|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 A PROFILE OF SPECIAL EDUCATION TEACHERS: WHO STAYS? By MELISSA R. DUNN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010
2 2010 Melissa R. Dunn
3 To my husband Bill
4 ACKNOWLEDGMENTS There are many individuals without whom I would not have been able to fulfill my goal of completing my doctorial work. First and foremost is my family. I wish to thank my husband for his patience, support, and ongoing encouragement. I wish to thank my children and grandchildren for their understanding when I missed birthdays and celebrations. They have been a blessing. I also thank my mother and father for their encouragement and confidence in my abilities. Many thanks are necessary to my program cohort members who made the journey meaningful. I appreciate my committee members and their depth of knowledge as they have assisted me through the process. I owe special thanks to Jean Crockett. She has been instrumental in pushing me to think and explore ideas. Her editing skills have also been invaluable and have helped me become a more thoughtful writer. There are many individuals who have encouraged me throughout this process and to all I owe a debt of gratitude.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................. 4 TABLE OF CONTENTS .................................................................................................. 5 LIST OF TABLES ............................................................................................................ 8 LIST OF FIGURES ........................................................................................................ 10 ABSTRACT ................................................................................................................... 11 1 INTRODUCTION .................................................................................................... 13 Teacher Shortages ................................................................................................. 13 Retention or Supply? ........................................................................................ 14 Geographic Location and Demographics ......................................................... 16 High Need Subject Areas ................................................................................. 16 Policy Initiatives ................................................................................................ 17 Teacher Quality ...................................................................................................... 18 Teacher Preparation and Certification Routes ........................................................ 21 Research Problem .................................................................................................. 22 Theoretical Perspective .......................................................................................... 24 Purpose of the Study .............................................................................................. 25 Research Questions ............................................................................................... 26 Definition of Terms .................................................................................................. 26 Significance of the Study ........................................................................................ 27 Limitations ............................................................................................................... 28 2 REVIEW OF LITERATURE .................................................................................... 30 Federal and State Policy ......................................................................................... 30 Teacher Labor Market ............................................................................................. 32 Teacher Shortage: Retention or Supply? ................................................................ 33 Why Teachers Leave .............................................................................................. 35 Working Conditions .......................................................................................... 35 Individual Characteristics .................................................................................. 36 School Characteristics ...................................................................................... 37 Teacher Supply ....................................................................................................... 39 Alternative Certification Routes ............................................................................... 41 Characteristics of Effective Alternative Certification Routes ................................... 45 Alternative Route Teachers in HighNeed Schools .......................................... 47 Special Education Alternative Routes to Certification ....................................... 48 Equity Issues .......................................................................................................... 49 Teacher Quality ................................................................................................ 49
6 Impact of Turnover on Student Achievement ................................................... 52 Hiring Practice s ....................................................................................................... 53 Job Choice Theory .................................................................................................. 55 Realistic Job Previews ..................................................................................... 56 PersonOrganization Fit Theory ....................................................................... 57 Summary ................................................................................................................ 58 3 METHODOLOGY ................................................................................................... 61 Research Design .................................................................................................... 62 Setting and Context ................................................................................................ 66 Sample ............................................................................................................. 67 Data Sources and Collection ............................................................................ 67 Definition of Variables ...................................................................................... 68 School and Teaching Assignment Variables .................................................... 69 School poverty level ................................................................................... 69 Percentage of minority students ................................................................. 70 Job Code .................................................................................................... 71 Teacher Preparation Variable ........................................................................... 71 Teacher Preparation ......................................................................................... 72 Data Analysis .......................................................................................................... 72 PersonPeriod Data Set .......................................................................................... 73 Summary ................................................................................................................ 76 4 ANALYSIS AND RESULTS .................................................................................... 78 Description of Sample ............................................................................................. 78 Research Questions ............................................................................................... 81 Predictor Variables ................................................................................................. 81 Teacher Characteristics .................................................................................... 82 Teacher Preparation ......................................................................................... 83 School Assignment ........................................................................................... 84 Survival Analysis ..................................................................................................... 86 Teacher Characteristics .................................................................................... 89 Change Variable ............................................................................................... 90 Teacher Race ................................................................................................... 91 5 SUMMARY AND DISCUSSION ............................................................................ 106 Findings ................................................................................................................ 107 Discussion ............................................................................................................ 107 Job Choice Theory ......................................................................................... 118 Urban Districts ................................................................................................ 119 Implications for Practice ................................................................................. 120 Recommendations for Future Research ......................................................... 121 LIST OF REFERENCES ............................................................................................. 122
7 BIOGRAPHICAL SKETCH .......................................................................................... 132
8 L IST OF TABLES Table page 3 1 Personperiod data set ....................................................................................... 77 4 1 Frequency table: Descriptive data for invariant teacher characteri stics .............. 93 4 2 Frequency table: Descriptive data for teachers with differing types of teacher preparation ......................................................................................................... 94 4 3 Frequency table: Descri ptive statistics providing the school demographic data where the newly hired special education teachers are assigned to teach .. 95 4 4 Frequency table: Descriptive statistics of the racial breakdow n of the special education teachers and the minority level of their school assignment ................ 95 4 5 Teacher gender and race: Does the sample teachers gender or race impact whether they stay in or leave teaching after their first year? ............................... 97 4 6 Teacher age: Does the sample teachers age impact whether they stay in or leave teaching after their first year? .................................................................... 97 4 7 Teaching experience: Does the sample teachers prior teaching experience impact whether they stay in or leave teaching after their first year? ................... 98 4 8 Teacher Preparation: Does the sample teachers level or type of teacher preparation impact whether they stay in or leave teaching after their first year? .................................................................................................................. 98 4 9 School Minority Level: Does the minority population of the school assignment impact whether the sample teachers stay in or leave teaching after their first year? .................................................................................................................. 99 4 10 School Poverty Level: Does the level of poverty in the school impact whether the sample teachers stay in or leave teaching after their first year? ................... 99 4 11 Exceptionality Assignment: Does the exceptionality assignment impact whether the sample teachers stay in or leave teaching after their first year? ..... 99 4 12 Change in teaching assignment: Does a change in teaching assignment impact whether the sample teachers stay in or leave teaching ? ....................... 100 4 13 Job change: Does the level or type of teacher preparation impact whether the sample teachers change teaching assignments ? ............................................. 100 4 14 Spell one survival anal ysis: Distribution of event occurrence for special education teachers beginning 2003 ending 2009 ............................................. 100
9 4 15 Survival analysis: How prior teaching experience impacts staying in teaching 101 4 16 Survival analysis: Job change and the impact on leaving teaching .................. 101 4 17 Spell two survival analysis: Distribution of event occurrence for special education teachers beginning 2003 ending 2009 ............................................. 102 4 18 Survival analysis: How the teachers race impacts a return to teaching ........... 102
10 LIST OF FIGURES Figur e page 4 1 Survival distribution of the sample teachers over time. ..................................... 103 4 2 Survival distribution of teachers with varying levels of pr ior teaching experience. ....................................................................................................... 103 4 3 Survival distribution for teachers changing teaching assignment and leaving teaching. ........................................................................................................... 104 4 4 Survival distribu tion of the returning sample teachers over time. ...................... 104 4 5 Survival distribution of the returning sample teachers and race. ...................... 105
11 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy A PROFILE OF SPECIAL EDUCATION TEACHERS: WHO STAYS? By Melissa R. Dunn August 2010 Chair: David M. Quinn Cochair: Jean B. Crockett Major: Educational Leadership The purpose of this study was to understand more about the characteristics of special education teachers who are retained in a large urban school district. Although teacher shortages and teacher retention have been subjects of research for some time, much has changed in recent years regarding who is entering the field and how they are becoming qualified to teach. The demographic of the teaching force is changing dramatically with a n increasing number of men, minority candidates, and midcareer changers taking advantage of alternative certification route programs that allow them to begin teaching and complete certification requirements simultaneously. Given the sig nificant impact that teachers have on student achievement, research into who is teaching, how they are prepared, and whether they stay has added importance. This study utilized survival analysis methodology to examine teacher retention behavior over a per iod of six years within the context of their varying teaching assignments in a large urban district. The study attempted to identify significant predictor variables that would indicate which teacher characteristics might signify a greater likelihood of the teacher staying in teaching. An analysis of the of the level and
12 type of teacher preparation revealed that the teachers who stayed in teaching and the teachers who left teaching did not differ significantly with regard to their teacher preparation. The fi nding that the risk for leaving teaching did not differ significantly for teachers with special education training is not consistent with other research, which has found that more extensive preparation in special education is associated with higher retenti on, demonstrating the need for continued research. The survival analysis provided statistics consistent with current research regarding retention. The findings indicated a 71% probability that teachers would survive through their first year and a 44% probability of surviving through their third year. By their sixth year, a little more than a third of the teachers were expected to still be teaching. A variable of interest that arose during the process of analyzing the data was job change. Among the sample t eachers who changed job assignments, the retention was twice as high as that of the teachers who stayed in the same assignment. The study discusses the importance of supportive hiring practices and working conditions for special education teachers.
13 CHAPTER 1 INTRODUCTION Many states across the United States are facing serious budget challenges, which result in education programs competing with other essential programs for adequate funding. Considering the billions of dollars that are spent on recruit ment, training, and retention efforts every year we must focus attention on ways to address chronic teacher shortages. Why do some teachers stay and some leave? This research study explored relationships that exist between special education teacher charac teristics, teacher preparation, teaching assignments, school assignments, and retention. Teacher Shortages Teacher shortages exist in varying degrees and types throughout the nation. Research has also demonstrated that the supply of teachers differs in both quality and quantity according to geographic location, socioeconomic factors, and subject area ( Council of the Great City Schools, 2000; Ingersoll, 2003; Murphy, DeArmond, & Guin, 2003). There has been a pattern of steadily increasing demand for both general education teachers and special education teachers for some time (American Association for Employment in Education [AAEE], 2006; Cook & Boe, 2007). The Teacher Attrition and Mobility study which used 200405 Teacher Follow Up survey data showed that the percentage of teachers leaving the profession grew from 5.6 % during the 198889 survey to 8.4 % during the most current survey. The statistics for teachers changing schools were relatively stable, growing from 7.9% in 198889 to 8.1% in the 200405 data. However, the demand for special education teachers grew at
14 a much higher rate (38%) than the demand for general education teachers (26%) from 198788 to 199900 (U.S. Department of Education, 2004). Retention or Supply? There are differing perspecti ves about whether teacher shortages are attributable to retention or supply. Ingersoll (2001) posits that the difficulty schools experience with staffing classrooms is due primarily to the revolving door caused by excessive turnover. Ingersoll and other re searchers contend that the supply of teachers is adequate to meet the demand (Hanushek, Kain, & Rivkin, 2004; National Commission on Teaching and Americas Future [NCTAF], 2003). Teacher characteristics and school demographics have been shown to impact re tention. Teachers leave in greater numbers within the first few years of their career and the numbers increase again as teachers near retirement. Factors such as inadequate support from the school administration, lack of adequate induction support and preparation, low salary, as well as increased career opportunities for women negatively impact teacher retention (Ingersoll, 2003; Johnson, 2006; Strunk & Robinson, 2006). Billingsley (2003) noted that teacher attrition is a major contributor to the shortage of special education teachers. In addition to the working conditions that general education teachers face, special education teachers have the added responsibilities of paperwork, unmanageable client loads, and issues with those in leadership positions who may have little expertise in special education. In contrast, Cook & Boe (2007) characterize the shortage of qualified teachers as an imbalance in supply and demand. The majority of the teaching force is comprised of continuing teachers, 92% based on 199900 Public School Teacher Questionnaire
15 (PSTQ) surveys. The remaining 8% is filled by first time teachers and returning teachers. Considering the size of the teacher workforce, an 8% loss translates to a considerable number of teachers who must be replaced. The 200405 Teacher Follow Up Survey (TFS) data report a work force of 3,214,900 and a loss of 8.4% or 269,600 teachers (U.S. Department of Education, 2004). Returning teachers are part of what is referred to as the reserve pool, which is composed of experienced teachers who left teaching and those who were prepared to teach but delayed entering the field. Approximately half of the vacancies in special and general education are filled from the reserve pool. Little is actually know about the size and qualit y of this source, which supplies a large portion of new hires annually (Cook & Boe, 2007). However, studies have shown that former educators are most likely to return to teaching if they do so soon after leaving. The longer a teacher stays out of the classroom the less likely they are to return. Earlier studies demonstrated that special education teachers typically return at a higher rate, and Black educators were shown to have a 50% higher probability of returning to teaching than their White counterparts (Singer, 1993, Willett & Singer, 1995). There has been a decline in the supply of special education graduates since 199798. Each year special education teacher vacancies are filled by general education teachers and individuals pursuing alternative routes to teacher certification. This supply does not mitigate the need to replace thousands of special education teachers who are not qualified for their current placements (Cook & Boe, 2007; McLeskey, Tyler, & Flippin, 2004).
