knowledge of tobacco use prevention programming 5. Curriculum-level: Schools that have providers who perceive that the tobacco use prevention curriculum has shown evidence or been endorsed as being effective in preventing adolescent tobacco use and requires instructional methods that are already familiar to the provider, and that the provider uses skills compatible with interactive instructional methods (i.e., discussion, activities, role playing) used to deliver the curriculum 6. Demographic-level: Schools that are located in urban areas, are smaller in population size, have higher poverty and percentages of minorities, and are middle/junior high schools 7. Multi-level (Interactions): (a.) Schools that have an interactive combination of strong principal leadership of tobacco use prevention education and strong provider self-efficacy to teach tobacco use prevention education; Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 48 (b.) Schools that have an interactive combination of strong principal leadership of tobacco use prevention education and that the provider uses skills compatible with interactive instructional methods (i.e., discussion, activities, role playing) used to deliver the curriculum; and (c.) Schools that have an interactive combination of strong provider self- efficacy to teach tobacco use prevention education and strong provider perception that the curriculum has been demonstrated or endorsed as being effective Study Two: Longitudinal Study Specific Hypothesis Changes in each adolescent tobacco-use related outcome (e.g., smoking behavior, attitudes and knowledge) will be associated with: 1. an interactive combination of a higher level of school capacity to implement innovatiave programming (e.g., principal leadership, provider expectations, school climate, etc.) with a higher degree of implementation of evidence-based tobacco use prevention programming (according to CDC guidelines) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 49 METHODS Overview Based on the theoretical and empirical research described above, this research study had two separate but related aims. First, to examine the correlates and interactions between correlates of adoption and implementation of evidence-based tobacco use prevention curricula and programs in schools using primary data collection and analysis (study 1); and second, to investigate schools’ capacity to implement innovative programming as a moderator of the effects of implementation of evidence-based tobacco use prevention programming on changes in adolescent tobacco use-related outcomes using secondary data analysis (study 2). Data for both studies were obtained from schools that participated in the large-scale Independent Evaluation study of the California Tobacco Control Program (Independent Evaluation Consortium, 1998, 2001,2003; Rohrbach et al., 2002). Specifically, data for study 2 were obtained directly from the data sets that were collected and used for the purposes of the Independent Evaluation, whereas data for study 1 were collected, in part, specifically for this study, from a subset of the study population included in study 2. Therefore, the methodological characteristics of the Independent Evaluation study of the California Tobacco Control Program (Independent Evaluation Consortium, 1998, 2001, 2003) are described first. Independent Evaluation Overview, Design and Sampling The Independent Evaluation of the California Tobacco Control Program was a large multi-method study undertaken to determine the effectiveness of California’s Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 50 tobacco control activities by examining relationships between program implementation and exposure, and program outcomes (Independent Evaluation Consortium, 1998, 2001,2003; Rohrbach et al., 2002). The primary goal of the tobacco control program was to prevent and reduce tobacco use across the state of California by altering the social-political environment in which tobacco use occurs. To reach this goal, the tobacco control program was comprised of three primary program components utilizing multiple interventions and multiple modalities. The first component was community programming, designed to involve local community leaders, organizations and county health departments in the initiation and implementation of local tobacco control initiatives, policy development and public education programs. The second component was the statewide media campaign, which provided anti-tobacco advertisements through television, radio, print media, and billboards. The third component was the school-based Tobacco Use Prevention Education (TUPE) program, which provided funding for local educational agencies to implement tobacco prevention, education, and cessation programming for youth. The evaluation study employed a repeated cross-section design (longitudinal at the school and county levels), which included three waves of data collection separated by 18-month intervals. The first wave (Wave 1) of data collection occurred from October 1996 to February 1997, and focused on tobacco control activities in 1995 and 1996. The second wave (Wave 2) of data collection was conducted from March to July 1998, and focused on tobacco control activities implemented during 1997 and 1998. The third wave (Wave 3) of data collection took place from October Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 51 1999 to February 2000, and focused on tobacco control activities during 1998 and 1999. Sources of data included community opinion leaders, law enforcement staff, community opinion leaders, adults, school-district administrators (i.e., TUPE coordinators), school principals and teachers, and 5* 8th , 10ft and 12th grade students. Multiple data collection methods were used, including self-report questionnaires, telephone interviews, and coding of archival records. The conceptual framework for the evaluation has been described elsewhere (Independent Evaluation Consortium, 2001). To obtain a sample for the school-based component of the California Tobacco Control Program, the Independent Evaluation utilized a multi-stage sampling strategy. First, a sample of 18 California counties was selected from among California’s 58 counties by pre-selecting five counties that overlap with the state’s five largest media markets, and randomly selecting four to five counties from within three clusters (strata) based on county population density (population per square mile) and percent rural area. These 18 counties represent approximately three- quarters of the population of California. Second, school districts were randomly selected in each of these counties. Third, schools were randomly selected within each district. However, due to the structure of the funding mechanisms for TUPE interventions, which provides funds to schools that serve grades 4-8 on an entitlement basis and provides funds to schools that serve grades 9-12 on a competitive grants basis, a different sampling approach was used for the high schools relative to the other types of schools. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 52 For schools that serve grades 4 through 8, a sampling pool was first created by stratifying school districts by enrollment size in grades 4-8. Two strata were formed, based on a cut-point (varying by county) that represented the median enrollment in grades 4-8 combined. One district was randomly selected from each stratum, yielding a sample of 36 school districts. Of these chosen districts, 92% agreed to participate and the remaining 8% were replaced with other school districts that were willing to be included in the study. From the larger school districts, three elementary and three middle/junior high schools were randomly selected; whereas, within smaller school districts, only one elementary and one middle/junior high school was randomly selected. At Wave 1, a final sample of 70 elementary schools and 68 middle/junior high schools was obtained. From Wave 1 to Wave 2, one elementary school was dropped but replaced with another, resulting in a sample of 70 elementary schools at Wave 2 (99% of original sample retained). From Wave 1 to Wave 2, two middle/junior high schools were dropped but replaced with two others, and two additional schools had data lost by the data collectors, resulting in a sample of 66 middle/junior high schools at Wave 2 (94% of original sample retained). From Wave 2 to Wave 3, one elementary school was dropped but replaced with another, resulting in a sample of 70 elementary schools at Wave 3 (97% of original sample retained). From Wave 2 to Wave 3, no middle/junior high schools were dropped or added; however, Wave 3 data were available for the two schools that had lost Wave 2 data, resulting in a sample of 68 middle/junior high schools at Wave 3 (94% of original Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 53 sample retained). For each wave, two 5th -grade classrooms were randomly selected from elementary schools, and four 8* -grade classrooms (in a required topic area) were randomly selected from middle/junior high schools for participation in the survey. In order to select schools that serve grades 9 through 12, two sampling pools were first created without regard to the school district in which the school was located. For each county, pools were formed of: (1) regular high schools that received competitive TUPE grant funds (“grantees”); and (2) regular high schools that did not receive grant funds (“non-grantees”). Two grantee and three non-grantee schools were randomly selected per county. Schools that decided to drop out were replaced with other schools that were willing to participate. At Wave 1, a final sample of 65 high schools was obtained. From Wave 1 to Wave 2, four high schools were dropped and 19 others were added, and one additional school had data lost by the data collectors, resulting in a sample of 79 high schools at Wave 2 (92% of original sample retained). From Wave 2 to Wave 3, no high schools were dropped and six were added; further, Wave 3 data were available for the one school that had lost Wave 2 data, resulting in a sample of 86 high schools at Wave 3 (94% of original sample retained). It should be noted that, compared to the elementary and middle schools, the sample of high schools changed more over the three waves due to changes in schools’ TUPE grantee status over time. For each wave, five 10th -grade classrooms (in a required topic area) were randomly selected from high schools for Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 54 participation in the survey. At Waves 1, 2 and 3, the number of retained grantee schools in the sample was 41, 37, and 54, respectively. In each participating school, self-report data were collected from multiple sources, including: students in grades 5, 8, 10, and 12; principals and teachers in those same elementary, middle and high schools; and corresponding school district- level tobacco prevention coordinators. In each wave, the sample is representative of approximately 95% of public school students in grades 5, 8, and 10, 94% of public schools, and 92% of public school districts in California. In the present study, we focus on the teacher and student data. Youth Data. In all schools, an implied (“passive”) parental consent procedure was used to obtain parental permission to participate in the student self-report survey. That is, parents were assumed to permit their child to participate in the study if the parent did not sign a form declining participation. The student survey assessed students’ tobacco-related beliefs, attitudes, knowledge and behaviors, and exposure to tobacco control programs in schools, communities, and the mass media. The questionnaires were administered during a single, regular classroom period by researcher-trained data collectors following a standardized protocol. Students were briefed about the purpose of the study and informed that the survey was anonymous and voluntary. In Waves 1, 2, and 3, implied parental consent rates for 5th grade students were 93%, 92% and 94%, respectively. Implied parental consent rates were equally as high for the other grades, with respective rates of 99%, 99%, and 97% for 8th grade students, and 99% at all three waves for 10* grade students. Further, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 55 student consent (i.e., consent among students whose parents provided consent) was obtained for 99% of students at each grade level for all three waves. Thus, in Waves 1, 2, and 3, sample sizes were 3139, 3065, and 3060 for 5th -grade students; 5870, 5457, and 6238 for 8th - grade students; and 6505, 8226, and 8604 for 10ft-grade students, respectively. Classroom Teacher Data. During each wave of data collection, self-report questionnaires were also administered on-site to 5th -, 8th -, and 10th -grade teachers in whose classrooms students were assessed. In Waves 1, 2, and 3, teacher response rates were 89% (n=381), 90% (n=443), and 90% (n=408), respectively. In addition to these data, a survey was mailed to a random sample of teachers who the school principal identified as being responsible for tobacco prevention education. Response rates for mailed questionnaires for Waves 1,2, and 3 were 45% (n=145), 43% (n=151), and 54% (n=66), respectively. Study One: Cross-sectional Study Design and Sample Study 1 of this dissertation builds upon recent diffusion research that has investigated factors associated with the adoption and implementation of evidence- based tobacco prevention programs in California schools (as described in the introduction; Rohrbach & Skara, in preparation, 2003). Cross-sectional data from a sub-sample of tobacco use prevention teachers and school principals in schools that participated in the Independent Evaluation were utilized for analyses. Specifically, the target population comprised the set of 142 middle/junior and high schools (n=61 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 56 middle/junior and 81 high schools) that provided teacher data in at least one wave of the Independent Evaluation. In the present study, target respondents included one teacher and the principal of the school. Data collection for the teacher surveys was conducted during the 2002-2003 school year. Principal surveys were collected during the 2002-2003 and 2003-2004 school years. The assessment and the consent procedures were approved by the Institutional Review Board at the University of Southern California. Classroom Teacher Data. Before beginning data collection for the teacher sample, a letter was sent to the school district tobacco coordinator and school principal of each of the target schools (n=142). The letter introduced the study, explained its purpose, and solicited their support. Next, school principals were contacted by telephone and asked to identify all of the teachers at the school who provide tobacco education (generally science, health, or physical education teachers), are familiar with the school’s non-classroom tobacco prevention activities, and have a good understanding of the school’s tobacco prevention program. If more than one teacher was identified, one name was randomly selected to participate. The target sample of teachers (n=142) was sent a letter of introduction to the study, an informed consent letter, the self-report questionnaire, a stamped return self-addressed envelope, and a small compensation ($10 gift certificate to Barnes and Noble bookstore). The informed consent letter explained that they had been identified by their school principal as the primary tobacco prevention educator, the purpose of the study, that participation was voluntary, that their responses would be confidential, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 57 and any relevant instructions for the survey. If the survey was not returned by the deadline, a second questionnaire was mailed to nonrespondents. If teachers failed to complete and return the second survey, reminder telephone calls were made. Further, if the second surveys were not returned by the second deadline, a survey was then mailed to a different teacher (identified by the school principal as a tobacco prevention educator) at the same school using a similar protocol as described above. A total of 118 teachers returned completed surveys. Thus, the percentage of overall participant recruitment, i.e., the percentage enrolled in the study of those approached, was 83% for the teacher sample. However, seven schools had two teachers return a survey. These second observations were deleted, yielding a final sample of 111 teachers in 45 middle/junior and 66 high schools. School Principal Data. The eligible principal sample included principals from schools that returned completed teacher surveys (n=l 11). The principal was sent a letter of introduction to the study, an informed consent letter, the self-report questionnaire, and a stamped return self-addressed envelope. The informed consent letter explained the purpose of the study, that participation was voluntary, that their responses would be confidential, and any relevant instructions for the survey. If the survey was not returned by the deadline, a second questionnaire was mailed to nonrespondents. If principals failed to complete and return the second survey, reminder telephone calls were made. A total of 75 principals returned completed surveys. Thus, the percentage of overall participant recruitment, i.e., the percentage enrolled in the study of those approached, was 68% for the principal sample. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 58 Analytic Sample fo r Study 1. To address the primary aim of study 1, primary analyses were conducted on schools that provided data from both teachers and principals. Thus, the final sample of schools was comprised of a total of 75 schools (30 middle/junior and 45 high schools). Measures The teacher and principal measures provided information on organizational, provider, curriculum and demographic factors related to the adoption and implementation of evidence-based curricula. A multi-method approach was used such that items from both teacher and principal sources were combined to create scales used in the analyses. Study 1 utilized many of the existing items created for the Independent Evaluation, and examined in study 2, to provide consistency across study 1 and study 2 of this dissertation. The majority of the new scales were derived from other school-based substance use prevention and school improvement research studies, including the Keys Project (Keys to Excellence in Your Schools undertaken by the National Education Association, 2003) and the Chicago Annenberg Research Project (Consortium on Chicago School Research, 2000, 2002). Specifically, the Keys Project initiative was conducted by the National Education Association to explore empirical connections between school conditions, teaching and learning, and student outcomes while emphasizing the interactive nature of organizational, curricular, and instructional processes. The Chicago Annenberg Research Project Research Project was founded in 1990 to assess the progress of school improvement Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 59 and reform through a biannual survey of teachers and students in 434 of Chicago’s schools. These school-based substance use prevention and school improvement research investigations have provided evidence of good internal consistency (Cronbach alphas above .60 for both individual- and school-level reliability) for many of their scales (e.g., see Consortium on Chicago School Research reports, 2000, 2002); however, this research study adapted many items and combined them into new scales. The rationale for altering existing, proven scales was to limit the overall length of the questionnaire, especially for those administered to principals, a group that has shown lower response and completion rates in previous data collection efforts (Independent Evaluation Consortium, 1998, 2001, 2003). As a result, these new theoretically-driven scales do not have previously reported reliability statistics. Study 1 measures of each variable are summarized as follows, and questionnaire items measuring these variables are provided in Appendices B and C (2002-2003 Teacher Survey and 2002-2003 School Site Administrator Survey, respectively). Dependent Variables Adoption and Implementation o f Evidence-based Tobacco Use Prevention Curricula. A binary dependent variable was created from teacher questionnaire data to represent the implementation or non-implementation of evidence-based tobacco prevention curricula. Participants were asked to indicate which commercial or published curricula they used the most recent time they taught tobacco prevention Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 60 lessons in the classroom. Response options specified in the questionnaire included a list of 26 classroom-based substance use prevention curricula. An “other” category was also provided to allow respondents to write in any curricula that were not named in the questionnaire. Curricula were classified as being evidence-based if they were identified as a research-validated, exemplary, or model program by one or more of the following agencies: the Center for Substance Abuse and Prevention (CSAP; 2001), National Institute on Drug Abuse (NIDA; 1997), Drug Strategies, Inc. (2002), United States Department of Education’s (2001) Expert Panel, the University of Colorado’s Center for the Study and Prevention of Violence (2002), or the California Department of Education’s (2000) publication Getting Results. As shown in Table 3, a total of 6 curricula that were specified on the teacher questionnaires were identified as being evidence-based, including Project ALERT, Life Skills Training, Project TNT (Towards No Tobacco Use), Quest: Skills fo r Adolescence, Minnesota Smoking Prevention Program and Reconnecting Youth: A Peer Group Approach to Building Life Skills. Schools that had teachers who reported that they had used at least one of the curricula determined to be evidence-based were classified as having implemented an evidence-based program. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 61 Table 3. Study One: Prevalence ofUse of Tobacco Use Prevention Curricula for Junior/Middle and High School Grades (N-75 Schools)____________ Curricula N (%) Other (including health textbook, district-developed curriculum, etc.) 25 (33.33) Tobacco Free! Tobacco-Use Prevention Lessons 23 (30.67) Here’s Looking at You, 2000 17 (22.67) PARE 12 (16,00) Get Real About Tobacco 10 P -- * V : ii * O 00} Life Skills Training*’ ^ eJ Discover: Skills for Life : ■ : - oo) Growing Healthy 3 ( 4.00) Health Skills for Life 3 ( 4.00) IN-DEPTH Drug Education and Prevention Tools for Your Health 's ( 4 00) Project TNT (Towards No Tobacco Use)a £ ^J > , " * < } ; Drugs, Alcohol, and Tobacco “Totally Awesome” Teaching Strategies 2 ( 2.67) Discover: Decisions for Health 2 ( 2.67) Learning to Live Drug Free 2 ( 2 67) Quest; Skills for Adolescence8 af 2 ( 2,67) Know Your Body ! ( . i Minnesota Smoking Prevention Program - Tobacco-Free Tee l ! ’ ..D Personal/Social Skills Lessons: The Missing Link in Prevention Curricula - High School 1 ( 1.33) Project Life 1 ( 1.33) Youth Media Network - An Interactive Tobacco Education Curriculum 1 ( 1.33) Media Sharp: Analyzing Tobacco and Alcohol Messages 0 ( 0.00) McGruffs Drug Prevention and Child Protection 0 ( 0.00) Project SCAT (Schools and Community Against Tobacco) if t ;){■:()) Reconnecting Youth. A Peer Group Approach to Building Life S k ills ^ 0 , 00: ? Quest: Skills for Growing 0 ( 0.00) Quest: Skills for Action 0 ( 0.00) Note. Highlighted curricula were classified for this study as being evidence-based. "Recommended by Center for Substance Abuse and Prevention (CSAP, 2001) ‘ ’ Recommended by National Institute on Drug Abuse (NIDA, 1997) "Awarded an ‘A’ by Drug Strategies. Inc. (2002) ‘ ‘ identified as an exemplary program by United States Department of Education's Expert Panel (2001) "Identified as a model program bv University o f Colorado’s Center for the Study and Prevention o f Violence (2002) ‘ identified as an exemplary program by California Department o f Education’s publication Getting Results (2000) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 62 Adoption and Implementation o f Evidence-based Tobacco Use Prevention Programming. In addition to the binary dependent variable, which solely measured the implementation or non-implementation of specific evidence-based tobacco use prevention curricula, a continuous dependent variable was created from teacher questionnaire data to represent the degree of adoption and implementation of evidence-based tobacco use prevention programming overall. The decision to examine this second dependent variable was made due to the small percentage of teachers (i.e., 20% or 15 of 75 teachers) who reported adopting and implementing one of the evidence-based tobacco use prevention curricula assessed on the questionnaire. This low rate of use of evidence-based curricula may have been the result of a combination of two factors, the specific curricula which were assessed on the questionnaire and the grade level at which teachers who comprised the final analytic sample provided tobacco lessons. That is, only six of the 26 curricula specified on the questionnaire were classified as being evidence-based. Of these evidence-based curricula, five targeted junior/middle school students; whereas, only one curriculum (i.e., Reconnecting Youth: A Peer Group Approach to Building Life Skills') was designed to be delivered to high school students. However, the final analytic sample of 75 schools included mostly high schools (i.e., 60% of the school sample). Therefore, it is possible that additional high school teachers would have reported use of an evidence-based curriculum if the questionnaire specifically included, by name, more evidence-based curricula that were designed for high school students. It should be noted that most evidence-based substance use prevention Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 63 programs have been designed for junior/middle schools, not high schools (Sussman et al, 1995). Thus, the present study’s selection of specific curricula which were assessed on the questionnaire reflects the dearth of effective high school programs. The second dependent variable that was assessed was more general in scope than the first because it represented a comprehensive index of the critical elements of effective tobacco use prevention and cessation programming, in and out of the classroom, including policy, curriculum delivery and teacher training. A standardized implementation index, adapted from a similar index developed in previous research (Rohrbach, de Calice et al., 2001, 2002), was created for each school, representing the degree of implementation of effective tobacco use prevention programming. The index was based on 5 of 7 of the guidelines from the CDC Guidelines for School Health Programs to Prevent Tobacco Use and Addiction (1994; see Table 1). Specifically, the implementation index was compri sed of the teacher self- report data, and was created in the following manner. First, analyses were limited to data derived only from S^-grade teachers who taught health and/or science, which are the middle school subject areas in which tobacco lessons are typically taught. Similarly, analyses were restricted to include data obtained only from lO^-grade teachers who taught health and/or physical education, which are the most common subject areas for tobacco instruction at the high school level. Second, to construct the implementation index, items from the surveys were selected to address each of the CDC guideline elements. Third, for each element of the index that was created from Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 64 multiple measures (i.e., items), internal consistency was calculated using Cronbach’s coefficient alpha to determine if items should be combined. Items that did not show good internal consistency with the remainder of the scale (alphas below .60) were dropped. Items that showed good internal consistency (alphas above .60) were selected for inclusion in the index. Fourth, Cronbach’s coefficient alpha was calculated for the final implementation factor, to determine how well the elements fit together within the index. Fifth, all elements of the index were standardized to a mean of 0 and a standard deviation of 1 (to ensure that equal weight was given to each element) and summed to create a final continuous score for each school. Sixth, analyses were conducted on observations having non-missing data on all of the variables used to create the index. The final implementation index consisted of 5 variables, tobacco use policy enforcement and violation consequences, tobacco instruction, teacher program- specific training, parent involvement and cessation support, which were selected to represent 5 of the 7 guidelines from the CDC Guidelines for School Health Programs to Prevent Tobacco Use and Addiction (1994; see Table 1). Cronbach’s alpha for the final implementation index was .62. Tobacco use policy enforcement (Guideline #1) was measured with two items that indicated the teacher’s perception of the extent of policy enforcement at their school (using a 4-point scale ranging from l=none to 4=a great deal), and the principal’s report of the types of policy violation consequences for students who are caught smoking (scored as 0=punishment only, l=punishment and remedial, 2=remedial only). Tobacco instruction (Guideline #2) was assessed by Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 65 responses to two items indicating the number of tobacco use prevention lessons taught to each class/section (scored on a continuous scale) and whether they had included the recommended instructional content (a sum of physiological effects, social influences, social consequences, peer norms, and refusal skills; each scored 0=no or l=yes). Teacher training (Guideline #4) was measured by responses to two items that asked whether teachers had participated in tobacco use prevention in- service training in the last five years (scored 0=no or l=yes), and whether the training was program-specific (scored 0=no or l=yes). Parent involvement in support of tobacco use prevention programming (Guideline #5) was assessed with one item indicating the extent of teachers’ efforts to involve parents in tobacco use prevention programs (using a 4-point scale ranging from l=none to 4=a great deal). Cessation support (Guideline #6) was measured with one item that indicated whether the school had an on-site cessation program for student smokers (scored 0=no or l=yes). The guideline regarding the grade span for tobacco programs (Guideline #3) was excluded because the focus of this study was middle and high schools only, and the guideline regarding regular program evaluation (Guideline #7) was excluded because the vast majority of school districts reported that they had conducted recent evaluations (Rohrbach et al., 1998). All items used to create the implementation index were derived from the Independent Evaluation study (Independent Evaluation Consortium, 1998). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 66 Independent Variables Organizational-level Variables. The first domain of variables hypothesized to influence the implementation of evidence-based curricula and programming encompassed the organizational-level. Organizational-level components included variables grouped into three sub-domains: principal characteristics, structure and processes, and school culture. Principal Characteristics. Characteristics of the school principal were assessed with three factors, principal leadership of tobacco use prevention education, principal’s commitment to tobacco use prevention education, and principal’s knowledge of effective tobacco use prevention programs. Principal leadership was measured with an index that was created from three principal survey items (each originally scored on a 5-point scale ranging from l=strongly agree to 5=strongly disagree). These items asked the principal to indicate to what extent they agreed with the following statements: “As the principal (or assistant principal) at this school, I”: (a) “Encourage teachers of health, science, and/or physical education to take professional development on tobacco prevention education ”; (b) “Have communicated a clear mandate to teachers regarding tobacco prevention education instruction.”; (c) “Provide or help teachers get the resources they need to effectively teach tobacco prevention education.” Principal leadership items were adapted from the Keys Project questionnaire for school site administrators (National Education Association, 2003). Cronbach’s alpha for this three-item index was .79. The principal’s commitment to tobacco use prevention education was assessed with three Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 67 items. The first item asked principals about the priority that tobacco use prevention education has at their school (originally scored on a 5-point scale ranging from l=the highest priority to 5=the lowest priority). The second and third items asked teachers to indicate how supportive the school principal has been of tobacco use prevention education (originally scored on a 4-point scale ranging from l=very supportive to 4=not at all supportive), and how much input the school principal has in making decisions about which tobacco use prevention curricula are used (originally scored on a 4-point scale ranging from l=a great deal of input to 4=no input). The first principal commitment item was derived from the Independent Evaluation study (Independent Evaluation Consortium, 1998), whereas the two other items were created for the purposes of this study. Cronbach’s alpha for this three-item index was .61. Principal’s knowledge of tobacco use prevention curricula was measured with two items derived from the principal surveys. The first item asked, “Are you aware of exactly which tobacco prevention curricula are research-tested and proven to work?” Response options included, “Yes, definitely,” “Yes, somewhat,” or “No.” The second item asked, “Have you ever used the guide called “Getting Results: An Action Guide to Tobacco Use Prevention Education” (developed by the California Department of Education)?” Response options were, “Yes,” “No,” and “Em not sure.” The first principal knowledge item was created for the purpose of this study, and the latter item was derived from the Independent Evaluation study (Independent Evaluation Consortium, 1998). The correlation coefficient (r) for this two-item principal commitment scale was .41, p<.G0O3. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 68 Structure and Processes. The structure and processes of the school were assessed with four variables, management structure, decision-making processes, communication, and resources (training, teaching materials, and time). To assess whether schools have a decentralized management structure, three items from the principal questionnaire were used. The first item asked, “At what level are decisions made within your school (e.g., decisions about budget, personnel, curricula, etc.)?” The second item asked, “At what level are decisions made within your school district (e.g., decisions about budget, personnel, curricula, etc.)?” Response options for both questions included, “Mostly at top,” “Policy at top, some delegation,” “Broad policy at top, more delegation,” and “Throughout, but well integrated.” The third item asked principals to indicate (on a 5 point-scale ranging from l=strongly agree to 5=strongly disagree) the extent to which they agreed with the following statement: “Our school district administration does not interfere with our ability to make tobacco prevention education decisions at this school” The first two items were adapted from the “Profile of Organizational Characteristics” questionnaire measures developed by Rensis Likert for research in organizational development (Likert, 1967), and the third item was developed for the purposes of this study. Cronbach’s alpha for this three-item index was .64. Decision-making was assessed with four items derived from both principal and teacher questionnaires. The same items asked principals and teachers, separately, to indicate (on a 5 point-scale ranging from l=strongly agree to 5=strongly disagree) the extent to which they agreed with the following statements: “The principal usually consults with teachers before making Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 69 decisions that affect them.” and “Teachers are involved in making decisions that affect them.” Both items were derived from the Independent Evaluation study (Independent Evaluation Consortium, 1998). Cronbach’s alpha for this four-item decision-making index was .65. Communication was assessed with two items (both originally scored on a 5-point scale ranging from l=strongly agree to 5=strongly disagree) that asked principals to indicate the extent to which they agreed with the following statements: “Teachers at the school feel free to communicate with the principal.” and “As the principal (or assistant principal) at this school, I keep well- informed about the tobacco prevention curriculum and instructional program that students are receiving.” The first item was derived from the Independent Evaluation study (Independent Evaluation Consortium, 1998), whereas the second item was adapted from the Keys Project questionnaire for school site administrators (National Education Association, 2003). Cronbach’s alpha for this two-item communication index was .60. The availability of adequate resources was measured with three teacher questionnaire items (each scored yes or no) that asked respondents to indicate whether inadequate training, inadequate teaching materials, or other demands on class time made teaching about tobacco difficult for them. Responses to the resources items were summed to create a continuous scale. The resource items were both derived from the Independent Evaluation study (Independent Evaluation Consortium, 1998). School Culture. Culture was assessed with four variables, shared mission and/or goals, innovativeness, climate and collaboration. Shared mission and/or goals Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 70 was measured with three items. The first item assessed the extent to which principals agreed that teachers and principals share beliefs and values about what the central mission of their school should be. The second and third items asked principals and teachers, separately, to indicate the extent to which they agreed that in their school there is a feeling that everyone is working together toward common goals. All three shared mission and/or goal items were measured on a 5-point scale ranging from l=strongly agree to 5=strongly disagree. The first item was adapted from the Keys Project questionnaire for school site administrators (National Education Association, 2003), and the latter two items were derived from the Independent Evaluation study (Independent Evaluation Consortium, 1998). Cronbach’s alpha for this three-item shared mission and/or goals scale was .61. Innovativeness was measured with three items derived from the principal questionnaire. The first two items asked principals to indicate the extent to which they agreed that teachers at their school resist change and that teachers are eager to try new ideas. The third innovativeness item asked principals to indicate the extent to which they agreed with the following statement: “Most changes introduced in this school are often more trouble than they are worth.” All three items were measured on a 5-point scale ranging from l=strongly agree to 5=strongly disagree. The first two items were both derived from the Independent Evaluation study (Independent Evaluation Consortium, 1998), and the third item was adapted from the Chicago Annenberg Research Project site administrator surveys (Consortium on Chicago School Research, 2000, 2002). Cronbach’s alpha for this combined three-item index was .68. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 71 School climate was assessed with six items derived from both principal and teacher questionnaires. Three items with the same wording asked principals and teachers, separately, to indicate (on a 5 point-scale ranging from l=strongly agree to 5=strongly disagree) the extent to which they agreed with the following statements: “Teachers are supportive of one another.”; This school seems like a big family; everyone is close and cordial.”; and “In general, teachers at this school are treated fairly.” All of these school climate items were adapted from the Keys Project questionnaire for school site administrators and teachers (National Education Association, 2003). Cronbach’s alpha for this combined six-item index was .62. Collaboration was assessed with three items derived from both principal and teacher questionnaires. Two items with the same wording asked principals and teachers, separately, to indicate (on a 5 point-scale ranging from l=strongly agree to 5==strongly disagree) the extent to which they agreed that teachers frequently consult with and help one another. The third item asked only principals to indicate (on a 5 point-scale ranging from l=strongly agree to 5=strongly disagree) the extent to which they agreed that administrators and teachers collaborate to make their school run effectively. The collaboration items were derived from the Independent Evaluation study (Independent Evaluation Consortium, 1998). Cronbach’s alpha for this combined three-item collaboration index was .60. Provider-level Variables. The second domain of variables hypothesized to influence the implementation of evidence-based curricula and programming encompassed characteristics of the provider, i.e., those who actually implement Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 72 tobacco use prevention programming. Provider-level components included the following five variables: commitment to tobacco use prevention education, knowledge of effective tobacco use prevention programs, perceived mandate to use effective tobacco use prevention curricula, perceived expectations and values regarding tobacco use prevention programming, and perceived self-efficacy to implement the programming. Provider’s commitment to tobacco use prevention education was assessed with three items that asked teachers about the priority that tobacco use prevention education has at their school (originally scored on a 5-point scale ranging from l=the highest priority to 5=the lowest priority), how much input classroom teachers have in making decisions about which tobacco use prevention curricula are used (originally scored on a 4-point scale ranging from l=a great deal of input to 4=no input), and how supportive classroom teachers have been about tobacco use prevention education (originally scored on a 4-point scale ranging from l=very supportive to 4=not at all supportive). The first item was derived from the Independent Evaluation study (Independent Evaluation Consortium, 1998), and the second and third items were developed for this study. Cronbach’s alpha for this three-item index was .65. Provider’s knowledge of effective tobacco use prevention curricula was measured with two items derived from the teacher surveys. The first item asked, “Have you ever received a copy of the “Guidelines for School Health Programs to Prevent Tobacco Use and Addiction” published by the U.S. Centers for Disease Control and Prevention?” Response options included, “Yes,” “No,” or “I’m not sure.” The second item asked, “Have you ever used the guide called “Getting Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 73 Results: An Action Guide to Tobacco Use Prevention Education” (developed by the California Department of Education)?” Response options were, “Yes,” “No,” and “I’m not sure.” Both items were derived from the Independent Evaluation study (Independent Evaluation Consortium, 1998). Responses to these items were summed to create a continuous 0-2 scale. Providers’ perceived mandate to use effective tobacco prevention curricula was assessed with one item that asked teachers, “During the past three years, to what extent has your school district encouraged you to use a tobacco prevention curriculum that is effective or research-based?” Respondents were given a 4-point scale for this item ranging from l=a great deal to 4^=not at all. The item was developed for this study. To measure the provider’s perceived expectations and values regarding tobacco use prevention programming, teachers were asked one item: “How effective do you think your tobacco use prevention lessons are in preventing or reducing tobacco use among your students?” Response options ranged from l=very effective to 4=not at all effective. This item was adapted from the Independent Evaluation study (Independent Evaluation Consortium, 1998). Providers’ perceived self-efficacy to implement tobacco use prevention education was assessed with two teacher questionnaire items. The first item asked teachers about how comfortable they are with various teaching strategies, including role plays and small group discussions (originally scored on a 4-point scale ranging from l=very comfortable to 4=not at all comfortable), and how confident they are that they are doing a good job teaching tobacco use prevention lessons (originally scored Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 74 on a 4-point scale ranging from l=very confident to 4=not at all confident). Both items were adapted from the Independent Evaluation study (Independent Evaluation Consortium, 1998). The correlation coefficient (r) for this two-item index was .38, p< 002. Curriculum-level Variables. The third domain of variables hypothesized to influence the implementation of evidence-based curricula and programming dealt with characteristics of the curriculum. Curriculum measures were derived from questionnaire items that referred to the curriculum the teacher used the most recent time they taught tobacco use prevention lessons in the classroom. Curriculum-level components included three variables that addressed whether the provider perceives that the curriculum has shown evidence or been endorsed as being effective in preventing adolescent tobacco use, whether it requires instructional methods that are already familiar to the provider, and the extent to which the provider uses skills compatible with the interactive instructional methods (i.e., discussion, activities, role playing) emphasized the curriculum. To assess the extent to which the provider perceives that the curriculum has shown evidence or been endorsed as being effective in preventing adolescent tobacco use, two items were used. Teachers were asked, “How important were each of the following factors in influencing your decision to use this/these tobacco prevention curriculum(a)?”: (a) “Demonstrated to be effective in evaluation studies.” and (b) “Endorsed by the California Department of Education.” Responses for each item included a 4-point scale ranging from l=very important to 4==not at all important. Both of the items were developed for the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 75 purposes of this study. The correlation coefficient (r) for this two-item index was .43, p< 01. The provider’s perception that the curriculum requires instructional methods that are familiar to the provider were assessed with two items. Teachers were asked, “How important were each of the following factors in influencing your decision to use this/these tobacco prevention curriculum(a)?”: (a) “Utilized teaching methods familiar to you.” and (b) “Minimal training required.” Responses for each item included a 4-point scale ranging from l=very important to 4=not at all important. Both of the items were developed for the purposes of this study. The correlation coefficient (r) for this two-item index was .54, p< 000. To assess the extent to which the provider uses skills compatible with the interactive instructional methods (i.e., discussion, activities, role playing) emphasized in the delivery of the curriculum, three questions were used. Teachers were asked, “How much did you use the following instructional strategies in your tobacco prevention lessons?”; (a) “Small group activities.”; (b) “Student worksheets (seat work).”; and (c) “Role plays or skills practice.” Responses for each item included a 4-point scale ranging from l=a great deal to 4=not at all. These three items were adapted from the Independent Evaluation study (Independent Evaluation Consortium, 1998). Cronbach’s alpha for this three-item index was .63. School Demographic-level Variables. In addition to the primary data collected from the teachers and principals, school-district and school-level demographic data on the sample of schools were obtained from three sources. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 76 Information regarding population density (e.g., rural, suburban, urban) of the area served by the school was obtained from the National Center for Educational Statistics website (2001). The website for the California Basic Education System (CBEDS) provided information on school size (number of students enrolled) and percentage of students eligible to receive free or reduced cost lunch (used as a proxy measure for socio-economic status; California Department of Education, 1998). Student ethnicity (percentage of non-white student composition) and school type (high or junior/middle school) were derived from the Wave 3 (year 2000) Independent Evaluation student data sets. These five demographic measures were used as covariates in this study. Each variable was based on a school-level measure (i.e., individual student or school scores were aggregated to create school means). Population density was examined using an ordinal scale (0=rural, l=suburban, 2=urban). Both enrollment size and percentage of students eligible for free lunch (used as a proxy measure for socio economic status) were continuous scales (as provided by the CBEDS data set). The student ethnicity variable represented the proportion of non-white students in the school. The data were recoded to represent 0==white and l=non-white (i.e., any ethnic group other than white). School type was recoded as 0=high and l==junior/middle school. Data Analysis Plan Data Analysis Overview. Data analyses were conducted on primary data (as described below) to examine the various correlates of adoption and implementation Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 77 of evidence-based tobacco use prevention curricula and programming, as well as the importance of the hypothesized interactive influence of the combination of organizational-, provider-, and curriculum-level factors on the adoption and implementation of evidence-based tobacco use prevention curricula and programming by schools. All statistical analyses were performed using the SAS System (version 8.02) software program (SAS Institute, 2001), and all testing was carried out at the traditional .05 significance level unless otherwise noted. For all variables of interest, univariate frequencies were run to detect outliers, and inconsistent or missing data. Questionable values were compared against hardcopy surveys for accuracy and re entered if appropriate, or set to missing if the correct response could not be determined. A ttrition Analyses. Attrition analyses were conducted to examine differential attrition from the study on all school demographic variables. Schools that provided complete data from both teachers and principals (i.e., the analytic sample of 75 junior/middle and high schools) were compared to all other schools (that provided data from the teachers or principals, but not from both; n=36 junior/middle and high schools). This type of analysis helps determine how representative the analytic sample is of the population of schools. T-tests and GLM procedures were used to compare 75 schools used versus 36 schools not used in the analyses. Sample Characteristics. Demographic statistics were calculated for the analytic sample (n=75), by school, teacher and principal groups. That is, for Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 78 categorical variables, frequencies and percentages were computed; and for continuous measures and scales, means and standard deviations were computed. Descriptive Analyses o f Variables. Descriptive statistics were computed for each of the dependent and independent variables. That is, means, standard deviations, ranges and sample sizes were calculated, when applicable. All distributions were examined for adequate variability and possible violations of normality assumptions (by skewness and kurtosis). Results indicated that all variables were normally distributed, thus, transformations or robust statistical methods were not considered in all further bivariate or multivariate analyses. Further, for the independent variables, means, standard deviations and ranges were calculated for all 75 schools combined, then by whether or not the school used an evidence- based tobacco prevention curriculum. T-tests were then performed on all predictor variables to analyze whether statistically significant differences existed between schools that had not used, versus schools that had used an evidence-based curriculum. Pearson correlation analyses were also conducted to examine relationships among all dependent and independent measures. Hypothesis Testing. To test the hypotheses, two sets of a four-stage regression analysis protocol were employed. The first set of models was used to identify which independent variables were significantly associated with the adoption and implementation of evidence-based tobacco use prevention curricula, and the second set of models was used to identify which independent variables were significantly Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 79 associated with the implementation of evidence-based tobacco use prevention programming. For the first set of models, analyses were performed as follows. First, univariate logistic regression analyses were conducted by regressing the dependent variable (use or non-use of an evidence-based curriculum) on each hypothesized correlate of implementation (including organizational, provider, curriculum and demographic factors). Odds ratios and 95% confidence intervals were calculated for each variable. Second, multivariate analyses were conducted to determine which variables were most important at each level of factors (i.e., organizational, provider, etc.). That is, predictors within each domain (e.g., among principal characteristics only) were simultaneously entered into a multivariate logistic regression model, controlling for all statistically significant demographic variables. Adjusted odds ratios and 95% confidence intervals were calculated for each variable by domain. Third, all statistically significant variables (p< 10) from the second-stage multivariate analyses were simultaneously entered into a multivariate logistic regression model (that combined predictors from all domains), controlling for all statistically significant demographic variables. Adjusted odds ratios and 95% confidence intervals were calculated for each variable. Fourth, to enable further examination of how the multiple levels of factors were related, hypothesized interactions were examined. Specifically, selected variables from different domains were entered into a multivariate logistic regression model along with the corresponding hypothesized Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 80 interaction term, controlling for all statistically significant demographic variables. Adjusted odds ratios and 95% confidence intervals were calculated for each variable. For the second set of models, analyses were performed as follows. First, univariate linear regression analyses were conducted by regressing the dependent variable (the degree of implementation of evidence-based programming) on each hypothesized implementation correlate (including organizational, provider, curriculum and demographic factors). Parameter estimates (standardized and raw), standard errors and p-values were calculated for each variable. Second, multivariate analyses were conducted to determine which variables were most important at each level of factors (i.e., organizational, provider, etc.). That is, predictors within each domain (e.g., among principal characteritics only) were simultaneously entered into a multivariate linear regression model, controlling for all statistically significant demographic variables. Parameter estimates (standardized and raw), standard errors and p-values were calculated for each variable by domain. Third, all statistically significant variables (p< 10) from the second-stage multivariate analyses were simultaneously entered into a multivariate linear regression model (that combined predictors from all domains), controlling for all statistically significant demographic variables. Parameter estimates (standardized and raw), standard errors and p-values were calculated for each variable. Fourth, to enable further examination of how the multiple levels of factors were related, hypothesized interactions were examined. Specifically, selected variables from different domains were entered into a multivariate linear regression model along with the corresponding hypothesized Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 81 interaction terms, controlling for all statistically significant demographic variables. Parameter estimates (standardized and raw), standard errors and p-values were calculated for each variable. Statistical Power. Statistical power analyses were conducted to determine the appropriate number of subjects (or units) needed to detect existing statistically significant effects. Based on previous research studies that have investigated relationships between school characteristics (at the organizational, provider and curriculum level) and the quality of program implementation, it was reasonable to expect correlations in this study in the .20 to .50 range (Rohrbach & Skara, in preparation, 2003). Thus, given a sample of 75 schools, study 1 had sufficient power to detect significant effects at power of .8 and at a two-sided alpha level of .05. Study Two: Longitudinal Study Design and Sample To address the primary aim of study 2, secondary analyses were conducted on all three waves of data collected from 1996 to 2000 for the school-based component of the Independent Evaluation of the California Tobacco Control Program. This study utilized data collected from samples of classroom teachers and youth in the subset of middle and high schools that had both teacher and student data available for all 3 waves. Analytic Sample fo r Study 2. This study excluded 5 grade students, most of whom were between 10 to 11 years old, because smoking prevalence rates are relatively lower in this age group, and survey items were considerably different from Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 82 those for 8th and 10* graders. Twelfth grade students were also excluded from analyses because data were collected for these students at the final wave (Wave 3) only. Further, this study included only the high schools that received grant funds (in either wave 1 and/or wave 2). Thus, the final sample was comprised of 70 schools (44 middle/junior and 26 high schools) that provided both teacher and student data at all 3 waves. Measures Over 30 items from the Independent Evaluation Wave 2 teacher and Wave 3 student self-report surveys were relevant to the analyses for study 2 (See Appendices D and E, respectively). The selected measures provided teacher self-reported data on organizational, provider and curriculum characteristics associated with adoption and implementation of evidence-based tobacco prevention curricula, as well as student self-reported data on program outcomes. The measures that were used are summarized as follows. Dependent Variables Adolescent Tobacco Use-related Program Outcomes. Student program outcome variables were derived from student self-report data and were used as dependent variables in the analyses. The outcome measures for this study included seven variables measured at two time-points to assess changes in tobacco-related beliefs, attitudes, skills and behaviors from 1996-1997 to 1998-1999 (i.e., from wave 1 to wave 3). The variables assessed were negative outcome expectancies of smoking, negative attitudes towards the tobacco industry, perceived peer norms Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 83 (prevalence) of smoking, tobacco refusal self-efficacy, lifetime (ever) smoking, thirty-day smoking and smoking quit attempts in the last year. Individual student scores were aggregated to create school means for each program outcome measure. Negative outcome expectancies of tobacco use was a composite index (ranging from 0 to 5) that averaged 5 items regarding tobacco effects (e.g., cigarette smoking makes people smell bad), each originally rated on a 4-point scale (l=no to 4=yes, definitely). Cronbach’s alpha for the negative consequences index was .77. Negative attitudes toward the tobacco industry were assessed by creating a continuous index (ranging from 0 to 4) that averaged 4 items regarding tobacco industry practices (e.g., “Tobacco companies try to get young people to start smoking by using advertisements that are attractive to young people.”), each originally rated on a 4-point scale (l=no to 4=yes, definitely). Cronbach’s alpha for the negative attitudes index was .59. The perceived peer norms (prevalence) of smoking item asked students to assess the proportion of peers who smoke monthly (0-100%); thus, responses already reflected a perceived prevalence percentage estimate (0-100%). Tobacco refusal self-efficacy was measured with an item that asked how difficult it would be for the student to say ‘no’ to a best friends’ cigarette offer (l=very hard to 4=very easy); responses were not recoded. Lifetime (ever) smoking was assessed with one question that asked students whether they had ever tried cigarette smoking, even a few puffs; responses were recoded to a dichotomous (0=no use vs. l=ever use) variable representing the proportion of youth who ever smoked a cigarette in their lifetime. Thirty-day Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 84 cigarette smoking was measured by asking students about the number of days in the past 30 days that they smoked cigarettes (l=none to 7=all 30 days); responses were recoded to a dichotomous (0=no use vs. l=any use) variable representing the proportion of youth who smoked a cigarette on at least one day during the 30 days prior to the survey. Smoking quit attempts were restricted to those students who reported they had smoked 100 cigarettes or more in their lifetime. The quit attempt item asked whether students had tried in the past year to quit smoking for longer than one day; responses were recoded to a dichotomous (0=no or l=yes) variable representing the proportion of ever-smokers who attempted quitting. Independent Variables Implementation o f Evidence-based Tobacco Use Prevention Programming. As in Study 1 of this dissertation, a standardized implementation index was created for each school based on the degree of implementation of evidence-based tobacco use prevention programming. The index was based on 5 of 7 of the guidelines from the CDC Guidelines for School Health Programs to Prevent Tobacco Use and Addiction (1994; see Table 1). However, the index differed from the one used in Study 1 because Study 2’s index: (a) was comprised of data from the 1997-98 school year (Wave 2); and (b) individual teacher scores for the implementation index were aggregated to create school mean scores; thus, each school had a score based on responses from 1 to 5 teachers at each school. For details, please refer to pages 63-65 of this dissertation. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 85 School Capacity to Implement Innovative Programming. A standardized capacity to implement innovative programming index was created for each school to assess the extent to which schools possessed the potential to implement innovative programming during the 1997-98 school year (Wave 2). Previous research indicates that organizations differ considerably in terms of their ability and willingness to change, evolve, and adapt to the need for continuous change. Thus, an organization’s capacity to change is an indicator of its ability to adopt as well as use technological innovations. In the case of schools, various factors (e.g., principal leadership, school climate) have been identified that could affect a school’s capacity to change, and therefore, its ability to adopt and implement effective tobacco use prevention programming. Organizational-, provider- and curriculum-level factors that affect program adoption and implementation variables were selected to create the school capacity to implement innovative programming index. The school capacity to implement innovative programming index was comprised of teacher self-report data collected during wave 2, and was created in the following manner. First, analyses were limited to data derived only from 8& -grade teachers who taught health and/or science, which are the middle school subject areas in which tobacco lessons are typically taught. Similarly, analyses were restricted to include data obtained only from 10*-grade teachers who taught health and/or physical education, which are the most common subject areas for tobacco instruction at the high school level. Second, to create the capacity to implement innovative programming index, items from the surveys were selected, when available, to Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 86 represent a construct that was based on previous theoretical and empirical research findings, as well as all statistically significant univariate predictors in Study 1. Third, for each element of the index that was created from multiple measures (i.e., items), internal consistency was calculated using Cronbach’s coefficient alpha to determine if items should be combined. Items that did not show good internal consistency with the remainder of the scale (alphas below .60) were dropped. Items that showed good internal consistency (alphas above .60) were selected for inclusion in the index. Fourth, Cronbach’s coefficient alpha was calculated for the final capacity to implement factor, to determine how well the elements fit together within the index. Fifth, individual teacher scores for each school capacity to implement innovative programming index were aggregated to create school mean scores; thus, each school had a score based on responses from 1 to 5 teachers at each school. Sixth, all elements of the index were standardized to a mean of 0 and a standard deviation of 1 (to ensure that equal weight was given to each element) and summed to create a final continuous index score for each school. Seventh, analyses were conducted on observations having non-missing data on all of the variables used to create the index. The final capacity to implement innovative programming index consisted of 7 variables that were selected to represent a construct that was based on previous theoretical and empirical research findings, as well as all statistically significant correlates in Study 1. The variables included were school leadership (specific to tobacco use prevention education), school commitment to teach tobacco use prevention education, school support for teaching tobacco use prevention, school Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 87 climate, school collaboration, teacher expectations and values regarding tobacco use prevention programming, and teacher self-efficacy to implement tobacco use prevention programming. Of these 7 variables selected for the index, four were found to be statistically significant correlates of adoption and implementation of evidence-based tobacco use prevention curricula in Study 1, including school leadership, school collaboration, teacher expectations and values regarding tobacco use prevention programming, and teacher self-efficacy to implement tobacco use prevention programming. Cronbach’s alpha for the final capacity to implement innovative programming index was .62. School leadership (specific to tobacco use prevention education) was measured with two items derived from the teacher survey. The first item asked, “Does your school-site administration expect you to teach tobacco prevention lessons as part of your curriculum. Response options included “Yes,” “No,” and “I’m not sure.” The second item asked a similar question that was focused on the school district administration, “Does your school district administration expect you to teach tobacco prevention lessons as part of your curriculum. Response options included “Yes,” “No,” and “I’m not sure.” The correlation coefficient (r) for this two-item index was .91, p< 0001. School commitment to teach tobacco use prevention education was assessed with one item that asked teachers about the priority that tobacco prevention education has at their school (originally scored on a 5-point scale ranging from l=the highest priority to 5=the lowest priority). School support for teaching tobacco use prevention was derived from two items, the first focusing on Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 88 the support of the school-site administration, and the latter on the school district administration. These two items were: “To what extent has your school-site administrators) supported you in your teaching of tobacco prevention lessons?” and “To what extent has your school district administration supported you in your teaching of tobacco prevention lessons?” Responses for both items were given on a 4-point scale ranging from l=a great deal to 4=mot at all. The correlation coefficient (r) for this two-item index was .74, p< 0001. To assess school climate, 4 items from the teacher questionnaire were used. Each item asked the teacher to indicate (on a 5 point-scale ranging from l=strongly agree to 5=strongly disagree) to what extent they agreed with the following statements. “Teacher morale is high at our school this year,” “Teachers at this school demonstrate a great deal of school spirit,” “Students at this school feel teachers are on their side,” and “Teachers at this school are not burned out.” Cronbach’s alpha for the climate index was .85. School collaboration was assessed with one item. Teachers were asked to indicate (on a 4 point-scale ranging from l=a great deal to 4=not at all) to what extent had their colleagues or a master teacher supported them in their teaching of tobacco prevention lessons. To assess teacher expectations and values of tobacco prevention programming, teachers were asked one item that indicated to what extent they feel that tobacco prevention education is a valuable use of student time (originally scored on a 4-point scale ranging from l=very valuable to 4=not at all valuable). Teacher perceived self-efficacy to implement tobacco prevention programming was assessed with one item that asked, “Overall, to what Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 89 extent do you feel you are prepared to teach tobacco prevention lessons?” Responses were given on a 4-point scale ranging from l=a great deal to 4=not at all. School Demographic-level Variables. As in Study 1, school-district and school-level demographic data on the sample of schools were obtained from three sources. However, demographic data differed from that used in Study 1 because Study 2’s student ethnicity (percentage of non-white student composition) and school type (high or junior/middle school) data were derived from the Independent Evaluation youth data sets for Wave 2. For details, please refer to pages 75-76 of this dissertation. Data Analysis Plan Data Analysis Overview. A series of analyses were conducted on secondary data (as described below) to examine the importance of the interactive effect between the degree of implementation of effective tobacco use prevention programming and the newly-developed school capacity to implement innovative programming construct, with subsequent changes in adolescent tobacco use-related outcomes (from 1996 to 2000). All analyses were conducted using school as the unit of analysis. Ultimately, this study’s purpose was to draw conclusions about the effects of program implementation factors on program outcomes among adolescents, rather than to generalize to California as a whole; thus, unweighted analyses were conducted. All statistical analyses were performed using the SAS System (version 8.02) software program (SAS Institute, 2001), and all testing was carried out at the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 90 traditional .05 significance level unless otherwise noted. For all variables of interest, univariate frequencies were run to detect outliers, and inconsistent or missing data. Attrition Analyses. Attrition analyses were conducted to examine differential attrition from the study on all school demographic variables. Schools that provided complete data for all 3 waves (i.e., the analytic sample of 70 middle and high schools) were compared to all other schools (that did not provide data for all 3 waves) on Wave 2 data. This type of analysis helps to determine how representative the analytic sample is of the population of schools. T-tests and GLM procedures were used to compare 70 schools used versus 73 schools not used in the analyses. Sample Characteristics. Demographic statistics were calculated for the analytic sample (n=70), by school, teacher and student groups. That is, for categorical variables, frequencies and percentages were computed; and for continuous measures and scales, means and standard deviations were computed. Descriptive Analyses o f Variables. Descriptive statistics were computed for each of the dependent and independent variables. That is, means, standard deviations, ranges and sample sizes were calculated, when applicable. All distributions were examined for adequate variability and possible violations of normality assumptions (by skewness and kurtosis). Results indicated that all variables were normally distributed, thus, transformations or robust statistical methods were not considered in all further bivariate or multivariate analyses. Further, for the dependent variables, means, standard deviations and ranges were calculated for the 70 schools at Wave 1 and at Wave 3. T-tests were then performed to analyze Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 91 whether statistically significant differences existed between the two waves for any of the seven adolescent tobacco use-related program outcome variables. Pearson correlation analyses were also conducted to detect relationships among all dependent and independent measures. Hypothesis Testing. To test the hypotheses, multivariate analyses were conducted in two stages for each of the seven student outcome variables. First, using the general linear model (GLM) procedure, each program outcome variable was regressed on each independent variable (implementation of evidence-based tobacco use prevention programming and school capacity to implement innovative programming indices; Wave 2) while including the baseline (Wave 1) outcome variable (i.e., conditional model). All of these models also included the covariates of population density, student enrollment, poverty, percentage non-white students and school type. Second, multivariate regression models, using the GLM procedure, were employed to examine the hypothesized moderating effect of school capacity to implement innovative programming on the relationship between implementation of evidence-based tobacco use prevention programming and each program outcome variable, controlling for the corresponding baseline (Wave 1) outcome variable and all five of the demographic covariates. Statistical Power. Insufficient power may result in a Type II error, whereby there is failure to reject the null hypothesis when indeed it should be rejected. Therefore, statistical power analyses were conducted to determine the appropriate number of subjects (or units) needed to detect existing statistically significant effects. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Based on previous research studies that have investigated relationships between program implementation and program outcomes, it is reasonable to expect correlations in this study in the .20 to .30 range (Rohrbach et al, 1993, 1998). Thus, given a sample of 70 schools, study 2 had sufficient power to detect significant effects at power of .8 and at a two-sided alpha level of .05. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 93 RESULTS Overview This dissertation investigated the diffusion process of diffusion of evidence- based tobacco use prevention programming through use of two studies that were comprised of two separate samples. Study 1 examined the adoption and implementation phases, by focusing on identification of variables (i.e., organizational, provider, curriculum and demographic) associated with the adoption and implementation of evidence-based tobacco use prevention curricula and programming. This first study utilized a cross-sectional sample comprised of both teachers and principals. Study 2 focused on the adoption and implementation phases, along with the program outcome phase, by examining the effect of an interaction between a newly-developed school capacity to implement innovative programming construct (derived from organizational-, provider- and curriculum-level variables) and the degree of implementation of evidence-based tobacco use prevention programming, on changes in adolescent tobacco use-related beliefs, attitudes, skills and behaviors. This second study utilized a longitudinal sample (at the school level) comprised of both teachers and students. Study One: Cross-sectional Study Attrition Analyses Table 4 shows the demographic characteristics of schools that were included and excluded from the study. In order to assess the extent to which the analytic sample was representative of all schools surveyed, schools that provided complete Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 94 data from both teachers and principals (i.e., the analytic sample of 75 junior/middle and high schools) were compared to all other schools (that did not provide data from both teachers and principals; n=36 junior/middle and high schools). Results indicated that no significant differences existed on any of the school demographic characteristics across these groups. However, schools used in the analyses (i.e., the analytic sample) had slightly lower mean scores for population density (0=rural, l=suburban or 2=urban; M = 1.16 vs. M= 1.19 and SD=0.64 vs. SIM).67, respectively), school enrollment size (M = 399.71 vs. M= 410.14 and SD=227.95 vs. SD=248.40, respectively), and school type (0=high, 1 junior/middle school; M = 38.67 vs. M= 44.44 and SD=49.03 vs. SD=50.04, respectively) compared to schools not used in the analyses. Further, schools used in the analyses also had a lower proportion of students eligible to receive free or reduced cost lunch (i.e., a proxy measure for socioeconomic status; 36.02% vs. 42.58%), and a higher proportion of non-white student composition (50.53% vs. 45.15%) when compared to schools not used in the final analyses. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 95 Table 4. Study One: Comparison of Characteristics of Schools Used in Analyses vs. Schools Not Used in Analyses_______________ __________________________ Schools Used in Analyses (N=75) Schools Not Used in Analyses (N-36) Variable Population density (0=rural, l=suburban, 2=nrban) Mean (SD) 1.16 (.64) 1.19 (.67) School size (number of students enrolled) Mean (SD) 399.71 (227.95) 410.14 (248.40) Socioeconomic status (percentage of students eligible to receive free or reduced cost lunch) % 36.02 42.58 Ethnic minority (percentage of non white students) % 50.53 45.15 School type (0=high, l^junior/middle) Mean (SD) 38.67 (49.03) 44.44 (50.40) *p<05 Sample Characteristics School Data. The full analytic sample consisted of 75 schools that provided data from both teachers and principals. Table 5 shows the demographic characteristics of the full analytic sample of schools. The majority (58.67%) of the schools were located in suburban areas, followed by urban (28.00%) and rural (13.33%) areas. The mean number of students enrolled in these schools was 399.71 (SD=227.95). Over one-third (36.02%) of the students in these schools were eligible to receive free or reduced cost lunches. Approximately half (50.53%) of the students Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 96 at these schools were of minority (i.e., non-white) ethnicity, including Black, Hispanic, Asian, and American Indian backgrounds. Sixty percent (n=45) of the schools were high schools and 40% (n=30) were junior/middle schools. Table 5. Study One: Demographic Characteristics of the Analytic School Sample tn=75 Schools)____________________________________________________ Variable N (%) Mean (SD) Population Density Rural 10(13.33) Suburban 44 (58.67) Urban 21 (28.00) School size (number of students enrolled) 399.71 (227.95) Socioeconomic status (percentage of students eligible to receive free or reduced cost lunch) (36.02) Ethnic minority (percentage of non-white students) (50.53) School Type High school 45 (60.00) Junior/middle school 30 (40.00) Teacher Data. The demographic characteristics of the teacher sample (n=75) at these schools are provided in Table 6. The sample included only those respondents who reported that they had taught tobacco prevention education to students in the classroom, including classroom teachers (84.00%), school TUPE Coordinators (6.67%), and those occupying other (9.33%) positions (e.g., school nurse). Almost Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 97 two-thirds (62.50%) of the participants were female. The sample was predominantly white in ethnicity (approximately 79%), with only a small representation of Hispanics (10.67%), Blacks (2.67%), Asians (4.00%), and Others (4.00%). Participants reported a mean age o f45.66 years (SD=10.87). The vast majority (83.10%) of the teachers had 3 or more years of experience in teaching tobacco use prevention education at any school, with a mean of 10.04 years (SD=9.53). In terms of teaching tobacco use prevention education at their current school, a mean of 7.73 years (SD=8.30) was reported for the sample. Overall, the participants reported high educational achievement, with approximately 54% of the sample reporting earning a Bachelor’s degree, and approximately 43% reporting having a Master’s degree. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 98 Table 6. Study One: Demographic Characteristics of the Analytic Teacher Sample (n-75) Variable N (%) Mean (SD) Position Classroom Teacher 63 (84.00) School TUPE Coordinator 5 ( 6.67) Other 7 ( 9.33) Gender Male 27 (37.50) Female 45 (62.50) Ethnicity White 59 (78.67) Hispanic 8 (10.67) African-American 2 ( 2.67) Asian 3 ( 4.00) Other 3 ( 4.00) Age (years) 45.66 ( 10.87) 20-30 6 ( 8.10) 31-40 22 (29.73) 41-50 14 (18.92) 51-60 28 (37.84) 61+ 4 ( 5.41) Years of teaching tobacco use prevention at any school 10.04 ( 9.53) 0-2 12 (16.90) 3-5 23 (32.39) 6-8 7 ( 9.86) 9-19 15(21.13) 20+ 14 (19.72) Years of teaching tobacco use prevention at this school 7.73 ( 8.30) 0-2 19 (26.76) 3-5 22 (30.99) 6-8 9(12.68) 9-19 11 (15.49) 20+ 10 (14.08) Highest degree B.A./B.S. 40 (54.05) M.A. /M.S. 32 (43.24) Ph.D. / Ed.D. 2 ( 2.70) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 99 Principal Data. Table 7 presents demographic information for the principal sample (n=75). The sample consisted of principals (66.67%), assistant principals (24.00%) and those occupying other (9.33%) positions, including deans and guidance counselors. Almost two-thirds (63.51%) of the principal sample was male. Similar to the teacher sample, the principal sample was predominantly white in ethnicity (approximately 81%), with only a small representation ofHispanics (10.67%), Blacks (5.33%), and Asians (2.67%). Participants reported a mean age of 50.0 years (SD=8.02). The vast majority (86.30%) reported having their position (e.g., principal or assistant principal) at any school for 3 years or more, with a sample mean of 8.15 years (SD=6.85). In terms of holding their position at their current school, a mean of 4.99 years (SD=4.58) was reported for the sample. Educational achievement was very high in this group, with over 95% reporting earning a Master’s degree or greater. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 0 0 Table 7. Study One: Demographic Characteristics of the Analytic Principal Sample (n=75) Variable N (%) Mean (SD) Position Principal 50 (66.67) Assistant Principal 18 (24.00) Other 7 ( 9.33) Gender Male 47(63.51) Female 27 (36.49) Ethnicity White 61 (81.33) Hispanic 8 (10.67) African-American 4 ( 5.33) Asian 2 ( 2.67) Other 0 ( 0.00) Age (years) 50.00 ( 8.02) 20-30 1 ( 1.56) 31-40 8(12.50) 41-50 18(28.12) 51-60 35 (54.69) 61+ 2 ( 3.13) Years of position at any school 8.15 ( 6.85) 0-2 10 (13.70) 3-5 23 (31.50) 6-8 16(21.92) 9-19 17 (23.29) 20+ 7 ( 9.59) Years of position at this school 4.99 ( 4.58) 0-2 23 (33.33) 3-5 24 (34.78) 6-8 13 (18.84) 9-19 8(11.59) 20+ 1 ( 1.45) Highest degree B.A. /B.S. 3 ( 4.22) M.A. /M.S. 59 (83.10) Ph.D. / Ed.D. 9(12.68) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 101 Descriptive Analyses of Variables Descriptive Analyses of Dependent Variables Adoption and Implementation o f Evidence-based Tobacco Use Prevention Curricula. The implementation or non-implementation of evidence-based tobacco use prevention curricula was classified based on teacher data indicating which commercial or published curricula they had used the most recent time they taught tobacco use prevention lessons in the classroom. Table 3 presents the prevalence of use of each curriculum. As shown, relatively few teachers indicated they used an evidence-based curriculum. Two teachers indicated that they had selected two of the evidence-based curricula, and one teacher indicated that he/she had selected three of the evidence-based curricula, yielding a total of 15 teachers (only 20% of the analytic sample) who selected at least one evidence-based curriculum. The evidence- based curriculum used most often by teachers was Project ALERT (9.33%; n=7), followed by Life Skills Training (8.00%; n=6), Project TNT (4.00%; n=3), Quest: Skills fo r Adolescence (2.67%; n=2) and Minnesota Smoking Prevention Program - Tobacco Free Teens (1.33%; n=T). None of the teachers reported using Reconnecting Youth: A Peer Group Approach to Building Life Skills. The majority of teachers used curricula that were not evidence-based, with 33.33% (n=25) of the teachers using a curriculum classified as “other.” This category mostly included school health textbooks and district-developed curricula. The second most-used curriculum, selected by 30.67% (nr=23) of the teachers, was Tobacco Free! Tobacco Use Prevention LeMOMs, wfficti are infusion lessons that the state of Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 102 California developed but did not evaluate. The three other curricula most frequently selected by teachers were Here’ s Looking at You, 2000 (22.67%; n=T7), DARE (16.00%; n=12) and Get Real About Tobacco (13.33%; n=10). Adoption and Implementation o f Evidence-based Tobacco Use Prevention Programming. The degree of implementation of evidence-based tobacco use prevention programming was measured with a continuous index based on five of the seven CDC Guidelines for School Health Programs to Prevent Tobacco Use and Addiction (Centers for Disease Control and Prevention, 1994; see Table 1). This second dependent variable was more general in scope than the first because it represented a composite of the critical elements of effective tobacco use prevention programming, in and out of the classroom, including policy, curriculum delivery and teacher training. Table 8 shows the means (with standard deviations) or percentages for each of the five variables that comprised the standardized index. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 103 Table 8. Study One: Means (SD) and Percentages for Variables Used in Adoption and Implementation of Evidence-based Tobacco Use Prevention Programming (Continuous) Composite Index (n=51 Schools'!_______ __________ _________ Range Mean (SD) or % Variable Policy Enforcement 1-4 3.59 (.61) Consequences 0-2 .53 (.50) Tobacco Instruction Number of tobacco use prevention lessons taught 0-15 4.45 (3.23) Taught recommended content 0-5 4.45 (.92) Program-specific training (% yes) 0-100% 19.61 Parent involvement 1-4 1.75 (.77) On-site cessation program (% yes) 0-100% 35.29 Adoption and implementation of effective tobacco use prevention programming index (standardized) Cronbach’s alpha =.62 -9.32 to 7.95 0.00 (3.41) Regarding tobacco use policy enforcement, teachers reported relatively high mean levels (M=3.59 on a scale of 1-4) of this variable, indicating that teachers perceived that their school’s no-tobacco use policy was being enforced a great deal. The mean score for policy violation consequences, however, was relatively low (0.53 on a scale of 0-2), indicating that principal’s reported that students most often received punitive-type (e.g., suspension or parents are called) rather than remedial- type (e.g., required to attend a special tobacco education class) consequences for being caught smoking or using smokeless tobacco at school. As for tobacco Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 104 instruction, teachers reported teaching an average of 4.45 tobacco use prevention lessons to their students, and these same teachers also indicated that they taught the majority of recommended instructional content (M=4.45 on a scale of 0-5; a sum of physiological effects, social influences, social consequences, peer norms, and refusal skills; scored 0=no or l=yes). The prevalence of teacher training in tobacco use prevention was particularly low, with only 19.61% of the teachers reporting that they had received program-specific tobacco use prevention in-service training. These teachers also reported relatively low mean levels (M=T .75 on a scale of 1-4) for trying to get students’ parents involved in tobacco use prevention education. Further, 35.29% of teachers reported that their schools had on-site classes, groups, or programs for students who wanted to quit using tobacco. The final index had a Cronbach’s alpha of .62 and scores that ranged from -9.32 to 7.95 (with a mean of 0.00). Descriptive Analyses of Independent Variables Table 9 presents a summary of the means, standard deviations, and ranges of the correlates of adoption and implementation of evidence-based tobacco use prevention curricula for all 75 schools combined, as well as by use or non-use of an evidence-based tobacco use prevention curriculum. For the most part, across the total sample of 75 schools, principals and teachers reported relatively high levels of the various organizational-, provider- and curriculum-type variables that were hypothesized to be associated with the adoption and implementation of evidence- based tobacco use prevention curricula and programming. For principal Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. characteristics, participants reported relatively high levels of leadership of tobacco use prevention education (M = 3.80 on a scale of 1-5) and commitment to teach tobacco use prevention education (M = 3.03 on a scale of 1-4.33), whereas knowledge of effective tobacco use prevention programs was reported to be only slightly above the mean midpoint (M = 1.37 on a scale of 1-2.5). Organizational structure and processes variables were also reported as existing at relatively high levels for the full sample, with means well above the midpoint on decentralized management (M = 3.17 on a scale of 1-4.33), participatory decision-making (M = 3.96 on a scale of 1-5), communication (M = 3.86 on a scale of 1-5) and adequate resources to teach tobacco use prevention education (M = 2.17 on a scale of 0-3). Regarding school culture, on a scale from 1-5, participants were well above the midpoint on mean levels of shared vision/goals (M = 4.02), innovativeness (M = 3.65), climate (M = 3.93) and collaboration (M = 4.07). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 106 Table 9. Study One: Means and Standard Deviations of All Analytic Correlates of Adoption and Implementation of Evidence-based Tobacco Use Prevention Curricula (n=75 Schools) ___________ _________________ __________ _______ All Schools N=75 Schools Not Used Evidence- based Curriculum Schools Used Evidence- based Curriculum N=15 Range Mean (SD) Mean (SD) Mean (SD) © 1 n 1 1 Principal Leadership of tobacco use prevention education 3.80 (.86) 3.67 (.85) 4.36 (.69)** 1-5 Commitment to tobacco use prevention education 3.03 (.61) 2.96 (.58) 3.32 (.67)* 1-4.33 Knowledge of effective tobacco use prevention programs 1.37 (.47) 1.32 (.44) 1.57 (.53) 1-2.5 Structure & Processes Management (decentralized) 3.17 (.67) 3.12 (.65) 3.36 (.75) 1-4.33 Participatory decision-making 3.96 (.68) 3.91 (.69) 4.13 (.66) 1-5 Communication 3.86 (.78) 3.83 (.73) 3.97 (.95) 1-5 Adequate resource (training, materials, time) 2.17 (.91) 2.15 (.94) 2.27 (.80) 0-3 School Culture Shared vision/goals 4.02 (.61) 3.90 (.57) 4.51 (.50)*** 1-5 Innovativeness 3.65 (.75) 3.53 (.74) 4.13 (.59)** 1-5 Climate 3.93 (.52) 3.88 (.51) 4.13 (.51) 1-5 Collaboration 4.07 (.68) 3.98 (.68) 4.42 (.60)* 1-5 Provider-level Components Commitment to tobacco use prevention education 3.10 (.68) 3.08 (.70) 3.18 (.60) 1-4.33 Knowledge of effective tobacco use prevention programs .15 (.39) .12 (.32) ?7 ( 59) 0-2 Perceived mandate to use effective tobacco use prevention curricula 2.41 (1.14) 2.16 (1.13) 3.13 (.83)** 1-4 Expectations and values regarding tobacco use prevention programming 3.00 (.61) 2.89 (.60) 3.33 (.49)* 1-4 Self-efficacy to implement tobacco use prevention programming 3.56 (.45) 3.49 (.48) 3.77 (.26)* 1-4 Curriculum-level Comnonents Provider perception that curriculum has been demonstrated or endorsed as being effective 3.14 (.72) 3.04 (.75) 3.35 (.66) 1-4 Utilized teaching methods familiar to provider, minimal training required 3.17 (.62) 3.23 (.64) 3.04 (.56) 1-4 Compatibility with interactive instructional methods 3.06 (.62) 2.97 (.61) 3.33 (.59) 1-4 Note. Tested for statistically significant differences between schools that had not used an evidence- based curriculum vs. schools that had used an evidence-based curriculum. *p<.05, **p<.01, ***p<001 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 107 Regarding provider-level components, respondents reported high levels on four of the five variables, with means well above the midpoint on commitment to teach tobacco use prevention education (M - 3.10 on a scale of 1-4.33), perceived mandate to use effective tobacco use prevention curricula (M = 2.41 on a scale of 1- 4), expectations and values regarding tobacco use prevention programming (M = 3.00 on a scale of 1-4), and self-efficacy to implement tobacco use prevention programming (M = 3.56 on a scale of 1-4). However, knowledge of effective tobacco use prevention programs was reported as being particularly low (M = 0.15 on a scale of 1-2.5), suggesting that teachers were much less informed on this subject than principals. For curriculum-level components, the sample was generally well above the midpoint on mean levels of provider’s perception that the curriculum has been demonstrated or endorsed as being effective (M= 3.14 on a scale of 1-4), utilized teaching methods familiar to provider (M= 3.17 on a scale of 1-4), and provider regularly uses skills compatible with interactive instructional methods (i.e., discussion, activities, role playing) used to deliver the curriculum (M= 3.06 on a scale of 1-4). Overall, compared to schools that had not used an evidence-based curriculum (n=60), schools that had used an evidence-based curriculum (n=15) reported higher mean scores for all 19 correlates (organizational-, provider- and curriculum-level) except one, curriculum utilized teaching methods familiar to provider (minimal training required). Specifically, regarding principal characteristics, schools that had used an evidence-based curriculum reported significantly higher levels of leadership Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 108 of tobacco use prevention education (M - 4.36 vs. M =3.67, p< 01) and commitment to teach tobacco use prevention education (M = 3.32 vs. M = 2.96, p<05) compared to schools that had not used an evidence-based curriculum. Knowledge of effective tobacco use prevention programs was also higher among those schools that had used an evidence-based curriculum, although the difference was not statistically significant (M= 1.57 vs. 1.32). All four organizational structure and processes variables were also reported at higher levels (although not statistically significant) for schools that had used an evidence-based curriculum versus schools that had not used an evidence-based curriculum, including decentralized management (M = 3.36 vs. M = 3.12), participatory decision-making (M = 4.13 vs. M = 3.91), communication (M = 3.97 vs. M = 3 .83) and adequate resources to teach tobacco use prevention education (M = 2.27 vs. M = 2.15). Regarding school culture, shared vision/goals (M = 4.51 vs. M = 3.90, p< 001), innovativeness (M = 4.13 vs. M = 3.53, p< 01) and collaboration (M = 4.42 vs. M = 3.98, p<05) were significantly higher for schools that had used an evidence-based curriculum. Climate was also higher among these schools (M = 4.13 vs. M = 3.88); however, the difference was not statistically significant. Regarding provider-level components, perceived mandate to use effective tobacco use prevention curricula (M = 3.13 vs. M = 2.16, p<01), expectations and values regarding tobacco use prevention programming (M = 3.33 vs. M = 2.89, p< 01), and self-efficacy to implement tobacco use prevention programming (M = 3.77 vs. M = 3.49, p< 01) were all significantly higher among schools that had used Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 109 an evidence-based curriculum compared to schools that had not used an evidence- based curriculum. Commitment to teach tobacco use prevention education (M = 3.18 vs. M = 3.08) and knowledge of effective tobacco use prevention programs (M = 0.27 vs. M = 0.12) were also higher among the schools that had used an evidence- based curriculum; however, these differences were not statistically significant. For cuniculum-level components, schools that used an evidence-based curriculum reported higher mean levels (although not statistically significant) on the provider’s perception that curriculum had been demonstrated or endorsed as being effective (M = 3.35 vs. M = 3.04) and provider’s use of skills compatible with interactive instructional methods (i.e., discussion, activities, role playing) used to deliver the curriculum (M = 3.33 vs. M = 2.97), compared to schools that had not used an evidence-based curriculum. As indicated above, among all of the correlates, there was only one (i.e., curriculum utilized teaching methods familiar to provider) that had a higher mean score for schools that had not used at least one evidence- based curriculum compared to schools that had used at least one evidence-based curriculum (M = 3.23 vs. M = 3.04); however, the difference was not statistically significant. Correlations among All Analytic Variables Table 10 shows correlations among the analytic variables, including all dependent, independent and demographic variables. Inspection of these correlations revealed statistically significant associations between the dependent and independent variables. Regarding correlations among the independent variables, many Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 1 0 statistically significant associations existed among variables in the same domain (i.e., organizational, provider, curriculum and demographic) as well as many significant associations among variables across domains. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 10. S tudy One: Intercorrelations o f A ll Analytic Variables (n=75 Schools). I ll . P r s i . ,• * a n o q 2^00 0 o o eq so eq h v aq V © eq « n 8 q a a q S 3 e q S3 e q # 00 r- O s O O n * t- O «s 9 9 O © M * * * ^ * m in . * ■ * ■ ^ o r - * ^ . a n . . . (S w . P C ? n $ « n s - „ * nj • • P * P V O V O fn . . ■ « * ■ » * * * * * * * f Nm « N V O r- 00 * * “« C N • m - s - I 5 © O P tS P o o o 2 . cs fS V O a n o o o q o •8- # « • * \© * * * o * * 0 0 © v « N ■ * ? $ • ro * t * * « n T t * ^ 2 V O O O O « — I ■ 9 © c * q © < n arj * £ > * n o o o g o * ■ # O s O m - & • ■ & - s . • & < y * } I S g S g S o o -t o <s °. ,• 9 ^ rt - « t v o « n 8 * a n 3- « n C N anOOt^CV^-a^Ovv-HOO^vOanTt iOcNq©«— '©q©?-**^* V O « ■ ■ » ■ s - i t & ~ ■ & 3 0 , * n o o n 9 9 « n ‘ * m © v ^ o o cn o ^ < N v o a n v o h q o o o < n » n c -4 i— * * * - * V O o < N .* * # * - i f - * f - H ^ * V O* rf < N O 0 0 0 0 T f e s i q * q c n * * « ■ ■ K - ■ K - * r- r < 5 ■ & • % > .16 ^ s < " ! o n .03 « - 0 0 ■ S ' - i i - V O * r- < £ O O a n m m n « n n O O C S O o es o v \o -h o 9 1 — < o O p - a * ” 1 a - a V O V O " 4 " ® 5 < n o m * — • . * v * * P cn r ~ * ^ s - o © o © * “ H C N | O » " H # „ V O « ■ 5^ - H « S ! “ * m C N M ’ t n r" « n N « H - H o o " S f r * O es co * ' * • * * * © © * s ■ & n n w m * »n * - f r vo «n ^ * O * - < oo es * < n < n ’ — a 3 L c n q O .* <N t - N '* # “ * S an O < N an MO-HO ""t * N l s j s a s s s 2 ^ g ^ s 2 i § 3 r , r» BqeqeqecieqfiqBqspeqeqsqoO Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission. Cl C2 C3 C 4 cs D1 D2 D3 El E 2 E3 E4 ES A1 A2 El B2 S3 B4 BS B6 B7 BS B9 B IO B ll Cl C2 .17 - C3 .22 .39** - C 4 .35** .39** . C S .03 .23 .15 .34** - 1 )1 -.12 -.05 .09 .01 -.31* - D 2 -.06 -.04 -.01 -.24 .04 .14 - D3 .05 .07 .09 .36** .42*** .31* -.03 - El .14 .13 .36** .08 .08 .01 -.08 -.21 - E2 .12 .02 .15 .11 -.13 .05 -.01 -.38** .56*** - E3 .10 -.18 -.06 .00 -.09 .26 -.02 .15 .18 .02 E4 -.15 .13 -.02 -.09 .06 -.20 -.01 .02 -.45*** -.38*** E S .12 -.10 ,11 .11 .12 -.12 ,05 .28* -.06 -.29** -.48*** .47*** -.12 *p<.05, **p<.01, ***p<.001 Dependent variables A1 Adoption and implementation of evidence-based tobacco use prevention curriculum A2 Adoption and implementation of effective tobacco use prevention programming to Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission. Table 10. Study One: Intercorrelations of All Analytic Variables (n=75 Schools'). Continued. Independent variables (organizational-, provider-, and curriculum level components’ ) Organizational- level Components B1 Leadership of tobacco use prevention education B2 Commitment to tobacco use prevention education B3 Knowledge of effective tobacco use prevention programs B4 Management (decentralized) B5 Participatory decision-making B6 Communication B7 Adequate resource (training, materials, time) B8 Shared vision/goals B9 Innovativeness BIO Climate B ll Collaboration Provider- level Components Cl Commitment to tobacco use prevention education C2 Knowledge of effective tobacco use prevention programs C3 Perceived mandate to use effective tobacco use prevention curricula C4 Expectations and values regarding tobacco use prevention programming C5 Self-efficacy to implement tobacco use prevention programming Curriculum- level Components D 1 Provider perception that curriculum has been demonstrated or endorsed as being effective D2 Curriculum utilized teaching methods familiar to provider, minimal training required D3 Compatibility with interactive instructional methods PemograBhic variables E1 Population density (rural, suburban or urban) E2 School size (number of students enrolled) E3 Socioeconomic status (percentage of students eligible to receive free or reduced cost lunch) E4 Ethnic minority (percentage of non-white students) E5 School type (high or junior/middle school) u > 114 Dependent Variables. Many of the organizational-, curriculum-, provider-, and demographic-level variables were significantly correlated with the dependent variables. Specifically, the binary dependent variable, adoption and implementation of evidence-based tobacco use prevention curricula, was significantly correlated with 10 independent variables, including principal leadership of tobacco use prevention education (r=.32, p< 01), principal commitment to teach tobacco use prevention education (r=.24, p< 05), shared vision/goals (r=.41, p< 001), innovativeness (r=.32, p< 01), collaboration (r=.26, p< 05), provider perceived mandate to use effective tobacco use prevention curricula (r=.38, p< 01), provider expectations and values regarding tobacco use prevention programming (r=.32, p< 01), provider self-efficacy to implement tobacco use prevention programming (r=.27, p< 05) provider regularly uses skills compatible with interactive instructional methods (i.e., discussion, activities, role playing) used to deliver the curriculum (r=.25, p<05), and school type (junior/middle or high school; r=.27, p< 05)., Furthermore, the continuous dependent variable, adoption and implementation of evidence-based tobacco use prevention programming, was significantly correlated with 6 independent variables, including adequate resources (training, materials and time; r=.43, p<01), provider commitment to teach tobacco use prevention education (r=.30, p< 05), provider knowledge of effective tobacco use prevention programs (r=.30, p<05), provider perceived mandate to use effective tobacco use prevention curricula (r=.32, p<05), provider expectations and values regarding tobacco use prevention programming (r=.52, p< 001), and provider Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 115 regularly uses skills compatible with interactive instructional methods (i.e., discussion, activities, role playing) used to deliver the curriculum (r=.29, p< 05), four of which were provider-level variables. Organizational-level Components. Most of the organizational level variables correlated well with one another. For example, principal leadership of tobacco use prevention education was significantly and positively correlated with several organizational-level variables, including principal commitment to teach tobacco use prevention education (r=.27, p< 05), principal knowledge of effective tobacco use prevention programs (r=.42, p<.001), communication (r=.59, p<001), shared vision/goals (r=.48, p< 001), innovativeness (r=.46, p< 001), climate (r=.37, p< 001) and collaboration (r=.38, p<001). As another example, communication was also significantly correlated with commitment to teach tobacco use prevention education (r=.26, p< 05), principal knowledge of effective tobacco use prevention programs (r=.58, p<.001), participatory decision-making (r=.29, p<.01), shared vision/goals (r=.44, p<.001), innovativeness (r=.39, p<.001), climate (r=.28, p<.05) and collaboration (r=.36, p< 01). Conversely, decentralized management and adequate resources (training, materials and time) were correlated with relatively few organizational-level variables. That is, decentralized management was correlated with only adequate resources (r=.25, p<G5) and innovativeness (r=.25, p< 05), and adequate resources was fiirther correlated with only innovativeness (r=.26, p<,05). The six strongest correlations among all of the 19 independent variables (including demographic variables) were found among the organizational-level Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 116 variables, particularly among the four school culture variables. Shared vision/goals was highly correlated with innovativeness (r=.62, p< 001), climate (r=.73, p< 001) and collaboration (r=.72, p<.001), while innovativeness was highly correlated with climate (r=.62, p< 001) and collaboration (r=.63, p<.001), and climate was highly correlated with collaboration (r=.78, p<.001). Further, each of the four school culture variables was consistently correlated with most all of the organizational variables, with the exception of decentralized management and adequate resources. A few of the organizational-level variables correlated with variables across domains. For example, principal leadership of tobacco use prevention education, an organizational-level variable, was significantly correlated with provider-level variables, including provider commitment to teach tobacco use prevention education (r=.48, p<001), provider perceived mandate to use effective tobacco use prevention curricula (r=.32, p< 01), and provider expectations and values regarding tobacco use prevention programming (r=.29, p< 05). Whereas innovativeness, another organizational-level variable, was significantly correlated with provider perceived mandate to use effective tobacco use prevention curricula (r=.32, p<01), provider expectations and values regarding tobacco use prevention programming (r=.25, p< 05), provider self-efficacy to implement tobacco use prevention programming (r=.29, p< 05) and provider regularly uses skills compatible with interactive instructional methods (i.e., discussion, activities, role playing) used to deliver the curriculum (r=.25, p< 05). Moreover, shared vision/goals was significantly and Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 117 positively related to school 0=high, l=junior/middle school; r=.26, p<.05), a demographic variable. Provider-level Components. The provider level variables also correlated well with one another. For example, provider expectations and values regarding tobacco use prevention programming was significantly correlated with all of the other provider-level variables, including provider commitment to teach tobacco use prevention education (r=.35, p< 01), knowledge of effective tobacco use prevention programs (r=.41, p< 001), provider perceived mandate to use effective tobacco use prevention curricula (r=.39, p< 01) and provider self-efficacy to implement tobacco use prevention programming (r=.34, p< 01). As noted above, some of the provider level variables correlated with the organizational-level variables, as well as with curriculum-level variables. For example, provider self-efficacy to implement tobacco use prevention programming was significantly and inversely related to the curriculum-level variable, provider’s perception that curriculum has been demonstrated or endorsed as being effective (r= -.31, p< 05), and significantly and positively related to provider regularly uses skills compatible with interactive instructional methods (i.e., discussion, activities, role playing) used to deliver the curriculum (r=.42, p< G Q 1). Moreover, provider perceived mandate to use effective tobacco use prevention curricula was also significantly related to population density (r=.36, p<01), a demographic variable. Curriculum-level Components. Of the curriculum-level variables, only two of the three variables were significantly correlated with one another. For instance, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 118 provider’s perception that curriculum has been demonstrated or endorsed as being effective was significantly correlated with provider regularly uses skills compatible with interactive instructional methods (i.e., discussion, activities, role playing) used to deliver the curriculum (r=.31, p< 05). As noted above, some of the provider level variables were also correlated with organizational-level variables, as well as with demographic variables. For example, provider regularly uses skills compatible with interactive instructional methods used to deliver the curriculum was significantly related to two demographic variables, school enrollment size (r= -.38, p< 01) and school type (0=high, l=junior/middle school; r=.28, p<G5). Demographic Variables. The demographic variables were found to be consistently correlated with one another. School enrollment size was significantly and positively correlated with population density (rural, suburban or urban; r= .56, p< 001), and significantly and inversely related to ethnic minority (percentage of non-white student composition; -.38, p<001) and school type (high or junior/middle; r= -.29, p< 01). Ethnic minority was also inversely correlated with population density (r= -.45, p< 001) and socioeconomic status (r= -.48, p<001). Whereas, socioeconomic status was also positively correlated with school type (r=.47, p<01). Also, as noted above, a few of the demographic variables were also correlated with organizational-, provider- and curriculum-level variables. Hypothesis Testing A four-stage regression analysis protocol was completed on cross-sectional data for both dependent variables, the adoption and implementation of evidenee- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 119 based tobacco use prevention curricula and the implementation of evidence-based tobacco use prevention programming. Adoption and Implementation o f Evidence-based Tobacco Use Prevention Curricula. The core analyses in this study assessed the adoption and implementation of evidence-based tobacco use prevention curricula in schools. This binary dependent variable was created from teacher questionnaire data and measured the implementation or non-implementation of tobacco use prevention curricula that have been identified as evidence-based. Thus, the logistic regression procedure was utilized for this entire set of models. First-stage Models. The first set of models examined the association of the dependent variable (the adoption and implementation of evidence-based tobacco use prevention curricula) and each hypothesized implementation correlate (i.e., 11 organizational-level variables, 5 provider-level variables, 3 curriculum-level variables and 5 school demographic variables). These univariate models permitted elimination (in subsequent models) of those variables that did not have a significant relationship with the adoption and implementation of evidence-based tobacco use prevention curricula. Table 11 presents the odds ratios, 95% confidence intervals, Wald’sX2 estimates, p-values, and n sizes for these relationships. The n in most of these models was 75; however, the models involving the curriculum-level variables ranged from 40-60. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 120 Table 11. Study One: Univariate Logistic Regression Results of Factors Associated with the Adoption and Implementation of Evidence-based Tobacco Use Prevention Curricula in Schools OR Cl’s Wald’s V2 P value N Orcanizational-level Comnonents Principal Leadership of tobacco use prevention education 3.60 1.37, 9.46 6.73 .01** 74 Commitment to tobacco use prevention education 3.22 1.03, 10.06 4.06 .04* 75 Knowledge of effective tobacco use prevention programs 2.87 .92, 8.98 3.27 .07 75 Structure & Processes Management (decentralized) 1.72 .71, 4.19 1.43 .23 75 Participatory decision-making 1.75 .66, 4.64 1.25 .26 75 Communication 1.26 .59, 2.66 .36 .55 75 Adequate resource (training, materials, time) 1.16 .60, 2.23 .20 .65 75 School Culture Shared vision/goals 20.44 3.90, 107.11 12.75 .00*** 75 Innovativeness 4.49 1.45, 13.86 6.81 .01** 75 Climate 2.94 .83, 10.44 2.78 .10 75 Collaboration 3.96 1.17, 13.45 4.87 .03* 75 Provider-level Components Commitment to tobacco use prevention education 1.24 .52, 2.92 .23 .63 75 Knowledge of effective tobacco use prevention programs 2.26 .65, 7.91 1.64 .20 75 Perceived mandate to use effective tobacco use prevention curricula 2.40 1.26, 4.56 7.07 .01** 58 Expectations and values regarding tobacco use prevention programming 3.90 1.26, 12.05 5.57 .02* 61 Self-efficacy to implement tobacco use prevention programming 5.62 1.05, 30.21 4.04 .04* 61 Curriculum-level Conmonents Provider perception that curriculum has been demonstrated or endorsed as being effective 1.96 .68, 5.66 1.57 .21 40 Utilized teaching methods familiar to provider, minimal training required .60 .21, 1.75 .88 .35 46 Compatibility with interactive instructional methods 3.04 .96, 9.56 3.60 .05* 60 DemoeraDhic Variables Population density (rural, suburban or urban) 2.36 .89, 6.26 2.96 .09 75 School size (number of students enrolled) 1.00 .99, 1.00 .04 .84 75 Socioeconomic status (percentage of students eligible to receive free or reduced cost lunch) 4.36 .48, 39.81 1.71 .19 75 Ethnic minority (percentage of non-white students 1.05 .13, 8.62 .00 .96 75 School type (high or junior/middle school) 4.00 1.20,13.28 5.12 .02* 75 *p< 05, **fX.01, ***p<001 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 121 A total of 10 variables (5 organizational-level variables, 3 provider-level variables, 1 curriculum-level variable, and 1 school demographic variable) were significantly associated with the adoption and implementation of evidence-based tobacco use prevention curricula. In addition, 3 other variables (2 organizational- level variables, and 1 school demographic variable) approached levels of statistical significance (p< 10). Based on unadjusted analyses, principal leadership of tobacco use prevention education (OR=3.60, p< 01), principal commitment to teach tobacco use prevention education (OR=3.22, p<05), shared vision/goals (OR=20.44, p< 001), innovativeness(OR=4.49, p<01), collaboration(OR=3.96, p<05), provider perceived mandate to use effective tobacco use prevention curricula (OR=2.40, p<.01), provider expectations and values regarding tobacco use prevention programming (OR=3.90, p<05), provider self-efficacy to implement tobacco use prevention programming (OR=5.62, p<05), provider regularly uses skills compatible with interactive instructional methods (i.e., discussion, activities, role playing) used to deliver the curriculum (OR=3.04, p<05) and school type (0=high, l=junior/middle school; OR=0.25, p< 05) were significantly associated with the adoption and implementation of evidence-based tobacco prevention curricula. As shown, among these significant variables, the highest odds ratio was found for shared vision/goals, an organizational-level (school culture) variable. Specifically, schools that had teachers and principals who scored high on the shared vision/goals index had over 20 times the odds (95% 0=3.90, 107.11) of adopting and implementing an evidence-based tobacco prevention curriculum compared to Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 122 those schools that had teachers and principals who scored low on the shared vision/goals index. Furthermore, school type was the only demographic factor that was significantly associated with the outcome variable and, therefore, was included in subsequent analyses. This association revealed that schools that were junior/middle schools were significantly more likely (OR=4.00, 95% CI=1.20, 13.28) to adopt and implement an evidence-based tobacco prevention curriculum than schools that were high schools. Second-stage Models. At the second stage, multivariate analyses were conducted to determine which variables were most important at each of the level of factors (i.e., organizational, provider, etc.). That is, predictors within each domain (e.g., among principal characteristics only) were simultaneously entered into a multivariate logistic regression model, controlling for the only statistically significant demographic variable, school type. Adjusted odds ratios, 95% confidence intervals, Wald’s X2 estimates, p-values, and n sizes were calculated for each variable by domain, along with a Wald’s X2 estimate and corresponding p-value for each overall model. The results are shown in Table 12. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 123 Table 12. Study One: Multivariate Logistic Regression Results of Factors Associated with the Adoption and Implementation of Evidence-based Tobacco Use Prevention Curricula in Schools OR C l’s Wald’s X * P value Oreanizational-level Comoonents Principal Leadership of tobacco use prevention education 3.11 .98, 9.89 3.69 .05* Commitment to tobacco use prevention education 2.22 .64, 7.69 1.58 .21 Knowledge of effective tobacco use prevention programs 2.06 .50, 8.51 1.00 .32 School type (high or junior/middle school) 4.06 .95, 17.24 3.60 .06 Overall model (n=74) 10.31 .04* Structure & Processes Management (decentralized) 2.16 •77, 6.07 2.16 .14 Participatory decision-making 1.39 .51, 3.77 .41 .52 Communication .95 .40, 2.22 .02 .89 Adequate resource (training, materials, time) .91 •43, 1.96 .05 .82 School type (high or junior/middle school) 4.85 1.30, 18.02 5.55 .02* Overall model (n=75) 7.62 .18 School Culture Shared vision/goals 41.49 3.65, 471.59 9.02 .00** Innovativeness 5.96 1.06, 33.52 4.11 .04* Climate .04 .00, .78 4.52 .03* Collaboration 1.51 19, 12.02 .15 .70 School type (high or junior/middle school) 6.66 1.23, 35.97 4.85 .03* Overall model (n=75) 15.02 .01** Provider-level ComDonents Commitment to tobacco use prevention education .50 .15, 1.69 1.23 .27 Knowledge o f effective tobacco use prevention programs .67 .09, 5.14 .15 .70 Perceived mandate to use effective tobacco use prevention curricula 2.32 1.04, 5.12 4.29 .04* Expectations and values regarding tobacco use prevention programming 4.76 ■ 5 9 , 38.28 2.15 .14 Self-efficacy to implement tobacco use prevention programming 2.99 .29, 30.99 .84 .36 School type (high or junior/middle school) 5.73 1.13, 29.00 4.46 .03* Overall model (n=5§) 11.42 .08 Curriculum-level Comoonents Provider perception that curriculum has been demonstrated or endorsed as being effective 1.95 .42, 9.11 .72 .40 Utilized teaching methods familiar to provider, minimal training required .44 .09, 2.14 1.04 .31 Compatibility with interactive instructional methods 6.11 .68, 54.99 2.61 .11 School type (high or junior/middle school) 4.03 .65, 24.96 2.25 .13 Overall model (n=38) 7.87 .10 *p<.05, **p<.01, ***p<001 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 124 The second stage of analysis produced results that differed from those of the first stage. For example, regarding (organizational-level) principal characteristics, only principal leadership of tobacco use prevention education remained statistically significant (OR=3.11, p< 05) in the multivariate model which combined all principal characteristics; whereas, principal commitment to teach tobacco use prevention education and school type were no longer significant (OR=2.22 and OR=4.06, p=.21 and p=.06, respectively). With regard to school culture, shared vision/goals (OR=41.49, p< 00), innovativeness (OR=5.96, p= 04) and school type (OR=6.66, p=,03) remained significant; however, climate became significant (OR=.04, p<.03) and collaboration was no longer significant (OR=1.51, p=.70). It should also be noted that climate had become negatively associated with the outcome in these multivariate analyses. Regarding provider-level variables, perceived mandate to use effective tobacco use prevention curricula (OR=2.32, p=.04) and school type (OR=5.73, p=.03) were the only variables that remained significant; whereas, provider expectations and values regarding tobacco use prevention programming (OR=4.76, p= 14) and provider self-efficacy to implement tobacco use prevention programming (OR=2.99, p=.36) were no longer significant when placed in a multivariate model. As for the curriculum-level variables, no variables were significant in the multivariate model. That is, provider regularly uses skills compatible with interactive instructional methods (i.e., discussion, activities, role playing) required to deliver the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 125 curriculum (OR=6.11, p— . 11) and school type (OR=4.03, p=. 13) were no longer significant in this model that combined all curriculum characteristics. Third-stage Models. The results of the third stage multivariate model are shown in Table 13. This stage of analysis placed all significant correlates from the second-stage model in the same simultaneous multivariable regression model, controlling for the only statistically significant demographic variable, school type. Adjusted odds ratios, 95% confidence intervals, Wald’s X2 estimates, p-values, and n sizes were calculated for each variable, along with a Wald’s X2 estimate and corresponding p-value for each overall model. To the extent that a variable’s coefficient in this model decreases from that of the second-stage model, the variable’s influence must be either indirect, through one or more other correlates in this model, or spurious. Table 13. Study One: Multivariate Logistic Regression Results of Factors Associated with the Adoption and Implementation of Evidence-based Tobacco Use Prevention Curricula in Schools OR C l’s Wald’s Xs P value All Significant (p<.10) Variables from Multivariate Analyses Principal leadership of tobacco use prevention education 1.37 .21, 8.97 .11 .74 Shared vision/goals 84.31 2.80, >99.99 6.51 .01** Innovativeness 5.10 48, 54.19 1.83 .18 Climate .05 .00, 1.44 3.07 .08 Provider perceived mandate to use effective tobacco use prevention curricula 2.64 .84, 8.31 2.74 .10 School type (high or junior/middle school) 7.54 -92, 61.68 3.55 .06 Overall model (n=S7) 11.88 .06 *p<-05, **p<01, ***p<0Ol Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 126 Of the 6 variables placed in this model, only one (shared vision/goals) remained statistically significant (OR=84.31, p< 01); whereas, three others (climate, provider perceived mandate to use effective tobacco use prevention curricula, and school type) approached statistical significance (all p values equal or less than . 10). Again, it should be noted that climate had become negatively associated with the outcome in these multivariate analyses Principal leadership of tobacco use prevention education (OR=l .37, p=.74) and innovativeness (OR=5.10, p=. 18) were no longer significant. Fourth-stage Models. To enable further examination of how the multiple levels of factors were related, hypothesized interactions were examined. Specifically, selected variables from different domains were entered into a multivariate logistic regression model along with the corresponding hypothesized interaction term, controlling for the only statistically significant demographic variable, school type. Adjusted odds ratios, 95% confidence intervals, Wald’s X2 estimates, p-values, and n sizes were calculated for each variable, along with a Wald’s X2 estimate and corresponding p-value for each overall model. The results are shown in Table 14. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 127 Table 14. Study One: Multivariate Logistic Regression Results of Hypothesized Interactions Associated with the Adoption and Implementation of Evidence-based Tobacco Use Prevention Curricula in Schools OR C l’s Wald’s X2 P value PrinciDa! leadership of tobacco use prevention education X provider self-efRcacv to teach tobacco use ureventiou education Principal leadership of tobacco use prevention education .10 <00, >99.99 .13 .72 Provider self-efficacy to teach tobacco use prevention education .09 <.00, >99.99 .10 .75 Interaction of principal leadership of tobacco use prevention education and provider self-efficacy to teach tobacco use prevention education 2.86 .09, 94.57 .35 .56 School type (high or junior/middle school) 4.24 .95, 18.90 3.58 .06 Overall model (n=6©) 10.91 .03 PrinciDal leadership of tobacco use Drevention education X comoatibiMtv with interactive instructional methods Principal leadership of tobacco use prevention education 1.37 .00, >99.99 .01 .92 Compatibility with interactive instructional methods .40 <.00, >99.99 .05 .82 Interaction of principal leadership of tobacco use prevention education and compatibility with interactive instructional methods 1.46 ■ 2 1 , 9.96 .15 .70 School type (high or junior/middle school) 4.08 .94, 17.70 3.52 .06 Overall model (n=59) 9.43 .05 Provider self-efficacv to teach tobacco use nrevention education X provider perception that curriculum has been demonstrated or endorsed as being effective Provider self-efficacy to teach tobacco use prevention education .13 <.00, >99.99 .08 .77 Provider perception that curriculum has been demonstrated or endorsed as being effective .03 <.00, >99.99 .18 .67 Interaction of provider self-efficacy to teach tobacco use prevention education and provider perception that curriculum has been demonstrated or endorsed as being effective 3.47 .05, >99.99 .32 .57 School type (high or junior/middle school) 6.38 1.26, 32.38 5.00 .03 Overall model (n=40) 8.64 .07 *p<05, **p<01, ***p<001 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 128 No significant associations (other than the main effect of school type) were found between adoption and implementation of evidence-base tobacco curricula and each of the three hypothesized interaction models: (a.) Schools that have an interactive combination of strong principal leadership of tobacco use prevention education and strong provider self-efficacy to teach tobacco use prevention education; (b.) Schools that have an interactive combination of strong principal leadership of tobacco use prevention education and providers who regularly use skills compatible with interactive instructional methods (i.e., discussion, activities, role playing) specified in the curriculum; and (c.) Schools that have an interactive combination of strong provider self-efficacy to teach tobacco use prevention education and strong provider perception that the curriculum has been demonstrated or endorsed as being effective. Adoption and Implementation o f Evidence-based Tobacco Use Prevention Programming. A second dependent variable was also created from teacher questionnaire data to represent the degree of adoption and implementation of evidence-based tobacco use prevention programming overall. The decision to examine this second dependent variable was made due to the small percentage of providers (i.e., 20% or 15 of 75 teachers) who reported adopting and implementing at least one of the evidence-based tobacco use prevention curricula that were specified on the questionnaire. This second dependent variable was more general in scope than the first because it represented a comprehensive index of the critical elements of effective tobacco use prevention programming, in and out of the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 129 classroom, including policy, curriculum delivery and teacher training. A continuous implementation index was created for each school based on the degree of implementation of evidence-based tobacco prevention programming as outlined by the CDC Guidelines for School Health Programs to Prevent Tobacco Use and Addiction (see Table 1). Thus, the linear regression procedure was utilized for this entire set of models. First-stage Models. The first set of models examined the association of the dependent variable (adoption and implementation of evidence-based tobacco use prevention programming) and each hypothesized implementation correlate (i.e., 11 organizational-level variables, 5 provider-level variables, 3 curriculum-level variables and 5 school demographic variables). These univariate models permitted elimination (in subsequent models) of those variables that did not have a significant relationship with the adoption and implementation of evidence-based tobacco use prevention programming. Table 15 presents the parameter estimates (standardized and raw), standard errors, p-values and n sizes for these relationships. The n in most of these models was 51; however, the models involving the curriculum-level variables ranged from 34-51. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 130 Table 15. Study One: Univariate Linear Regression Results of Factors Associated with the Adoption and Implementation of Evidence-based Tobacco Use Prevention Programming in Schools______ ____ __________ _________ ____________ Standardized Estimate Parameter St Error P value N Organizational-level Comnonents Principal Leadership of tobacco use prevention education .21 .85 .59 ,21 50 Commitment to tobacco use prevention education .06 .37 .83 .66 51 Knowledge of effective tobacco use prevention programs .26 1.92 1.02 .07 51 Structure & Processes Management (decentralized) .14 .67 .68 .32 51 Participatory decision-making -.18 -.86 .69 .21 51 Communication .01 .07 .64 .92 51 Adequate resource (training, materials, time) .43 1.57 .48 Q Q * * * 51 School Culture Shared vision/goals .17 .88 .74 .24 51 Innovativeness .24 1.08 .61 .08 51 Climate .08 .56 .96 .56 51 Collaboration .18 .91 .72 .21 51 Provider-level Comuonents Commitment to tobacco use prevention education .30 1.66 .75 .03* 51 Knowledge of effective tobacco use prevention programs .30 2.52 1.16 .03* 51 Perceived mandate to use effective tobacco use prevention curricula .32 .98 .42 .02* 49 Expectations and values regarding tobacco use prevention programming .52 3.02 .72 00*** 51 Self-efficacy to implement tobacco use prevention programming .21 1.66 1.11 .14 51 Curriculum-level Comuonents Provider perception that curriculum has been demonstrated or endorsed as being effective .13 .59 .79 .46 34 Utilized teaching methods familiar to provider, minimal training required -.21 -1.12 .85 .20 38 Compatibility with interactive instructional methods .29 1.65 .78 .04* 51 Demographic Variables Population density (rural, suburban or urban) -.03 -.19 .83 .82 51 School size (number of students enrolled) .06 .00 .00 .69 51 Socioeconomic status (percentage of students eligible to receive free or reduced cost lunch) -.04 -.59 2.03 .77 51 Ethnic minority (i.e., non-white) student composition -.11 -1.43 1.77 .42 51 School type (high or junior/middle school) -.08 -.57 .98 .56 51 *p<.05, **p<.01, ***p<.001 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 131 A total of 6 variables (1 organizational-level variable, 4 provider-level variables, and 1 curriculum-level variable) were significantly associated with the adoption and implementation of evidence-based tobacco use prevention programming. In addition, 2 other organizational-level variables approached levels of statistical significance (p< 10). Based on unadjusted analyses, adequate resources (training, materials and time) to teach tobacco use prevention education (standardized estimate = 0.43, p<00), provider commitment to teach tobacco use prevention education (standardized estimate = 0.30, p<05), provider knowledge of effective tobacco use prevention programs (standardized estimate = 0.30, p<05), provider perceived mandate to use effective tobacco use prevention curricula (standardized estimate = 0.32, p<05), provider expectations and values regarding tobacco use prevention programming (standardized estimate = 0.52, p< 00), and provider regularly uses skills compatible with interactive instructional methods (i.e., discussion, activities, role playing) specified in the curriculum (standardized estimate = 0.29, p<05) were significantly associated with the adoption and implementation of evidence-based tobacco use prevention programming. As shown, among these significant variables, the highest standardized estimate was found for provider expectations and values regarding tobacco use prevention programming (a provider-level variable). Specifically, a statistically significant positive association existed between teachers’ beliefs that tobacco use prevention programming was very effective in preventing or reducing tobacco use among students, and adoption and implementation of evidence-based tobacco use Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 132 prevention programming at the school. Adequate resources to teach tobacco use prevention education (an organizational-level structure and processes variable) was the second strongest correlate. None of the demographic factors were significantly associated with the outcome variable and, therefore, none were included in subsequent analyses. Second-stage Models. At the second stage, multivariate analyses were conducted to determine which variables were most important at each domain of factors (i.e., organizational, provider, etc.). That is, variables within each domain (e.g., among principal characteristics only) were simultaneously entered into a multivariate linear regression model. Parameter estimates (standardized and raw), standard errors, p-values and n sizes were calculated for each variable by domain, along with a p-value for each overall model. The results are shown in Table 16. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 133 Table 16. Study One: Multivariate Linear Regression Results ofFactors Associated with the Adoption and Implementation of Evidence-based Tobacco Use Prevention Programming in Schools_______________ __________ _________ __________ Standardized Estimate Parameter Estimate St. Error P value Organizational-level Comoonents Principal Leadership of tobacco use prevention education .12 .50 .61 .42 Commitment to tobacco use prevention education .01 .05 .82 .95 Knowledge of effective tobacco use prevention programs .26 1.88 1.07 .09 Overall model (n=50) .16 Structure & Processes Management (decentralized) .01 .06 .64 .92 Participatory decision-making -.23 -1.12 .67 .10 Communication .13 .58 .62 .35 Adequate resource (training, materials, time) .44 1.63 .50 .00** Overall model (n=51) .00** School Culture Shared vision/goals .06 .31 1.33 .82 Innovativeness .24 1.06 .83 .21 Climate -.24 -1.60 1.67 .34 Collaboration .18 .90 1.27 .48 Overall model (n=51) .42 Provider-level Components Commitment to tobacco use prevention education .18 .97 .72 .18 Knowledge of effective tobacco use prevention programs .09 .70 1.21 .57 Perceived mandate to use effective tobacco use prevention curricula .14 .43 .44 .34 Expectations and values regarding tobacco use prevention programming .32 1.88 .85 .03* Self-efficacy to implement tobacco use prevention programming .15 1.67 1.02 .26 Overall model (n=49) .00** Curriculum-level Components Provider perception that cuniculum has been demonstrated or endorsed as being effective -.04 -.19 .79 .81 Utilized teaching methods familiar to provider, minimal training required -.19 -.99 .87 .27 Compatibility with interactive instructional methods .42 2.46 1.04 .03* Overall model (n=32) .08 *p<.05, **p< 01, ***p< 001 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 134 The second stage of analysis produced results that differed from those of the first stage. Most notably, for example, are the differences found for the provider- level variables. Only provider expectations and values regarding tobacco use prevention programming remained statistically significant (standardized estimate = 0.32, p< 05) in the multivariate model that combined all provider characteristics; whereas, provider commitment to teach tobacco use prevention education (standardized estimate = 0.18, p=. 18), provider knowledge of effective tobacco use prevention programs (standardized estimate = 0.09, p=.57) and provider perceived mandate to use effective tobacco use prevention curricula (standardized estimate = 0.14, p=.34) were no longer significant. Third-stage Models. The results of the third stage multivariate model are shown in Table 17. This stage of analysis placed all significant correlates from the second-stage model in the same simultaneous multivariable regression model. Parameter estimates (standardized and raw), standard errors, p-values and n sizes were calculated for each variable by domain, along with a p-value for each overall model. To the extent that a variable’s coefficient in this model decreases from that of the second-stage model, the variable’s influence must be either indirect, through one or more other correlates in this model, or spurious. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 135 Table 17. Study One: Multivariate Linear Regression Results of Factors Associated with the Adoption and Implementation of Evidence-based Tobacco Use Prevention Programming in Schools_____________ __________ _________ ____________ Standardized Estimate Parameter Estimate St. Error P value All Significant (jkJO ) Variables from Multivariate Analyses Principal knowledge of effective tobacco use prevention programs .23 1.70 .82 .04* Participatory decision-making -.28 -1.35 .55 .02* Adequate resource (training, materials, time) .27 .98 .44 .03* Provider expectations and values regarding tobacco use prevention programming .43 2.50 .73 .00*** Compatibility with interactive instructional methods .05 .30 .70 .67 Overall model (n=51) .00*** — — ■ » » ■ ■ ■ ■ ■ Note: Overall model R -.46. *p<05, **p<.01, ***p<001 Of the 5 variables placed in this model, four remained statistically significant, including principal knowledge of effective tobacco use prevention programs (standardized estimate = 0.23, p<05), participatory decision-making (standardized estimate = -0.28, p<05), adequate resources to teach tobacco use prevention education (standardized estimate = 0.27, p< 05) and provider expectations and values regarding tobacco use prevention programming (standardized estimate = 0.43, p< 00); whereas, one variable (provider regularly uses skills compatible with interactive instructional methods specified in the curriculum) was no longer statistically significant. It should be noted that participatory decision-making was negatively associated with the outcome in these multivariate analyses, as well as in the univariate analyses. Fourth-stage Models. To enable further examination of how the multiple levels of factors were related, hypothesized interactions were examined. Specifically, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 136 selected variables from different domains were entered into a multivariate linear regression model along with the corresponding hypothesized interaction term. Parameter estimates (standardized and raw), standard errors, p-values and n sizes were calculated for each variable by domain, along with a p-value for each overall model. The results are shown in Table 18. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 137 Table 18. Study One: Multivariate Linear Regression Results of Hypothesized Interactions Associated with the Adoption and Implementation of Evidence-based Tobacco Use Prevention Programming in Schools_______ _____ Standardized Estimate Parameter Estimate St Error P value Principal leadership of tobacco use Drevention education X provider self-efficacv to teach tobacco use Drevention education Principal leadership of tobacco use prevention education .20 .82 5.30 .88 Provider self-efficacy to teach tobacco use prevention education .22 1.71 5.70 .77 Interaction of principal leadership of tobacco use prevention education and provider self-efficacy to teach tobacco use prevention education .02 .02 1.45 .99 Overall model (n=50) .20 PrinciDal leadership of tobacco use Drevention education X compatibility with interactive instructional methods Principal leadership of tobacco use prevention education .47 1.95 3.19 .54 Compatibility with interactive instructional methods .55 3.03 4.11 .47 Interaction of principal leadership of tobacco use prevention education and compatibility with interactive instructional methods -.44 -.42 1.07 .70 Overall model (n=50) .14 Provider self-efficacv to teach tobacco use Drevention education X provider perception that curriculum has been demonstrated or endorsed as being effective Provider self-efficacy to teach tobacco use prevention education .97 7.21 8.35 .39 Provider perception that curriculum has been demonstrated or endorsed as being effective 1.46 6.55 9.46 .49 Interaction of provider self-efficacy to teach tobacco use prevention education and provider perception that curriculum has been demonstrated or endorsed as being effective -1.22 -1.47 2.50 .56 Overall model (n=34) .27 *p<05, **p<01, ***p<0Ol No significant associations were found between adoption and implementation of evidence-based tobacco use prevention programming and each of the three hypothesized interaction models; (a.) Schools that have an interactive Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 138 combination of strong principal leadership of tobacco use prevention education and strong provider self-efficacy to teach tobacco use prevention education; (b.) Schools that have an interactive combination of strong principal leadership of tobacco use prevention education and providers who regularly use skills compatible with interactive instructional methods (i.e., discussion, activities, role playing) specified in the curriculum; and (c.) Schools that have an interactive combination of strong provider self-efficacy to teach tobacco use prevention education and strong provider perception that the curriculum has been demonstrated or endorsed as being effective. Study Two: Longitudinal Study Attrition Analyses Table 19 presents the differences in attrition (at Wave 2) on all school demographic variables for the analytic sample of schools (n=7Q) that provided complete data for all 3 waves of the Independent Evaluation of the California Tobacco Control Program compared to all other schools that did not provide data for all 3 waves (n=73). This type of analysis helps to determine how representative the analytic sample is of the population of schools which participated in the study. As shown, schools that were used in the analyses had higher rates (although not statistically significant) for population density (e.g., rural, suburban, urban) of the area served by the school (M=1.26 vs. M-1.08 and SD=.67 vs. SD= .66, respectively), school enrollment size (M-398.87 vs. M=378.71 and SD=225.41 vs. SD= 245.39, respectively) and student ethnic minority composition (54.47% vs. 47.49%). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 139 Table 19. Study Two: Comparison of Characteristics of the Schools Used in Analyses vs. Schools Not Used in Analyses_____________________ Schools Used in Analyses (N=70) Schools Not Used in Analyses (N=73) Variable Population density (0=rural, 1-suburban, 2: =urban) Mean (SD) 1.26 (.67) 1.08 (.66) School size (number of students enrolled) Mean (SD) 389.87 (225.41) 378.71 (245.39) Socioeconomic status (percentage of students eligible to receive free or reduced cost lunch) % 41.91 33.78* Ethnic minority (percentage of non white student composition) % 54.47 47.49 School type (CMiigh, 1 -junior/middle) Mean (SD) 62.86 (48.67) 27.40 (44.91)*** *p<.05, **p< 01, ***p<001 Statistically significant differences were found, however, for the percentage of students eligible to receive free or reduced cost lunch (used as a proxy measure for socio-economic status) and school type (high or junior/middle). Specifically, schools used in the analyses had significantly higher rates of students who were eligible to receive free or reduced cost lunch (41.91% vs. 33.78%, p<05), indicating that these schools had higher rates of poverty compared to schools that were excluded from the analyses. Further, schools included in the analyses had a significantly higher mean for school type (M=62.86 vs. M=27.40 and SD=48.67 vs. SD= 44.91, respectively; Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 140 p< 001), indicating that this group of schools was comprised of a higher proportion of junior/middle schools relative to high schools compared to the group of schools not used in the analyses (which had a higher proportion of high schools relative to junior/middle schools). Sample Characteristics School data. The M l analytic sample consisted of 70 junior/ middle and senior high schools that provided both classroom teacher and student data for all 3 waves. Table 20 shows the demographic characteristics of the M l analytic sample of schools. Almost half of the schools (48.57%) were located in suburban areas, followed by urban (38.57%) and rural (12.86%) areas. The mean number of students enrolled in these schools was 389.87. More than 40% of the students in these schools were eligible to receive free or reduced cost lunches. Approximately 55% of the students at these schools were of minority (i.e., non-white) ethnicity, including Black, Hispanic, Asian, and American Indian backgrounds. Over 60% (n=44) of the schools were junior/middle schools and nearly 40% (n=26) were high schools. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 141 Table 20. Study Two: Demographic Characteristics of the Analytic School Sample fn=70 Schools)________________________________________ _______________ Variable N (%) Mean (SD) Population Density Rural 9(12.86) Suburban 34 (48.57) Urban 27 (38.57) School size (number of students enrolled) 389.87 (225.41) Socioeconomic status (percentage of students eligible to receive free or reduced cost lunch) (41.91) Minority (i.e., non-white) student composition (54.47) School Type High school 26 (37.14) Junior/middle school 44 (62.86) Teacher Data. The demographic characteristics of the teacher sample (n=97 teachers at 70 schools) are provided in Table 21. More than two-thirds (67.37%) of the participants were female. The majority (76.40%) of the teachers had 3 or more years of experience in teaching at their school, with a sample mean of 8 .0 years (SD= 6.92). In terms of hours of teaching tobacco use prevention education in the last year, over two-thirds (68.89%) of the respondents taught 0-5 hours, although a mean of 6.20 hours (SD= 4.97) was reported for the entire sample. Approximately 55% of these teachers were certified by the State of California to teach health education. Regarding smoking history, approximately one-third of the sample (33%) reported Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 142 that they had smoked at least 100 cigarettes in their lifetimes. Current smoking, however, was reported for only a very small percentage (1.06%; n=l) of the teachers, with the rest of the teachers (98.94%; n=93) indicating that they do not smoke at all. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 143 Table 21. Study Two: Demographic Characteristics of the Analytic Teacher Sample f11=97 Teachers at 70 Schools) Variable N (%) Mean (SD) Gender Male 31 (32.63) Female 64 (67.37) Years of teaching at this school 8.00 ( 6.92) 0-2 21 (23.60) 3-5 26 (29.21) 6-8 11 (12.36) 9-19 23 (25.84) 20+ 8 ( 8.99) Hours taught tobacco use prevention last year 6.20 ( 4.97) 0-2 37(41.11) 3-5 25 (27.78) 6-8 14(15.56) 9-19 9 (10.00) 20+ 5 ( 5.56) Certified by State of California to teach health education Yes 51 (55.43) No 41 (44.57) Smoked 100 cigarettes in lifetime Yes 31 (33.00) No 63 (67.00) Now smoke cigarettes Jivery day 1 ( 1.06) Some days 0 ( 0.00) Not at all 93 (98.94) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 144 Student Data. Table 22 presents demographic information for the student sample (n=6921students at 70 schools). Gender was evenly distributed across the sample. A majority (57.12%) of the students were enrolled in 8th grade, whereas 42.88% were enrolled in 10th grade. Mean age for the sample was 14.50 years (SD=1.16). More than half (55.37%) of the students were of minority (i.e., non- white) ethnicity, including Black, Hispanic, Asian, and American Indian backgrounds. Over two-thirds (69.42%) of the students reported that they spoke only or mostly English at home, whereas 24.00% spoke half English and half another language, and 6.58% spoke only or mostly another language. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 145 Table 22. Study Two: Demographic Characteristics of the Student Sample (n=6921 Students at 70 Schools)_______________________________________________ Variable N (%) Mean (SD) Gender Male 3420 (49.62) Female 3472 (50.38) Grade 8th 3953 (57.12) 10th 2968 (42.88) Age 14.50 ( 1.16) 13 or under 1667 ( 24.21) 14 2091 (30.37) 15 1470(21.35) 16 1513 (21.97) 17 or above 145 ( 2.11) Ethnicity White 3089 (44.63) Non-white 3832 (55.37) Language spoken at home Only English 3801 (56.07) Mostly English 905 (13.35) Vi English & V z other language 1627 (24.00) Mostly another language 311 ( 4.59) Only another language 135 ( 1.99) Descriptive Analyses of Variables Descriptive Analyses of Dependent Variables The means and/or percentages for the seven tobacco used-related program outcomes measured at two time points, 1996 and 2000, are summarized in Table 23. Among the 8th and 10th grade students surveyed, all of the changes in outcome Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. variables from 1996 to 2000 were statistically significant. Overall, students reported significant improvements in students’ tobacco-related beliefs, attitudes, skills, tobacco use prevalence and tobacco cessation. Specifically, over the four-year program period, students reported significant increases in their agreement with statements about the negative consequences of tobacco use (from a mean of 3.40 to 3.47 on a 1-4 point scale, p<01) and negative attitudes towards the tobacco industry (from a mean of 3.46 to 3.49 on a 1-4 point scale, p< 01); whereas, the perceived peer norms of tobacco use among peers (i.e., students’ estimates of the proportion of peers who smoke monthly) decreased significantly (from 47.24% to 41.63% on a 0- 100% scale, p<01). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 147 Table 23 . Study Two: Means fSD) and Percentages of Adolescent Tobacco Use- related Program Outcome Variables. 1996 and 2000 (N-70 Schools i _________ 1996 (Wave 1) 2000 (Wave 3) Mediating Variables Negative outcome expectancies of smokinga Mean (SD) 3.40 (.13) 3.47 (.11)** Negative attitudes towards the tobacco industrya Mean (SD) 3.46 (.17) 3.49 (.13)** Perceived peer norms of smoking (mean % peers who smoke) 0-100% 47.24 41.63** Tobacco refusal self-efficacy a Mean (SD) 3.44 (.22) 3.54 (.17)** Behavioral Variables Lifetime (ever) smoking (%) 0-100% 54.43 45.99** Thirty-day smoking (%) 0-100% 22.47 16.68** Smoking quit attempts in past year (%) 0-100% 58.04 61. 13** *p<05, **p<.01 a 4-point scale (range of 1-4) At the same time, significant increases were observed for tobacco refusal self-efficacy skills, with students reporting that they believed it would be easy to refuse an offer of tobacco from a friend (from a mean of 3.44 to 3.54 on a 1-4 point scale, p< 01). Further, students reported significant decreases in the prevalence of lifetime cigarette smoking (from 54.43% to 45.99%, p<01) and 30-day cigarette smoking (from 22.47% to 16.68%, p<01). In addition, there was a significant Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 148 increase in the proportion of student smokers who had tried to quit in the past year (from 58.04%to 61.13%,p<01). Descriptive Analyses of Independent Variables Implementation o f Evidence-based Tobacco Use Prevention Programming. Table 24 presents information regarding the extent to which 5 of the 7 CDC Guidelines for School Health Programs to Prevent Tobacco Use and Addiction were applied in this sample of California schools during the 1997-98 school year (Wave 2). Regarding tobacco use policy enforcement, teachers reported relatively high mean levels (M=3.69 on a scale of 1-4), indicating that they perceived that their school’s no-tobacco use policy was being enforced a great deal. As for tobacco instruction, over three-quarters (84.50%) of teachers reported teaching tobacco use prevention lesson to their students, and these same teachers also indicated that they were fairly comprehensive in terms of teaching the recommended instructional content (M=3.70 on a scale of 0-5; a sum of physiological effects, social influences, social consequences, peer norms, and refusal skills; scored 0=no or l=yes). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 149 Table 24. Study Two: Means (SD) and Percentages for Variables Used in Implementation of CPC Guidelines Index (n-70 Schools)______ ______ Variable Range Mean (SD) or % Policy enforcement 1-4 3.69 ( .61) Tobacco Instruction Implemented lessons 0-100% 84.50 Taught recommended content 0-5 3.70(1.17) Program-specific training 0-100% 22.86 Parent involvement Amount 1-4 1.73 ( .76) Types 0-6 .74 ( .90) On-site cessation program 0-100% 42.86 Implementation of CDC Guidelines (standardized index; Cronbach’s alpha = .64) -7.38 to 11.96 .00 (3.58) The prevalence of teacher training was particularly low, with only 22.86% of the teachers reporting that they had received program-specific tobacco prevention in- service training. These teachers also reported relatively low mean efforts (M=1.73 on a scale of 1-4) to get students’ parents involved in tobacco use prevention education. Regarding the types of parent involvement they obtained, teachers reported very low mean scores (M=0.74 on a scale of 0-6), indicating that they rarely involved parents in homework assignments, held meetings with parents of student smokers, distributed newsletters or educational materials to parents, or provided information on smoking cessation to parents. Further, in the matter of cessation classes, 42.86% Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 150 of teachers reported that their school had on-site classes, groups, or programs for students who wanted to quit using tobacco. The final index had a Cronbach’s alpha of .64 and scores that ranged from -7.38 to 11.96 (with a mean of 0.00). School Capacity to Implement Innovative Programming. Table 25 presents a summary of the means, standard deviations, and ranges of the variables that comprised the school capacity to implement innovative programming index which assessed the extent to which schools possessed the potential to implement innovative programming during the 1997-98 school year (Wave 2). For the most part, across the total sample of 70 schools, teachers reported relatively high levels of the various organizational- and provider-type variables that were hypothesized to be associated with improvements in students’ tobacco-related beliefs, attitudes and skills, tobacco use prevalence, and tobacco cessation. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 151 Table 25. Study Two: Means and Standard Deviations for Variables used in School Variable Range Mean (SD) School leadership of tobacco use prevention education 0-1 .88 ( .25) School commitment to teach tobacco use prevention education 1-5 3.52 ( .70) School support for teaching tobacco use prevention 1-4 3.03 ( .70) School climate 1-5 3.50 ( .65) School collaboration 1-4 2.82 ( .90) Provider expectations and values regarding tobacco use prevention programming 1-4 3.41 ( .57) Provider self-efficacy to implement tobacco use prevention programming 1-4 3.38 ( .50) Capacity to implement innovative programming (standardized index; Cronbach’s alpha = .62) -8.79 to 6.56 0.00 (3.70) Regarding school leadership, participants reported relatively high levels of leadership of tobacco use prevention education (M = 0.88 on a scale of 0-1). Similarly, other organizational-type variables were also reported as existing at relatively high levels for the sample, with means well above the midpoint on school commitment to teach tobacco use prevention education (M = 3.52 on a scale of 1-5) and school support for teaching tobacco use prevention education (M = 3.03 on a scale of 1-4). School climate (M = 3.50 on a scale of 1-5) and school collaboration Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 152 (M = 2.82 on a scale of 1-4) were also reported to exist above the mean midpoint. Regarding provider-type variables, on a scale from 1-4, participants were well above the midpoint on mean levels of provider expectations and values regarding tobacco use prevention education (M = 3.41) and provider self-efficacy to implement tobacco use prevention education (M = 3.38). The final index had a Cronbach’s alpha of .62 and scores that ranged from -8.79 to 6.56 (with a mean of 0.00). Correlations among All Analytic Variables Table 26 shows correlations among the analytic variables, including all dependent, independent and demographic variables. Inspection of these correlations revealed statistically significant associations between the dependent and independent variables. Regarding correlations among the independent variables, many statistically significant associations existed among variables of the same major constructs (i.e., implementation of evidence-based tobacco use prevention programming and school capacity to implement innovative programming) as well as many significant associations among variables across the major constructs. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. o o - S o zn o r - - c n 1 c 3 > o »1—i » '■ " " I o c f l a .2 J 5 , 1 3 b 8 < 5 " S i — i b £ H •5 3 0 0 S O C M a j H < n * * I J ^ C N l * ( S C< # ■ C S O *-H « 4 * ^ ^ ♦ ^ S a * ^ <N OS <N . c o _*. co vo *L ^ - © : < n * I I Q Q h * * & £* 2 2 o f~ JJ c o c o • > * 1 c s o C 'l f-H ■ i f - $ ■ * * * * O n # t « t O n ( N l -t < N | . * s o - - t- ■ ' * « . * ~ ~ ■ » J, * i M S M • n ~ t r ■ i f b * N n * S s° © © * . (N ,- ,• m ■ ’ ' IT, C l o o o o r ~ P o < - H t ~ ~ V O O O n V O so t" - ,—I pH i-H « — H O ^ * O o i co - i f - i f * * * * « • o OS *-H CO co «n ■ i f o * * * <•o $ - i f * * t" co o * a jo * cs co O * c o V O C S C N C S r - H © < N - i f « - i f : x h co »n ^ 2 o T - 1 o » — H t —I Ov H N ^ H p r — on P o o ■ * £-- r-H 0*0 r ^ 1 ^ O s 0 0 > O O f-H -if * •if * CN VO o s CO # $- # 00 a n o p-H r - Os n <N * n n -0“ r-H O 1 v-H r - O p 9 — ^ c s a * o p-H p O p-H O fH f-H > C i CN CO CO l ’ a * a * CO a * a" a * * -if Os 3f « n €N a’ CS ■ if -if CN VO -if -if a n VO 00 <N 0 0 •if VO V0 CN CO a n ■ if O s CO p o s * CO CO CN « * a * VO CO a * o o p a " O a’ o a’ Os e s O a * O O a * CN O a * p •if * # * CO r - CO *Ti Os •if •if CO 0 0 o CO 0 0 VO o Os l > •if •if * -if o s CO CO CO CO s’ p M O t“ H a * r - •* 3 “ O a * o s “H o 5 “H O rH r*H O a n CO VO CO a * • i f • i f •T) CO • i f • i f • i f O * * * * if t * * « ! < '! Jl 3 ^ * s * t * t* in *n * • S £; * * O o ’ - < ^ » * C O OvNfjs! p « " » *? 7 -1 > 1 ' t* h- ^ ^ n sn O O O « -H v o O s CO o if « n eN O O o p * -3 - ^rs * * ” « - h r * * * o ! rf * 0 ■ i f , ° CO * CO O «N -if r-H i-H C ^ CO * CO 0 0 * t - H if '• § ’ O * M ’^S0n\0Swrf.(f) '0 fl0 w X Q O ^ l<lQs^ P S P r H . P S P i - S o o P P o P o P o P S S c O l ■ i i • i • • • f l u * a ' i s • • - i f > * SO —' > 55 ? ? « S ■ > ' S O * # * "t # * < — » ^ T f • < * t C O } ’ '^tcsvoSDS92°°2',:S ‘ O iP ^ J r r r * r • • r t* co Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 26. S tudy Two: Intercorrelations o f A ll Analytic Variables (n= 70 Schools). Continued. 154 > C j § § Si o n * * T-S. I T f r 4 5 - O <S t-H »-H * O • * ■ < N rj S n O N H (S O fS co * * * i O N J ^ O ° - * 0 g «n oo ^ — o ■ H * ■ H N O * * © 1 * ^ - N O co t o o \0 * O CO « - « n < n u 3 * h : * o o ^ S o i * r b * b * • r # * _ ^ S O * r h d N m * n • * r- m H )0 f S J C N O °. i-4 o * r- O M 00 O « * co ' . Sx «n S 8 < N * CO 'O m n 5 oo • r- »^eN Sco><o'o^«NiPo>V5^©^oe^^Po>?fi^©^®si^^| fn>^ Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. *p<.05, **p<.01, ***p<.001 Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission. Table 26. Study Two: Intercorrelations of All Analytic Variables (n=70 Schools). Continued. Dependent variables (Adolescent Tobacco Use-related Program Outcomes'! At Negative outcome expectancies of smoking A2 Negative attitudes toward the tobacco industry A3 Perceived peer norms (prevalence) of smoking A4 Tobacco refusal self-efficacy A5 Lifetime smoking A6 Thirty-day smoking A7 Smoking quit attempts in the past year Predictor variables (CPC Implementation Index items) B1 Policy enforcement B2 Tobacco instruction (implemented lessons) B3 Tobacco instruction (taught recommended content) B4 Program-specific training B5 Parent involvement (amount and types) B6 On-site cessation program B7 CDC Implementation Index (Standardized) Predictor variables (Capacity to Implement Innovative Programming Index items! Cl School leadership C2 School commitment to teach tobacco prevention education C3 School support for teaching tobacco prevention C4 School climate C5 School collaboration C6 Teacher expectations and values of tobacco prevention programming C7 Teacher self-efficacy to implement tobacco prevention programming C8 Capacity to Implement Innovative Programming Index (Standardized) Demographic variables D 1 Population density (rural, suburban or urban) D2 School size (number of students enrolled) D3 Socioeconomic status (percentage of students eligible to receive free or reduced cost lunch) D4 Ethnic minority (i.e., non-white) student composition D5 School type (high or junior/middle or high school) 156 Dependent Variables. Many of the individual variables that comprised either of the independent variables (i.e., implementation of evidence-based tobacco use prevention programming or school capacity to implement innovative programming indices) were significantly correlated with the dependent variables. Specifically, for implementation of evidence-based tobacco use prevention programming, perceived policy enforcement was positively related to students’ negative outcome expectancies of smoking (r=.30, p< 05). Policy enforcement was also negatively related to students’ perceived peer norms of smoking (r=-.38, p< 01), lifetime smoking (r=-.25, p<05) and 30-day smoking (r=-.37 p<.01). Regarding tobacco instruction, implementation of tobacco use prevention lessons in the last year was associated with a higher proportion of student smokers who had tried to quit in the past year (r=.24, p< 05). Unexpectedly, however, implementation of the recommended instructional content (physiological effects, social influences, social consequences, peer norms and refusal skills) was associated with higher student rates of lifetime smoking (r=,33, p< 01). Another example of a significant correlation was between teachers’ efforts to get parent involvement in tobacco use prevention and students’ negative outcome expectancies of smoking (r=.33, p< 01). School capacity to implement innovative programming variables were also significantly correlated with some of the dependent variables. For instance, provider expectations and values regarding tobacco use prevention education was positively associated with both perceived peer norms of smoking (r=.25, p<05) and lifetime Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 157 smoking (r=.29, p< 05). Both of these correlations were in the unexpected direction, indicating that teachers’ perceptions that tobacco use prevention education is a valuable use of student time were positively associated with higher student estimates of the proportion of peers who smoke, and higher rates of lifetime smoking. School demographic variables were consistently associated with several of the dependent variables. School enrollment size was significantly and positively related to both perceived peer norms of smoking (r=.27, p< 05) and tobacco refusal self-efficacy (r=.35, p< 01). Socioeconomic status was inversely correlated with students’ negative attitudes about the tobacco industry (r=-.47, p<001) and tobacco refusal self-efficacy (r=-.36, p<.01). The proportion of ethnic minority students was positively correlated with negative attitudes about the tobacco industry (r=.34, p< 01). Further, school type was significantly and positively correlated with 5 of the 7 dependent variables. More specifically, compared to high schools, junior/middle schools had higher student agreement with statements about the negative consequences of tobacco use (r=.34, p<01), estimates of the proportion of peers who smoke (r=.62, p< 001), refusal self-efficacy (r=.67, p< 001), and rates of lifetime (r=.54, p<001) and 30-day smoking (r=.61, p<001). Implementation o f Evidence-based Tobacco Use Prevention Programming. Variables that comprised the implementation of evidence-based tobacco use prevention programming index were found not to be consistently correlated with one another. In particular, only two significant correlations existed; that is, the positive correlations between having taught recommended instructional content and parent Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 158 involvement (r^.30, p<.01) and having taught recommended instructional content and existence of an on-site cessation program (r=.39, p< 001). As expected, all six of the variables that comprised the index were significantly and positively correlated (with correlation coefficients ranging from .38 to .63) to the final standardized index. Several of the variables that comprised the implementation of evidence-based tobacco use prevention programming index correlated with variables across constructs. For example, having taught recommended instructional content was significantly and positively correlated with school capacity to implement innovative programming variables, including school leadership of tobacco use prevention education (r=.30, p<05), school commitment to teach tobacco use prevention education (r=.37, p<001), school collaboration (r=.24, p<05) and provider expectations and values regarding tobacco use prevention programming (r=.48, p< 001). Parent involvement in tobacco use prevention education was significantly and positively correlated with school leadership of tobacco use prevention education (r=.24, p<05), school commitment to teach tobacco use prevention education (r=.44, pc.001), school support for teaching tobacco use prevention education (r=.24, p< 05), provider expectations and values regarding tobacco use prevention programming (r=.32, p<01) and provider self-efficacy to implement tobacco use prevention programming (r=.26, p<05). As another example of significant correlations across the various constructs, existence of an on-site cessation programming (for students who want to quit smoking) was positively correlated with Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 159 two demographic variables, school enrollment size (r=.28, p< 05) and school type (high or junior/middle; r=.53, p< 001). School Capacity to Implement Innovative Programming. The school capacity to implement innovative programming variables correlated well with one another. For example, school support for teaching tobacco use prevention education was significantly correlated with school climate (r=.31, p<.01), school collaboration (r=,51, p<001), provider expectations and values regarding tobacco use prevention programming (r=.24, p<.05) and provider self-efficacy to implement tobacco use prevention programming (r=.26, p<05). As another example, school commitment to teach tobacco use prevention education was significantly correlated with school support for teaching tobacco use prevention education (r=.31, p<05) and provider expectations and values regarding tobacco use prevention programming (r=.38, p<01). As expected, all seven of the variables that comprised the index were significantly and positively correlated (with correlation coefficients ranging from .39 to .75) to the final standardized index. As noted above, some of the school capacity to implement innovative programming variables correlated with the implementation of evidence-based tobacco use prevention programming variables, as well as with demographic variables. For example, school collaboration was significantly and positively correlated with population density (rural, suburban or urban; r= .31, p< 01) and school enrollment size (r= .29, p<05). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 160 Demographic Variables. The demographic variables were found to be consistently correlated with one another. School enrollment size was significantly and positively correlated with population density (rural, suburban or urban; r= .42, p<.001) and school type (high or junior/middle; r=.48, p< 001), and significantly and inversely related to the proportion of ethnic minority students (r= -.25, p<.05). The proportion of ethnic minorities was also inversely correlated with population density (r= -.61, p< 001) and socioeconomic status (r= -.46, p<001). Whereas, socioeconomic status was also inversely correlated with school type (r= -.33, p< 01). Also, as noted above, a few of the demographic variables were correlated with variables from other major constructs. Hypothesis Testing A two-stage general linear model (GLM) analysis protocol was completed on longitudinal data. The first-stage models assessed the effects of each independent variable (implementation of evidence-based tobacco use prevention programming or school capacity to implement innovative programming indices) on each of the seven student program outcome variables. The second-stage models assessed the hypothesized moderating effect of school capacity to implement innovative programming on the relationship between implementation of evidence-based tobacco use prevention programming and each of the seven student program outcome variables. All of these models also included the covariates of the corresponding baseline (Wave 1) outcome variable, population density, student enrollment size, socioeconomic status, percentage non-white students and school type. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 161 First-stase Models Implementation o f Evidence-based Tobacco Use Prevention Programming. The effects of implementation of evidence-based tobacco use prevention programming on 2000 adolescent tobacco use-related program mediators and behavioral outcomes are shown in Table 27. F-tests and p-values indicated that three of the seven relationships were statistically significant; however, one other variable approached levels of statistical significance (p< 10). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 162 Table 21 . Study Two: Effects of Implementation of Evidence-based Tobacco Use Prevention Programming Index on 2000 Adolescent Tobacco Use-related Program Mediators and Behavioral Outcomes F-test P value Mediating Variables Negative outcome expectancies of smokinga 14.61 00*** Negative attitudes towards tobacco industry * .66 .42 Perceived peer norms of smoking (mean % peers who smoke) 3.57 .06+ Tobacco refusal self-efficacya .01 .91 Behavioral Variables Lifetime (ever) smoking (%) 6.09 .02* Thirty-day smoking (%) 8.82 .00** Smoking quit attempts in past year (%) .20 .65 Note. All models adjusted for 1996 program outcomes, population density, school enrollment size, school socioeconomic status, school ethnicity, and school type. +p<.10, *p<.05, **p<.01, ***p<.001 a 4-point scale (range of 1-4) The strongest effect of implementation of evidence-based tobacco use prevention programming was on negative outcome expectancies of smoking (F=14.61, p<001). This result indicated that a higher degree of implementation of evidence-based tobacco use prevention programming (based on 5 of the 7 CDC Guidelines for School Health Programs to Prevent Tobacco Use and Addiction; see Table 1) was predictive of significant increases (over the four-year program period) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 163 in students’ agreement with statements about the negative consequences of tobacco use (from a mean of 3.40 to 3.47 on a 1-4 point scale). Likewise, a higher degree of implementation of evidence-based tobacco use prevention programming was predictive of significant decreases (over the four-year program period) in the prevalence of lifetime cigarette smoking (from 54.43% to 45.99%; F=6.09, p=.02) and 30-day cigarette smoking (from 22.47% to 16.68%; F=8.82, p<.01). A marginally significant relationship existed between implementation of evidence- based tobacco use prevention programming and perceived peer norms of smoking, suggesting that a higher degree of implementation of evidence-based tobacco use prevention programming was predictive of significant decreases (over the four-year program period) in students’ estimates of the proportion of peers who smoke monthly (from 47.24% to 41.63% on a 0-100% scale; F=3.57, p=.06). Implementation of evidence-based tobacco use prevention programming had the weakest predictive relationship with tobacco refusal self-efficacy (F=.01, p=.91), followed by smoking quit attempts in the last year (F=.20, p=.65) and then negative attitudes towards the tobacco industry (F=.66, p=42). That is, a higher degree of implementation of evidence-based tobacco use prevention programming as prescribed by the CDC was not predictive of subsequent increases in students’ beliefs that they could refuse a best friends’ cigarette offer, students’ agreement with negative statements about the tobacco industry, and the proportion of student smokers who had tried to quit smoking in the past year. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 164 School Capacity to Implement Innovative Programming. Table 28 presents the effects of school capacity to implement innovative programming on 2000 adolescent tobacco use-related program mediators and behavioral outcomes. F-tests and p-values indicated that only one of the seven relationships was statistically significant. That is, school capacity to implement innovative programming was significantly predictive of students’ tobacco refusal self-efficacy (F=5.21, p=.03). This result indicated that a higher degree of a school’s capacity to implement innovative programming was predictive of significant increases (over the four-year program period) in students’ beliefs that they could refuse a best friends’ cigarette offer (from a mean of 3.44 to 3.54 on a 1-4 point scale). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 165 Table 28. Study Two: Effects of School Capacity to implement Innovative Programming Index on 2000 Adolescent Tobacco Use-related Program Mediators and Behavioral Outcomes F-test P value Mediating Variables Negative outcome expectancies of smokinga .03 .85 Negative attitudes towards tobacco industry8 .05 .82 Perceived peer norms of smoking (mean % peers who smoke) 1.54 .22 Tobacco refusal self-efficacy a 5.21 .03* Behavioral Variables Lifetime (ever) smoking (%) 1.35 .25 Thirty-day smoking (%) 2.37 .13 Smoking quit attempts in past year (%) .06 .81 Note. All models adjusted for 1996 program outcomes, population density, school enrollment size, school socioeconomic status, school ethnicity, and school type. +p< 10, *p<05, **p<01, ***p<.001 a 4-point scale (range of 1-4) Second-stage Models The results of the second-stage multivariate models are shown in Table 29. This stage of analysis assessed the effects of the interaction models (implementation of evidence-based tobacco use prevention programming index X school capacity to implement innovative programming index) on 2000 adolescent tobacco use-related Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 166 program mediators and behavioral outcomes. Beta estimates, standard errors and p- values are presented for the implementation of evidence-based tobacco use prevention programming index, the school capacity to implement innovative programming index, and the interaction term for each of the seven student program outcome variables. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 167 Table 29. Study Two: Effects of Interaction Models (Implementation of Evidence- based Tobacco Use Prevention Programming Index X School Capacity to Implement Innovative Programming Index) on 2000 Adolescent Tobacco Use-related Program Mediators and Behavioral Outcomes__________ Beta Estimate St. Error P value Mediating Variables Negative outcome expectancies of smoking * Implementation index .014 .005 01** Capacity to implement index -.004 .004 .20 Interaction .001 .001 .39 Negative attitudes towards tobacco industry Implementation index .011 .006 .08 Capacity to implement index -.003 .005 .55 Interaction .000 .001 .88 Perceived peer norms of smoking (mean % peers who smoke) Implementation index -.132 .295 .66 Capacity to implement index -.020 .236 .93 Interaction -.120 .061 .05* Tobacco refusal self-efficacy Implementation index .003 .007 .68 Capacity to implement index .000 .005 .98 Interaction .001 .001 .68 Behavioral Variables Lifetime (ever) smoking (%) Implementation index -.004 .005 .43 Capacity to implement index -.001 .004 .75 Interaction .000 .001 .63 Thirty-day smoking (%) Implementation index -.006 .003 .04* Capacity to implement index .003 .002 .21 Interaction -.000 .001 .55 Smoking quit attempts in past year (%) Implementation index .003 .015 .87 Capacity to implement index -.004 .011 .76 Interaction -.002 .004 .62 Note. All models adjusted for 1996 program outcomes, population density, school enrollment size, school socioeconomic status, school ethnicity, and school type. *p<05, **p<.01, ***p<.0Ol a 4-point scale (range of 1-4) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 168 As shown, only one significant interaction effect was observed. School capacity to implement innovative programming had a significant moderating effect on the relationship between implementation of evidence-based tobacco use prevention programming and perceived peer norms of smoking (beta estimate—. 120, p< 05); although, no significant main effects existed for either the predictor (implementation of evidence-based tobacco use prevention programming) or the moderator (school capacity to implement innovative programming). More specifically, results indicated that perceived peer norms of smoking was significantly reduced in schools with both a high degree of implementation of evidence-based tobacco use prevention programming and a high degree of school capacity to implement innovative programming. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 169 DISCUSSION Overview This dissertation study contributes further evidence to the growing body of diffusion research indicating a relationship between theory-based organizational-, provider-, and curriculum-level constructs and the adoption and implementation of evidence-based tobacco use prevention programming in schools. In particular, this investigation contributes new knowledge to this area of investigation because it investigated the interrelationships among multiple predictors of implementation by testing their interactive effects on subsequent adolescent tobacco-use related program mediators and behavioral outcomes. Specifically, Study 1 utilized a cross-sectional sample to examine the adoption and implementation phases of the diffusion process, by identifying the potential importance of organizational, provider, and curriculum influences on the implementation of evidence-based tobacco use prevention curricula, as well as evidence-based tobacco use prevention programming. Study 2 utilized a longitudinal sample to examine the adoption and implementation phases, along with the program outcome phase, by focusing on the importance of the interactive effect between a newly-developed school capacity to implement innovative programming construct (derived from organizational-, provider- and curriculum-level variables) and the degree of implementation of evidence-based tobacco use prevention programming, with subsequent changes in adolescent tobacco use-related beliefs, attitudes, skills and behaviors. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 170 Study One: Cross-sectional Study Results of the cross-sectional study found support for the primary hypothesis that variables from the organizational, provider, curriculum and demographic domains would be associated with the adoption and implementation of evidence- based tobacco use prevention curricula and programming. However, it was found that adoption and implementation of evidence-based tobacco use prevention curricula and programming were not positively associated with each individual variable listed in Table 2, and with each level of factors (i.e., organizational, provider, curriculum, and demographic) as hypothesized. Further, the patterns of associations were markedly different for the two outcome variables, suggesting that the factors affecting adoption and implementation of an evidence-based tobacco use prevention curriculum differ from those affecting adoption and implementation of an evidence-based tobacco use prevention programming practices school-wide. Adoption and Implementation o f Evidence-based Tobacco Use Prevention Curricula. As posited, variables from the organizational, provider, curriculum and demographic domains were significantly associated with the adoption and implementation of evidence-based tobacco use prevention curricula; however, in the univariate first-stage analyses, only 10 of 24 variables were significantly associated. Unadjusted analyses revealed that more of the organizational-level variables than provider- or curriculum-level variables were correlates. Specifically, significant relationships were found for 5 organizational-level variables, 3 provider-level variables, 1 curriculum-level variable and 1 school demographic variable, including Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 171 principal leadership of tobacco use prevention education, principal commitment to teach tobacco use prevention education, shared vision/goals, innovativeness, collaboration, provider perceived mandate to use effective tobacco use prevention curricula, provider expectations and values regarding tobacco use prevention programming, provider self-efficacy to implement tobacco use prevention programming, provider regularly uses skills compatible with interactive instructional methods (i.e., discussion, activities, role playing) specified in the curriculum and school type (being a junior/middle school). Three other variables (2 organizational- level variables, and 1 school demographic variable) approached levels of statistical significance (p< 10), including principal knowledge of effective tobacco use prevention programs, school climate, and population density (being located in an urban area). Among these significant first-stage univariate relationships, the highest odds ratio was found for shared vision/goals, an organizational-level (school culture) factor. Specifically, schools that had teachers and principals who scored high on the shared vision/goals index had over 20 times the odds of adopting and implementing an evidence-based tobacco prevention curriculum compared to those schools that had teachers and principals who scored low on the shared vision/goals index. Other organizational variables that had a strong relationship with adoption and implementation of evidence-based tobacco use prevention curricula were the school culture variables, innovativeness and collaboration. Schools that had teachers and principals who scored high on innovativeness and collaboration were approximately Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 172 4 times as likely to adopt and implement an evidence-based tobacco prevention curriculum compared to those schools that had teachers and principals who scored low on these measures. The hypothesis related to principal leadership as the key correlate of whether an evidence-based curriculum was adopted and implemented was not fully supported, although univariate results did provide evidence that this variable was strongly related. This is somewhat consistent with research indicating that the school principal is central and most closely associated with leading and supporting change and improvement within schools (Fullan, 1991, 1992; Hall & Hord, 1987; Hord, 1992; Miles, 1983), such as successfully adopting and implementing effective research-based prevention programs. The second- and third-stage multivariable models provided only some support for the hypothesis that variables from each domain were significantly associated with the adoption and implementation of evidence-based tobacco use prevention curricula. In the second-stage analysis, organizational- (principal characteristics and school culture), provider- and curriculum-level variables were significantly related with the adoption and implementation of evidence-based tobacco use prevention curricula. However, many of the variables significant in the first stage of analysis were no longer significant. School culture variables, as a group, remained as strong correlates of the adoption and implementation of evidence-based tobacco use prevention curricula. In the third-stage multivariable model, where all statistically significant predictors from the second-stage model Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 173 were entered simultaneously, only one variable remained statistically significant, shared vision/goals (an organizational school culture variable). This finding is supported by previous research which indicates that characteristics of the school culture are important influences in the adoption and implementation of schoolwide change efforts. Studies have shown that in order to make significant and lasting school change, the culture of a school must be addressed (Bolman & Deal, 1993; Conley, 1997; Cunningham & Gresso, 1993; Stoll & Fink, 1996). However, reform efforts typically have focused on changing the formal organizational structures of the school, while neglecting its informal and cultural dimension. It is believed that systemic change will not occur by introducing some specific modification, policy, or regulation. Rather, transformational change will be most successful if it is made at a deeper level, which would entail changing the culture of the school. School culture resides in the members of the organization, and it develops and evolves over time. It is the basic beliefs and assumptions, which operate unconsciously, that the administrators, teachers, and support staff hold about each other, the school, the students, and the community. One explanation for the way school culture, and more specifically shared vision/goals, influence decisions of adoption and implementation of tobacco use prevention curricula may be that teachers must feel a sense of coherence and continuity if meaningful and substantive change is to occur in their classrooms. Specifically, teachers may be more likely to adopt and implement evidence-based curricula if they feel that their school possesses clear and well understood goals that Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 174 are shared among school members. In turn, teachers may gain a sense that their personal choices in their classroom are all related to a common, generally understood, non-fragmented overall purpose. Research on school change has long supported the idea that common and mutually understood goals are vital for any successful change effort (Conley, 1997; Fullan, 1993; Hall & Hord, 1987; Hord, 1992; Nanus, 1992). Goals help provide coherence and focus for change efforts. The setting of clear goals may help to guide teachers’ behaviors in a way that increases the likelihood that goals will actually be achieved. As important as it is to set goals, it is perhaps even more critical that goals be shared by all school members. The literature also suggests that collaboration in setting goals bolsters teacher efficacy; that is, the extent to which teachers believe they can influence student learning (Fullan, 1991; Hall & Hord, 1987; Stoll & Fink, 1996). Fourth-stage models indicated that there was no support for any of the hypothesized interactions. Specifically, no significant associations (other than the main effects of school type) were found between adoption and implementation of evidence-base tobacco curricula and each of the three hypothesized interaction models: (a.) Schools that have an interactive combination of strong principal leadership of tobacco use prevention education and strong provider self-efficacy to teach tobacco use prevention education; (b.) Schools that have an interactive combination of strong principal leadership of tobacco use prevention education and that the provider regularly uses skills compatible with interactive instructional methods (i.e., discussion, activities, role playing) specified in the curriculum; and (c.) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 175 Schools that have an interactive combination of strong provider self-efficacy to teach tobacco use prevention education and strong provider perception that the curriculum has been demonstrated or endorsed as being effective. These interaction results suggest that different values of the moderating variable do not have an effect on the relationship of interest. For example, higher scores of the moderator (provider self- efficacy to teach tobacco use prevention education) did not change the relationship between the degree of principal leadership of tobacco use prevention education and the implementation or non-implementation of an evidence-based tobacco use prevention curricula. Adoption and Implementation o f Evidence-based Tobacco Use Prevention Programming. Results indicated some support for the primary hypothesis that variables from the organizational, provider, curriculum and demographic domains would be significantly associated with the adoption and implementation of evidence- based tobacco use prevention programming (as recommended by the CDC). However, in the univariate first-stage analyses, only 6 of 24 variables were significantly associated with the adoption and implementation of evidence-based tobacco use prevention programming. Unadjusted analyses revealed that provider- level variables were most important in this relationship, followed by organizational- level variables. Specifically, significant relationships were found for 1 organizational-level variable, 4 provider-level variables and 1 curriculum-level variable, including adequate resources (training, materials and time) to teach tobacco use prevention education, provider commitment to teach tobacco use prevention Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 176 education, provider knowledge of effective tobacco use prevention programs, provider perceived mandate to use effective tobacco use prevention curricula, provider expectations and values regarding tobacco use prevention programming, and provider regularly uses skills compatible with interactive instructional methods (i.e., discussion, activities, role playing) specified in the curriculum. The strongest relationship among these significant first-stage univariate associations was found for provider expectations and values regarding tobacco use prevention programming (a provider-level variable) and the adoption and implementation of evidence-based tobacco use prevention programming at the school. Specifically, a statistically significant and positive association existed between teachers’ beliefs that tobacco use prevention programming was very effective in preventing or reducing tobacco use among their students, and a higher degree of adoption and implementation of these effective tobacco use prevention programming practices at the school. Adequate resources to teach tobacco use prevention education (an organizational-level structure and processes variable) was the second strongest correlate, indicating that adequate training, teaching materials and time to teach tobacco use prevention education, as reported by teachers, were significantly associated with adoption and implementation of evidence-based tobacco use prevention programming practices at the school. Results from the second- and third-stage multivariable models provided only some support that variables from each of the domains were significantly associated with the adoption and implementation of evidence-based tobacco use prevention Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 177 programming. In the second-stage analysis, organizational- (structure and processes), provider- and curriculum-level variables were significantly related to adoption and implementation of evidence-based tobacco use prevention programming; however, many of the variables significant in the first stage of analysis were no longer significant. At this second stage, adequate resources to teach tobacco use prevention education showed the strongest association with the adoption and implementation of evidence-based tobacco use prevention programming. In the third-stage multivariable model, where all statistically significant predictors from the second- stage model were entered simultaneously, four of the 5 variables remained statistically significant, including principal knowledge of effective tobacco use prevention programs, participatory decision-making, adequate resources to teach tobacco use prevention education, and provider expectations and values regarding tobacco use prevention programming. Specifically, these third-stage results indicated that high levels of principal knowledge of effective tobacco use prevention programs, adequate resources to teach tobacco use prevention education, and provider expectations and values regarding tobacco use prevention programming, combined with low levels of participatory decision-making (an unanticipated finding), were associated with adopting and implementing evidence-based tobacco use prevention programming practices at the school. All but one of these significant variables (low levels of participatory decision-making) has been supported in the research literature on school change. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 178 In particular, previous research has supported the proposition that knowledge about an innovation is a key criterion for the adoption of new educational programs (Fullan, 1992; Paulussen et al., 1995). Principals who have sufficient valid information are in a better position to make well informed choices, and thus, are expected to make decisions that will advance their organization toward its goals. Research studies have also shown the provision of adequate resources is a necessary antecedent of successful adoption and implementation of educational programming. Material supports, such as funds to pay for program materials or training workshops, have been routinely documented in the school change literature to facilitate implementation of programming (Fullan, 1992; Huberman & Miles, 1984; Miistein, 1993; Yin & White, 1984). Specifically, schools that have administrators who provide assistance in the form of support personnel (coordinator or consultants), or release time to attend program trainings, have proved to be more likely to continuously implement programming (Fullan, 1992; Huberman & Miles, 1984). Further, considerable research indicates that an individual’s expectations and values help to determine an individual’s behavior (Bandura, 1986, 1997). Beliefs about outcomes likely to occur if one performs a behavior influence the types of behavior one chooses to engage in, the level of perseverance, and the amount of effort expended to enact the behavior. The more highly valued the expected outcome, the more likely the person will perform the behavior to yield the outcome. Thus, intervention research has shown that providers with high expectations and values Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 179 regarding tobacco use prevention programming have higher adoption rates of tobacco prevention education programs (Parcel, O’Hara-Tompkins et al, 1995). In contrast, the current study’s finding that a low level of participatory decision-making was associated with a higher degree of the adoption and implementation of evidence-based tobacco use prevention programming is not consistent with the literature. In fact, research has consistently shown the importance of subordinate participation in decision making regarding innovations. Shared decision making offers a feeling of control to the individuals that are affected by the change efforts. Empirical studies have found that schools that engage in participatory decision-making have higher decision quality, increased ownership, commitment and satisfaction, and reduced resistance to change, and thus, will have the greatest impact on the degree to which an innovation is successfully adopted and implemented (Bernd, 1992; Boyd, 1992; Stoll & Fink, 1996). One explanation for this unexpected result may be that a school must have a principal who is very knowledgeable about effective tobacco use prevention programs, able to obtain sufficient resources (training, materials and time) for the teachers, and is more, not less, authoritative when it comes to making decisions for the school. In other words, the school must have a strong principal who knows what to do, can garner the resources needed, and makes unilateral decisions about the tobacco use prevention programming at the school, all of which, in turn, facilitates and supports their teachers’ efforts to implement an innovation. This rationale is supported by school research that has concluded that principal influence by itself Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 180 does not directly account for school improvement; instead, principals are key players in facilitating the process of change (Hall & Hord, 1987). Regarding the fourth-stage models, results were similar to those found for the dichotomous outcome variable of implementation (or non-implementation) of evidence-based tobacco use prevention curricula; that is, there was no support for any of the hypothesized interactions. Specifically, no significant associations were found between adoption and implementation of evidence-base tobacco use prevention programming and each of the three hypothesized interaction models: (a.) Schools that have an interactive combination of strong principal leadership of tobacco use prevention education and strong provider self-efficacy to teach tobacco use prevention education; (b.) Schools that have an interactive combination of strong principal leadership of tobacco use prevention education and that the provider regularly uses skills compatible with interactive instructional methods (i.e., discussion, activities, role playing) specified in the curriculum; and (e.) Schools that have an interactive combination of strong provider self-efficacy to teach tobacco use prevention education and strong provider perception that the curriculum has been demonstrated or endorsed as being effective. Summary of Results for Study One Overall, the results provided support for the primary hypothesis that variables from the organizational, provider, curriculum and demographic domains would be associated with the adoption and implementation of evidence-based tobacco use prevention curricula and programming. However, it was found that adoption and Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 181 implementation of evidence-based tobacco use prevention curricula and programming were not positively associated with each individual variable listed in Table 2, and with each level of factors (i.e., organizational, provider, curriculum, and demographic) as hypothesized. Further, results were markedly different for the associations with the two outcome variables. That is, a different combination of correlates was found for the adoption and implementation of evidence-based tobacco use prevention curricula than for the adoption and implementation of evidence-based tobacco use prevention programming (as recommended by the CDC). For example, the hypothesis that organizational variables would be the most important group of correlates was supported for the dichotomous adoption and implementation of evidence-based tobacco use prevention curricula outcome variable, but not for the adoption and implementation of evidence-based tobacco use prevention programming variable. Regarding the hypothesis related to principal leadership as the strongest correlate of whether evidence-based curricula or programming was adopted and implemented, the relationship was supported in the univariate results for the adoption and implementation of evidence-based tobacco use prevention curricula, but not for the adoption and implementation of evidence-based tobacco use prevention programming. Moreover, the final multivariate models provided further evidence that different sets of factors are associated with the two outcome variables. That is, schools that have a high level of shared vision/goals (an organizational-level characteristic) are more likely to adopt and implement evidence-based tobacco use Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 182 prevention curricula. Whereas, schools that have certain organizational characteristics (i.e., strong principal knowledge of effective tobacco use prevention programs, more resources, and lower levels of participatory decision-making) and provider characteristics (i.e., strong teacher expectations and values about tobacco use prevention programming) are more likely to adopt and implement evidence- based tobacco use prevention programming. A possible explanation as to why a different combination of correlates was found for the two outcome variables (adoption and implementation of evidence- based tobacco use prevention curricula and adoption and implementation of evidence-based tobacco use prevention programming) concerns decision-making. That is, the teacher is the sole decision-maker of whether an evidence-based curriculum is actually adopted and implemented (the dichotomous outcome) because the choice is made within the classroom. Whereas, multiple decision-makers (including the principal, teachers, and staff) are involved in making decisions about, and that lead to, adoption and implementation of school-wide programming (the continuous outcome), which includes policy enforcement, teacher training, parent involvement and the existence of on-site cessation programs for student smokers. As discussed earlier, school culture variables may be the key influence for the teacher. However, a broader combination of variables (including principals’ knowledge of effective programs, adequate resources, and teachers’ expectations and values regarding tobacco use prevention programming) may be more important for multiple decision-makers. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 183 Study Two: Longitudinal Study The longitudinal study sought to provide further information about the diffusion process by examining the importance of the interactive effect between the degree of implementation of evidence-based tobacco use prevention programming and a newly-developed school capacity to implement innovative programming construct (derived from organizational-, provider- and curriculum-level variables), with subsequent changes in adolescent tobacco use-related beliefs, attitudes, skills and behaviors. Although a few significant main effect relationships were found among the predictor and outcome variables, results provided only minimal support for the hypothesis that school capacity to implement innovative programming would moderate the relationship between implementation of evidence-based tobacco use prevention programming and each of the seven student program outcome variables. In particular, the first-stage models showed significant main effects for both predictors (implementation of evidence-based tobacco use prevention programming and school capacity to implement innovative programming indices) on adolescent tobacco use-related program mediators and behavioral outcomes. However, the implementation of evidence-based tobacco use prevention programming index had more utility than the school capacity to implement innovative programming index in predicting improvements (over the four-year program period) in student program outcomes. Implementation o f Evidence-based Tobacco Use Prevention Programming. First-stage model results indicated that three of the seven relationships were Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 184 statistically significant. That is, the degree of implementation of evidence-based tobacco use prevention programming had an effect on students’ negative outcome expectancies of smoking, lifetime cigarette smoking and 30-day cigarette smoking. The strongest predictive relationship existed between implementation of evidence- based tobacco use prevention programming and negative outcome expectancies of smoking. Specifically, this finding suggests that a higher degree of implementation of evidence-based tobacco use prevention programming (based on 5 of the 7 CDC Guidelines for School Health Programs to Prevent Tobacco Use and Addiction) was predictive of significant increases (over the four-year program period) in students’ agreement with statements about the negative consequences of tobacco use (from a mean of 3.40 to 3.47 on a 1-4 point scale). Likewise, a higher degree of implementation of evidence-based tobacco use prevention programming was predictive of significant decreases (over the four-year program period) in the prevalence of lifetime cigarette smoking (from 54.43% to 45.99%) and 30-day cigarette smoking (from 22.47% to 16.68%). School Capacity to Implem ent Innovative Programming. Results from the first-stage models indicated that school capacity to implement innovative programming was predictive of only one of the student outcomes, tobacco refusal self-efficacy. This finding revealed that a higher degree of a school’s capacity to implement innovative programming was predictive of significant increases (over the four-year program period) in students’ beliefs that they could refuse a best friends’ cigarette offer (from a mean of 3.44 to 3.54 on a 4 point scale). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 185 SeCond-stage model results provided only minimal support for the main hypothesis that school capacity to implement innovative programming would have a significant moderating effect on the relationship between implementation of evidence-based tobacco use prevention programming and improvements in program outcomes. Specifically, a significant interaction effect was found for only one of the seven outcome variables. School capacity to implement innovative programming was found to have a significant moderating effect on the relationship between implementation of evidence-based tobacco use prevention programming and perceived peer norms of smoking; although, there was no significant main effect for either the predictor (implementation of evidence-based tobacco use prevention programming) or the moderator (school capacity to implement innovative programming). More specifically, results from the interaction models indicated that the degree of implementation of evidence-based tobacco use prevention programming was not predictive (over the four-year program period) of decreases in students’ estimates of the proportion of peers who smoke monthly (from 47.24% to 41.63% on a 0-100% scale). Further, the degree of school capacity to implement innovative programming also did not predict the significant decreases in this student outcome. Thus, neither a high degree of implementation of evidence-based tobacco use prevention programming nor a high degree of school capacity to implement innovative programming, by itself, predicted decreases in perceived peer norms of smoking. However, perceived peer norms of smoking was significantly reduced in Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 8 6 schools with both a high degree of implementation of evidence-based tobacco use prevention programming and a high degree of school capacity to implement innovative programming. These findings are in line with considerable research that has shown that the implementation of empirically-validated research-based curricula and prevention program practices produces consistent and significant improvements in adolescent tobacco use-related knowledge, beliefs, attitudes, skills and, most importantly, behaviors (Botvin et al., 1995; California Department of Education, 2000; Centers for Disease Control and Prevention, 1994; Ellickson et al., 1993; Flay et al., 1985b; Pentz et al., 1989; Sussman et al., 1995). Further, these findings also support the very important need for research studies to investigate the interrelationships among multiple predictors of implementation (through utilization of interaction models, not only main effects models), or else risk that significant existing relationships may go undetected (Greenberg et al., 2000). Summary of Results for Study Two Overall, the results provided only minimal support for the main hypothesis that school capacity to implement innovative programming would have a significant moderating effect on the relationship between implementation of evidence-based tobacco use prevention programming and program outcomes. Only one significant interaction effect existed; that is, a significant moderating effect was found for decreases in students’ estimates of the proportion of peers who smoke monthly. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 187 One explanation for the lack of support for the hypothesis that school capacity to implement innovative programming would have a significant moderating effect on the relationship between implementation of evidence-based tobacco use prevention programming and program outcomes may be due to the measures selected to represent a school’s capacity to implement innovative programming. Perhaps variables other than the 7 used (school leadership, school commitment to teach tobacco use prevention education, school support for teaching tobacco use prevention, school climate, school collaboration, teacher expectations and values regarding tobacco use prevention programming, and teacher self-efficacy to implement tobacco use prevention programming) would be a better measure of a school’s capacity to implement innovative programming. For example, the index may need to include more variables that address teacher- and curriculum-related constructs. Limitations Certain methodological limitations of the overall study need to be considered. First, Study 1 ’s use of cross-sectional data raises questions about causation. Correlational data do not provide evidence of whether organizational, provider, and curriculum characteristics play a causal role in the adoption and implementation of evidence-based tobacco use prevention curricula or whether other unidentified variables may account for these relationships. A longitudinal design is needed to help establish a temporal relationship indicating which variable precedes and predicts which other variable. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 138 Second, data in this study were generated from self-report surveys, the accuracy of which could not he independently verified . Thus, it is impossible to assess the extent to which such data may be biased, particularly in regard to social desirability among teachers, principals and students. Regarding teachers and principals, it is possible that bias entered into self-reports consciously and unconsciously. Respondents may knowingly bias their responses in one direction because they want to portray a flattering picture of themselves or they may be inaccurate because they forget what happened. Because many of the measures in this study required that teachers and principals rate their own skills and behaviors (e.g., participator decision-making, communication, collaboration, etc.), social desirability may have been a major contaminant. Future studies may try to remedy this type of bias by obtaining information from sources other than the teachers and principals, such as other teachers and staff at the school or trained observers. However, with a final sample of over 200 teachers and principals across Study 1 and Study 2, self-report measures provided the most feasible assessment of knowledge, beliefs, attitudes, skills and behavior, in terms of ease and cost, for this dissertation study. Regarding youth and the accuracy of self-report measures, it is possible that many students may have underreported their smoking status. Unfortunately, this study did not utilize collateral measures of smoking, such as reports from other sources (such as teachers, peers or family members) or biochemical validation. Biochemical validation of adolescent tobacco use is of questionable validity because Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 189 adolescents may metabolize nicotine differently than adults (Stacy et al, 1990). However, these physiologic measures can be useful in conjunction with self-report questionnaires by increasing the validity of self-report measures. Research on biochemical validation including “pipeline” procedures, in which subsamples of participants actually have their samples tested by a biochemical device, suggests that the knowledge that physiological indicators ofbehavior will be assessed may increase the accuracy of self-reports ofbehavior (Sussman et al, 1995). Increasing accuracy in adolescent self-report behavior is extremely important in determining the effectiveness of adolescent prevention interventions. Thus, this type of validation is necessary and should be used, whenever possible, in smoking studies conducted among youth. Nonetheless, past studies have supported the validity and reliability of self-report measures of non-normative adolescent behavior, including cigarette use (Graham et al., 1984). Further, considerable research indicates that the accuracy of self-reporting cigarette or other drag use among adolescents, for example, is mainly determined by whether they have been assured anonymity and confidentiality (Hansen et al., 1985). Third, although this study achieved a high response rate for teachers overall (over 80% in both studies), the response rate for teachers who taught health, science, and/or physical education, which are the subject areas in which tobacco lessons are typically taught, was relatively low. Further, Study 1 ’s response rate for principals was under 70%. Even though a random sampling technique was utilized, this low response rate may have lead to a sample that was not representative of the population Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 190 chosen. When there is high nonresponse, a random sample will begin to resemble a sample recruited through volunteering. Fourth, another limitation of this study concerns reliability of the measures in Study 1. Cronbach’s alpha reliability for many predictor scales was acceptable (i.e., .60), but not high. This may have been due to averaging of teachers’ and principals’ responses. It may limit ability to detect true relationships between these variables and outcomes. However, from a validity perspective, it enhanced the validity of measures (e.g., on principals’ commitment to tobacco use prevention education) to get both teachers’ and principals’ perspectives. Fifth, some of the measures in this study were limited in terms of the specific measures selected, and this may have affected internal validity. For example, the content validity of Study l ’s measure of a provider’s perceived mandate to use effective tobacco use prevention curricula may have been insufficient due to the availability of only one item in the data set to measure this construct. Past research indicates that theoretically expected associations have not been detected when only one item was used to assess a psychological construct (Rushton et al., 1983). Thus, it is possible that the item selected to represent a provider’s perceived mandate to use effective tobacco prevention curricula did not capture the construct, and a more comprehensive measure would have been more effective in detecting a relationship with the adoption and implementation of evidence-based tobacco use prevention curricula and programming. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 191 Finally, the results of Study 1 should be viewed and interpreted with caution due to issues of multicollinearity among the independent variables under investigation. Results from the correlational analyses indicated that many of the independent variables were, in fact, correlated with each other. When independent variables are highly correlated, it is difficult to disentangle the unique effects of each independent variable on the dependent variable. Further, a factor analysis using the Promax rotation method utilized for intercorrelated variables was performed (but not reported in this paper) which indicated that many of these intercoirelated variables were loading on only one or two factors. An examination of only one or two super factors would not be very useful in interpreting results; instead, this study sought to retain its original analysis plan of examining the contribution of each variable individually and then by each domain. Moreover, the original plan was followed primarily because it is believed that the selected variables are measuring distinct dimensions, as supported by previous theoretical and empirical research. For example, regarding principal characteristics, numerous research studies on educational administration, school change, and school effectiveness lend empirical support to the assumption that each of the three principal variables in this dissertation study (i.e., principal leadership, principal’s commitment to tobacco use prevention education, and principal’s knowledge of tobacco use prevention curricula) are assessing discrete constructs (Fullan, 1991, 1992; Hall & Hord, 1987; Hord, 1992; Miles, 1983). Specifically, in their extensive research on principals as facilitators of change in schools, Hall & colleagues (1987) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 192 operationally-defined different attributes of the principal into various categories and found that the distinct types of principal traits, styles, and behaviors affected program implementation differently, and at different stages, indicating that they are measuring distinct dimensions. Future studies may address the issue of multicollinearity and discriminant validity by using an analysis plan that differs from the one used in this dissertation. For example, the present study conducted multivariate analyses (at the second stage) to determine which individual variables were most important at each domain of factors (i.e., organizational, provider, etc.). That is, variables within each domain (e.g., among principal characteristics only) were simultaneously entered into a multivariate regression model. Future analysis plans may want to determine the importance of the variables after a factor analysis is performed on all variables within a domain. That is, at the second stage, factors (determined from a factor analysis), not individual variables, are entered into a multivariate regression model. Another method that may be appropriate to deal with multicollinearity of the variables is to employ structural equation modeling techniques (that account for multicollinearity of variables simultaneously) on longitudinal data. Implications and Future Directions Despite these limitations, the results of this study provide new evidence regarding the relative importance of theory-based organizational-, provider-, curriculum- and demographic-level constructs on the adoption and implementation of evidence-based tobacco use prevention curricula and programming in schools, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 193 along with subsequent changes in adolescent tobacco-use related beliefs, attitudes, skills and behaviors. In particular, several findings of this study have important implications for diffusion of innovations in school settings, including the approach of future empirical studies investigating predictors of adoption and implementation of evidence-based tobacco use prevention curricula and programming, the design of effective interventions to encourage adoption and implementation of evidence-based tobacco use prevention curricula and programming, and governmental and school district-level policy decisions regarding tobacco use prevention education in our nation’s schools. Im plications and Future D irections fo r the Design o f Em pirical Studies. Although the results provided support for the primary hypothesis that variables from the organizational, provider, curriculum and demographic domains would be associated with the adoption and implementation of evidence-based tobacco use prevention curricula and programming, not all variables, including the interactions, were significantly associated as hypothesized. Thus, this study’s lack of findings identifies the need for more investigations of the diffusion of innovations within schools, in particular studies that focus on implementation phase because this is the stage where the program actually is put into use and gets delivered to the targeted population, the students. The implementation phase is of critical importance at this point in time, in light of recent national data indicating only small and sporadic declines in adolescent drug use over the last decade (Johnston, 2002). Specifically, the widespread Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 194 implementation of ineffective or unevaluated programming may account, in part, for the lack of significant decline in smoking prevalence among youth in California during the early and mid 1990’s (Elias, 1997). These weak or negative findings may eventually cause governmental policymakers, school administrators, and the general public to lose faith in school-based substance use prevention programs. Thus, studies are needed that provide evidence that if effective-based programs are implemented with high fidelity, they will achieve maximum impact. Further, this study’s positive significant findings or lack of findings warrant replication in other samples, incorporating additional measures and potential predictors of adoption and implementation of evidence-based tobacco use prevention curricula and programming. Variables other than the ones chosen for this investigation may account for a substantial amount of the variance in the adoption and implementation of evidence-based tobacco use prevention curricula and programs. A potentially important group of factors that were unavailable, and thus not addressed, in this dissertation study are those at the school district-level. Thus, future studies also need to examine the role of district-level variables on the adoption and implementation of school-based tobacco use prevention programs and subsequent changes in youth tobacco use behavior. To date, numerous research studies have provided empirical evidence of the various school district-level characteristics that are associated with successful change and improvement within schools. However, relatively few of these studies have investigated factors specific to the prediction of adoption and implementation of innovative school-based tobacco Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 195 use prevention programs; rather, factors have been investigated largely in terms of other types of successful school change and improvement, such as providing professional development to staff or increasing student test scores. It is believed, however, that the same factors and processes that determine various changes within schools may also determine the successful adoption and implementation of tobacco use prevention programs in schools. In terms of research specific to the implementation of school-based tobacco use prevention curricula, only two intervention studies (Smith et al., 1993 and Goodman et al., 1991) investigated district-level variables associated with the implementation of tobacco use prevention programs in schools; however, findings were mixed. Smith and colleagues (1993) found that the presence of supportive district-level administrators was significantly associated with implementation of the curriculum. Further, teachers in districts with administrative support implemented a greater percentage of tobacco use prevention curriculum activities than teachers without district-level support. One other study conducted by Goodman and colleagues (1991) tested an intervention aimed at increasing the diffusion of school- based tobacco use prevention programs, while focusing on all four stages of the diffusion process, dissemination, adoption, implementation, and sustainability/maintenance. However, district-level variables were tested at only the dissemination and sustainability/maintenance phases. At the sustainability/maintenance stage, the strategy of process consultation focused on the political skills of those individuals who championed the program. Participants were Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 196 given skills instruction on how to build coalitions of program advocates so that the program could become entrenched within the school system. Results revealed that this strategy resulted in comparably low levels of maintenance in intervention and comparison districts (McCormick, et al., 1995). The literature indicates that district support, whether in the form of commitment and support, communication, direct assistance, structure and processes, or planning and assessment, figures largely in the positive impact of changes made in schools. Regarding commitment, it has been found that general and verbal support from the school-district level administrators will not suffice; instead, priority must be given to the change effort if implementation of change is expected to occur (Fullan 1991, 1992). Having a supportive environment is also an important determinant of successful school change. Numerous studies have found that school districts must help to create an overall environment that supports school efforts to improve (Berends et al., 2001, 2002; Bodily et al, 1998; Deal et al., 1975). The elements of supportive environments include helping schools build leadership, trust, ownership, and a shared vision of change among all school staff. School districts that possess features of a positive organizational climate, such as openness, free communication and trust, are more likely to have their schools implement change efforts successfully (Bodilly et al., 1998; Deal et al., 1975). Similar to the current study’s findings for school-level shared vision/goals, school districts that have a clear vision and mission are also more successful at school change (Arnold et al. 1999). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 197 Other studies have shown that dear and consistent communication and pressure from district-level administrators is needed, both initially and during implementation, in order to achieve greater rates of school change (Fullan 1991, 1992). However, it has also been found that administrative pressure must be accompanied by assistance to enable the change (Fullan, 1991; Huberman & Miles, 1984). Assistance requires adequate resources for training, release time, materials, and technical assistance throughout the change effort. Studies have shown that school districts that were effective in mobilizing district resources were more likely to be successful in implementation of technological innovations (e.g., computers) in the classroom (Arnold et al. 1999; Corbett et al., 1984; Fullan, 1991; Huberman & Miles, 1984). This finding regarding adequate resources provided by the school district mirrors the current study’s finding that the provision of adequate resources from within the schools was associated with the adoption and implementation of evidence-based programming. Some researchers have emphasized the relationship between the characteristics of school district structure and processes and the adoption and implementation of innovations. The degree of decentralization of power and authority from school districts to schools is viewed as an important structural factor that influences an organization’s capacity to change. It is believed that the traditional command and control model of organizational structure is not suitable for school reform efforts because top-down management reinforces fear, distrust and internal competition, and reduces collaboration and cooperation. It may lead to compliance, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 198 but commitment is required for lasting change. Power sharing generates increased empowerment, and subsequent motivation and interest in school improvement among teachers (Fullan, 1992). Accordingly, studies have revealed that school districts that support organizational autonomy and decentralized authority are more likely to have their schools implement innovations (Berends et al., 2001; Bodilly, 1998; Deal et al., 1975; Louis & Miles, 1990). The importance of subordinate participation in decision-making regarding innovations has also been given great importance in the school improvement literature. Although not consistent with the current study’s findings, other research has found that school districts that engage in participatory decision-making with their schools results in higher decision quality, increased ownership, commitment and satisfaction, and reduced resistance to change within schools, and thus, will have the greatest impact on the degree to which an innovation is successfully adopted and implemented (Bemd, 1992; Bodilly, 1998; Deal et al., 1975; Louis & Miles, 1990). Regarding planning and assessment of change efforts, research indicates that successful school improvement has been associated with school district-level support such as creating strategic plans for improvement, using data to drive reform, and measuring progress so that the process of change becomes a cycle of continuous improvement (Berends et al., 2001; Bodilly, 1998). The process of matching strategies to school needs is most effective when school districts have a detailed understanding of the needs of a particular school and student population. School districts must help schools document their efforts and refine their strategies as Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 199 needed in order for the school to meet higher expectations. A district-wide emphasis on continuous evaluation has been found to be an important factor in helping schools monitor the change process and assess whether students are achieving goals (Berends et al., 2001). Again, while it is believed that the same factors and processes that determine various changes within schools may also determine the successful adoption and implementation of school-based prevention programs, much more empirical research is needed to establish the specific relationship between school district-level factors and implementation of tobacco use prevention programs in schools. Im plications and Future Directions fo r the Design o f Effective Interventions. Despite the resources currently expended on empirically-validated school- based tobacco use prevention programs, these programs will not be successful in improving adolescent tobacco use outcomes unless they are systematically adopted and implemented by teachers in schools. The results of this study have important implications for the design of effective interventions to encourage adoption and implementation of evidence-based tobacco use prevention curricula and programming. The current study’s findings may help to suggest exactly what types of interventions may be most effective. For example, based on the specific finding indicating that provider expectations and values regarding tobacco use prevention programming was significantly associated with the adoption and implementation of evidence-based programming, an intervention may be most beneficial if it is Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 200 designed to increase teachers’ beliefs that tobacco use prevention programming is very effective in preventing or reducing tobacco use among their students. To achieve maximum impact from an intervention, the concept ofbaseline x treatment interaction effects as indicated by Kellam and colleagues (Kellatn et al, 1994; Kellam & Van Horn, 1997) may be utilized when designing an intervention. A baseline-treatment interaction represents systematic (non-random) differences among individuals at different baseline levels in the outcome benefit they receive as a result of an intervention. The most relevant relationship is between an individual’s initial baseline status and the change due to an intervention. In other words, of interest is whether treatment effect of an intervention is different depending on the baseline value of a covariate. Baseline x treatment interactions are often examined in intervention prevention evaluations that show no main effects, or in studies where the data indicate that different groups may benefit from the intervention differently. A hypothesized training intervention (based on the current study’s findings) directed at increasing levels of adoption and implementation of evidence-based tobacco use prevention curricula among teachers may be useful in illustrating this. Teachers with low expectations and values regarding tobacco use prevention programming may benefit more from an intervention than teachers with high levels of this construct. Teachers at different baseline levels may require different levels of intervention. Further, an intervention program that universally targets teachers may not be vigorous enough for those with high baseline levels and may be irrelevant for those in the low baseline level group. Thus, in this case, treatment may only be Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 201 effective for teachers in the moderate baseline. Thus, an intervention might be designed to appropriately target intervention training to each subgroup. That is, teachers could be categorized into groups (low, medium, or high) depending on their baseline level of expectations and values regarding tobacco use prevention programming, and given targeted training that would include different dosage and instructional methods. This type of targeting would suggest whether the effects of the interventions were for the most part due to changes in the specific processes/practices targeted for change. Thus, these results would help to support the theoretical model underlying the intervention (Ialongo et al., 1999). If baseline x treatment interactions, as in the example described above, are found to exist, they may inform future intervention design and implementation. They may indicate for whom the intervention is effective, under what conditions the intervention would be effective, and exactly which interaction effects there are between the individual (including background characteristics and experiences) and the treatment. This information may enable researchers to develop appropriately- targeted interventions, which are more efficient and cost effective. Furthermore, these interaction effects may provide information to better understand the change in the underlying processes as a result of the intervention. Methodological approaches to baseline x treatment interaction