Copyright by Florence Linelle Clark 2002 The Dissertation Committee for Florence Linelle Clark Certifies that this is the approved version of the following dissertation: ZERO-TOLERANCE DISCIPLINE: THE EFFECT OF TEACHER DISCRETIONARY REMOVAL ON URBAN MINORITY STUDENTS Committee: Lonnie H. Wagstaff, Supervisor Jay D. Scribner William Moore, Jr. Charles T. Clark Joao H. Vargas ZERO-TOLERANCE DISCIPLINE: THE EFFECT OF TEACHER DISCRETIONARY REMOVAL ON URBAN MINORITY STUDENTS by Florence Linelle Clark, B.A., M.S. Dissertation Presented to the Faculty of the Graduate School of The University of Texas at Austin in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy The University of Texas at Austin August, 2002 Dedication This dissertation is dedicated to my children, Courtney Marie, and Julius Wesley Clark. A special dedication is extended to my daughter Courtney who endured my long hours of study and frequent absences. I also dedicate my accomplishment to Gigi Bryant, and Rosie Luiz who acted as a surrogate mother to Courtney during my time of need. This work is dedicated to all the black and brown children who are at-risk of being subjected to the web of zero-tolerance. I hope this study will honor their tenacity to rise above the injustice. Acknowledgements While the road to this point has not been without bumps and curves due to my many competing commitments, I could not have successfully made this journey without the support from the many Professors within the Department of Education Administration. I wish to specifically acknowledge Dr. Lonnie Wagstaff whose high expectations helped make me become a better researcher and writer. Dr. Wagstaff exemplified good teaching, and as his protégé, I am truly grateful for his instructive modeling and constructive critique of my work. Dr. Wagstaff demanded the best from me, and for this I am eternally appreciative. Dr. Jay Scribner my mentor and friend, who from the first day I entered the doctorate program provided me with continuous encouragement and support. Words could not describe the appreciation and the lessons learned from his words of wisdom and guidance. Dr. Scribner taught me the importance of the continuation of my research. Dr. Terry Clark and Sung-kwan Yang of the Education Productivity Council, I wish to express my sincere appreciation for the tremendous assistance and time you both provided me to organize and test the data. The timeliness of my completion of the dissertation and the quality of this product is directly attributed to v their technical assistance. I am honored to not only to have received such wonderful support from Dr. Clark and Sung-kwan, but more honored to be able to call them my friends. Dr. Moore and Dr. Vargas, I wish to acknowledge for exposing me to social and value issues that confront educational leaders. The course materials presented in Societal Factors and Critical Race Theory helped shape the framework for this research study, and forced me to reexamine my ontology. vi ZERO-TOLERANCE DISCIPLINE: THE EFFECT OF TEACHER DISCRETIONARY REMOVAL URBAN MINORITY STUDENTS Publication No._____________ Florence Linelle Clark, Ph.D. The University of Texas at Austin, 2002 Supervisor: Lonnie H. Wagstaff The principal theoretical proposition from which this work begins is that schools are a place of cultural reproduction and that school discipline policies are racialized resulting in a negative consequence for African American and Hispanic children. The purpose of this study is to determine the effect of teacher discretionary removal on urban minority students. This research examined the impact for African American and Hispanic elementary students when a state vii expands its zero tolerance discipline policy to allow for the discretionary removal of students from the traditional learning environment. The research questions were: (1) What are the trends in student suspension rate subsequent to the 1996 adoption of a discretionary discipline removal policy, and the general socio-demographic characteristics of students who are subjected to a discretionary removal? (2) Are the achievement test scores, Texas Academic Achievement (TAAS) / Texas Learning Index (TLI) reading and math scores, of student's subjected to a discretionary removal significantly different from the test scores of comparable peer students not subjected to a discretionary removal? (3) Is there a relationship between discretionary referral reason and ethnicity? The methodology used for this study was a quasi-experimental approach. The district selected for this study was a large urban Texas school district. Data used for this study was obtained from records maintained by the school district's alternative education facility for students suspended because of a discretionary removal. Descriptive and inferential statistics were used to examine relationships between variables and the application of discretionary removal and suspension. viii The results of the investigation clearly indicated African American and Hispanic students were subjected to discretionary removals at higher rates than non-minorities. The findings determined numerous trends in discretionary removal based on student ethnicity, gender, special education identification, and socioeconomic status. The study also determined a negative relationship between discretionary removal and academic achievement growth. On average, students subjected to a discretionary removal had lower achievement test scores in comparison to peers not subjected to a discipline removal. ix Table of Contents List of Tables ........................................................................................ xiii List of Figures ..........................................................................................xv CHAPTER 1 Introduction................................................................................................1 School Discipline Policies & Theories of Education............................1 Discipline Policies & School Safety ....................................................4 Zero- Tolerance Discipline Policies & School Safety ..........................5 The Goals of Zero-Tolerance Discipline Policies.................................7 Statement of the Problem ...........................................................................8 Purpose ....................................................................................................13 Research Questions ..................................................................................15 Definition of Terms..................................................................................16 Importance of the Study ...........................................................................19 CHAPTER 2 Review of Literature.................................................................................23 Zero-Tolerance Discipline Policies & Academic Achievement..........25 Zero-Tolerance Discipline & The Psychological Implications ...........30 Zero-Tolerance Discipline & Disproportional Application ................33 Socioeconomic Status & Discipline................................................................35 Gender& Discipline..........................................................................37 Ethnicity & Discipline ......................................................................38 Referral Reasons..............................................................................41 Summary..................................................................................................43 CHAPTER 3 Research Design & Methodology .............................................................44 Design .............................................................................................47 Participants ..............................................................................................49 Schools............................................................................................49 Subjects...........................................................................................50 x Data Collection................................................................................51 Research Questions ..................................................................................56 Procedures................................................................................................57 T-test ...............................................................................................57 Multiple Regression Analysis ..........................................................58 Limitations...............................................................................................61 CHAPTER 4 Findings ...................................................................................................63 Section-1 (Trends in Application)...........................................................64 Enrollment, Removal, Ethnicity & Gender.......................................64 Socioeconomic Status......................................................................73 Special Education ............................................................................79 Grade Level.....................................................................................84 Section-2 (Academic Implications).........................................................89 T-test Findings ...............................................................................96 Achievement & Ethnicity ...............................................................98 Achievement & Socio-economic Status........................................101 Achievement & Special Education ...............................................104 Regression Model ........................................................................107 Section-3 (Referral Reasons...... ...... ..............................................114 Administrator Response to Zero-Tolerance and Discretionary Removal .....................................................................................123 Common Themes........................................................................132 CHAPTER 5 Conclusions.............................................................................................133 Introduction..................................................................................133 Ethnicity & Gender ......................................................................134 Socioeconomic Status ..................................................................135 Special Education Identification ...................................................136 Grade Level .................................................................................137 Academic Achievement & Discretionary Removals .....................138 xi Discretionary Referral Reason & Ethnicity...................................140 Discretionary Removals and Other Factors...................................142 Theoretical Implications...............................................................143 Policy Implications..................................................................................144 Future Research ......................................................................................146 General Comments.................................................................................147 REFERENCES ......................................................................................149 VITA .....................................................................................................159 xii List of Tables Table 4.1. Table 4.2. Table 4.3 Table 4.4. Table 4.5. Table 4.6. Table 4.7. Table 4.8. Table 4.9. Enrollment by Ethnicity & Number of Discretionary Removals .....67 Discretionary Removal by Gender .................................................72 Discretionary Removal by Socio-economic Status .........................73 1996-97 Math & Reading Independent Sample t-Test ....................95 1997-98 Math & Reading Independent Sample t-Test ...................95 1998-99 Math & Reading Independent Sample t-Test ....................95 Discretionary Removal on Mean Math Scores ...............................97 Discretionary Removal on Mean Reading Scores...........................97 Discretionary Removal by Ethnicity on Mean Math Scores ..........100 Table 4.10. Discretionary Removal & Socioeconomic Status on Mean............103 Math Scores Table 4.11. Effect of Discretionary Removal & Socioeconomic Status on Mean Reading Scores .............................................................................103 Table 4.12. Effect of Discretionary Removal & Special Education Status on Mean Math Scores ......................................................................106 Table 4.13. Effect of Discretionary Removal & Special Education Status on Mean Reading Scores...................................................................106 xiii Table 4.14. Results Predicting 1999 Math Scores ........................................111 Table 4.15. Results Predicting 1999 Reading Scores ...................................113 Table 4.16. 1996-97 ANOVA (# Referrals by Ethnicity) .............................117 Table 4.17. 1997-98 ANOVA (# Referrals by Ethnicity) .............................117 Table 4.18. 1998-99 ANOVA (# Referrals by Ethnicity) .............................118 Table 4.19. 1999-00 ANOVA (# Referrals by Ethnicity) ..............................118 xiv List of Figures Figure IV-1 Discretionary Removals 1996-97 ...............................................68 Figure IV-2 Discretionary Removals 1997-98 ...............................................69 Figure IV-3 Discretionary Removals 1998-99 ...............................................70 Figure IV-4 Discretionary Removals 1999-00 ................................................71 Figure IV-5 Socio-economic Status by Ethnicity 1996-97 .............................75 Figure IV-6 Socio-economic Status by Ethnicity 1997-98 .............................76 Figure IV-7 Socio-economic Status by Ethnicity 1998-99 ..............................77 Figure IV-8 Socio-economic Status by Ethnicity 1999-00 ..............................78 Figure IV-9 Special Education Status by Ethnicity 1996-97 ...........................80 Figure IV-10 Special Education Status by Ethnicity 1997-98 .........................81 Figure IV-11 Special Education Status by Ethnicity 1998-99 .........................82 Figure IV-12 Special Education Status by Ethnicity 1999-00 .........................83 Figure IV-13 Discretionary Removal by Grade Level and Ethnicity 1996-97 ...........85 Figure IV-14 Discretionary Removal by Grade Level and Ethnicity 1997-98..............86 Figure IV-15 Discretionary Removal by Grade Level and Ethnicity 1998-99 ..............87 xv Figure IV-16 Discretionary Removal by Grade Level and Ethnicity 1999-00 ..............88 Figure IV-17 Referral Status & Academic Achievement 1997........................91 Figure IV-18 Referral Status & Academic Achievement 1998........................92 Figure IV-19 Referral Status & Academic Achievement 1999........................93 Figure IV-20 Type of Offenses 1996-97.......................................................119 Figure IV-21 Type of Offenses 1997-98.......................................................119 Figure IV-22 Type of Offenses 1998-99.......................................................120 Figure IV-23 Type of Offenses 1999-00.......................................................120 Figure IV-24 Offenses by Ethnicity 1996-97 ................................................121 Figure IV-25 Offenses by Ethnicity 1997-98 ................................................121 Figure IV-26 Offenses by Ethnicity 1998-99 ................................................122 Figure IV-27 Offenses by Ethnicity 1999-00 ................................................122 xvi CHAPTER 1 Introduction School Discipline Policies & Theories of Education The principal theoretical proposition from which this work begins is that schools are a place of cultural reproduction and that school discipline policies are racialized resulting in a negative consequence for African American and Hispanic children. Discipline policies are a means of enforcing rules and habits so that children will develop into adults who conform to the general social expectations of mainstream culture, absorbing its basic attitudes and beliefs and thus becoming law-abiding citizens. From a social theorist perspective, schools are expected to produce students who will be in harmony with the values and needs of society and collective interest. Therefore, it comes as no surprise when schools are viewed as unsafe and classrooms unruly the public's demand for reform is inevitable and social policies designed to control behavior is forthwith. This demand for reform has recently come in the form of zero-tolerance for student behavior deemed unsafe or disruptive to the learning environment. Hence, legislation addressing school 1 safety expounding what student behaviors will not be tolerated flows from the federal to state level of government and subsequently to local school districts and school campuses. While it is true schools have a political and social purpose, what occurs in public schools also has economic implications for society. For example, if a zerotolerance discipline policy limits the academic and social experiences of children and youth, the net result may be a decrease in the number of workers who can meet the education and technical demands of the market place. Hence, from a Weberian economic perspective, while schools may serve to educate and socialize the populace, schools may also help shape the type and color of the U.S. workforce. This potential control of the labor market access directly conflicts with civil right policies designed to enforce an equal education for all children. More importantly for minorities student education is seen as a mechanism of access to societal wealth. As noted by Fanon (1967), racist structures are permanently embedded in the psychology, economy, society, and culture of the modern world. As noted by Derrick Bell (1992), because bias is masked in unofficial practices and "neutral" standards one must question whether race or some individual failing is the reason 2 for the negative outcome for a person of color. Therefore, when examining the lack of academic success for children of color in light of Civil Rights laws, for African American and Hispanic children one can not help question if so called neutral discipline policies rather than undergird goals of equality, serve to increase social and economic inequality. From a critical race theoretical framework it appears that societal interest and current school zero-tolerance discipline policies that limit a student's access to the established curriculum serves to strengthen educators dominant position of power over students and their economic potential. As pointed out by Gary Orfield (2000), elementary schools are usually the first institutions to feel the impact of any major change in American society. Therefore, it comes as no surprise for elementary students, ones age and maturity does not exempt them from this lack of tolerance, but rather ones ethnicity might. In Chapter 2, a review of the literature, an effort will be made to elaborate how zero-tolerance discipline policies have become racilized and illustrate how schools continue to serve as institutions of cultural reproduction with negative implications for African American and Hispanic children caught in the web of zero3 tolerance. However, before examining the current findings that support this research, one must first understand why policy makers and educators reached a point of "zero-tolerance" for the behavior of children? In other words, why has student behavior traditionally addressed by the classroom teacher necessitated the intrusion of federal and state policy makers? Discipline Policies & School Safety Concerns about school safety are not new but when school shootings such as the one that occurred at Columbine High School are repeatedly portrayed on the nightly news the public cannot help but question, how safe are our schools? Violence not only poses a threat of physical injury it also threatens the entire educational process because of the disruption to teaching and