16 Geographic Location and Demographi cs Teacher shortages vary widely within each state and between geographic regions (Murphy, et al., 2003). Research has shown a tendency for new teachers to seek out positions in schools that are in close proximity to their hometowns, or in demographically similar communities (Boyd, Lankford, Loeb, & Wyckoff, 2005). The demographic composition of communities and schools plays an important role in teacher recruitment and retention. This is evident in urban school districts where they continuously struggle to staff challenging schools and in the recruitment and retention efforts of rural districts. Turnover is higher in these hard to staff schools ( Council of the Great City Schools, 2000; Murphy, et al., 2003). The teachers race also plays a role. White teachers are more likely to turnover than their Black counterparts if the schools they are teaching in have a high proportion of minority students. Black teachers have also been shown as more likely to return to teaching after a spell out of teaching (Singer, 1993). Whether this is a function of racial composition, or the high poverty conditions that typically are characteristic of urban schools, is unclear (Hanushek, et al., 2004; Strunk & Robinson, 2006). These issues have policy implications considering the large numbers of new teachers that many urban school districts hire from the surrounding suburbs (Boyd, et al., 2005). High Need Subject Areas There are chronic teacher shortages in districts across the country, particularly in highneed subject areas such as mathematics, sciences, foreign languages, and special education. As immigrant populations increase, many states are also experiencing a critical shortage of English as a second language (ESL) teachers (Ingersoll, 2001; Tyler, Yzquierdo, Reyna, & Flippin, 2004). These shortages vary widely within each state, and
17 it is no surprise that highpoverty schools experience shortages in highneed subject areas the most acutely (Ingersoll, 2001; Urban Teacher Challenge, 2000). There is general agreement among researchers that a chronic and pervasive shortage of qualified special education teachers exists across all geographic regions in the country with 98% of the states reporting considerable shortages (Billingsley, 2003; Brownell, Hirsch, & Seo, 2004; Rosenberg & Sindelar, 2005). Projections from the Bureau of Labor Statistics indicate that the number of special education teachers is expected to increase by 15% between 2006 and 2016. A 20% increase in the need for special education teachers in preK, kindergar ten, and elementary is projected during this period. The increases are due to the growing number of students who require special education services, which is attributed to earlier diagnosis and advances in medical treatments. Educational reforms requiring higher standards for graduation and an increased emphasis on training and employment for the disabled population are also credited with the growing need for services and teachers (U.S. Department of Labor, 2008). Policy Initiatives State and federal mandat es have an impact on the supply and demand for teachers. Class size amendments in states such as California and Florida compound their teacher shortages by increasing the number of teachers required to reduce the teacher student ratio. Federal legislation such as the 2001 No Child Left Behind Act (NCLB) mandates that teachers meet specific requirements to be deemed highly qualified. These requirements may exacerbate the difficulties school districts have recruiting teachers, particularly special education and science teachers who are often
18 responsible for teaching multiple subject areas (Dai, Sindelar, Denslow, Dewey, & Rosenberg, 2007; Johnson, Birkeland, & Peske, 2005). Alternative routes to certification have proliferated in a political environment that h as demanded a greater supply of qualified teachers. In 2003, Congress appropriated $41.65 million for the Transition to Teaching program to enable midcareer persons to pursue alternative routes into the classroom (Mikulecky, Shkodriani, Wilner, 2004). All 50 states and the District of Columbia now have some type of alternate route to teacher certification with approximately 485 alternate route programs being implemented (Feistritzer, 2008). One of the more visible alternative route programs is Teach For Am erica (TFA). A program founded in 1989 with the mission of closing the achievement gap in high poverty schools. TFA recruits nationally for seniors and recent graduates from competitive colleges who have strong academic backgrounds and demonstrated leaders hip qualities. TFA core members recruited in 2003 had an average SAT score of 1310 and 3.5 GPA. It is important to note that core members commit to teaching for a minimum of two years (Decker, Mayer, Glazerman, 2004). This limited commitment is one of the perceived drawbacks for high need schools striving to create a stable learning environment. Teacher Quality Studies focusing on teacher quality and student achievement have examined attributes that are frequently identified as indicators of teacher qualit y: teacher preparation, licensure, years of experience, advanced degrees, and academic proficiency (DarlingHammond, 1999; Eide, Goldhaber, & Brewer, 2004; Nougaret, Scruggs, & Mastropieri, 2005). The results have been mixed, and there is no strong consens us regarding the value of pedagogical preparation for teachers found in the
19 literature. It seems that the quality and content of teacher training programs vary so greatly that the impact is not always clear (CochranSmith & Fries, 2005; Goldhaber & Anthony 2003). It is evident that more research is required. Wilson, Floden, & Ferrini Mundy (2001) provided recommendations for future research that emphasized a set of research design principles to ensure the establishment of a foundation of credible teacher preparation research. In a recent study Boe, Shin, & Cook (2007) found that extensive teacher preparation for both general and special education teachers contributed greatly to a number of teacher qualification indicators. Extensive preparation was defined as completing 10 weeks or more of practice teaching in addition to the following components: coursework in adapting instructional materials and educational psychology, observing other classroom teaching, and receiving feedback on teaching. Although this definition demonstrates a relatively low bar for extensive preparation, Boe, et al., (2007) found these indicators to be connected with becoming fully certified, teaching in field, feeling prepared to plan effectively, being prepared to handle classroom management, and feeling prepared to use a variety of instructions methods. All of which are attributed to more qualified teachers and the tendency to stay in teaching. The specific indicator of classroom management preparation was positively correlated with a beginning teachers likelihood of staying in teaching (Boe, Cook, & Sunderland, 2008a). Studies of teacher licensure as an indicator of teacher quality have also produced contradictory results. Some researchers have found that students of teachers certif ied through alternative routes do at least as well in math and reading as students whose
20 teachers are fully statecertified. For example, in an experimental study in which students were randomly assigned to TFA teachers and a mix of other novice teachers a nd certified veteran teachers, TFA teachers had a statistically significant positive impact on math achievement. Reading achievement was about the same for TFA students and control students (Decker, et al., 2004). However, other studies have found that ful ly licensed teachers are more effective (DarlingHammond, Holtzman, Gatlin, & Heilg, 2005). Goldhaber and Anthony (2003) concluded that there is not a strong enough research base from which to draw definitive conclusions about the relationship between teac her licensure and student learning. Some research findings have indicated that there is not a strong correlation between teachers having advanced degrees and student achievement. The exceptions were mathematics teachers and reading teachers (DarlingHammo nd, 1999; Wayne & Youngs, 2003). Years of experience have been positively related to student outcomes, indicating that first year teachers are generally less effective in improving student test scores (Boyd, Lankford, Loeb, Rockoff, & Wyckoff, 2007). A pos itive relationship between teachers' academic proficiency and student achievement has also been demonstrated (Wayne & Youngs, 2003) Goldhaber & Anthony (2003) noted in their Indicators of Teacher Quality that the majority of the studies on teacher academi c proficiency and student learning show that the teachers academic proficiency may represent one of the best predictors of teacher quality. It is generally recognized that teacher quality is one of the most important factors impacting student achievement (Sanders, & Rivers, 1996; Wright, Horn, & Sanders, 1997), and research has demonstrated that teaching experience is positively correlated
21 with student achievement (Boyd, et al., 2007; Johnson, 2006; Kane, Rockoff, & Staiger, 2006; Rockoff, 2004). Therefor e, children with disabilities are not only affected by the lack of qualified special education teachers; high turnover rates mean the students will have less experienced teachers as well (Nougaret, et al., 2005). Of first time SETs hired in 199000, only 46% were extensively prepared to teach in special education ( Boe, Cook, & Sunderland, 2007, p. 36). Teacher Preparation and Certification Routes Amid efforts to address the ongoing problem of teacher shortages there are disagreements about what it means for teachers to be qualified and about how teachers should be prepared. Many in the field of education see alternative routes to teacher certification programs as a serious threat and a means of deregulating the profession (Brownell, Ross, Colon, & McCallu m, 2005; DarlingHammond, et al., 2005). Others would like to consider teacher preparation more broadly, focusing on pathways into the classroom for talented individuals to become skilled teachers rather than focusing on program types ( Boyd, Grossman, Hamm erness, et al., 2007; Humphrey, Wechsler, & Hough, 2008). Much like traditional teacher preparation programs, there is tremendous variation among alternative certification programs: their goals, selection process, preservice experiences, induction support links to researchbased professional development, cost, and time (Humphrey & Wechsler, 2007; Rosenberg, Boyer, Sindelar, & Misra, 2007; Zeichner & Schulte, 2001). This variation in programs makes it difficult to compare outcome measures and provide the r igorous research necessary to determine the most effective ways to prepare quality teachers (Wilson, et al., 2001; Zeichner & Conklin, 2005).
22 It is becoming apparent that there is as much variation within programs as between them (Boyd, Grossman, Lankford, Loeb, & Wyckoff, 2006). In their study of effective characteristics of alternative certification programs, Humphrey and Wechsler noted that how participants experience their program depends a great deal on what they bring to the program. It is a combinat ion of the program, person, and contextual elements that determines outcome, Clearly, much more needs to be known about alternative certification participants and programs and about how alternative certification can best prepare highly effective teachers (2007 p. 523). Research Problem The shortage of special education teachers is both chronic and pervasive. McLeskey, et als., (2004) review of research on special education teacher supply and demand utilized data from the U.S. Department of Education Offi ce of Special Education Programs, National Center for Education Statistics Data, and the American Association for Employment in Education in order to investigate the extent of the special education teacher shortage. Data from the 20002001 school year indi cate that approximately 11% of all SETs were not fully certified, translating into approximately 808,000 students taught by less than fully qualified teachers. More recent data for teachers providing special education and related services to students ages 6 through 12 under IDEA, Part B show 11% were not fully certified in 2003, these numbers dropped to 10% in 2004 and 2005. The classification was changed to not highly qualified in 2006, when again the data indicate that 11% of the teachers were not highly qualified. Teacher shortages in many subject areas are due to excessive demand from attrition, insufficient supply, or both (Boe, Cook, & Sunderland, 2008b). Multiple factors that contribute to the lack of qualified special education teachers have been i dentified.
23 Inadequate numbers of qualified new special education teachers are prepared to teach, Of firsttime SETs hired in 199000, only 46% were extensively prepared to teach in special education ( Boe, et al., 2007, p. 36). Many first time teachers who have been prepared to teach general education students accept special education assignments to gain entry into teaching. Therefore, it is not surprising when these teachers leave the field to take a position for which they are better qualified. The increasing demand for qualified special education teachers is also a result of an increasing population of children identified as having disabilities (Cook & Boe, 2007; McLeskey, et al., 2004). Teacher turnover, defined as teachers leaving teaching, switching t eaching areas, or moving to a different school, increased by over a third during 199192 to 20002001. That equates to 2223% of public special education teachers and general education teachers turning over annually during this period. The effects of teacher turnover are longreaching and contribute to a lack of stability in the teaching force, which impacts the functioning of schools and most importantly student achievement, particularly for the vulnerable population of students with disabilities (Boe, et al., 2007). The financial impact of teacher turnover is also an important aspect that should be taken into consideration. Based on U.S. Department of Labor turnover cost estimates, the Alliance for Excellent Education (2005) computed replacement costs of about $4.9 billion annually for teachers who leave teaching employment and who move to another school. Considering the tight budgets that many school districts operate within, this is funding that could be put to use in other areas of need. The financial i ncentive associated with improving teacher turnover is an important issue for research and education policymakers (Boe, et al., 2008b).
24 Theoretical Perspective Research in organizational behavior and management provides the theoretical framework that grounds this study. Jobchoice theory and organizationperson fit theory are considered as additional lenses through which school districts can examine their recruitment and hiring practices. Given the challenges of recruiting and retaining qualified special education teachers, the value of a good fit between the teacher and the school setting takes on greater importance. Ingersolls (2001) analysis of hiring patterns indicates that staffing problems result not just from a shortage of qualified teachers but because many new recruits leave within a short time of hire. School districts and principals, as employers, should consider post hire outcomes such as retention rates, performance, and job satisfaction at the beginning of the hiring process as part of their recruitment planning and objectives. The way employers recruit and place new teachers can influence post hire outcomes (Breaugh & Starke, 2000). The organizational behavior and management research literature provides insight into how job choice, organizati on fit, and job fit theories can help inform the pre and post hiring process (Liu & Johnson, 2006). Kristof (1996) defines personorganization fit as the compatibility between people and organizations that occurs when: (a) at least one entity provides what the other needs, (b) they share similar fundamental characteristics, or (c) both. We know that teacher attrition varies a great deal based on teacher characteristics, the subject area assignments, and the school settings themselves (Ingersoll, 2003). Therefore, a carefully considered match between the prospective teacher and the teaching job has implications for job satisfaction and retention. School recruiting representatives should recognize that career changers have become an important pool of candidates for
25 teacher recruitment in both special and general education (Feistrizer, 2005; Rosenberg, et al., 2007). Consequently, the match between these prospective teachers and teaching assignments takes on added importance, given that second career individuals may have other employment options (Dai, Sindelar, Denslow, & Dewey, 2007). Purpose of the Study The purpose of this study was to understand more about the characteristics of special education teachers who are retained in a large urban school district. The overarching goal was to ascertain who stays, who leaves, and who returns. Although teacher retention has been studied extensively, the demographic of the teaching force is changing dramatically ( Feistritzer, 2005; Johnson, 2004) Alternative certifica tion routes have attracted talented and diverse career changers from a variety of backgrounds who may not have otherwise chosen to enter the field of teaching. In addition to contributing important information to the retention literature, identifying a pr ofile of individuals who chose to teach special education and tended to be retained at higher rates, provides valuable information to teacher preparation program providers, recruitment, and staffing managers. The study examined the personal and academic background characteristics of newly hired special education teachers to determine if relationships existed between individual characteristics, teaching assignments, specific school assignments, and retention. The study also sought to predict how many years after hire a special education teacher may teach before leaving. The annual retention and school assignment status for all special education teachers hired beginning 20032004 through 20052006 were reviewed for a period of six years.
26 Research Questions The research questions guiding this study were: 1. Does the risk of leaving teaching differ with respect to a special education teachers personal characteristics? 2. Does the risk of leaving teaching differ with respect to a special education teachers level or ty pe of preparation? 3. Does the risk of leaving teaching differ with respect to a special education teachers school assignment (percentage of poverty and minority student enrollment)? 4. Does the risk of leaving teaching differ with respect to a special educator s teaching assignment? 5. How many years on average do newly hired special education teachers continue to teach in a large urban school district? 6. What is the probability of a special education teacher returning to teaching after leaving? The research quest ions will focus on the following variables: (a) teacher characteristics race, gender, age; (b) teacher preparation route; (d) special education teaching assignment students disability classification; (e) school demographics percentage of free and reduced lunch, percentage of minority students served. Definition of Terms The definition of teacher retention and turnover behaviors will be the same as those outlined in Boe, et als., (2007) study of teacher turnover in special and general education, and used i n SASS and TFS data tables. The retention components of attrition and migration are differentiated. Attrition refers to leaving teaching employment. In this study attrition will refer only to leaving teaching in the district. School Attrition means to leave a particular school.
27 Teaching Area Attritionrefers to teachers changing fields (e.g., leaving a teaching assignment in special education for some other teaching assignment). Leavers refers to teachers who leave teaching employment. School Migratio n indicates that a teacher has moved from one public school to a different school. These teachers are referred to as Movers and teachers who stay in the same school are referred to as Stayers Switchers refers to teachers who transfer from one teaching area to a different area (such as from special education to elementary education) Switching is distinguished from other forms of transfer such as migrating to a different school. Teaching area transfer can coincide with school migration. Traditional Teacher Preparationrefers to four and fiveyear, state approved, teacher education degree programs with a supervised internship component leading to eligibility for a standard teaching credential. Alternative Teacher Certificationrefers to state approved programs that allow individuals with the minimum of a bachelors degree in a field other than education to become certified through competency based programs that may be provided through school districts, or college of education programs that provide streamli ned post baccalaureate certification programs. Significance of the Study There is general consensus among researchers that teacher turnover and the shortage of qualified teachers in highneed subject areas and schools are serious issues. Continuous turnover and the practice of hiring under certified teachers in hard to staff schools impacts the ability of schools to meet the needs of their students, particularly the vulnerable population of special needs students as well as the staff and community. Considering the chronic shortages of special education teachers and the number of alternatively certified people teaching special education, it is important to take an indepth look at the individuals who choose to enter the field in order to address the questions of who stays and who leaves. Additional study of teacher turnover and retention, as a function of teacher characteristics, qualifications, and teacher training is needed. This study provides a
28 district level perspective that can be considered along with the current research based on national aggregated data. Examining special education teacher characteristics, career paths, and retention over time may help to inform the hiring process and placement guidelines for school organizations. These data may also provide useful information for teacher preparation program designers both traditional and alternative. Limitations This study was conducted in a large urban school district in the southeast that has a state approved alternative certification program. The sampling frame limits the generalization of findings. The school district maintains a detailed, computerized personnel tracking system that was used to gather the data necessary to study the special education teachers. However, the tracking system provides limited information about teachers who leave the district. The sample, special education teachers hired within the 200304, 200405, and 200506 school years, included teachers with years of experience who were returning to the classroom. Careful consideration of how the more experienced teachers data influenced the retention findings is important to overall conclusions. The personnel tracking system provided data specific to the teacher, and the schools in which they were assigned. However, the individu al classroom data is limited to the job code, which indicated the exceptionality assignment. The data set does not provide details such as number of students or the possible range of students disabilities for which the teacher was responsible all important factors for satisfactory working conditions that impact retention. It has long been recognized that mentorship and administrative support in particular, are critically important to the retention of new teachers (Boe, et al., 2008a;
29 Johnson, 2004). This study did not measure the presence or quality of mentorship and administrative support that the special education teachers had access to as beginning teachers.
30 CHAPTER 2 REVIEW OF LITERATURE The purpose of this chapter is to provide an overview of the r ecent literature related to the supply and retention of teachers. The shortage of highneed subject area teachers and the differing perspectives on the causes of these shortages are discussed with a focus on special education. The importance of preparing and retaining qualified teachers is reflected in the teacher quality research, which underscores the equity issues surrounding the impact of teacher turnover for highneed students and achievement. Additionally, organizational behavior and management theori es regarding job choice and organizational fit are discussed as additional lenses through which school districts may examine their hiring practices. This chapter provides a framework for understanding the relationships between the qualities and characteris tics of schools and teachers, which can serve as predictors for retention. Federal and State Policy The impact of federal and state mandates on the teacher labor market has both intended and unintended consequences. The Individuals with Disabilities Educ ation Improvement Act (IDEA) is a case in point. The IDEA was enacted to ensure that special needs students have access to necessary services. However, one outcome of this mandate has been an increased demand for qualified special education teachers. The f ield has been trying to catch up with this demand for over two decades (McLeskey, et al., 2004). The IDEA mirrors requirements in the No Child Left Behind Act (NCLB) reauthorized in 2001, which mandates accountability by requiring academic achievement for all students. In addition, the law requires that all teachers, including special education teachers, be highly qualified for the core academic subject areas they
31 teach. In order to be designated as highly qualified, teachers must have a minimum of a bachel ors degree, full state certification, and also demonstrate subject matter knowledge and teaching skill in each core academic subject they teach ( H.R. Rep. No. 108779, 2004; Thornton, Peltier, & Medina, 2007). The IDEA and NCLB legislation both commit to providing resources and requiring accountability measures for ensuring that the educational needs and rights of children with disabilities are met. However, NCLB is widely criticized as an unfunded mandate, and many states struggle to meet the laws requi rements. Despite these well intentioned policy initiatives, and in some settings because of them, chronic shortages of qualified special education teachers persist, and special needs children continue to be served by under qualified teachers (Boe & Cook, 2006; Katsiyannis, Zhang, & Conroy, 2003; McLeskey, et al., 2004). Mandates such as the constitutional amendments to limit class size, which many states have passed in an effort to increase the quality of education and student achievement in their schools also have unintended outcomes. The class size amendments such as those passed in California and Florida place additional burdens on school districts to find funding, recruit, and retain sufficient numbers of highly qualified teachers. These mandates have resulted in teacher shortages, which are reflected in the hardto staff subject areas such as mathematics, the sciences, and special education, and budget shortfalls in already cash strapped districts (Florida Tax Watch, 2002; Johnson, 2006).
32 Teacher Labor Market Teacher labor market statistics show a variation in levels of supply and demand across the geographical regions in the country. The American Association for Employment in Education (AAEE) 2006 report data show that of the 64 education fields sur veyed, 50% now report some or considerable shortage. Additionally, the shortage of qualified special education teachers is shown as greater than any other subject area throughout the country, particularly in the largest cities and highpoverty schools. A complex range of factors contributes to the shortage of qualified special education teachers. The population of children identified as having disabilities has grown faster than the general population of students. Advances in medical technology, increased awareness and earlier diagnosis, as well as expanded eligibility criteria have combined to increase the number of children eligible for services (Florida Legislature, 2003; U.S. Department of Education, 2005). Some states have seen tremendous growth in the special education population. For example, Florida experienced a 167% increase in enrollment between 198384 and 200203 as compared to the overall student population growth of 70%. There is variation among exceptionalities as well. Two of the fastest grow ing exceptionalities in Florida, autism, and other healthimpaired students, grew 129% and 425% respectively from 199798 and 200203 (Florida Legislature, 2003). Although there is variation in the need for teachers among the exceptionalities, as well as within and across states, the chronic shortage of highly qualified special education teachers is well documented. Whether these ongoing shortages are caused by excessive turnover or an insufficient supply of new teachers continues to be a subject of contr asting viewpoints (Boe, et al., 2008b; McLeskey, et al., 2004).
33 Teacher Shortage: Retention or Supply? The differing perspectives on the causes of the teacher shortage are a dilemma for policy makers and school district administrators. Researchers in recent studies have posited that the seemingly contradictory findings may result from a combination of factors: (a) variations in the definition of attrition, (b) the use of different data sets, and (c) the inclusion of different variables, for exampleprivate and public teachers, full and part time teachers. Some studies have focused on teachers leaving the field while others have considered a combination of leavers and teachers migrating to other schools, districts, or fields. If research is to inform practic e, consumers must weigh the findings carefully. (Boe, et al., 2008a; Cook & Boe, 2007; Harris & Adams, 2007; Ingersoll, 2001, 2003). Teacher turnover has been referred to as a revolving door. The negative effects that the lack of stability in the teaching force can have on student achievement, particularly in difficult to staff schools are an ongoing concern. Teacher turnover, not the lack of prepared teachers, has been implicated as the primary cause for the shortage of qualified teachers in the nations classrooms (Ingersoll, 2001, 2003). Based on an analysis of the 199091 Schools and Staffing Survey (SASS) and its supplement the 199192 Teacher Follow up Survey (TFS), I ngersoll (2001) reported a 14% turnover rate for teachers. He compared this rate wit h an 11% turnover rate for all occupations published by the Bureau of National Affairs and a 12% turnover rate reported for nurses in hospitals in the mid1990s in order to demonstrate that the teaching profession has a higher turnover rate than other prof essions. It is important to note however, that Ingersolls definition of turnover includes teacher migration, and that a comparison of turnover rates from different data sources and time periods impacts the interpretation of
34 his findings (Boe, et al., 2008b; Guarino, Santibanez, & Daley, 2006; Harris, & Adams, 2007). Nonetheless, whether teachers leave or migrate, the impact to the schools that lose these teachers is the same (Johnson, Berg, & Donaldson, 2005). In an analysis of the research literature on s pecial education teacher retention and attrition, Billingsley (2004) identified attrition as an important factor in the shortage of special education teachers. The attrition rate of special education teachers has been reported as higher than that of general education teachers, and when combined with the number of special education teachers who transfer to the field of general education, the total attrition rate for special education is nearly double that of general education teachers (McLeskey, et al., 2004). Interestingly, recent research analyzing SASS and TFS data over a tenyear period from 1990 through 2001 indicates that the attrition rate of general education and special education teachers has been about the same (Boe, et al., 2008b). It is important to note that aggregated national data provides one perspective and does not reflect the considerable variation in turnover that exists at the state and local level (Guarino, et al., 2006; Villegas & Clewell, 1998). How attrition is defined is important wi th regard to how studies are interpreted and compared. Billingsley (2004) provides a four category definition for special education teacher retention, transfer, and attrition. The first category, retention, refers to teachers remaining in the same school as the previous year. The second category, transfers to another special education position, includes those who stay in special education teaching but transfer to another position whether in the same or different district. The third category, transfers to general education teaching, pertains to teachers who leave
35 special education, but remain in teaching, and the fourth category, exit attrition includes those who leave teaching altogether. Why Teachers Leave Work environment conditions, personal characteris tics, and qualifications each play integral roles in teacher retention. Special education teachers in particular have challenges that are specific to their field. Although certification status, perceived stress, school climate, and age are all identified as strong predictors for special educator retention and attrition (Miller, Brownell, & Smith, 1999), job design also has relevance. The concept of job design provides a framework within which to study the nature of a special educators job. Does the job, w ith all that it entails, make sense? Is it feasible? Is it one that well trained, interested, special education professionals can manage in order to accomplish their major objectiveenhancing students academic, social and vocational competence? (Gersten, Keating, Yavanoff, & Harniss, 2001 p. 551). Poor job design affects teachers negatively, resulting in stress, decreased job satisfaction, and a weakened sense of efficacy (Billingsley, 2004). Working Conditions Working conditions are very important to teachers and can affect their decision to stay, move to another school, or leave teaching altogether (Billingsley, 2003; Ingersoll, 2003; Johnson & Birkeland, 2003; Johnson, 2006). Working conditions are particularly important to special education teacher ret ention efforts. School climate and collegial relationships have been shown to be more important to the special educators than to the general educators, job responsibilities and the extent to which their schools are caring and supportive of students and st affaffect teachers confidence and intent to stay in the profession (Carlson, Brauen, Klein, Schroll, & Willig, 2002, p. 1).
36 Workplace conditions such as reasonable teaching assignments, collaborative colleagues, meaningful professional development, and safe facilities have been identified as important factors for both recruitment and retention (DeAngelis & Presley, 2007; Gersten, et al., 2001; Johnson, 2006). Administrative support has come to the forefront as one of the more important factors in teacher retention and sense of efficacy. Although the importance of mentoring and collegial support for new teachers has long been recognized (Billingsley, 2004; Miller, et al., 1999), recent research has highlighted support from the principal or building admini strator as critically important (Boe, et al., 2008b). Individual Characteristics Considerable retention research has focused on the individual characteristics of teachers in an effort to understand why people stay or leave the classroom. Characteristics s uch as age, gender, race, academic preparation, and teaching field have been identified as predictors for attrition. Age has been shown to have a ushaped relationship with leaving the classroom, indicating greater attrition rates within the first 3 to 5 y ears and again after 20 years, coinciding with child rearing years and retirement (Harris & Adams, 2007; Imazeki, 2005; Strunk & Robinson, 2006) It is important to note that many of these teachers contribute to the reserve pool of experienced teachers who later return to the field. This reserve pool is a rich source for recruitment and provides a substantial percentage of experienced teachers (Cook & Boe, 2007). Age is the only demographic variable that is consistently linked to attrition in the special e ducation literature. Special education teachers show the same pattern of early attrition as their general educator counterparts younger teachers are more likely to
37 leave than older teachers. Younger special educators are also more likely to transfer than older special education teachers (Billingsley, 2004). The age at which a teacher begins teaching appears to impact attrition differently for men and women. Imazeki (2005) found that men who begin teaching when they are older are more likely to leave than w omen. Conversely, women who begin teaching after 30 years of age have a lower exit attrition rate than women who begin teaching prior to the age of 30. The relationship between teacher characteristics such as age and gender and retention has implications f or programs that specifically recruit older, second career individuals. These alternative certification route programs can help high need districts that typically have difficulty staffing schools. However, it is important to consider that some of these teachers will have high opportunity costs, and may be more likely to leave for higher paying nonteaching positions (Dai, et al., 2007; Lui & Johnson, 2006). Women with advanced degrees and those hired to teach in large districts have also shown a greater tendency to leave (Harris & Adams, 2007; Imazeki, 2005). School Characteristics The characteristics of a schools student population and socioeconomic status are directly related to its ability to recruit and retain teachers. Retention varies according to sc hool poverty levels and the percentage of minority student enrollment (Elfers, Plecki, & Knapp, 2006; Hanushek, et al., 2004; Imazeki, 2005; Scafidi, Sjoquist, & Stinebrickner, 2007; Strunk & Robinson, 2006). In a study of school characteristics and teacher turnover in Georgia school districts, Scafidi, et al. (2007) found that teachers who begin teaching in schools with low student achievement, high poverty rates, or in schools with a high percentage of minority students, are more likely to leave or move t o other schools within the district. Similar to Hanushek, et als., (2004) study in Texas
38 school districts, the researchers found strong evidence that the Georgia schools with large percentages of Black students have much higher attrition rates than other schools. However, when the teacher is Black he or she is more likely to stay at schools with a large percentage of Black students. Similar findings are reported regarding Hispanic teacher retention in schools with higher percentages of Hispanic students. Research on teacher salary and mobility in Wisconsin school districts indicate that turnover rates are higher in urban and rural districts that serve the most low income students. The teachers race is also identified as a factor in retention. Minority teachers demonstrate a higher attrition rate overall. Although earlier studies demonstrated that Black educators were shown to have a 50% higher probability of returning to teaching after a spell out of teaching than their White counterparts (Singer, 1993, Wi llett & Singer, 1995). Minority teacher retention rates also show an increase when they are assigned to schools with larger proportions of nonwhite students, while White male teachers demonstrate a greater likelihood of leaving when there is a greater proportion of nonwhite students (Imazeki, 2005). Loeb, DarlingHammond, & Luczak (2005) examined predictors of high rates of turnover in California schools: We find that the racial, ethnic, poverty, and language composition of a schools student body influences a schools turnover, along with its difficulty filling vacancies and proportions of beginning teachers. However, we also find that working conditions add substantial predictive power to models of turnover and that, when these working conditions are added, the influence of student demographics on reported turnover and hiring problems is reduced. (p. 65) A study of North Carolina teachers documented the tendency for more qualified and experienced teachers to seek positions in schools that serve more aff luent and higher performing student populations, which presumably have more favorable working
39 conditions. The study did not focus on retention, however it lends credence to the idea that working conditions and student characteristics play an important role in teacher turnover. Retention is an important issue to address if school districts wish to ensure that disadvantaged populations have access to qualified, experienced teachers (Clotfelter, Ladd, & Vigdor, 2006). Teacher Supply In contrast to views ident ifying turnover as the root of the teacher shortage problem, other researchers have found that the attrition rates in education, both general and special education, are relatively stable and no higher than comparable professions. Although there is agreement that any attempts to improve retention will serve to benefit the profession as a whole; recent studies demonstrate that a supply side effort is needed to ensure that the ongoing need for teachers is met (Boe, et al., 2008b; Harris & Adams, 2007). An anal ysis of trends in the rate of general education and special education teacher attrition, transfer, and migration from 199192 through 200001 indicates that turnover increased by more than a third during the nine year period. Clearly, teacher turnover has been high nationally: 2223% of public SETs and GETs either left teaching, switched teaching area, or migrated to a different school annually during the 1990s (p. 23). While these numbers are high, it is important to note that just 30% of this turnover r esulted from attrition. Transfer and migration among teachers accounted for the majority of the turnover. Regardless, the impact at the school level is the same whether a teacher leaves or transfers. These teachers do however, fill needs in other schools, and are not lost to the profession (Boe, et al, 2008b).
40 Contrary to studies claiming that high turnover, not lack of prepared teachers is the cause of shortages, analysis of national datasets show that attrition among public school teachers is no higher t han other occupations and is not unreasonable to expect. Increasing the supply of highly qualified teachers is needed to address the teacher shortages particularly in the area of special education (Boe, et al., 2007, 2008b; Harris & Adams, 2007). The supply of highly qualified special education teachers is a longstanding problem. In the 19992000 school year, more than 12,000 positions went unfilled or were filled with substitute teachers because of a shortage of qualified applicants. In addition, 33,262 special education teachers were not fully certified for their main teaching assignment, requiring administrators to request 5,369 class size waivers and 5,480 caseload waivers (Carlson, et al., 2002). It is evident that an insufficient number of new speci al education teachers are prepared annually. Nearly twice as many elementary teachers are prepared for each available position as special education teachers (McLeskey, et al, 2004). During the 19992000 school year, 18% of the first time special education teachers hired had been prepared for general education teaching (Boe, et al., 2008b). A review of the degrees awarded to special education majors in 2005 06 shows a total of 7,977 bachelors and 15,586 masters degrees. In elementary education 44, 374 bachelors, and 19,771 masters degrees were conferred (U. S. Department of Education, 2007). Typically, the majority of individuals earning masters degrees in special education are already employed as teachers. Therefore, these numbers do not reflect additional supply. In light of the fact that many individuals who complete education degrees either delay
41 entrance into teaching or do not enter the teaching field at all, administrators must look to other sources to staff their classrooms (Cook & Boe, 2007). The majority of the teaching force, about 92%, will continue in their current positions, or move to another teaching position. The vacancies created by teachers who leave must be filled with individuals who are reentering the profession from the pool of peopl e who either deferred entry or left for other reasons, or newly prepared teachers (Boe, et al., 2008b). Former teachers are most likely to return if they do so soon after leaving. The longer a teacher stays out of teaching, the less likely they are to ret urn. Singer (1993) indicates that up to 34% of the special education teachers who leave teaching return within years. Alternative Certification Routes Alternative routes to teacher certification have become an integral part of the teacher preparation landscape by providing an additional source of newly prepared teachers (Humphrey, et al., 2008; Johnson, Birkeland, et al., 2005) On average, one in five of the countrys new teachers is certified through alternative routes rather than pursuing traditional ed ucation degrees. The National Center for Education Information (NCEI) survey data indicate that all 50 states and the District of Columbia report having some type of alternative route program for teacher certification (Feistritzer, 2008). Alternate route t eachers are teaching subjects that are in greatest demand for teachers. Onefifth (20 percent) of teachers coming through alternate routes, compared to 6 percent of all public school teachers, teach mathematics. Fourteen percent of alternate route teachers compared to 8 percent of all teachers, teach general special education. The pattern continues for all high demand subject areas. (Feistritser, 2005, p. 13)
42 Alternate route programs vary in multiple ways, such as who administers them, program requirements program components, as well as length and cost, among others. These variations contribute to inconsistency in program definition regarding what is and what is not an alternative route and impact the ability of researchers to accurately study, compare and report findings on outcomes (Feistrizer, 2005; Humphrey, et al., 2008; Rosenberg & Sindelar, 2005; Wilson, et al., 2001; Zeichner & Schulte, 2001). Similar issues affect research on traditional programs. A great deal of variability can also be found among traditional teacher preparation programs in schools and colleges of education (Levine, 2006; Zeichner & Conklin, 2005). Clear and consistent descriptions of teacher preparation programs are required if research is to inform policy and practice. The defini tion for alternative certification adopted by the American Education Research Association Panel on Research and Teacher Education (2005) states that teacher education programs that enroll noncertified individuals with at least a bachelors degree and offer shortcuts, special assistance, or unique curricula, which lead to eligibility for a standard teaching credential, are classified as alternative certification (CochranSmith, & Zeichner, 2005). The AERA definition reflects the broad nature of the types of programs available. The National Center for Education Information (NCEI), which has provided data on alternative routes for teacher certification since the first programs emerged in the 1980s, gives a more detailed description. The center developed a clas sification system in 1990 to categorize the differing routes used in each state for certifying teachers (Feistritser, 2005).The NCEI classification system, which is listed below, provides a common framework that can be used to differentiate alternative cer tification programs.