effects can be applied to the study of baseline x treatment intervention interaction effects (Ialongo et al., 1999; Kellam et al., 1994). Specifically, stratified designs and statistical analysis procedures have been identified that are useful in assessing baseline x Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 202 treatment interaction effects (Kellam et al., 1994; Muthen & Curran, 1997). For example, if the effect of treatment is expected to vary substantially across important pre-specified subgroups (e.g., teachers with low, medium, or high expectations and values), then stratifying for these subgroups can help in interpreting the treatment effect and its consistency across these subgroups. Stratified designs can also enhance the credibility of some subgroup analyses that are a priori of high interest. If such an interaction is expected, then the intervention study should be designed to have adequate power to detect treatment effects within specific subgroups. Furthermore, use of multiple-group latent growth modeling techniques may be most useful in assessing treatment effects (Curran, 2000; Muthen & Curran, 1997). Im plications and Future D irections fo r Governmental and School D istrict-level Policy Decisions. Results from this study also may offer may offer guidance for policymakers, including governmental policymakers who are involved in drug use prevention research and practice, and local policymakers such as district-level administrators. For example, results indicate that schools that have teachers who indicated that they had adequate resources (training, materials and time) to teach tobacco use prevention education were more likely to adopt and implement evidence-based tobacco use prevention programming. Therefore, in this example, this finding could inform governmental policymakers (including those at the federal, state, and local levels) of the need to allocate more drug use prevention education funds to schools specifically for training and materials. Whereas, local policymakers, such as district-level Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 203 administrators who might have significant hands-on discretion over day-to-day operations and tobacco use prevention education funds, may help to allocate the funds exactly where they might be needed most (i.e., the provision of time for teachers to teach tobacco use prevention versus additional training). Further, results of this study provided support for the hypothesis that providers who perceived a mandate to use effective tobacco use prevention curricula were more likely to adopt and implement an effective or research-based curricula. This finding points to the need for strong governmental policy that not only endorses but requires that schools use evidence-based tobacco use prevention curricula and programming. One such policy that is currently in effect is the Principles of Effectiveness. The Principles ofEffectiveness were developed by the U.S. Department of Education’s Safe and Drug Free Schools Program and implemented in 1998 to help teachers, school administrators, and prevention program developers achieve safe learning environments where students would be free from fear of violence and the influence of drugs. Compliance with the Principles was a prerequisite for public schools to continue receiving funds through the Safe and Drug Free Schools and Community Act (U.S. Department of Education, 1998). The Safe and Drug Free Schools Program is the single largest source of funding for school drug and violence prevention programs. The four principles required a school district to ensure that the program or activity: (1) be based on an assessment of objective data regarding the incidence of violence and illegal drug use in elementary schools, secondary schools, and communities to be served; (2) be based on performance Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 204 measures aimed at ensuring that these schools and communities have a safe, orderly, and drug-free environment; (3) be grounded in scientifically-based research that provides evidence that the program to be used will reduce violence and illegal drug use; and (4) be evaluated periodically relative to their goals and objectives. It was believed that these principles would prompt school districts to abandon their use of non-evidence based prevention programs and activities and adopt and implement science-based practices. Unfortunately, data collected in 1999 from public and private school districts indicated that only 34.6% of public schools and 12.6% of private schools used at least one of the ten curricula specified on the questionnaire (Ritigwali et al., 2002). The most popular curriculum being implemented was DARE (53.5%), followed by M cG ruff’ s Drug Prevention and C hild Protection (15.9%), and H ere’ s Looking at You 2000 (15.8%). Project ALERT(14.6%) and Life Skills Training (12.8%) were the most prevalent among the curricula identified as being effective. These findings suggest that schools had largely ignored the U.S. Department of Education’s Principles of Effectiveness, which required recipients of funding through the Safe and Drug Free Schools and Community Act to design and implement evidence-based drag use prevention research practices. However, another study conducted by Hallfors and Godette (2002) indicated that school districts that had implemented the U.S. Department of Education’s Principles ofEffectiveness policy were more likely to select a research-based program compared to school districts that had not implemented the policy. Results Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 205 also indicated that 59% of the Safe and Drag Free Schools Coordinators reported use of one or more of the six research-based programs identified on the questionnaire, including Life Skills Training, Project ALERT, Project STAR or 1-STAR, and Reconnecting Youth: A Peer Group Approach to Building Life Skills. Rohrbach and colleagues’ (in press) research also assessed the influence of governmental guidelines on the adoption and implementation of effective programming. Results provided evidence demonstrating that school districts that had consulted informational materials obtained from credible sources such as federal and state substance use prevention agencies (e.g., National Institute on Drug Abuse and Center for Substance Abuse Prevention) were significantly more likely to decide to adopt an evidence-based substance use prevention curriculum. Thus, from these few studies following the U.S. Department of Education’s promulgation of the Principles of Effectiveness on July 1, 1998, it appears that policy, with its incentive of funding, and other federal and state guidelines do influence school districts to reconsider their process for selecting and using evidence-based tobacco and other drug use prevention curricula. More recent developments in public policy have given the federal government another opportunity to help shape the way school-based tobacco and other drug use prevention programs are being adopted and implemented. Specifically, In January 2002, President Bush signed the No Child Left Behind Act o f2001, which is a reauthorization of the Elementary and Secondary Education Act of 1965 (U.S. Department of Education, 2003). This far-reaching education law was designed to Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 206 strengthen the federal pressure on all states to pursue a standards-based reform agenda. This includes high academic standards for all students; extra support to help students and schools meet those standards; increased flexibility for local schools in order for them to do so; and greater accountability for the results, particularly as measured by student performance on standardized tests. Under Title IV of the No Child Left Behind Act (which consolidates the Safe and Drug Free Schools and Communities Act (SDFSCA) and the 21st Century Community Learning Centers programs), local prevention programs and activities are now required to meet the Principles ofEffectiveness (U.S. Department of Education, 1998). This is the first time that the Principles ofEffectiveness have been codified by law. In fact, one of the significant changes in the SDFSCA is a requirement that State and local prevention programs and activities meet the Principles of Effectiveness. Under the reauthorized SDFSCA, the Principles of Effectiveness include a requirement that fimds be used to support only programs grounded in scientifically based research. Although the more recent No Child Left Behind legislation has promise to affect adoption and implementation of evidence-based substance use programs and activities as indicated in the above-mentioned studies, some reservations of its potential success still remain. Additional reasons exist to question the new law’s effectiveness in promoting effective research-based practices, which may be one of its lowest priorities. For example, to accomplish its many goals, the NCLB Act places greater demands on states and school districts than ever before. States must Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 207 meet certain regulatory requirements and procedures, expand their testing programs, analyze and report test results in new a new uniform management information and reporting system, provide technical assistance to underperforming districts and schools, and help teachers become better qualified, among other duties. School districts must raise test scores, close achievement gaps, design improvement strategies and interventions for underperforming schools, and create or expand public school choice programs, and much more. This may require states and school districts to deal forcefully with schools that are not improving student achievement. Thus, it remains to be seen how these numerous demands will be accepted and carried out. It is possible that states and school districts could turn from dedication to making this education reform successfiil to mere technical compliance with the law’s many detailed requirements. Furthermore, this law proposes that the federal government continue to provide only 7% of the total funding for public schools, and very limited technical assistance; however, it demands 100% accountability. Successful implementation of this law may also be threatened due to the fact the most states are in fiscal crises and overloaded with demands, and thus, may not be willing to accept the heavy accountability and demands of this law for the very limited federal aid. Finally, the provision of adequate resources is also a necessary antecedent of successful adoption and implementation of educational programming. Material supports, such as funds to pay for program materials or training workshops, have been routinely documented in the school change literature to facilitate implementation of programming (Fullan, 1992; Huberman & Miles, 1984; Milstein, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 208 1993; Yin & White, 1984). Specifically, schools that have administrators who provide assistance in the form of support personnel (coordinator or consultants) or release time to attend program trainings have also proved to be more likely to continuously implement programming (Fullan, 1992; Huberman & Miles, 1984). However, the No Child Left Behind Act proposes that schools implement evidence- based substance use prevention programming without much greater federal financial and technical assistance than before. Based on this information, it would seem reasonable to believe that we should expect higher prevalence rates for adoption and implementation of effective substance use prevention curricula or activities now that the Principles of Effectiveness have been codified by law in the No Child Left Behind Act. However, future policies and/or amendments should include more explicit federal guidance, with stronger enforcement of policy violations, and should also provide substantial federal or state resources and funding, if successful adoption and implementation of evidence-based tobacco use prevention curricula and programming is to be achieved. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 209 REFERENCES Ajzen, I. (1988) Attitudes, personality, and behavior. Chicago: Dorsey Press. Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50, 179-211. Arnold, M., Simms, N.G., Wilber, D M. (1999). Innovative approaches to maximizing resources. In Noteworthy perspectives on comprehensive school reform, (pp. 49-55). Aurora, Co: Mid-continent Regional Educational Laboratory. Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84, 191-215. Bandura, A. (1986). Socialfoundations o f thought and action: A social cognitive theory. New York: Prentice Hall. Bandura, A. (1997). Self-efficacy: The exercise o f control. New York: W.H. Freeman. Barth, R. (1988). School: A community of leaders. In A Lieberman (Ed.), Building a professional culture in schools (pp. 129-147). New York: Teachers College Press. Basen-Engquist, K., O’Hara-Tompkins, N., Lovato, C.Y., Lewis, M.J., Parcel, G.S., & Gingiss, P. (1994). The effect of two types of teacher training on implementation of Smart Choices: A tobacco prevention curriculum. Journal o f School Health, 64, 334-339. Battistich, V., Solomon, D., Kim, D., Watson, M., & Schaps, E. (1995). Schools as communities, poverty levels of student populations, and students* attitudes, motives, and performance, a multilevel analysis. American Educational Research Journal, 32(3), 627-658. Berends, M., Bodilly, S.J., & Kirby, S.N. (2002). Facing the challenges o f whole- school reform. New American schools after a decade. Santa Monica, C A: RAND. Berends, M., Kirby, S.N., Naftel, S. & McKelvey, C. (2001). Implementation and performance in New American Schools: Three years into scale-up. Santa Monica, CA: RAND (MR-1145-EDU). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 210 Berman, P. & MdLauglin, M.W. (1977). Federal programs supporting educational change, vol. VII: Factors affecting implementation and continuation. Santa Monica, CA: RAND. Berman, P. & McLauglin, M.W. (1978). Federal programs supporting educational change, vol. VIII: Implementing and sustaining innovations. Santa Monica, CA: RAND. Bernd, M. (1992). Shared decision making requires effective instructional leadership. NASSP Bulletin, 76(546), 64-69. Beyer, J.M., & Trice, H.M. (1978). Implementing change: Alcoholism programs in work organizations. New York: Free Press. Block, P. (1987). The empowered manager. San Francisco: Jossey-Bass. Bodilly, S.J. (1998). Lessons from New American Schools ’ scale-up phase: Prospects for bringing designs to multiples schools. Santa Monica, CA: RAND (MR-1777-NAS). Bolman, L.G., & Deal, T.E. (1993). The path to school leadership. Newbury Park, CA: Corwin Press. Botvin, G.J., Baker, E., Dusenbury, L., Botvin, E.M., & Diaz, T. (1995) Long-term follow-up results of a randomized drug abuse prevention trial in a white middle-class population. JAMA, 273, 1106-1112. Botvin, G. I., Eng, A., & Williams, C.L. (1980). Preventing the onset of cigarette smoking through life skills training. Preventive M edicine, 9(1), 135-43. Boyd, V. (1992). School context: Bridge or barrier to change? (Office of Educational Research and Improvement Contract No. RP 1002003). Austin, TX: Southwest Educational Development Laboratory. Brink, S.G., Basen-Engquist, K.M., O’Hara-Tompkins, N.M., Parcel, G.S., Gottlieb, N.H., & Lovato, C.Y. (1995). Diffusion of an effective tobacco prevention program. Part 1: Evaluation of the dissemination phase. H ealth Education Research, 10(3), 283-295. Bruvold, W.H. (1993). A meta-analysis of adolescent smoking prevention programs. American Journal o f Public Health, 83(6), 872-80. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 211 California Department of Education. (1998). California Board o f Education data set, 1998. Available at: http://www. cde. ca. gov/demographics/files. Accessed September 16. 1998. California Department of Education (2000). Getting results. California action guide to tobacco use prevention education. Sacramento, CA: California Department of Education, Healthy Kids Program Office. California Department of Health Services (1998). A m odelfor change: The California experience in tobacco control. Sacramento, CA: California Department of Health Services. Centers for Disease Control and Prevention (1994). Guidelines for school health programs to prevent tobacco use and addiction. M orbidity and M ortality Weekly Reports (Recommendations and Reports), 43 (RR-2), 1-18. Centers for Disease Control and Prevention (1998). Youth risk behavior surveillance— United States, 1997. M orbidity and M ortality Weekly Reports, 47 (SS-3), 1-97. Centers for Disease Control and Prevention. (1999). Tobacco Use in the United States. Available at: http://www.cde.gov/tobacco/tobus_us.fatml. Accessed October 17.1999. Center for the Study and Prevention of Violence. (2002). Blueprints model programs overview. Available at http://www.colorado.edu/cspv/blueprints/model/overview. fatm l. Accessed Mav 18. 2002. Center for Substance Abuse and Prevention. (2001). CSAP’ sprevention portal: M odel programs. Available at http://www.samhsa.gov/centers/csap/modelprograms/programs cfm Accessed on May 18. 2001. Conley, D.T. (1997). Roadmap to restructuring: Charting the course o f change in American education. Eugene, OR: ERIC Clearinghouse on Educational Management, University of Oregon. Consortium on Chicago School Research. (2000). Public use data set user’ s manual. June 2000. Available at: www.consortium-chicago.org. Accessed on November 14, 2002. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 212 Consortium on Chicago School Research. (2002). Educational technology: A vailability and m e in Chicago’ s Public Schools. Available at: www.consortium-chicago.org. Accessed on November 14, 2002. Cooke, M., Mattick, R.P., & Campbell, E. (1999). The dissemination of a smoking cessation program to 23 antenatal clinics: the predictors of initial program adoption by managers. Australian & New Zealand Journal o f Public Health, 23(1), 99-103. Corbett, H. D., Dawson, I. & Firestone, W. (1984). School context and school change: Im plications fo r effective planning, Philadelphia, PA: Research for Better Schools. Cunningham, W.G., & Gresso, D.W. (1993). Cultural leadership. Boston: Allyn & Bacon. Curran, P.J. (2000). A latent curve framework for studying developmental trajectories of adolescent substance use. In J. Rose, L. Chassin, C. Presson, & J. Sherman (Eds.), M ultivariate Applications in Substance Use Research, (pp. 1-42). Hillsdale, NJ: Erlbaum. Deal, T. E., Meyer, J. W. & Scott, W. R. (1975). Organizational influences on educational innovation. In J. V. Baldridge & T. E. Deal (Eds.) M anaging Change in Educational Organizations. Berkeley, CA: McCutchan Publishing Corporation, 109-132. Donaldson, Jr., G.A. (2001). Cultivating leadership in schools: Connecting people, purpose, and practice. New York: Teachers College Press. Drug Strategies, Inc. (1999). M aking the grade: A guide to school drug prevention programs. Washington, D C.: Drug Strategies, Inc. Eicholz, G., & Rogers, E.M. (1964). Resistance to the adoption of audio-visual aids by elementary school teachers: Contrasts and similarities to agriculture innovation. In M. Miles (Ed.), Innovation in Education (pp.299-316). New York: Teachers College Press, Columbia. Elias, M.J. (1997). Reinterpreting dissemination of prevention programs as widespread implementation with effectiveness of fidelity. In R.P. Weissberg & T.P. Gullotta (Eds.), Healthy Children 2010: Establishing preventive services. Issues in children’ s andfam ilies ’ lives, Vol. 9 (pp.253-289). Thousand Oaks, CA: Sage. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 213 Ellickson, P.L., Bell, R.M., & Harrison, E.R. (1993). Changing adolescent propensities to use drugs: results from Project ALERT. Health Education Quarterly, 20, 227-242. Ennett, S.T., Ringwalt, C.L., Thorne, J., Rohrbach, L.A., Vincus, A. A., Simons- Rudolph, A., & Jones, S.M. (2003). A comparison of current practice in school-based substance use prevention programs with meta-analysis findings. Prevention Science, 4(1), 1-14. Evans, R.I. (1976). Smoking in children: Developing a social psychological strategy of deterrence. Preventive M edicine, 5(1), 122-127. Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior: An introduction to theory and research, Addison-Wesley. Flay, B.R. (1985a). Psychosocial approaches to smoking prevention: a review of findings. Health Psychology, 4(5), 449-88. Flay, B.R., Ryan, K.B., Best, A., Brown, K.S., Kersell, M.W., d’Avemas, J R., & Zanna, M.P. (1985b). Are social-psychological smoking prevention programs effective? The Waterloo Study. Journal o f Behavioral M edicine, 8, 37-59. Flore, M.C., Novotny, T.E., Pierce, J.P., et al. (1989). Trends in cigarette smoking in the United States: The changing influence of gender and race. JAMA, 261, 49-55. Fullan M.G. (1991). The new meaning o f educational change. New York: Teachers College Press. Fullan, M.G. (1992). Successful school improvement: The implementation perspective and beyond Philadelphia, PA: Open University Press. Fullan, M.G. (1993). Innovation, reform, and restructuring strategies. In G. Cawelti (Ed.), Challenges and achievements o f American Education. Alexandria, VA: Association for Supervision and Curriculum Development. Fullan M.G. (2001). The new meaning o f educational change, 3r d edition. New York: Teachers College Press. Fullan, M.G., & Hargreaves, A. (1998). W hat’ s worth fig h tin g fo r out there. New York: Teachers College Press. Fullan, M.G., & Miles, M. (1992). Getting reform right: What works and what doesn’t. Phi D elta Kappa, 73(10), 745-752. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 214 Gingiss, P.L., Gottlieb, N.H., & Brink, S.G. (1994a). Increasing teacher receptivity toward use of tobacco prevention education programs. Journal o f Drug Education, 24(2), 163- 176. Gingiss, P.L., Gottlieb, N.H., & Brink, S.G. (1994b). Measuring cognitive characteristics associated with adoption and implementation of health innovations in schools. American Journal o f Health Promotion, 8, 294-301. Glickman, C. (1993). Renewing America ’ $ schools: A guide fo r school-based action. San Francisco: Jossey-Bass. Gold, R.S., Parcel, G.S., Wallberg, H.J., Luepker, R.V., Portnoy, B., & Stone, E.J. (1991). Summary and conclusions of the THTM evaluation: The expert work group perspective. Journal o f School Health, 61, 39-42. Goodman, R.M., Smith, D.W., Dawson, L., & Steckler, A. (1991). Recruiting schools districts into a dissemination study. Health Education Research: Theory and Practice, 6(3), 373-385. Goodman, R.M., Steckler, A., & Kegler, M.C. (1997). Mobilizing organizations for health enhancement: Theories of organizational change. In K. Glanz, F.M. Lewis, & B .K. Rimer (Eds.), H ealth behavior and health education: Therory, research, and practice (2n d ed., pp. 287-312). San Francisco, CA: Jossey- Bass. Goodman, R.M., Tenney, M., Smith, D.W., & Steckler, A. (1992). The adoption process for health curriculum innovations in schools: A case study. Journal o f H ealth Education, 23, 215-220. Graham, J.W., Flay, B.R., Johnson, C.A., Hansen, W.B., Grossman, L. Sobel, J.L. (1984). Reliability of self-report measures of drug use in prevention research: Evaluation of the Project SMART questionnaire via the test-retest reliability matrix. Journal o f Drug Education, 14(2), 175-193. Greenberg, M.T., Domitrovich, D., Graczyk, P. A., & Zins, J.E. (2000). The study o f implementation in school-based prevention research: Theory, research, and practice. Report to the Center for Mental Health Services (SAMHSA). Available at: http://www.prevention.osu.edu/. Accessed on August 15, 2002. Guskey, T.R. (1988). Teacher efficacy, self-concept, and attitudes toward the implementation of instructional Innovation. Teaching and Teacher Education, 1, 63-69. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 215 Hall, G.E., & Hord, S.M. (1987). Change in schools: Facilitating the process. Albany: State University of New York Press. Hall, G.E., & Hord, S.M., & Griffin, T.H. (1980). Implementation a t the school building level: The development and analysis o f nine mini-case studies (Report No. 3098). Austin: The University of Texas at Austin, Research and Development Center for Teacher Education (ERIC Document Reproduction Service No. ED 207 170). Hallfors, D ., & Godette, D. (2002). Will the “Principles of Effectiveness” improve prevention practice? Early findings from a diffusion study. Health Education Research, 17(4), 461-70. Hansen, W.B. (1992). School-based substance abuse prevention: a review of state of the art in curriculum, 1980-1990. H ealth Education Research: Theory and Practice, 7(3), 403-30. Hansen, W.B., Malotte, C.K„ & Fielding, J.E. (1985). The bogus pipeline revised: The use of the threat of detection as a means of increasing self-reports of tobacco use. Journal o f Applied Psychology, 70, 789-792. Hord, S.M. (1992). Facilitative leadership: The imperative fo r change. (Office of Educational Research and Improvement Contract No. RP 91002003). Austin, TX: Southwest Educational Development Laboratory. Huberman, M., & Miles, M. (1984). Innovation up close. New York: Plenum. Ialongo, N.S., Werthamer, L., Kellam, S.G., Brown, C H., Wang, S. & Lin, Y. (1999). Proximal impact of two first-grade preventive interventions on the early risk behaviors for later substance abuse, depression, and antisocial behavior. American Journal o f Community Psychology, 27(5), 599-641. Independent Evaluation Consortium (1998). F inal report o f the Independent Evaluation o f California Tobacco Control Program: Wave I data, 1996- 1997. Rockville, MD: The Gallup Organization. Independent Evaluation Consortium (2001). Interim report: Independent Evaluation o f the California Tobacco C ontrol Prevention and Education Program. Wave 2 data, 1998. Wave I and Wave 2 D ata Comparisons, 1996-1998. Sacramento, CA: California Department of Health Services. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 1 6 Independent Evaluation Consortium (2003). Final report. Independent Evaluation o f the California Tobacco Control Prevention and Education Program: Waves 1, 2, and 3 (1996-2000). Sacramento, CA: California Department ofHealth Services. Johnston LD, O'Malley PM, Bachman JG. (2002) The M onitoring the Future National Survey results on adolescent drug use: overview o f key findings, 2001. Bethesda, MD: National Institute on Drug Abuse, 2002 [NIH Publication No. 02-5105]. Joyce, B.R., & Murphy, C. (1990). Epilogue: the curious complexities of cultural change. In B.R. Joyce (Ed.), Changing school culture through sta ff development. Alexandria, VA: Association for Supervision and Curriculum Development. Kam, C-M., Greenberg, M.T., & Walls, C.T. (2003). Examining the role of implementation quality in school-based prevention using the PATHS curriculum. Prevention Science, 4(1), 55-63. Kellam, S.G., Rebok, G.W., Mongo, N., & Mayer, L.S. (1994). The course and malleability of aggressive behavior from early first grade into middle school: Results of a developmental epidemioiogically-based preventive trial. Journal o f C hild Psychology and Psychiatry, 35(2), 259-281. Kellam, S.G., & Van Horn, Y.V. (1997). Life course development, community epidemiology, and preventive trials: A scientific structure for prevention research. American Journal o f Community Psychology^ 25(2), 177-188. Lee, V.E., & Smith, J.B. (1995). Effects of high school restructuring and size on early gains in achievement and engagement. Sociology o f Education, 68(4), 241-270. Lee, V.E., & Smith, J.B. (1997). High school size: Which works best and for whom? Educational Evaluation and Policy Analysis, 19(3), 205-227. Likert, R. (1967). The Human Organization. New York: McGraw-Hill. Little, J.W. (1990). The persistence of privacy: autonomy and initiative in teachers’ professional relations, Teachers College Record, 91(4), 509-36. Louis, K.S., & Kruse, S.D., & Associates. (1995). Professionalism and community: Perspectives on reform ing urban schools. Thousand Oaks, CA: Corwin Press. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 217 Louis, K.S., & Miles, M.B. (1990). Improving the urban high school: What works and why. New York: Teachers College Press and (1991) London: Cassell. McCormick, L.K., Steckler, A., & McLeroy, K.R. (1995). Diffusion of innovation in schools: A study of adoption and implementation of school-based tobacco prevention curricula. American Journal o f Health Promotion, 9, 210-219. Miles, M.B. (1964). Educational innovation: The nature of the problem. In M.B. Miles (Ed.), Innovation in Education (pp. 19-22). New York: Teachers College, Columbia University, 1964. Miles, M.B. (1983). Unraveling the mystery of institutionalization. Educational Leadership, 41(3), 14-19. Milstein, M.M. (1993). Restructuring schools: Doing it right. Newbury Park, CA. Corwin. Muthen, B.O., & Curran, PJ. (1997). General longitudinal modeling of individual differences in experimental designs: A latent variable framework for analysis and power estimation. Psychological M ethods, 2, 371-402. Nanus, B. (1992). Visionary leadership: Creating a compelling sense o f direction fo r your organization. San Francisco: Jossey-Bass. National Education Association. (2003). Keys to excellence fo r your schools. Available at: http://www.keysonline.org/guide/index.cfm. Accessed on March 31. 2003. National Cancer Institute (1987). Integrating tobacco education into the school system: Request fo r research grant proposals. Bethesda, MD: Author. National Center for Education Statistics. (2001). D igest o f Education Statistics. Elementary and Secondary Education. Available at http://nces. ed. gov/pubs2001/digest/dt 1 OO.html. Accessed December 14. 2001. National Institute on Drug Abuse (1997). Preventing drug use among children and adolescents: A research-based guide. Rockville, MD: National Institute on Drug Abuse. Nias, J., Southworth, G., & Yeomans, R. (1989). S ta ff relationships in the prim ary school: A study o f organizational cultures. London: Cassell. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 218 Owens, R.G. (1970). Organizational behavior in schools. Englewood Cliffs, N.J.: Prentice-Hall. Palestini, R.H. (2000). Ten steps to educational reform: M aking change happen. Lanham, Maryland: The Scarecrow Press, Inc. Parcel, G.S., Eriksen, M.P., Lovato, C.Y., Gottlieb, N.H., Brink, S.G., & Green, L.W. (1989). The diffusion of school-based tobaeco-use prevention programs: Project description and baseline data. Health Education Research, 4, 111-124. Parcel, G.S., O’Hara-Tompkins, N.M., Harrist, R.,B., Basen-Engquist, K.M., McCormink, L.K., Gottlieb, N.H., & Eriksen, M.P. (1995). Diffusion of an effective tobacco prevention program. Part 2: Evaluation of the adoption phase. Health Education Research, 10,297-307. Parcel, G.S., Taylor, W.C., Brink, S.G., Gottlieb, N., Engquist, K., O’Hara, N.M., & Eriksen, M.P. (1989). Translating theory into practice: Intervention strategies for the diffusion of a health promotion innovation. Family & Community Health, 12(3), 1-13. Paulussen, T., Kok, G., Schaalma, H P. (1994). Antecedents to adoption of classroom-based AIDS education in secondary schools, Health Education Research, 9(4), 485-496. Paulussen, T., Kok, G., Schaalma, H P, & Parcel, G.S. (1995). Diffusion of AIDS curricula among Dutch secondary school teachers. Health Education Quarterly, 22(2), 227-243. Pentz, M.A. (2003). Evidence-based prevention: Characteristics, impact, and future direction. Journal o f Psychoactive Drugs, Supplement 1, 143-152. Pentz, M.A., Dwyer, J.H., MacKinnon, D.P., Flay, B.R., Hansen, W.B., Wang, E.Y., & Johnson, C.A. (1989). A multicommunity trial for primary prevention of adolescent drug abuse. JAMA, 261, 3259-3266. Pentz, M.A., Trebow, E.A., Hansen, W.B., MacKinnon, D.P., Dwyer, J.H., Johnson, C.A., Flay, B.R., Daniels, S., & Cormack, C. Effects of program implementation on adolescent drug use behavior. The Midwestern Prevention Project (MPP). Evaluation Review, 14(2), 264-289. Perry C.L., Kelder S.H. (1992). Models for effective prevention. Journal o f Adolescent Health, 13(5), 353-363. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 219 Ringwalt C.L., Ennett S., Vincus A., Thome J., Rohrbach L.A., & Simons-Rudolph A. (2002). The prevalence of effective substance use prevention curricula in U.S. middle schools. Prevention Science, 3(4), 257-265. Rohrbach, L.A., de Calice, S., & Unger, J.B. (2001; December). The effectiveness o f tobacco prevention programs in California schools. Presented at the Tobacco-Related Disease Research Program annual meeting, Los Angeles, CA. Rohrbach, L.A., de Calice, S., & Unger, J.B. (2002; May). Implementation o f evidence-based practices fo r tobacco prevention in schools: Effects on outcomes. Presented at the Annual Meeting of the Society for Prevention Research, Seattle, WA Rohrbach, L.A., Dent, C., Johnson, C.A., Unger, J., & Gunning, G. (1998). Evaluation of the school Tobacco Use Prevention Education Program. In: Independent Evaluation Consortium, F inal Report. Independent Evaluation o f the California Tobacco Control Program. Wave 1 Data, 1996-1997. Rockville, MD.: The Gallup Organization, pp. 78-98. Rohrbach, L.A., D'Onofrio, C.N., Backer, T.E., & Montgomery, S B. (1996). Diffusion of school-based substance abuse prevention programs. American Behavioral Scientist, 39(7), 919-934. Rohrbach L.A., Graham, J.W., & Hansen, W.B. (1993). Diffusion of a school-based substance abuse prevention program: predictors of program implementation. Preventive M edicine, 22, 237-260. Rohrbach L.A., Howard-Pitney, B., Unger, J.B., Dent, C.W., Howard, K.A, Cruz, T.B., Ribisi, K.M., Norman, G.J., Fishbein, H., & Johnson, C.A. (2002). Independent Evaluation of the California Tobacco Control Program: Relationships between program exposure and outcomes, 1996-1998. American Journal o f Public Health, 92(6), 975-983. Rohrbach L.A., Ringwalt C.L., Ennett S. T., & Vincus A.A. (In press). Factors affecting adoption of evidence-based substance use prevention curricula in U.S. school districts. Rohrbach, L.A. & Skara, S. (2003; June). Factors affecting implementation o f evidence-based tobacco prevention practices in schools. Presented at the Annual Meeting of the Society for Prevention Research, Washington, DC. Rohrbach, L.A, & Skara, S. (In preparation). Factors affecting implementation of evidence-based tobacco prevention practices in schools. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 220 Rohrbach, L.A., Skara, S., Unger, J., & Dent, C. (2001). The school-based Tobacco Use Prevention Education (TUPE) program. In: Independent Evaluation Consortium, Final Report. Independent Evaluation o f the California Tobacco Control Program. Wave 3 Data, 2001. Rockville, MD.: The Gallup Organization. Rohrbach, L.A., Unger, J.B., & Skara, S. (2002; November). Implementation o f tobacco prevention programs in California schools. Presented at the National Conference on Tobacco or Health, San Francisco, CA. Rogers, E.M. (1983). D iffusion o f innovations. New York, NY: The Free Press. Rogers, E.M. (1995). D iffusion o f innovations (4th Edition). New York, NY: The Free Press. Rushton, J.P., Brainerd, C.J., & Pressley, M. (1983). Behavioral development and construct validity: The principle of aggregation. Psychological Bulletin, 94, 18-38. Sammons, P., Hillman, I. & Mortimore, P. (1995). Key characteristics o f effective schools: A review o f school effectiveness research. London: Office for Standards in Education. Senge, P. (1999). The practice of innovation. In F. Hesselbein & P. Cohen (Eds.), Leader to leader (pp.57-68). San Francisco: Jossey-Bass. Sergiovanni, T. (1992). M oral leadership: G etting to the heart o f school improvement. San Francisco: Jossey-Bass. Sergiovanni, T. (1996). Leadership fo r the schoolhouse. San Francisco: Jossey-Bass. Sizer, T. (1986). H orace’ s compromise: The dilemma o f the American high school. Boston: Houghton Mifflin. Skara S., & Sussman S. (2003). A review of 25 long-term adolescent tobacco and other drug use prevention program evaluations. Preventive M edicine, 37, 451-474. Smith, D.W., McCormick, L.K., Steckler, A.B., & McLeroy, K.R. (1993). Teachers’ use of health curricula: Implementation of Growing Healthy, Project SMART, and the Teenage Health Teaching Modules. Journal o f School Health, 63, 349-354. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 221 Smith, D.W., Steckler, A., McCormick, L.K., & McLeroy, K.R. (1995). Lessons learned about disseminating health curricula to schools. Journal o f Health Education, 26(1), 37-43. Stacy, A.W., Flay, B.R., Sussman, S., Brown, K.S., Santi, S., & Best, J.A. (1990). Validity of alternative self-report indices of smoking among adolescents. Psychological Association, 2, 442-446. Steckler, A, Goodman, R.M., McLeroy, K.R., Davis, S., & Koch, G. (1992). Measuring the diffusion of innovative health programs. American Journal o f H ealth Promotion, 6,214-224. Stoll, L. & Fink D. (1996). Changing our schools: Linking school effectiveness and school improvement. Philadelphia, PA. Open University Press. Sussman, S., Dent, C.W., Burton, D., Stacy, A.W., & Flay, B.R. (1995). Developing school-based tobacco use prevention and cessation programs. Thousand Oaks, CA: Sage Publications. Sussman, S., Sun, P., McCuller, W.J., & Dent, C.W. (2003). Project Towards No Drug Abuse: Two-year outcomes of a trial that compares health educator delivery to self-instruction. Preventive M edicine, 37, 155-162. Tobler, N.S., Roona, M R , Ochshom, P., Marshall, D.G., Streke, A.V., & Stackpole, K.M. (2000). School-based adolescent drag prevention programs; 1998 meta analysis. Journal o f Primary Prevention, 20(4), 275-336. Tobler, N.S., & Stratton, H.H. (1997). Effectiveness of school-based drag prevention programs: A meta-analysis of the research. Journal o f Primary Prevention, 18(1), 71-128. U.S. Department of Education. (2003). Safe and Drug-Free Schools and Communities Act. State Grants. Guidance o f state and local programs. Available at: http://www.ed. go v/programs/dvpformula/guidance. doc. Accessed on June 25. 2003. U.S. Department of Education, Safe and Drug-Free Schools Program. (2001). Safe and drug-free schools and communities program model programs. Available at: http;//www.ed.gov/offices/OESE/SPFS/programs.html. Accessed September 18. 2001. U.S. Department of Education, Safe and Drug-Free Schools Program. (1998). Notice of final principles of effectiveness. Federal Register, 63, 29901-29906. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 222 U.S. Department of Health and Human Services. (1989). Reducing the health consequences o f smoking: 25 years o f progress. A report o f the Surgeon General. U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control, Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health. DHHS Publication No. (CDC) 89-8411. U.S. Department of Health and Human Services. (2000). Reducing tobacco use: A . Report o f the Surgeon General. Atlanta, GA: US Department of Health and Human Services, Public Health Service, CDC, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health. Vaill, P. (1998). Spirited leading and learning: Process wisdom fo r a new age. San Francisco: Jossey-Bass. Yin, R.K., & White, J.L. (1984). M icrocomputer implementation in schools. Washington, D C.: COSMOS Corporation. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Contributions of interpersonal communication within a mass media environment: Exploring the role of social influence in a statewide tobacco control media campaign