learning (Hill & Drolet, 1999). When violence occurs at a school or when there is fear that violence shall occur at a school, both students and teachers are victimized. As a result, the public not only demands but also expects an immediate response to this real or perceived concern surrounding school safety. In response to this concern that schools are no longer safe, President Clinton signed into law the Gun-Free Schools Act (GFSA) in 1994. This law required states 4 to develop and enforce school safety policies and associated punishments to curtail school violence or risk the loss of federal funds. The GFSA act's definition of weapons includes guns, bombs, grenades, rockets, and missiles and broader definitions as prescribed by each State. While the 1994 Gun Free Schools Act (GFSA) required the expulsion for one calendar year of any student bringing a weapon to school, the law also allowed for a case-by-case modification of the one-year expulsion by the "chief administrative officer" of the school district. However, many school districts as illustrated in the literature, for whatever reason did not appear to readily utilize this clause. Rather many school districts appeared to have embraced a punitive, no exception, and no tolerance approach towards student discipline and expanded its application to include non-violent behavior. Zero- Tolerance Discipline Policies & School Safety The 1994 GFSA does not use the term "zero tolerance", the slogan "zero tolerance" first came to wide public attention in 1986 when the U.S. attorney in San Diego used the slogan to explain his seizure of boats carrying illegal drugs (Ficus, 2000). The slogan was later used by state policy makers and educators as a method 5 of defining school safety policies and demonstrating to a concerned public that school district's would not tolerate behavior deemed disruptive to school safety and orderliness. The threat of violence that makes the news is only part of the picture when there is a discussion surrounding school safety. Disorder, disrespect, vandalism and other relatively minor crimes are also constant problems that many schools are attempting to address (Skiba & Peterson, 1999). Therefore, in response to the perceived lax in school discipline and order, educators often address a wide range of student infractions by using a zero tolerance discipline policy. Unlike weapon and drug offenses that typically mandate an expulsion per federal statue, zero tolerance discipline actions imposed for other types of student behavior are classified as "discretionary" because each school district may define the types of behavior that will not be tolerated. As described by Skiba (1999) the term "zero tolerance" is being used more and more to describe the degree to which society is willing to accept (or tolerate) certain kinds of student behavior. Hence, by 1997, ninety percent of the nation's schools had zero tolerance policies for 6 firearms and other types of behavior deemed disruptive to the learning environment (Fiscus, 2000). The Goals of Zero-Tolerance Discipline Policies The two main goals of a zero tolerance discipline policy are: (1) ensure the safety of staff and students, and (2) create an environment conducive to learning. Therefore in response to these goals, many states have passed legislation that allows for the discretionary removal of students from the traditional academic setting for any behavior deemed disruptive to the learning environment. Noguera (1995) argues that a zero tolerance discipline policy is adopted less for its effectiveness than for its symbolic value. In other words, a zero tolerance discipline policy attempts to reassure administrators, parents, and teachers that strong actions are a necessary response to a perceived breakdown in school order. Proponents of zero-tolerance discipline policies appear to embrace a paternalistic-authoritarian approach towards student behavior, thus presenting the argument that students are immature beings that need to coerced to be educated and disciplined without regards to their rights (Morris, 1980). As we enter the 21st 7 century, the sentiment of proponents of a zero tolerance discipline policy appear to be saying, "we are going to create an atmosphere where students can learn, and if a student acts up, we'll move them out of there" (Portner, 1995, p.1). A "zero tolerance" discipline policy serves the objective of removing disruptive students so that other students can learn. It also serves the purpose of removing potentially violent dangerous students from the general population and thereby providing for a safe learning environment. As the debate surrounding safe schools continues, states such as Texas have established programs that segregate students viewed as discipline problems from the other members of the student body. Statement of the Problem Society often attempts to control and define acceptable behavior by imposing rules designed to maintain a rational and civilized society. The application of a "zero tolerance" discipline policy is one example of society's attempt to control student behavior and respond to the public's perception that schools are not safe. However, as with any rule there is often the risk of unintended consequences, as documented in the research literature highlighting the 8 significant controversy surrounding the use of "zero tolerance" discipline policies especially when the policy is broadened to allow for discretionary definition of unacceptable behaviors. Controversy related to the application of zero-tolerance discipline policies and the discretionary removal of students is often sparked by the continuum of offenses for which the policy is applied. For example, some school districts have extended zero tolerance to fighting, homework completion, general misbehavior, and disrespect just to name a few of the infractions cited as reasons for the removal of a student. Therefore, in these cases a zero tolerance discipline policy tends to punish both major and minor incidents (Sikba & Peterson, 1999). Research illustrates the importance of access to academic experiences through the curriculum, teachers, and other school activities are of particular importance for minority students who are already marginalized in school settings (Davis & Jordan, 1995; Finn & Cox, 1992; Irvine, 1990; Sanders & Reed, 1995). Therefore, any policy that involuntarily denies a student established instructional time and suspends the established curriculum even for brief periods of time should be subjected to strict scrutiny. As documented in the research, there is an academic 9 achievement "gap" between minority and non-minority students, specifically for African American and Hispanic students. While several researchers have offered many possible explanations for this achievement gap, there does not appear to be one specific reason. However, what is clear in the research is the need for students to have continuous access to the established curriculum and classroom instruction. Hence, a discretionary removal policy may unintentionally undermine the goal of closing the achievement gap. When discretionary removal policies are applied to all students regardless of age or maturity, particularly for elementary students, the potential negative emotional and academic ramifications must be carefully considered. As Hyman and Snook (2000) have argued, the increase and automatic use of punishment as opposed to prevention of misbehavior and violence in schools, often makes the schoolhouse toxic for children. In other words, zero tolerance discipline policies that allow for the discretionary removal of students have been used as a "gate keeping" mechanism in determining which students will have access to public schools. 10 Davis and Jordan (1994) determined that suspension and expulsion the two primary types of punishments associated with a zero-tolerance policy, negatively impact minority student academic achievement at a greater rate than EuroAmericans. Hyman & Snook, 2000 had similar findings that out-of-school suspension is linked to several negative outcomes including academic failure, grade retention, and negative school attitudes. The association of negative academic achievement and discipline practices should come as no surprise in light of the documented inequities and associated negative schooling experiences for minority students. The application of a zero tolerance discipline policy and the discretionary removal of students from the traditional learning environment also raise equity concerns in its application, particularly for those with the least amount of power within a school system. The research shows those most negatively impacted by the application of a zero tolerance discipline policy are minority and poor students. Researchers identified a correlation between a teacher's prejudice and negative expectations for minorities and more frequent and severe application of punishment for school related behavior (Townsend, 2000). The findings of a correlation 11 between a teacher's prejudice and expectations should not be surprising if one considers how issues of racial ideology and power continue to matter in America. In other words, a zero tolerance approach to student behavior allows those in power to demand social conformity from minorities, specifically from African American and Hispanic students. While there is a need to address student discipline, the degree of discretion and the embracement of a "zero tolerance" approach towards student behavior and the demand for the discretionary removal of disruptive students tend to address the concerns of those least impacted by this lack of tolerance or the associated consequence. Rather, zero-tolerance discipline policies appear to segregate and relegate students to a specific academic track. This differentiation in academics may result in a negative initial school experience and future expectation of failure and non-conformity. As Michael Casserly (1996) points out, "schoolchildren are tracked, sorted, labeled, and pigeonholed. Some are chronically detained, expelled, suspended, or removed. Either they are "pushed out" or they are graduated knowing very little. Either way, they have failed and have been failed". Educational experiences often serve as antecedents to many of the social and 12 economic ills students face later in life. Hence, inequities in schooling experiences underpinned by the application of discretionary discipline removal policies may have potentially broad consequences for students' future educational attainment, employment, and family relations (Davis, 1994). Purpose The purpose of this study is to determine the effect of teacher discretionary removal on urban minority students. This research will examine the impact for African American and Hispanic elementary students when a state expands its zero tolerance discipline policy to allow for the discretionary removal of a student from the traditional learning environment. In other words, this study will specifically examine the implications for minority students when teachers are given the discretionary authority to determine who will have access to their classrooms. For example, in Texas, legislation was adopted to expand the states zero tolerance discipline policy to give teachers the discretionary authority to remove students who were documented as being disruptive from their classrooms. 13 Chapter 37 (section 37.002 (b) of the Texas Education Code allows a teacher to remove a student from a classroom when the student has been determined to meet any of the following conditions. (1) Has been documented by the teacher to repeatedly interfere with the teacher's ability to communicate effectively with the students in the class or with the ability of the student's classmates to learn; or (2) Whose behavior the teacher determines is so unruly, disruptive, or abusive that it seriously interferes with the teacher's ability to communicate effectively with the students in the class or with the ability of the student's classmates to learn; or If a teacher removes a student from class under Subsection (b), the principal may place the student into another appropriate classroom, into in-school suspension, or into an alternative education program. The principal may not return the student to the teachers' class without the teacher's consent unless it is determined by a committee that such placement is the best or only alternative available. 14 If the removal of a student is classified as discretionary, the principal or an assistant principal makes the decision regarding the specific punishment to be imposed. Dependent upon the perceived severity of the offense and the district's and campus established student code of behavioral conduct, the punishment can range from timeout within a specific location at the school to a suspension. However, unlike a mandatory removal for serious offenses such as drugs and weapons, per Chapter 37 of the Texas Education Code, the suspension cannot exceed three consecutive days for one offense. Research Questions This study will seek to determine the extent, if any, the effects of teacher discretionary removal on urban minority students by answering the following questions: 1. What are the trends in student suspension rate subsequent to the 1996 adoption of a discretionary discipline removal policy, and the general sociodemographic characteristics of students who are subjected to a discretionary removal? 15 2. Are the achievement test scores (TAAS/TLI reading and math scores), of student's subjected to a discretionary removal significantly different from the test scores of comparable peer students not subjected to a discretionary removal? 3. Is there a relationship between discretionary referral reason and ethnicity? Definition of Terms To provide clarity and consistency in interpretation of major terms utilized throughout this report, the usage of these terms should be construed to mean the following: Discretionary Removals & Suspension (Chapter 37, Section 37.005, Texas Education Agency)- State law does not require the discretionary removals of students. However, because of violations of discipline management policies, a school can choose to enforce a zero tolerance policy and suspend a student. A discretionary removal and suspension cannot be made for longer than 3 days on the basis of multiple offenses that occur during the same time period. For example, a student may commit the following acts of assault and disrespect, while the student 16 has potentially violated two provisions of a student code of conduct, since the offense occurred on the same day a suspension should not exceed three days. Disciplinary Alternative Education Program/ Facility- A separate school / designated campus for students who have been suspended or expelled from school for a disciplinary reasons. For purposes of this study, the state of Texas does not require alternative discipline facilities to continue the educational program / curriculum for students who are suspended as the result of a discretionary removal. Minority Student- For purposes of this study, a minority student refers to individuals whose ethnic membership is African American and/or Hispanic. Socioeconomic Status Indicator (SES)- This is an index that represents family financial resources based on whether a student participates in the free or reduced cost of lunch program. Access to free or reduced lunch is defined by the family applying for that benefit and qualifying for it based on the total family income not exceeding a defined income level. Texas Assessment of Academic Skills (TAAS)- In the state of Texas, the Texas Assessment of Academic Skills (TAAS) is the primary measurement device used in the educational accountability system. The TAAS is a criterion-referenced test. A 17 Texas Learning Index (TLI) scaled score on the TAAS is a standardized score based on a scale centered at the level of minimum expectations (70) and a standard deviation of 15 scale score points. The TAAS is administered at grades 3, 4, 5, 6, 7, 8, and 10th in Texas. Texas Learning Index (TLI)- A score that allows for comparison both across years and across grades within subject areas. A TLI score of 70 corresponds to the minimum expectations/passing level and remains constant across administrations of different forms of the test. Special Education- students tested and identified as having some type of disability (physical or emotional) that may impede their learning. Any applicable designators (labels), for example, (LD) learning disability will be specifically identified as the research data are presented and discussed. Zero-Tolerance Discipline Policy- Provides a known consequence if a given act is committed, for example, a mandatory expulsion for bringing a weapon or drugs to school. 18 Importance of the Study There is no disagreement that violent and disruptive behavior cannot be tolerated in schools. The benefit of this study is to illustrate that a one-size fits all approach to student behavior does not appear to be a rational one to address the complex issue of school safety. What may have started, as a narrowly tailored policy to address specific behaviors unfortunately appears to have become the norm rather the exception for educators to handle student classroom management issues? In other words, the application of a zero tolerance discipline policy to address minor as well as major school safety concerns may promote and support inequity in the delivery of education services. School experiences have potentially broad consequences for students future educational attainment, employment, and family relations. Access to academic experiences through the curriculum, teachers, and other school activities is of particular importance for minority students who are often marginalized in school settings (Finn & Cox, 1992; Irvine, 1990; Sanders & Reed, 1995). As pointed out in the literature, little is know about salient school contextual and structural 19 variables and how they affect the educational outcomes for minority students (Davis & Jordan, 1994). As documented in the research, negative school experiences and outcomes for students are often products of school contextual and structural factors that limit learning opportunities, especially for minority males (Ferguson, 1991; Polite, 1993b). School environments often help students develop a sense of order, place, and expectations (Adler, Kless, & Adler, 1992). If the purpose of schools is to promote intellectual development and socialization, (Davis & Jordan, 1994) then arguably schools are falling short of their goals for students who are subjected to discretionary removals. The documented research that minorities, especially African American males have traditionally been subjected to disparate treatment in the application of discipline policies has primarily examined the impact on middle and high school students. However, zero-tolerance discipline discretionary policies tend not to discriminate based on age or maturity of children. Hence, this research is significant in examining the impact for elementary minority students, especially if one were to consider that due to their age and maturity, a discretionary removal 20 could prove even more detrimental to their perception of school and academic learning. Discriminatory treatment of minority students in school discipline is not an isolated phenomena, but rather part of a complex of inequity that appears to be associated with both special education overrepresentation and school dropout (Gordon, Della, Piana, & Keleher, 2000). The source of institutional inequity in the public education system is typically unconscious (Skiba, Michael, Nardo, & Peterson, 2000). Therefore studies such as this are important to bring to the attention of policy makers and educators that the academic success of minority students may be impeded by several factors to include those purported as benign school safety policies. As pointed out by Bowditch (1993), whether or not discrepancies in school discipline are in fact racially motivated does not eliminate the concern surrounding the overrepresentation of minorities and those of lower socioeconomic status in school discipline referrals; as well as the concern that discipline removals contribute to racial stratification in school and society. Racial bias in the practice of school discipline is only part of a broader discourse concerning the presence of institutional racism (Hanssen, 1998) or 21 structural inequity (Nieto, 2000) in education. This inequity could also be attributed to what Bowles & Gintis, 1976 and others have defined as cultural reproduction. The theory of cultural reproduction argues that schools serve as institutional mechanisms for the transmission and perpetuation of differential social class values. Hence, for elementary students whose behavior does not appear to conform, they are at increased risk of being subjected to a discretionary removal. In summary, this study is important because it will seek to illustrate how the application of discretionary removals may have unintended negative consequences for minority students. This study will further illustrate that for minority students it is not one variable but several intertwined variables that may impede their academic and social success. And more importantly, this study will seek to illustrate the "problem" may not be inherent in the student, rather a consequence of the education system. 22 CHAPTER 2 Review of the Literature The purpose of this study is to investigate the impact on minority students in elementary grades (k-5) when educators have the discretionary authority to broadly enforce a zero tolerance discipline policy to control student behavior. Student misbehavior is not new for classroom teachers; however, what is new is the degree of tolerance and responsibility teachers are willing to accept for the management of student behavior within their classroom, and the age at which students are being suspended. This review of the literature illuminates research that analyzes the effect on African American and Hispanic children's academic and social development when a zero-tolerance discipline policy is used to address student behavior. As the debate surrounding safe schools and classroom discipline continue, states such as Texas, Arkansas, Delaware, Hawaii, Louisiana, Mississippi, and Oregon established programs that segregate students viewed as discipline problems. These states have embraced the argument that society's degree of 23 acceptance for student misbehavior should be "zero". Therefore, faced with perceived disruptive and aggressive behavior, a typical response has been the punishment and exclusion of students exhibiting challenging behavior (Skiba & Peterson, 1999). Even in light of recent data illustrating a decrease in school violence, fears of school crime continues to permeate society and drive legislative policy (Portner, 2000). As a result, many legislative bodies and school officials endorse punishment over prevention and intervention, and have neglected to consider the potential negative consequence for children. The application of a zero-tolerance discipline policy adopted by policy makers and adopted by schools as a method of cultural reproduction as illustrated in the research literature has a negative impact on academic achievement due to the removal from the academic environment. The literature review will also illustrate the message a zero tolerance discipline sends to minority students who have a higher probability of being negatively affected by such policies. A review of the literature also illustrates the emotional and psychological implications for students and their perception of school and learning. Arguably, a child who does not feel 24 good about "self", "school" or "learning" will undoubtedly have a higher probability of not successfully moving through the educational system. Hence, while the policy is purported as being designed to serve the needs of society, as illustrated in the literature, policies designed to control behavior more likely than not negatively impact those with the least amount of power, the poor and ethnic minorities. Zero-Tolerance Discipline Policies & Academic Achievement Schools have increasingly adopted law enforcement models rather than educational models to address concerns surrounding school violence. This overdependence on a law enforcement model has often resulted in a variety of unintended negative consequences that make the schoolhouse toxic for too many children (Hyman and Snook, 2000). For example, several researchers have found a negative relationship between suspension rate, academic achievement growth, dropout rate and social achievement (Hyman and Snook, 2000). Davis & Jordan, 1994; Foster, 1993 & Hamilton, 1997 found a correlation between rates of suspensions and expulsions and increased school failure. A study conducted by 25 Costenbader and Markson in 1994 reported that of all the students who dropped out of school, between 51% and 55% had been suspended more than once. Costenbader and Markson (1998) also found that disruptive behavior within the classroom setting is predictive of less academic engagement time, which contributes to lower grades and poor performance on standardized test. In a recent study conducted by Rodney, L., Crafter, et.al. (1999), using a sample of 243 African American males between the age of 13-17 living in a midwestern city, found the number of suspensions from school to be the strongest predictor associated with grade retention. Rodney, et al. used a multiple-regression analysis to test 22 school-related variables, including alcohol usage, on school retention. While there was no significant finding between alcohol usage and grade retention, of the 22 school-related variables investigated three were positively associated with grade retention, the number of suspensions, violence against others, and the lack of discipline at home, with the strongest predictor for grade retention to be associated with suspension recidivism. In direct relationship to Rodney's research, Ledingham & Schwartzman, in a 1984 study concluded that elementary age children who demonstrate aggressive 26 behaviors are not only at risk being suspended, but are also at risk for grade retention and special education placement. Several researchers have also found specifically for African American males who are often over-identified as being aggressive, there is a positive correlation between suspension, grade retention, special education referrals and dropout rates. Nationally, African American boys score lower than any other group on standardized tests and are three times more likely in comparison to white students to be placed in special education classes, particularly at the elementary level (Jackson, 1975; Ohio Office of Black Affairs, 1990), the application of a zerotolerance discipline policy has a very significant consequence. While the predominant subjects of school discipline research has been with African Americans, this does diminish the implications for Hispanic students who make up the largest percentage of students who dropout of high school. As reported in the February 2002 United States General Accounting Office (GAO) on school dropouts, Hispanic students dropout of school at higher rates than other ethnic groups. In addition, Hispanics born outside the country are nearly three times as likely to drop out as those born in the United States. The research has shown that 27 multiple factors such as low grades, absenteeism, disciplinary problems, frequent school changes and grade retention are associated with students dropping out of school (GAO report, 2002). In terms of a negative impact of the punishments such as out of school suspension, which is often associated with a zero tolerance discipline policy, there is significant research illustrating a negative correlation between suspension, and grade retention. Royer, (1995) found that both suspensions and expulsions have short and long-term negative effects on social development, academic achievement and high school completion. Ekstrom, Goertz, Pollack, and Rock, (1986) determined in the High School and Beyond study, over 30% of sophomores who dropped out of school had been suspended, a rate three times that of peers who stayed in school (GAO report, 2002). However, there is debate surrounding the direction of the relationship between suspension and grade retention. For example, there are those who present the argument the strongest predictor of grade retention is a student's academic ability rather than the number of behavior referrals. In response to this debate regarding the direction of the relationship between academic ability, suspension 28 and grade retention, Jimerson, Carlson, Rotert, Egeland, and Soufe (1997) performed a group comparison study of (low-achieving promoted) to a similar retained group, they found the groups did not differ significantly on measures of intellectual functioning at age 64 months and at the end of third grade. Nikalson (1984) and Sandoval (1984) had similar findings that retained students did not exhibit lower cognitive ability in comparison with promoted students. Nikalson and Sandoval also determined retained students did not differ from a group of lowachieving but promoted peers. Therefore, these findings suggest that student discipline may be more influential than intellectual ability in the determination of a child's academic track and grade retention. Oakes (1994) research findings concluded that student's who are repeatedly suspended have a higher probability of being referred to a special education programs, or lower-academic tracks. Another possible explanation for the academic gap for minority and non-minority students may also be attributed the number of discipline referrals and subsequent suspensions, and the insufficient continuation of the academic requirements at alternative discipline education facilities. For example, to date only 26 states require alternative education 29 discipline facilities to provide for any type of academic instruction. However, while many of these facilities may offer some type of academics many do not offer a continuation of the established curriculum, or have certified teachers who provide academic instruction (Harvard Civil Rights Project, 2000). Therefore, while we would like to believe that all children have a choice and all parents have a voice regarding the academic track desired for their child, the application of a zero-tolerance discipline policy that allows for discretionary removals, places educators in the primary position of power to determine early in a child's school experience, his or her academic track. Zero-Tolerance Discipline & The Psychological Implications Suspension is a very severe form of punishment in that once students are served notice that their behavior is inappropriate and have no place in school, many students cannot view sanctions against their behavior as being separate from sanctions against their individual self. When educators suspend students, specifically elementary students, the research has shown that out of school suspension programs can lead to emotional and psychological trauma and recurring behavioral problems that can invite a cycle of behavior and expectation that is 30 difficult for both the student and the school system (Nicols, Ludwin, & Iadicola, 1999). For example, in the elementary school years students who have poor social skills may alienate teachers and peers. In turn, this alienation may result in the child being labeled as a "problem child" and subsequently resulting in his or her removal from the classroom or ignored, and as a result the child typically has below average academic skills. Hence, educator's low-expectations first begin to materialize at the elementary grades. In other words, what started as a desire to correct an undesirable behavior and the negative consequence becomes the expectation for both the teacher and student. Therefore, by middle school, these youngsters tend to become less interested in school and begin to seek out antisocial peers (Skiba & Peterson, 2000). Hence, the research illustrates that the involuntary removal of a student from the academic learning environment to be ineffective in changing disruptive behavior (Skiba and Peterson, 1999) and perhaps is the most powerful message of rejection contributing greatly to student disengagement (Felice, 1981; Wheelock & Dorman, 1988). 31 The application of a zero tolerance approach and the discretionary removal of a student regardless of age and maturity may also result in unintended outcomes. For example, Costenbader and Markson (1997) found that students who are suspended were either angry with the person who recommended the suspension or happy to get out of the situation. Costenbader and Markson also found that 32% of the students reported the suspension had not helped at all and they would probably be suspended again. Therefore, the message of "zero-tolerance" especially for African American males who are suspended at rates two and three times their percentage in the population (Townsend, 2000) may result in African American males believing they are not capable of abiding by school's social and behavioral codes. In other words, a self-fulfilling prophecy may develop as a result of these negative messages (Rosenthal & Jacobson, 1968). One must also consider, while suspension is considered a behavioral management technique to reduce aversive behavior by withdrawing reinforcing stimuli for a specified period of time this assumes that the environment one is removed from was in fact stimulating. As the literature shows, for many minorities student's schools are not always environments in which they feel wanted or 32 welcomed. Hence, suspension may function for children either as negative reinforcement of maladaptive behavior, or an escape mechanism from a classroom and/or school (Skiba & Peterson, 2000). Sullivan (1989) notes that punishment without meeting student needs for academic tutoring and other behavioral change seldom provides motivation for reform. Punishments in particular are temporary and transitory. Once the punishment is over, the student has "served his/her time" and is " free and clear" from further responsibility (Marshall, 1998), which may also help explain the suspension recidivism in spite of the intent of the policy to serve as a deterrent. Zero-Tolerance Discipline & Disproportional Application Typically, schools have two primary functions: (1) promoting and structuring the intellectual development of students; and (2) socializing young people for their roles and responsibilities in society. However, racial bias in schools is part of a broader discourse concerning the continuing presence of institutional racism in education (Hanssen, 1998). Institutions of education have been identified as systems that serve as institutional mechanisms for the transmission and perpetuation of differential social class values and serves to 33 perpetuate inequity within the education system (Hanssen, 1998). In schools there is a demand for the erasure of difference in student behavior, for students not aware of these expectations or unable to model the expected behavior, may often find themselves subjected to disciplinary actions. Hence, discipline policies can be used as a method used by teachers to maintenance their power and perpetuate cultural reproduction (Banister & Maher, 1998). Therefore, students are not only tracked based on academics but also for perceived social conformity to the rules established by those in power. One explanation for this dissonance and teacher expectation of minority student behavior may be attributed to a cultural mismatch. As described by several researchers, there is often a cultural mismatch between teacher expectations and minority student behavior and in response to this disconnect between teachers increased discipline actions (Foster, 1990; 1993; Ladson-Billngs, 1995; Villegas, 1991). Townsend (2000) suggests that many teachers, especially non-minority teachers may be unfamiliar and even uncomfortable with the different communication styles associated with African American children, specifically African American males. Unfamiliar listeners may interpret the documented 34 emotive manner associated with African Americans as combative or argumentative which may also explain the disproportional number of discipline referrals. There is also research that suggests that the relationship between suspension and the number of students that dropout of school may not be accidental. In several ethnographic studies school disciplinarians reported that suspension was sometimes used as a tool to encourage "troublemakers" or those perceived as unlikely to succeed in school to leave (Bowditch, 1993). Wehlage and Rutter (1986) found that in a national longitudinal study, 44% of the African American dropouts, 31 % of Hispanic dropouts, and 26% of white dropouts had been suspended or put on probation at least once before dropping out of school. Hence, for those students who do not master the overt and covert rules and standards defined by those in power, the schoolhouse may truly be a toxic environment. Socioeconomic Status & Discipline One troubling characteristics of a zero tolerance discipline policy is that a disproportionate number of those at-risk for a range of school punishments are poor and African American (Ward, 1998). Even when all socio-economic indicators are held constant, research has shown that African-American children are still 35 suspended and expelled at a much higher rate than their white peers. Brantlinger (1991) reported that disciplinary sanctions at the secondary level did in fact unfairly target low-income students as reported by both high and low income students. Wu, Pink, Crain, and Moles (1982) reported that while student behavior and attitude were correlated with suspension, school characteristics such as school governance, teacher attitude towards students, and student ethnicity was a greater determinant in how discipline was enforced and who was suspended. While there appears to be some debate regarding the degree of significance socioeconomic status contributes to ones probability of being suspended, the research is consistent that minorities, specifically African American males have a higher probability of being subjected to discipline policies than white students. While the annual suspension rate for all students nearly doubled between 1974 and 1998 from 3.7 percent to 6.9 percent, African Americans and Hispanics continued to be suspended at higher rates than whites. The largest disparity existed in the suspension rates for African Americans who comprised approximately 17 percent of all students in 1998-99, but accounted for 33 percent of all students 36 suspended (Harvard Civil Rights Project Report, 2000). The inequity in the suspension of minority students has been a consistent finding (Costenbader & Markson, 1994; Kaeser, 1979; McCarthy & Hoge, 1987; McFadden, Marsh, Price, & Hwange, 1992; Skiba et al., 1997; Wu, Pink, Crain & Moles, 1982). As Bowidtch (1993) points out, whether or not the inequity in the application of disciplinary sanctions was conscious or unconscious, the overrepresentation of minorities and those of lower socioeconomic status who are suspended and/or expelled from schools serves to only contribute to the continued racial stratification in schools and in society. Gender & Discipline There also appears to be consistent evidence of overrepresentation of boys referred for disciplinary sanctions (Skiba, Michael, and Nardo, 2000). A number of studies have found that males are over four times as likely as girls to experience a disciplinary referral and punishment such as suspension (Bain & McPherson, 1990; Cooley, 1995; Gregory, 1996; Imich, 1994). In a study of 13,879 secondary (middle and high school) students, male students were 1.93 times more likely to 37 receive a discipline referral than female students (Nichols, Ludwin, and Iadicola, 1999). Ethnicity & Discipline The research also illustrates that ethnicity appears to make a contribution to disciplinary outcome independent of socioeconomic status or student behavior (Wu, et al., 1982). One possible explanation for this racial disparity may be attributed to the race relations in the United States. While one would like to believe teachers' leave their racial biases at the schoolhouse door, the reality is that individuals are influenced by their experiences and all have their own view of the world. For example, Ahlquist, 1991; Haberman, 1992; King, 1991; Sullivan, 1989 & Tatum, 1994 determined, racist attitudes in the United States are just as common among teachers as within the general population. Foster (1993) found in a survey of both educators and non-educators that the majority in both groups held stereotypical attitudes and prejudices toward AfricanAmerican males in the categories of athletics, crime, education, music, attitude, family, and personality. 38 Skiba, et al. (2000) research determined, teachers who are prone to accepting stereotypes of African American males as threatening or dangerous may overreact to relatively minor threats to authority. Several researchers also found higher rates of punishment for minority students to correspond to a teacher's prejudice beliefs and negative expectations (Foster, 1993; Fremon & Hamilton, 1997). Davis and Jordan (1994) had similar findings when using national data, the investigation of various school contextual and structural factors affecting African American male success in middle and high school and during the transition between these two academic levels. Davis and Jordan observed three significant results: (1) African American males in urban schools have lower achievement than those elsewhere; (2) an emphasis on discipline at school is associated with lower achievement; and (3) teachers who assign more work seem also to issue higher course grades. The aforementioned findings suggest that students who have increased discipline referrals / action are typically placed on a lower academic track and/or special education programs, therefore, one would not expect these students to receive the benefit of more work. One must also be concerned the abundant 39 influence of cultural stereotypes of race, which are pervasive and well learned (Devine, 1989). Consensual knowledge about cultural racial stereotypes has been demonstrated to be similar across levels of racial prejudice. Research also shows that knowledge of stereotypes can and often does have automatic influences on perceptions, judgments, and even behaviors (Devine, 1989). A variety of measures of the racial stereotype / racial attitude are connected to ones memories or cognitive structures, and tend to surface with or without awareness and this knowledge tends to affect ones basic social judgments (Dovidio et al., 1997; Fazio, Jackson, Dunton, & Williams, 1995; Maurer, Park, & Rothbart, 1995). The net effect of these research findings is that race is a salient basis of categorization that leads to exaggerated perceptions of differences and is reflected in biases that favor the racial category of one's own group. These effects can and often do occur without awareness; therefore one is often convinced that race had no effect on one's judgments. Hence, one could argue that educators are guilty of "objective fallacy" in the application of discipline policies (Darley and Gross, 1983). 40 Referral Reasons In spite of the research findings of a high correlation between repeated suspensions and subsequent grade retention (Safer, 1986) and dropout rates (Diem, 1988; Quinn, 1991), schools continue to suspend student's even for non-violent offenses, in the name of zero-tolerance for student misbehavior. For example, documented reasons students are suspended include: disrespect, noncompliance, defiance, and general school disruption (Imich, 1994; McFadden, Marsh, Price, & Hwang, 1992). In contrast, Imich (1994) found that about half of all suspensions were prompted by verbal or physical aggression against peers. A review of the literature also found a lack of consensus of the definition of aggressive behavior and other offenses. Skiba, Peterson & Williams (1997) study of reasons for student referrals in two mid-western cities across multiple schools found behaviors that led to office referral were primarily for student noncompliance and disrespect. In addition, they found little evidence of a consistent relationship between seriousness of the offense and severity of the consequence. Costenbader & Markson, 1994 determined that discipline actions were often tied to a student's prior disciplinary history, academic 41 record, and degree of remorse for the unacceptable behavior. However, even in light of these variables, ethnicity appeared to be the most significant determinant if and what type of punishment would be imposed. For example, African-American students appear to be referred for disciplinary sanctions for less severe behavior and receive harsher punishments (McFadden et al., 1992; Shaw & Braden, 1990). Costenbader and Markson (1994) investigated 349 schools representing 55% rural, 20% urban, and 24% suburban areas in 10 states in which they found that African American students were being suspended disproportionately to their total enrollment, and the most common infraction resulting in suspension was physical aggression. One explanation for the disparity in the application of discipline behavior patterns presented by (Gregory, 1997) was that African American boys have assimilated assertive behavior patterns that are adaptive outside of school but are maladaptive in the classroom. The difficulty with Gregory's assumption is that Hispanic students do not fare much better in the successful navigation of the U.S. education system. For example, the dropout rate for Hispanics is significantly greater than that of African Americans (Snyder & Hoffman, 1994). 42 Summary Many zero tolerance policies, especially a policy that allows for the discretionary removal and suspension of a student are often unfair and arbitrary and therefore may have no place in the education of children. The application of this type of policy that focuses on a punitive approach to modify behavior as illustrated in the research literature is often fraught with negative consequences for children. While it is true teachers must maintain classroom order, the question of how much "tolerance" should students and parents, expect teachers and school staff to have for children? 43 CHAPTER 3 Research Design & Methodology The purpose of this study was to investigate the impact of zero tolerance discipline policy that allows for the discretionary removal of students for behavior deemed unacceptable by a classroom teacher or administrator. Specifically, this study will examine the implications of this policy for minority elementary students in grades k-5th within one large urban school district in the State of Texas. In the State of Texas teachers have the discretionary authority to remove and suspend a student for the following reasons: (1) Students who has been documented as repeatedly interfering with the teacher's ability to communicate effectively with the students in the class or with the ability of the student's classmates to learn; or (2) Whose behavior the teacher determines is so unruly, disruptive, or abusive that it seriously interferes with the teacher's ability to 44 communicate effectively with the students in the class or with the ability of the student's classmates to learn. The primary methodology used for this study is a quasi-experimental approach to examine the effect of discretionary removal on academic achievement. The reason this methodology was used is that it allows the effect of specific variables such as discretionary removal rate on student academic achievement to be measured. This methodology allows comparison of reading and math scores of students subjected to a discretionary removal to their peers not subjected to a discretionary removal. A quasi-experimental design was used to statistically control for third-variable explanations by holding different variables constant. Quasi-experimental designs unlike experimental designs do not include random assignment to conditions (Cook & Campbell, 1979). Although quasi-experimental designs do not involve a rigorous test of cause- effect hypotheses, as do experimental designs, they do seek to answer the question whether an independent variable is an "indicator" of whatever the real cause may be (Dane, 1990). In other 45 words, the use of a quasi-inferential statistics would identify the strength and significance of relationships between variables and academic achievement. An analysis of the discretionary removals for students in grades K-5th during the1996-1997, 1997-1998, and 1999-2000 academic school years was performed. The rationale for an examination of the trends in discipline removals was due to the expectation that one year following the 1995 adoption of the discretionary removal policy one might expect a higher rate of referrals. However, one would expect if the policy were effective in preventing disruptive student behavior there would be a decline in the number of referrals in later years. In addition to the use of inferential and descriptive statistics to examine the implication of discretionary discipline policy, the research findings will be shared with a sample of ethnically diverse Principals. The Principals will be asked to respond to the findings and the implications of discretionary removal policy. Five Principals from ethnically diverse backgrounds with a minimum of five years experience as an Administrator in Texas were individually interviewed, two African Americans, two whites, and one Hispanic. This study relied heavily on a database maintained by the school districts 46 Alternative education facility for students suspended as a result of a discretionary removal. A limitation of the database is students subjected to a discretionary removal could select to have their short-term "removal" occur at home. Due to incomplete documentation of at-home removals only discretionary removals for student's to an established alternative education facility for elementary was considered for this study. The alternative education facility maintained complete records documenting student's demographic information, dates of referrals, reason for referral, and number of days associated with each removal. The data were compiled into one single database for purpose of analysis. Design A mixed (Between-Subject and Within-Subject) design is used to answer the research questions regarding the effect of teacher discretionary removal on urban minority students. To examine trends in discipline removals and the types of offenses committed by students in grades k-5, a within-subjects experimental design comparison of students subjected to a discretionary removal in 1996-1997, 1997-1998, and 1999-2000 by ethnicity, gender, socio-economic status, special education status, and grade level was examined. While the information gained 47 from this design methodology is useful the findings cannot be generalized beyond the sample group. Therefore, to answer the research question regarding the effect of discretionary removal on academics, a between- group comparison of students was performed. A between- group comparison is used as a means to determine if discretionary removal and not some other variable caused the observed effects on academic achievement. Inferential statistics such as a T-test and a Multiple Regression model was used to test the between-group comparison of academic achievement. Since the Texas Assessment of Academic Skills (TAAS) test is not administered to students in grades k-2, the effect of suspension on academic achievement growth for this group of students will not be addressed. An Independent T-test was used to determine if there was a significant difference in the group mean academic test scores of students subjected to a discretionary removal and their peers not subjected to a removal. However, during this phase of testing the effect of discretionary removal on academics, potential interceding variables were not subjected to control. Hence, a multivariate analysis was employed to enhance the understanding of the relationships between discretionary removals and academic achievement while incorporating and 48 statistically controlling for student demographics, socioeconomic status, and special education status. An Analysis of Variance and descriptive statistics were used to examine referral reasons by ethnicity. The Analysis of Variance statistic compared group mean differences between the number of referrals, and ethnicity. Participants Schools The district selected for this study is a large urban Texas school district with a student population that exceeds 75,000 students. The ethnic population is approximately 16 % African American, 2.5% Asian, 48% Hispanic, 0.3% Native American, and 34% Anglo. As with other large Texas urban school districts, this district has sufficient numbers of elementary students subjected to suspension to warrant the establishment and maintenance of a separate disciplinary facility for this population. The number of elementary schools that could remove students to an alternative disciplinary education facility was sixty-eight (68) between 1996-1998 and increased to seventy (71) with the opening of three new elementary schools between 1998 and 2000. 49 Subjects The data were aggregated so that students became the unit of analysis. The subjects used for analysis purposes were students in grades k-5 who were subjected to discretionary removal and suspension between the academic 1996/97 and 1999/00 school year. Due to the small number of Pre-kindergarten students suspended between 1996/97 and 1999/00 and the majority of elementary schools did not go beyond the 5th grade, these students will not be included in the analysis. The total average number of elementary students who could have been subjected to discretionary removal and suspension between 1996/1997 and 1999/2000 academic school years was 37,404. The total number of elementary students subjected to a discretionary removal and suspension between 1996-1997 and 1999/2000 was 5,414. To examine the impact of discretionary removal on a student's academic achievement scores (TAAS/TLI reading and math scores) for a cohort of students in the third grade in 1996-97 with three consecutive TLI readings and math scores 50 will be used as the comparison group N=3,735. The comparison group N=3,735, will be matched based on ethnicity, socio-economics, gender, and special education status to students subjected to a discretionary removal in 1996-97, 1997-98, and 1998-99 for purposes of comparing TLI reading and math differentiations over a three year period of time. Students subjected to a discretionary removal were sorted based on five different criteria in order to identify comparable students within the cohort group for matching purposes. For matching purposes students subjected to a discretionary removal between 1996-1999, N=112 (1996-97); N= 165 (1997-98); N= 187 (1998-99) were sorted based on: Not Economic Disadvantaged / Not Special Ed., Not Economic Disadvantaged / Special Ed., Economic Disadvantaged/ Not Special Ed., and Economic Disadvantaged / Special Ed. Data Collection The 1996-1997 to 1999-2000 school years discipline data used for this research was provided by an alternative learning facility located within the subject school district. Prior to 1998 schools provided the alternative learning center with handwritten referral information that identified students and demographic 51 information, the name of the referring school, the number of days for the suspension, not to exceed three days for discretionary removals, and a description of the offense(s) based on the district's disciplinary policy, as outlined in the disciplinary handbook. During this period there were 12 primary reasons annotated on the form submitted by campus personnel to the alternative discipline education facility. This information was entered into attendance notebooks maintained by the alternative discipline learning facility. Beginning in the 1997-1998 school year campuses were required to electronically enter discipline information/ codes to conform to the Texas Education Agency PEIMS reporting requirements, which expanded the category of offenses from 12 to 15. The following offense key was developed in compliance with the Texas Education Agency PEIMs reporting requirements. All Texas schools for discipline reporting purposes currently use these offense codes. These 1998-1999 offense codes attempt to be more descriptive and extend beyond behaviors more readily associated with elementary students. The data used for academic achievement implications were obtained from the University of Texas Education Productivity Council (EPC). The EPC has 52 access to the PEIMs accountability information reported by all Texas school districts to the Texas Education Agency. The 1996-1997 socio-economic student data were provided by the participating school district. OFFENSES KEY 1998-1999- TEA PEIMS CODING 01 Attendance- class cutting, tardy, leaving class without authorization, failure to return to class, leaving school without authorization 02 Tobacco- possession, use 03 Insubordination- failure to follow instructions, refusal to go to D-hall, refusal to go to In-school suspension, forgery of permit, Misrepresentation on campus w/o permission, violation of law, policy or rule 04 Disruption- misbehaving, talking in class, running, throwing objects, class disruption, disruptive behavior, disruption of educational process 05 Abusive Conduct (student) ­ rude/profane language / gestures, harassment/intimidation, extortion/blackmail, threat/incite threat, coercion, fighting, assault, aggravated assault 06 Abusive Conduct (adults)- rude/ profane language/gestures, harassment/intimidation, extortion/blackmail, threat/incite threat, coercion, fighting, assault, aggravated assault, retaliation Property- unauthorized entrance, robbery/theft/stealing, possession of stolen property, vandalism, arson, destruction, criminal mischief 08 Illegal Weapon- firearm, illegal knife, club, other 09 Alcohol- possession, consumption, influence, sale distribution 07 53 10 Dangerous drug/ controlled substance- possession, consumption, influence, sale distribution 11 Other- cheating, gambling, fireworks, unclassifiable, aggravated kidnapping, murder, terrorist threat, legal knife Sexual Behavior- indecent exposure, lewd behavior, sexual assault, aggravated sexual assault, indecency with a child Conduct punishable as a felony in the community Glue or Aerosol- consumption, influence, distribution Serious & persistent misbehavior after removal to Alternative Educational Placement 12 13 14 15 Since there is only a slight variation between the earlier offense code and the subsequent 1998-99 modification, the two discipline code tables were collapsed and student referral reasons fell within one of the following (12) categories. (Offense Key To Be Used for Analysis) ASP Assault against peer to include fighting ASS Assault against staff DR Drugs (added from 1998 offense category) 54 GM General Misbehavior- acting out in class, throwing things, running around; Temper Tantrum, disruptive GV LC PD RFR SX General Verbal- cussing, threats Left Class / Left Campus Property Destruction Refusal to follow Rules- defiant act towards teacher, ignoring teacher Sexual offense- inappropriate touching others, displaying body Parts ST TR W Stealing Truancy Weapon While the subsequent 1998 code of offenses was more detailed, a comparison of the discipline codes entered at the campus level often did not match the corresponding descriptive incident information. Hence, a comparison was made of the discipline code and descriptive information. The ultimate determinant in the referral reason code was incident description. In addition, while there are similarities between general misbehavior and refusal to follow rules, based on the narratives of the incidents provided by the referring teachers, refusal to follow rules 55 typically entailed offensive behavior directed specifically at an adult. In contrast, general misbehavior appeared to be directed to no specific person or person(s). After obtaining the applicable district clearance and removing individual identifying information, all other applicable discretionary removal / suspension information was entered into a single database for analysis purpose. Research Questions This study is designed to determine the extent, if any, the effects of teacher discretionary removal on urban minority students. The questions to be answered by the study and the corresponding statistical analysis to be used for this purpose are presented and described in this section. 1. What are the trends in student suspension rate subsequent to the 1996 adoption of a discretionary discipline removal policy and the general sociodemographic characteristics of students who are subjected to a discretionary removal? 2. Are the achievement test scores (TAAS/TLI reading and math scores), of student's subjected to a discretionary removal significantly different from 56 the test scores of comparable peer students not subjected to a discretionary removal? 3. Is there a relationship between discretionary referral reason and ethnicity? Procedures The T-test will be used to determine if there is a significant difference in the mean reading and math test scores of students subjected to a discretionary removal and a comparable cohort not subjected to a discretionary removal. (T-test) The t-test is a univariate procedure used to test two-group situations (e.g., minority students suspended compared to students not suspended and the impact on academic achievement growth). The t-statistic is the ratio of difference between the sample means (ľ1-ľ2) to their standard error. T statistic= ľ1- ľ2 SEľ1ľ2 ľ1= mean of group 1 ľ2 = mean of group 2 SEľ1ľ2 = standard error of the difference in group means 57 T6By forming the ratio of the actual difference between the means to the difference expected due to sampling error one can quantify the amount of the actual impact of the treatment (suspension) that is due to random sampling error. If the t value is sufficiently large, then statistically we can say that the difference was not due to sampling variability, but represents a true difference. Multiple Regression Analysis Multiple regression analysis is a form of general linear modeling. It is used to examine the relationship between a single dependent variable and a set of independent variables. In this study, multiple regression analysis will be used to determine if socioeconomic status, special education identification, ethnicity and referral rates impact academic achievement. The flexibility and adaptability of multiple regressions allows for its use with almost any dependence relationship (Hair, Anderson, Tatham and Black, 1998). Assumptions of the regression model (Hair, et. al., 1998) are: No relevant independent variables have been excluded. No irrelevant independent variables have been included. 58 - The variables are accurately measured. For each observation, the expected value of the error term is zero. The variance of the error term is constant for all values of the independent variable. - The error terms are uncorrelated. The independent variables are uncorrelated with the error term. The error term is normally distributed. The absence of perfect multicollinearity- i.e., none of the independent variables are perfectly correlated with another independent variable. In regression analysis, the estimates for the expected values will be "good" if the model is true. A model is true if the expected values can be expressed as a linear combination of observable values and unknown parameters in the way described by the model (Ward & Jennings, 1979). Hence, conclusions are limited to whether the sample data produce the expected values that are related the way the model describes the relationship. 59 To test whether suspension rate, socioeconomic status, and ethnicity impact math and reading Texas Learning Index math and reading scores, the following regression models will be performed: Model 1: Y (Math TLI in 1999) = constant +b1X1 (Math TLI in 1997) + b2X2 (Socioeconomic status) + b3X3 (Special Education) + b4X4 (African American) + b5X5 (Hispanic) +b6X6 (Other Ethnicity) + b7X7 (Removed-1) + b8X8 (Removed2) + b9X9 (Removed-3) + b10X10 (Removed-4) + E(error) b7=1 if the student was subjected to discretionary removal 1 time and 0 if not b8=1 if the student was subjected to discretionary removal 2 times and 0 if not b9=1 if the student was subjected to discretionary removal 3 times and 0 if not b10= 1 if the student was subjected to discretionary removal 4 or more times and 0 if not Model 2: Y (Reading TLI in 1999) = constant +b1X1 (Reading TLI in 1997) + b2X2 (Socioeconomic status) + b3X3 (Special Education) + b4X4 (African American) + b5X5 (Hispanic) +b6X6 (Other Ethnicity) + b7X7 (Removed-1) + b8X8 (Removed-2) + b9X9 (Removed-3) + b10X10 (Removed-4) + E(error) b7=1 if the student was subjected to discretionary removal 1 time and 0 if not 60 b8=1 if the student was subjected to discretionary removal 2 times and 0 if not b9=1 if the student was subjected to discretionary removal 3 times and 0 if not b10= 1 if the student was subjected to discretionary removal 4 or more times and 0 if not Limitations One significant limitation of this study is the inability to control for student "maturity" especially for the younger grades (k-2) when examining issues of disproportional numbers of suspensions (discipline removals) by student demographics. Another limitation of this study is prior to 1997 the tracking and maintenance of disciplinary referrals was performed by hand and all entries were not always legible or complete and therefore omitted. In addition, as previously mentioned, records of suspensions that occurred at home were not reliable and therefore not considered in this study. Hence, while the sample is large it is not reflective of all of the suspensions resulting from a discretionary removal between 1996-2000. Another limitation of this study is the modification in the behavior offense codes. Prior to 1997 the codes while broad in interpretation, provided fewer 61 categories of offenses that could be indicated. In contrast, after 1997 while the categories remained the same, codes of subcategories were made available for purposes of providing a more detailed description of the inappropriate behavior. For example, Insubordination (03) while identified as an offense prior to 1998, after 1998 the subcategories available to further describe the type of insubordination included: failure to follow instructions, refusal to go to D-hall, refusal to go to In-school suspension, forgery of permit, misrepresentation, on campus without permission, and violation of law, police or rule. To resolve the change in the offense key, only categories applicable to the subjected population was used for this study. 62 CHAPTER 4 Findings The results of descriptive and inferential statistics were used to determine the extent, if any, the effects of teacher discretionary removal on urban minority students. The findings include an examination of the data for trends in discretionary removals for the 1996-97, 1997-98, 1998-99, and 1999-2000 academic school years for elementary students in grades (K-5). Inferential statistics were used to determine if there was a relationship between discretionary removals on academics, but more specifically the amount of effect. The research questions and findings within this chapter were divided into three sections. Included in Section 1 are the results of the descriptive analysis of the trends in discretionary removals for students in grades K-5 from 1996 to 2000. Included in Section 2 are the results of inferential statistics used to examine group mean differences in math and reading scores of students subjected to a discretionary removal between 1996-97 and 1998-99 school year in comparison to a cohort of students not subjected to a discretionary removal. Included in Section 3 63 are descriptive and inferential statistic results of discipline removal reasons by ethnicity. Also included in this section were the perspectives of five school administrators from the subject school district regarding the research findings, and the use of a discretionary removal policy. Section 1 Question 1: What are the trends in the application of a zero-tolerance discipline policy for one urban Texas school district between 1996 and 2000 after the policy is expanded to allow for the discretionary removal of children deemed disruptive to the learning environment? Enrollment, Removal, Ethnicity & Gender On average (4%) of the total yearly average, N=37,404 students enrolled in grades K-5 during the 1996- 97, 1997-98, 1998-99, and 1999-2000 academic school years were subjected to a discipline removal. Breakdowns of the discretionary removals by ethnicity were detailed in Table 4.1, Figures IV-1, IV-2, IV-3 and IV-4. The findings indicated that while the enrollment rate for African American students remained relatively stable between 1996-97, 1997-98, 1998-99, 64 and declined slightly in 1999-2000, the number of discipline removals continued to increase yearly. This same pattern of disparity does not appear to hold true for white or Hispanic students. The average total enrollment for African American students between 1996/97 and 1999/2000 was 6,062. The average discipline removal rate was 598 (10%) of the average total African American student enrollment. The percentage of African American students (10%) subjected to a discretionary removal exceeded the 4% average discipline removal rate for the total average K-5 student enrollment between 1996/97 and 1999 2000. In comparison, for Hispanics the average fouryear academic enrollment was 16,890, of which 3% of the total student enrollment was subjected to a discretionary removal. For white students the average enrollment between 1996/97 and 1999/2000 was 13,377 of which, less than 5% of their total population was subjected to a discretionary removal. The findings as illustrated in Table 4.1 indicate the percentage of discretionary removals for African American students almost doubles between 1996-97 and 1999-2000 from 8.6% to 13.37%. In comparison, for Hispanic and white students the percentage of removals either decreased or remained relatively 65 stable. For Hispanic students the highest percentage (3.2%) N=16,401 occurred in 1997-98. However, during the highest enrollment N=17,841 in 1999-2000, the percentage of Hispanic students subjected to a discretionary removal decreased. In terms of discretionary removal and gender, on average, males were six times more likely to be subjected to a discretionary removal than females. Between 1996-97 and 1999- 2000, males represented 85% of the total average number of students subjected to a discretionary removal (Table 4.2). 66 Table 4.1 Enrollment by Ethnicity & Number of Discretionary Removals__ 1996-97 1997-98 1998-99 1999-00 Enrollment Removal Enrollment Removal Removal Enrollment Removal Enrollment Black 6342 543 (8.6%) 6301 564 (8.9%) 6417 590 (9.2%) 5191 694 (13.37%) Hispanic 15897 465 (2.9%) 16401 496 (3.2%) 17422 512 (2.9%) 17841 527 (2.9%) White 14060 264 (1.9%) 13590 242 (1.8%) 13425 238 (1.8%) 12436 221 (1.8%) Other 1127 7 (.62%) 994 9 (.90%) 1106 17 (1.5%) 1068 25 (2.3%) Total 37426 1279 37286 1311 38370 1357 36536 1467 Note: % = Percent of total ethnic subgroup enrollment 67 FIGURE IV- 1 DISCRETIONARY REMOVALS 1996-1997 600 500 400 300 200 Count 100 0 BLACK HISPANIC OTHER WHITE ETHNICITY 96-97 68 FIGURE IV- 2 DISCRETIONARY REMOVALS 1997-1998 600 500 400 300 200 Count 100 0 BLACK HISPANIC OTHER WHITE ETHNICITY 97-98 69 FIGURE IV- 3 DISCRETIONARY REMOVALS 1998-1999 700 600 500 400 300 200 Count 100 0 BLACK HISPANIC OTHER WHITE ETHNICITY 98-99 70 FIGURE IV- 4 DISCRETIONARY REMOVALS 1999-2000 800 600 400 200 Count 0 BLACK HISPANIC OTHER WHITE ETHNICITY 99-00 71 TABLE 4.2 Discretionary Removals by Gender Year Males Females Total 1996-97 1997-98 1998-99 1999-00 1048 (81.9%) 1088 (83%) 1214 (89.5%) 1260 (85.9%) 231(18.1%) 223 (17.0%) 143 (10.5%) 207 (14.1%) 1279 1311 1357 1467 72 Socio-economic Status In 1996, the largest percentage of the students subjected to a discretionary removal was not economically disadvantaged. However, beginning in 1997-98 there was a significant increase in the percentage of students from low socioeconomic status subjected to a discretionary removal. The percentage of low socioeconomic status students subjected to a discretionary removal jumped from 36.2% to over 90% (Table 4.3). The increase in students identified as being of low socioeconmomic status is most notable for Hispanic and African American students. The findings as illustrated in Figures IV-5, IV-6, IV-7, and IV-8 also indicated between 1996-97 and 1999-2000, African Americans represented the largest percentage of low socio-economic students. A plausible explanation for the significant increase after 1996-97 in the number of students of low socio-economic status subjected to a discretionary removal could be attributed to data entry and/or reporting error. However, another contributing factor to this disparity in socio-economic status over time could possibly be attributed to an exodus of middle class families to surrounding suburban communities and schools. As determined by Orfield (2000) enrollment in 73 American schools is undergoing a major transformation resulting in the segregation of schools by both race and poverty. To date over 50% of the total student population within this school district is identified as economically disadvantaged. Hence, it is not necessarily surprising that a high percentage of removed students were identified as economically disadvantaged. The trend from 1997-2000 in the number of economically disadvantaged students subjected to a discretionary removal continued to increase among all ethnic groups. Table 4.3 Discretionary Removal by Socio-economic Status Year 1996-97 1997-98 1998-99 1999-00 Disadvantaged 463 (36.2 %) 1184 (90.3 %) 1277 (94.1) 1377 (93.9%) 90 Not Disadvantaged 816 (63.8%) 127 (9.7%) 80 (5.9) (6.1%) Total 1279 1311 1357 1467 74 FIGURE IV-5 Socio-economic Status by Ethnicity 700 600 500 400 300 ETHNICITY 96-97 BLACK HISPANIC 200 Count 100 OTHER WHITE DISAD NOT DISAD 0 ECON 96-97 75 Figure IV-6 Socio-economic Status by Ethnicity 700 600 500 400 300 ETHNICITY 97-98 BLACK HISPANIC 200 Count 100 OTHER WHITE ECON DISAD NOT DISAD 0 ECON98 76 FIGURE IV-7 Socio-economic Status by Ethnicity 700 600 500 400 300 ETHNICITY 98-99 BLACK HISPANIC 200 Count 100 OTHER WHITE ECON DISAD NOT ECON DISAD 0 ECON 98-99 77 Figure IV-8 Socio-economic Status by Ethnicity 700 600 500 400 300 ETHNICITY 99-00 BLACK HISPANIC 200 Count 100 OTHER WHITE ECON DISAD NOT DISAD 0 ECON00 78 Special Education According to Figures IV-9, IV-10, IV-11, and IV-12, between 1996-2000 the largest number of students subjected to a discretionary removal did not receive special education services. However, the largest percentage of students subjected to a disciplinary removal that were identified as receiving special education services were primarily African American or Hispanic. The special education findings were also interesting in that on average African Americans represented the largest number of students identified as having some type of special education disability. Based on the available research findings students identified as special education are often on a lower academic track; therefore, it could be anticipated this variable would have a significant negative effect on this subgroup manifesting high academic achievement. Hence, if the current research bears truth, it would be anticipated that statistical tests would reveal that Special Education students would have lower TLI reading and math scores in comparison to students not identified as Special Education. 79 Figure IV-9 Special Education Status by Ethnicity 700 600 500 400 300 ETHNICITY 96-97 BLACK HISPANIC 200 Count 100 OTHER WHITE NOT SPED SPED 0 SPED 96-97 80 Figure IV-10 Special Education Status by Ethnicity 700 600 500 400 300 ETHNICITY 97-98 BLACK HISPANIC 200 Count 100 OTHER WHITE NOT SPED SPEC ED 0 SPED 97-98 81 Figure IV-11 Special Education Status by Ethnicity 700 600 500 400 300 ETHNICITY 98-99 BLACK HISPANIC 200 Count 100 OTHER WHITE NOT SPED SPEC ED 0 SPECIAL ED 98-99 82 Figure IV-12 Special Education Status by Ethnicity 700 600 500 400 300 ETHNICITY 99-00 BLACK HISPANIC 200 Count 100 OTHER WHITE NOT SPED SPED 0 SPED 99-00 83 Grade Level A student's grade, age or maturity did not tend to exempt him/her from being subjected to a discretionary removal. As shown in Figures IV-13 through IV-16 students as young as five years of age were subjected to discretionary removals. One year after the adoption of a discretionary removal policy, African Americans represented the largest number of students across all grade levels to be subjected to a discretionary removal, followed by Hispanics. In addition, for African Americans the number of removals continued to increase over the fouryear period, especially in the 4th and 5th grades. For example, between 1998-99 and 1999-2000 the number of removals for African American 4th graders increased by 63%, from 110 to 176, and for 5th graders by 78%, from 135 to 173. For Hispanic students, unlike African American students, there was a significant increase in the number of removals beginning in the 5th grade. This finding would suggest that African American students committed more serious offenses at an earlier age to warrant both the number of removals at such an early age and the consistency 84 across time in the removal rate. This finding will be further examined using inferential statistics. Figure IV-13 Discretionary Removal by Grade Level And Ethnicity 700 600 500 400 300 ETHNICITY 96-97 BLACK HISPANIC 200 Count 100 OTHER WHITE 0 1 2 3 4 5 0 GRADE 96-97 85 Figure IV-14 Discretionary Removal by Grade Level And Ethnicity 700 600 500 400 300 ETHNICITY 97-98 BLACK HISPANIC 200 Count 100 OTHER WHITE 0 1 2 3 4 5 0 GRADE 97-98 86 Figure IV-15 Discretionary Removal by Grade Level And Ethnicity 700 600 500 400 300 ETHNICITY 98-99 BLACK HISPANIC 200 Count 100 OTHER WHITE 0 1 2 3 4 5 0 GRADE 98-99 87 Figure IV-16 Discretionary Removal by Grade Level And Ethnicity 700 600 500 400 300 ETHNICITY 99-00 BLACK HISPANIC 200 Count 100 OTHER WHITE 0 1 2 3 4 5 0 GRADE00 88 Section 2 Question 2: Were the achievement test scores TLI reading and math of students subjected to a discretionary removal significantly different from the test scores of comparable peer students not subjected to discretionary removal? A cohort was selected from N= 3,735 third grade students in 1996. All students in the cohort had three consecutive years (1997, 1998 and 1999) TLI reading and math scores. Students selected for the cohort were identified for purposes of matching to a select group of students subjected to a discretionary removal between 1996 and 1999. The student matching criteria used were: grade, ethnicity, gender, socioeconomic status, and special education identification. However, prior to examining the effect on academic achievement for a comparable, i.e., matched students, a T-test was performed to determine if there were mean score differences in math and reading scores between students subjected and those not subjected to a discretionary removal. The findings illustrated in Figures IV-17-19 show the mean math and reading 1997, 1998, and 1999 TLI scores for students subjected to a discretionary removal on average was 20 or more points lower in comparison to students not 89 subjected to a discretionary removal. However, these findings do not account for demographic variables such as special education status, ethnicity, etc. or differences in socio-economic status. Hence, in an attempt to better understand the implications of the differences in reading and math TLI scores, a T-test was performed using different variables compared the mean score differences between students subjected to a discretionary and their peer comparison group not subjected to a discretionary removal. 90 Figure IV-17 Referral Status and Academic Achievement 80 70 69 70 60 50 44 Mean 40 43 LINMTH97 LINRED97 30 No Referral Referral REFER97R 91 Figure IV-18 Referral Status and Academic Achievement 80 77 70 71 60 57 50 52 Mean LINMTH98 LINRED98 No Referral Referral 40 REFER98R 92 Figure IV-19 Referral Status and Academic Achievement 80 77 75 70 60 54 50 Mean 49 LINMTH99 LINRED99 No Referral Referral 40 REFER99R 93 Illustrated in Tables 4.4, 4.5, and 4.6 are t-tests for Equality of Means for student math and reading scores of students subjected and those not subjected to a discretionary removal. The findings indicated a significant relationship between referral status and academic achievement in 1996-97, t =3.26, p= .001 for math; and for reading, t = 3.21, p=. 002. Application of Levene's Test for Equality indicated 47% of math and 35% of the reading score variation was explained by an Independent t-test. In 1997-98, t = 1.99, p =. 047 for math; and for reading, t =2.071, p=. 039. Application of Levene's Test for Equality indicated 26% of the math variation and 28% of the reading variation was explained by the Independent t-test. In 1998-99, t = .784, p=. 433, indicated no significant relationship between math TLI scores for students subjected to a discretionary removal and those not subjected to a removal. However, for reading there was a significant relationship, t =1.95, p =. 052. The test of 1998-99 math and reading TLI score comparison accounted for 86% of the variation in math scores and 62% of the variation in reading scores. 94 Table 4.4 1996-1997- Math & Reading Independent Sample t-Test Source Math Reading t 3.261 3.212 df 220 220 Sig. (2-tailed) .001 .002 Mean Difference 13.81 14.25 Note: 95 % Confidence Interval of the difference; p < .05 Table 4.5 1997-1998- Math & Reading Independent Sample t-Test Source Math Reading t 1.993 2.071 df 327 327 Sig. (2-tailed) .047 .039 Mean Difference 6.21 6.44 Note: 95 % Confidence Interval of the difference; p <.05 Table 4.6 1998-1999- Math & Reading Independent Sample t- Test Source Math Reading t .784 1.95 df 372 372 Sig. (2-tailed) .422 .052 Mean Difference 2.68 7.15 Note: 95 % Confidence Interval of the difference; p < .05 95 T-Test Findings An Independent ­Samples t test was performed for each sample year 1997, 1998, and 1999 to further examine group mean differences in math and reading for students subjected and those not subjected to a discretionary removal when controlling for ethnicity, socio-economic, and special education status. The largest achievement gap in math and reading scores between students subjected to a discretionary removal and those not subjected to a removal occurred in 1996-97 (Tables 4.7 and 4.8). Students subjected to a discretionary removal in 1996-97 scored approximately 7 points lower in math and 14 points lower in reading in comparison to students not subjected to a removal. There were also findings in 1997 and 1999 of a gap in reading and math achievement for students subjected to a discretionary removal and students not subjected to a removal. In 1997-98, on average for every gain in math and reading made by students not subjected to a discretionary removal, students subjected to a discretionary removal scored 6 points lower. In 1998-99 while students not subjected to a removal continued to score 96 lower in reading and math than students not subjected to a removal, the greatest gap 8 points was in the area of reading. Table 4.7 Effect of Discretionary Removal on Mean Math Scores Year 1996-97 NoRemoval N= 110 (58.13) Removal N= 110 (51.92) Mean Difference 7.21 1997-98 N= 165 (58.13) N= 165 (51.92) 6.21 1998-99 N= 187 (56.66) N=187 (53.97) 2.69 Table 4.8 Effect of Discretionary Removal on Mean Reading Scores Year 1996-97 No Removal N= 110 (57.03) Removal N= 110 (42.78) Mean Difference 14.25 1997-98 N= 164 (63.63) N= 187 (55.88) N= 165 (57.19) N=187 (47.83)) 6.44 1998-99 8.05 97 Achievement & Ethnicity The data contained within Table 4.9, display the 1997, 1998, and 1999 differences in mean math scores by ethnicity for students subjected to a discretionary removal and students not subjected to a removal. In 1997, white students subjected to a discretionary removal scored on average 30 points lower in comparison to white students not subjected to a discretionary removal. However, this gap in math achievement for white students subjected to a discretionary removal and white students not subjected to a discretionary removal began to close over the next three years. For African Americans, the mean difference in math scores for students subjected to a discretionary removal and African Americans not subjected to a discretionary removal did not fluctuate significantly over the three-year period, ranging from 11 to a 12 point score differentiation. For Hispanic students the largest gap in math achievement between students subjected to a removal and Hispanic students not subjected to a removal occurred in 1996. In 1997 and 1999 while the 98 number of Hispanic students subjected to a removal increased the difference in mean math scores to Hispanics not subjected to a removal was relatively small. A comparison of the mean reading scores for African Americans, Hispanic and white students subjected to a discretionary removal to their peers within their ethnic subgroup not subjected to a removal is illustrated in Table 4.9. For African Americans, students subjected to a discretionary removal had TLI reading scores that ranged 10-15 points lower in comparison to African Americans not subjected to a discretionary removal. For Hispanic students, the greatest mean difference 15 points occurred in 1996, a year after the adoption of a discretionary removal policy. However, the gap in reading scores for Hispanic students subjected to a discretionary removal and those not subjected to a removal decreased in 1998 and 1999. White students experienced a 20-point differentiation in reading mean scores for students subjected and not subjected to a discretionary removal, with this gap decreasing in later years. 99 Table 4.9 Effect of Discretionary Removal by Ethnicity on Mean Math Achievement African American Year 1996-97 Hispanic Difference 10.32 White Difference 29.65 No Removal Removal N= 54 (51.81) N= 70 (59.57) N= 64 (52.30) N= 55 (41.49) N= 71 (47.30) N=64 (64.15) No Removal Removal Difference No Removal Removal N=34 (55.65) N=53 (54.60) N=76 (56.88) N=34 (45.75) N=53 (57.08) N=76 (56.80) 9.90 N=21 (75.29) N=39 (61.10) N=46 (63.41) N=22 (45.64) N=39 (54.05) N=46 (64.15) 12.27 2.48 6.95 1997-98 11.85 .08 .74 1998-99 100 Achievement & Socio-economics Status The findings in Tables 4.10 and 4.11 support the research literature of a relationship between a student's socioeconomic status and academic achievement. As indicated in Table 4.11, between 1996 and 1999 economically disadvantaged students not subjected to a discipline removal in comparison to non-disadvantaged scored on average 22 points, 19 points, and 16 points respectively, lower in math in comparison to non-economic disadvantaged students. A similar gap was also prevalent in the area of reading achievement. The findings also illustrated in 1996 economically disadvantaged students subjected to a removal on average scored 14 points lower in math and reading in comparison to economically disadvantaged students not subjected to a removal (Tables 4.10 and 4.11). A gap in math and reading for economically disadvantaged students subjected to a discipline removal and students not subjected to a removal was also prevalent in 1997-98 and 1998-99. The implications of a discipline removal for non-disadvantaged students subjected to a removal are most notable in mean reading scores. As 101 indicated in Table 4.11, in 1996-97 non-disadvantaged students subjected to a discipline removal on average scored almost 16 points lower in reading in comparison to non-disadvantaged students not subjected to a removal. This subpopulation also experienced almost a 12-point gap in reading in 199798, with the gap significantly decreasing to 5 points in 1998-99 102 Table 4.10 Effect of Discretionary Removal & Socioeconomic Status on Mean Math Scores Disadvantaged Year 1996-97 Not Disadvantaged Difference No Removal Removal Difference 13.54 N=15 (76.80) N=36 (72.89) N=42 (71.17) N=15 (63.40) N=36 (66.94) N=42 (68.14) 13.40 No Removal Removal N= 95 (54.76) N=128 (53.98) N=145 (52.46) N= 95 (41.22 ) N=128 (47.71) N=145 (49.87) 6.27 5.95 1997-98 2.59 3.03 1998-99 Table 4.11 Effect of Discretionary Removal & Socioeconomic Status on Mean Reading Scores Disadvantaged Year 1996-97 Not Disadvantaged No Removal N= 95 (54.47) N=128 (59.07) N=145 (50.50) Removal Difference No Removal Removal Difference N= 95 (40.97 ) N=128 (54.10) N=145 (42.66) 13.50 N=15 (73.20) N=36 (79.83) N=42 (74.45) N=15 (57.40) N=36 (68.69) N=42 (69.71) 15.80 1997-98 4.97 11.14 1998-99 7.84 4.74 103 Achievement & Special Education From 1996-1999 special education students on average scored 40 or more points lower in math than students not identified as special education. As can be seen in Table 4.12, when controlling for discretionary removal and comparing math scores of special education and non-special education, the greatest gap in math scores occurred in 1996-97. In 1996-97, non-special education students subjected to a discretionary removal scored 15 points lower than their peers. Gaps in math scores were prevalent across time both within and between groups. Indicated in Table 4.13 were the effects of a discretionary removal on the reading scores for non-special education and special education students. The results show in 1996-97, non-special education students subjected to a removal on average scored 15 points lower in reading than their peers not subjected to a removal. However, beginning in 1997 the gap in reading achievement between non-special education students subjected to a discretionary removal and their peers not subjected to a removal decreased to 6-points and subsequently to a 3-point difference. Special education students subjected to a discretionary removal in comparison to their peers not subjected to a discretionary removal experienced the 104 largest gap (10-points) in reading achievement in 1996-97. However, the gap in reading for special education students in comparison to their peers, decreased to 7 points beginning in 1997-98, and to 4- points in 1998-99 (Table 4.13) 105 Table 4.12 Effect of Discretionary Removal & Special Education Status on Mean Math Score Not Special Education Year 1996-97 Special Education No Removal Removal Difference No Removal Removal Difference N= 70 (73.19) N= 98 (71.54) N= 97 (76.76) N= 70 (57.99) N= 99 (65.67) N=98 (74.31) 15.2 N=40 (30.78) N=66 (38.21) N=90 (34.99) N=42 (20.57) N=66 (31.29) N=89 (31.58) 10.21 1997-98 5.87 6.92 1998-99 2.45 3.41 Table 4.13 Effect of Discretionary Removal & Special Education Status on Mean Reading Score Not Special Education Year 1996-97 Special Education No Removal Removal Difference No Removal Removal Difference N= 70 (77.16) N= 98 (78.28) N= 97 (79.10) N= 70 (58.90) N= 99 (70.67) N=98 (71.09) 18.26 N=40 (21.80) N=66 (41.88) N=90 (30.86) N=42 (15.90) N=66 (36.97) N=89 (24.11) 5.90 1997-98 7.61 4.91 1998-99 8.01 6.75 106 Regression Model The Independent T-test provided information regarding mean differences. A multiple-regression model was used to help elaborate group differences. A multiple regression analysis was conducted to predict the overall significance of socioeconomics status, ethnicity, special education classification, and discipline referral rate on math and reading academic achievement growth. The following models were used: Model 1: Y (Math TLI in 1999) = constant +b1X1 (Math TLI in 1997) + b2X2 (Socioeconomic status) + b3X3 (Special Education) + b4X4 (African American) + b5X5 (Hispanic) +b6X6 (Other Ethnicity) + b7X7 (Removed-1) + b8X8 (Removed-2) + b9X9 (Removed-3) + b10X10 (Removed-4) + E(error) b7=1 if the student was subjected to discretionary removal 1 time and 0 if not b8=1 if the student was subjected to discretionary removal 2 times and 0 if not b9=1 if the student was subjected to discretionary removal 3 times and 0 if not b10= 1 if the student was subjected to discretionary removal 4 or more times and 0 if not 107 Model 2: Y (Reading TLI in 1999) = constant +b1X1 (Reading TLI in 1997) + b2X2 (Socioeconomic status) + b3X3 (Special Education) + b4X4 (African American) + b5X5 (Hispanic) +b6X6 (Other Ethnicity) + b7X7 (Removed1) + b8X8 (Removed-2) + b9X9 (Removed-3) + b10X10 (Removed-4) + E(error) b7=1 if the student was subjected to discretionary removal 1 time and 0 if not b8=1 if the student was subjected to discretionary removal 2 times and 0 if not b9=1 if the student was subjected to discretionary removal 3 times and 0 if not b10= 1 if the student was subjected to discretionary removal 4 or more times and 0 if not Test of the significance of the relationship between the following multiple correlation coefficients, number of referrals, ethnicity, special education status, and 1997 math TLI score was not only statistically significant but also rather large in magnitude thereby indicating a relationship between the combination of all predictors and the criterion, 1999 math TLI scores. The obtained R _ indicates 57% of the variance in math TLI achievement was shared by this combination of variables. These results suggest the individual differences among math TLI scores were either caused by the combination of the predictors themselves, or by other 108 factors linked to these predictors. Each coefficient can be interpreted as the amount of change that is expected to occur in the criterion (math TLI score) per unit change in that predictor when statistical control has occurred for all other variables. In other words, students from low socio-economic backgrounds math scores will decrease approximately 4 points in comparison to students not from a low socioeconomic background. The findings also suggest when controlling for special education status there is an 18-point difference in the math scores of special education and non-special education students. In addition, as indicated in the regression models the most significant predicator of low academic achievement was a special education placement (Tables 4.14 and 4.15). An interesting finding was the effect of discipline removal rate on academic achievement. While there was a negative relationship between discipline removal rate and math achievement, the relationship was not statistically significant for the 1st or 3rd removal. In other words, while students subjected to one discipline removal may have scored almost 3 points lower statistically one could not attribute the low score to being subjected to one discipline removal. However, there was a statistically significant relationship between a student's low math score and being 109 subjected to a second removal, and for students subjected to four or more removals. There was no known explanation for the lack of statistical significance for a third removal and math achievement, this finding warrants additional research. The results of a multiple regression model (Table 4.14) illustrated African Americans on average scored almost 4 points lower in math in comparison to mathematical gains made by whites. Hispanic's in comparison to whites, scored on average, 2 points lower in math. 110 Table 4.14 Results: Predicting 1999 Math Scores Predictors Constant (97 Math TLI) MTH97 SES97 Sped97 African American Hispanic Referral ­1 Referral-2 Referral-3 Referral ­4 Unstandardized t Coefficients 46.438 38.320 .487 -3.57 -18.24 -3.912 -1.23 -2.27 -10.62 -3.67 -11.65 Significance .000 35.6820 .000 38.32 -19.02 -4.35 -1.72 -1.72 -4.38 -1.02 -5.17 .000 .000 .000 .086 .087 .000 .308 .000 Summary statistics: R = .756, R_ =. 572. The findings illustrated in Table 4.15 indicated the predictor variables, number of referrals, ethnicity, special education status, and 1997 reading TLI scores were not only statistically significant but also rather large in magnitude; 111 thereby indicating a large relationship between the combination of all predictors and the criterion, 1999 reading TLI scores. The obtained R _ indicates 63% of the variance in reading achievement was attributed to the combination of predictor variables: ethnicity, socio-economic status, special education identification, and discipline removal rate. These results suggest a good deal of the individual differences among reading TLI scores were caused by either by the combination of the predictors themselves or by other factors that were causally linked to these predictors. Analysis of TLI reading scores over time revealed a statistical significant relationship between all of the predictor variables and reading scores. The most significant coefficient negatively affecting reading scores was the rate of discretionary removal. The findings indicate on average students subjected to a removal experienced a 5-point decrease in reading scores in comparison to students not subjected to a removal. Provided in Table 4.15 is a detailed description of the negative differences in reading scores attributed to the different predictor variables. Results of the regression model (Table 4.15) predicting 1999 reading achievement 112 indicated that special education students would score approximately 18 points lower in reading for every gain made by non-special education students. Table 4.15 RESULTS: PREDICTING 1999 READING SCORES Predictors Constant- 97 Reading TLI RED97 SES97 Sped97 African American Hispanic Referral ­1 Referral-2 Referral-3 Referral ­4 Unstandardized t Significance Coefficients 45.916 36.038 .000 .544 -5.69 -17.82 -4.55 -3.24 -4.60 -8.04 -13.06 -14.59 38.992 .000 -7.67 -16.54 -4.70 -4.10 -3.22 -3.08 -3.36 -5.99 .000 .000 .000 .086 .001 .002 .001 .000 Note: Summary statistics R= .789, R_ =.623 113 Section 3 Question 3: Is there a relationship between discretionary referral reason and ethnicity? An Analysis of Variance (ANOVA) comparing group mean differences indicated a relationship between the number of referrals and ethnicity. The ANOVA was significant for all four academic years; however, the effect size of the relationship was weak, F (3) =3.329, p=.019; F (3) = 7.964, p=.000; F (3) =2.736, p=.042, and F (3) = 3.222, p = .022. (Tables 4.16, 4.17, 4.18, and 4.19). Student offenses and discretionary removal reasons committed between 1996-97 and 1999-2000 are illustrated in Figures IV-20 through IV-23. The two most prevalent reasons for the discretionary removal of students in grades K-5 were assault against peers and general misbehavior. A review of the discipline data indicated that assault against peers at the higher grades, third through fifth often involved fighting. However, in grades k-2nd, the definition of assault varied considerably but always involved some type of touching of another student. As shown in Figure IV-24, African Americans on average were subjected to the largest 114 number of removals due to general misbehavior. The discipline removal category of general misbehavior is one of several subjective offenses used by teachers to justify the removal of a student. Since, there was no definition either in policy or legal statue identifying specific behavior that would constitute general misbehavior, a few examples of the types of behaviors identified by teachers as constituting general misbehavior of African American, Hispanic and white students are outlined below: Teacher Description of Discipline Referral Reasons K K 1st 1st 1st 1st 1st 2nd 2nd 2nd 2nd 2nd 3rd 4th 4th 4th 4th 5th 5th Afr. American Afr. American Hispanic Hispanic Afr. American Anglo Anglo Anglo Hispanic Hispanic Afr. American Afr. American Anglo Hispanic Afr. American Afr. American Hispanic Hispanic Anglo Noncompliance refusal to talk, fighting, disruptive disruption and noncompliance noncompliance, disruption kicking, screaming, tearing up things, tantrum on way to ISS disruptive in ISS disrespectful, not following directions He slammed a door so hard that a clock fell off the wall, very disruptive etc. very disruptive disruptive noncompliance, wandering halls, insubordination disruption, hitting throwing things etc.... noncompliance disruption kicked his chair, disruptive, broke contract raging, tantrum, disrespectful misbehavior in lunchroom, throwing food disruptive arguing, disrespectful, disrupting class, throwing stuff 115 As illustrated in the example of student actions falling within the category of general misbehavior, many of the descriptions used to describe the student behavior are vague. For example adjectives such as "disruptive", "disrespectful", and "noncompliance", without any further explanation are sufficient language and reason to deny a student continuous access to the learning environment. 116 Table 4.16 1996-97- One-Way ANOVA (#Referrals by Ethnicity) Components Sum of Squares Between 55.396 groups (combined) Within 7071.189 Groups Total 70126.585 Degrees of Freedom 3 Mean Squares 18.465 F-Ratio 3.329 Significance .019 1275 1278 5.546 .019 Note: the effect size index= .008 Table 4.17 1997-98 One-Way ANOVA (#Referrals by Ethnicity) Components Sum of Squares Between 97.276 groups (combined) Within 5321.494 Groups Total 5418.770 Degrees of Freedom 3 Mean Squares 32.425 F-Ratio 7.964 Significance .000 1307 1310 4.072 Note: the effect size index= .018 117 Table 4.18 1998-99 One-Way ANOVA (#Referrals by Ethnicity) Components Sum of Squares Between 24.631 groups (combined) Within 4059.932 Groups Total 4084.563 Degrees of Freedom 3 Mean Squares 8.210 F-Ratio 2.736 Significance .042 1353 1356 3.001 Note: the effect size index= .006 Table 4.19 1999-00 One-Way ANOVA (#Referrals by Ethnicity) Components Sum of Squares Between 24.049 groups (combined) Within 3640.148 Groups Total 3664.196 Degrees of Freedom 3 Mean Squares 8.016 F-Ratio 3.222 Significance .022 1463 1466 2.488 Note: the effect size index= .007 118 Figure IV-20 OFFENS97_ 600 500 400 300 200 Frequency 100 0 ASP ASS GM GV LC PD RFR ST SX TR W OFFENS97_ Figure IV-21 OFFENSE 97-98 40 30 20 Percent 10 0 ASP ASS DR GM GV LC PD RFR ST SX TR W OFFENSE 97-98 119 Figure IV-22 OFFENSE99 500 400 300 200 Frequency 100 0 ASP ASS DR GM GV LC PD RFR ST SX TR W OFFENSE99 Figure IV-23 OFFEN00 600 500 400 300 200 Frequency 100 0 ASP ASS GM GV LC PD RFR ST SX W OFFEN00 120 Figure IV-24 Offense Type by Ethnicity 300 200 ETHNICITY 96-97 100 BLACK HISPANIC Count OTHER 0 ASP ASS GM GV LC PD RFR ST SX TR W WHITE OFFENS97_ Figure IV-25 Offense by Ethnicity 300 200 ETHNICITY 97-98 100 BLACK HISPANIC Count OTHER 0 ASP ASS DR GM GV LC PD RFR ST SX TR W WHITE OFFENSE 97-98 121 Figure IV-26 Offense by Ethnicity 300 200 ETHNICITY 98-99 100 BLACK HISPANIC Count OTHER 0 ASP ASS DR GM GV LC PD RFR ST SX TR W WHITE OFFENSE99 Figure IV- 27 Offense by Ethnicity 300 200 ETHNICITY 99-00 100 BLACK HISPANIC Count OTHER 0 ASP ASS GM GV LC PD RFR ST SX W WHITE OFFEN00 122 Administrator Response to Zero-Tolerance and Discretionary Removal The research findings were shared individually with five Principals, two African Americans, one Hispanics, and two whites; then each Principal was asked to respond to the following questions: 1. What are your perceptions and recommendations regarding the discretionary removal of students? "Children are not suffering academically because of home life or community", it is what schools are doing to children". Not surprised the largest percentages of removals are African Americans, specifically the removal of African American males. A lot has to do with classroom instructional practice. In early grades, K-2 interactive / responsive learning is encouraged, however in grades 3-5 interactive / responsive learning is discouraged. For children who are not able to make this shift in pedagogy (i.e., become less active), these children are at risk of being removed from the classroom. 123 "Zero-tolerance discipline policies are important", can't allow children to be disruptive to the learning process, however, "we must seek to understand the behavior of the child receiving the consequence", "need to investigate the reasons for the behavior" " Study aligned with national research" " Discretion should be with the Administrator, who must understand the consequence of the decisions (punishment)". "Zero-tolerance leaves a lot of latitude," "a persons racial identify and view of the world impacts how teachers view children". "When a child is not in school, (classroom) they are not learning" "No surprise that African Americans represent the largest number of removals", a lot has to do with "teacher tolerance". "Regardless of where the removal occurs, at 124 home, alternative setting, in the Principal's office", the child is still missing out on learning. 2. What is your observation regarding the finding that the largest number of students in special education classes (the most prevalent label, emotionally disturbed) are African Americans? " Surprised that elementary students being identified as emotionally disturbed" " Not surprised", "there is no good sense of what constitutes emotionally disturbed". An example was provided where the Principal asked behavior diagnosticians to explain how an emotionally disturbed special education determination was made for a particular group of students. The response from the diagnostician was, "they all come from the same neighborhood", i.e., (foster home). One Principal provided this example to illustrate how those who make decisions regarding student's, often operate from their own ontology which often is from a White / middle class frame of reference. This white/ middle class frame of 125 reference is often used to define what is "normal". In lieu of marveling at a child's resilience in spite of life's obstacles, it is automatically assumed the child has a deficit and therefore must be emotionally disturbed. As pointed out by the Principal, it is okay for a teacher to have external difficulties, " okay for a teacher to fall apart, that is normal", but not normal for a child to be upset by circumstances beyond their control, and acceptable for "teachers have little tolerance" when children display emotional difficulties. "Teachers are not familiar with cultural differences and instead of trying to understand a child's life, we attempt to explain (different) behavior by labeling" " Special Education is no longer for academics, current special education model tries to change social behavior". Students who are not able to shift from an interactive to a less interactive learning model are subjected to removal from the classroom. "These children are then pulled out of the classroom for the purpose of receiving help with social skills". 126 3. Do you think teachers abuse the discretionary removal policy? "Maybe", "but it is the Administrators responsibility to make certain that it is addressed by showing teachers the data". "Yes", "that is why it is important to disaggregate the data and discuss with teachers" "No", "it is often a situation where "everyone needs to calm down", and teachers must be given opportunity to discuss with the Principal their concerns. "Teacher and Principal must have a shared voice" " Policy abuse is dependent upon the Administrator", Administrators must provide clear expectations regarding teacher responsibility and classroom management "Policy not abused", teachers do not necessarily understand cultural differences, "misunderstandings" 127 "Yes, policy is being abused and used to get rid of children, and we house them in special education classes" 4. What is your opinion / observation regarding the finding of a significant increase in discipline removals for Hispanic students occur around the 4th and 5th grades? " Family expectations change specifically for male children around the 4th and 5th grade. Around this age, "Hispanic males are expected to be independent". This shift in roles may conflict with the role the child is expected to play at school. "Hispanic children experimenting, social expectations change, puberty, and peer pressure" "Cultural differences, male and female roles become more defined even to a point of how a male child is permitted to speak to his mother". This difference in (male to female) communication in the school environment may be viewed as insubordinate by teachers. "Hispanic male defending his manhood" 128 " Mom / Dad working student unsupervised", "role of child / degree of responsibility increases and the role may conflict with teacher view of how a child should act". "Family culture", 4th / 5th grade males expected to have more leadership. The way a male child communicates with his mother changes. " There is a shift in roles for Hispanic males, and the student may not be aware of how they sound to a teacher". "Hispanic males may have difficulty shift from one role at home to a different role at school" " Cultural difference", machismo, Hispanic males more independent from their mother around this time, and the child's home social learning may be in conflict with school. "The male child is moving into the role of a man, and does not know how to shift / change role back to child in school". 5. What type of help/ assistance should be provided teachers? "Track discipline referral reasons, discuss as a team what strategies could be tried at the classroom level". "Work with teachers in understanding poverty, the roles 129 and responsibilities children of poverty often must assume at an early age, teach children how to communicate, i.e., shift roles within the school environment." " Have teachers examine their biases and generate communication". Require teachers to clearly articulate how they reached a decision that a child warrants a discipline referral. Principal must investigate; ask what interventions the teacher used before invoking a discipline action. "Give teachers a shared voice" "Require teachers to handle discipline issues at the classroom level, provide them with a plan that outlines levels of offenses with specific examples and interventions". Provide teachers with clear classroom management expectations and clear definition of the types of behavior that warrant a referral to the Principal. "Share discipline referral data with teachers, start a conversation, share internal and external classroom behavior management best practices". Provide teachers with different behavior management techniques. 130 "Help teachers learn different methodology (relevant pedagogy) that addresses the needs of children from different cultures". "Provide teachers with support to learn new integrated strategies, cultural differences and balanced curriculum". 131 Common Themes While the views and experiences of the five Principals regarding zerotolerance discipline policy and discretionary removals varied, two primary themes emerged: 1. A need to have discipline data that can be analyzed and shared with staff. 2. Communication, shared learning and understanding of cultural differences. All Principals articulated that Hispanics had a unique culture that may be in conflict with the expectations of teachers. 132 CHAPTER 5 Conclusions Introduction The aim of this study was to determine the effect of discretionary removal on urban minority students. The primary motivation for this study was to determine whether and how the expansion of the zero discipline policy to allow for discretionary removal contributed to the documented academic achievement gap between minority and non-minority students. The use of a mixed subject design methodology strengthened the value and validity of the findings by comparing the effect of discretionary discipline removal within and between the three different student ethnic populations. The results of this study clearly indicated minority students were subjected to discretionary removals at higher rates than non-minorities. The findings illustrated numerous patterns in discretionary removals based on student ethnicity, gender, special education identification, and socio-economic status. This study also determined a negative relationship between discretionary removal and academic 133 achievement; on average, students subjected to a discipline removal had lower achievement test scores in comparison to peers not subjected to a discipline removal. In terms of a relationship between discretionary referral reasons and ethnicity, the findings indicated a relationship between ethnicity and the number of discipline referrals. On average, minority students were subjected to more discipline removals than non-minority students. Descriptive analysis of the discretionary discipline removal reasons established a pronounced relationship between ethnicity and referral reasons. Patterns of Discretionary Removals Ethnicity & Gender There were several notable patterns in discretionary discipline removals between 1996-97 and 1999-2000 for students in grades K-5th. For example, the number of African American students subjected to discretionary removals increased each year. In addition, when African American student enrollment declined by 24% between 1998-99 and 1999-2000, the discretionary removal rate 134 jumped from 9.2% to 13.37%. In contrast, the number of Hispanics and whites subjected to a discretionary removal remained stable. Between 1996-97 and 19992000, the percentage of discipline removals for Hispanics and whites was less than 5% of their total group enrollment. Another discretionary discipline removal pattern in the study was that a large majority (80%) of the students subjected to discipline actions were males. These findings of males being subjected to disproportionate discipline actions are aligned with current research findings, for example, Skiba, Micheal, Nardo (2000) and others have also documented males are subjected to more discipline actions in comparison to females. Socioeconomic Status Several researchers have determined disproportional numbers of children of low-socioeconomic backgrounds are subjected to discipline actions. This study had similar findings of the over-representation of poor children subjected to discipline actions. Beginning in 1997, over 90% of the students in this study subjected to a discretionary removal were from a low socio-economic background, of which 135 African American students represented the largest number (N=542), followed by Hispanic students (N= 477). Special Education Identification In regards to the number of special education students subjected to a discretionary removal, unlike other research findings, the largest numbers of students subjected to a discretionary removal in this study were not special education students. The grade of the subjects used for this study may have attributed to this deviation from existing research. Typically, children are not tested for special education services until the 3rd grade, and this study included students in grades K-2. Between 1996-97 and 1999-2000, a total of (N=1,499) special education students were subjected to a discretionary removal, of which African Americans represented 43%, Hispanics 34%, and whites 22%. This finding is consistent with the research of Ledingham & Schwartzman (1984) and Jackson (1975), and others who determined a relationship between discipline referrals of minority students and special education placement. 136 Grade Level In terms of patterns in the removal by grade level, the number of African American and Hispanic students in grade K-5 subjected to a discretionary removal increased every year between 1996-97 and 1999-2000. However, for Hispanic and white students, unlike African American students, the most notable increased rates of removal occurred in the 5th grade. One significant concern in the pattern of removals by grade level was the increased percentage of removals between kindergarten and second grade. On average between 1996-97 and 1999-2000, discretionary discipline removals increased over 45% between kindergarten and first grade, and 80% between first and second grade. These findings are of particular concern due to the age and maturation of the students and the potential psychological harm of a discretionary removal. As noted in the research literature, when educators suspend elementary students, many cannot view sanctions against their behavior, as separate from sanctions against their individual self. As a result, this type of action may invoke emotional and psychological trauma (Nicols, Ludwin, & Iadicola, 1999). 137 Academic Achievement & Discretionary Removals Student academic and social failure is an evolving process; children do not enter the schoolhouse doors with aspirations of academic failure, a desire to be labeled "disruptive", or to be ejected from classrooms. However, on average, this is what happened to 10% of the African American and 3% of the Hispanic students enrolled within the subject Texas urban school district. Also illustrated in the research, the subjective use of teacher discretionary discipline removal of students from the academic learning environment is not without negative consequence, especially for African American and Hispanic students. In addition, for students with pre-existing risk factors associated with low academic achievement, such as poverty and special education placement, discipline removals exacerbated low academic achievement. For example, in 1999, special education student's math scores on average would be 18 points lower in comparison to non-special education students, and when students are subjected to one discipline referral the math scores are further decreased by 3 points. 138 Consistent with the findings of Costenbader and Markson (1998), that disruptive student behavior is predictive of less academic engagement time and contributes to lower grades, and poorer performance on standardized test, the findings of this study indicated discretionary discipline removal had a negative effect on student future academic achievement. As demonstrated in the study, there was a causal relationship between prior student achievement scores (1997), discipline referral rates, and future (1999) academic student achievement. The research illustrated the1997 discipline removal rate had different negative effects on student's 1999 math and reading scores. The findings indicated that two or more discretionary removals increased the likelihood a student would have a math score 10 or more points lower than a peer not subjected to a discipline removal. In contrast, any number of discipline removals had a negative impact on a student's reading achievement. Students subjected to one discipline removal on average scored 5 points lower in reading than peers not subjected to a discipline removal. The gap in reading doubled with a second removal and increased with subsequent removals. 139 While the results of this study indicated one discipline removal might not negatively effect a student's future math achievement, the findings were clear that any number of removals had a negative effect on reading. This would imply there is a need for elementary students to have continuous access to reading instruction and curriculum. Discretionary Referral Reason & Ethnicity An analysis of variance determined between 1996-97 and 1999-2000 there was a measurable difference between ethnicity and the number of discipline referrals; however, the effect size of the relationship between ethnicity and discipline referral reasons was small. The number of similar offenses committed by African American and Hispanic students may have contributed to the small effect size. However, descriptive analysis of the trends in referral reasons by ethnicity provided more substantive findings. For example, on average African American students, followed by Hispanic students, committed the largest number of assaults against peers. This finding was consistent with research conducted by 140 Imich in 1994 which determined approximately half of all suspensions were prompted by verbal or physical aggression against peers. Another prevalent reason for the removal of African American and Hispanic students was for general misbehavior and disrespect. The definition of general misbehavior and disrespect varied significantly from teacher to teacher. Future research may examine how teachers have characterized and defined student misbehavior and disrespect, especially in light of the disproportionate number of minority students subjected to discretionary removals. As determined by Skiba (2000) and other researchers, teachers are prone to accept stereotypes of different racial and ethnic groups that may influence their application of student discipline. Researchers have also determined many teachers, especially non-minority teachers, may be unfamiliar with the different communication styles of minority children which may help explain the disproportional discipline actions taken against minority children. 141 Discretionary Removals and Other Factors While the purpose of the interviews with Principals was not to answer the research questions posed by the study, the information garnered from them provided additional explanations of the findings. One common theme that emerged from the Principal interviews was an explication of cultural differences between teachers and minority students. For example, Principals attributed increased discretionary removals of Hispanic fifth graders to teacher-student cultural clashes over male role identity and family expectations. In contrast, there was no consensus as to the possible reasons African American students and teachers may clash. Three out of the five Principals acknowledged negative teacher perceptions of African American students. As noted by a Principal during an interview, "students bring different experiences and views of the world to school that could either help or hinder their ability to successfully transverse the education system, "normal" behavior is defined using a white middle class ontology". Therefore, students not exhibiting "normal" behavior experienced increased discipline actions, 142 and special education placement. This finding is also consistent with established research that racial stereotypes influenced teacher perceptions, judgments and behavior, Devine (1989). In summary, while Principals offered no one explanation for the disproportional application of discretionary removals, they all agreed students denied access to the established academic curriculum and instruction are placed at increased risk of academic failure. Theoretical Implications As pointed out by Hanssen (1998), schools and teachers serve as mechanisms of the transmission and perpetuation of different social class values. The research confirmed discretionary discipline policies as currently used by schools tend to perpetuate cultural reproduction, with teachers as the instrument to transmit the accepted culture. While there was acknowledgement by Principals of cultural differences and cultural clashes between students and teachers, this did not mitigate the removal of students from classrooms. In other words, a teacher's ontology defined the accepted culture and behaviors. Hence, students had the 143 burden of behavior change and cultural conformity. In addition, the disproportionate application associated with discretionary removals between ethnic groups demonstrated how zero-tolerance discipline policies are racialized and in direct conflict with Civil Right laws established to promote equality and educational access for minorities. Policy Implications There is no disagreement schools must be safe, and classrooms free of disruption, but teachers need to be monitored for disproportional applications of the policy on poor and minority students. The Civil Rights Act of 1964 established a legal standard known as the "adverse impact" doctrine. This doctrine dictates when a purported racially neutral policy or practice produces a disproportionate harmful impact on students of color, the burden shifts to the school system to justify its policy or practice (Harvard Civil Rights Project, 2000). Based on the documented disproportionate number of African American and Hispanic elementary students subjected to discretionary removals and the lack of clear definition of the types of student behaviors that warrant a discretionary discipline removal, this school 144 district as well as other Texas school districts may be in violation of the adverse impact doctrine. Hence, Texas policymakers in light of the evidence should consider rescinding and/or revising the zero-tolerance discipline policy, especially at the elementary grades. Texas policymakers should also consider the number of campus discretionary discipline removal actions as part of State accountability and school performance ratings. In other words, suspension rates attributed to discretionary removals should have a negative effect on school performance ratings. This action would require campus and district level administrators to monitor student discretionary removals. Hence, Principals would be held accountable for providing counseling, professional development, and the removal of teachers who have difficulty successfully communicating with and teaching ethnically diverse students. Typically for minority students it is not "one" variable that impedes academic achievement, but rather a multitude of factors that must be addressed if we are to move beyond a band-aid approach to closing the academic achievement 145 gap between minority and non-minority students. It can be hypothesized that, students who are performing well academically and positively engaged in learning are at a decreased risk of being subjected to a discretionary discipline removal. Therefore, school administrators may consider the establishment of multidisciplinary problem solving teams composed of teachers, counselors, social workers, a parent specialist, school nurse, and other campus personnel tasked to identify intervention and prevention strategies to help both teachers and students have a successful school experience. Future Research The data presented in this report raises several questions; one of these questions is whether there is a relationship between discipline removals of African American male students in K-2nd grade, and the over identification of African American students as emotionally disturbed. This future research question is also a by-product of the Principals interviews. As noted by a Principal, "African American males subjected to discretionary removals are at increased risk of being labeled emotionally disturbed and placed into special education classes for behavior 146 management training". The Principal also pointed out when African American students first enter special education classes most are on academic grade level. However, as documented in this research, by the third-grade the most pervasive predictor of low math and reading scores beyond a discipline removal was special education placement. A second question to consider, based on Principal interviews, personal observations, and review of discipline actions at the elementary, middle and high school levels, is the majority of discretionary referrals made by the same and relatively small number (less than 3) teachers within a school. Third, would similar findings occur, if this study were replicated at suburban and/or rural school districts. General Comments In summary, the question remains that even if student behavior warrants removal, should punishment of a behavior result in academic punishment? As indicated in the findings, there is a significant negative correlation between discipline removals and academic achievement. The need for continued access to the learning environment is of particular concern for students in grades K-2 who may not have 147 necessarily entered school as academically prepared as others. Therefore, it is not a question of the location of the removal; it is a question of academic access. As noted by Sullivan (1989), punishment without meeting student needs for academic tutoring seldom provides motivation for reform. If the intent of educators and policymakers were to close the academic achievement gap for African American and Hispanic students, and to "Leave No Child Behind", then an examination of zero-tolerance discipline policies that deny students continuous access to the established curriculum and classroom instruction would be a good starting place. 148 REFERENCES Ahlquist, R. (1991). Position and imposition: Power relations in multicultural class. Journal of Negro Education, 60(2), 133-146. Adler, P.A., Kless, S.J., & Adler, P. (1992). Socialization to gender roles: Popularity among elementary school boys and girls. Sociology of Education, 65, 169-187. Banister, J. & Maher, M. (1998). Recentering multiculturalism moving toward community. Urban Education, 43(2), 182-217. Bear, G. (1998). School discipline in the United States: Prevention, correction and long-term social development. School Psychology Review, 27(1), 14-32. Bell, D. (1992). Faces at the bottom of the well. New York: Basic Books. Blair, F. E. (1999). Does zero tolerance work? Children's Health & Safety, 36-37. Bowles, S., & Gintis, H. (1976). Schooling in capitalist America. New York: Basic Books. Bernstein, B. (1977). Class, Codes, and control, Vol., 3. London: Routledge and Kegan Paul. Bowditch, C. (1993). Getting rid of troublemakers: High school disciplinary procedures and the production of dropouts. Social Problems, 40, 493507. 149 Brantlinger, E. (1991). Social class distinctions in adolescent's reports of problems and punishment in school. Behavioral Disorders, 17, 36-46. Casserly, M. (1996). Discipline & Demographics. Education Week on The Web, [Online] Available: http://www.edweek.org/ew/. Cook, T.D., & Campbell, D.T. (1979). Quasi-experimentation: Design & analysis for field settings. Chicago: Rand McNally. Costenbader, V., & Markson, S. (1998). School suspension: A study with secondary school students. Journal of School Psychology, 36(1), 59-82. Dane, F. (1990). Research Methods. Belmont, CA: Brooks/Cole. Darley, J., & Gross, P.H. (1983). A hypothesis confirming bias in labeling effects. Journal of Personality and Social Psychology, 44, 20-33. Davis, J. E., & Jordan, W. J. (1994). The effects of school context experiences on African American males in middle and high school. Negro Education, 63(4), 570-87. Delpit, L. (1988). The silenced dialogue: Power and pedagogy in educating other people's children. Harvard Educational Review, 58, 280-298. Delpit, L. (1995). Other people's children: Cultural conflict in the classroom. New York: The New York Press Diem, R.A., (1988). On campus suspensions: A case study. The High School Journal, 72, 36-39. 150 Dovido, J.F., Gaertner, S.L., Validzic, A., Johnson, B., & Frazier, S. (1997). Extending the benefits of recategorization: Evaluations, self-disclosure, and helping. Journal of Experimental Social Psychology, 33, 401-420. Ekstrom, Goertz, Pollack & Rock (1986). Who drops out of high school and why: Findings from a national study. Teachers College Record, 87, 357-73. Fanon, F. (1967). Black skins, white mask. New York: Groves Press. Fazio, R.H., Jackson, J.R., Dunton, B.C. & Williams, C.J. (1995). Variability in automatic activation as unobtrusive measure of racial attitudes: A bona fide pipeline? Journal of Personality and Social Psychology, 69, 1013-1027. Felice, L. (1981). Black student dropout behavior: Disengagement from school, rejection and racial discrimination. Journal of Negro Education, 50(4), 415-424. Ferguson, R. (1991). Racial patterns in how school and teacher quality affect achievement and earnings. Challenge: A Journal of Research on Black Men, 2, 1-26. Finn, J.D., & Cox, D. (1992). Participation and withdrawal among fourthgrade pupils. American Educational Research Journal, 29, 141-162. Fiscus, J. (2000). Effective policy or display of administrative machismo? The Safety Zone, 2 (1). [Online] Available: http://www.safetyzone.org Foster, M. (1990). The politics of race: Through the eyes of African American teachers. Journal of Negro Education, 173 (3), 122-141. 151 Foster, M. (1993). Resisting racism: Personal testimonies of AfricanAmerican teachers. In L. Weis and M. Fine (Eds.) Beyond silenced voices: Class, race, and gender in United States schools (pp. 273-288). Albany: State University of New York Press. Fremon, C. & Hamilton, S.R. (April, 1997). Are schools failing black boys? Parenting, 116-32. Glickman, C. (Fall,1998). Revolution, education, and the practice of democracy. The Educational Forum, (63), 16-22. Gonzalez, N., Moll, L., Tenery, M. Rivera, A., Rendon, P., Gonzalez, R., & Amanti, C. (1995). Funds of knowledge for teaching in Latino households. Urban Education, 29 (4), 443-470. Gordan, R., Della Piana, L., & Keleher, T. (2000). Facing the consequences: An examination of racial discrimination in U.S. Public Schools. Oakland, CA: Applied Research Center. Gregory, J. F. (1997). Three strikes and they're out: African American Boys and American Schools' responses to misbehavior. International Journal of Adolescence and Youth, 7, 25-34. Haberman, M. (1992). Does direct experience change education students perceptions of low-income minority children? Midwestern Educational Research, (5), 28-31. Harvard Civil Rights Project (2000). Opportunities Suspended: The devastating consequences of zero tolerance and school discipline policies. [Online] Available: http://www.law.harvard.edu 152 Hanssen, E. (1998). A white teacher reflects on institutional racism. Phi Delta Kappan, 79, 694-698. Hyman, I.A. & Snook, P. (March, 2000). Dangerous schools and what you can do about them. Phi Delta Kappan, 489-501. Imich, A.J. (1994). Exclusions from school: Current trends and issues. Educational Research, 36, 3-11. Irvine, J.J. (1990). Black students and school failure. Westport, CT: Greenwood Press. Jackson, G (1975). The research evidence on effects of grade retention. Review of Education Research, 45, 613-635. Jimerson, S., Carlson, E., Rotert, M., Egeland, B., & Sroufe, A. (1997). A prospective, longitudinal study of the correlates and consequences of early grade retention. Journal of School Psychology, 35(1), 3-25. Kaeser, S.C. (1979). Suspensions in school discipline, Education and Urban Society, 11, 465-484. King, J. (1991). Dysconscious racism: Ideology, identity, and miseducation of teachers. Journal of Negro Education, 60 (2), 133-146. Kohn, A. (1996). Beyond Discipline: From compliance to community. Alexandria, VA: Association for Supervision & Curriculum Development. Ladson-Billings, G. (1995). Toward a theory of culturally relevant pedagogy. American Educational Research Journal, 32 (3), 465-491. Latham, A. (1998). Rules & learning. Educational Leadership, 104-105. 153 Ledingham, J.E., & Schwartzman, A.E. (1984). A 3-year follow up of aggressive and withdrawn behavior in childhood: Preliminary findings. Journal of Abnormal Child Psychology, 12, 157-168. Marshall, M. (1998). Empower rather than overpower. Education Week, 17 (7), 32-36. Maurer, K.L., Park, B., & Rothbart, M. (1995). Subtyping versus subgrouping processes in stereotype representation. Journal of Personality and Social Psychology, 69(5), 812-825. Mayer, G.R., (1995). Preventing antisocial behavior in the schools. Journal of Applied Behavior Analysis, 28, 467-468. McCarthey, J.D., & Hoge, D. R. (1987). The social construction of school punishment: Racial Disadvantaged out of universalistic process. Social Forces, 65, 1101-1120. McFadden, A.; Marsh, G.; Price, B.; and Hwange, Y. (1992). A study of race and gender bias in the punishment of school children. Education and Treatment of Children, 15 (2), 140-146. Mizell, M.H. (1978). Designing and implementing in-school alternatives to suspension. The Urban Review, 10,213-226. Moles, O.C., (Ed.). (1989). Strategies to reduce student misbehavior. Washington, D.C.: Office of Educational Research and Improvement. Morgan- D'Atrio, C., Northup, J., LaFleur, L., & Spera, S. ( 1996). Toward prescriptive alternatives to suspensions: A preliminary evaluation. Behavioral Disorders, 21(2), 190-200. 154 Morris, A. A. (1980). The Constitution and American education. Minneapolis, MN: West Publishing Company. Morrison, G.M., Furlong, M.J., & Morrison, R.L. (1994). School violence to school safety: Reframing the issue for school psychologists. School Psychology Review, 23, 236-256. Nicols, J., Ludwin, W., & Iadicola, P. (1999). A darker shade of gray: A year-end analysis of discipline and suspension data. Equity & Excellence in Education, 32(10), 43-55. Niklason, L. (1984). Nonpromotion: A pseudoscientific solution. Psychology in the Schools, 21, 485-499. Noguera, P. (1995). Preventing and producing violence: A critical analysis of responses to school violence. Harvard Educational Review, 189-212. Oakes, J. (1994). Tracking, inequality, and the rhetoric of reform: Why schools don't change. In J. Kretovics & E. J. Nussel (eds.), Transforming urban education (pp. 146-164). Needham, MA: Allyn & Bacon. Oppenheimer, J., & Ziegler, S. (1988). Suspension, alternatives to suspension and other approaches to discipline, (189) Toronto, Canada: Toronto Board of Education. Orfield, G. (2000). Our resegregated schools. Principal, 79 (5 ) 6-11. Polite, V. (1993b). If we knew then what we know now: Foiled opportunities to learn in suburbia. Journal of Negro Education, 62(1), 12-23. Portner, J. (2000). School violence down, but worry high [Online] Available: Education Week on the Web http//www.edweek.org.ew/. 155 Quinn, T. (1991). The influence of school policies and practices on dropout rates. NASSP Bulletin, 75, 73-83. Rodney, L.; Crafter, B.; Rodney, E.; and Mupier, R. (1999). Variables contributing to grade retention among African American adolescent males. The Journal of Educational Research, 92 (3), 185-190. Royer, E. (1995). Behavior disorders, exclusion and social skills: Punishment is not education. Therapeutic Care and Education, 4(3), 32-36. Safer, D. J. (1986). The stress of secondary school for vulnerable students. Journal of Youth and Adolescence, 15(5), 405-417. Sanders, E.T. W., & Reed, P.L. (1990). An investigation of the possible effects of an immersion as compared to a traditional program for African-American males. Urban Education, 30, 93-112. Sandoval, J. (1984). Repating the first grade: How the decision is made. Psychology in the Schools, 21, 457-462. Shaw, S.R., & Braden, J.P. (1990). Race and gender bias in the administration of corporal punishment. School Psychology Review, 19, 378-383. Shores, R.E. Gunter, P. L., & Jack, S. L. (1993). Classroom management strategies: Are they setting events for coercion? Behavioral Disorders, 19, 92-102. Skiba, R., & Peterson, R., (1999). The dark side of zero tolerance. Phi Delta Kappan, 80 (5), 372- 382. Skiba, R., & Peterson, R. (2000) School discipline at a crossroads: From zero tolerance to early response. Exceptional Children, 66 (3), 335-347. 156 Skiba, R., Peterson, R.L., and Williams, T. (1997). Office referrals and suspension: Disciplinary intervention in middle schools. Education and Treatment of Children, 20(3), 1-21. Skiba, R., Michael, R., Nardo, C., and Peterson, R. (2000). The color of discipline: Sources of racial and gender disproportionality in school punishment. Policy Research Report (SRS1): Indiana Education Policy Center. Snyder, T. & Hoffman, C. (1994). Digest of Education Statistics. National Center for Educational Statistics ( 94-115): U.S. Department of Education, Office of Educational Research and Improvement: Washington, D.C. Spring, J. (1972). The rise and fall of the corporate state. Boston: Beacon Press. Sullivan, J. S. (1989, April). Elements of a successful in-school suspension program. NASSP Bulletin, 33-38. Tatum, B. D. (1994). Teaching white students about racism; The search for white allies and the restoration of hope. Teachers College Record, 95 (4), 462-476. Townsend, B. (2000). Disproportionate discipline of African American children and youth: culturally responsive strategies for reducing school suspensions and expulsions. Exceptional Children, 66, 381-391. U.S. General Accounting Office. (2002). School dropouts education could play a stronger role in identifying and disseminating promising prevention strategies (GAO Publication No. 02-240). Washington D.C. : Author. Villegas, A. (1991). Culturally responsive pedagogy for the 1990's and beyond. Washington DC: ERIC Clearinghouse on Teacher Education. 157 Ward, C. (1998). Student discipline and alleviating criminal behavior in the inner city. The Urban Review, 30 (1), 29-48. Wehlage, G. & Rutter, R. (1986). Dropping out: How much do schools contribute to the problem. Teachers College Record, 87 (3), 374-392. Wehlage, G., Rutter, R., Smith, G. A., Lesko, N., & Fernandez, R. (1989). Reducing the risk: Schools and communities of support. New York: Falmer Press. Wheelock, A., & Dorman, G. (1988). Before it's too late: Dropout prevention in the middle grades. Boston: Massachusetts Advocacy Center and the Center for Early Adolescence. Williams, J. (1989). Reducing the disproportionately high frequency of disciplinary actions against minority students: An assessment-based policy approach. Equity and Excellence, 24(2), 31-37. Wu, S., Pink,W., Crain, R., & Moles, O. (1982). Student suspension: A critical reappraisal. The Urban Review, 14 (4), 245-304. 158 Vita Florence Linelle Clark was born in Washington, D.C. on August 18, 1957, the daughter of Annette M. Ellis. After graduation from Largo High School, Largo, Texas in 1974, she briefly attended Towson State College, and Prince George's Community College, eventually transferring to the University of Maryland in College Park and graduating in 1978, with a Bachelor of Arts degree. Upon graduation, she entered the United States Air Force (USAF) as a Commissioned Officer serving on active duty for six years, and as a reservist for eight years. Upon departing the military in 1984 after serving in Desert Storm, she held various management positions within the private and public employment sector. In 1995, she accepted a position with the Austin Independent School District as their Community Business Liaison, and later selected as the Dropout Prevention/ Reduction Coordinator. Permanent address: 9624 Copper Creek Dr. This dissertation was typed by Ms. Sara Cale. 159