43 CLASS A: is the category reserved for those routes that meet the following criteria: The alternative teacher certification route has been designed for the explicit purpose of attracting talented individuals who already have at least a bachelors degree in a field other than education into elementary and secondary school teaching. The alternate route is not restricted to shortages, secondary grade levels or subject areas. These alternative teacher certification routes involve teaching with a trained mentor, and any formal instruction that deals with the theory and practice of teaching during the school year, and sometimes in the summer before and/or after. CLASS B: Teacher certification routes that have been designed specifically to b ring talented individuals who already have at least a bachelors degree into teaching. These routes involve specially designed mentoring and some formal instruction. However, these routes either restrict the route to shortages and/or secondary grade levels and/or subject areas. CLASS C: These routes entail review of academic and professional background, and transcript analysis of the candidate. They involve specially (individually) designed in service and coursetaking necessary to reach competencies requi red for certification, if applicable. The state and/or local school district have major responsibility for program design. CLASS D: These routes entail review of academic and professional background, and transcript analysis. They involve specially (indivi dually) designed inservice and coursetaking necessary to reach competencies required for certification, if applicable. An institution of higher education has major responsibility for program design. The following classes are for exceptions and rarely us ed.
44 CLASS F: These programs are basically emergency routes. The prospective teacher is issued some type of emergency certificate or waiver, which allows the individual to teach, usually without any onsite support or supervision, while taking the traditio nal teacher education courses requisite for full certification. CLASS G: Programs in this class are for persons who have few requirements left to fulfill before becoming certified through the traditional approved college teacher education program route, e. g., persons certified in one state moving to another; or persons certified in one endorsement area seeking to become certified in another. CLASS H: This class includes those routes that enable a person who has some "special" qualifications, such as a wellknown author or Nobel Prize winner, to teach certain subjects. CLASS I: These states reported that they were not implementing alternatives to the approved college teacher education program route for licensing teachers. CLASS J: These programs are desi gned to eliminate emergency routes. They prepare individuals who do not meet basic requirements to become qualified to enter an alternate route or a traditional route for teacher licensing. CLASS E: These post baccalaureate programs are based at an instit ution of higher education. As studies on alternate routes to teacher certification continue, researchers are finding that there is as much variation within the individual programs and the way each participant experiences it, as there is between the indiv idual programs ( Boyd, Grossman, Hammerness, et al., 2007; Humphrey, et al., 2008; Johnson, Birkeland, et al., 2005). The differences in teacher preparation programs, in addition to the lack of
45 clear and consistent program definitions and descriptions, com bine to underscore the counterproductive practice of comparing programs to determine which is the best (CochranSmith & Fries, 2005, chap. 2; Humphrey, et al., 2008; Zeichner & Schulte, 2001). Researchers emphasize the need for more indepth study of the multiple pathways that provide teacher training and certification to address the question: What are the relative contributions of the various components of these multiple pathways into teaching? More comprehensive data about program content and how parti cipants learn is needed in order to determine how these multiple pathways tailor programs to address the needs of a diverse set of participants while providing highquality teacher preparation with the goal of positively impacting student achievement (Wils on, et al., 2001; Zeichner & Schulte, 2001). Characteristics of Effective Alternative Certification Routes As a result of the increasing role that alternative certification routes play in the preparation of the nations new teachers, understanding what c onstitutes an effective program has become increasingly important (Boyd, Grossman, Hammerness, et al., 2007; Humphrey, et al., 2008). There are several recurrent themes in the literature that impact program quality including: (a) selection criteria, (b) pl acement criteria, (c) teacher training, (d) mentor support, and (e) induction. Rigorous selection criteria and high standards for choosing participants are critical to the process. Examples of screening tools used in some programs are strong subject area m astery, minimum grade point average, test scores, and college selectivity (Peske, 2003; Wilson, et al., 2001). In addition, previous classroom experience such as substitute teaching or work as a paraprofessional is considered an asset. It is also
46 important to select placements strategically. Careful placement of participants in schools with strong leadership, a culture of support, and adequate resources is identified as one of the most important components of an effective alternative teacher certification program (Humprey, et al., 2008). There has been considerable debate regarding the efficacy of pedagogical training. Some researchers have identified a positive relationship between teacher training, certification, and student achievement (CochranSmith & Ze ichner, 2005; Darling Hammond, et al., 2005; Wilson, et al., 2001). Other researchers have conducted studies that do not find a significant difference between traditionally certified new teachers and new teachers entering the classroom through an alternati ve route (Decker, et al., 2004; Kane, et al., 2006). Successful teacher training programs, whether alternative or traditional, have been found to provide timely and relevant coursework that is adapted to the needs of the participants, including training in instruction, management, curriculum, and working with a diverse student population (Peske, 2003; Wilson, et al., 2001). Effective mentor support is an important component of teacher preparation and development. Strong teacher induction programs provide t ime and resources for new teachers and their mentors. Mentors are encouraged to plan with new teachers, to share curricula, and model lessons. Timely coaching and feedback in conjunction with frequent classroom observations provide essential support. These induction activities are particularly important for alternative teacher certification program participants (Humprey, et al., 2008; Johnson, Birkeland, et al., 2005).
47 Alternative Route Teachers in HighNeed Schools Teacher turnover varies considerably within and between school districts. Typically, urban school districts lose teachers at a higher rate than their suburban neighbors. In a study of New York City schools, approximately 44% of elementary and 55% of middle school teachers leave or migrate from th eir initial placement within two years. Many of these teachers tend to migrate to higher achieving schools with fewer minority students (Boyd, Grossman, Lankford, Loeb, & Wyckoff, 2007). In Illinois school districts about 44% of new teachers leave their initial school within their first two years, and 67% leave within five years. Roughly onethird of the teachers who leave during their first five years return to teachbut not in the most disadvantaged schools (DeAngelis & Presley, 2007). Typically, new and returning teachers prefer to teach in communities similar to the ones in which they live or grew up (Boyd, et al., 2005). As a result districts, primarily urban and rural, are turning to alternatively certified teachers who are willing to teach in highneed schools (Cook & Boe, 2007; Dai, et al., 2007; Feistritzer, 2005). Alternative route programs have also demonstrated the ability to attract a more diverse pool of prospective teachers in terms of age, gender, and ethnicity (Feistritzer, 2005, 2008; Zeichner & Schulte, 2001). Although many program participants continue to reflect the racial composition of their local labor market (Humphrey, 2008), these routes do provide an important contribution given the differences between the racial makeup of the student population and that of the teaching workforce. The demographic that comprises the majority of the newly hired teacher workforce is white females under the age of 30 (Sachs, 2004), and over 40% of all public schools have no minority teachers (Rosenburg & Sindelar, 2005). The differences in sociocultural identities between
48 teachers and students influence teacher retention and success in urban schools (Sachs, 2004). These disparities also impact minority students who would benefit from the opportunity to have teachers who understand their culture and can be seen as role models (Villegas & Clewell, 1998). Special Education Alternative Routes to Certification Alternative certification routes for special educators have provided an increasing number of teachers for hard to staff settings and exceptionalities such as emotional and behavioral disorders (Katsiyannis, et al., 2003; Tyler, et al., 2004). Data from the 2002 Study of Personnel Needs in Special Education (SPeNSE) indicate that approximately 7% of current special education teachers earned their certification through an alternative route, compared to 4.5% of their general education counterparts. About 12% of current special education teachers for students with emotional disturbance have achieved certification through an alternative route (Carlson, et al., 2002). Alternative certification routes have also been credited with bringing more culturally and linguistically diverse (CLD) teachers, a significantly underrepresented group, into the field of special e ducation. This is particularly beneficial in special education where there are large numbers of minority and ESL students (Feistritzer, 2005; McLeskey, et al, 2004; Tyler, et al., 2004). Although continued study is needed, research on alternative certifi cation routes for special education teachers has begun to provide evidence that well designed programs can and do produce effective teachers. The nations chronic shortage of highly qualified special education teachers has resulted in a myriad of alternati ve certification programs, which vary in content and implementation. Rosenberg & Sindelar (2005) identified program components such as (a) meaningful collaboration between
49 stakeholders, (b) adequate program length and rigor, and (c) highquality supervision as indicators of a well designed program. These well designed alternative route programs provide a more diverse pool of well prepared teacher candidates and are an important component of the supply effort (Sindelar, Daunic, & Rennells, 2004). Unfortunat ely, fast track programs whose purpose is to move individuals into the classroom quickly without adequate support also exist, contributing to the number of under qualified teachers who fill many classrooms (Rosenburg, et al., 2007). Equity Issues Teacher quality is considered to be one of the most important factors impacting student achievement in the classroom (Sanders & Rivers, 1996). Teacher quality contributes more strongly to student learning than class size or ethnic and socioeconomic status (Hanushek et al., 2002; Wright, et al., 1997). The strong relationship between teacher quality and student achievement brings equity issues to the forefront for hard to staff schools that frequently turn to under certified teachers to fill vacancies. Teacher Quali ty The federal government highlights teacher quality in the NCLB legislation, which mandates that all students will have access to highly qualified teachers. State policies also focus on teacher quality through mandates to improve teacher recruitment, educ ation, certification, and professional development (DarlingHammond, 1999). A number of states and school districts have begun using student achievement data to measure and reward teacher quality. Some districts have gone even further by implementing contr oversial teacher performance pay programs based on student learning gains (Eide, et al., 2004).
50 Although there is no clear consensus on a definition, the research literature identifies two general ways to conceptualize teacher quality teacher quality defi ned by teacher qualifications or inputs and teacher quality defined by student achievement or outcomes (CochranSmith & Zeichner, 2005; Goe, 2007; Eide, et al., 2004). Common a ttributes, which are frequently identified as indicators of teacher quality suc h as degree level, teacher preparationpedagogical versus content knowledge, licensure, years of experience, and academic proficiency, are not easily correlated with teacher effectiveness. However, student achievement can be measured and attributed to the teacher (Clotfelter, et al., 2006; Goe, 2007; Rivkin, Hanushek, & Kain, 2005). Reflecting two decades of research on the impact of school inputs on students achievement, Hanushek defined teacher quality quite simply, good teachers are ones who get large gains in student achievement for their classes; bad teachers are just the opposite (CochranSmith & Zeichner, 2005, p. 40). States ensure that teachers demonstrate a minimum level of competency by requiring licensure or certification. The majority of teachers meet these certification requirements through traditional teacher preparation programs, which require varying degrees of field experience, coursework in pedagogy and subject area, as well as passing the requisite licensure exams States also permit districts to employ nontraditionally licensed teachers who achieve licensure through alternative routes. These programs may be very similar to a traditional teacher preparation program with preservice field experiences and coursework, or a fast track progr am that is basically onthejob training with concurrent professional development. The ability to circumvent tradition preparation in order to enter the teaching field has resulted in debates about
51 the value of teacher education, and whether candidates wit h strong academic backgrounds might be at least as effective as teacher education graduates (DarlingHammond, et al., 2005; Eide, et al., 2004). Researchers conclude that there is not a strong enough research base from which to draw definitive conclusions about the relationship between teacher licensure and student learning or the value of pedagogical preparation for teachers. The quality and content of teacher training programs vary greatly, and the impact is not always apparent (Goldhaber & Anthony, 2003; Wilson, et al., 2001). There is general consensus that a degree major or the number of math courses a teacher has taken matters for mathematics teachers, particularly at the secondary teaching level (Goldhaber & Brewer, 2000). However, evidence is lacking to support a similar relationship in other subjects. Researchers have also found that advanced degrees do not appear to increase teachers skills or raise the quality of teaching ( Clotfelter et al., 2006; Harris & Sass, 2007; Rivkin, et al., 2005), an in teresting finding in light of district policies providing pay increases for advanced degrees. There is considerable evidence showing that a teachers ability increases with years of experience as measured by his or her contribution to student learning. Although Clotfelter et al., (2006) found that veteran teachers continue to gain skill in later stages of their careers; other researchers have found that there are diminished returns on experience after five years. Subsequent years of experience do not appear to contribute additional impact (Hanushek, Kain, OBrien, & Rivkin, 2005; Harris & Sass, 2007; Kane, et al., 2006; Rockoff, 2004).
52 Rockoffs (2004) study of elementary teachers and students found that teacher experience is a significant predictor of tes t scores for both reading subjects and math computation. Evidence for gains in math was weakest. The study indicated that two years of teaching experience positively impacts student math scores, but that subsequent years of experience may actually have a negative impact on math scores. Harris and Sass (2007) also had mixed results in their analysis of the effects of teacher experience. More experienced teachers produced greater student gains, but only in elementary and middle school reading. Goldhaber (2007) found that students assigned to a teacher who has one to two years of experience outperform students with novice teachers by 3% to 7% of a standard deviation, and students with teachers who have three to five years of experience tend to outperform those with one to two years of experience by an additional 2% of a standard deviation. However, little evidence of significant gains for students has been associated with teaching experience beyond five years (Eide, Goldhaber, & Brewer, 2004; Rivkin, et al., 2005 ). A teachers academic ability evidenced by teacher tests, has been identified as an indicator for teacher quality. Goldhaber (2007) found that North Carolina teachers who scored in the top quintiles on the Praxis II Curriculum test appear to be signific antly more effective at producing student achievement for different types of students. Teacher test scores have also been positively related to student achievement, particularly for math (Clotfelter, et al., 2006). The findings suggest that teacher test performance is indicative of teacher effectiveness. Impact of Turnover on Student Achievement There is a growing body of literature demonstrating that teachers sort very unequally across schools. The least experienced teachers and those with the poorest
53 a cademic records are often found in schools with the highest concentrations of low income, low performing and minority students (Clotfelter et al., 2006). Highneed students are taught by teachers with the weakest credentials, such as certification status and exam scores, SAT scores, ranking of undergraduate college, and perhaps most importantly, teaching experience (Boyd, Lankford, et al., 2007, p. 2). Teacher turnover plays a major role in the shortage of highly qualified classroom teachers, particular ly in difficult to staff schools (Hanushek, et al., 2004; Ingersoll, 2001, 2003). Studies also indicate that urban and rural low socioeconomic schools, which are typically difficult to staff, are disproportionately impacted by teacher turnover. These schools and students are subjected to a continuous stream of inexperienced and less qualified teachers, raising serious equity issues in light of teacher quality research (Boyd, et al., 2002, 2005; Darling Hammond, 1999; Villegas, & Clewell, 1998). When teacher s leave they take their knowledge of the students, their families, and the practices of the school with them. The loss of school specific, accumulated knowledge compromises the schools capacity and ability to do its work. The most at risk populations typi cally experience the greatest impact from teacher turnover and inexperienced teachers, which perpetuates a cycle of unequal access (Johnson, 2006). Hiring Practices Similar to other professions, the initial years in teaching are a time of higher attrition as individuals determine whether or not they have a good fit with their school organization. New teachers are assessing the fit of their new profession and acclimating to what for many is unfamiliar terraindifferent racial, ethnic, socioeconomic backgrounds, and geographical location. Administrators are also making retention decisions about new teachers as they assess whether the individual fits within the
54 culture of the organization (Schein, 2004). Attrition takes on added significance because of the ski ll and knowledge that new teachers must acquire within their first years of experience (Elfers, et al., 2006; Kristof, 1996; Lankford, Loeb, & Wyckoff, 2002). Wellplanned, supportive hiring practices can increase the likelihood of a good match between a new teacher who has individual skills, knowledge, and talents, and a school with its unique set of challenges and opportunities (Liu & Johnson, 2006). Supportive hiring practices are identified as (a) school based, to ensure exposure to the culture and needs of the school; (b) information rich, to provide multiple opportunities for gathering information from interviews and observations; and (c) having early hiring timelines, to allow ample time for the new teacher to prepare (Johnson, Kardos, Kauffman, Liu, and Donaldson, 2004). Research in organizational behavior and management provides support for the value of ensuring personorganization and personjob fit (Cable & Judge, 1996; Kristof, 1996). A poor match may impact a teachers sense of efficacy, job satisfaction, and ultimately retention. To the extent that a poor fit compromises a new teachers effectiveness on the job and therefore her sense of success, it may contribute to her leaving her school or exiting teaching altogether (Liu & Johnson, 2006, p. 325). Hiring is a two way process and the extent to which there are opportunities available for the interviewers and interviewees to learn enough about the other to make informed decisions has important consequences (Pounder & Merrill 2001). In a study of teacher hiring in four states, Liu and Johnson (2006) found that most teachers had limited interactions with school based personnel. Few schools provided or required
55 classroom observations or teaching demonstrations, each of which could yield valuable information pertaining to job fit (Breaugh & Starke, 2000). The timing of hiring decisions is a major obstacle for principals. Union contracts, budgeting processes, and student enrollment information, all delay hiring of new teachers until the summer months and even after the school year begins. Urban schools consistently miss out on qualified teachers willing to teach in challenging school settings as a result of the delays in hiring. Lengthy recruitment periods and delays in contacting candidates sends mixed messages and often results in the loss of the most attractive applicants (Levin & Quinn, 2003; Rynes, Bretz, & Gerhart, 1991). Job Choice Theory School district human resource departments and school administrators work hard to meet the challenge of staffing their classrooms with highly qualified teachers year after year. These staffing challenges result not only from a shortage of qualified teachers, but also because many new recruits leave within a short time of hire (Ingersoll, 2001). It is import ant that employers consider post hire outcomes such as retention rates, performance, and job satisfaction at the beginning of the hiring process as part of their recruitment objectives. The way employers recruit can influence post hire outcomes (Breaugh & Starke, 2000). The organizational behavior and management research literature provides employers with insight into how job choice, organization fit, and job fit theories can help inform their pre and post hiring activities. Job choice theory provides a conceptual framework within which to consider job selection behavior. Behling, Labovitz, and Gainer (1968) originally conceptualized job choice theory in their work with recruiting college graduates. The theory was later used in educational settings with teachers and
56 administrators (Young, Rinehart, & Place, 1989; Pounder, Merrill, 2001). Behling, et al. (1968) proposed three separate theories of job choice: objective theory, subjective theory, and critical contact theory. Objective theory posits that candidates seek to maximize their economic status and make job decisions based on economic factors such as pay, benefits, advancement prospects, and other factors, which are objective and measurable. In contrast, subjective theory states that job choice is determined based on the candidates perception as to whether the organization will meet his or her psychological needs. The choice of one school or district over another might therefore be influenced by the organizations climate and culture (Liu & Johnson, 2006; Pounder & Merrill, 2001; Schein, 2004). Critical contact theory proposes that candidates usually have limited knowledge and contact with the hiring organization. Job choice decisions are therefore made based on the initial contact with the recruiter. The recruiters knowledge and ability to communicate the specific job requirements and expectations influences the candidates job choice decision (Young, et al., 1989). Candidates prefer information that conveys a realistic view of the job and connect wit h representatives who are personable and credible (Breaugh & Starke, 2000). Realistic Job Previews Employers who are concerned with post hire outcomes such as job satisfaction, performance, and retention ensure that candidates have an accurate perception of the job responsibilities and expectations. Providing descriptions of both positive and negative aspects gives an information rich, realistic job preview, which tempers unrealistic or inflated candidate expectations. It is important that candidates also have
57 an accurate perception of their own abilities and aspirations (Breaugh & Starke, 2000; Johnson, et al, 2004; Popvich & Wanous, 1982). However well prepared and committed they may be, teachers have no assurance that they will succeed in the classroom because teaching, by its very nature, is unpredictable work (Johnson & Birkeland, 2003, p. 584). Realistic job previews in which candidates visit the job sitethe school, and have the opportunity to talk with individuals who currently hold the position i ncrease the probability of the applicants job expectations being met. Less informed applicants tend to have less job satisfaction and may be more likely to quit (Breaugh & Starke, 2000; Meglino, Ravlin, & DeNisi, 2000). Unfortunately, many teachers are of ten hired after a single interview and accept positions for which they are unprepared (Liu & Johnson, 2006). PersonOrganization Fit Theory Broadly defined, personorganization fit is the compatibility between a person and an organization. This compatibil ity occurs when, (a) at least one entity provides what the other needs, (b) they share similar fundamental characteristics, (c) both (Kristof, 1996, p 4). We know that teacher attrition varies a great deal based on the differences in teachers, their subj ect areas, and the schools where they are hired (Hanushek, et al., 2004; Ingersoll, 2003; Johnson, Berg, & Donaldson, 2005). People are attracted to organizations whose goals are similar to their own or will enable them to attain their personal goals. A better match between the prospective teacher and the teaching job has implications for job satisfaction, efficacy, and retention. Individuals are also attracted to an organizations culture, climate, and values. Values congruence is a significant form of fit Values are fundamental to organizational
58 culture and guide employee behavior. Principals who consider fit when hiring teachers increase the likelihood of maintaining a stable faculty, and fulfilling the mission of the school. An individual is more likely to stay if the environment meets his or her needs; conversely individuals with low levels of congruence with the needs, values and culture of the school are more likely to leave (Cable & Judge, 1996; Judge, Higgins, & Cable, 2000; Kristof, 1996; Schein, 2004). A good match or fit takes on additional importance when recruiting and retaining second career individuals who may have other career options available to them. These are important considerations when recruiting teachers for highneed schools and hav e implications for retention and ultimately student achievement (Dai, et al. 2007; Liu & Johnson, 2006). Summary The teacher labor market reflects a history of shortages in high need subject areas. In many districts principals have turned to other sources to fill theses vacancies, hiring under certified teachers and those who come to teaching through alternative certification routes. These practices are most evident in districts with high poverty urban and rural schools. A complex range of factors including legislative mandates, inadequate supply, and turnover has been identified as contributing to teacher shortages. Although there are differing perspectives as to whether teacher supply or retention is the underlying cause of teacher shortages, it is import ant to understand that the impact at the school level is the same. Consequently, holistic efforts targeting both supply and retention are needed to address the ongoing shortages. Working conditions, personal characteristics, and individual qualifications all play important roles in both recruitment
59 and retention efforts. Therefore, the availability of high quality traditional and alternative teacher preparation programs, information rich hiring practices, and retention program initiatives take on added importance. In addition to subject area assignments, personal and school characteristics can be predictors of teacher retention. School characteristics such as poverty level, the percentage of minority students, and geographic location influence both recruitm ent and retention of nonminority teachers. Alternative certification route programs have proven successful in recruiting a more diverse pool of prospective teachers who are willing to teach high need subject areas in high poverty schools, which typically have a greater population of minority students. Welldesigned alternative certification route programs that have rigorous selection criteria, strong training, and support can produce competent teachers. However, many alternative routes are fast track programs that move teachers into the classroom quickly with minimal supports. The use of fast track programs and the practice of hiring under certified teachers raise serious issues of equity for the students of these teachers. Despite the fact that teacher quality is one of the most important factors impacting student achievement in the classroom, we find that the least experienced and least qualified teachers are teaching in highneed settings. The organizational behavior and management literature regarding personorganization fit and job fit provides principals with insight about teachers who are more likely to stay, as well as those who are most likely to leave. Understanding the personal and school characteristics that can serve as predictors for teacher r etention can inform hiring decisions and post hiring outcomes. Principals who consider fit and preparation
60 when hiring teachers may increase the likelihood of maintaining a stable faculty and positively impacting student achievement.
61 CHAPTER 3 METHODOLOGY The purpose of this chapter is to discuss the research design used to study the retention of special education teachers in a large urban school district. The overarching goal of the study was to ascertain who stays, who leaves, who returns, and why. Devel oping a greater understanding of teachers who are more likely to be retained in special education assignments has important implications for recruitment, training, and retention. This chapter explains how the study was conducted. First the quantitative re search design is discussed, next the setting and context are described, the sample selection is explained, and finally data sources and variable definitions are provided. A discussion of the data analysis follows. The research questions guiding this study were: 1. Does the risk of leaving teaching differ with respect to a special education teachers personal characteristics? 2. Does the risk of leaving teaching differ with respect to a special education teachers level or type of teacher preparation? 3. Does t he risk of leaving teaching differ with respect to a special education teachers school assignment (percentage of poverty and minority student enrollment)? 4. Does the risk of leaving teaching differ with respect to a special educators teaching assignment ( students disability classification)? 5. How many years on average do newly hired special education teachers continue to teach in a large urban school district? 6. What is the probability of a special education teacher returning to teaching after leaving?
62 R esearch Design This quantitative study provided descriptive data on a sample of special educators hired during three academic years from July 1, 2003 through January 30, 2006. The demographic data on the teachers age, gender, and race were recorded. In a ddition, the level and type of teacher preparation for each educator was identified, and any previous teaching experience was noted. Data describing the school assignment demographics and teaching assignment details were also collected. The study tracked t he career history of the special education teachers for a period of six years ending June 30, 2009. The resulting data provided a rich source of information for analysis. Frequency tables were used to summarize the categorical data and provide descriptive statistics about the sample. The focus of this study centered on event occurrencewhether a teacher leaves or returns to teaching and, if so, why. Researchers have identified survival analysis or event history analysis as a useful method for investigatin g event occurrence and identifying event predictors (Allison, 1982; Singer & Willett, 2003; Willett & Singer, 1993). The survival analysis model requires three methodological features. First, the study must have a target event, one whose occurrence is recorded. Second, a starting time is required, and third, there must be a meaningful scale for measuring time when recording event occurrence. Event occurrence is defined as moving from one state to another state i.e., a teacher is either teaching or not teac hing. Survival analysis requires that states not overlap, and that an individual occupy only one state during the period of study. Once the event does occur the individual moves to another state, which in this study was identified as a spell (Singer & Will ett, 2003).
63 Time scales for event occurrence can be classified as either continuous or discrete. This distinction is important given that the methods applied to continuous time may not apply to discrete time. The use of continuous time as a measurement sc ale requires that the exact time of event occurrence is known or that the interval is small enough that it can be considered continuous. Discrete time can be accounted for in two ways. The event occurrence may happen at any point during the time interval or period being measured, or the event may occur at discrete points in time during the time interval. Graduation and student retention at the end of a school year are examples of situations when discrete time would be an appropriate time scale. Discrete tim e analysis is appropriate for studying time intervals in which many individuals will experience the target event in the same period of time (Allison, 1982; Mayson, 2003; Singer & Willett, 2003). Standard analytic methods are not well suited to dealing with time varying explanatory variables and missing or censored data, both of which are typical in the study of event history. The use of continuous time or discrete time analysis can address these problems and have proven to be useful tools for studying event history. Censoring occurs when an individual in the study does not experience the target event or is unaccounted for by the time the study ends. It is not possible to know when the individual will experience the target event, only that they have not experienced it by the end of the study. This is referred to as right censoring. Censored data contributes only partial data to the investigation when using standard statistical methods of analysis (Allison, 1982; Singer & Willett, 2003).
64 Attempts to account f or censoring have included leaving censored data out of the analysis, imputing data such as the average time to event, or using the ending date of the study instead of not counting the person. All of these methods may bias the findings. Survival analysis m ethods can deal with both known and censored event times thereby providing a more accurate analysis. Censored individuals do contribute important information to the study. The fact that they do not experience the event may provide insight about others who do (Singer & Willett, 2003). This study utilized discrete time survival analysis to examine special education teacher retention. The longitudinal data provided the opportunity to analyze events in the teachers career histories such as leaving or returning to teaching and changing school or teaching assignments. Advantages of the model are that it easily accommodates timeinvariant predictors such as age, gender, and race, as well as timevarying predictors such as teaching assignment or school setting. The effects of predictors may also vary over time (Singer & Willett, 2003; Willett & Singer, 1995). The event history of every individual in the risk set (at risk for experiencing the target event) was divided into spells. Each spell corresponded with an event occurrence. In this study each teachers career history was divided into spells that indicated when and if they left teaching, and when and if they returned to teaching. For example, entering the first spell indicated being hired to teach; entering a sec ond spell indicated leaving teaching; entering a third spell indicated a return to teaching, and so forth listing as many spells as necessary to record the repeated events for each teacher during the course of the six year study.
65 The discrete time metric was an academic year. The choice of an annual metric provided six periods or years to observe event occurrence. A teacher may or may not have experienced the target event within one or more periods (academic years). Only those teachers who left teaching or reentered teaching were shown as moving to the next spell. Teachers who were hired to teach during the threeyear period from July 1, 2003 through January 30, 2006, and who did not leave teaching at any time during the data collection period that ended June 30, 2009, were censored. This survival analysis used SAS programming (version 9.2, Cary, N.C.) to conduct the tests and analysis necessary to address the research questions. The SAS LIFETEST Procedure provided life tables and survival curves that were used to summarize and examine the event occurrence data. The SAS Product Limit Survival Estimates table provided data for the following: observed event time, estimate of the survivor function, estimate of the cumulative distribution function of the failure time, survival standard error, number of observed event times, and the number of event times that remain to be observed. The observed event time indicated which year the teacher left teaching or returned to teaching. The survivor function provides another way to illustrate event occurrence data by aggregating the periodby period risk of the event not occurring. This probability showed what percentage of teachers survived or were still teaching at the end of each successive time period, for example, by the end of year one, by the end of year two, and so on until the end of the sixth year of data collection. The cumulative distribution function of the failure time or hazard function identified the chronological
66 pattern of the probability of teachers in the risk set experiencing the event of interest. It tells us whether and when an event occurs. The lifetable provided the data necessary to determine the median lifetime or center of the distribution data. It showed the point at which approximately half of the sample had experienced the target event and half have not. The median lifetime statistic answers the question how long does the average teacher teach (Singer & Willett, 2003). Summarizing the event occurrence data provided information regarding how long the sample teachers stayed in teaching. The data also indicated whether teachers returned to teaching after a period of time. These data however, did not help answer why they returned, which is important information for developing retention and support programs. Determining why some teachers left teaching or returned to teaching required a statistical model that would identify the relationships between the event occurrence and predictor variables. After examining the survival distribution function graphs, l og rank and Wilcoxon tests were used to test for significant differences (Singer & Willett, 2003; Willett & Singer, 1993, 1995). Setting and Context This study was conducted in a large urban school district in the State of Florida. The district currently s erves approximately 125,000 students and employs over 8,000 teachers, of which approximately 900 are special education teachers. The districts challenges are typical of many other large urban school districts: high poverty, a persistent achievement gap between the minority and majority student populations, high failure rate, and a graduation rate that is less than the state average (Villegas & Clewell, 1998).
67 The districts student population is diverse. The 2007 demographic information posted on the Flor ida Department of Education website listed 56.4% minority, 15.3% exceptional student education, and 4.2% English as Second Languagemore than double the percent in 2004. Minority enrollment has also increased steadily over the past five years, both in total number and as a percentage of the total population. Sample The sample selected for study included all teachers hired for special education assignments during three consecutive academic years beginning July 1, 2003 and ending January 30, 2006. Collecti ng data within this timeframe ensured that teachers hired during the summer recruiting season, as well as those hired during the first semester of the school year, were included in the data set. Teachers are often hired after the start of the school year i n urban school districts (Levin & Quinn, 2003; Liu & Johnson, 2006). Career histories were documented for six academic years ending June 30, 2009. The sample included 603 special education teachers with a wide variety of demographic and education backgrounds. Frequency tables were developed to summarize and display descriptive statistics for the variables related to the teachers personal characteristics and career history data. It is important to note that the district has a threeyear transfer policy. Teachers may request transfer to another school after three years of service at the hiring school. Data Sources and Collection After requesting and obtaining permission to conduct this study from the University of Floridas Institutional Review Board and t he school districts Office of Research, Assessment, and Evaluation, the Office of Human Resource Services was contacted to obtain the necessary teacher data. A request for relevant demographic,
68 education, certification, and job assignment data on the teac hers hired for special education assignments from July 1, 2003 through January 30, 2006 was submitted. These data were provided without the teachers names. Therefore informed consent was not required. As this protocol did not involve the use of human par ticipants in research, it was exempt from further review. Statistics for school demographic data were obtained from the Florida Department of Education (FLDOE) website. The percentage of minority student enrollment for the districts schools was provided for each school year in the data collection period. The percentage of free and reduced lunch participation served as a proxy for poverty level in the schools. These data were also available on the FLDOE website for each school year examined. Definition of Variables The dependent variable in this study was teacher retention. In Billingsleys 2004 analysis of the literature on special education teacher retention and attrition, she noted that there are a variety of definitions used for both retention and attr ition in the literature. In this study Billingsleys (1993) four category definition of retention was used. In the first category, retention pertained to teachers who remained in the same teaching assignment and the same school as the previous year. The second category, transfers to another special education teaching position, included those who stayed in special education teaching but transferred to another position (in either the same or a different district). The third category, transfers to general education teaching, was of concern because this group reflected a loss to the special education teaching force. The fourth group, exit attrition, included those who left teaching altogether that is, retired, returned to school, stayed home with young children, or took nonteaching positions in education.
69 In order to understand more about the risk of special education teachers leaving teaching, the study examined relationships among the following variables which may serve as predictors: teacher demographics, level and type of teacher preparation, teaching assignment, and school assignment demographics. The event occurrence data identifying whether and when a teacher leaves or returns to teaching was recorded for the purpose of the discrete time survival analysis. The categories of retention, such as transfers were reflected in the teachers teaching assignment and school assignment data. School and Teaching Assignment Variables The poverty level of the school was determined based on the percentage of students eligible for free and reduced lunch. Federal requirements for Title One designation stipulates that at least 40% of the student body must be eligible for free and reduced lunch in order to qualify for federal funding. However, for the urban district i n this study, the cutoff percentage for eligibility was usually not lower than 60%. The district has a range of 60 to 70 public schools designated as Title One from year to year. The categories for the poverty level variable were defined to show low, average, high, and very high levels of poverty. School poverty level 0 = 81% 97% 1 = 49% 80% 2 = 26% 48% 3 = 5% 25% The districts student minority population is 59%, with Black students composing the largest proportion at 44%. The categories for t he minority level variable were defined to show low, average, high, and very high levels of minority student enrollment.
70 Percentage of minority students 0=70% 100% 1=50% 69% 2=30% 49% 3=14% 29% The teaching assignments in the district are identi fied by job code, which are used as proxies for describing the special education teaching assignment. The categories for this study were defined based on similar job responsibilities and whether the teaching assignment was in a special education self contained classroom or resource room setting. The disability classifications used in the districts job codes and the attendant settings for service delivery were current at the time of hire. (Inclusive service delivery in the district for students with disabil ities was limited during the years for which data were collected.) Special educators teaching assignments were coded into three categories. Teachers assigned to Category 0 taught students considered to have emotional and behavioral disorders (EBD), which were classified previously in Florida regulations as emotionally handicapped (EH) and severely emotionally disturbed (SED). Most students with these disability classifications were taught in self contained settings with teacher assistants. Teachers assigned to Category 1 taught students considered to have highlevel disability related needs. Students with intellectual and developmental disabilities were previously classified as mentally handicapped, which included subclassifications reflected in the job codes as educable (EMH), trainable (TMH), and profound (PMH). Most students with these disability classifications were taught in self contained settings with teacher assistants; some EMH students were mainstreamed for a portion of the day. Students with p hysical disabilities referred to in the job code as PH, were placed
71 for instruction according to the level of their support needs. The majority of the districts students with Autism were taught in self contained settings with teacher assistants, with the higher functioning students being mainstreamed for a portion of the day. Teachers assigned to Category 2 taught students with specific learning disabilities (SLD ) or students with varying exceptionalities (VE), which included students with EBD and SLD. These students who were taught in either self contained classrooms or in resource settings designed as pull out programs in which students left their general education class for a portion of the day to focus on individualized instruction in their designated area of need. The prekindergarten special education population was taught in self contained settings with a teacher assistant. Job Code 0 = EH, SED 1 = Autistic, EMH, TMH, PMH, PH 2 = SLD, VE, Resource, Pre KH 3 = General Ed 4 = District Resource/ Counselor Teacher Preparation Variable Teachers with degrees in a field of special education or general education typically receive more extensive preparation through field experience and practice teaching required by the traditional programs. These programs are defined as approved teacher education degree programs, which lead to eligibility for a standard teaching credential. Individuals entering education from other fields pursue teacher certification through alternative routes defined as preparation pathway s that provide streamlined programs, special assistance, or unique curricula leading to eligibility for a standard teaching credential.
72 Teacher Preparation 0 = Incomplete or no prior teacher preparation (degree major in field other than education/ hired under temporary certificate) 1 = Special education preparation (special education degree major or special education apprenticeship program requiring extensive field experience) 2 = General education teacher preparation (college of education degree major ot her than special education) Data Analysis The sample special education teachers demographic information, academic background, certification status, teaching assignment, school assignment, years of experience, and employment status for each of the six years studied were compiled. The level of teacher preparation for each teacher was determined based on their degree major. Teachers with special education degree majors were coded as 1 for having more extensive preparation, teachers with college of education degree majors other than special education were coded as 2 for high level of preparation. The sample data were compared to the districts Alternative Certification Program database to determine which teachers were enrolled in or had completed the distr ict program. These teachers had been issued temporary certificates to teach special education. The Florida Department of Education provides individuals who have the minimum of a bachelors degree (in any field) who pass the states licensing subject area exam for K 12 Exceptional Education, eligibility to receive temporary certification (FLDOE Certification). These special education teachers were coded 0 for minimal level of teacher preparation. Further comparison of the data with the districts Transition to Teaching program database identified the teachers who had completed a year long exceptional student education apprenticeship program. This certification program
73 required both coursework and onthejob training while assigned to work with a veteran special education teacher. These teachers were coded 1 for extensive teacher preparation. All other special education teachers with a degree major in a field other than education hired under a Florida temporary teaching certificate were coded as 0 indicat ing minimal preparation. The school assignment demographic data were collected from the FLDOE Florida Schools Indicator Reports. These annual reports identify the percentage of minority student enrollment and the percentage of students participating in a free and reduced school lunch program, which serves as a proxy for poverty level (FLDOE, 2009). PersonPeriod Data Set In order to summarize the special education teachers data and the event occurrence of interest (entering or leaving teaching), the long itudinal data set was transformed into a personperiod data set. The data set has separate records or rows for each sample teacher and every academic school year in which they were at risk of experiencing a target event. Data were collected for teachers hi red during the following timeframes: July 1, 2003 through January 30, 2004; July 2004 through January 30, 2005; and July 30, 2005 through January 30, 2006. The beginning date for each teacher was represented by the year of hire variable which indicated the first year or discrete time period that they were at risk of experiencing a target event. The data set included a spell variable and a period variable to specify the number of school years a teacher remained in or out of teaching. The event variable indic ated whether the event of interest had occurred, a 0 indicated no event occurred, and 1 indicated the event did occur either the teacher left or returned to teaching.
74 Table 31 is an example of the personperiod table and contains records for the firs t three teachers in the sample. The personperiod data set contains demographic information for each teacher, the level and type of teacher preparation, the school teaching assignment, which indicates exceptionality, and the school assignment characteristi cs. The spell describes the length of time in or out of teaching and coincides with the occurrence or event of interest. The period variable indicates the number of years within each spell. Table 31 shows that teacher 1 left after the first year in teaching and did not return. Teacher 2 spent two years in teaching interrupted by one year out of teaching. Teacher 3 taught 2 years, left teaching and then returned. Teacher 3 was still teaching at the end of data collection and was censored. Therefore, it is unknown whether and when the teacher left teaching. Once the personperiod table was constructed, SAS programming (version 9.2, Cary, N.C.) was used to analyze the data. Frequency tables were created to describe the data set and provide descriptive statist ics for each variable of interest. The SAS FREQ procedure was conducted to provide data on the frequency and percentage of special education teachers who experienced the event of interest specific to each variable. Univariate analysis was conducted using c hi square testing to identify variables that demonstrated a relationship with teacher retention and warranted further analysis. After identifying the significant variables, the next step was to conduct the survival analysis. The SAS LIFETEST Procedure was conducted using the variable of time to build a life table and provide survival estimates for the sample teachers and a baseline with
75 which to compare other variables of interest (Singer & Willett, 2003; Willett & Singer, 1995). The SAS Product Limit Survival Estimates table tracked the career history of the sample of special education teachers for the six year data collection period. Data were provided for the estimate of the survivor and hazard function, as well as the survival standard error. The number of observed event times and the number of event times that remain to be observed provided the data necessary to determine the median lifetime or center of the distribution data. It showed the point at which approximately half of the sample had experienced the target event and half had not. The median lifetime statistic answers the question how long does the average teacher teach (Singer & Willett, 2003). In order to understand more about why some teachers leave teaching and some stay, it was necessary to determine whether the probability of event occurrence differed systematically among the special education teachers. Did the risk of leaving vary according to age, race, or gender? Were special educators who completed traditional teacher training programs ret ained at a higher level than those who become certified through an alternate route? What were the variables that predicted a greater risk for leaving teaching? Statistical models were used to test these relationships. The SAS Survival Analysis provided the within group survival distribution function which plotted the survival data and displayed the effects of the predictor variable which assisted in the analysis and communication of findings. In order to determine whether the special education teachers ri sk differed significantly with regard to the predictor variables the LIFETEST Procedure was used to provide the Test for Equality over Strata (variable), which produced logrank and
76 Wilcoxon significance test results (Allison, 1982; Singer & Willett, 2003; Willett & Singer, 1995). These tests were conducted for each of the variables of interest providing the necessary data to address the studys research questions, as well as raise additional questions for further research. Summary This chapter explains the discrete time survival analysis research design used to study the retention of special education teachers in a large urban school district. The overarching goal of the study was to discover who stays, who leaves, who returns, and why. The rationale for using the survival analysis methodology and the conceptual application has been discussed in relation to the research questions. This chapter provides details regarding the sample, data sources, data collection, and an overview for data analysis. The next chapter will present details regarding the analysis and findings.
77 Table 31. Personperiod data set ID Teacher Preparation Race Gender Age School Poverty Level School Minority Level Assignment Years Experience Spell Period Event 1 1 1 1 2 2 2 0 5 1 1 1 2 1 1 1 4 1 0 2 5 1 1 1 2 1 1 1 4 3 3 2 5 2 1 1 2 1 1 1 4 3 3 2 5 3 1 1 3 0 1 1 2 2 2 2 5 1 1 0 3 0 1 1 2 2 1 2 5 1 2 1 3 0 1 1 2 3 2 2 5 2 1 1 3 0 1 1 2 2 2 2 5 3 1 0 3 0 1 1 2 2 2 2 5 3 2 0 Note. Example of the personperiod table containing records for the first three teachers in the sample. Teacher 1 left after the first year in teaching and did not return. Teacher 2 spent two years in teaching interrupted by one year out of teaching. Teacher 3 taught 2 years, left teaching and then returned. Teacher 3 was still teaching at the end of data collection.
78 CHAPTER 4 ANALYSIS AND RESULTS This chapter presents the results of an analysis of retention behavior for a sample of special education teachers over a period of six years. Data regarding the personal characteristics and the school teaching assignments of the sample special education teachers who stayed in teaching and those who left were examined in an effort to understand more about the teachers who are successfully retained in a large urban school district. A description of the sample is provided and descriptive statistics are given for the variables of interest. The research questions are addressed throughout the discussion of the study results. Description of Sample The sample selected for this study included all teachers hired for special education assignments in a large urban school district over the course of three consecutive academic years, beginning with the 200304 academic year and ending with the 200506 academic year. The teachers career histories were documented for six academic years with an ending date of June 30, 2009. The sample captured a total of 603 special education teachers and represents individuals with a variety of education backgrounds and demographic representations. Personal and career history data were documented for each teacher. The personal characteristics examined were race, gender, and age. Any prior teaching experience was also recorded. Data reflecting each teachers level of teaching preparation were collected according to three categories (a) no prior training, (b) having a degree major in special education, or (c) having a degree major in general education.
79 The data reveal a sample that is mostly female (77%) with a racial demographic that is almost evenly split between white teachers (53%) and minority teachers (41% Black, 6% Hispanic, Asian or American Indian). The ages of the teachers ranged from 25 to over 56. Although the highest percentage of the teachers hired fell in the 2535 age range (35%), the majority were 36 and older. The data on prior teaching experience showed that 67% of the teachers had less than three years teaching experience and 49% were teaching for the first time. The remaining teachers captured in the sample had previous teaching experience and were returning to teaching after an unspecified amount of time out of teaching. The data show that 54% of the special education teachers hired held degrees in subject areas other than education and had no teacher preparation. There were 34% of the teachers who had more extensive special education preparation and 12% who had more extensive preparation in general education. Table 41 provides a breakdown of these invariant teacher characteristics. An examination of the data with regard to the l evel of teacher preparation provided additional details about the characteristics of the sample teachers. Male teachers comprised 12% of the teachers trained in special education, 24% of those prepared for general education, and 30% of the teachers who ent ered teaching with no preparation. Higher numbers of minority teachers entered teaching from fields other than education. And over half (56%) of the first year teachers entered with no prior preparation. The data show very similar initial school and teachi ng assignments for teachers regardless of the level or type of preparation. Table 42 provides descriptive data on the level and
80 type of teacher preparation, teacher personal characteristics, school assignments, and teaching assignment characteristics. T he school assignment characteristics for the sample teachers were documented. The poverty and minority composition of the student population were also recorded each year for each teacher. These data showed that 60% of the teachers were hired to teach in sc hools with the highest levels of poverty where 49% to 97% of the student population received free and reduced breakfast and lunch. Schools with a high minority population also tend to have high levels of poverty. The data show that of the sample teachers hired, 72% were assigned to the schools where 50% to 100% of the student population was minority. Table 43 provides data on school assignment. An examination of the data showed that 51% of the Black teachers were hired in schools with 70% 100% minority population, whereas 28% of the White teachers were hired in these schools. Table 44 provides a breakdown based on teacher race and school minority level. Data were also collected on the teaching assignments, as identified by the disability classification of the students each teacher was assigned to teach. Seventy one percent of the teachers were hired for teaching assignments in Category 2, which included working with students with specific learning disabilities, varying exceptionalities, and also prekindergarten students with disabilities (see job code and category descriptions in Chapter 3). Twenty two percent of the teachers were hired for teaching assignments in Category 1, which included students with Autism and students with intellectual and developmental disabilities previously classified as educable, trainable, and profoundly handicapped. The smallest number (7%) of the teachers was
81 hired for teaching assignments in Category 0, working with the population of students previously classified as emotionally handicapped and severely emotionally disturbed. Research Questions The following research questions were addressed in this study: 1. Does the risk of leaving teaching differ with respect to a special education teachers personal characteristics? 2. Does the risk of leaving teaching differ with respect to a special education teachers level or type of teacher preparation? 3. Does the risk of leaving teaching differ with respect to a special education teachers school assignment (demographic composition and academic performance indicators)? 4. Does the risk of leaving teaching differ with respect to a special educators teaching assignment (students disability classification)? 5. How many years on average do newly hired special education teachers continue to teach in a large urban school district? 6. What is the probability of a special education teacher returning to teaching after leaving? Predictor Variables The variables identified for this study represented both invariant characteristics specific to each teacher and variables that were specific to each school and job assignment, which could vary from year to year. Each variable was examined in order to determine whether it could serve as a predictor for teachers who may be at risk for leaving teaching. Chi square statistical analysis was used to identify the variables that demonstrated a relationship with teacher retention behavior whether and when the teachers left or returned to teaching.
82 Teacher Characteristics The following research questions regarding teacher characteristics: (a) Does the risk of leaving teaching differ with respect to a special education teachers personal characteristics, and (b) Does the risk of leaving teaching differ with respect to a special education teachers level or type of preparation were addressed using SAS programming (version 9.2, Cary, N.C.). The SAS FREQ procedure was conducted to provide data on the frequency and percentage of all sample teachers who stayed and those who left teaching specific to each variable of int erest. A univariate analysis was then conducted for each predictor variable to determine whether it should be included in further analysis. Chi square statistics were run to determine whether the probability of special education teachers leaving or returni ng to teaching differed with respect to the teachers personal characteristics or level of teacher preparation. Research question 1: Does the risk of leaving teaching differ with respect to a special education teachers personal characteristics? The variables of gender, race, age, and prior teaching experience were examined. The chi square statistic showed that the probability of a teacher leaving teaching did not differ with respect to teacher gender, X2 (1, n = 603) = 0.57, p = 0.45; or race, X2 (4, n = 603) = 1.94, p = 0.75. These variables were not considered for further analysis. Table 45 provides statistics for gender and race. However, the chi square statistic indicated that the age variable did impact the risk of the special education teacher leav ing teaching, X2 (3, n = 603) = 10.44, p = 0.02. The youngest teachers, age 25 35, left at the highest rate (70%) after their first year of teaching. The rate decreased with age until teachers reached the age range of 56+, at
83 which point the rate increas ed to 65% leaving. Table 46 provides data on the percentage of teachers who stayed and who left teaching in each age range category. The personal characteristic of prior teaching experience also had a statistically significant relationship with the teachers risk for leaving, X2 (6, n = 603) = 19.10, p = 0.00, indicating that early career teachers leave teaching at a higher rate. The percentage of teachers leaving teaching spiked again for teachers with 6 to10 years of prior teaching experience upon hire. Table 47 provides information on the teaching experience categories and the percentage of teachers who stayed and those who left teaching during the collection period. Teacher Preparation The variable of teacher preparation was examined. Three categories were identified (a) incomplete or no prior teacher preparation, (b) more extensive preparation in special education teaching, and (c) more extensive preparation in general education teaching. The special education teachers, who had no prior teaching preparation, were hired based on successful completion of required subject area examinations and allowed to complete certification requirements as they taught. Teachers with more extensive preparation held education degree majors in either special education or another education field. Research question 2: Does the risk of leaving teaching differ with respect to a special education teachers level or type of teacher preparation? The level or type of teacher preparation was not shown to impact the probability t hat the teacher would leave teaching. The chi statistic X2 (2, n = 603) = 1.66, p = 0.44, indicated no significant difference in teachers leaving or staying in teaching relative to their level of teacher preparation. Table 48 provides a breakdown of teacher
84 preparation level and percentages for teachers staying and those leaving after their first year of hire. Over 50% of the sample special education teachers hired had incomplete or no prior preparation for teaching special education students. This categor y of teachers stayed in teaching at rates similar to the teachers with more extensive preparation in teaching. School Assignment Data regarding the sample teachers school assignments were recorded each year of data collection. The characteristics of the school demographics were examined to determine whether the probability of leaving teaching differed with respect to placement. The percentage of minority student enrollment and percentage of students qualifying for free and reduced breakfast and lunch were identified and documented for the schools assigned. Research question 3: Does the risk of leaving teaching differ with respect to the school assignments demographic composition? The chi square statistic, X2 (3, n = 603) = 3.94, p = 0.27, indicated that the occurrence of higher percentages of minority students did not significantly impact the risk of teachers leaving. Table 49 provides data on the percentage of minority students in the school population and the percentage at which teachers stay or leav e after their first year of hire. Data on the socioeconomic level of the school assignment were also collected. The percentage of students eligible for free and reduced lunch programs served as a proxy for the poverty level of the schools. The chi square statistic, X2 (3, n = 603) = 1.68, p = 0.64, showed that the poverty level variable did not influence the risk of sample teachers leaving teaching. Table 410 provides data on the poverty level of the
85 student population and the percentage at which teachers stay or leave after their first year of hire. The nature of the teaching assignment was examined to determine whether the risk of leaving differed with respect to the type of exceptionality of the students taught. Common job responsibilities were used to group job assignments into five categories. Three of the five were specific to special education assignments that varied from self contained settings to pullout resource programs (categories 02). Although all of the sample teachers were hired to fill special education assignments, after their first year some teachers changed to general education, counseling, or district positions (categories 34). Research question 4: Does the risk of leaving teaching differ with respect to the special education teaching assignment? The chi square statistic, X2 (2, n = 603) = 1.00, p = 0.60, indicated that the teachers who left did not differ with regard to their special education teaching assignment. Table 411 provides statistics on the percentage at which teachers stay or leave after their first year of hire for each of the exceptionality assignment categories. However, a variable that did demonstrate a highly significant relationship with teacher retention was job change, X2 (1, n = 603) = 24.81, p = 0.00. The teacher s who moved from one job assignment to another left teaching at a much lower rate than the teachers who stayed in their same job assignment. Table 412 shows that 72% of the teachers who did not change their teaching assignment left teaching, while 52% of the teachers who did change their assignments left teaching.
86 Descriptive statistics for the 218 teachers who stayed in teaching for the duration of the data collection period indicated that 47% did not change special education teaching assignments and 24% changed from special education to general education assignments. The remaining teachers changed special education teaching assignments or left for district level positions. Table 413 provides a frequency distribution of the teaching assignment and school assignment changes. The finding that teachers tended to stay in teaching if they changed jobs prompted an analysis of whether the change was related to the teachers type of preparation. The chi square statistic, X2 (2, n = 603) = 0.66, p = 0.72 indicated that there was no relationship between the teachers type of preparation and the probability of changing teaching assignments. Table 41 4 provides data on the percentage of teachers who changed jobs after their first year of hire. Survival Analysis Af ter identifying the significant variables through statistical analysis, the next step was to conduct the survival analysis. These analyses provided information regarding whether and when teachers stayed, left, or returned to teaching. The findings address the remaining research questions: a) How many years on average do newly hired special education teachers continue to teach in a large urban school district? b) What is the probability of a special education teacher returning to teaching after leaving? The SAS LIFETEST Procedure was conducted using the variable of time to provide survival estimates for the sample teachers and a baseline with which to compare other variables of interest. Table 41 5 summarizes the distribution of event occurrence, i.e. leaving teaching in the first spell or time in teaching. The time column indicates the metric of time used to identify whether and when the event of interest
87 occurred, which in this study was an academic year. Each observation was conducted on June 30th of the sc hool year. The Survival column shows the percentage of teachers who survived or were still teaching at the time of the observation. The Failure column indicates the percentage of teachers who left teaching by the time of the June 30th observation each year The Failure column provides the hazard rate or risk of teachers leaving by the time of observation. Finally, the Number Failed and Number Left columns show the cumulative data for the number of teachers who left and the number of teachers who stayed over the course of the six year study. The spell one survival analysis showed a 71% probability that a teacher would survive through the first year, 54% through the second year, and a 44% probability of surviving through the third year. By the sixth year, there was a 34% probability of survival. The hazard probability identifies the risk for leaving teaching for each year. The analysis showed that there was a 29% risk that the teacher would not survive beyond the 1st year, a 56% risk of leaving in the third year, and in the 6th year there was a 66% risk that the teacher would leave teaching. It is important to note that censoring began to impact the data in the fourth time period or year of study. The teachers hired in the 2005 2006 school year were observed for a period of four years ending with the last observation on June 30, 2009. The 78 teachers who did not leave teaching by the end of data collection were listed as censored, we cannot know whether and when they left teaching. The SAS Survival Estimates output does not specify the number of teachers censored. The Number Left column reflects the difference. After time period four, 78 were censored, the 72 surviving teachers hired in 200405 were censored at the end of
88 time period five, and the 68 surviving teachers hired in 20032004 were censored at the end of time period six. There were 218 (36%) teachers who did not leave teaching by the end of data collection in 2009 and were censored. The survival analysis methodology has the advantage of the hazard f unction, which takes these censored teachers into consideration. The hazard function is computed on each years risk set, which includes all the teachers who have not left teaching up to the time of observation. Research question 5: How many years on average do newly hired special education teachers continue to teach in a large urban school district? The summary statistics on the time variable for teachers in their first spell of teaching showed the median survival time for the sample teachers was 3 years ( 95% CI = 2.0 to 3.0), indicating that at the year 3 observation, half of the teachers remained in teaching and half had left. The median survival statistic is the result of a skewed distribution. It has a lower bound but not an upper bound, indicating that the average teaching time would be longer. The mean survival time indicated the average number of years teachers continue to teach M = 3.43, SE = 0.09. It was noted that the mean survival time and its standard error were underestimated because the largest observation was censored and the estimation was restricted to the largest event time. The SAS Survival Distribution Function demonstrates the probability of the sample teachers continuing to teach through each period or year of data collection, from year one through year six. The use of time as the variable provided a baseline with which to compare how other variables of interest differ. The survival curve in figure 41 provides a visual representation of the data to assist with the communication of the f indings.
89 Teacher Characteristics The initial analysis of teacher characteristic variables indicated that there was a relationship between the special education teacher having prior teaching experience and staying in teaching. A survival analysis provided data showing that teachers with prior experience upon hire tended to stay in teaching at a higher rate than those with no prior experience. Table 41 6 shows the number of teachers who left teaching, or Failed at each level of prior experience. The Censored column shows the number of teachers who did not leave teaching during data collection. Teachers in their first and second year of teaching were retained at similar rates. Teachers with three to five years experience stayed at higher rates. With the except ion of stratum 5, which is comprised of teachers with 6 10 years of prior teaching experience, the percentage of teachers who stayed in teaching increased with the level of prior teaching experience. The survival curve in figure 42 provides a visual representation, which shows that the probability of teachers with prior teaching experience staying in teaching increases over time. Both the logrank and Wilcoxon tests demonstrated a significant difference in survival experience among teachers with varying levels of prior teaching experience. The rank tests for homogeneity indicate a significant difference among the sample teachers with regard to experience and staying in teaching, p = 0.01 for the logrank test and p = 0.01 for the Wilcoxon test. Initial variable analysis also identified teacher age as having an impact on the probability of leaving teaching. Further examination of the age variable through survival analysis showed that there was no statistically significant relationship. Both the logrank, p = 0.08 and Wilcoxon, p = 0.39, tests demonstrated no statistically significant
90 difference in survival experience among teachers of differing age during the period of data collection. Although not statistically significant, the data indicated that teacher s did tend to stay in teaching as age increased with the highest retention in the 4655 age range. This drops off as teachers enter the 56+ age group. Change Variable Change in teaching assignment was not initially considered for analysis. However, duri ng the analysis of the school assignment variables and their relationship to retention, change emerged as a variable that should be examined. An analysis of change in the teaching assignment showed a statistically significant impact on whether teachers stayed in teaching or left. The teachers who did not experience a change in their teaching assignment were found to leave teaching at higher rate than the teachers who did change. Table 41 7 provides summary statistics for the job change variable. The Failed column represents teachers who left teaching. The Censored column represents the teachers who did not leave teaching prior to the end of data collection. The survival curve in figure 43 provides a visual representation of the teachers who changed exceptio nalities or fields. The rank tests for homogeneity of survivor functions across strata indicated a significant difference among the sample teachers with regard to changing job assignments and staying in teaching, p = 0.00 for the logrank test and p = 0.00 for the Wilcoxon test. The teachers who changed job assignments stayed in teaching significantly longer than those who do not change assignments. Table 44b shows the job change data for the 218 special education teachers who did not leave teaching during data collection.
91 Research question 6: What is the probability of a special education teacher returning to teaching after leaving? The sample data show that of the 603 special education teachers studied, 385 teachers left teaching. A total of 126 or 33% of these teachers returned for a second spell in teaching. The SAS LIFETEST Procedure was conducted using the variable of time to provide survival estimates for the sample teachers who returned. Table 41 8 summarizes the distribution of the event occurr ence of interest, i.e. returning to teaching for a second spell or time in teaching. The spell two survival analysis showed a 30% probability that a teacher would return in the first year after leaving, a 31% probability that a teacher would return to teaching by the second year, and a 33% probability that a teacher would return by the third year out of teaching. A total of 259 teachers did not return by the end of data collection and were censored. The SAS Survival Distribution Function demonstrates the probability of the sample teachers returning to teaching for a second spell in teaching. The time variable provides a baseline with which to compare how other variables of interest differ. The survival curve in figure 44 provides a visual representation of the data to assist with the communication of the findings. Teacher Race An initial analysis of teacher characteristics and school assignment variables that may be related to a sample teacher returning to teaching yielded only one variable which indicated a relationship. The teachers race and returning for a second spell in teaching was statistically significant, chi statistic X2 (4, n = 385) = 19.31, p = 0.00. A survival analysis provided data showing that Black teachers tended to return to teaching
92 aft er a period out of teaching at a much higher rate. Table 41 9 shows the number and racial demographic for the teachers who returned to teaching and those who did not return during data collection. Forty four percent of the Black teachers returned to teachi ng and 25% of the White teachers returned for a second spell. Stratum three, four, and five show the remaining minority teachers, but are represented by small sample sizes. The survival curve in figure 45 shows that teacher race impacts the probability of teachers returning for a second spell in teaching. Both the logrank and Wilcoxon tests demonstrated a significant difference for teachers with regard to race. The rank tests for homogeneity indicate a significant difference for the sample teachers with r egard to race and returning to teaching, p = 0.00 for the logrank and p = 0.00 for the Wilcoxon test. A review of the descriptive data showed that 26 of the 51 (51%) White teachers who returned for a second spell in teaching were hired in schools with a 50% or higher level of minority students as compared to 48 of the 72 (67%) Black teachers hired in similar schools. For schools where 70% or more of the student enrollment was minority, 13 of the 51(25%) White teachers who returned were hired as compared to 30 of the 72 (42%) Black teachers.