PDF

Implicit cognition and dissociative experiences as predictors of adolescent substance use

PDF

Adolescent exposure to alcohol advertising: A prospective extension of Strickland's model

PDF

Cigarette smoking, metabolic syndrome, and subclinical atherosclerosis

PDF

Correlates of smoking behaviour among Egyptian adolescents

PDF

Aggression and victimization as predictors of drug use in early adolescence

PDF

Correlates of compliance with mammography screening guidelines among low-income Latinas: An exploratory study

PDF

Understanding visceral fat, subcutaneous abdominal fat, and total body fat in children: Longitudinal changes and relations to metabolic risks of type 2 diabetes

PDF

Emotional intelligence and smoking risk factors in early adolescents

PDF

Patterns and risk factors for adolescent smoking progression

PDF

The influence of future orientation on the relationship between condom use and violence among homeless/runaway youth: Implications for the prevention of HIV and violence

PDF

The relationship between high school anti-smoking policies and adolescent cigarette use

PDF

Self-help and problem-proneness as predictors of weapon carrying among adolescents

PDF

Stressful life events, social support, and smoking behaviors in United States And Chinese adolescents: A cross-cultural comparison

PDF

Alternatives to condom use and not sharing needles among young injection drug users, their sex partners, and the social network members

PDF

Enabling clinically based knowledge discovery in pharmacy claims data: An application in bioinformatics

PDF

Effects of a brief safer sex intervention on provider counseling behaviors, attitudes, and the provider -patient relationship

PDF

Diet, body size and menarche in a multiethnic cohort

PDF

Self -care diabetes behaviors among Latinos with type 2 diabetes

PDF

The effect of potential stressors on early atherosclerosis: The Los Angeles Atherosclerosis Study

Asset Metadata

Creator Skara, Silvana Nicolle (author)

Core Title Adoption and implementation of evidence-based tobacco use prevention curricula and programming

School Graduate School

Degree Doctor of Philosophy

Degree Program Preventive Medicine - Health Behavior Research

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag health sciences, public health,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c16-418800

Unique identifier UC11335743

Identifier 3145288.pdf (filename),usctheses-c16-418800 (legacy record id)

Legacy Identifier 3145288.pdf

Dmrecord 418800

Document Type Dissertation

Rights Skara, Silvana Nicolle

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA