xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001956412
007 cr mnu|||uuuuu
008 080910s2007 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002204
Hellman, Deborah W.
Implementing differentiated instruction in urban, Title I schools:
b effects of facilitated support groups and program fidelity on student achievement
h [electronic resource] /
by Deborah W. Hellman.
[Tampa, Fla.] :
University of South Florida,
ABSTRACT: This study presents the results of a mixed methodology study and pilot that investigated the effects of facilitated teacher support groups and differentiated instruction on student achievement at two urban, Title I middle schools. Both general education and students with special needs being served in a collaborative co-taught setting were included in the study. Implications for research to practice and effective inclusive strategies were addressed and the field-tested Differentiated Instruction: Fidelity Implementation Tool (DI: FIT) used to assess teacher fidelity is included. During the first year, the principal investigator developed and field-tested the DI: FIT observation tool, field-tested a facilitated support group, and collected student achievement data to determine the feasibility of the implementation of differentiated instruction research design. During this second year, two matched urban, Title I middle schools were purposively selected to serve as research sites. At each of the two school sites, 13 to 15 teachers were selected to participate in the treatment group and 13 to 14 teachers in the control group. The teachers selected were balanced among the three grade levels within each school. A triangulation of data from monthly, 2-hour, facilitated support group meeting minutes (group's perspective), teacher implementation logs (individual's perspective), and differentiated instruction observations (observer's perspective) were utilized to determine the impact of differentiated instruction on teacher implementation fidelity. Finally, the effects of teacher use of differentiated instruction with fidelity on the reading and mathematics achievement scores of approximately 906 students (461 in the treatment group and 445 in the control group) that were part of the combined sample population at the two school sites were assessed using ANOVA procedures. Cohen's (1977) f effect sizes are included.
Dissertation (Ph.D.)--University of South Florida, 2007.
Includes bibliographical references.
Text (Electronic dissertation) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 192 pages.
Adviser: Albert J. Duchnowski, Ph.D.
Students with disabilities.
x Special Education
t USF Electronic Theses and Dissertations.
Implementing Differentiated Instruction in Urban, Title I Schools: Effects of Facilitated Support Groups and Program Fidelity on Student Achievement by Deborah W. Hellman A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Exceptional Student Education College of Education University of South Florida Major Professor: Albert J. Duchnowski, Ph.D. David H. Allsopp, Ph.D. Elizabeth M. Doone, Ph.D. William F. Benjamin, Ph.D. D ate of Approval: October 19, 2007 Keywords: implementation, inclusion, middle school, professional development, and students with disabilities Copyright 2007, Deborah Hellman
Dedication I would like to dedicate this dissertation to my husba nd, Greg Hellman. Our love has produced 23 years of a magnificent marriage and two wonderful children. I want to thank you for being my biggest supporter. You are always there for me and you made sure everything at home was taken care of so I could focus o n furthering my education. Without you, I could not have completed this arduous task. Words cannot express how much your love and support has meant to me on this journey and in my life. I am who I am because of you.
Acknowledgements I would like to thank my committee chair and members, Dr. Albert Duchnowski, Dr. David Allsopp, Dr. Elizabeth Doone, and Dr. William Benjamin. Your unwavering support and guidance throughout this process has been invaluable as you helped me conceptualize and implement thi s study. I would especially like to thank my chair, Dr. Albert Duchnowski, for your assistance, direction, encouragement, and hours of conference time. Your knowledge of evidence based practice, fidelity studies, and general research knowledge has been ind ispensable to me. Others who deserve special thanks are Charles Dixon, Dr. Joseph Brown, Jennifer Apgar, and Dianne Williams and all of the teachers at the school sites for making this study a success. Without your cooperation and your willingness to part icipate, I could not have completed my study. I also want to thank my USF family for helping me whenever I asked. Thank you to all of the Exceptional Student Education instructors and office staff, especially Dr. James Paul and Jaye Berkowitz. Finally, ye t importantly, this acknowledgement would not be complete if I did not thank my very special cohort family: Teri Crace, Sandra May, Sharon Ray, and Anne Townsend. The two years we spent together in classes as scholars would not have been the same without t he mutual support we all shared.
i Table of Contents List of Tables ................................ ................................ ................................ .................... vii List of Figures ................................ ................................ ................................ ................... xii Abstract ................................ ................................ ................................ .............................. xv Chapter One: Introduction ................................ ................................ ................................ ... 1 Statement of the Problem ................................ ................................ ......................... 1 Theoretical/ Conceptual Framework ................................ ................................ ........ 5 Purpose ................................ ................................ ................................ ................... 12 Research Questions ................................ ................................ ................................ 12 Significance of St udy ................................ ................................ ............................. 13 Operational Definitions ................................ ................................ .......................... 14 Curriculum ................................ ................................ ................................ 14 Differentiated Instruction ................................ ................................ ........... 14 Facilitated Support Group ................................ ................................ .......... 15 Flexible Grouping ................................ ................................ ...................... 15 Implementation Fidelity ................................ ................................ ............. 15 Inclusion ................................ ................................ ................................ ..... 15 Mathematics Achievement ................................ ................................ ......... 15 Middle School ................................ ................................ ............................ 16
ii Mild to Moderate Disabili ties ................................ ................................ .... 16 Professional Development ................................ ................................ ......... 16 Reading Achievement ................................ ................................ ................ 16 Standards Based Test ................................ ................................ ................. 16 Tiered Assignments ................................ ................................ ................... 17 Title I Schools ................................ ................................ ............................ 17 Title II Schools ................................ ................................ ........................... 17 Delimitations ................................ ................................ ................................ .......... 17 Limitations ................................ ................................ ................................ ............. 18 Chapter Two: Review of the Literature ................................ ................................ ............. 20 Overview ................................ ................................ ................................ ................ 20 Literature Search ................................ ................................ ................................ .... 21 Professional Development ................................ ................................ ..................... 22 Differentiated Instruction ................................ ................................ ....................... 25 Facilitated Support Groups/ Teacher Study Groups ................................ .............. 38 Instructional Fidelity ................................ ................................ .............................. 50 Summary ................................ ................................ ................................ ................ 56 Chapter Three: Method ................................ ................................ ................................ ...... 57 Overview ................................ ................................ ................................ ................ 57 Pilot Data ................................ ................................ ................................ ............... 57 Overview of Pilot ................................ ................................ ....................... 57 DI: FIT Observat ion Tool ................................ ................................ .......... 58 Student Achievement ................................ ................................ ................. 58
iii Student Reading Data ................................ ................................ ................ 59 Student Mathematics Data ................................ ................................ ......... 59 Summary of Pilot ................................ ................................ ....................... 63 Research Design ................................ ................................ ................................ ..... 63 Sample Selection and Assignment ................................ ............................. 65 Facilitated Support Group Model ................................ .............................. 66 rnals ................................ 68 Teacher Observations ................................ ................................ ................. 69 Population and Sample ................................ ................................ .......................... 70 Independent Variables ................................ ................................ ........................... 72 Experimental Group, Level O ne: Treatment Group ................................ .. 72 Experimental G roup, L evel T wo: Control Group ................................ ...... 73 School Site ................................ ................................ ................................ 73 Grade Level ................................ ................................ ................................ 73 Dependent Variables ................................ ................................ .............................. 73 DI: FIT Fidelity Observation Score ................................ ........................ 73 FCAT Mathematics Predictor Test Scores ................................ ................ 74 Stanford Diagnostic Reading Test Scores ................................ .................. 74 Instruments/ Measurement Tools ................................ ................................ ........... 74 Differentiated Instruction: Fidelity Implementation Tools (DI: FIT) ........ 74 FCAT Mathematics Predict or Tests ................................ ........................... 77 Stanford Diagnostic Reading Test Fourth Edition (SDRT 4) ................. 80 Data Collection Procedures ................................ ................................ .................... 80
iv Facilitated Support Groups and Focus Group ................................ ............ 80 Teacher Journals ................................ ................................ ........................ 81 DI: FIT Teacher Observations ................................ ................................ ... 82 FCAT Mathematics Predictor Tests Form B a nd C and Stanford Diagnostic Reading Test Fourth Edition ................................ ..... 82 Confidentiality ................................ ................................ ................................ ....... 83 Data Analysis ................................ ................................ ................................ ......... 84 Question 1 ................................ ................................ ................................ .. 84 Question 2 ................................ ................................ ................................ .. 85 Question 3 ................................ ................................ ................................ .. 85 Question 4 ................................ ................................ ................................ .. 86 Chapter Four: Results ................................ ................................ ................................ ........ 90 Overview ................................ ................................ ................................ ................ 90 Demographics of Participants ................................ ................................ ................ 91 Student De mogra phics ................................ ................................ ............... 91 Teacher D emographics ................................ ................................ .............. 95 Student Achievement Data Analyses Overview ................................ .................... 97 Reading Achievement Analyses ................................ ................................ ............ 97 Within Site Comparisons at Time 1 ................................ ........................... 97 Between School Comparisons at Time 1 ................................ ................. 102 Time 2 Analyses ................................ ................................ ...................... 102 Change Score Analyses ................................ ................................ ............ 105 Pooled Reading Achievement Data ................................ ......................... 111
v Mathematics Achievement Analyses ................................ ................................ ... 119 Within Site Comparisons at Time 1 ................................ ......................... 119 Between School Comparisons at Time 1 ................................ ................. 123 Combined Mathematics Data at Time 1 ................................ .................. 124 Time 2 Analyses ................................ ................................ ...................... 126 Change Score Analyses ................................ ................................ ............ 130 Teacher Fidel ity Observation Data Analyses ................................ ...................... 136 Relationship Between Fidelity and Achievement ................................ ................ 138 Support Group Analyses ................................ ................................ ...................... 141 Summary ................................ ................................ ................................ .............. 145 Chapter Five: Discussion ................................ ................................ ................................ 147 Purpose ................................ ................................ ................................ ................. 147 Method ................................ ................................ ................................ ................. 147 Results ................................ ................................ ................................ .................. 149 Limitations ................................ ................................ ................................ ........... 153 Significance ................................ ................................ ................................ .......... 153 Implications for Practice ................................ ................................ ...................... 155 Implications for Research ................................ ................................ .................... 157 References ................................ ................................ ................................ ........................ 160 Appendices ................................ ................................ ................................ ....................... 170 Appendix A: Strategies for Differentiating Instruction ................................ ....... 171 Appendix B: Comparison of the Traditional vs. Differentiated Classroom ........ 172
vi Appendix C: Qualities of a Supportive Classroom Envi ronment for Differentiation ................................ ................................ ......................... 173 Appendix D: Differentiated Instruction: Fidelity Implementation Tool (DI: FIT) ................................ ................................ ................................ ... 174 Library ................................ ................................ ................................ ...... 176 Appendix F: Facilitated Support Group Feedback Fo rm ................................ ..... 178 Appendix G: The Differentiate d Classroom Observation Form .......................... 179 Appendix H: Teacher/ Peer R eflection on Differentiation ................................ .. 180 Appendix I: DI Support Group Feedback ................................ ............................ 183 Appendix J: Informed Consen t to Participate in Research ................................ .. 185 Appendix K: District In service Evaluation Summary: School A ....................... 189 Appendix L: District In service Evaluation Summary: School B ....................... 191 About the Author ................................ ................................ ................................ ... End Page
vii List of Tables Table 1. Descriptive Statistics of FCAT Reading Developmental Scale Change Scores (N = 353) ................................ .............................. 60 Table 2. Analysis of Variance Summary Table for FCAT Reading Developmental Scale Change Scores (N = 353) ............................ 61 Table 3. Descriptive Statistics of FCAT Mathematics Developmental Scale Change Scores (N = 353) ................................ ............................... 62 Table 4. Analysis of Variance Summary Tabl e for FCAT Mathematics Developmental Scale Change Scores (N = 353) ............................ 63 Table 5. Descriptive Statistics of the Student Population who Participated in the State Standardized Assessments at the Two School Sites During the 2006 2007 School Year (N = 1026 ) .................... 72 Table 6. Table of Research Questions, Data Collected, and Analyses Conducted ................................ ................................ ...................... 87 Table 7. Demographic Characteristics of Student Participants: Frequency a nd Percentage by School Site (N = 906) ................................ ...... 92 Table 8. Demographic Characteristics of Student Par ticipants: Frequency a nd Percentage by Group (N = 906) ................................ .............. 94
viii Table 9. Demographic Characteristics of Teacher Participants: Frequency and Percentage by Group (N = 55) ................................ ................ 96 Table 10. Descriptive Statistics of Reading Scores for School A: Mean, Standard D eviation, Skewness, Kurtosis, Minimum, and Maximum by Group at Time (N = 472) ................................ ......... 98 Table 11. Descriptive Statistics of Reading Scores for School B: Mean, Standard Deviation, Skewness, Kurtosis, Minimum, and Maximum by Group at Time (N = 427) ................................ ......... 99 Tabl e 12. Site Comparison at Time 1 (N = 472) ................................ .......... 101 Table 13. Site Comparison at Time 1 (N = 427) ................................ .......... 101 Table 14. Descriptive Statistics of Stud ent Reading Scores at Time 1 by Treatment Level Between School Sites (N = 899) ...................... 102 Table 15. Descriptive Statistics of Reading Scores for School A: Mean, Standard Deviation, Skewness, Kurtosis, Minimum, and Maximum by Group at Time 2 (N = 472) ................................ .... 103 Table 16. Descriptive Statistics of Reading Scores for School B: Mean, Standard Deviation, Skewness, Kurtosis, Minimum, and Maximum by Group at Time 2 (N = 427) ................................ .... 104 Table 17. Descriptive Statistics of Reading Change Scores for School A: Mean, Sta ndard Deviation, Skewness, Kurtosis, Minimum, and Maximum by Group (N = 472) ........................... 106
ix Table 18. Descriptive Statistics of Reading Change Scores for School B: Mean, Standard Deviation, Skewness, Kurtosis, Minimum, and Maximum by Group (N = 427) ........................... 106 Tabl e 19. Change Scores (N = 472) ................................ ............................. 110 Table 20. Change Scores (N = 427) ................................ ............................. 110 Table 21. Descriptive Statistics of Combined Reading Scores: M ean, Standard Deviation, Skewness, Kurtosis, Minimum, and Maximum by Group at Time 1 (N = 899) ................................ .... 111 Table 22. Analysis of Variance Summary Table for Combined Reading Scores at Time 1 (N = 899) ................................ .......................... 113 Table 23. Descriptive Statistics of Combined Readin g Scores: Mean, Standard Deviation, Skewness, Kurtosis, Minimum, and Maximum by Group at Time 2 (N = 899) ................................ .... 114 Table 24. Analysis of Variance Summary Table for Combined Reading Scores at Time 2 (N =899) ................................ ........................... 115 Table 25. Descriptive Statistics of Readi ng Change Scores Combined Schools: Mean, Standard Deviation, Skewness, Kurtosis, Minimum, and Maximum by Group (N = 899) ........................... 117 Table 26. Reading Change Scores (N = 899) ................................ .............. 119
x Table 27. Descript ive Statistics of Mathematics Scores for School A: Mean, Standard Deviation, Skewness, Kurtosis, Minimum, and Maximum by Group at Time 1 (N = 469) ................................ .... 120 Table 28. Descriptive Statistics of Mathematics Scores for School B: Mean, Standard Deviation, S kewness, Kurtosis, Minimum, and Maximum by Group at Time 1 (N = 423) ................................ .... 121 Table 29. Comparison at Time 1 (N = 469) ................................ ................. 122 Table 30. Comparison at Time 1 (N = 423) ................................ ................. 123 Table 31. Descriptive Statistics of Student Mathematics Scores at Time 1 Pooled by Treatment Level Between School Sites (N = 892) ................................ ................................ ...................... 124 Table 32. M ean, Standard Deviation, Skewness, Kurtosis, Minimum, and Maximum by Group at Time 1 (N = 892) ............................. 125 Table 33. Descriptive Statistics of Mathematics Scores School A: Mean, Standard Deviation, Skewness, Kurtosis, Minimum, and Maximum by Group at Time 2 (N = 469) ................................ .... 127 Table 34. Descriptive Statistics of Mathematics Scores School B: Mean, Standard Deviation, Skewness, Kurtosis, Minimum, and Maximum by Group at Time 2 (N = 423) ................................ .... 128
xi Table 35. cores: Mean, Standard Deviation, Skewness, Kurtosis, Minimum, and Maximum by Group at Time 2 (N = 892) ............................. 129 Table 36. Descriptive Statistics of Mathematics Change Scores Combined Sites: Mean, Standard Deviation, Skewness, Kurtosis, Minimum, and Maximum by Group (N = 892) ........................... 131 Table 37. Mathematics Change Scores (N = 892) ................................ ....... 135 Table 38. Descriptive Statistics of Teacher DI: FIT Observation Mean Scores: Mean, Standard Deviation, Skewness Kurtosis, Minimum, and Maximum by Group (N = 55) ............................. 137 Table 39. Analysis of Variance Summary Table for Teacher DI: FIT Observation Scores by Treatment Level and by School (N=55) ................................ ................................ .......................... 139 Table 40. n Score, School (N = 55) ................................ ................................ ............ 140
xii List of Figures Figure 1. Concept Map of Differentiated Instruction ................................ ................ 10 Figure 2. Balancing the Equation to Make Differentiation Work ............................. 29 Figure 3. FC AT Reading Developmental Scale Change Scores Figure 4. FCAT Mathematics Developmental Scale Change Scores Figure 5. Supporting and Assessing Teacher Implem entation Fidelity: Triangulation of Teacher Data ................................ ....................... 64 Figure 6. Design of Teacher Participants Assignment by School and Grade Level (N = 55) Figure 7. Mean Grade Level Reading Test Scores at School A by T reatment Group at Time 1 (N = 472) ................................ ......... 100 Figure 8. Mean Grade Level Reading Test Scores at School B by Treatment Group at Time 1 (N = 427) ................................ ......... 100 Figure 9. Mean Grade Level Reading Test Scores at School A by Treatment Group at Time 2 (N = 472) ................................ ......... 104 Figure 10. Mean Grade Level Reading Test Scores at School B by Treatment Group at Time 2 (N = 427) ................................ ......... 105
xiii Figure 11. Level and Grade Level (N = 472) ................................ ................ 108 Figure 12. est Scores by Treatment Level and Grade Level (N = 427) ................................ ................ 108 Figure 13. Time 1 (N = 899) ................................ ................................ ......... 112 Figure 14. Time 2 (N = 899) ................................ ................................ ......... 114 F igure 15. Time 2 (N = 899) ................................ ................................ ......... 116 Figure 16. Level (N = 899) ................................ ................................ ............ 118 Figure 17. Mean Grade Level Mathematics Test Scores at School A b y Treatment Group at Time 1 (N = 469) ................................ ......... 121 Figure 18. Mean Grade Level Mathematics Test Scores at School B by Treatment Group at Time 1 (N = 423) ................................ ......... 122 Figure 19. Level at Time 1 (N = 892) ................................ ........................... 126 Figure 20. Level at Time 2 (N = 892) ................................ ........................... 130 Figure 21. by Grade Level (N = 892) ................................ ............................ 132
xiv Figure 22. Line Graph of Mean Mathematics Scores at Tim e 1 and Time 2 (N = 892) ................................ ................................ ..................... 133 Figure 23. Mean Teacher DI: FIT Observation Scores by Treatment and by School (N = 55) ................................ ................................ ............ 138
xv Implementing Differentiated Instruction in Urban, Title I Schools: Effects of Facilitated Support Groups and Program Fideli ty on Student Achievement Deborah W. Hellman ABSTRACT This study presents the results of a mixed methodology study and pilot that investigated the effects of facilitated teacher support groups and differentiated instruction on student achievement at two u rban, Title I middle schools. Both general education and students with special needs being served in a collaborative co taught setting were included in the study. Implications for research to practice and effective inclusive strategies were addressed and t he field tested Differentiated Instruction: Fidelity Implementation Tool (DI: FIT) used to assess teacher fidelity is included. During the first year, the principal investigator developed and field tested the DI: FIT observation tool, field tested a facil itated support group, and collected student achievement data to determine the feasibility of the implementation of differentiated instruction research design. During this second year, two matched urban, Title I middle schools were purposively selected to s erve as research sites. At each of the two school sites, 13 to 15 teachers were selected to participate in the treatment group and 13 to 14 teachers in the control group. The teachers selected were balanced among the three grade levels within each school. A triangulation of data from monthly, 2 hour, facilitated
xvi perspective) were utilized to det ermine the impact of differentiated instruction on teacher implementation fidelity. Finally, the effects of teacher use of differentiated instruction with fidelity on the reading and mathematics achievement scores of approximately 906 students (461 in the treatment group and 445 in the control group) that were part of the combined sample population at the two school sites were assessed using ANOVA f effect sizes are included.
Chapter One Introduction Statement of the Problem Unquestionably, there are many problems facing education today. Two of these students. First, the traditional one day professional development opportunities provided to teachers are expensive and have not demonstrated a transfer of practices to the classroom (CEPRI, 2005; Greenwood & Abbott, 2001; Gregory, 2003; Guskey, 1986; Joyce & Showers, 2002; Klingner, Ahwee, Pilonieta, & Menendez, 2003; Little, 1993; PCESE, 2002; R ichardson, 1997; Showers, Joyce & Bennett, 1987; Spencer & Logan, 2003; U.S. DOE, NCES, 2000). Second, many of the traditional modes of instruction currently used by teachers are inadequate to meet the varied needs of learners, especially struggling learne rs and diverse learners (Brandt, 1998; Chapman & King, 2005; NAS, 2002; PCESE, 2002; Tomlinson, 1995; Tomlinson, 1999; Tomlinson, 2003a; Tomlinson, 2003b; U.S. DOE, NCLB, 2002; U.S. DOE, OPSE, 2005). on on Excellence in Special Education inadequate for a number of reasons, including lack of substantive and research based content, the lack of systematic follow up necessary for sustaina
2 generations (CEPRI, 2005). Klingner, Ahwee, Pilonieta, and Menendez (2003) further state that although we know more about how to conduct professional development than we did twenty years ago, we have not learned how to increase the implementation of evidence based practices on a broad scale. Leading researchers i n the field attribute this gap between research and practice to limited implementation fidelity perpetuated by insufficient administrative support, inadequate follow up support, little teacher collaboration at individual school sites, the pressures of high stakes testing, and a general lack of time (CEPRI, 2005; Guskey, 1986; Joyce & Showers, 2002; Klingner, Ahwee, Pilonieta, & Menendez, 2003; Little, 1993; Showers, Joyce & Bennett, 1987; Spencer & Logan, 2003; U.S. DOE, NCES, 2000). An additional factor t hat educators must address is the increased diversity in the general education environment. Recent data from the National Center for Educational very diverse group of students. Within the same classroom, teachers typically have students from a variety of cultural, language, and religious backgrounds who have population is comprised of 60% Caucasian and 40% students of color with 10% of the students receiving services for English language learners, 13% receiving services for students with disabilities, and 36% receiving free or reduced lunch (U.S. Department of Education, 2004). In u rban districts, the diversity percentages are usually higher. The population of students in the district selected for this study are 43% Caucasian and 57%
3 students of color with 13% of the students receiving services for English language learners, 16% rece iving services for students with disabilities, and 50% receiving free or be able to tail (U.S. Department of Education, 2005, p. 13). Contemporary classrooms are clearly more diverse than they were ten years ago and, with the current standards and accountability mandates, many teachers feel ill equipped to meet the needs of all of their students (Grant, 2000). This is especially relative to educating students with special needs. The No Child Left Behind legislation (U.S. Department of Education, 2002) requires that all st udents, including students with special needs, meet or at least make adequate yearly progress toward uniform benchmarks. In addition, parents, school districts, and the state and federal governments are requiring schools to implement the least restrictive environment provision of the Individuals with Disabilities Education Act (1997). Consequently, more students with disabilities are able to access the general education curriculum and classrooms. With the challenge of meeting the needs of diverse learners, teachers require strategies and an educational philosophy that will help them to meet those requirements. Willis and Mann (2000) state: Every child is unique. Although we may rejoice in this fact, it poses a dilemma for educators. When students are diverse teachers can either teach to the middle and hope for the best, or they can face the challenge of diversifying their instruction. (p. 1)
4 Some of the evidence based and promising practices that educators are encouraged to implement in their classroom to me et the needs of the diverse learner are early intervention programs, cooperative learning, peer tutoring, direct instruction, mnemonic strategies, teaching reading comprehension, scientific inquiry, formative evaluation, and differentiated instructional st rategies (Joyce & Showers, 2002; Forness, 2001; Klingner, Ahwee, Pilonieta, & Menendez, 2003; USDOE, OPSE, 2005). Of these, one particular instructional philosophy that is gaining support is differentiated instruction (DI). Differentiated instruction is a proactive, student centered approach for teaching diverse learners in a supported, heterogeneous environment in which assessment drives the instruction. The differentiated instruction philosophy utilizes student assessment data to provide multiple learning opportunities for students that vary the content, process, and product in a blend of whole class, group, and individual instruction according to the readiness, learning profile, and interests of the students (Tomlinson, 2000). The students must be present ed with respectful tasks that are both engaging and challenging in flexible Differentiated instruction has become a priority topic among educators because of its potential to transform class room environments and to motivate students (ASCD, differences of auditory, visual, and kinesthetic learners, motivates students, and taps into gory, 2003; Willis & Mann, 2000). They also require strategies to help create a classroom atmosphere that accepts and celebrates diversity (Fullan, 2001; Tomlinson, 2003b). The differentiated instruction philosophy supports the
5 current school philosophies of teaming, celebrating diversity, community building, and supporting the needs of all children (Lawrence Brown, 2004; Tomlinson, 2003b). However, as previously mentioned, in order to assess the impact of differentiated instruction on classroom learners, professional development facilitators and administrators must provide adequate on going teacher support to ensure sustainability and to evaluate fidelity of implementation in the classroom (Blozowich, 2001; Boyd, 2001; Greenwood & Abbott, 2001; McAdamis, 2 001; NAS, 2002; PCESE, 2002; Spencer & Logan, 2003; U.S. Department of Education, 2002). The literature on differentiated instruction and teacher support groups recommend further investigation. This study will contribute to the knowledge base for different iated instruction, facilitated support groups, and implementation fidelity. This study will explore the effect of facilitated support groups on the degree of teacher implementation of differentiated instruction strategies in middle school, inclusive classr ooms. This study will also evaluate the relationship among classrooms that utilize the differentiated instruction philosophy, the degree of teacher fidelity, and high stakes academic state tests. Theoretical/ Conceptual Framework ork guided the investigations of this study. Constructivist researchers strive to gain consensus through their methodology by designing all constructions to be as real and accurate as possible and then by comparing and contrasting the constructions from mu ltiple perspectives. Consequently, this study employed both quantitative and qualitative methods that naturally match the setting and research questions and provide multiple opportunities for the confirmation of findings.
6 First, it is widely accepted that the traditional professional development workshops that utilize a top down approach in which educators come together for a day and receive information and materials with no follow up or support have limited impact on teacher implementation (Greenwood & Ab bott, 2001; Gregory, 2003; Guskey, 1986; Joyce & Showers, 2002; Klingner, Ahwee, Pilonieta, & Menendez, 2003; Little, 1993; Richardson, 1997; Showers, Joyce & Bennett, 1987; Spencer & Logan, 2003; U.S. DOE, CPRE, 1995; U.S. DOE, NCES, 2000). Over the last twenty years, professional development research has specifically focused on successful professional development programs in the hope of identifying components that will help close the research to practice gap. Because of their research efforts, there is no w a list of factors that researchers generally agree facilitate the implementation of knowledge and skills learned during teacher in service opportunities. Effective components include on going teacher support, the use of assessment to inform instruction, time for collaboration, necessary materials, coaching, and feedback on implementation (Joyce & Showers, 2002; Klingner, Ahwee, Pilonieta, & Menendez, 2003; Little, 1993; Richardson, 1997; Showers, Joyce & Bennett, 1987; Spencer & Logan, 2003). This study w as designed to investigate the use of on going, facilitated, teacher support groups, a practice that in previous research studies has claimed to increase the transfer of practices from the professional development opportunities to the classroom (Joyce & Sh owers, 2002; Klingner, Ahwee, Pilonieta, & Menendez, 2003; Little, 1993; Richardson, 1997; Showers, Joyce & Bennett, 1987; Spencer & Logan, 2003). Research findings have shown that the use of teacher support programs help schools and districts to ensure th at the money, time, and other resources spent on teacher in service are not
7 wasted (Blozowich, 2001; Boyd, 2001; Greenwood & Abbott, 2001; McAdamis, 2001; NAS, 2002; PCESE, 2002; Spencer & Logan, 2003; U.S. Department of Education, 2002). An effective teac reduce their stress and burnout, increase retention, and improve teaching effectiveness Duffy, Prohn, Ray, & Herzog, 2005, p. 8). Impleme nting a new instructional program or philosophy is difficult. According to Fullan (2001), the use of an on going teacher support program will help offset the half of t he teachers involved with change began with enthusiasm and confidence; but when they ran into challenges, their enthusiasm and confidence waned, and they typically gave up. The use of on going teacher support groups can minimize the implementation dip effe ct (Osborne, 1993; Richardson, 1997; Sparks, 2001). Tomlinson (2005) points out from seven to fifteen years. Within a teacher support group, members encourage each o ther, engage in collaborative problem solving, share ideas and successes, and explore available literature and resources. To maximize the benefit of the support group, participation should be voluntary and nonjudgmental (Westling, Cooper Duffy, Prohn, Ray, & Herzog, 2005). On going, facilitated teacher support/study groups have the potential to aid in the students (Boyd, 2001; Davis, 2003; Joyce & Showers, 2002; Klingne r, Ahwee, Pilonieta,
8 & Menendez, 2003; Little, 1993; Martin, 2000; Pfaff, 1999; Richardson, 1997; Showers, Joyce & Bennett, 1987; Spencer & Logan, 2003). Quality professional development and the implementation of instructional strategies that address the d factors of the educational equation (Brandt, 1998; Chapman & King, 2005; NAS, 2002; PCESE, 2002; Tomlinson, 1995; Tomlinson, 1999; Tomlinson, 2003a; Tomlinson, 2003b; U.S. DOE, NCLB, 2002; U.S. DOE OPSE, 2005). Differentiated instruction is an instructional philosophy that combines numerous elements of evidence based practices into a holistic instructional model. This philosophy has been gaining popularity since the 1990s (Hodge, 1997; Tomlinson, 1 995; Tomlinson, 1999; Tomlinson, 2000) and is being adopted by many school districts across the country. The No Child Left Behind Act (NCLB) (U.S. Department of Education, 2002) and the Excellence in Special Education (PCESE) (200 2) both call for the increased implementation of evidence based practices. The U.S. Department of Education, Office of Post Secondary Education (2005) even makes specific references to the recommended use of differentiated instruction in urth Annual Report on Teacher Quality: A Highly Qualified Teacher in Every Classroom. However, while there is much interest in the model, there is a paucity of empirical investigations evaluating the effect of differentiated instruction on student achievem fidelity to the model. This lack of evidence is partly because the multi facets of the differentiated instruction philosophy make it difficult to operationalize. Differentiated instruction is a proactive, student ce ntered approach for teaching diverse learners in a supported,
9 heterogeneous environment. In addition, assessment drives the instruction, which provides multiple opportunities to vary the content, process, and product in a blend of whole class, group, and i ndividual instruction according to the readiness, learning profile, and interests of the students. The following widely respected conceptual map (see Figure 1) developed by Tomlinson and Allan (2000) provides a concise visual. When differentiating their i nstruction teachers use knowledge about each of their students to help them create lessons that provide two to four learning options. Teachers also employ a balance of strategies and approaches so each learner can be successful, e.g., stations, choice boar ds, curriculum compacting, cubing (see Appendix A). A differentiated classroom has some obvious differences when compared to a traditional classroom (see Appendix B). To begin with, traditional classrooms are designed for organized, left brain learners; wh ereas, differentiated classrooms support multiple learner profiles (Willis & Mann, 2000). Specifically, differentiated instruction is a proactive, student centered approach for teaching diverse learners in a supported, heterogeneous environment (see Append ix C) (Heacox, 2002). Assessment drives the instruction, which provides multiple opportunities to vary the content, process, and product in a blend of whole class, group, and individual instruction (Tomlinson & Allan, 2000; Tomlinson, 2003b). Differentiate philosophy. It provides an opportunity for both the student and teacher to learn together and from each other (Lawrence Brown, 2004; Tomlinson, 2001).
10 Differentiated Instruction guided by general principles of differentiation, such as respectful on going assessment tasks and adjustment flexible grouping Teachers can differentiate Content Process Product Readiness Interests Learning Profile through a range of instructional and management strategies such as multiple intelligences tiered lessons 4MAT jigsaw tiered centers varied questioning taped materials tiered products strategies anchored activities learning contracts interest centers varying organizers small group instruction interest groups varied supplementary orbitals compacting materials independent study varied journal prompts literature circles complex instruction Figure 1. Concept map of differentiated instruction. Note From Leadership for Diffe rentiating Schools & Classrooms by C. A. Tomlinson and S. D. Allan, 2000, Alexandria, VA: Association for Supervision and Curriculum Development, p.3. Reprinted with permission from ASCD, February 2006.
11 In order for individual classrooms and schools to support differentiated instruction, school districts are encouraged to provide differentiated instruction in service opportunities, administrative support, money for additional resources, teacher support groups, and site based facilitators (Willis & Mann, collaborative situations with a coaching component that includes study teams and opportunities to problem solve with supportive colleagues have an 80 to 90 percent chance of applying the innovation into their classroom repertoir Finally, there is the issue of implementation fidelity. It is generally supported that for an instructional strategy or intervention to produce the desired effect, the classroom application must closely match the conditions an d implementation procedures of the original research study (Cook & Campbell, 1975; Mokrue, Elias, & Bry, 2005; Seachrest, West, Phillips, Redner, & Yeaton, 1979). In the classroom, however, it is very difficult to determine the effectiveness of an instruct ional delivery program without also assessing the fidelity of implementation. When fidelity to the model is not monitored and maintained, researchers should always question whether the reported effects of the intervention are possibly the result of outside interventions and influences or whether the lack of desired results may possibly be due to an improper or inconsistent implementation of the intervention (Mokrue, Elias, & Bry, 2005). In order to implement instructional practices adequately, administrati on and support team members must agree that program fidelity will be observed and documented. The use of an observation measurement tool further helps teachers and administrators measure and maintain instructional fidelity (Gregory, 2003; Tomlinson & Allan 2000). Conducting on going assessments of teacher fidelity to the model is critical
12 to student and program success (Webster Stratton, 2003). Without this final component of a school program, the school and/or district will never know if the program is tr uly effective or if the outcomes are the result of chance. The concern for assessing program fidelity is not new and it is definitely not waning (Cook & Campbell, 1975; Greenwood l., 2003; Seachrest, West, Phillips, Redner, & Yeaton, 1979; U.S. Department of Education, 2002). Regardless of the intervention, teachers, facilitators, and administrators need to encourage a high degree of program fidelity and to collect and analyze impl ementation data on all interventions, treatments, and evidence based practices. Purpose The goals of this study were (1) to investigate the effect of differentiated instruction on the mathematics and reading achievement of urban, middle school students; (2) to monitor teacher fidelity to the differentiated instruction model; (3) to assess the effect of facilitated support groups on teacher fidelity; and (4) to evaluate the relationship between the degree of teacher implementation of differentiated instruc tion and student achievement scores. This study contributes to the limited body of research addressing classroom implementation of the differentiated instruction model. Research Questions 1. What were the effects of differentiated instruction with teacher su pport during a five month period on the academic achievement outcomes of urban, Title I, middle school students? 2. What were the statistical differences among teacher groups who participated in facilitated support groups and those who did not with respect to their
13 implementation of differentiated instruction as measured by the Differentiated Instruction: Fidelity Implementation Tool (DI: FIT) (see Appendix D) observation tool? 3. impleme change scores? 4. Using qualitative data and feedback provided by the teachers in the treatment and thei r instructional growth? Significance of the Study In the context of accountability, schools across the United States need to implement research based practices (Brandt, 1998; Greenwood & Abbott, 2001; et al., 2003; Joyce & Showers, 2002; Klingner, Ahwee, Pilonieta, & Menendez, 2003; Little, 1993; Richardson, 1997; Showers, Joyce & Bennett, 1987; Spencer & Logan, 2003; U.S. Department of Education, NCLB, 2002). Differentiated instruction is an emerging practice that requires further investigation. Equally important, a gap still exists between research and practice accurately monitored, administrators must organiz e and endorse on going, facilitated support groups so teachers can work with other professionals and problem solve site based solutions to the inevitable challenges of implementation. Finally, because instructional fidelity is difficult to maintain this st udy incorporates a component of continuous assessment where fidelity to the model is observed and measured.
14 In this period of emphasis on high stakes testing, administrators and educators are being held accountable for the progress of all students, and tea chers are desperately searching for effective instructional strategies. Currently, many educators find themselves struggling to teach their increasingly diverse classes. Furthermore, differentiated instruction combines many evidence based practices for tea ching within one educational philosophy, which provides opportunities for increased social interaction, appropriate learning strategies, helpful feedback, and a positive learning environment (Brandt, 1998; Chapman & King, 2005; Hornsby & Diket, 1999; Tomli nson, 2005). Professional standards warrant that all teachers provide students with a positive, interesting, challenging, collaborative, and supportive learning environment and that they can no longer afford not to differentiate their instruction and administrators can no longer collaborative efforts, and thus, all students will be afforded opportunities to learn. Operational Definitions Curriculum. The term curriculum refers to the prescribed content, skills, values, and attitudes that schools a nd teachers are accountable for teaching (U.S. Department of Education, 2002). Differentiated instruction. Differentiated instruction is a proactive, student centered approach for teaching diverse learners in a supported, heterogeneous environment. Assessm ent drives the instruction, which provides multiple opportunities to vary the content, process, and product in a blend of whole class, group, and individual
15 instruction according to the readiness, learning profile, and interests of the students (Tomlinson, 2000). Facilitated support group A facilitated support group, also commonly referred to as a teacher study group, is represented by a small group of teachers, usually led by a facilitator, who come together on a regular basis to learn about a particular topic and to provide each other with support, information, and suggestions relating to implementation of a common instructional focus. Flexible grouping. With flexible grouping, the students are continuously grouped and regrouped, to prevent struggling students from being singled out. Teachers can differentiate student groups by content, process, or product according to student readiness, interest, or learning profile. Students are expected to eventually work with all students in the classroom and teache rs should not overuse any particular student grouping. Implementation fidelity. Implementation fidelity is the degree to which a teacher maintains the integrity of a particular instructional program (reliability) and implements it in the classroom. Inclusi on. Educating students with individual educational plans in the general education classroom for all or part of the school day while providing the appropriate supports as needed. Mathematics achievement. Mathematics achievement is an estimate of the student number and operations, algebra, geometry, measurement, data analysis and probability,
16 problem solving, and reasoning as defined by the National Council of Teachers of Mathema tics (2005). Middle school. Middle school refers to the school configuration that includes sixth grade through eighth grade. In middle school, students are attached to heterogeneous teams that are taught by an interdisciplinary team of teachers. Mild to mo derate disabilities. Students who are identified as members of this category are students with disabilities that are cognitive, emotional, and/or physical and do not severely limit their ability to benefit from inclusion in the general education environmen emotional/ behavioral disabilities. Professional development. Professional development is also frequently referred to as staff development or in service training. It is used to describe the professional training experiences in which teachers participate in order to improve their instructional skills and/ or knowledge. Teachers usually receive stipe nd pay and in service points, which help satisfy recertification requirements. Reading achievement. to respond to standardized assessment items that measure phonics, phonemic awareness, vocabulary comprehension, and fluency as defined by the U.S. Department of Education. Standards based test. A test based on student learning standards. They are standardized achievement tests with criterion referenced interpretations.
17 Tiered assignments A lesson i n which all students are working toward the same key concept even though they are purposely divided into groups that are adjusted according to student readiness. Title I schools According to the No Child Left Behind legislation (U.S. Department of Educati on, 2002), the public schools in each district that have student income families are eligible to receive Title I, federal assistance funds. The school districts usually distribute the mone y based on the percentage of students eligible for free or reduced lunch at each site beginning with the schools with the highest percentages of eligibility downward until all funds have been expended. The purpose of the additional funds is to target the a cademic achievement of children from low income families. Title II schools Some schools designated as Title II schools are eligible to receive supplemental funds in addition to their Title I funding (U.S. Department of Education, 2002). These schools have the highest percentages of students who qualify for free or reduced lunch and are in need of academic assistance. Title II schools in the district selected for this research study must have 90% or more of their students eligible for free or reduced lunch. Teachers who chose to work at these schools receive a supplement. Delimitations The results of this study may be generalized to diverse, urban middle school student populations in the southern United States and to the teachers who teach these populations using a team and inclusionary approach. The facilitated support group function results are appropriate for facilitators wishing to implement single site
18 differentiated instruction support groups with teachers who are differentiated instruction novices. Li mitations The interpretation of results of this study may be limited by the following possible threats to internal validity. First, there was the factor of history and the fact that some students can be exposed to instructional strategies and information outside of those used by their main teachers that may affect their achievement scores. However, in order to control for this threat, purposive comparison and treatment groups were selected from within each school site, which will help control for any schoo l or neighborhood level factors. In this way, if a school starts a new educational initiative, it will tend to have an even effect on both groups and minimize any differences. Second, even though the data collection period for this study was five months, a ny maturation changes in the students and teachers was similar for both the comparison and treatment groups and has little effect on the results of this study. Third, because the Stanford Diagnostic Reading Test Fourth Edition and FCAT Mathematics Predic tor/ Benchmark Tests are administered every year to students in Florida, the Florida Department of Education has taken great care to make sure that the different forms of the test have equally high reliability and validity estimates without sensitizing the students to the questions or testing process. Fourth, to combat any reliability issues with the DI: FIT, the observation tool was field tested prior to this study and all observers were methodically trained and monitored with periodic checks for inter rat er reliability to ensure a high degree of reliability and validity. Furthermore, because there was no possible way to control for the loss of students to particular classes, the large number of students and teachers that were a part
19 of this study helped to minimize the effect of experimental mortality. In addition, the achievement scores from any students who are not present during the entire study were removed from the data set and not included in the final analyses. Both school sites for this research pro ject were carefully selected and matched on multiple factors to ensure that the student populations were as closely matched as possible. For example, the schools were matched according to school size, SES, percentage of minorities, percentage of students w ith disabilities, region, overall school grade, and percentage of students not making adequate yearly progress. Finally, the bias of the experimenter or experimenter effect was controlled for by having a senior researcher who oversaw the entire study and b y conducting frequent member and observer checks.
20 Chapter Two Review of the Literature Overview Due to the multi faceted qualities of differentiated instruction, it is difficult to research. Differentiated Instruction is not one entity, but a synthesi s of many educational theories and practices. This holistic, student centered approach combines many of differentiated instruction as a whole, many of its common sense co mponents and strategies are well grounded in decades of research on effective instructional practices. For example, the approaches of differentiated instruction include research supported e, most of the current empirical research that is available has been done on gifted students. However, there is a growing body of individual research cases that demonstrate the effectiveness of differentiated instruction on students with disabilities in an inclusion setting (Baumgartner, Lipowski, & Rush, 2003; Good & Weaver, 2001; McAdamis, 2001; Tomlinson, 2000; Wertheim & Leyser, 2002).
21 Literature Search In order to locate any pertinent research that might pertain to this study an exhaustive search was conducted using electronic searches, manual searches, and utilized to search Cambridge Scientific Abstracts (CSA) Illumina, SAGE, Educational Resources Information Cente r (ERIC), Wilson Select, all of the education full text databases, and the dissertation database in an attempt to locate all articles pertaining to combinations of these terms. This search yielded several hundred articles, however upon reading the abstracts it was found that most of the articles did not pertain to the focus of and reviewed for possible inclusion into the literature review. Research articles that pertained to middle school and/ or urban students were given priority. Recent issues of popular peer reviewed research journals that focus on staff development and classroo m instruction were also manually searched for the same topics that were previously listed. However, only a few more articles were located using this method. Finally, discussions were conducted with prominent researchers in the field who gave suggestions of books for background information. This researcher also conducted an audit of the references listed for any article selected for inclusion in this review.
22 Professional Development Professional development for educators is an integral and costly part of t he instructional process. The Council for Educational Policy, Research, and Improvement (CEPRI) (2005), an appointed council under the Office of Legislative Services for the Florida Department of Education, recently reported that during the 2002 2003 schoo l year approximately $182 million was spent on staff development for teachers in Florida, not including the cost of substitutes. At the time, this amount equated to approximately $1,150 per teacher. Based on the percentages, the committee estimates that in 2005 also points out that there is currently no systematic way to assess the benefits of in service education and its impact on student achievement (CEPRI, 2005). It i s interesting to note that twenty years ago Guskey (1986) researched the major ineffective and unconnected to student achievement. At that time, he recommended that e up after the initia l training (p. 9 Showers (2002), who have also been researching this problem for the last twenty years, similarly state that teachers and facilitators tend to underestimate the cogni tive aspects of instructional implementation. The ability of staff development personnel and facilitators to effect the transfer of skills from the professional in service to the classroom continues to be an area of research that is under investigation. Th ey state that, unfortunately, most
23 going support, coaching, and the use of assessment to inform instruction (Joyce & Showers, 2002). The researchers further highlight the use of these effective implementation practices in projects such as Just Read, Read to Succeed, Success for All, The River City Experience, and the Schenley School Project. Statist ics conducted their own study on teacher preparation and professional development. They surveyed 5,253 teachers from all 50 states using a stratified sample procedure. In the report, they specifically criticized traditional one day in service opportunities as inadequate. Further, with regard to the types of in service attended performance st proficien administration encouraged them to apply their new skills in the classroom. This report serves to confirm the lack of quality professional development activities that are available to te achers that encourage instructional strategies that support diversity and the lack of administrative support for professional development. Lester (2003) in a study of 93 secondary teachers in Louisiana studied the components of effective professional deve lopment. Using a questionnaire, interviews, observations, and a reflective writing activity, she concluded that small, on going
24 collaborative support groups that met on a regular basis and provided opportunities for reflection were perceived as effective b y the teachers in helping them to grow Klingner, Ahwee, Pilonieta, and Menendez (2003) studied professional implement ation of evidence based practices in inclusion classes. They specifically focused on the barriers and facilitators to implementation. For their study, they collected data from 29 teachers from six different elementary schools in Miami, Florida using interv iews, teacher journals, and classroom observations. Each of the participants attended a two week professional development program that targeted the use of four researched based reading strategies: partner reading, collaborative strategic reading, making wo rds, and phonological awareness. After the initial training, each teacher received on going support for the remainder of the school year. Particular attention was made to help each teacher adapt the strategies to fit his or her particular teaching style. R esearchers used a qualitative analysis of coding the chunks of data into five a priori categories for the interview and observation data and then they analyzed the teacher journals using SPSS 10.1. In a comparison of high implementers (HI), moderate imple menters (MI), and low time was the most frequent complaint by the teachers, although i t was noted that the teachers who implemented at a low level cited this reason more often than the teachers who were observed to be high implementers (HI). In addition, the low implementers (LI)
25 often felt overwhelmed and wanted more support and modeling o f the strategy. All of the teachers reported that the strategy facilitators were very helpful in supporting the implementation of the strategies. Most of the teachers also responded that when the students liked a particular strategy or when the students pe rformed well that knowledge with 9 teachers considered to be high implementers (HI), 9 as moderate implementers (MI), and 11 as low implementers (LI). The researchers fe lt that with a little more time and continued support the moderate implementers (MI) could shift to high implementers (HI). More than a third of the teachers implemented the strategies at a low implementation level, even after a year of support and many te achers modified the strategies, leaving out important components. The factors that did help the teachers transfer the strategy knowledge from the in service to the classroom with fidelity and helped the district to utilize their in service budget more effe ctively were on going This study provided some very important aspects to consider when implementing evidence based practices in general education classrooms that also provid e inclusion services for students with disabilities. In the discussion, the authors use five examples to point out that even the best professional development studies have mixed results ( Klingner, Ahwee, Pilonieta, & Menendez, 2003) Differentiated Instru ction The support for differentiated instruction is rapidly growing. However, while there exists a great deal of support for each of the educational practices integrated into differentiated instruction, there are a relatively small number of studies that look at the
26 package as a whole. The following represent the studies that most closely match the intended focus of this study. Beginning as early as 1990, Tomlinson, et al. (1994) began investigating the impact of differentiated instruction on high achievin g, gifted learners, pre service teachers, and teachers of the gifted. She primarily used qualitative research analysis to investigate the educational requirements of pre service teachers, the impact of differentiated instruction on students, and the qualit ies of effective differentiated instruction in services. First, she interviewed 70 pre service teachers multiple times over the course of a year as they participated in a training program designed to prepare them to teach in a mixed ability classroom. She found that pre service teachers are not sufficiently prepared by their college educational programs to teach in a mixed ability classroom and that specific training in differentiated instructional strategies is warranted. The use of multiple reviewers an d the multiple interviews over the course of one convenience sample was used, the results are typical of pre service teachers and programs in all regions of the United States. Consequently, administrators, mentors, and support group facilitators should consider the limited instructional strategy awareness of new teachers when planning staff development and instructional support opportunities. Tomlinson (1995) followed the study in 1994 on the impact of differentiated instruction on gifted learners, pre service teachers, and teachers of the gifted with another 18 month qualitative study focused on the ability of in service facilitators to change the attitude and practices of middl e school teachers with respect to differentiated instruction. For this study, she spent the entire year working with a typical mid sized suburban
27 school. She also fostered credibility and dependability by collecting data from multiple sources. She presente d staff in services, observed teachers in classrooms, participated in their small learning communities, and interviewed teachers, administrators, and parents. She again concluded that teachers do not differentiate their instruction without specific instruc tion and support and that on going administrative support is key to the change. Her research reinforced the growing belief that teachers require specific training and support in order to incorporate differentiated instruction into their instructional philo sophy. Clearly, if administrators want to make sure that teachers implement differentiated instruction with fidelity, they should formulate multi faceted programs that provide staff development, on going support, classroom observation with feedback, and sm all learning communities. As a culmination to her research, Tomlinson wrote several books outlining the principles of the differentiated instructional philosophy. In a co authored book, Leadership for Differentiating Schools & Classrooms, Tomlinson and Al lan (2000) postulate that the following key principles govern effective differentiated instruction. The differentiated instruction concept map, shown earlier (Figure 1), is useful in gaining a holistic picture of the process. The authors state that, primar ily, the classroom teacher must be flexible in his/her instruction, procedures, grouping, and assessment. It is only through on going assessment that the teacher will be able to target the specific needs and interests of the students. This continual assess ment process helps the teacher to plan future instruction. Third, the teacher adjusts the content, product, and process according aspect of differentiated instructio n will provide each student with a wide range of
28 learning opportunities and encourage the embracing of diversity. There are many grouping combinations possible using a differentiated approach. Lessons can be differentiated by content, process, or product a is possible to create variety for both the students and teacher. Fifth, the teacher must ensure that all students are wor king on assignments that are both meaningful and engaging. In addition, the skill being taught must be challenging but not frustrating. Finally, the teacher and the students must both collaborate in the learning process. Students should be allowed the oppo rtunity to make choices about their learning; it provides empowerment (Tomlinson & Allan, 2000; Tomlinson, 2003a). In the same book, Tomlinson and Allan (2000, p. 134) also provide a helpful o educators can successfully implement differentiated instruction in their classroom (see Figure 2). Some of these key components will be integrated into the current study. For example, they s, Fleming and Baker (2002) investigated the interactive role between lesson planning, student teaching participants in this study were five pre service teachers who were p laced at three rural Ohio middle schools, and the goal was to collect evidence of their transfer of the
29 differentiated instruction knowledge gained during their college pre service methods class into their field teaching experiences. The What High level, idea based instruction using key skills to understand and apply the ideas employing key principles of differentiation: o Flexible gr ouping o Respectful activities o On going assessment and adjustment Modifying content, process, and product based on student readiness, interest, and learning profile using a range of student centered, meaning making instructional str ategies Coaching for individual growth with the goal of moving each student as far and fast as possible Assessing student growth at least in significant measure according to personal growth The How Clarity of purpose and vision Systemic efforts Generalist/ specialist partnerships for classroom application Time and support for collaboration Structured lesson (curriculum) planning and instructional evaluation Focused staff development with plans for transfer Incentives for classro om application Aligned and focused policies and initiatives Coherent leadership Integration with professional growth and accountability Formative and summative evaluation of efforts and use of findings Involvement of parents in understanding and contributi ng to assessment of change Persistence over time Figure 2. Balancing the equation to make differentiation work. Note. From Leadership for Differentiating Schools & Classrooms, by C. A. Tomlinson and S. D. Allan, 2000, Alexandria, VA: Association for Supe rvision and Curriculum Development, p. 134. Reprinted with permission from ASCD, February 2006.
30 The participants in the Fleming and Baker (2002) study were purposefully selected based on their completion of the college methods class. For data collection pu rposes, each participant was asked to submit six lesson plans, three from their college classes and three from their classroom field experience. After these documents were reviewed for evidence of differentiated instruction according to content, process, o r product, each participant was observed twice during the last two weeks of their field experience, by two different observers. Once they had finished their student teaching, each participant completed a short survey and was interviewed using four open end ed questions. The participants were not told that the researchers were specifically looking for evidence of differentiated instruction so as not to bias the results. In the final report, pseudonyms were used to reference each teacher to insure confidential ity. In order to insure dependability and confirmability of the results, both researchers separately reviewed all data prior to collaboration. Each of the researchers also represented different areas of expertise; one taught general education while the oth er taught special education. The study was conducted over a one year period to increase credibility of their findings, and both method and data triangulation were utilized which further enhanced the dependability and confirmability of the report findings. Finally, to address transferability, the researchers provided very clear, thick rich descriptions of their methods and results so the results could be replicated and/ or applied to other situations; however, the small sample size and geographic location of the study will greatly limit the generalizability of the results. The results showed that there still was clearly a gap in the expectations of the university supervisors and the classroom supervising teachers. Several of the pre service
31 teachers experienc ed resistance by the classroom supervising teachers who did not understand the differentiated instruction philosophy and did not require the pre service teachers to write lesson plans. Most of the pre service teachers were still confusing differentiated in struction with simple classroom accommodations. It was also noted that the pre service teachers limited their differentiation to product options, which are the easiest to implement, with an absence of differentiation by content or process. All of the parti cipants commented in the interviews that differentiating in the classroom was harder and more time consuming than they had expected. Again, the importance of providing teachers with support and time for collaborating and developing differentiated lessons i s key to implementation. It is also very clear from this study that new teachers, even those who have participated in a formal educatio nal preparation program are ill equipped to differentiate their instruction in a mixed ability classroom. ptive study in 2004 provided limited expansion of previously known information for researchers hoping to develop their differentiated instruction school or district program. This mixed method study used a questionnaire, a focus group, and walk through clas sroom observations to collect data at a single semi rural middle school in the Shenandoah Valley in Virginia. It was noted that this district is so small that it only had one middle school, so random selection was not possible. Even though the school sel ected for this study was semi rural, the diversity of the in the sample were 61% Caucasian and 39% minority students. The non English speaking students represented 34% o f the student body and over 40% of the students qualified for free or reduced lunch. A high percentage of the students were gifted (12%), although, no
32 information was available on the percentage of students who qualified for special education. Fifty five t eachers were part of the targeted convenience pool of participants, however, only 30 teachers returned the questionnaire and became part of the study. Triangulation of evidence was used in order to increase the internal validity of the study. For the quant Teacher/ Peer Differentiation Reflection Instrument questionnaire and observation checklist. Reliability for the questionnaire, which was reviewed by a panel of exp erts with minimal changes, was established using the split half questionnaire, participants were enticed with a cash drawing. For the observation data, 15 teachers were s elected for the unannounced walk through observation, which only lasted 15 minutes. Another limitation to the observation data was that the researcher did not specify how the participants were selected. For the qualitative portion of the study, seven teach ers from the staff voluntarily participated in the focus group. Furthermore, the participants in the focus group were enticed with the promise of a cash drawing, a factor that may affect the internal validity of the results. All instruments were field test ed first using a modified pilot. The results of the self reported questionnaire and observations indicated that most teachers who differentiated their lessons utilized content differentiation. However, the seven teachers who participated in the focus group responded that they used product differentiation more often than content differentiation. Most of the teachers in the focus planning instruction. The majority of the se teachers also felt that they were comfortable
33 using materials other than the textbook for lessons. From a pragmatic standpoint, this study did not yield any particular patterns in teacher use of differentiated instruction that were not expected. Instead it was found that there is a very wide discrepancy between the types of strategies used and the degree of implementation. As expected, the author recommended continued staff development and support within the school. In 2002, Vanfleet focused specifical ly on the use of differentiated instruction as a means to facilitate inclusion for students with mild to moderate disabilities. Using a mixed method design, she explored the effectiveness of professional development differentiated instruction training with 43 secondary school teachers in a suburban area of Alabama. The quantitative portion of the design consisted of an analysis of data collected using the Data Survey of Secondary School Teachers. The pretest posttest data from this survey were compared usin g a t to whether or not they were adequately meeting the needs of students with disabilities in the general education classroom were effected by the differentiated instruction training they receive d to a statistically significant degree ( p < .05). The qualitative portion of her study revealed that the participants with the most recent formal education had the most collaborative partnerships, improved preparation and attitude toward diversity and stu dents with disabilities, and a greater experience with differentiated instructional strategies. of differentiated instruction staff development on student achievement, perce ptions of parents, and teacher attitudes. Using a t test procedure, she analyzed the reading and math achievement data of students in grades two through six at a suburban, Alabama public
34 elementary school and found that the students who received differenti ated instruction made statistically significant gains in mathematics but not in reading [t (92) = 2.24, t critical = 1.66]. Student academic growth was measured using the Stanford Diagnostic Test: Reading and Mathematics Batteries developed in 1983. Survey s were also administered to 44 teachers and 160 parents with 79% of the parents responding. An analysis of the survey data found no statistical significant differences between the perceptions of the treatment and control groups of parents and the treatment and control diverse learners in the classroom. However, the study presents limited validity or generalizability for the diverse populations in the urban centers of Amer ica. The cluster sample of 160 students (94 treatment, 96 control) used for this quasi experiment was 98% Caucasian, upper socioeconomic status (median family income of $80,366), with only 10.7% of the students in need of academic remediation. In fact, the mean academic achievement scores of the students at the selected elementary school were well above the national average, based on state assessment results, prior to the study. Furthermore, of the teachers used in the sample, 60% held advanced degrees with a mean of 12.5 years of teaching experience, again above the national averages. There is much to be learned from the staff development and support that the participate in a 16 hour staff development with Carol Ann Tomlinson in 1996, followed up by seven months of on going support and quarterly staff development. Throughout the study, the teachers videotaped themselves, watched the tapes, and reflected on ways to
35 improve thei r instruction. They also observed teachers within the school that were considered successful and confident with their differentiation of instruction. Affholder (2003) expanded the research vein of differentiated instruction by investigating its use with al l learners in inclusive classrooms. She used a case study design with branching interviews and questionnaires to examine the implementation of differentiated instruction and the factors required to support this approach based on a district supported initia tive. The Blue Valley School District in Kansas began the initiative six years prior and made sure that the appropriate supportive components were in place by providing differentiated instruction staff development, time and resources, opportunities for col laboration, and shared decision making. The focus of this study (Affholder, 2003) was the perceptions of 26 elementary school teachers, 12 administrators, and a school board member. In addition to the interview, the selected elementary school teachers also responded to the Stages of Concern Questionnaire developed by Hall and Hord in 2001. All of the teachers in the study participated in a differentiated instruction in service two years earlier and the researcher collected evidence as to the degree of imple mentation that each teacher maintained. The branching interviews served as a vehicle to locate those teachers who had the highest levels of implementation. The consensus of the data revealed that the teachers all expressed the desire for on going support a nd staff development in addition interviewed in this study mentioned time as a critical factor for the implementation of differentiated instruction, time for lesson planning and preparation, time for collaboration, and a student contact time sufficient for assessment and instruction of
36 had a pronounced sense of responsibility for stude nt growth, familiarity with the curriculum, and a willingness to try new instructional approaches. This study is limited, as was the previous study, by the homogeneity of the sample. The small suburban Blue Valley District is primarily Caucasian (90%) and only 4% of the students qualify for special education, compared with the national average of 13% (U.S. Department of Education, 2004). In the future, if this study is replicated, a more diverse population would help to generalize the results. McAdamis (200 1), the coordinator of staff development for Rockwood School District in Missouri, is one of the few educators who seems to have pulled many of the effective components of instructional implementation into a complete design. In her article in the Journal o f Staff Development using a narrative style, she recounts the year action research process. The district began by adopting a policy of supporting differentiated instruction in order to address the needs of all learners. First, a cadre of volu nteer teachers was trained in differentiated instruction and peer coaching strategies. The group met five times during the first year, during which they learned new strategies, shared successes and challenges, developed new lessons, and practiced reflectio n. They also observed each other and that provided support for other teachers at their individual school site. Next, the district began offering differentiated instru ction workshops on a continual basis to all teachers and administrators. In addition, the district added on going instructional support in the form of release time for teachers to work on lessons, teacher support/ study groups, peer
37 coaching, and classroom level, action research support. In the end, each school was required to add a differentiated instruction goal to its school improvement plan. McAdamis (2001) admits that the change did not happen overnight, the entire process was five years in the making ; but in the end, the payoff for everyone involved increased their academic performance on the state standardized tests. Overall, the percentage of under performing student s, bottom quartile, decreased by 8% in reading and language arts, 5% in math, and 7% in science. They also increased the percentage of students scoring in the top quartile. Rockwood is a suburban district, part of the greater St. Louis area. The district e mploys approximately 3,000 teachers and has an enrollment of 22,000 students who are represented by the following percentages: 83% Caucasian, 14% African American, and 3% other. The students who qualify for free or reduced lunch represent 15% of the popula tion, and students with special needs represent 8%. Although this study was not a rigorous quantitative study and it did not represent the population targeted for this study, it does represent the power of differentiated instruction coupled with on going teacher support, teacher support/ study groups, and observation with feedback. As evidenced by the previous research reviews, there is limited information on the academic achievement effect of differentiated instruction in American classrooms. Sadly, the re are little rigorous data available on it s implications with urban middle school students. There is, however, ample evidence that observation and on going support is key to an effective differentiated instruction school program.
38 Facilitated Support Group s/ Teacher Study Groups Having a good facilitated support group is a key ingredient to initiating an effective instructional program (Westling, Cooper Duffy, Prohn, Ray, & Herzog, 2005). There has been an exceptionally strong push throughout the United St ates due to both the A New Era: Revitalizing Special Education for Children and Their Families (2002), for the field of educ ation to include more research based practices in the classroom. However, one of the primary reasons that many of these practices are absent from the classroom instructional repertoire of teachers is that the majority of teacher in service programs have li ttle to no follow up support. Consequently, when challenges arise, if teachers feel unsupported and unsure of their abilities, they will revert to what is comfortable (Osborne, 1993; Richardson, 1997; Sparks, 2001; Spencer & Logan, 2003). Many of the previ ously discussed differentiated instruction studies emphasized the characteristics and necessity of having a well run professional support group for teachers. An additional qualitative case study conducted by Hale (1999) investigated, on a smaller scale, th e use of a facilitated support group as a vehicle for professional development. For this study, seven enhancement specialists, from different urban schools, and progres s, a university facilitator met with this formal support group throughout the entire study. The initial task for the group members was to evaluate, plan, and implement research based strategies as well as finding a productive way to work together and suppo
39 The author, using thick, rich, narrative, documented the entire process as these specialists developed their collaborative process and the effect that the collaborative process had on their work and th eir relationships. As a result of the collaborative experience, the author was able to identify affective and supportive factors that such as external events, prof essional background, disposition, and expectations were key experience. Even though this was a small qualitative design with voluntary participants, the study did prov ide guidance for future researchers regarding the process and possible problems that can be avoided when facilitating professional support groups. For example, the group benefited from an experienced facilitator, group norms need to be established at the f irst meeting (be on time, respect the input from all members), notes should be taken during each meeting, member checks as to the accuracy of the notes need to be conducted immediately after each session, all members need to agree to make the meetings a pr iority and to not schedule other activities during the agreed meeting times, make sure the meeting place is comfortable, the group size should be kept small, and the group should meet at least once a month. Although this study may have limited transferabil ity, it provides a very clear picture of the affective factors that must be considered when planning a teacher support group such as the one for this research project. On a more in depth and slightly larger scale, Pfaff (1999) further investigated the eff ect of professional study groups on teacher efficacy. For her study, she used a mixed method design, primarily qualitative, to answer multiple questions regarding the effect of
40 a year long professional study group on the perceptions of seven staff members from a small, rural elementary school located in Carroll County, a north central Maryland district. and was therefore purposefully selected. All of the elementary schools fr om the district were invited to participate, however, only three principals responded. Of the three, two already had study groups in place, which might bias the teacher participants. The school selected was the smallest elementary school in Carroll County with only 417 students and 25 teachers. Again, the entire staff was invited to participate, however, only eight members volunteered and one teacher dropped out early in the process due to scheduling conflicts. Of the seven remaining members, two were ident ified as resource specialists and five were classroom teachers. The mean number of years of experience represented The results of this study therefore have limited tran sferability due to the small, purposeful, non typical sample of participants. The student body of the school was predominantly middle class and Caucasian. They had no students who were eligible for the Limited English Proficiency program, only 8% who qua lified for the free and reduced lunch program, and 15.3% who received special education services. The professional study group met for 90 minutes per month and each teacher was paid a small in service stipend for putting extra hours beyond the standard sc hool day. At the end of the year, questionnaires and interviews were utilized to collect qualitative data
41 teaching efficacy because of their participation in the professiona l study group. The results supported the belief that the group participation facilitated a sense of security and helped to increase their metacognitive awareness of their own abilities. In addition, using a pretest posttest, factorial ANOVA analysis, the responses of the teachers in the professional study group on The Teacher Efficacy Scale (TES) were compared to the responses of a control group of teachers selected f rom the same site. The author provided the reliability and internal validity scores for the TES instrument as reported by the creator of the instrument. The results of this analysis did not show a significant difference between the two groups. This lack of difference was rationalized to be directly related to the fact that the teachers in the treatment group scored high on the instrument during the pretest and therefore had little room for growth. Regression to the mean was a factor to be considered. It was further noted that the teachers in the study group maintained their high efficacy ratings until the end of the year, whereas the mean efficacy rating of the teachers in the control group declined over the course of the year. The results of this study pro vide important aspects to consider when creating teacher support/ study groups. The impact on student achievement was missing from this study and is recommended to be added in future studies. In order for the results to be more broadly transferable, the sa mple of teachers, schools, and student body should also be more representative of the typical classrooms of America. In another qualitative, longitudinal, case study, Boyd (2001) investigated the implementation of inclusionary practices at a moderately siz ed, urban middle school in a Central Florida county over the course of four years. The primary focus of her research
42 was to explore the effect of the climate, best practices, administrative support, and staff attitudes on the inclusion of students with dis abilities at the school, which educated an average of 899 students, including 142 varying exceptionality students (15.8%), 178 students with limited English proficiency (LEP) (19.8%), and 458 students on free and reduced lunch (50.9%). The mean enrollment by ethnicity was 37.6% Caucasian (n = 338.5), 14.7% African American (n = 131.8), 39.5% Hispanic (n = 355), and 8.2% other (n = 73.7). For this study, the researcher increased dependability by triangulating data from multiple sources. She conducted surveys observations, focus groups, a document review, and interviews from a pool of 38 staff members including 31 teachers, an administrator, 3 deans, 2 guidance counselors, an ESE resource specialist. She also increased her data credibility by collecting these data over a four year period. Anonymity of all participants was assured for this voluntary study. Of the 38 possible participants, 34 responded to the survey, providing good credibility of the information. From the pool of 31 teacher participants, the sev enth and eighth grade content area teachers were considered the primary focus of the study. Therefore, only 18 seventh and eighth grade teachers were selected to participate in the interviews, 11 seventh and eighth grade teachers were selected for observat ions, and 10 seventh and eighth grade teachers participated in the two focus groups. The author claimed that the 63 question, Likert scale survey was adapted from provided. The researcher followed standard development protocol while developing the classroom observation tool. Each teacher was observed nine times for approximately 45 minutes per observation. The open ended interview questions were purposefully
43 developed in ord er to address four main categories: school climate, administrative support and preparation, attitudes of the participants, and best practices. The interview participants were provided with a copy of the questions during the interview to ease their fears an d to guide the process. Member checks were utilized to verify the emergent themes. As a final confirmation piece, the focus groups helped to provide clarity of issues raised during the interviews and informal conversations on campus with the researcher. Fr Respondents indicated that over the years the program had deteriorated and they had been left o ut of the decision making process. Most of the staff agreed on the survey that the differentiated instruction classroom practices, like cooperative learning, multiple intelligences, activity based learning, portfolio assessment, and peer tutoring were all important; however, when observed, most teachers utilized a purely traditional teaching style. She also found that continuous staff development including peer coaching, mentoring, small learning communities, teacher observations, and a focus on best practi ces were components of an effective inclusive support program and should be added in the future research studies. More recently, Davis (2003) conducted a larger, mixed method research study in which she explored the relationship between the use of study g roups and the implementation of knowledge and strategies gained during professional development workshops. For this study, the sample consisted of 57 elementary school teacher participants from 14 study groups within a suburban, Pennsylvanian school distri ct. The transfer of skills from the study group to the classroom was measured using self reported
44 data using a pretest posttest design. The use of self reported data has limitations and decreases the reliability of the conclusions. The participants also co mpleted a Learning Style Inventory designed by Kolb in 1999 and a Study Group Participant Questionnaire designed by the author. In addition, both participant and facilitator interviews were gement and structure, group resources and technical support, group dynamics and interpersonal relationships, were conducted using regression analysis, ANCOVA, and MANOVA p rocedures. The statistically significant ( p < .05) results of the combined analysis revealed that all 14 study groups perceived that they were able to transfer the skills learned during professional development workshops to the classroom. However, the resu lts of the teacher learning style data provided inconsistent results that were not statistically skills. It is therefore not sufficient to rely solely on the use of inter views and self reported instructional repertoire needs to be quantified with observational and student data. Although Murphy and Lick (2001) conducted their research on who le faculty study groups, their pivotal work provides insight and guidelines that would also benefit researchers planning to conduct research on stand alone collegial teacher study groups and was therefore selected for inclusion in this literature review. I n their book, they summarized their sizeable research that was collected over a five year period and utilized data from over 2,000 study groups at over 200 schools. The authors begin by clearly stating that for study groups to be successful, the individual group members must first
45 agree on a guiding question, guiding principles, and procedural guidelines. Murphy and Lick (2001) emphasize that these criteria cannot be sacrificed or the group will flounder. In order for the study groups to be effective, they provide the following additional guidelines: Keep the size of the group to no more than six Establish and keep a regular schedule Establish group norms at the first meeting of the study group Agree on an action plan for the study group Complete a journal entry after each study group meeting Encourage members to keep individual journals for their personal reflections Establish a pattern of study group leadership Give all study group members equal status Ha ve a curriculum and instructional focus Plan ahead for transitions Make a comprehensive list of learning resources, both material and human Evaluate the effectiveness of the study group Establish a variety of co mmunication networks and systems (p. 51 59) The authors further emphasize the key role of the facilitator or leader. They state that the facilitator has an essential role in the functioning of the study group. In order for the study group to be effective, the facilitator must take on the additional roles of organizer, recorder, and liaison.
46 Based on the research by Murphy and Lick (2001), Martin (2000), an elementary school principal, conducted a qualitative descriptive multi case study investigating the ef development workshops, the interaction of the study group members, and the effects of the facilitators on the study groups. The sample consisted of 24 teachers from a r ural suburban school district in Pennsylvania. Martin (2000) collected data on two treatment study groups of seven and five members each who were matched to another 12 teachers in a control group. The 12 teachers who voluntarily participated in the study g roups were matched to teachers in the control group who had attended the same in service opportunity but chose not to participate in the study group. For her data collection, she utilized interviews, questionnaires, and an innovation checklist. She had ori ginally intended to conduct a secondary statistical analysis on some of the data, but she was unable to do this as her sample size was too small. Martin (2000) reported that the majority of the teachers in the two study groups believed that their classroom instruction had improved as a direct result of their organization, on going communication, and interpersonal functioning of the group. In a components necessary for teacher integration of professional development skills were:
47 characteristics of the participants, althou gh her descriptions of the facilitative process of the study groups would be very beneficial to researchers choosing to implement a support/ study group. Because she used self reported data, the validity of the findings may be limited. She admits that the group would have been more productive if they had kept a formal meeting log to document their progress and accomplishments. Furthermore, observations by a trained observer would have added another source of data verification. Recently, another pair of rese archers (Spencer & Logan, 2003) also investigated the use of on going teacher study groups, but added components that were missing from some of the previously discussed studies. Spencer and Logan (2003) used a time series experiment with a treatment and a comparison group to study the effects of a school based staff development model that utilized the support of a Research Lead Teacher (RLT) to help general education teachers develop and maintain instructional fidelity to For this study, Spencer and Logan (2003) selected a large elementary school (1100 students) in a large suburban school district. The location of the school was not provided. The student body was representative of the U.S. Depar NCES (2004) national averages for elementary schools with 59% Caucasian, 18% African American, 12% Asian, 8% Hispanic, and 3% other. The percentage of students enrolled in special education was unavailable, however, it was noted that 2 5% of the students qualified for free or reduced lunch. The mean number of years of teaching experience for the intervention group was 9.7 years, while the comparison group averaged 11.3 years. All teachers had between two and four students with mild to
48 mo derate special needs included in the general education environment in addition to First, all of the 42 general education K 5 teachers in the selected elementary school attended the traditional half day in service provided b y the Research Lead Teacher (RLT). Then all of the teachers were invited to participate in the RTL Model that going teacher study group, coaching, observations, and data based volunteered, nine became part of the treatment group and nine became the comparison group. The data collection began during the 9 weeks before the beginning of the school year, during which the nine teachers in the treatment group attended nine 60 minute, weekly teacher study group sessions as part of a voluntary summer training opportunity. The treatment group of teachers also agreed to be observed on a weekly basis with follow up feedback provided by the RLT. The RLT also provided in class modeling and co aching of the Benchmark Strategy Instruction Process for each of the teachers as needed. For their participation, the teachers received ten staff development credit hours. The nine teachers in the comparison group, without on going support, were told from the beginning the purpose of the study. They also agreed to be observed during the baseline, treatment, and maintenance weeks. The comparison group of teachers knew that their observation data would be compared to teachers who attended the same in service with the additional coaching support. Therefore, no deception was involved. The researchers presented lessons to the best of their ability, possible John Henry Effect, thereby illustrating the impact of the RLT Model.
49 After the initial 9 week training period, all 18 teachers were observed four times during the 3 week baseline period to determine the degree of implementation by each teacher prior to the intervention from specialists and observational feedback. All observations lasted approximately 30 to 40 minutes. This phase was followed by a 9 week treatment period, during which all teachers were observed nine times. The comparison group received no feedback while the tr eatment group received specific feedback, modeling, and support. During the next 3 week rest period, no observations were conducted. This period was followed by the final 2 week maintenance check in which all teachers were again observed twice. Two trained observers were used to collect the classroom implementation data. The observers were not aware which teachers were receiving the intervention in order to avoid observer bias. Using the RLT Model, all of the teachers in the treatment group mastered the imp lementation of the 15 steps of the Benchmark Strategy Instruction Process. The mastery of the instructional strategy did not happen in isolation. All of the teachers in the treatment group required between three and nine weeks of on going support including modeling, coaching, feedback, and study group participation before they were able to demonstrate mastery. On the other hand, none of the teachers in the comparison group was able to achieve mastery of the strategy and their observation data documented a d eclining trend in their performance over the study period. Based on the data, the authors concluded that the RLT Model was an effective method of supporting the implementation of evidenced based practices. Although the components of the program cannot be s eparated, all of the teachers felt that the on going support, observations with feedback, and teacher study group sessions were critical to
50 their overall implementation fidelity. Unfortunately, no student data were collected during the study. These data wo uld have helped the researchers support their claims of especially important to note that even though all of the teachers attended the same in service, the teachers in the comparison group had more experience, and the comparison group teachers were aware of the purpose of the study, they still were not able to successfully implement the strategy without the support of the RLT Model. As evidence by the above studies, it is vital to provide systematic, on going support for teachers who are implementing a new instructional strategy using a well designed teacher support/study group in addition to observations with specific feedback. Although none of the studies specifically add ressed the urban, middle school populations, most of the studies do illustrate successful research strategies for investigating facilitated support groups. Instructional Fidelity Instructional fidelity is vital to the implementation of any evidence based instructional program. Without an assessment of fidelity, the reliability of the results will always be in question. One such qualitative study by Blozowich (2001) investigated the implementation of differentiated instruction strategies in ten sixth grade middle school, classrooms in a rural school district in eastern Pennsylvania. The researchers selected a moderately sized middle school for this study with approximately 700 students; the SES and ethnicity percentages were unavailable. The students at th is school are placed on interdisciplinary teams in which they were heterogeneously grouped according to their ability except for
51 mathematics. The students are tracked for mathematics. Students with disabilities nd students with mild to moderate disabilities were included into the general education classroom for most of their instruction. The researcher collected data using multiple assessments including a professional development survey, a differentiated instruct ion survey, an unannounced classroom observation using a checklist, and follow up interview. Although, the researcher developed all of the instruments, no reliability or construct validity information was provided, thus findings must be interpreted with gr eat caution. Through the planned variety of assessment tools, he did however allow for a triangulation of results. He also built into his design member checks with each of the participants to confirm the validity of the results and peer examination of the interpreted data themes to remove the effect of researcher bias. The results of the study (Blozowich, 2001) revealed disappointing data and themes. First, the survey results did show that although the teacher participants had been exposed to differentiate d instruction strategies through county professional development activities, the teachers as a whole did not make any effort to learn any more beyond what they had learned at the workshops. Furthermore, only three of the ten incorporated differentiated ins tructional strateg ies into their classroom lesson plans, more than half did not participate in a learning community or collaborate with other teachers, more than half stated they were satisfied with their current teaching strategies, and more than half kne w about differentiated instruction and/ or did not wish to learn additional information. From the classroom observations, it was also found that most of the teachers who claimed to be using complex differentiated strategies on the surveys did not demonstra te any evidence
52 of use when observed. The results from the interviews corroborated this finding. It was found that, despite a school board policy for teachers to include differentiated instructional strategies into their lesson plans, the majority of the t en teachers continued to teach their class like the traditional tracked classroom with very little learner differentiation. Although this study had a very small sample size with limited reliability and internal validity, it is consistent with previous res earch findings that indicate if teachers are left on their own without on going support, despite a district policy, they will continue to teach the same way they always have taught. Thus, researchers should be very cautious of data based on teacher survey responses in relation to the fidelity of instructional practices. After exhausting the databases for articles on instructional fidelity of research based practices in the classroom, the literature search was broadened to include mental health studies that focused on the fidelity of behavioral training programs for students. In particular, Webster The Incredible Years program reported key findings learned from ten years of research using ra ndom controlled trials that could easily be adapted for the implementation of instructional strategies in schools. The focus of the chapter, The Incredible Years Parent, Teacher, and Child teacher interactions regulation competence Stratton, 2004, p. 1). Throughout the entire chapter, the auth or stresses the importance of
53 fidelity when implementing evidence based interventions, frequently referred to as treatment fidelity. She also references their comprehensive teacher training intervention in which they help teachers implement classroom manag ement and discipline strategies that promote social competence. As reported by Webster Stratton (2004) the five key components for effective program implementation with fidelity are: 1. Standardization of treatment delivery using comprehensive clinician manua ls, well articulated protocols, videotapes, and materials for parents, teachers, and children 2. Standardized quality training for group leaders delivering the intervention 3. Effective supervision of group leaders 4. On going fidelity monitoring and certification 5. Agency or administrative support (p. 2) Although this study represents only one particular area of mental health research, researchers in this field have been investigating implementation fidelity since the mid as a lot to learn. In a similar fashion, this proposed study hopes to demonstrate that when teachers are provided adequate resources and support combined with observations and feedback they can effectively implement evidenced based practices into the class room. Another notable mental health study by Mokrue, Elias, and Bry (2005) investigated the effectiveness of a video series, Talking with TJ which is intended to be used by teachers to encourage the development of positive social and emotional skills wit h urban, predominantly minority, elementary school children. The sample for the
54 study was 655 second and third graders from 30 classrooms at six urban elementary schools in the Plainfield, New Jersey, School District. The descriptive information from the d istrict revealed that at the time of the study 60% of the students qualified for free or reduced lunch and the district had a high percentage of students of color with 82% African American, 14% Hispanic, and 4% other. The instruments used, in addition to a demographic information intake sheet, were the Social Skills Rating Scale (SSRS) developed by Gresham and Elliot in 1990 to measure social competence, the Piers Concept Scale developed by Piers and Harris in 1984 to measure self con cept, and a Teacher Implementation Survey designed by the authors to measure the degree of fidelity of each instructor. Reliability and validity information was provided for both the SSRS and the Piers Self Concept Scale. The SSRS was rep orted to have an internal consistency reliability of .96, a test retest reliability of .68 to .87, and validity of .75 to .81 when compared to the Child Behavior Checklist developed by Achenbach and Edelbrock in 1983. The Piers Conce pt Scale was reported to have an internal consistency of .85 and a test retest reliability coefficient of .73. Because this program was part of the county required curriculum, before data collection could begin passive consent letters were sent home to the families that out option for them if they did not want their child to participate. Only one family chose for their child to not participate in the study. For data collection purpo ses, the student surveys were administered before and after the video intervention. To assist in the process, two trained research assistants went together to each of the 30 classrooms and administered the
55 student surveys using scripted directions during t wo 30 to 45 minute periods so as not to tire the young children. All items were read aloud to the students to accommodate for any reading difficulties in students. This process was then repeated after the students viewed and participated in the interventio n program, which lasted approximately four months. The teachers were also asked to complete SSRS forms on each student before and after the intervention. The teachers were paid $22 per hour for completing the surveys on their own time. Based on the mean of reported implementation fidelity scores, an ANCOVA procedure whil in high implementation groups had higher ratings of social skills and lower ratings of problem behaviors during post assessment period while their counterparts received Mokrue, Elias, & Bry, 2005, p. 68). It is important to note that the initial comparison of the reported, self co ncept scores revealed no significant differences between the two groups. Had the researchers not measured the fidelity of the teachers and conducted a second weighted analysis of the data they would have concluded that the program had no effect on the stud concept. The necessity for researchers to measure the fidelity of implementation is key to determining the true success of an instructional program. However, the results of this particular study should be interpreted with caution b ecause the researchers relied on self
56 reported data. Finally, the results are not generalizable to the majority of schools in America. The results can, in fact, only be generalized to other urban elementary schools. Summary From this research synthesis, i t is evident there are a limited number of published studies investigating the implementation of differentiated instruction with fidelity on urban middle school students. There does, however, seem to be a growing body of evidence that report success with u sing teacher support/ study groups to assist teachers with implementing instructional practices learned during professional development inservices. Furthermore, although investigations with differentiated instruction as a complete philosophy and instructio n model began with the highly able, gifted learners, recently more researchers have begun to focus on its use and ability to effect the more diverse populations including students with special needs and struggling learners. Currently there is also a predom inance of qualitative studies being conducted in this area, precipitating a need for more quantitative and mixed method studies. Data remain scarce regarding the effect of differentiated instruction on student achievement. Finally, while it may be apparent that teachers are receptive to utilizing a supportive differentiated instructional philosophy, the bottom line is that until there is a body of evidence illustrating its impact on academic achievement and a viable approach to measuring instructional fidel ity, district personnel and administrators will be reluctant to support its implementation.
57 Chapter Three Method Overview The goals of this study were (1) to investigate the effect of differentiated instruction on the mathematics and reading achieve ment of urban, middle school students; (2) to monitor teacher fidelity to the differentiated instruction model; (3) to assess the effect of facilitated support groups on teacher fidelity; and (4) to evaluate the relationship between teacher implementation of differentiated instruction and student achievement scores. This study incorporated data through a mixed methods design that evaluated the effectiveness of facilitated teacher support groups on the implementation of differentiated instruction in two urba n, middle school settings. In addition, this study contributes to the limited body of research that addresses classroom implementation of design, population and sample, variables, measurement tools, data collection procedures, and data analysis. Pilot Data Overview of pilot. During the previous year, the principal investigator conducted a pilot to this research study at an urban, middle school in Florida. The purpose of t he pilot was to develop and field test the DI: FIT observation tool, field test a facilitated support
58 group, and collect student achievement data to determine the feasibility of the implementation of the research design. The pilot had a design similar to t he current study in that the treatment and control groups were based on interdisciplinary team membership. There were four teams of five teachers each, two seventh grade and two eighth grade. Special effort was made to insure that the groups were similar i n size and demographics. The study was conducted over the entire school year and the student achievement data were assessed using the students individual FCAT Developmental Scale Scores (DSS). These ordinal scores are based on the FCAT Scale Scores and po ssible scores range from 0 to 3000. As with the current study, teacher fidelity to the differentiated instruction philosophy was encouraged and supported through a facilitated differentiated instruction support group, access to a resource library, and clas sroom fidelity observations using the DI: FIT once each nine weeks. DI: FIT observation tool The ten teachers in the treatment group were observed four times each by trained observers and the correlation statistic for the DI: FIT observation tool was calc ulated to be .86 ( p < .0013) with a 95% confidence interval ranging from .67 to .96 [CI 95 = (.67, .96)]. A correlation statistic of .85 or greater is considered good (Cohen, 1992) Student achievement. Because the middle school students had a wide range of initial academic abilities, student achievement was measured as their improvement or change over the course of the year with respect to their FCAT Developmental Scale Change Score, posttest minus pretest. With regard to the student academic achievement pi lot data, the results showed a great deal of promise for future studies.
59 Because the reading and mathematics achievement scores are individual, discrete scores, they were analyzed separately. Students with missing data were removed from the data set prior to analyses. The reading and mathematics FCAT Developmental Scale Change Scores were analyzed using a 2x 2 factorial ANOVA by treatment level and by grade, with an alpha level of less than .05 considered significant. Student reading data. The descriptive statistics of the four reading subgroups are presented in Table 1 and illustrate that the four groups were similar in size and distribution with a total sample size of 353. Figure 3 shows the side by side comparison tal scale mean change score by group, from which it can be easily seen that the students whose teachers were part of the treatment group improved their scores more than the students whose teachers were part of the control group at both grade levels. The an alysis of variance data (see Table 2) further support this statement as the treatment effect was found to be statistically significant (F(1, 349) = 5.41, p 1992) effect size ( f ) wa f is calculated by taking the square root of the product of the degrees of freedom times the F statistic divided by the total sample size: f f of .1 is considered small, .25 medium, and .4 large. Student mathematics data As with the reading data, the descriptive statistics of the four mathematics subgroups are presented in Table 3 and illustrate that the f our groups were similar in size and distribution with a total sample size of 353. Figure 4 shows the side by scale mean change score by group, from which it can be easily seen that the student s
60 whose teachers were part of the treatment group improved their FCAT Developmental Scale Scores more than the students whose teachers were part of the control group at both grade levels. The analysis of variance data (see Table 4) further support this sta tement as p = f ) was calculated to be .17, which is considered to be a small to medium effect size Table 1 Descriptive Stati stics of FCAT Reading Developmental Scale Change Scores (N = 353) Statistic 7 th grade with support (n= 92) 7 th grade without support (n = 92) 8 th grade with support (n = 78) 8 th grade without Support (n = 91) Mean 163.6 122.5 149.3 113.1 Median 158.0 141 .5 115.0 104.0 Range 892.0 805.0 704.0 728.0 Interquartile Range 212.5 202.5 166.0 199.0 Standard Deviation 169.2 160.0 148.1 142.9 Skewness .3 .3 .3 .2 Kurtosis .5 .1 .1 .1 Standard Error Mean 17.6 16.7 16.7 15.0
61 Figure 3. FCAT reading developmental scale change scores by grade level (N = 353). Table 2 Analysis of Variance Summary Table for FCAT Reading Developmental Scale Change Scores (N = 353) Source D f SS MS F P Grade 1 12318.7 12318.7 .51 .48 Treatment 1 131221.3 131221.3 5.41 .02* Grade x Treatment 1 505.4 505.4 .02 .88 Within Group (Error) 349 8462321.2 24247.3 p < .05
62 Table 3 Descriptive Statistics of FCAT Mathematics Developmental Scale Change Scores (N = 353) Statistic 7 th grade with s upport (n = 92) 7 th grade without support (n = 92) 8 th grade with support (n = 78) 8 th grade without support (n = 91) Mean 152.5 119.5 123.2 101.3 Median 137.0 124.0 103.5 98.0 Range 557.0 828.0 555.0 551.0 Interquartile Range 112.5 162.5 140.0 107.0 Standard Deviation 103.1 132.0 121.0 89.2 Skewness .6 .1 .9 .3 Kurtosis 1.4 1.1 .4 1.7 Standard Error Mean 10.8 13.8 13.7 9.4 Figure 4. FCAT mathematics developmental scale change scores by grade level (N = 353).
63 Table 4 Analysis of Variance Summary Table for FCAT Mathematics Developmental Scale Change Scores (N = 353) Source D f SS MS F p Model 3 123607.4 41202.5 3.27 .02* Grade 1 49729.5 49729.5 3.95 .04* Treatment 1 660.68.5 66068.5 5.24 .02* Grade x Tr eatment 1 2704.6 2704.6 .21 .64 Within Group (Error) 349 4396704.2 12598.0 p < .05 Summary of pilot. The success of the pilot was critical in laying the groundwork for this study. It demonstrated that the DI: FIT was a viable tool for assessing teache r fidelity to the differentiated instruction model and it provided preliminary evidence that the use of differentiate d instruction strategies could a ffect student achievement. Further, it contributed to the development of the current support group model. Research Design A mixed methods design, with a quasi experimental design in the quantitative component, was utilized to evaluate the multiple themes of this research study over a five month period. First, qualitative and quantitative methods were used to investigate the impact of facilitated teacher support groups, teacher reflection, and fidelity observations triangulation of data from facilitated support group min
64 group model on teacher implementation fidelity (see Figure 5). The use of multiple end of the five month period, an ANOVA procedure was conducted to quantitatively determine the relationship between the degree of teacher implementation of differentiated instruction and student achievement. Figure 5. Supporting and assessing teacher implementation fidelity: Triangulation of teacher data. Facilitated Support Groups Teacher Journals Differentiated Instruction: Fidelity Implementation Tool (DI: FIT) Teacher Implementation Fidelity Fidelity Perspective Perspectives
65 Sample selection and assignment. During this second research phase, two matched urban, Title I middle schools were purposively selected to serve as research sites. First, permission was received from the district research and compliance office to use the school sites and then the princip al at each site was contacted to explain the study and gain permission. Once permission and support were obtained from each principal and the appropriate IRB permissions were obtained, the study was explained to both faculties. In School A 28 teachers vol unteered to participate and in School B 27 teachers volunteered. The participants were divided into a treatment group and control group based on currently existing interdisciplinary teams within each school because these teachers would be in daily contact with each other and often share strategies. This method of assignment was utilized because it would reduce the amount of cross contamination of the treatment. When determining which teachers would purposely be assigned to the treatment group, preference w as given to the content area teams that had the greatest number of students with special needs included in their general education classrooms. The number of teachers selected was distributed among the three grade levels within each school and kept as balan ced as possible. The remaining teachers were matched by grade level and content area within each site and assigned to serve as the control group (see Figure 6). After much discussion and research, this type of assignment was determined to be the most effec tive because a school vs. school comparison would have produced data with poor reliability and validity. Assigning a treatment and a control group within each site was desirable because it minimized possible extraneous variables and the nesting
66 effect of i ndividual school factors, such as school wide reading programs, a strong administration, and extended learning programs. School A School B 28 teacher participants 27 teacher participants 15 13 13 14 treatment control treatment control participants participants participants participants 5 (6 th Grade) 4 (6 th Grade) 4 (6 th Grade) 4 (6 th Grade) 5 (7 th Grade) 5 (7 th Grade) 5 (7 t h Grade) 6 (7 th Grade) 5 (8 th Grade) 4 (8 th Grade) 4 (8 th Grade) 4 (8 th Grade) Figure 6. Design of teacher participants assignment by school and grade level (N = 55). The demographics of each group were kept as balanced as possible. This is espe cially important because urban, Title 1 schools typically have a high teacher turnover and a large percentage of new teachers. The number of new teachers were dispersed among the treatment and control groups as much as possible and will be addressed in the results section. Facilitated support group model. As part of the study, the teachers in the treatment and control group first attended the standard, district, seven hour differentiated
67 instruction in service workshop. During the next five months of the s chool year, each teacher in the treatment group attended five monthly, two hour, facilitated support group instruction book study in that each teacher has access to r esources, support from the facilitator and the other members of the group, lesson and strategy support, and they could earn in service points. The fifth and final meeting was a focus group which assisted the facilitator in obtaining qualitative feedback on the support group model. The groups began with the typical formalities of introduction; and then after group norms were established, i.e., being on time, taking roll, respecting the opinion of others, and bringing the teacher reflective journal to each m eeting so they could take notes and make connections to their classroom instruction, the format for the sessions was standardized. Each meeting began with teachers sharing their classroom differentiated instruction experiences and providing feedback to the group in a round robin fashion. Teachers were encouraged to share both successes and challenges so others could learn from them. Teachers also discussed future differentiated instruction lessons to get ideas from the group. The researcher acted as the gro members had an opportunity to speak. Each month, the facilitator also made sure that the meeting space was reserved and arranged in a man discussion. The facilitator also provided the participants with snacks, refreshments, and an article that highlighted various differentiated classroom strategies in order to facilitate discussion and expand their teach ing repertoire. The table portion of the session closed with questions from the participants in a round robin style and a reminder of the next
68 instruction resource library and select a new differentiated instruction resource and/or strategy book (see Appendix E). minutes and teacher comments. In order to ensure the reliability of the data coll ected at the facilitated support group sessions, the minutes were e mailed out on the following day for verification. All of the teachers in the treatment group were also provided with a Facilitated Support Group Feedback Form (see Appendix F) after each s ession. The form served two purposes, to verify that each member read the minutes and to obtain written feedback from each member regarding the accuracy of the minutes. If there were any changes, the minutes were amended and sent out again to e nsure consen sus by all of the members. This process was repeated until all members felt that the minutes were accurate and served as a member check. As a backup, the sessions were digitally recorded to insure the accuracy of the information and comments collected. If there was any information was used in the final report to assure the anonymity of all participants. This method was field tested during the pilot study and it was found t hat it provided the necessary information needed for the purposes of this study. At the final group meeting, the facilitator asked the participants to reflect and discuss what they liked and what they would like to change about the facilitated support grou p sessions and to suggest ways to improve upon the support group model/format for the upcoming year. The purpose of the differentiated instruction resource library for the teachers in the treatment grou ps was to
69 provide an additional opportunity for teachers to find new ideas and strategies and to At the conclusion of each meeting, participants traded in their book fro m the previous month and selected a new resource book to review for at least 30 minutes. They were encouraged to implement at least one new differentiated instruction strategy during the next three to four weeks in their classroom. In addition, teachers in the treatment group maintained implementation journals based on their differentiated instruction experiences. Although the teachers in the comparison groups have been exposed to the elements of differentiated instruction through standard district in servi ce opportunities, they did not receive any of the treatment interventions or additional supports ( i.e., feedback from observations, participation in support groups, copies of minutes, use of reflective journals, or access to the reference library. ) Teacher observations. Prior to beginning the observations, each observer was pre trained using tapes and the process was practiced until each rating team reached an inter rater reliability of .85 or better. In order to compare fidelity of the two groups, the pri ncipal investigator and research assistant observed and assessed the teachers in both the treatment and the comparison groups each nine weeks to determine the degree of fidelity that each teacher demonstrated with respect to differentiated instruction usin g the Differentiated Instruction: Fidelity Implementation Tool (DI: FIT). Inter rater reliability was re checked each nine weeks. If the correlation statistic between the trained observer and the principal investigator fell below .85 then the observing tea m participated in a re training program with additional observations and follow up discussions until a correlation of .85 or better was obtained. Additional training on the use of the DI: FIT
70 assessment tool was conducted if necessary. In this way, the res earcher was able to empirically examine the difference in the differentiated instruction implementation scores of the two groups of teachers. This comparison was especially important because some teachers who have attended cooperative learning workshops ma y utilize some form of group work and/ or learning profile accommodations which does not make the lesson differentiated, but to the untrained observer it may look like differentiated instructional strategies. Population and Sample The targeted sample for t his study was urban middle school teachers and students in Florida. Specifically, this research project was designed to provide strategies, support, and assessment tools for teachers who use a team approach and support the inclusion of diverse student popu lations especially students with disabilities. The students selected for this study represent the diverse populations of students who typically live in the inner city areas of large urban cities. Large percentages of these students are usually from low in come families and are primarily minorities. Further, the student population included disabilities. For the current research study, data were collected at two urban middle sc hools. The two schools were selected because they were closely matched based on 14 different requirements for the past three years. The two middle schools selected met each of the following criteria:
71 Teach students in sixth through eighth grades Student enrollment with approximately 500 students Located within the central portion of the school district, inner city Located within 10 miles of each other More than 80% of the students receive free lunch, Title I More than 80% of the students are identified as minority More than 10% of the students receive exceptional student education services Less than 50% of the students met the state high standards in reading (3 or above on the FCAT) Less than 50% of the students met the state high standards in mathematics (3 or above on the FCAT) More than 40% of the students did not make gains in reading More th an 30% of the students did not make gains in mathematics More than a third of the fa culty has less than three years experience The school did not meet the federal requirements for Adequa te Yearly Progress Within each of the two school sites, three grade level teams of teacher participants were selected based on interest and assigned to the treatment group and three teaching teams of matched teachers were assigned to the comparison group. This method provided a total sample of 28 teacher participants in the treatment group and 27 in the control group. All of the grade level groups were relatively balanced with the largest having six members and the smallest group having four members. Furth er, there were a large
72 number of students in the sample. Because each of the six teaching teams provided instruction for a pool of approximately 80 students, approximately 480 students were part of the treatment group pool and 480 students were part of the comparison group pool, for a total estimate of 960 students. Some students were excluded because of excessive absences (>21 days during the study period) or missing data. Table 5 Descriptive Statistics of the Student Population who Participated in the Sta te Standardized Assessments at the Two School Sites During the 2006 2007 School Year (N = 1026) School Number of Students Enrolled in 2006 2007 % of Students on Free or Reduced Lunch % of Minority Students % of Students with Disabilities % Meeting High S tandards in Reading % Meeting High Standards in Math % Making Reading Gains % Making Math Gains % of Lowest 25% Making Learning Gains in Reading School Grade 2004 2005 2006 A 519 85 86 11 42 49 53 66 59 C B C B 507 92 87 12 30 34 53 56 68 C C C Source. Data based on information reported by the FLDOE, retrieved August 16, 2007, from http://schoolgrades.fldoe.org/default.asp Independent Variables Experimental group, level one: Treatme nt group The treatment group consisted of the middle school teacher participants who received a seven hour differentiated instruction in service workshop, which emphasized philosophy, underlying research, strategies, videos, and logistics of a differentia ted classroom. The workshop was followed by five monthly facilitated support group sessions, on going teacher support (in the form
73 of a mentor), and implementation observations using the DI: FIT with constructive feedback. There were a total of 28 teachers in the treatment group, 15 at School A and 13 at School B. Experimental group, level two: Control group. The control group consisted of middle school teacher participants who did not participate in any of the differentiated support activities but still p articipated in all standard whole school activities. The control group of teachers taught their classes without the benefit of the facilitated teacher support group. Each teacher in the comparison group was observed at least once per nine weeks using the D I: FIT. The student achievement scores and teacher DI: FIT scores will serve as a comparison with the student achievement and teacher fidelity score of the treatment group during the analysis phase of this study. There were a total of 27 teachers in the co mparison group, 13 at School A and 14 at School B. School site. Thi s nominal assigned variable consist ed of two urban middle were analyzed first within each school site and then pooled between the two school sites. Grade level. This o rdinal assigned variable was limited to sixth, seventh and eighth grade designation. Dependent Variables DI: FIT fidelity observation score. Once each nine weeks trained observers using th e DI: FIT observation tool observed each teacher in the study. Each observation produced a discrete, ordinal observation score ranging from 0 to 20 and served as a numerical representation of the number of observable differentiated instruction
74 instructiona l strategies utilized by each teacher during the observation and follow up conference. FCAT Mathematics Predictor Test scores requirement for schools to assess and report the continuous progress of their students, espec ially the lower performing students, the district now requires all schools to assess and report reading and mathematics progress scores on all students. The FCAT Mathematics Predictor Tests were administered in January (Form B) and in May of 2007 (Form C). The FCAT Mathematics Predictor tests yield a discrete, interval percentile score ranging from 1 to 99. A pre post comparison was calculated using the student scores from Form B and Form C, respectively. Stanford Diagnostic Reading Test Scores. The Stanfor d Diagnostic Reading Test Fourth Edition (SDRT 4) (Psychological Corporation, 1995) was automatically administered in September by district personnel to all middle school students. This score was utilized as a pre test reading score. These same students were then re tested in April to determine their growth. This score was used as the post test reading score. The assessment report provided each student with a discrete, interval percentile score ranging from 1 to 99 and an approximate grade level equivalen t. This benchmark test score is has met the district benchmark for promotion. Instruments/ Measurement Tools Differentiated Instruction: Fidelity Implementation Tool ( DI: FIT). The DI: FIT assessment tool was developed and field tested during the first phase of this research project in 2005. This observation tool consists of 20 differentiated instruction indicators
75 ption (see Appendix D). The collaboration, and lesson planning. In the development of this instrument, a literature search was first conducted to see if any differe ntiated instruction fidelity observation tools were already developed and available. From this investigation, two observation tools developed by leaders in the field were located. The first, The Differentiated Classroom Observation Form (see Appendix G), w as developed by Chapman and King (2005) and the second, the Teacher/ Peer Reflection on Differentiation (see Appendix H), was developed by Tomlinson and Allan (2000). Both of these evaluation tools allow the observer to mark on a scale the degree to which a teacher is implementing a particular strategy or demonstrating a behavior. Chapman and King (2005) further subdivided their tool into the following areas: tools were not available, both of these instruments provide the user with a great deal of For the purposes of this study, a dichotomous observation tool was desired so multiple observers would be able to obtain a higher degree of agreement on the same observation. Because there is no middle ground, it requires the observer to select that the indicator was either evident or not evident. A higher inter rater reliability will help increase the overall rel iability of the observation data and reduce the chances of observer
76 differences. The dichotomous fidelity observation instrument entitled Fidelity Instrument for Measuring the Use of Evidence Based Academic Strategies in Special Education Classrooms (2005) was used as a guide for layout and wording that would be specific and measurable Based on a review of the previously mentioned instruments and the field experiences of differentiated instruction district trainers, a new instrument was developed, the DI: FIT. Next, the DI: FIT was submitted to several professors in the College of Education at the University of South Florida. Specific feedback and suggestions were obtained regarding language, operational descriptions, and feasibility. Based on this expert f eedback, several items were changed and specific numbers were added to make items more quantifiable and observable. During this revision, two of the ded because there may not be an opportunity to observe those two indicators in all lessons. Later, the DI: FIT was submitted to Tomlinson at the University of Virginia. Again, specific feedback was requested concerning each item, the overall structure, and the validity item content. She suggested clarifying the terminology of tiered lessons and incorporated into the final version. In order to assess teacher fidelity t o the differentiated instruction model, a teacher who previously attended a differentiated instruction professional development workshop presented by the researcher was trained on how to use the instrument and specific evidence for each indicator was discu ssed. Then the researcher and the trained teacher
77 observed a differentiated lesson together, completing the instrument independently. After the observation a follow up interview with the teacher who taught the lesson was conducted by the two observers. Th en the two observers compared their completed instruments. On the first comparison there were three discrepancies, which were discussed until agreement was reached. A second classroom observation was scheduled and the process was repeated. This time the t wo observers only differed on one indicator. This item was discussed at length until both observers felt confident that they completely agreed on how to code this item during future observations. During the pilot study, the two researchers then used the in strument to observe ten teachers participants. Because the teachers were several months into the pilot, the scores were high ranging from 16 to 19 out of a possible 20 points. The correlation statistic for the DI: FIT was calculated to be .86 ( p < .0013) w ith a 95% confidence interval ranging from .67 to .96 [CI 95 = (.67, .96)]. A correlation statistic of .85 or greater is considered good (Cohen, 1992) FCAT Mathematics Predictor Tests. The FCAT Predictor Tests, Form B and Form C, were developed by the Stat e of Florida to assess the Sunshine State Standards so the districts and schools could assess, monitor, and report the academic progress and skills of students as part of the continuous progress model. The items were modeled after the FCAT and many items a re directly taken from previous test versions. The two tests are parallel test forms and each test consists of 25 items (24 multiple choice and 1 short response/ think, solve, and explain). The standardized mathematics assessment items s ability to respond to items that test number sense, number and operations, algebra, geometry, measurement, data analysis and probability, problem
78 solving, and reasoning as defined by the National Council of Teachers of Mathematics (2005). According to th and Accountability Briefing Book (2005), the FCAT, a standards based test, was described in the Sunshine State Standards. In order to ensure the content validity of FCAT, the Florida Department of Education (2004) implemented the following steps for all FCAT items: Educators and citizens judged the standards and skills acceptable. Item specifications were writte n. Test items were written according to the guidelines provided by the item specifications. The items were pilot tested using randomly selected groups of students at appropriate grade levels. All items were reviewed for cultural, ethnic, language, and g ender bias and for issues of general concern to Florida citizens. Instructional specialists and practicing teachers reviewed the items. The items were field tested to determine their psychometric properties. The tests were carefully constructed with ite ms that met specific psychometric standards. The constructed tests were equated to the base test to match both content coverage and test statistics. (p. 26)
79 In May 1996, the Florida Department of Education contracted CTB/ McGraw Hill to develop the origin al form of the FCAT test for grades 4, 5, 8, and 10. In 1999, the Harcourt Educational Measurement Company was hired to develop the test for grades three through ten. In addition to the use of commercial testing companies and the establishment of the previ ously mentioned standards for test items, Florida DOE personnel collaborated with practicing Florida educators (e.g., teachers, curriculum specialists) in an effort to promote strong content validity across both the criterion referenced and norm referenced measures of the FCAT. Correlations between .70 and .81 were obtained for students tested in the aforementioned grades (Florida Department of Education, 2004). Four kinds of reliability coefficients were used in the development of the FCAT: internal con sistency, test retest reliability, inter rater reliability, and reliability of classifications. For any measure of reliability, the reliability coefficient can range from zero to one (0.0 1.00), with a zero score showing a lack of reliable results and a on e reflecting extremely consistent results. The most common ly used measure of reliability with the FCAT is internal consistency, because it involves utilizing only one test administration per student. Internal consistency reliability is reported for the FCA T using Richardson 20 (KR 20) p. 25). The KR 2 component, which scores items between 0 and 4, and the KR 20 for the NRT comparison
80 part. The data on Mathematics SSS and NRT show strong reliability coefficients between .80 and .90. Stanford Diagnostic Reading Test Fourth Edition (SDRT 4) (Psychological Corporation, 1995) Karlsen and Gardne r, two leaders in the field of assessment, updated and re normed the SDRT in 1995. The resulting SDRT 4 was developed using the highest diagnostic standards. It can be administered to groups of students or individually. The test administration for the midd le school level takes 85 minutes. All test items are in a multiple choice format. The score report provides both criterion referenced and norm comprehension skills. The diagnostic report provides sub scores with reference to the comprehension of functional and recreational reading material, and reading rate. Reliability of greater than .85 was reported using internal consistency measures. Test validity of greater than .85 was determined using the OLSAT 8. Data Collection Procedures Facilitated support groups and focus group. Over the course of a five month period, each teacher attended monthly, 120 minute, facilitated support group sessions. and teacher comments. In order to ensure the reliability of the data collected at the facilitated support group sess ions, the minutes were e mailed out on the fol lowing day for verification and a feedback form was put in box Then, each participant was asked to respond if he/she felt the i nformation was accurate, saw any changes, and/or had suggestions fo r the next meeting. If there were any changes, the minutes were
81 amended and sent out again to insure consensus by all of the members as a member check. As a backup, the sessions were recorded to insure accuracy of the information and comments collected. Th e facilitator also kept a reflection journal that was completed immediately after each session. In the reflective journal, the facilitator recorded specific events that might be important, connections that participants made during the meetings, notes on ho w to improve the process, and any general them es or emergent meaning that beca me apparent during the facilitated support group sessions. At the final group meeting a focus group was conducted, the session was taped, and later transcribed. The facilitator provided a question/graphic organizer (see Appendix I) for the participants to c apture what the teachers felt were the most and least valuable aspects of the facilitated support group sessions, suggested changes for future groups, and feedback on their ove rall differentiated instruction implementation experience. Teacher journals. During the five month period, each teacher maintained an implementation journal based on their experiences with the differentiated instruction philosophy and lessons. These journa ls helped to provide an alternative avenue for teachers to provide personal feedback on the study to the researcher. The y were encouraged to write in a free response style in the books after concluding differentiated lessons, glue examples of student work, comment about the support group meetings, note strategies and lesson ideas for future use, and to reflect on how their students responded to differentiated lessons. The facilitator monitored the journals at the monthly meetings to ensure that the teachers were maintaining them. One teacher really liked expressing herself using the journal format and her journal was used as an example for other participants. The journals were collected at the last teacher support group for analysis.
82 DI: FIT teacher observa tions. Once per nine weeks, each teacher participant, from both the treatment and comparison groups, was observed using the DI: FIT observation tool. In addition to the researcher, two resource specialists, one at each school site, were trained to observe the teachers in the study. First, the observers attended a seven hour training, taught by the researcher, so consistency of information could be maintained. Then, the observers practiced simultaneously observing and scoring a teacher who was not part of th e study. Following the observation, the scoring of each indicator was compared and any discrepancies were discussed until everyone agreed. This process was repeated until the inter rater reliability of the observers was greater than or equal to .85. Throug hout the study, if the DI: FIT observation scores ever differed by more than a point then the observers repeated the inter rater reliability process to insure consistency and the validity of the scoring process. FCAT Mathematics Predictor Tests Form B an d C and Stanford Diagnostic Reading Test Fourth Edition. post data were collected on all students whose teachers were participants in the study as procedure. All student assessment tests were supervised and administered following the state required procedures. All students were supervised to insure independence of the results. Once the student scores were obtained at each school, school personnel sor ted the data by grade and by team and removed the student identifiers before providing the deidentified data to the researcher. Students who did not have achievement scores available from both the pre and post tests were removed from the data set, because they were not present for the entire treatment period. In addition, a high degree of student absence could possibly limit the potential for
83 academic improvement due to the treatment. Therefore, because the state of Florida considers students who miss 21 or more days of school in one year excessive and a criteria for determining good schools, students who missed more than 21 days of school during the treatment period were also removed from the data set because they would have had limited exposure to the trea tment. Confidentiality All written data, audiotapes, and videotapes were anonymously coded and stored professor had access to the data. The researcher maintained physical possession of the data and ensured the safety of participants and confidentiality of the data. Data were safely stored after each observation and monthly facilitated support sessions. The signed informed consent forms (see Appendix J) will be stored for t hree years in a secured file these documents will be shredded or destroyed. In accordance with the IRB 2006 requirements: The results of this study may be published However, the data obtained from the participants will be combined with data from others in the publication. The published results will not include names or any other information that would personally identify the participants in any way. Furthermore, the privacy and research records of all participants will be kept confidential to the extent of the law. Authorized research personnel, employees of the Department of Health and Human Services, and the USF Institutional Review Board, and any other
84 individuals acting on behalf of USF, may inspect the records from this research project. Data Analysis Question 1: What were the effects of differentiated instruction with teacher support during a five month period on the academic achievement outcomes of urban, Title I, middle school students? A quasi experimental design was used to quantitatively compare the impact of differentiated instruction strategies on the reading and mathematics achievement scores of middle school students whose teachers were part of the inter vention or comparison groups. This analysis was accomplished using the pretest posttest data as measured by the FCAT Mathematics Predictor Tests Form B and C and Stanford Diagnostic Reading Test Fourth Edition The scores were then analyzed using an 2 (t reatment/ comparison) x 3 (6 th grade/ 7 th grade/ 8 th grade) factorial Analysis of Variance (ANOVA), the level of significance was set at .05 using Statistical Analysis System (SAS). The data from students who were not present for the entire study and did n ot have both pre and post test scores available were removed from the data set. Students who missed 21 or more days of school during the treatment period were also removed from the data set. Finally, the f effect size was calcula ted on comparisons that were found to be statistically significant (see Table 6).
85 Question 2: What were the statistical differences among teacher groups who participated in facilitated support groups and those who did not with respect to their impleme ntation of differentiated instruction as measured by the Differentiated Instruction: Fidelity Implementation Tool (DI: FIT) observation tool? In this quantitative analysis, the DI: FIT observation scores at Time 1 and Time 2 for all teacher participants s eparated by treatment, grade, and school and then basic descriptive statistics were run in order to examine the differences among all groups (see Table 6). Question 3: implementatio scores? In order to answer this question, the mean teacher DI: FIT observation scores were analyzed by treatment group and school site using a 2x 2 factorial Analysis of Variance (ANOVA). Thi s analysis of the DI: FIT observation scores provided information as to whether on not a statistical difference existed among the teacher differentiated instruction implementation fidelity scores by treatment group and by school. The Statistical Analysis S ystem (SAS) software was utilized for this analysis with f effect size was calculated on comparisons that were found to be statistically significant idelity implementation scores were compared to the reading and mathematics achievement scores of their students. Specifically, treatment and
86 DI: FIT observation scores analysis was completed using a SAS correlation procedure. Correlations that were significant at the .05 level would suggest that a relationship exists between implementation fidelity and student achiev ement scores (see Table 6). Question 4: Using qualitative data and feedback provided by the teachers in the treatment model and their instructional growth? This final question r equired the researcher to complete a qualitative analysis on the final focus group session. Using an inductive analysis in conjunction with a document review, all data were reviewed, meaningful units were identified, units of data were coded, and then the data were categorized in order to identify basic themes. Then, the data were furth er reduced through a constant comparison, a consolidation of any redundant categories, and an analysis of emergent themes. These multiple methods of data collection allowed for method and data triangulation and increased credibility (see Table 6).
87 Tab le 6 Table of Research Questions, Data Collected, and Analyses Conducted Research Question Data Analyses Question 1: What were the effects of differentiated instruction with teacher support during a five month period on the academic achievement outcomes of urban, Title I, middle school students? Student achievement mathematics and reading scores at Time 1 and Time 2 separated by treatment group, grade, and school Descriptive analysis of demographic and achievement data T 2 T 1 = Change Score T tes t analysis of T 1 scores between school sites and within sites to ensure that data are not statistically different before secondary analysis 2x3x 1 Factorial ANOVA f ) (Table continues)
88 Table 6 (Continued) Research Question Data A nalyses Question 2: What were the statistical differences among teacher groups who participated in facilitated support groups and those who did not with respect to their implementation of differentiated instruction as measured by the Differentiated Instr uction: Fidelity Implementation Tool (DI: FIT) observation tool? DI: FIT observation scores at Time 1 and Time 2 for all teacher participants separated by treatment, grade, and school Descriptive analysis of demographic and observation data (T 1 + T 2 ) / 2 = Mean Observation Score Question 3: What was the relationship differentiated instruction implementation scores as measured by the DI: FIT and the student achievement scores? DI: FIT Score Student s change scores by teacher 2x2x1 Factorial ANOVA f ) Correlation analysis R 2 (Table continues)
89 Table 6 (Continued) Research Question Data Analyses Question 4: Using qualitative data and feedback provided by the teachers in the treat ment groups, what were the facilitated support group model and their instructional growth? Monthly minutes Individual Teacher Feedback Forms Teacher Journals Journal Focus Group minutes and transcripts Documen t review Inductive analysis Constant comparative analysis Identification of emergent themes
90 Chapter Four Results Overview As previously stated, the goals of this study were (1) to investigate the effect of differentiated instruction on the mathematic s and reading achievement of urban, middle school students; (2) to monitor teacher fidelity to the differentiated instruction model; (3) to assess the effect of facilitated support groups on teacher fidelity; and (4) to evaluate the relationship between th e degree of teacher implementation of differentiated instruction and student achievement scores. However, prior to the analysis of the achievement data, a complete investigation of the demographics of the teacher and student samples at the two school sites had to be completed to ensure a like comparison of data. Further, the achievement data at Time 1 from the students had to be statistically compared to determine if the results could be compared within each site and/or combined between sites. Once all of t he teacher fidelity observations, support group meetings, and student pre and posttests were completed, then began the task of entering and analyzing thousands of achievement scores. The meticulous data entry process began by sorting scores by school, by g rade, and by treatment (teacher teams). Next, any students who missed more than 21 days of school during the treatment period or had missing data due
91 to an absence or attrition, being withdrawn, were removed from the data set before any further analyses we re completed. This resulted in the data from 46 students being removed from School A, leaving a total of 473 participants, and 74 students from School B, leaving a total of 433 participants. In order to ensure the accuracy of the data entry and increase re liability, the SA S reports were triple checked, t wice by the researcher and once by the trained teacher observer who had been providing support throughout this study. The descriptive data analyses and ANOVAs were all completed using the 2007 version of the Statistical Analyses Software (SAS, Release 9.1). The results are presented in several sections: demographic statistics of participants, student achievement data analyses, teacher fidelity observation data analyses, interaction between fidelity and achiev ement, and support group analyses. Demographics of Participants Student demographics. It is important to insure that the student populations of the two school sites were as similar as possible for future analyses. To assess the demographic composition of the student participants at each school a detailed frequency and percentage analysis was conducted with the assistance of school personnel. First, the data were analyzed by school and then pooled so a population comparison could be completed (see Table 7). The student demographics are remarkably similar, except that school site B has a heavier percentage (12.4%) of 6 th graders, 7.5% less 8 th graders, and 6.2% more students who are categorized as economically disadvantaged. Next, the student samples within e ach school site were further subdivided and analyzed by treatment group.
92 Table 7 Demographic Characteristics of Student Participants: Frequency and Percentage by School Site (N = 906) Variable School A (n = 473) School B (n = 433) Total (n = 906) f % f % f % Grade 6 th 113 23.9 157 36.3 270 29.8 7 th 167 35.3 132 30.5 299 33.0 8 th 193 40.8 144 33.3 337 37.2 Gender Female 237 50.1 212 49.0 449 49.6 Male 236 49.9 221 51.0 457 50.4 Ethnicity Caucasian 65 13.7 58 13.4 123 13.6 African American 303 64.1 295 68.1 598 66.0 Hispanic 97 20.5 79 18.2 176 19.4 Asian 6 1.3 1 .2 7 .8 Native American 2 .4 0 0 2 .2 Economically Disadvantaged 402 85.0 395 91 .2 797 88 English Language Learners 35 7.4 36 8.3 71 7.8 Students with Disabilities 41 8.7 49 11.3 90 9.9
93 When the student participants were analyzed by group within each school site (see Table 8) the consistency of data continued. Again, there w ere relatively few differences among the samples. The percentages of student ethnicities, gender, students who were identified as economically disadvantaged, and students who were identified as English Language Learners were extremely similar across all fo ur subgroups. Because the class rolls of each teach could not be manipulated, some small differences were expected to naturally exist. The only marginal differences were (1) within School B there were slight percentage differences by grade level (a differe nce of 16.9% in 6 th grade and 10.3% in 7 th grade) due to random assignment within classes and (2) within both sites, there were differences in the percentage of students with disabilities assigned to the groups. This was a direct result of the treatment as signments provision that teaching teams who support inclusion of students with disabilities in the general education environment were given preferential assignment into the treatment groups. Because the majority of students with disabilities are functionin g below grade level, this added challenge of having a larger percentage of students with disabilities on their team meant that to show improvement when compared to the control groups they would have to increase their academic score significantly over their non disabled peers. Thus is the hope that the differentiated instructional philosophy can help bridge the achievement gap for students with disabilities in the general education environment and increase access to the general education classroom.
94 Table 8 Demographic Characteristics of Student Participants: Frequency and Percentage by Group (N = 906) Variable School A School B Treatment Group (n = 245) Control Group (n = 228) Treatment Group (n = 216) Control Group (n = 217) f % f % f % f % Gr ade 6 th 66 26.9 47 20.6 60 27.8 97 44.7 7 th 82 33.5 85 37.8 77 35.6 55 25.3 8 th 97 39.6 96 42.1 79 36.6 65 30.0 Gender Female 121 49.4 116 50.9 105 48.6 107 49.3 Male 124 50.6 112 49.1 111 51.4 110 50.7 Ethnicity Caucasian 33 13.5 32 14.0 28 13.0 30 13.8 African American 157 64.1 146 64.0 147 68.1 148 68.2 Hispanic 52 21.2 45 19.7 41 19.0 38 17.5 Asian 2 1.2 4 1.8 0 0 1 .5 Native America n 1 .4 1 .4 0 0 0 0 Economically Disadvantaged 208 84.9 194 85.1 199 92.1 196 90.3 English Language Learners 15 6.1 20 8.8 17 7.9 19 8.8 Students with Disabilities 35 14.3 6 2.6 36 16.7 13 6.0
95 Teacher demographics. Since the study had the approval and support of both the school district and principal at each site, getting teachers to volunteer was relatively easy. Of all the possible interdisciplinary subject area teachers who were eligible for the study only one opted not to participate, which eased the commencement of the study and did not impact the study because she taught students who were cross teamed and would not have been selected for assignment due to possible cross contamination of treatment. Next, in order to complete the demo graphic analysis of participants, basic demographic data were also collected on the teacher participants. After consents were signed, information regarding each grade level assignment, gender, ethnicity, certification status, and years of teachin g experience was collected and is displayed in Table 9. The grade level assignment of teachers, grades 6 8, was purposely balanced and the resulting grade level percentages ranged from 28.6% to 42.9%. As expected, all groups were comprised of predominate ly female teachers, ranging closely from 71.4% to 86.7%. Conversely, the percentage of male teachers ranged from 13.3% to 28.6%. With regard to teacher ethnicities, the predominate category was Caucasian for all four groups (57.1% to 69.2%), followed by Af rican American teachers (30.8% to 38.5%). Certification data indicated a possible advantage for the control groups because they had only one or zero teachers that were uncertified, as opposed to the treatment groups who each had two teachers who were uncer tified. Uncertified teachers either are teachers who are out of field, working on certification, or are in an alternative certification program. Likewise, the treatment groups had a slightly higher percentage of teachers with zero to three years experience Thus, no significant differences existed that would warrant caution when interpreting the final data.
96 Table 9 Demographic Characteristics of Teacher Participants: Frequency and Percentage by Group (N = 55) Variable School A School B Treatment Group (n = 15) Control Group (n = 13) Treatment Group (n = 13) Control Group (n = 14) f % f % f % f % Grade 6 th 5 33.3 4 30.8 4 30.8 4 28.6 7 th 5 33.3 5 38.5 5 30.8 6 42.9 8 th 5 33.3 4 30.8 4 30.8 4 28.6 Gender Female 13 86.7 10 76.9 10 76.9 10 71.4 Male 2 13.3 3 23.1 3 23.1 4 28.6 Ethnicity Caucasian 10 66.7 9 69.2 8 61.5 8 57.1 African American 5 33.3 4 30.8 5 38.5 5 35.7 Hispanic 0 0 0 0 0 0 1 7.1 Asian 0 0 0 0 0 0 0 0 Native American 0 0 0 0 0 0 0 0 Certified Yes 13 86.7 13 100.0 11 84.6 13 92.9 No 2 13.3 0 0 2 15.4 1 7.1 Teaching Experience 0 3 years 9 60.0 6 46.1 5 38.5 4 28.6 4 9 years 3 20.0 4 30.1 5 38.5 7 50.0 >10 years 3 20.0 3 23.1 3 23.1 3 21.4
97 Student Achievement Data Analyses Overview In the following sections, the summarized data and results of all student academic achievement analyses are presented in order to address the first research question: What were the effects of differentiated instruction with teacher support during a five month period on the academic achievement outcomes of urban, Title I, middle school students? Due to the large amount o f data, reading and mathematics results will be separated into two sections and then further subdivided according to the data, time, and procedures that were performed. Reading Achievement Analyses Within site comparisons at Time 1. In order to assess the effect of the differentiated instruction and teacher support groups on reading achievement, all students were administered the Stanford Diagnostic Reading Test Fourth Edition pretest and then comparative analyses on data were completed. First, the prete st results of the student achievement data at the beginning of the study, Time 1, were compared within each school site to demonstrate that the treatment and the control groups were not statistically different at the beginning of the study. The general des criptive statistics from both school sites at Time 1 portrays a relatively normal distribution of data with nothing remarkable to note (see Tables 10 and 11). Additionally, bar graphs of the mean, grade level reading test scores at Schools A and B by treat ment group illustrates the closeness of the data at Time 1 (see Figures 7 and 8). The range of group means were relatively close with from 41.7 to 44.2. Finally, an analysi s of variance on the same data within each school
98 site demonstrated that the data were not statistically different at School A (F(5, 466) = 1.01, p = .4085) nor at School B (F(5, 421) = 0.30, p = 0.9119) (see Tables 12 and 13). Table 10 Descriptive Statis tics of Reading Scores for School A: Mean, Standard Deviation, Skewness, Kurtosis, Minimum, and Maximum by Group at Time 1 (N = 472) Group N M SD Skewness Kurtosis Minimum Maximum Grade 6 with Support 65 47.2 13.0 .20 .54 23 76 Grade 6 without Support 47 50.8 11.9 .42 .08 30 85 Grade 7 with Support 82 47.3 14.6 .58 .35 21 81 Grade 7 without Support 85 49.2 15.2 .32 .37 21 86 Grade 8 with Support 97 48.9 16.4 .03 .66 16 86 Grade 8 without Support 96 51.2 14.2 .25 .56 20 79
99 Table 11 Descripti ve Statistics of Reading Scores for School B: Mean, Standard Deviation, Skewness, Kurtosis, Minimum, and Maximum by Group at Time 1 (N = 427) Group N M SD Skewness Kurtosis Minimum Maximum Grade 6 with Support 60 41.7 16.0 .44 .17 13 82 Grade 6 Without Support 97 43.1 18.4 .56 .15 11 91 Grade 7 with Support 77 42.1 12.0 .40 .31 11 76 Grade 7 Without Support 55 44.1 14.6 .19 .52 13 74 Grade 8 with Support 76 42.2 11.6 .34 .22 22 73 Grade 8 without Support 62 44.2 17.4 .01 .75 8 79
100 Figure 7. Mean grade level reading test scores at School A by treatment group at Time 1 (N = 472). Figure 8. Mean grade level reading test scores at School B by treatment group at Time 1 (N = 427).
101 Table 12 (N = 472) Source df SS MS F p Model 5 1077.9 215.6 1.01 .4085 Grade 2 313.0 156.5 .74 .4794 Treatment 1 744.1 744.1 3.50 .0620 Grade x Treatment 2 46.2 23.1 .11 .897 0 Within Group (Error) 466 99039.2 212.5 Table 13 (N = 427) Source df SS MS F P Model 5 354.1 70.8 .30 .9119 Grade 2 58.8 29.4 .13 .8823 Treatment 1 325.2 325.2 1. 38 .2400 Grade x Treatment 2 7.7 3.9 .02 .9837 Within Group (Error) 421 98911.8 234.9
102 Between school comparisons at Time 1. For the next level of analysis, the reading pretest scores were compared between Schools A and B at Time 1 by treatment level. Although the means appear to be similar (see Table 14), a t test at a significance level of .05 showed that the support groups (t (455) = 4.5, p < .0001) and the control groups (t (440) = 4.5, p < .0001) between school sites were statistically different. Be cause of this statistical level of difference, the next level of analysis at Time 2 was completed on each school separately. Table 14 Descriptive Statistics of Student Reading Scores at Time 1 by Treatment Level Between School Sites (N = 899) Variable n M SD Treatment Groups School A 244 47.9 14.9 School B 213 42.0 13.1 Control Groups School A 228 50.4 14.1 School B 214 43.6 17.1 Time 2 analyses. At the end of the school year, all students were again administered the Stanford Diagnostic Read ing Test Fourth Edition The student reading achievement posttest results at Time 2 were then compared within each school site. The general descriptive statistics from both school sites at Time 2 portray a relatively normal distribution of data for most of the groups (see Tables 15 and 16), although a few of the
103 groups showed some minor deviation. For example, the sixth grade group without support at School B showed a slight positive skewness (.83) and four of the 12 sub groups were slightly platykurtic [ School A: sixth grade without support ( 1.24), seventh grade with support ( 1.07), and eighth grade without support ( 1.04); School B: seventh grade without support ( 1.36)]. However, none of these conditions was significant enough to cause concern. Additi onally, the bar graph of the mean, grade level reading test scores at Schools A and B by treatment group visually illustrate the improvement of the Figures 9 and 10). Tabl e 15 Descriptive Statistics of Reading Scores for School A: Mean, Standard Deviation, Skewness, Kurtosis, Minimum, and Maximum by Group at Time 2 (N = 472) Group n M SD Skewness Kurtosis Minimum Maximum Grade 6 with Support 65 61.6 16.3 .34 0.26 24 91 Grade 6 without Support 47 55.6 17.9 .07 1.24 24 86 Grade 7 with Support 82 63.8 14.7 .05 1.07 36 90 Grade 7 without Support 85 58.8 14.9 .36 .01 16 95 Grade 8 with Support 97 59.6 18.1 .11 .57 20 98 Grade 8 without Support 96 56.2 17.1 .14 1.04 22 92
104 Table 16 Descriptive Statistics of Reading Scores for School B: Mean, Standard Deviation, Skewness, Kurtosis, Minimum, and Maximum by Group at Time 2 (N = 427) Group n M SD Skewness Kurtosis Minimum Maximum Grade 6 with Support 60 52.5 17.8 .01 .74 15 85 Grade 6 Without Support 97 45.8 22.6 .83 .53 14 96 Grade 7 with Support 77 58.9 13.3 .18 .05 29 95 Grade 7 Without Support 55 54.2 19.2 .27 1.36 19 85 Grade 8 with Support 76 59.3 16.1 .19 .75 30 94 Grade 8 Without Support 62 49. 3 20.4 .14 .69 12 92 Figure 9. Mean grade level reading test scores at School A by treatment group at Time 2 (N = 472).
105 Figure 10. Mean grade level reading test scores at School B by treatmen t group at Time 2 (N = 427). Change score analyses. In order to determine the true impact of the differentiated instruction teacher support model, a comparison now needed to be completed on the the posttest, Time 2. The difference of these two scores will now be referred to as the change score (Time 2 Time 1 = Change Score). The basic descriptive statistics for the change scores from School A and School B are presented in Tables 17 and 18, resp ectively. The data illustrate a relatively normal distribution for all twelve subgroups, except two groups that were slightly platykurtic, School A: seventh grade without support ( .82) and School B: sixth grade without support ( .83). These two groups are both within the range of reasonable distributions and will have very little effect when combined with other groups.
106 Table 17 Descriptive Statistics of Reading Change Scores for School A: Mean, Standard Deviation, Skewness, Kurtosis, Minimum, and Maximum b y Group (N = 472) Group n M SD Skewness Kurtosis Minimum Maximum Grade 6 with Support 65 14.26 15.61 .23 .57 14 46 Grade 6 without Support 47 4.81 11.07 .10 .08 15 36 Grade 7 with Support 82 16.59 11.42 .04 .34 10 38 Grade 7 without Support 85 9 .39 15.78 .18 .82 19 43 Grade 8 with Support 97 10.78 13.87 .17 .33 18 40 Grade 8 without Support 96 4.96 11.80 .16 .02 24 37 Table 18 Descriptive Statistics of Reading Change Scores for School B: Mean, Standard Deviation, Skewness, Kurtosis, M inimum, and Maximum by Group (N = 427) Group n M SD Skewness Kurtosis Minimum Maximum Grade 6 with Support 60 10.85 12.62 .26 0.08 13 44 Grade 6 Without Support 97 2.77 15.52 .46 0.83 20 37 Grade 7 with Support 77 16.51 14.10 .27 0.27 15 48 Gra de 7 Without Support 55 9.80 15.44 .29 0.30 19 43 Grade 8 with Support 76 17.08 15.44 .13 0.12 17 53 Grade 8 Without Support 62 5.27 11.97 .38 0.64 18 30
107 When the mean reading change scores were graphically compared, the treatment groups had i ncreased their scores to a greater degree than the control groups at both sites (see Figures 11 and 12). In several grade level comparisons, the change for treatment group was as much as three to four times the growth of the control group. The least differ ence in reading scores between the treatment and control groups by grade level was 5.82 points at School A and 6.71 points at School B. The greatest difference in reading grade level scores was 9.45 points at School A and 11.81 points at School B. Figure 11. level (N = 472).
108 Figure 12. level (N = 427). In order to rule out chance, t data was to conduct a 2 (treatment) x 3 (grade level) factorial ANOVA on the resulting change scores using an alpha level of .05 to test for each effect. Before pr oceeding with the analysis, the assumptions of independence, normality, and homogeneity of variances were investigated. First, because the students worked individually on their assessments and trained teachers proctored the assessment, it is therefore reas onable to assume that the assumption of independence has not been violated. Although the sample sizes are not exactly equal, they are relatively similar and the within group degrees of freedom was 466 for School A and 421 for School B making the sample siz es large enough to expect
109 robustness to violations of the normality assumption, therefore the normality assumption does not appear to be violated (Cohen, 1992). Finally, when considering the homogeneity of variances, the largest variance ratio was 1.84, le ss than 2.0, which means the equal variance assumption does not appear to be violated. Therefore, we would expect the ANOVA to be relatively robust to violations of the homogeneity of variance assumption. Based on this analysis of the assumptions, it was r easonable to proceed with the factorial ANOVA (Cohen, 1992). The results of the two way factorial ANOVA are presented in Tables 19 and 20. For School A, the model was statistically significant (F(5,466) = 9.4, p (1977, 1988, 1992) effect size (f) was calculated to be .32, which is generally considered a medium effect. The main effect of the use of support for teachers was also found to be statistically significant using the Type III Sum of Squares data due to the unequal group sizes, (F(1 ,466) = 34.29, p f was calculated to be .27, which is generally considered a medium effect size. For School B, the model was also statistically significant (F(5,421) = 13.09, p < effect size (f) was again calculated and found to be .39, which is generally considered a large effect. The main effect of the use of differentiated instruction teacher support groups was also found to be statistically significant using the Type III Sum o f Squares data, (F(1,421) = 39.07, p <.0001). f was calculated to be .30, which is generally considered a medium effect size. It is notable that the reading achievement analyses from both schools indicated the effect of the diffe rentiated instruction teacher support group was statistically significant.
110 Table 19 Source df SS MS F p Model 5 8507.7 1701.5 9.4 <.0001** Grade 2 2384.8 1192.4 6.58 .0015 Treatment 1 6209.7 6209.7 34.29 <.0001** Grade x Treatment 2 229.4 114.7 .63 .5312 Within Group (Error) 466 84394.2 181.1 p < .01; ** p < .0001 Table 20 Source df SS MS F p Model 5 13547.5 2709.5 13.09 <.0001** Grade 2 2951.7 1475.9 7.13 .0009* Treatment 1 8085.4 8085.4 39.07 <.0001** Grade x Treatment 2 467.6 233.8 1.13 .3241 Within Group (Error) 421 87118.6 206.9 p < .001; ** p < .0001
111 Pooled Reading Achieve ment Data Because the results of the reading achievement data analyses were so similar between School A and School B, the data were pooled and the same analyses were re run. The descriptive statistics of the combined school data portrayed a relatively nor mal standard deviations, measures of skewness, and measure of kurtosis all fell within the normal ranges (see Table 21). Figure 13 is a graphic display of all of the grou reading test scores at Time 1 by grade level. Table 21 Descriptive Statistics of Combined Reading Scores: Mean, Standard Deviation, Skewness, Kurtosis, Minimum, and Maximum by Group at Time 1 (N = 899) Group n M SD Skewness Kurtosis Minimum Maximu m Grade 6 with Support 125 44.5 14.7 .22 .35 13 82 Grade 6 without Support 144 45.6 16.9 .32 .14 11 91 Grade 7 with Support 159 44.8 13.7 .60 .09 11 81 Grade 7 without Support 140 47.2 15.1 .30 .37 13 86 Grade 8 with Support 173 46.0 14.9 .30 .4 0 16 86 Grade 8 without Support 158 48.5 15.8 .24 .59 8 79
112 Figure 13. (N = 899). An analysis of variance procedure was then conducted on the pooled Time 1 data and the results showed that none of the comparisons were statistically significant (see Table 22). These results provide confirmation that the null would fail to be rejected and further analyses could be conducted because the groups were not statistically different at Time 1.
113 Table 22 Analysis of Variance Summary Table for Combined Reading Scores at Time 1 (N = 899) Source df SS MS F P Model 5 1668.2 333.6 1.44 .2065 Grade 2 704.8 352.4 1.52 .2185 Treatment 1 866.5 866.5 3.76 .0 532 Grade x Treatment 2 94.0 47.0 .20 .8162 Within Group (Error) 893 206548.8 231.3 and compared at Time 2. The results of the descriptive statistics are presented in T able 23. The results portray a relatively normal distribution except that the sixth grade group without support and the eighth grade group without support had a negative kurtosis close to 1, meaning that they had slightly less outliers in their distributi on. A bar graph of the
114 Table 23 Descriptive Statistics of Combined Reading Scores: Mean, Standard Deviation, Skewness, Kurtosis, Minimum, and Maximum by Group at Time 2 (N = 899) Group n M SD Skewness Kurtosis Minimum Maximum Grade 6 With Support 125 57.24 17.55 .20 .62 18 91 Grade 6 without Support 144 49.0 21.61 .50 .94 15 96 Grade 7 With Support 159 61.4 14.19 .15 .66 35 95 Grade 7 without Support 140 57.0 16.85 .41 .66 31 95 Grade 8 With Support 173 59.5 17.22 .002 .61 22 98 Grade 8 without Support 158 53.5 18.70 .09 .83 28 92 Figure 14. (N = 899).
115 An analysis of variance conducted on the combined Time 2 reading data showed that the model, the effect by grade, and the main effect of treatment were statistically significant (see Table 24). Of particular importance to this study is the treatmen t effect f) for this difference was calculated to be .17, which is generally considered a small effect. The f of 14 (F(2, 893) = 8. 25, p = .0003). Table 24 Analysis of Variance Summary Table for Combined Reading Scores at Time 2 (N =899) Source df SS MS F p Model 5 15004.7 3000.9 9.5 <.0001** Grade 2 5212.1 2606.1 8.25 .0003* Treatment 1 8626.5 8626.6 27.31 <.0001** Grade x Treatm ent 2 497.3 248.7 .79 .4555 Within Group (Error) 893 282109.3 315.9 p < .001; ** p < .0001 Figure 15 illustrates the interaction of the groups when tracked over time from Time 1 to Time 2. In both schools, the mean reading score of the treatment parti cipants started slightly below the control group at Time 1 and then by Time 2 their mean score was above the mean score of the control group students.
117 scores by grade level and treatment were examined. The results are reported in Table 25 and reflect a relatively normal distribution of scores with respect to standard deviation, skewness, n reading change scores by grade and treatment level. Table 25 Descriptive Statistics of Reading Change Scores Combined Schools: Mean, Standard Deviation, Skewness, Kurtosis, Minimum, and Maximum by Group (N = 899) Group n M SD Skewness Kurtosis Minimum Ma ximum Grade 6 with Support 125 12.62 14.30 .31 .31 14 46 Grade 6 without Support 144 3.44 14.22 .36 .64 20 37 Grade 7 with Support 159 16.55 12.75 .19 .20 15 48 Grade 7 without Support 140 9.55 15.59 .21 .65 19 43 Grade 8 with Support 173 1 3.55 14.87 .20 .19 18 53 Grade 8 without Support 158 5.08 11.83 .24 .30 24 37
118 Figure 16. (N = 899). When the analysis of variance procedur e was run on the combined schools reading change scores, again the model (F(5, 893) = 20.42, p < .0001), the effect by grade level (F(2, 893) = 10.11, p < .0001), and the effect of the treatment (F(1, 893) = 77.16, p < .0001), were all found to be statisti effect size ( f ) was calculated to be .34 for the model (a medium effect), .29 for the treatment effect (a medium effect), and .15 for the grade level effect (a small effect).
119 Table 26 Analysis of Variance (N = 899) Source df SS MS F p Model 5 19862.82 3972.56 20.42 <.0001* Grade 2 3933.56 1966.78 10.11 <.0001* Treatment 1 15008.83 15008.83 77.16 <.0001* Grade x Treatment 2 178.70 89.35 .46 63 Within Group (Error) 893 173705.57 194.52 p < .0001 Mathematics Achievement Analyses The following sections will describe the mathematics data that were collected during this study and the subsequent analyses. The student mathematics achievement d ata were submitted to the same rigorous analyses that were used on the reading achievement data. Within site comparisons at Time 1. First, in order to assess the effect of the differentiated instruction and teacher support groups on mathematics achieveme nt, all students were administered the FCAT Mathematics Achievement Predictor Test Form B pretest. Next, like the reading results, the pretest results of the student mathematics achievement data at the beginning of the study, Time 1, were compared within each school site to determine if the treatment and the control groups were statistically different at the beginning of the study. The general descriptive statistics from both school sites at
120 Time 1 portrays a relatively normal distribution with only a few of the groups being grade with support at School B, being slightly positively skewed (see Tables 27 and 28). Additionally, bar graphs of the mean, grade level mathematics test scores at Schools A and B by treatment group visually illustrate the closeness of the data at Time 1 (see Figures 17 and 18). The range of group ally, an analysis of variance on the Time 1 data within each school site demonstrated that the model was not statistically different at School A (F(5, 463) = 1.48, p = .1954) nor at School B (F(5, 417) = 0.81, p = 0.5440) (see Tables 29 and 30). Although, there was a statistical significance by grade level within School A F(2,463) = 3.35, p = .0359). Table 27 Descriptive Statistics of Mathematics Scores for School A: Mean, Standard Deviation, Skewness, Kurtosis, Minimum, and Maximum by Group at Time 1 (N = 469) Group N M SD Skewness Kurtosis Minimum Maximum Grade 6 with Support 66 56.9 22.2 .09 .91 11 97 Grade 6 without Support 45 55.6 17.1 .07 .79 24 91 Grade 7 with Support 82 52.7 14.5 .44 .46 28 87 Grade 7 without Support 84 53.0 13.5 .09 .22 2 1 88 Grade 8 with Support 97 56.5 16.3 .23 .70 22 98 Grade 8 without Support 95 58.1 17.3 .74 .36 19 88
121 Table 28 Descriptive Statistics of Mathematics Scores for School B: Mean, Standard Deviation, Skewness, Kurtosis, Minimum, and Maximum by Group a t Time 1 (N = 423) Group n M SD Skewness Kurtosis Minimum Maximum Grade 6 with Support 53 52.5 13.5 .23 .49 29 80 Grade 6 without Support 94 55.0 18.7 .44 1.00 28 96 Grade 7 with Support 77 49.7 15.0 1.07 .51 30 90 Grade 7 without Support 55 54.2 1 9.0 .01 .94 20 92 Grade 8 with Support 79 54.4 19.6 .12 1.02 22 92 Grade 8 without Support 65 53.3 16.6 .12 1.01 22 86 Figure 17. Mean grade level mathematics test scores at School A by treatment group at Time 1 (N = 469).
122 Figure 18. Mean grade level mathematics test scores at School B by treatment group at Time 1 (N = 423). Table 29 (N = 469) Source df S S MS F p Model 5 2085.4 417.1 1.48 .1954 Grade 2 1891.5 945.8 3.35 .0359* Treatment 1 5.4 5.4 .02 .8905 Grade x Treatment 2 148.3 74.2 .26 .7690 Within Group (Error) 463 130629.8 282.1 p < .05
123 Table 30 Analysis of Variance Summary Table for Scho (N = 423) Source df SS MS F p Model 5 1229.3 245.9 .81 .5440 Grade 2 242.0 121.0 .40 .6720 Treatment 1 334.1 334.1 1.1 .2951 Grade x Treatment 2 484.6 242.3 .80 .4514 Within Group (Error) 417 126794.3 304.1 Between school comparisons at Time 1. For the next level of analysis, the mathematics pretest scores were compared between Schools A and B at Time 1 by treatment level. The mean and standard deviation of the subgroups by treatment level and by school ar e presented in Table 31 and shows the similarity of the data from all four groups. A t test at a significance level of .05 showed that the support groups (t (455) = 1.93, p = .06) and the control groups (t (440) = .85, p = .39) between school sites were not statistically different. This statistical level of difference meant that the null should fail to be rejected and that the student mathematics achievement data could be pooled between sites. Therefore, all future analyses performed on the student mathematic s achievement scores will use combined school data to simplify the process and provide a larger sample.
124 Table 31 Descriptive Statistics of Student Mathematics Scores at Time 1 Pooled by Treatment Level Between School Sites (N = 892) Variable n M SD Treat ment Groups School A 245 55.3 17.6 School B 209 52.2 16.6 Control Groups School A 224 55.7 16.0 School B 214 54.3 18.1 Combined mathematics data at Time 1. The combined mathematics achievement data at Time 1 are presented in Table 32, which p ortray a similar distribution of means and standard deviations among groups. The data illustrate a relatively normal distribution for all subgroups, with the exception of a few groups that were slightly platykurtic, sixth grade without support ( .95), eigh th grade with support ( .83), and eight grade without support ( .77). Because these values were not less than 1.0, they were not considered a threat to the overall variation of data and allow for a continued analysis of combined data. Figure 19 visually d epicts the closeness of the means among all groups.
125 Table 32 Deviation, Skewness, Kurtosis, Minimum, and Maximum by Group at Time 1 (N = 892) Group n M SD Skewness Kurtosis Mini mum Maximum Grade 6 with Support 119 55.0 18.9 .11 .51 11 97 Grade 6 Without Support 139 55.2 18.1 .33 .95 24 96 Grade 7 with Support 159 51.2 14.7 .73 .13 28 90 Grade 7 Without Support 139 53.4 15.8 .08 .48 20 92 Grade 8 with Support 176 55. 6 17.9 .13 .83 22 98 Grade 8 Without Support 160 56.2 17.1 .47 .77 19 88
126 Figure 19. (N = 892). Time 2 analyses. At the end of the five mont h period, all students were administered the FCAT Mathematics Achievement Predictor Test Form C. The student mathematics achievement posttest results at Time 2 were then compared. The general descriptive statistics from both school sites at Time 2 portra y a relatively normal distribution of data for most of the groups (see Tables 33 and 34); although a few of the groups showed some minor deviation. For example, the eighth grade group without support at School A showed a slight negative skewness ( .84 ) and 10 of the 12 sub groups were slightly platykurtic which indicated less outliers and attests to the closeness of data. However, none of these conditions was significant enough to cause concern. When the data from both schools were combined, the similarity among the data sets
127 remained (see Table 35). Additionally, a bar graph of the combined mean, grade level mathematics test scores by treatment group visually illustrates the improvement of the he control groups, except in the eight grade where the resulting mean scores are very close (see Figure 20). Table 33 Descriptive Statistics of Mathematics Scores School A: Mean, Standard Deviation, Skewness, Kurtosis, Minimum, and Maximum by Group at Ti me 2 (N = 469) Group n M SD Skewness Kurtosis Minimum Maximum Grade 6 with Support 66 59.3 20.80 .19 1.19 23 96 Grade 6 without Support 45 56.3 21.3 .13 1.01 15 96 Grade 7 with Support 82 69.4 13.3 .13 .09 33 97 Grade 7 without Support 84 63.9 17. 2 .06 .72 26 98 Grade 8 with Support 97 68.6 15.4 .03 .80 39 99 Grade 8 without Support 95 68.9 18.6 .84 .08 18 97
128 Table 34 Descriptive Statistics of Mathematics Scores School B: Mean, Standard Deviation, Skewness, Kurtosis, Minimum, and Maximum by Group at Time 2 (N = 423) Group n M SD Skewness Kurtosis Minimum Maximum Grade 6 with Support 53 59.6 14.7 .05 1.07 33 87 Grade 6 without Support 94 50.3 20.0 .27 .99 17 92 Grade 7 with Support 77 71.2 11.9 .19 .65 43 93 Grade 7 without Suppor t 55 65.8 19.9 .16 1.17 32 98 Grade 8 with Support 79 64.2 18.9 .7 .92 26 98 Grade 8 without Support 65 62.1 16.0 .24 .98 29 92
129 Table 35 Deviation, Skewness, Kurtos is, Minimum, and Maximum by Group at Time 2 (N = 892) Group n M SD Skewness Kurtosis Minimum Maximum Grade 6 with Support 119 59.4 18.3 .15 .98 23 96 Grade 6 without Support 139 52.2 20.6 .23 .98 15 96 Grade 7 with Support 159 70.3 12.6 .18 .29 33 97 Grade 7 without Support 139 64.7 18.3 .09 .94 26 98 Grade 8 with Support 176 66.7 17.1 .13 .75 26 99 Grade 8 without Support 160 66.1 17.9 .55 .57 18 97
130 Figure 20. by grade level at Time 2 (N = 892). Change score analyses. In order to determine the resulting impact of the differentiated instruction teacher support model on student mathematics achievement, a ement from the pretest, Time 1, to the posttest, Time 2, referred to as the change score (Time 2 Time 1 = Change in Table 36. The data illustrate a relatively normal distribution in seventh and eighth grades. However, in sixth grade, the change scores were low which was consistent with scores from other middle schools within the district. It is notable that the sixth grade dicating that many students in that group actually scored lower on the posttest than the pretest.
131 Table 36 Descriptive Statistics of Mathematics Change Scores Combined Sites: Mean, Standard Deviation, Skewness, Kurtosis, Minimum, and Maximum by Group (N = 892) Group n M SD Skewness Kurtosis Minimum Maximum Grade 6 with Support 119 4.18 10.39 .14 .37 19 34 Grade 6 Without Support 139 3.17 10.25 .49 .25 22 28 Grade 7 with Support 159 19.65 12.60 .27 .56 10 47 Grade 7 Without Support 139 11.20 10.9 3 .09 .28 18 43 Grade 8 with Support 176 11.10 9.44 .32 .28 10 38 Grade 8 Without Support 160 9.87 9.31 .25 .33 19 34 When the pooled, mathematics change scores for all groups were graphically and numerically compared, the treatment groups had increased their scores to a greater degree than the control groups at all three grade levels (see Figure 21). The seventh grade had the largest difference of 8.45 percentage points and eighth grade had the smallest difference of 1.23 points.
132 Figure 21. (N = 892). Figure 22 illustrates the interaction of the groups when tracked over time from Time 1 to Time 2. Within both schools, the mean mathematics score of the treatment participants started slightly below the control group at Time 1 and then by Time 2 their mean score was above the mean score of the control group students. In fact, School B had a significant difference of 7.7 points at Time 2.
134 In order to rule out chance, t he final analysis on the student mathematics achievement data was to conduct a 2 (treatment) X 3 (grade level) factorial ANOVA on the resulting change scores using an alpha level of .05 to test for e ach effect. Before proceeding with the analysis, the assumptions of independence, normality, and homogeneity of variances were investigated. First, because the students worked individually on their assessments and trained teachers proctored the assessment, it is therefore reasonable to assume that the assumption of independence has not been violated. Although the sample sizes are not exactly equal, they are relatively similar both within each grade level and among groups. The total number of student partici pants in the mathematics treatment group was 454 and the total in the control group was 438 making the sample sizes large enough to expect robustness to violations of the normality assumption, therefore the normality assumption does not appear to be violat ed. Finally, when considering the homogeneity of variances, the largest variance ratio was 1.48, less than 2.0, which means the equal variance assumption does not appear to be violated. Therefore, we would expect the ANOVA to be relatively robust to violat ions of the homogeneity of variance assumption. Based on this analysis of the assumptions, it was reasonable to proceed with the factorial ANOVA. The results of the two way factorial ANOVA are presented in Table 37. The overall model was statistically sig nificant (F(5,886) = 77.36, p 1988, 1992) effect size (f) was calculated to be .66, which is generally considered a large effect. Due to grade level differences, the grade level effect of the use of teacher support groups was also found to be statistically significant using the Type III Sum of Squares data due to the unequal group sizes, (F(2,886) = 141.72, p
135 1992) f was calculated to be .56, which is again generally considered a large effect size. Of p articular significance to this research project, the treatment effect was also statistically significant (F(1,886) = 63.8, p effect size (f) was again calculated and found to be .38, which is generally considered a large effect. The interaction of the effect of both grade level and treatment level for the combined groups was significant (F(1, 886) = 10.8, p (1977, 1988, 1992) effect size ( f ) was .16, a small effect. Table 37 Analysis of Scores (N = 892) Source df SS MS F p Model 5 42880.2 8576.0 77.36 <.0001* Grade 2 31422.14 15711.07 141.72 <.0001* Treatment 1 7072.13 7072.13 63.80 <.0001* Grade x Treatment 2 2393.84 11 96.92 10.80 <.0001* Within Group (Error) 886 98219.06 110.86 p < .0001 The mathematics results were very encouraging. These data suggest that the use of differentiated instruction teacher support groups in urban, Title I schools was statistically sig nificant with regard to improving student mathematics achievement scores, even though there was a grade level effect.
136 Teacher Fidelity Observation Data Analyses In the following section, the summarized data and results of all teacher observation analyses are presented and evaluated in order to address the second research question: What were the statistical differences among teacher groups who participated in facilitated support groups and those who did not with respect to their implementation of differenti ated instruction as measured by the Differentiated Instruction: Fidelity Implementation Tool (DI: FIT) observation tool? First, all teachers were observed by trained observers to determine their degree of fidelity to the differentiated instruction model u sing the DI: FIT observation tool. Specifically, in order to confirm that the teachers in the treatment groups were applying the differentiated instruction philosophy to their classrooms and transferring the strategies learned in the differentiated instruc tion, facilitated teacher support groups, these teachers were observed once per nine weeks and given specific feedback on ways to improve their instruction. The teachers in the control group were also observed during each nine week period to assess whether or not they were differentiating their classroom instruction and, if they were, to what degree. observation scores revealed that all four groups had a similar sample size standard deviation, and minimum scores (see Table 38). The groups were different with respect to the mean scores (see Figure 23). As expected, the participants who were a part of the differentiated instruction monthly support groups scored an average of 3.4 points higher on the 20 point scale. In addition, the treatment group at School A had a few more high
137 scores (skewness = 1.41, kurtosis = 2.10). The control group of participants at School B had a slightly more flat distribution of scores (kurtosis = 1.28). It is important to note that the differences of the combined scores by school site are very similar with the mean scores differing by only .68 points. These data again provide evidence that the schools were similarly matched. Table 38 Descriptive Statistics of Teacher DI: FIT Observation Mean Scores: Mean, Standard Deviation, Skewness, Kurtosis, Minimum, and Maximum by Group and by School (N = 55) Group N M SD Skewness Kurtosis Minimum Maximum Treatment Group at School A 15 17.5 2.29 1.41 2.10 11.5 20.0 Control Group at School A 13 13.7 2.19 .26 .87 10.5 17.5 Treatment Group at School B 13 16.5 2.11 .63 .51 12.5 19.5 Control Group at School B 14 13.7 2.04 .35 1.28 10.5 16.5 Combined Treatment Groups 28 17.1 2.22 1.19 1.02 11.5 20.0 C ombined Control Groups 27 13.7 2.07 .29 .81 10.5 17.5 School A: Treatment + Control 28 15.75 2.92 .34 1.05 10.5 20.0 School B: Treatment + Control 27 15.07 2.49 .44 .58 10.5 19.5
138 Figure 23. Mean Teacher DI: FIT Obse rvation Scores by Treatment and by School (N = 55) Relationship Between Fidelity and Achievement In the following section, data for the analysis of the correlation between the hange scores are presented in order to address the third research question: change scores?
139 An analysi s of variance on the same observations scores revealed the overall model was statistically significant (F(3, 51) = 11.34, p < .0001) (see Table 39). The f effect size was calculated to be .78, which is considered a large effect s ize. The R 2 value reported by SAS was .40, which reflects the percentage of variance that was accounted for by the dependent variable. The use of the differentiated instruction teacher support groups, the treatment, was also statistically significant (F(1, 51) = 31.86, p f effect size of .76. The effect of the school site and the interaction were not statistically significant. These data allowed for the secondary analyses of the correlation between the teach Table 39 Analysis of Variance Summary Table for Teacher DI: FIT Observation Scores by Treatment Level and by School (N=55) Source df SS MS F p Model 3 159.0 53.0 11.34 <.0001* Tre atment 1 148.9 148.9 31.86 <.0001* School 1 3.28 3.28 .70 .4064 Treatment x School 1 3.06 3.06 .65 .4223 Within Group (Error) 51 238.4 4.67 p < .0001
140 c change score, and school variables shows that there was a very strong positive correlation between several of the variables (see Table 40). The was calculated to be 62 ( p <.0001), a moderate correlation. The correlation between score and the st statistic, R 2 was calculated and found to be .62, which is considered a moderate correlation and accounts for 62% of the variance. It is notable that again the correlation of the school site was not statistically significant and therefore had no effect on the other treatment correlations. Table 40 Academic Change Score, Treatment, and School (N = 55) Variable T DI: FIT Observation Score Academic Change Score School Treatment .62 (<.0001)* .63 (<.0001)* .05 (.69) DI: FIT Observation Score .79 (<.0001)* .13 (.36) Change Score .01 (.95) Note The values in the parentheses are p values. p < .0001
141 Support Group Analyses In the following section, a qualitative analysis on the detailed minutes from the facilit were examined in order to address the fourth research question: Using qualitative data and feedback provided by the teachers in the treatment groups, what were the teache and their instructional growth? In order to ensure that any instructional philosophy and teaching strategies learned are transferred to the classroom with fidelity, staff development personnel must ma ke certain that the teachers have on going support (Westling, Cooper Duffy, Prohn, Ray, & Herzog, 2005). In this study a facilitated, teacher support group model was initiated to help provide the necessary support teachers would need to change their teachi ng practices and sustain the newly implemented skills and strategies with fidelity. The data from the DI: FIT teacher observations documented that the teachers who participated in the groups did, in fact, implement the components of differentiated instruct ion to a greater degree than the teachers in the control group. Now, a closer analysis of the data gleaned from these meetings will help guide future meetings and to document the components that the teachers perceived as important. Using an inductive ana lysis in conjunction with a document review, all data were reviewed, meaningful units were identified, units of data were coded, and then the data were categorized in order to identify basic themes and views. Then, the data were further reduced through a c onstant comparison, a consolidation of any redundant
142 categories, and an analysis of emergent themes. The use of multiple methods of data collection allowed for method and data triangulation and increased credibility of the findings. All of the teachers re sponded that they thought the support group model was achievement, and the classroom learning community. One teacher even wrote in her as invited to be part of the DI Support Team. It has the 28 teacher in the treatment group, 96.4%, responded that they really liked the overall format of the facilitated, tea cher support groups. Further, based on the results from the district in service follow up questions (see Appendix K and L), when teacher participants were asked if the content of the differentiated support group meetings was appropriate and built upon the knowledge and experiences of the intended participants, 90.5% of the teachers responded that they strongly agreed and 9.5% responded that they agreed; no participants responded that they were undecided, disagreed, or strongly disagreed. During the final focus group session, 100% of the teachers responded that the length of each session, 2 hours, was adequate. However, 89.3% of the teachers requested that the groups meet more often during the beginning of the year. Specifically, 71.4% requested that the se ssions begin during pre planning so the teachers could begin to create differentiated lesson plans and create their flexible grouping student cards with scores. Then after the year began, 89.3% expressed an interest in meeting twice a month for the first three months.
143 On going support of the teachers is important before, during, and after support group meetings. In this study, this was accomplished in several ways. First, the use of the differentiated instruction, teacher resource library was appreciated by all of the support group, teacher participants, and 92.9% of the teachers agreed that they found ideas they could implement and/or encouraged the use of new differentiat ed instruction components the members of the support group with access to a DI c onference area on their email desktop. This provided the teachers with an opportunity to share lesson plans and internet links, ask questions, post concerns, receive support, and view the ideas and growth of others. In addition, it served as a daily remind er for teachers to strive toward increased implementation of the differentiated instruction principles. One hundred percent of the participants responded that they appreciated the DI conference area and expressed a desire for it to be continued into the ne xt year. Several improvements were suggested by the teachers during the final focus group to improve the process for the upcoming year. All of the teachers, 100%, supported the idea to make the support groups open to all faculty members. The majority of t he teachers, 92.9%, wrote on their DI Support Group Feedback sheets that they would like came up with several suggestions to accomplish this. First, they suggested ha ving differentiated lesson they were particularly proud of, they could invite the rest of the g peer
144 partner relationships so they would have the input of other classroom teachers and not just the trained observers. A theme that was expressed through the journals, the feedback sheets, and the focus groups was that the teachers genuinely enjoyed t he sheer process of sharing ideas lpful; it gave me ideas and inspiration and encouragement among peers. Teacher participants also referenced the benefit of the vertical and horizontal articulation among subject area teachers and across discipl ines as being a benefit. They stated that they just do not have time during the day to sit down and talk with other teachers about what they are doing in their classroom. When asked which elements of the study encouraged them to sustain their instructiona l fidelity throughout the study, there was a variety of responses. The majority of teachers referenced the support group meetings, the fact that they knew an observer m eetings kept my instructional focus and renewed my zeal to learn and implement new DI lessons. With other district in services, I lost interest and went back to the teaching ac
145 It is important to remember that the 50% of the teachers in the treatment group were new teachers (0 3 years of teaching experience) and 14.3% were not certified in the subject they were teaching, which is very representative of the teachers who commonly teach at Urban, Title I schools. In o very much about differentiating instruction at the beginning, but I feel more confident beginning of the year, I was leery of DI, but I really feel now that my students can really get behind ways the process worked well and ways it can be improved. For the future, the support group model holds promise as an important tool districts and in service personnel can use to help sustain implementation with fidelity of any instructional philosophy or program. Summary In this chapter, numerous data sources have been provided to assure that the treatment and contr ol groups were as closely matched as possible. Next, data on teacher fidelity and student achievement was presented so the impact of the differentiated instruction support group model could be substantiated. Finally, the support group model itself has been closely scrutinized. Of importance is the fact that although there was a school level effect with respect to the reading achievement and a grade level effect with
146 respect to mathematics, students, as a whole, whose teachers were participants in the differ entiated instruction support groups out performed their peers. Even though the difference was slight, in most cases, the change in academic achievement was statistically significant and a few points difference is all that many students need to meet state a nd district level benchmarks and standards.
147 Chapter Five Discussion Purpose From the onset, this mult ifaceted study had a myriad of purposes. First, it was the desire of the researcher to contribute to the research and knowledge base of differentiated instruction, facilitated support groups, and implementation fidelity of evidence based practices. Further, this study utilized and collected additional data on a new fidelity assessment tool for teachers, the DI: FIT, and evaluated teach facilitated support group model. Finally, the study examined the relationship between implementation of differentiated instruction with fidelity and the academic outcomes of urban, middle school adolescents. Method This study util ized a quasi experimental, mixed method design to investigate multiple components over a five month period. For this study, two matched urban, Title I, middle schools were selected as implementation sites; and within each of the sites, teams of teachers wh o taught the same body of students were purposefully assigned to the treatment or control group at each grade level. The demographics of both students and teachers were compared to assure that all comparisons were similar samples. The combined sample conta ined 55 teachers and 906 students. A triangulation of data from
148 instruct the support group model and instructional change. All teacher participants in the treatment group attended a two hour, monthly, differentiated instruction support group mee ting, had access to a differentiated instruction resource library, and received feedback and instructional suggestions following all observations, i.e., reinforcement on which differentiated components were observed, possible flexible grouping strategies, and ideas for future lessons. The final session of the year was a focus group so the researcher could obtain feedback from the teachers regarding which components of the support group they felt were the most valuable, which components need to be continued, and what components needed to be added. Reading achievement was assessed using the Stanford Diagnostic Reading Test Mathematics Predictor Tests. All student achievement results were then analyzed using basic descriptive statistical procedures in addition to an analysis of variance procedure. An effect size was also calculated on all statistically significant findings. All teacher participants were exposed to the different iated instruction model and then observed each nine week period to assess the degree of implementation fidelity. After all of the observations had been completed, the mean teacher fidelity observation scores were statistically compared using an analysis of variance by treatment and school level. Next, a
149 used to determine the relationship bet ween fidelity of implementation and student achievement. Results The results of this study are very encouraging. Both the reading and mathematics achievement change scores and the difference in the teacher fidelity observation scores, DI: FIT, by treatme nt group were statistically significant. A clear relationship also existed scores. In addition, the teachers who participated in the support group meetings clearly felt t hat the support was beneficial. between the treatment and control groups was 8.38 percentage points, which was statistically significant (F(1, 893) = 77.16, p <.0001, N = 899). The 1992) f effect size for the treatment comparison was calculated to have a medium effect size of .29. f of .1 is considered small, .25 medium, and .4 large (Cohen 1977, 1988, 1992.) The reading achievement results were simila r across all three grade levels and the difference between schools was not statistically significant. scores was 6.12 percentage points. This comparison by treatment level was al so statistically significant (F(1, 886) = 63.80, p 1988, 1992) f effect size of .38, a medium to large effect. With reference to the mathematical achievement data, grade level differences were found. The grade lev el effect was statistically significant (F(2,886) = 141.72, p f = .56) as was the interaction effect of grade level by treatment (F(1,886) = 10.8 p f =
150 .16). The overall mathematics achievement data were effected by the l ow performance of the sixth grade students. Sixth grade is usually a difficult transition year for students and the fact that three of the four sixth grade mathematics teachers were non tenured, explains some of the possible reasons for the lack of mathema tics growth in this grade. The fact that the majority of the sixth grade teachers were not tenured is a common scenario in middle schools. This is often a result of the more senior teachers selecting to teach the older, more mature students leaving the adm inistration with the task of having to fill the sixth grade positions with new, inexperienced teachers Although, the sixth grade students experienced limited growth in this study their change scores were consistent with the mean change scores of other si xth graders in the school district during the same period. In the final data collection, t here was also little difference between the mathematics achievement data in the eighth grade. This lack of difference in achievement res ults is po ssibly due to the high differentiated instruction implementation scores of the mathematics teachers as evidenced by the ir DI: FIT observations. Unlike other general education teachers in the control group, the eighth grade mathematics teachers im plemented many of the differentiated instruction components without the help of the support group Their high degree of implementation on their own is testimony to the fact that some teachers will implement evidence based strategies on their own; however, in this study they represent ed only 2 out of 27 teachers, 7.4%. Even with the closeness of data in eighth grade, the overall treatment effect was still statistically significant. In urban schools, where large percentages of students are academically below gra de level, these small gains
151 performance and provide some students with the few points they may need to pass their benchmarks in order to be promoted to the next grade. The teacher DI: FIT observation data also supported the conclusion that a statistically significant difference existed between the teachers who participated in the differentiated instruction support groups and the teachers who did not. While the overall mean difference by school was only .68 points, the overall mean difference by treatment level was 3.4 points out of a possible 20 points The implications are that the average teacher who participated in the support group utilized approximately 3 to 4 more differentiated instruction components in their classroom than the teachers who did not participate. The indicators that were often omitted by the control group were: #4 Lesson is differentiated by content, product, or process; #5 Lesson is differentiated according to rofiles; #9 Teacher uses anchor activities; #10 Teacher acts as a facilitator; #13 Teacher uses flexible and purposeful grouping; and #20 Teacher and students collaborate in the learning process. For research replication purposes both of the co observers were easily trained within a few hours. After a discussion of fidelity indicators and a practice coding session, a high degree of inter rater reliability was obtained. In the future, more teachers and administrators will need to be trained on the correct use of the DI: FIT instrument. The analysis of variance on the teacher mean observation scores by treatment level revealed a statistically significant relationship (F(1, 51) = 31.86, p <.0001), with a f effect size of .76, whic h is considered to be a large effect. academic change scores of their students, the resulting correlation value of .79 suggested
152 that there was a moderate correlation betwee n t hese two variables. This moderate correlation value accounts for 62% of the variance in the model (R 2 = .62). When teachers in the support group were provided an opportunity to share their perceptions of the differentiated instruction support group mode l, they provided some keen insight to common implementation difficulties and suggestions for future support group models. Further, many of the teachers shared that if they had not been a part of the support group model, they would have abandoned the differ entiated instruction philosophy early in the treatment period and opted for a more traditional approach to teaching because of the amount of time that it took to create project and lesson options and to change their pedagogy. All 28 of the teachers in the treatment group felt that the coaching and teacher to teacher sharing aspects of the support group model were key components in helping them maintain their instructional fidelity. The teacher participants also cited group accountability as a big motivator. Based on previous research, initial implementation, fidelity, and sustainability are common problem areas where district and school implementation projects experience their biggest challenges (Webster Stratton, 2003). Moreover, Joyce and Showers (2002), two notable researchers in the field of professional development have conducted and reviewed hundreds of studies on this topic and they caution purveyors that the most effective intervention will not produce desired effects if it is not implemented with fidelity. Their well known meta analysis on trai ning and coaching teachers indicated that the key components to implementation fidelity by practitioners were practice and feedback in training and on going coaching in the classroom (Joyce & Showers, 2002).
153 Limitations The results of this study are only applicable to urban, Title I, middle school teachers and students. The results are also limited to the tests that were utilized in this study. Further replications will need to be conducted with additional achievement tests to see if the results can be generalized. To compensate for some of the possible external threats to validity, a treatment and control group was selected within each school. All effects reported in this study have limited generalizability due to the specific demographics of the population studied and the sample of participants utilized. Although special care was used to ensure that student and teacher participant groups were closely matched and the study had a large number of participants, it was still a convenience sample; therefore the results must be viewed with caution. Significance In spite of these limitatio ns, this study provided data on the use of a viable model that enabled two urban schools to implement evidence based practices successfully. The need for such a model is well documented in the research (Brandt, 1998; Greenwood & Joyce & Showers, 2002; Klingner, Ahwee, Pilonieta, & Menendez, 2003; Little, 1993; Richardson, 1997; Showers, Joyce & Bennett, 1987; Spencer & Logan, 2003; U.S. Department of Education, NCLB, 2002). Even though there has been a wealth of research on effective teaching components, there continues to be a gap between researc h and practice in the field of education. In order to implement evidence based practices in schools, policy makers, trainers, coaches, and practitioners need a clear model that will help teachers implement and sustain their instructional fidelity. The teac her support group
154 model utilized in this study demonstrates the potential of this type of model in urban, Title I middle schools. This type of classroom environment holds many challenges for educators because many of the children have low scores and come from impoverish backgrounds. Since the model utilized in this study did provide statistically significant academic improvements in both reading and mathematics, it may possibly work for other populations of students in other regions. The current study al so provides additional data on effective professional development and instructional implementation practices that are sorely needed in the field of education The widely cited research of Guskey (1986), Joyce and Showers (2002), and the U.S. Department of (2000) document the limited transferability of instructional knowledge learned through in service opportunities to the classroom. Each year, school districts spend a great deal of money paying for traine rs, supplies, and participant salaries and then very little of the Showers, 2002). Based on the DI: FIT teacher observation data, the majority of the teachers who partic ipated in the on going support group during the five month period were able to implement many of the differentiated instructional strategies with fidelity They were also able to sustain their instructional enthusiasm through the support of their colleague s at the monthly teacher support group meetings Th e data obtained from the support for some possible implementation solutions that are missing in many professional development programs. Unquestionably, d istricts need to find a more cost efficient way to bridge the research to implementation gap ; and although the teacher sample size in this study was small the student data
155 supported the effectiveness of the teacher support model. The model utilized could easily be adjusted and the lessons learned can provide future guidelines for other districts wishing to expand their professional development programs. Furthermore, this study has demonstrated that when administrators, trainers, coachers, and teachers utilize a su pport group model in addition to fidelity observations with feedback, student achievement can be affected. Currently, many educators find themselves struggling to teach their increasingly diverse classes. The differentiated instruction philosophy combines many evidence based practices for teaching within one educational philosophy, which provides opportunities for increased social interaction, appropriate learning strategies, helpful feedback, and a positive learning environment (Brandt, 1998; Chapman & Kin g, 2005; Hornsby & Diket, 1999; Tomlinson, 2005). Teachers can no longer afford not to differentiate their instruction and administrators can differentiated instruction mo all students will be afforded opportunities to learn. Implications for Practice The two schools utilized for this study were purposefully selected beca use they characterize urban, Title I schools that typically have increased challenges due to large percentages of students who are functioning academically below grade level, increased discipline problems, high percentages of students from impoverished bac kgrounds, high teacher turn over, very di verse student populations, and high percentages of teachers who are either not certified or lack classroom experience. In the current study, the
156 differentiated instruction, support group model was a useful vehicle f or raising the reading and mathematics achievement scores of students at both school sites and raising was very cost efficient. After the initial costs of the district in service and the minimal costs for a reference library, the only continual cost was teacher time, which can be done either after school or during school planning times with the support of administration. If all of the components of the current model wer e implemented, it would require a person at each site assume the role of facilitator or organizer, the commitment of the participants, the training of participants and observers, the adoption of a fidelity observation tool, and coordinated meeting times fo r teams to plan and share ideas. In order to implement evidence based practices with fidelity, district supervisors, administrators, in service trainers, and peer coaches need to develop, utilize, and support on going, facilitated support groups at indivi dual school sites so teachers can work with other professionals and problem solve site based solutions to the inevitable challenges of implementation. Once a district or faculty adopts a particular evidence based practice, the next step is the initial impl ementation efforts coupled with monitoring fidelity, the development of positive peer coaching partnerships, and the alignment of district policy to support these innovative implementation sites. After the initial implementation, the next step for schools, such as the two in the current study, is school wide implementation and sustainability of the evidence based practice, which requires a long term commitment of time, effort, and training funds. Then over time, an evidence based practice can evolve into th e standard instructional practice of a school, recognized and supported by all members of the staff.
157 Implications for Research The first goal achieved by this research study was to add to the current knowledge and research base of studies that are availab le for purveyors and practitioners regarding the implementation of differentiated instruction, fidelity assessment, support group models, and implementation science. In order to bridge the research to implementation gap, researchers and practitioners must have two way conversations about what works and what does not work Researchers need to listen to needs of classroom teachers and These enhanced partnerships and practices between teache rs will help establish professional learning communities at actual implementation sites, which will further support the implementation of evidence based practices (Fix sen, Naoom, Blase, Friedman, & Wallace, 2005). Several possible research applications ca n be linked to the current study. First, response to intervention researchers are currently investigating promising models that promote the use of evidence based instruction in the general education environment. In the current study, the differentiated ins truction support group model has shown promise studies should be conducted in which the current model is utilized as a first tier of primary prevention in the general edu cation classroom to help remediate student deficits quickly and reduce the number of unnecessary evaluation s for special education services Second, in order to expand the current study, the next research steps could include: replication of the same model and method in other regions and with other student populations; further data collection on the use of the DI: FIT; the examining of sub
158 populations within the sample to see if the achievement varied by populations, i.e., students with disabilities; use of the model with random controlled trials; and a detailed analysis of differentiated instruction components to determine which components had the greatest impact on academic achievement i.e. flexible grouping, tiered lessons, use of on going assessment to g uide instruction The rigorous search for effective core intervention components will require much time because the same model and methods will need to be replicated many times with different combinations of components to see which components of the part icular evidence based practice, in this case differentiated instruction, hold the most promise for producing the desired effect. Unfortunately, this type of study will require the support of a foundation of federal grant money to carry out the detailed ana lysis over time. To achieve this, Fixsen, Naoom, Blas, Friedman, and Wallace (2005) recommend that, implementation strategies and methods that are grounded in research and elabora ted evidence based practices over time. The problem of sustainability is a complex one that will involve many layers of research. The initial perceptions and/or technical teaching ability effects their participation and fidelity. In the current study, there was a high percentage of non tenured teachers and yet they were able to help their stu dents make academic gains. In order to maintain implementation fidelity and achieve sustainability, what incentives will districts and administrators have to make available to classroom teachers and the designated school
159 facilitator? Will teachers continue to implement after the first year, or will they need an incentive?
160 References Affholder, L. P. (2003). Differentiated instruction in inclusive elementary classrooms (Doctoral dissertation, University of Kansas, 2003). ASCD website: About the Authors. (2004). Retrieved November 20, 2004 from the Association for Supervision and Curriculum Development Website: http://www.ascd.org/ publications/books Baumgartner, T., Lipowski, M. B., & Rush, C. (2003). Increasing reading achievement of primary and middle school students through differentiated instruction Master of Arts Research Project. University of Skylight Professional Development Field Blozowich, D. G. (2001). Differentiated instruction in heteroge neously grouped sixth grade classrooms (Doctoral dissertations, Immaculata College, 2001). Boyd, K. (2001). A case study: Climate, best practices, administrative support, and staff attitudes in a middle school inclusion program (Doctoral dissertation, Univ ersity of Central Florida, 2001). Brandt, J. (1998). Powerful learning Alexandria, VA: Association for Supervision and Curriculum Development. Chapman, C., & King, R. (2005). Eleven practical ways to guide teachers toward differentiation. Journal of Staff Development, 26 (4), 20 25.
161 Cohen, J. (1977). Statistical power analysis for the behavioral sciences (revised edition). New York: Academic Press. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2 nd edition). Hillsdale, NJ: Erlbaum Cohen. J. (1992). A power primer. Psychology Bulletin, 112 (1), 155 159. Cook, T. D., & Campbell, D. T. (1975). The design and conduct of quasi experiments and true experiments in field settings. In M. D. Dunette and J. P. Campbell, Handbook of Industrial and Organizational Research, 223 325. Chicago, IL: Rand McNally. Council for Educational Policy, Research, and Improvement (CEPRI). (2005). In service education: The challenge of determining cost and effectiveness. Retrieved July 14, 2006, from http://www .cepri.state.fl.us/pdf/2005%20Exec%20Summary %20and%20Full%20Inservice%20Report.pdf Davis, L. D. (2003). Exploring the relationship between study groups, transfer of professional development, and adult learning styles (Doctoral dissertation, Lehigh Univers ity, 2003). Fixsen, D., Naoom, S., Blas, K., Friedman, R., & Wallace, F. (2005). Implementation research: A synthesis of the literature Tampa, FL: University of South Florida, Louis de la Parte Florida Mental Health Institute, The National Implementation Research Network (FMHI Publication #231). Fleming, L. C., & Baker, P. H. (2002). Differentiating in the classroom: A study of student teachers (Report No. SP041709). Ashland, OH: Ashland University. (ERIC Document Reproduction Service No. ED479480).
162 Flori da Department of Education. (2005). Assessment and accountability briefing book: FCAT, school accountability, and teacher certification tests. Retrieved March 30, 2006, from http://firn.edu/doe/sas/fcat/pdf/fcataabb.pdf Florida Department of Education (200 7). Florida School Grades: Adequate Yearly Progress Report. Retrieved August 16, 2007, from http://schoolgrades.fldoe.org/ default.asp Forness, S. (2001). Special education and related services: What have we learned from meta analysis? Exceptionality, 9 (4) 185 197. Fullan, M. (2001). Leading in a culture of change. San Francisco, CA: Jossey Bass. Good, J., & Weaver, A. (2001). practices: A review of the literature. Paper presented at the Annual Meeting of the Mid South Educational Research Association: Little Rock, AR. (ERIC Document No. ED461637). Grant, S. G. (2000). Teachers and tests: Exploring teachers' perceptions of changes in the New York state testing program. Education Policy Analysis Archives 8 ( 14), 1 27. Retrieved October 5, 2005, from Education Policy Analysis Archives electronic journal: http:// epaa.asu.edu /epaa/v8n14.html (2003). Enhancing school based p revention and youth development through coordinated social, emotional, and academic learning. American Psychologist, 58 466 474. Greenwood, C. R., & Abbott, M. (2001). The research to practice gap in special education. Teacher Education and Special Educat ion, 24 (4), 276 289.
163 Gregory, G. H. (2003). Differentiated instructional strategies in practice: Training, implementation, and supervision Thousand Oaks, VA: Corwin Press. Guskey, T. (1986). Staff development and the process of teacher change. Educational Researcher, 15 (5), 5 12. Hale, M.S. (1999). Collaboration in a study group as a vehicle for professional development (Doctoral dissertation, University of North Carolina, 1999). Heacox, D. (2002). Differentiating instruction in the regular classroom: How to reach and teach all learners, Grades 3 12 Minneapolis, MN: Free Spirit Inc. Hobson, B. (2004). An analysis of differentiated strategies utilized by middle school teachers in a heterogeneously grouped classroom (Doctoral dissertation, Capella University 2004). Hodge, P. (1997). An analysis of the impact of a prescribed staff development program in differentiated instruction on student achievement and the attitudes of teachers and parents toward that instruction (Doctoral dissertation, University of Ala bama, 1997). classroom. The Delta Kappa Gamma Bulletin 65 (2), 5 10. Joyce. B. & Showers, B. (2002). Student achievement through staff development (3 rd Ed.). Alexandria, VA: A ssociation for Supervision and Curriculum Development. Klingner, J., Ahwee, S., Pilonieta, P., & Menendez, R. (2003). Barriers and facilitators in scaling up research based practices. Exceptional Children, 69 (4), 411 429.
164 Kutash, K., & Duchnowski, A. (200 5). Fidelity instrument for measuring the use of evidence based academic strategies in special education classrooms Instrument developed as part of Office of Special Education Programs (U.S.D.O.E.) Grant #H324T000019. Lawrence Brown, D. (2004). Differenti ated instruction: Inclusive strategies for standards based learning that benefit the whole class. American Secondary Education 32 (3), 34 62. Lester, J. (2003). Planning effective secondary professional development programs. American Secondary Education, 3 2 (1), 49 61. Educational Evaluation and Policy Analysis, 15 (2), 129 151. Martin, B. (2000). Exploring the effects of collegial study groups on the transfer of profess ional development training to classroom implementation (Doctoral dissertation, Leigh University, 2000). McAdamis, S. (2001). Teachers tailor their instruction to meet a variety of students needs. Journal of Staff Development 22 (2), 1 5. Retrieved November 27, 2004, from the National Staff Development Council Website: www.nsdc.org/library/ publications/ jsd/mcadamis222.cfm Mokrue, K., Elias, M. J., & Bry, B. H. (2005). Dosage effect and the efficacy of a video based teamwork building series with urban eleme ntary school children. Journal of Applied School Psychology, 21 (1), 67 97. Murphy, C., & Lick, D. (2001). Whole faculty study groups: A powerful way to change schools (2 nd ed.). Thousand Oaks, CA: Corwin Press.
165 The National Academy of Sciences (NAS). (2002 ). The national academy of sciences report: Minorities in special and gifted education. Washington, DC: National Academy Press. National Council of Teachers of Mathematics. (2005). Overview of principles and standards for school mathematics. Retrieved July 1, 2006, from http://www.nctm. org/standards/overview.htm Osborne, B. (1993, March 25). Understanding change in a time of change Paper presented at the Annual Conference on Creating the Quality School: Oklahoma City, OK. (ERIC Document No. ED358544). Pfa ff, M. E. (1999). The effects on teacher efficacy of school based collaborative activities structured s professional study groups (Doctoral dissertation, University of Maryland, 1999). 2). A New Era: Revitalizing Special Education for Children and Their Families. Washington, DC: U.S. Department of Education Office of Special Education and Rehabilitative Services. Retrieved May 24, 2005, from http://www.ed.gov/inits/ commissionsboards/ whspecialeducation/reports.html Richardson, J. (1997, March). Dealing with resisters biggest challenge for staff developers. The Developer Retrieved on November 27, 2004 from the National Staff Development Council website: www.nsdc.org/library/publicatio ns/ developer/dev3 97rich.cfm
166 Seachrest, L., West, S. G., Phillips, M. A., Redner, R., & Yeaton, W. (1979). Some neglected problems in evaluation research: Strength and integrity of treatments. Evaluation Studies Review Annual, 4, 15 35. Showers, B., Joyce B., & Bennett, C. (1987). Synthesis of research on staff development: A framework for future study and the state of the art analysis. Educational Leadership, 45 (3), 77 87. Sparks, D. (2001). Why change is so challenging for schools: An interview with Pet er Senge. Journal of Staff Development 22 (3), 1 6. Retrieved on November 27, 2004 from the National Staff Development Council website: www.nsdc.org/ library/publications/ jsd/senge223.cfm Spencer, S., & Logan, K. (2003). Bridging the gap: A school based s taff development model that bridges the gap from research to practice. Teacher Education and Special Education, 26 (1), 51 62. Tomlinson, C. A. (1995). Deciding to differentiate instruction in middle school: One Gifted Child Quarterly, 39 77 87. Tomlinson, C. A. (1999). The differentiated classroom: Responding to the needs of all learners. Alexandria, VA: Association for Supervision and Curriculum Development. Tomlinson, C. A. (2000). Reconcilable differences: Standards based teaching and differentiation. Educational Leadership, 58 (1), 6 11. Tomlinson, C. A. (2001). How to Differentiate Instruction in Mixed Ability Classrooms (2 nd Ed.). Alexandria, VA: Association for Supervision and Curriculum Development.
167 Tomlinson, C. A. (2003a). Decidin g to teach them all. Educational Leadership 61 (2), 6 11. Tomlinson, C. A. (2003b). Fulfilling the promise of the differentiated classroom: Strategies and tools for responsive teaching Alexandria, VA: Association for Supervision and Curriculum Developm ent. 2005 ASCD Summer Conference. Retrieved October 9, 2005, from ASCD website: http://2005ascdsummerconference.blogspot.com/2005/07/ you cant rush differentiated. html Tomlinson, C. A., & Allan, S. D. (2000). Leadership for differentiating schools & classrooms Alexandria, VA: Association for Supervision and Curriculum Development. Tomlinson, C. A., Tomchin, E. M., Callahan, C. M., Adams, C. M., Pizzat Tinnin, P., Cunni ngham, C. M., et al. (1994). Practices of pre service teachers related to gifted and other academically divers learners. Gifted Child Quarterly, 38, 106 114. U.S. Department of Education. (2002). The Elementary and Secondary Education Act (No Child Left Be hind Act of 2001). Retrieved October 9, 2005, from http://www.ed.gov/ policy/elsec/leg/esea02/index.html U. S. Department of Education, Consortium for Policy Research in Education (CPRE). (1995). Professional development today. Washington, DC: Institute of Educational Sciences. Retrieved July 14, 2006, from http://www.ed.gov/ pubs/CPRE/t61/index.html
168 U. S. Department of Education, National Center for Educational Statistics. (2000). Teacher preparation and professional development: 2000. Washington, DC: Inst itute of Educational Sciences. Retrieved July 14, 2006, from http://nces.ed.gov/surveys/frss/publications/2001088/ U.S. Department of Education, National Center for Education Statistics. (2004). Digest of education statistics tables and figures. Retrieved October 14, 2005, from http://nces. ed.gov/programs/digest/d04/lt2.asp#c2_1 U.S. Department of Education, Office of Post Secondary Education (OPSE). (2005). The every classr oom Retrieved October 14, 2005, from http://www.ed.gov/about/ reports/annual/ teachprep/2005Title2 Report.doc Vanfleet, K. P. (2002). An analysis of differentiated instruction applied to teaching academics in secondary schools as a means to facilitate the inclusion of students with mild to moderate disabilities in the general education classroom (Doctoral dissertation, The University of Alabama, 2002). Webster Stratton, C. (2003). Quality training, supervision, on going monitoring, and agency support: Key ingredients to implementing The Incredible Years programs with fidelity. Retrieved October 9, 2005, from The Incredible Years website: http://www.incredible years.com/research/ article quality key ingredients fidelity 04.pdf Wertheim, C., & Leyser, Y. (200 2). Efficacy beliefs, background variables, and differentiated instruction of Israeli prospective teachers. The Journal of Educational Research 96 (1), 54 63.
169 Westling, D. L., Cooper Duffy, K., Prohn, K., Ray, M., & Herzog, M. J. (2005, May/June). Building a teacher support program. Teaching Exceptional Children 37 (5), 8 13. Willis, S., & Mann, L. (2000, Winter). Differentiating instruction: Finding manageable ways to meet individual needs. Curriculum Update. Retrieved November 10, 2004 from Association f or Supervision and Curriculum Development (ASCD) website: http://www.ascd.org/cms/objectlib/ascdframeset/index.cfm?publication
171 Appendix A Strategies for Differentiating Instruction 1. Stations: The use of stations involves setting up different spots in the classroom where students work on various tasks simultaneously. These stations invite flexible grouping because not all students need to go to all stations all the time. 2. Compacting : This strategy encourages teachers to assess stude nts before beginning a unit of study or development of a skill. Students who do well on the pre assessment do not continue work on what they already know. 3. Agendas: Agendas are personalized lists of tasks that a student must complete in a specified time, us ually two to three weeks. 4. Complex Instruction: This strategy uses challenging materials, open ended tasks, and small instructional groups. Teachers move among the groups as they work, asking students questions and probing their thinking. 5. Orbital Studies: These independent investigations, generally lasting three to six weeks, revolve around some facet of the curriculum. Students select their own topics, and they work with guidance and coaching from the teacher. 6. Entry Points: This strategy from Howard Gardne r proposes student exploration of a given topic through as many as five avenues: narrational (presenting a story), logical quantitative (using numbers or deduction), foundational (examining philosophy and vocabulary), aesthetic (focusing on sensory feature s), and experiential (hands on). 7. Problem Based Learning: This strategy places students in the active role of solving problems in much the same way adult professionals perform their jobs. 8. Choice Boards: With this strategy, work assignments are written on ca rds that are placed in hanging pockets. By asking a student to select a card from a particular row of pockets, the teacher targets work toward student needs yet allows student choice. 9. 4MAT: Teachers who use 4MAT plan instruction for each of four learning p references over the course of several days on a given topic. Some lessons focus on mastery, some on understanding, some on personal involvement, and some on synthesis. Note. From The Differentiated Classroom: Responding to the Needs of All Learners, by Carol Ann Tomlinson, 1999, Alexandria, VA: Association for Supervision and Curriculum Development, p. 75 93. Reprinted with permission from ASCD, February 2006.
172 Appendix B Comparison of the Traditional vs. Differentiated Classroom Traditional Classroom Differentiated Classroom Student differences are masked or acted upon when problematic Assessment is most common at the end A relatively narrow sense of intelligence prevails A single definition of excellence exists Student interest is infrequently tapped Relatively few learning profile options are taken into account Whole class instruction dominates Coverage of texts and curriculum guides drives instruction Mastery of facts and skills out of contex t are the focus of learning Single option assignments are the norm Time is relatively inflexible A single text prevails Single interpretations of ideas and events may be sought The teacher directs student behavior The teacher solves p roblems The teacher provides whole class standards for grading A single form of assessment is often used Student differences are studied as a basis for planning Assessment is on going and diagnostic in order to make instruction more responsive to learner needs Focus on multiple forms of intelligence is evident Excellence is defined in large measure by individual growth from a starting point Students are frequently guided in making interest based learning choices Many learning profile options are provided Many instructional arrangements are used Student readiness, interest, and learning profile shape instruction Use of essential skills to make sense of and understand key concepts and principles is the focus of learning Multi option assignm ents are frequently used Time is used flexibly according to student need Multiple materials are provided Multiple perspectives on ideas and events are sought becoming more self reliant learners Students help other students solve problems Students work with the teacher to establish both whole class and individual learning goals Students are assessed in multiple ways Note From The Differentiated Classroom: Responding to the Needs of All Learners, by Carol Ann Tomlinson, 1999, Alexandria, VA: Association for Supervision and Curriculum Development, p. 16. Reprinted with permission from ASCD, February 2006.
173 Appendix C Qualities of a Supportive Classroom Environment for Differentiation A supportive classroom environment is vital to your success in differentiating instruction. Such an environment: Promotes acceptance of differences Affirms that all students have learning strengths Acknowledged that students learn at different rates and in different ways Recogni zes that for work to be fair, it must sometimes be different Acknowledges that success means different things to different people Allows students to work with various people for various purposes Recognizes that the key to motivation is interest, and that a ll students have different interests Promotes personal responsibility for learning Builds feelings of personal competence and confidence in learning Nurtures skills of independence Supports and celebrates student success in challenging work preferences Nurtures the creative spirit in all students Note From Differentiating Instruction in the Regular Classroom: How to Reach an d Teach All Learners, Grades 3 12, by Diane Heacox, 2002, Minneapolis, MN: Free Spirit Inc, p. 12 13. Reprinted with permission from ASCD, February 2006.
174 Appendix D Differentiated Instruction: Fidelity Implementation Tool (DI: FIT) Teacher: ____________ _________________ School: _____________________________ Observer: __________________________________ Date: __________________ Class: _________________________ Lesson: _________________________________________________________ # DI Strategy Evaluator Ta sk Circle One 1 Teacher ensures students understand the purpose of the lesson Observe if agenda and/ or objectives are posted or ask a student to explain the objective of the lesson. Goal(s) is (are) visible, two out of three students can verbalize the objective of the lesson 1 0 2 Teacher creates respectful assignments Observe the lesson The observed lesson is designed with at least one clear example of cultural, ethnic, and linguistic sensitivity 1 0 3 Teacher creates respectful assign ments Observe the lesson Assignments are designed so all students are working toward the same goal and/ or understanding. 1 0 4 Lesson is differentiated by content, product, or process Observe the lesson The observed lesson is differentiated by conte nt, product, or process 1 0 5 readiness, interests, or learning profiles Observe the lesson The observed lesson is designed to address readiness, interests, or learning profiles 1 0 6 Visible use of suppo rts Observe the lesson The teacher employs clear examples of supports e.g., organizers, peers, manipulatives, technology 1 0 7 Class functions as a community Observe the lesson There is evidence in the classroom of at least two of the following: positive reinforcement by peers, cooperative learning activities, teacher student collaboration, student student collaboration, peer support 1 0 8 Students demonstrate genuine interest in learning Observe the lesson Eighty percent or more of students appear to be engaged and interested in the lesson e.g., asking questions, participating, interacting with others 1 0 9 Teacher uses anchor activities Observe the lesson The teacher has established activities and routines for students who are finished 1 0 10 Te acher acts as a facilitator Observe the lesson The teacher guides learning and nurtures student independence 1 0 NA 11 Teacher promotes acceptance of differences Observe the lesson differences if there is an opportunity 1 0 NA
175 Appendix D (Continued) # DI Strategy Evaluator Task Circle One 12 All assignments provide a slight challenge for learners Ask the teacher to justify how he/she decided and designed the difficulty of the ass ignments for each child and/ or group Teacher can verbalize criteria and reasons for how the assignments were created and assigned. Answer either mentions or alludes that students will be challenged 1 0 13 Lesson is centered on key concepts or essential learning Ask the teacher to explain the purpose of the lesson and how he/she decided on what to include or not include Teacher can either verbalize or provide recent* lesson plans that specify the standards and/ or justify lesson importance 1 0 14 Teacher uses flexible and purposeful grouping Ask the teacher how he/ she selects student groups and to give examples Teacher can give multiple, recent* examples of diverse and flexible student groupings 1 0 15 Use of learning stations and/ or independ ent study Ask the teacher to describe or provide a recent* DI lesson plan that employed stations or independent study Teacher can provide recent* examples of student learning stations and/ or independent study 1 0 16 Students are given meaningful, learn ing choices Ask the teacher to describe or provide a recent* DI lesson plan that employed student choice options Teacher can provide recent* examples of meaningful, learning choices provided to students 1 0 17 Teacher provides a variety of product opti ons Ask the teacher to describe or provide a recent* DI lesson plan that employed product options Teacher can provide recent* examples of a variety of product options provided to students 1 0 18 Teacher uses a variety of assessment tools Ask the teach er to describe and provide examples of the types of assessment used Teacher can provide recent* proof of multiple forms of assessment used, e.g., portfolios, rubrics, traditional tests 1 0 19 Teacher utilizes on going assessment Ask the teacher to expla in how he/she uses assessment to guide instruction Teacher can provide specific examples of how he/ she used recent* assessments to guide instruction 1 0 20 Teacher and students collaborate in the learning process Ask the teacher how he/she involves t he students in the learning process Teacher provides students with the opportunity to be a stakeholder in the learning 1 0 recent (within one month) Total score (20 possible) Deborah Hellma n January 2006
176 Appendix E Chapman, C., & King, R. (2003). Differentiated instructional strategies for reading in the content area Thousand Oaks, CA: Corwin Press. Coil, C., & Merritt, D. (20 01). Solving the assessment puzzle: Piece by piece Marion, IL: Pieces of Learning. Cramer, K., Twyman, S., & Winholtz, W. (1998). 61 cooperative learning activities for science classes Portland, ME: J. Weston Walch. Dodge, J. (2006). Differentiation in a ction: A complete resource with research supported strategies to help you plan and organize differentiated instruction and achieve success with all learners. New York, NY: Scholastic, Inc. Drapeau, P. (2004). Differentiated instruction: Making it work : A practical guide to planning, managing, and implementing differentiated instruction to meet the needs of all learners. New York, NY: Scholastic, Inc. Forsten, C. (2003). Differentiating textbooks: Strategies to improve student comprehension & motivation. Pe terborough, NH: Crystal Springs Books. Forsten, C., Grant, J., & Hollas, B. (2002). Differentiated instruction: Different strategies for different learners Peterborough, NH: Crystal Springs Books. Heacox, D. (2002). Differentiating instruction in the regu lar classroom: How to reach and teach all learners, Grades 3 12 Minneapolis, MN: Free Spirit Inc.
177 Appendix E (Continued) Hollas, B. (2005). Differentiating instruction in a whole group setting Peterborough, NH: Crystal Springs Books. Paterson, K. (2005) Differentiated learning: Language and literacy projects that address diverse backgrounds and cultures Markham, Ontario: Pembroke Publishers. Tilton, L. (2003). tips, tools, and techn iques Shorewood, MN: Covington Cove Publications. Tomlinson, C. A. (1999). The differentiated classroom: Responding to the needs of all learners. Alexandria, VA: Association for Supervision and Curriculum Development. Tomlinson, C. A. (2001). How to diffe rentiate instruction in mixed ability classrooms (2 nd Ed.). Alexandria, VA: Association for Supervision and Curriculum Development. Tomlinson, C. A. (2003). Fulfilling the promise of the differentiated classroom: Strategies and tools for responsive teachi ng Alexandria, VA: Association for Supervision and Curriculum Development. Tomlinson, C. A. & Allan, S. D. (2000). Leadership for differentiating schools & classrooms Alexandria, VA: Association for Supervision and Curriculum Development. Wormeli, R. (20 06). Fair Isn't Always Equal: Assessing & Grading In the Differentiated Classroom. Portland, ME: Stenhouse Publishers.
178 Appendix F Facilitated Support Group Feedback Form Name: _____________________________________ Date: ___________________ I have read t Yes ___ No ___ I agree that the minutes are accurate. Yes ___ No ___ Corrections needed: ________________________________________________________________________ ________________________________________ ________________________________ ________________________________________________________________________ ________________________________________________________________________ Suggestions for next month: ___________________________________________________ _____________________ ________________________________________________________________________ ________________________________________________________________________
179 Appendix G The Differentiated Classroom Observation Form Check the appropriate box n ext to each item. Use the comment box to provide ideas for improvement in specific areas. If the form is completed during multiple observations, use tally marks. Review the results with the teacher as soon as possible to identify specific areas for improve ment and to praise strengths. Teacher: Grade Level/Subject Area: Observer: Date: Evidence of Implementation Often Sometimes Little or no Comments PHYSICAL ENVIRONMENT Presents an inviting, relaxed environment for learning Provides comfortable desks and work areas Contains individual, designated personal spaces for extra books and other items Is designed for quick and easy groupings of tables and chairs Is arranged for teacher and student movement during work sessions Provides work areas for individual needs, including knowledge/ability levels Reflects current content or skills through student displays and artifacts TEACHER BEHAVIORS Works with total groups, individuals, and small groups Monitors individuals and small groups Uses a variety of ongoing assessment tools such as checklists, surveys, and anecdotal records Applies assessment information to guide instruction Addresses academic, emotional, social, and physical st udent needs Provides time for students to actively process information Gives specific feedback to individuals and/or small groups STUDENT ENGAGEMENT Exhibits on task behavior while working alone Works effectively in small groups Wo rks on their individual knowledge or ability levels Feels respected and emotionally safe Uses self discipline MATERIALS/ RESOURCES Includes a variety of reading levels rela ted to the subject or topic Are accessible to students Supports the standards and topic Are age appropriate Are up to date Are available in an adequate number for the class size Include appropriate reference sources and materi als INSTRUCTIONAL STRATEGIES Uses a variety of assessment tools before, during, and after learning Uses a variety of instructional strategies and activities to teach standards Meets the divers needs of learners Engages students in vario us flexible grouping designs Uses centers and/or stations for individual and small group instruction Engages students with projects and/ or problem solving activities Presents students with choices in learning activities Note. From 11 P ractical Ways to Guide Teachers toward Differentiation (and an evaluation tool), by C. Chapman and R. King, 2005, Journal of Staff Development, 26 (4), p. 24. Used with permission of the National Staff Development Council, www.nsdc.org, 2007. All rights res erved.
18 0 Appendix H Teacher/Peer Reflection on Differentiation The following scale may be useful (1) little or no evidence (2) to some degree (3) demonstrates competence (4) demonstrates proficiency (5) demonstrates exemplary performan ce GENERAL Pre assesses students to determine level 1 2 3 4 5 of understanding. Assesses student interests. 1 2 3 4 5 1 2 3 4 5 Develops a student centered classroom. 1 2 3 4 5 Ensures re spectful assignments for all learners. 1 2 3 4 5 Consistently uses flexible grouping. 1 2 3 4 5 Varies the pace of learning for varying 1 2 3 4 5 learner needs. Utilizes active learning. 1 2 3 4 5 Demonstrates escalating expectat ions. 1 2 3 4 5 1 2 3 4 5 progress. CONTENT Differentiates using major concepts and 1 2 3 4 5 generalizations. Uses a variety of materials other than the 1 2 3 4 5 standard text. Var ious support mechanisms (e.g., reading 1 2 3 4 5 buddies, organizers, study guides). PROCESS Activities necessitate that students do something 1 2 3 4 5 with their knowledge (apply and extend major concepts and generalizations as opposed to just repeating it back). Uses higher level tasks for all learners (e.g., 1 2 3 4 5 application, elaboration, providing evidence, synthesis) to provide appropriate challenge.
181 Appendix H (Continued) PROCESS (continued) Uses tiered activities. 1 2 3 4 5 Activities involve all learners in both critical 1 2 3 4 5 and creative thinking. Varies tasks along continuum of the equalizer. 1 2 3 4 5 Varies tasks by students interests. 1 2 3 4 5 Varies tasks by learner profile. 1 2 3 4 5 PRODUCT Provides opportunities for student products to 1 2 3 4 5 be based upon the solving of real and relevant problems. Allows for a wide range of product alternatives 1 2 3 4 5 (e.g., oral, visual, kinesthetic, musical, spatial, creative, practical). Product assignments differ based on individual 1 2 3 4 5 (or group) readiness, learning needs, and interest. Teacher supports students in using a wide range 1 2 3 4 5 of varied resources. Product assignment necessitates that students 1 2 3 4 5 conduct research. Product assignment balances structure and choice. 1 2 3 4 5 Encourages students to use different avenues of 1 2 3 4 5 exploration and a variety of media. Works with individual students (or groups) to 1 2 3 4 5 determine what form the product will take. Necessitates that students apply key 1 2 3 4 5 understandings and skills of the subject to their own interest areas. Works with individual students to apply key 1 2 3 4 5 unders tandings and skills of the discipline by which the product will be judged. Uses both formative and summative evaluation. 1 2 3 4 5
182 Appendix H (Continued) INSTRUCTIONAL/MANAGEMENT STRATEGIES Uses compacting. 1 2 3 4 5 Uses student lear ning contracts. 1 2 3 4 5 Uses independent study. 1 2 3 4 5 Uses interest centers/groups. 1 2 3 4 5 Uses learning centers/groups. 1 2 3 4 5 Uses various instructional strategies to 1 2 3 4 5 differentiate (e.g. organize rs, cubing, etc.). Uses high level cooperative strategies (e.g., 1 2 3 4 5 complex instruction, group investigation). Other _________________________________ 1 2 3 4 5 Other _________________________________ 1 2 3 4 5 Note. Fr om Leadership for Differentiating Schools & Classrooms by C. A. Tomlinson and S. D. Allan, 2000, Alexandria, VA: Association for Supervision and Curriculum Development, pp. 144 146. Reprinted with permission from ASCD, February 2006.
183 Appendix I DI Suppo rt Group Feedback Was the number of meetings sufficient? Too many? Too few? _____________________ _______________________________________________________________________ _______________________________________________________________________ Was the lengt h of the support group meetings effective? Too short? Too long? ________________________________________________________________________ ________________________________________________________________________ ____________________________ ____________________________________________ Was the discourse, professional discussion with colleagues, helpful or not? How? Be specific. ________________________________________________________________ ____________________________________________________ ____________________ ________________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________ Was the DI confer ence area on your e mail desktop helpful? Explain. ________________________________________________________________________ ________________________________________________________________________ ____________________________________________________________ ____________ ________________________________________________________________________ Has participation in the on going support group helped to sustain your implementation of differentiated instruction? __________________________________________________ __ ______________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________ _________________________________________ _______________________________
184 Appendix I (Continued) What attributes of the DI Support Group Meetings do you think were helpful and would like to see continued? What areas or qualities of the DI Support Group Meetings do you think we need to change or improve on? Helpful or Positive Aspects to Keep Suggested Changes or Improvements
185 Appendix J Informed Consent to Participate in Research Information to Consider Before Taking Part in this Research Study Researchers at the University of South Florida (USF) study many topics. To do this, we need the help of people who agree to take part in a research study. This form tells you about this research study. We are asking you to take part in a research study that is called: Implementing Differentiat ed Instruction in Urban, Title I Schools: Effects of Facilitated Support Groups and Program Fidelity on Student Achievement The person who is in charge of this research study is Deborah Hellman. Other research personnel who you may be involved with includ e: ( names removed for security ). The research will be done at School A and School B ( names removed for security ). Purpose of the study The purpose of this research study is to collect and analyze data on the impact of differentiated instruction, facilitat ed support groups, and implementation fidelity on student achievement. Study Procedures You have been selected as one of 52 possible participants for this study. If you are selected as one of the approximately 26 teachers in the treatment group, you will participate in a series of five facilitated support groups that will last approximately 120 minutes each. You will have access to a library of support materials and will serve as part of the teacher support/ study group. At the conclusion of each meeting, you will select one or more resources to review for 30 minutes and then select at least one strategy to implement in the next three to four weeks in your classroom. At the next meeting, you will share their experience and provide feedback to the group conc erning the strategy implemented. During these sessions, the conversations will be digitally recorded to ensure accuracy of the information and comments collected, no names will be used in the final reports. Once per nine weeks, the principal investigator a nd/ or research assistant will observe you to determine the degree of fidelity that you demonstrate with respect to differentiated instruction. If you are selected as one of the approximately 26 teachers in the control group, you will carry out the normal requirements of your teaching position and you will be observed once per nine weeks to determine the degree of fidelity that you demonstrate with respect to differentiated instruction.
186 Appendix J (Continued) At the end of the school year, student performa nce data will be collected and analyzed from your students and the students of the other participants in the study. The effect of teacher fidelity to the differentiated instruction model will also be analyzed. Data collection and analysis will last from Ja nuary 2007 until December 2007. Alternatives You have the alternative to choose not to participate in this research study. Benefits The potential benefits to you are: by taking part in this research study, you may increase our overall knowledge of diffe rentiated instruction strategies. If you are selected as a participant in the treatment group you will be provided with direct support services, strategies, and access to resources. The principal investigator will also be available as a mentor throughout t he process. If you are selected as part of the control group, you will help to add to the body of evidence of implementation of evidence based practices. Risks or Discomfort There are no known risks to those who take part in this study Compensation We will not pay you for the time you volunteer while being in this study. Confidentiality We must keep your study records confidential. The data obtained from you will be combined with data from others in the publication. The published results will not inc lude your name or any other information that would personally identify you in any way. All written data, audiotapes, and videotapes will be anonymously coded and stored in a stored for After that time, they will be shredded. However, certain people may need to see your study records. By law, anyone who looks at your records must ke ep them completely confidential. The only people who will be allowed to see these records are: The research team, including the Principal Investigator, study coordinator, and all other research staff. Certain government and university people who need to know more about the study. For example, individuals who provide oversight on this study may need to look at your records. This is done to make sure that we are doing the study in the right way. They also need to make sure that we are protecting your righ ts and your safety.) These include:
187 Appendix J (Continued) o The University of South Florida Institutional Review Board (IRB) and the staff that work for the IRB. Other individuals who work for USF that provide other kinds of oversight may also need to loo k at your records. o The Florida Department of Health, people from the Food and Drug Administration (FDA), and people from the Department of Health and Human Services (DHHS). o recor ds to make sure the study was done correctly. We may publish what we learn from this study. If we do, we will not let anyone know your name. We will not publish anything else that would let people know who you are. Voluntary Participation / Withdrawal Y ou should only take part in this study if you want to volunteer. You should not feel that there is any pressure to take part in the study, to please the investigator or the research staff. You are free to participate in this research or withdraw at any tim e. There will be no penalty or loss of benefits you are entitled to receive if you stop taking part in this study. Your decision to participate or not to participate will not affect your or job status. If you have any questions, concerns or complaints abou t this study, call Deborah Hellman at 493 3302. If you have questions about your rights, general questions, complaints, or issues as a person taking part in this study, call the Division of Research Integrity and Compliance of the University of South Flori da at (813) 974 9343. If you experience an adverse event or unanticipated problem call Deborah Hellman at 493 3302. Consent to Take Part in this Research Study It is up to you to decide whether you want to take part in this study. If you want to take par t, please sign the form, if the following statements are true. I freely give my consent to take part in this study. I understand that by signing this form I am agreeing to take part in research. I have received a copy of this form to take with me. Signature of Person Taking Part in Study Date Printed Name of Person Taking Part in Study
188 Appendix J (Continued) Statement of Person Obtaining Informed Consent I have carefully explained to the person taking part in the study what he or she can expect. I hereby certify that when this person signs this form, to the best of my knowledge, he or she understands: What the study is about. What procedures/interventions/investigational drugs or devices will be used. What the potential benefits m ight be. What the known risks might be. I also certify that he or she does not have any problems that could make it hard to understand what it means to take part in this research. This person speaks the language that was used to explain this research. This person reads well enough to understand this form or, if not, this person is able to hear and understand when the form is read to him or her. This person does not have a medical/psychological problem that would compromise comprehension and therefore makes it hard to understand what is being explained and can, therefore, give informed consent. This person is not taking drugs that may cloud their judgment or make it hard to understand what is being explained and can, therefore, give informed consent. Signature of Person Obtaining Informed Consent Date Printed Name of Person Obtaining Informed Consent
189 Appendix K District In service Evaluation Summary: School A District Level Questions 1. Training content was appr opriate and built upon knowledge / experiences of intended participants. A. Strongly Agree 91% B. Agree 9% C. Undecided 0% D. Disagree 0% E. Strongly Disagree 0% 2. Content of training included information relevant and useful in my position. A. Strongly Agree 82% B. Agree 18% C. Undecided 0% D. Disagree 0% E. Strongly Disagree 0% 3. Training activities, assignments, and / or materials were related to course objectives. A. Strongly Agree 82% B. Agree 18% C. Undecided 0% D. Disagre e 0% E. Strongly Disagree 0%
190 Appendix K (Continued) 4. Trainer demonstrated knowledge and positive attitude toward content. A. Strongly Agree 92% B. Agree 8% C. Undecided 0% D. Disagree 0% E. Strongly Disagree 0% 5. Training environment w as appropriate for the course. A. Strongly Agree 90% B. Agree 10% C. Undecided 0% D. Disagree 0% E. Strongly Disagree 0% 6. What is the Primary Purpose of this course? A Add on Endorsement 10% B Alternate Certification 0% C Florida Educators Certificate Renewal 36% D Other Professional Certificate/License renewal 0% E Professional Skill Building 54% 7. Is this training aligned with your Individual Professional Development Plan (IPDP)? A Yes 100% B No 0% C N/A (Instructional S upport Only) 0%
191 Appendix L District In service Evaluation Summary: School B District Level Questions 1. Training content was appropriate and built upon knowledge / experiences of intended participants. A. Strongly Agree 90% B. Agree 10% C. Undeci ded 0% D. Disagree 0% E. Strongly Disagree 0% 2. Content of training included information relevant and useful in my position. A. Strongly Agree 80% B. Agree 20% C. Undecided 0% D. Disagree 0% E. Strongly Disagree 0% 3. Training activities assignments, and / or materials were related to course objectives. A. Strongly Agree 90% B. Agree 10% C. Undecided 0% D. Disagree 0% E. Strongly Disagree 0%
192 Appendix L (Continued) 4. Trainer demonstrated knowledge and positive attitude towar d content. A. Strongly Agree 90% B. Agree 10% C. Undecided 0% D. Disagree 0% E. Strongly Disagree 0% 5. Training environment was appropriate for the course. A. Strongly Agree 70% B. Agree 30% C. Undecided 0% D. Disagree 0% E. Strongly Di sagree 0% 6. What is the Primary Purpose of this course? A Add on Endorsement 0% B Alternate Certification 0% C Florida Educators Certificate Renewal 20% D Other Professional Certificate/License renewal 0% E Professional Skill Building 80% 7 Is this training aligned with your Individual Professional Development Plan (IPDP)? A Yes 90% B No 10% C N/A (Instructional Support Only) 0%
About the Author d has been a special education and inclusion co teacher for 27 years and is Nationally Board Certified. Mrs. Hellman is currently a district trainer for differentiated i nstruction, collaborative teaching, and teaming. In addition, she actively mentors other teachers and pre service teachers as an ESE Mentor, Model Classroom Teacher, and Professional Practice Partner with USF. She has received many honors and scholarships including Teacher of the Year, CEC Teacher of the Year, the Landis M. Stetler ESE Leadership Scholarship, and the Cathy Lynn Richardson Doctoral Scholarship. She began the doctora l program in the fall of 2004 and has been a graduate teaching instructor for EEX4070, Integrating the Exceptional Child for the past year.