93 Table 41. Frequency table: Descriptive data for invariant teacher characteristics Teacher characteristics Percent Frequency Gender Female Male 77 23 464 139 Race White Black Other 53 41 6 322 249 32 Age Range 2535 3645 4655 56+ 35 26 23 15 213 159 140 91 Teacher Preparation None Special Education General Educatio n 54 34 12 325 207 71 Experience at Hire 1st year 2nd year 3rd year 4th 5th year 6th 10th year 11th 15th year 16 th 20 th year 49 10 8 13 11 5 4 293 57 50 81 66 31 25 Note Three categories of teaching preparation (a) no prior training, (b) having a degree major in special education, or (c) having a degree major in general education. Sample: N = 603, percent reflects rounding
94 Table 42. Frequency table: Descriptive data for teachers with differing types of teacher preparation Characteristic No prior training Special education General education Percent Freq uency Percent Freq uency Percent Freq uency Gender Femal e Male 70 30 228 97 88 12 182 25 76 24 54 17 Race White Black Other 42 53 5 138 172 1 5 69 23 8 143 48 16 58 41 1 41 29 1 Years of experience 1st year 2nd year 3rd year 4th 5th year 6th 10th year 11 th + 56 9 9 14 9 4 182 29 29 44 28 13 44 11 6 13 14 13 91 22 13 26 29 26 28 8 11 15 13 24 20 6 8 11 9 17 Teaching assignment EH, SED MH, autistic SLD, VE, Pre K 8 21 70 27 69 229 6 23 71 13 48 146 7 20 73 5 14 52 School poverty percent 81 97 49 80 26 48 5 25 16 44 26 13 53 143 86 43 17 40 29 14 35 83 60 29 27 38 23 13 19 27 16 9 School minority percenta 70 100 50 69 30 49 14 29 40 36 16 7 131 118 51 23 30 32 29 8 63 66 61 17 46 27 17 10 33 19 12 7 Note. No Prior Training N = 325; Special Education N = 207; General Education N = 71 Percents reflect rounding aSchool Minority data missing for 2 female teachers with no prior training
95 Table 43. Frequency table: Descriptive statistics providing the school demographic data where the newly hired special education teachers are assigned to teach School demographic percent Newly hired special education teachers Percent Frequency School poverty range 81 97 49 80 26 48 5 25 18 42 27 13 107 253 162 81 School minority range a 70 100 50 69 30 49 14 29 38 34 21 8 227 203 124 47 Note. N=603, percent reflects rounding aSchool Minority data missing for 2 t eachers Table 44. Frequency table: Descriptive statistics of the racial breakdown of the special education teachers and the minority level of their school assignment School minority percent range White Black Other Percent Frequency Percent Fr equency Percent Frequency 70 100 28 91 51 127 28 9 50 69 33 106 33 83 44 14 30 49 28 91 10 25 25 8 14 29 10 33 5 13 3 1 Note. White N=322; Black N =249; Other N=32 (Asian 10, Hispanic 15, American Indian 7); missing school minority data for 1 White teacher and 1 Black teacher
96 Table 44b. Frequency table: Descriptive data for job change variable Placement Characteristics Percent Frequency School Assignment No change Change 62 38 135 83 Poverty Level No change Low to high High to low Multiple Changes 63 15 21 1 137 33 45 2 Minority Level No change Low to high High to low Multiple Changes 64 21 13 15 140 45 29 3 Teaching assignment No change Low to hi gh High to low Special education to General education District level Multiple changes 47 4 14 24 10 8 102 8 16 53 21 17 Note The low to high terminology indicates that the teacher moved from a school with a lower poverty or minori ty level to a higher poverty or minority level. High to low indicates the inverse. The low to high terminology for teaching assignment indicates that the teacher moved from a setting with students who required a greater number of services to a setting wher e students required a fewer number of services. District level positions are typically counseling or program resource. Sample: N = 218, percent reflects rounding, missing data for one teacher.
97 Table 45. Teacher gender and race: Does the sample teachers gender or race impact whether they stay in or leave teaching after their first year? Characteristics Staying Leaving p value Percent Frequency Percent Frequency Gender 0.45 Male 39 54 61 85 Female 35 164 65 300 Race a 0.75 White 37 120 63 202 Black 35 87 65 162 Hispanic 40 6 60 9 Asian 40 4 60 6 American Indian 14 1 86 6 Note. Gender: X2 (1, n = 603) = 0.57, p = 0.45; Race: X2 (4, n = 603) = 1.94, p = 0.75 aHispanic, Asian, and American Indian teachers are represented by small sample sizes Each teacher characteristic for gender and race is a mutually exclusive category and therefore provides one pvalue. *p < .05 Table 46. Teacher age: Does the sample teachers age impact whether they stay in or leave teaching after their first year? Age range Staying Leaving p value Percent Frequency Percent Frequency 25 35 30 63 70 150 0.02* 36 45 36 58 64 101 46 55 46 6 5 54 75 56+ 35 32 65 59 Note Age range: X2 (3, n = 603) = 10.44, p = 0.02* Each age range is a mutually exclusive category of teacher age and therefore one pvalue is provided. *p < .05
98 Table 47. Teaching experience: Does the sample teachers prior teaching experience impact whether they stay in or leave teaching after their first year? Prior experience Staying Leaving p value Percent Frequency Percent Frequency 1 st year 31 90 69 203 0.00** 2 nd year 30 17 70 40 3 rd year 3 6 18 64 32 4 th 5 th year 49 40 51 41 6 th 10 th year 35 23 65 43 11 th 15 th year 48 15 52 16 16 th 20 th year 60 15 40 10 Note. Teaching experience: X2 (6, n = 603) = 19.10, p = 0.00* Each teaching experience category is a mutually exclusive category of prior teaching experience and therefore one pvalue is provided. *p < .05. ** p<.01 Table 48. Teacher Preparation: Does the sample teachers level or type of teacher preparation impact whether they stay in or leave teaching after their first year? Preparation type Staying Leaving p value Percent Frequency Percent Frequency None 34 111 66 214 0.44 Special education 40 82 60 125 General education 35 25 65 46 Note. Teaching preparation: X2 (2, n = 603) = 1.66, p = 0.44 Each teaching preparation category is a mutually exclusive category of preparation and therefore one pvalue is provided. *p < .05. ** p<.01
99 Table 49. School Minority Level: Does the minority population of the school assignment impact wh ether the sample teachers stay in or leave teaching after their first year? Minority percent range Staying Leaving p value Percent Frequency Percent Frequency 70 100 35 80 65 147 0.27 50 69 33 66 67 137 30 49 43 53 5 7 71 14 29 40 19 60 28 Note. School minority level: X2 (3, n = 603) = 3.94, p = 0.27 Each minority range category is a mutually exclusive category of minority level and therefore one pvalue is provided. School minority data is missing for 2 teachers. *p < .05. ** p<.01 Table 410. School Poverty Level: Does the level of poverty in the school impact whether the sample teachers stay in or leave teaching after their first year? Poverty range percent Staying Leaving p value Percent Fr equency Percent Frequency 81 97 36 38 64 69 0.64 49 80 34 85 66 168 26 48 40 64 60 98 5 25 38 31 62 50 Note. X2 (3, n = 603) = 1.68, p = 0.64 Each poverty range is a mutually exclusive category of poverty level and therefore one p value is provided. *p < .05. ** p<.01 Table 411. Special Education Teaching Assignment: Does the assignment impact whether the sample teachers stay in or leave teaching after their first year? Teaching Assignment Staying Leaving p value Percent Frequency Percent Frequency EH, SED 42 19 58 26 0.60 MH, Autistic, PH 37 49 63 82 SLD, VE, Pre K 35 150 65 277 Note. X2 (2, n = 603) = 1.00, p = 0.60 Emotionally handicapped (EH), severely emotionally disturbed (SED); Mentally handicapped (MH) included profound, trainable, and educable; Autistic; Specific Learning disability (SLD), Varying Exceptionality (VE), Prekindergarten special education. Each teaching assignment is a mutually exclusive category of assignment and therefore one pvalue is provided. *p < .05. ** p<.01
100 Table 412. Change in teaching assignment: Does a change in teaching assignment impact whether the sample teachers stay in or leave teaching ? Change in job Staying Leaving p value Per cent Frequency Percent Frequency No Change 28 104 72 263 0.00* Change 48 114 52 122 Note. X2 (1, n = 603) = 24.81, p = 0.00* Each teaching assignment is a mutually exclusive category of assignment and therefore one pvalue is provided. *p < .05. ** p<.01 Table 413. Job change: Does the level or type of teacher preparation impact whether the sample teachers change teaching assignments ? Change in job No teacher preparation Special education preparation General education preparatio n p value Percent Frequency Percent Frequency Percent Frequency No Change 53 196 35 130 11 41 0.72 Change 55 129 33 77 13 30 Note. X2 (2, n = 603) = 0.66, p = 0.72 Each teaching category is a mutually exclusive category of teacher preparation and therefore one pvalue is provided. *p < .05. ** p<.01 Table 414. Spell one survival analysis: Distribution of event occurrence for special education teachers beginning 2003 ending 2009 Product Limit Survival Estimates Time Survival Failure Survi val standard error Number failed Number left 1 0.71 0.29 0.02 172 431 2 0.54 0.46 0.02 279 324 3 0.44 0.56 0.02 337 266 4 0.38 0.62 0.02 373 230 5 0.36 0.64 0.02 382 143 6 0.34 0.66 0.02 385 68 Note. Survival analysis column descriptors: Time column indicates the academic year; Survival column shows the percentage of teachers still teaching at the time of observation; Failure column show the percentage of teachers who left teaching by the time of observation (end of academic year); Number failed and number left columns show the cumulative data for the number of teachers who left and the number who stayed over the six year period. N = 603
101 Table 415. Survival analysis: How prior teaching experience impacts staying in teaching Summary of the N umber of Censored and Uncensored Values Stratum Teaching experience Total Failed Censored Percent censored 1 1st year 293 203 90 31 2 2nd year 57 40 17 30 3 3rd year 50 32 18 36 4 4 5 years 81 41 40 49 5 6 10 years 66 43 23 35 6 11 15 years 31 16 15 48 7 16 20 years 25 10 15 60 Total 603 385 218 36 Note. The stratum column represents each level of prior experience. The survival analysis showed the number of teachers who failed (left teaching) at eac h stratum by the observation date. The censored column shows the number of teachers who did not leave teaching during data collection. With the exception of stratum 5, the percentage of teachers who stayed in teaching increased with each stratum. Table 416. Survival analysis: Job change and the impact on leaving teaching Summary Statistics for Job Change Variable Stratum Total Failed Censored Percent Censored No Change 367 263 104 28 Change 236 122 114 48 Total 603 385 218 36 Note. The stratum column represents two categories teachers who did not change teaching assignments and teachers who experienced a change. The survival analysis showed the number of teachers who failed (left teaching) at each stratum by the observation date. The censored column shows the number of teachers who did not leave teaching during data collection.
102 Table 417. Spell two survival analysis: Distribution of event occurrence for special education teachers beginning 2003 ending 2009 Product Limit Survival Estimates T ime Out of teaching Return Out of teaching standard error Number returning Number remaining out of teaching 1 0.69 0.30 0.02 117 265 2 0.68 0.31 0.02 122 255 3 0.67 0.33 0.02 126 226 Note. Survival analysis column descriptors: Time column indicates the academic year; out of teaching column is the percentage of teachers who did not return at the time of observation; return column is the percentage who returned to teaching at the time of observation (end of academic year); Number returning and number out columns show the cumulative data. Two hundred twenty six teachers who did not return were censored after year three, 33 teachers were censored in years one through three. N = 385 teachers leaving after spell one, N = 126 teachers who returned for spel l two. Table 418. Survival analysis: How the teachers race impacts a return to teaching Summary of the Number of Censored and Uncensored Values Stratum Teacher race Total Return Censored Percent censored 1 White 202 51 151 75 2 Black 162 72 9 0 56 3 Hispanic 6 1 5 83 4 Asian 9 2 7 78 5 American Indian 6 0 6 100 Total 385 126 259 67 Note. The stratum column represents each race. The total column represents the total number of teachers who left teaching. The return column shows the number of teachers who returned to teaching at each stratum by the observation date. The censored column shows the number of teachers who did not return to teaching during data collection. Hispanic, Asian, and American Indian teachers are represented by small sample sizes
103 Probability 0.00 0.25 0.50 0.75 1.00 year 0 1 2 3 4 5 6 STRATA: Teaching_Experience=1 Censored Teaching_Experience =1 Teaching_Experience=2 Censored Teaching_Experience=2 Teaching_Experience=3 Censored Teaching_Experience=3 Teaching_Experience=4 Censored Teaching_Experience=4 Teaching_Experience=5 Censored Teaching_Experience=5 Teaching_Experience=6 Censored Teaching_Experience=6 Teaching_Experience=7 Censored Teaching_Experience=7 Figure 4 2. Survival distribution of teachers with varying levels of prior teaching experience. 0.00 0.25 0.50 0.75 1.00 year 0 1 2 3 4 5 6 Legend: Survival Curve Censored Obser vations Probability Figure 4 1. Survival distribution of the sample teachers over time.
104 0.00 0.25 0.50 0.75 1.00 YEAR SPELL 2 0 1 2 3 4 5 LEGEND: SURVIV AL CURVE CENSORED OBSERVATION S PROBABILITY Figure 4 4. Survival distribution of the returning sample teachers over time. Figure 4 3. Survival distribution for teachers changing teaching assignment an d leaving teaching. Probability 0.00 0.25 0.50 0.75 1.00 Year 0 1 2 3 4 5 6 STRATA: change_job=0 Censored change_job=0 change_job=1 Censored change_job=1
105 Probability 0.00 0.25 0.50 0.75 1.00 Year Spell 2 0 1 2 3 4 5 STRATA: race=1 Censored race=1 race=2 Censored race=2 race=3 Censored race=3 race=4 Censored race=4 race=5 Censored race=5 Figure 4 5. Survival distribution of the returning sample teachers and race.
106 CHAPTER 5 SUMMARY AND DISCUSSI ON The purpose of this study was to understand more about the characteristics of special education teachers who are retained in a large urban school district. The overarc hing goal was to determine who stays, who leaves, and who returns. Although teacher shortages and teacher retention have been subjects of research for some time, much has changed in recent years regarding who is entering the field and how they are becoming qualified to teach. The demographic of the teaching force is changing dramatically ( Feistritzer, 2005; Johnson, 2004), with an increasing number of men, minority candidates, and midcareer changers taking advantage of alternative certification route progr ams that allow them to begin teaching and complete certification requirements simultaneously. These programs have attracted a talented and diverse pool of prospective teachers from a variety of backgrounds who otherwise may not have chosen to enter teaching. Understanding more about the individuals who enter the field of special education and stay can provide important insights for teacher preparation program providers and district staffing representatives (Decker, et al., 2004; Humphrey, et al., 2008) Teacher turnover, which can indicate teachers leaving, switching teaching areas, or moving to a different school, can have a longreaching impact and contribute to a lack of stability in the teaching force (Billingsley, 2004; Boe & Cook, 2006; Ingersoll, 2001). The shortage of qualified special educators is of particular concern as growing numbers of special educators are being prepared through alternative route programs (Carlson, et al., 2002). Given the significant impact that teachers have on student
107 achiev ement (Hanushek, et al., 2002), continued research into who is teaching, how they are prepared, and whether they stay has added importance. Findings This study examined whether and when special education teachers in a large urban school district left teac hing and if they returned. The survival analysis methodology allowed for the study of the sample teachers retention behavior over a period of six years within the context of their varying teaching assignments. The study attempted to identify significant predictor variables that would indicate which teacher characteristics might signify a greater likelihood of the teacher staying in teaching. The following discussion summarizes the findings for each research question, examines how the analyses comport with current literature, and offers considerations for additional research. Discussion Research question 1: Does the risk of leaving teaching differ with respect to a special education teachers personal characteristics? The study findings regarding the sampl e teachers personal characteristics were mixed. The descriptive statistics identifying the race of the sample teachers showed that 53% were White, 41% were Black, and 6% were Hispanic, Asian, or American Indian. The findings, which indicated no significa nt difference in the risk of leaving teaching with regard to race, are of interest considering the higher percentage of Black teachers (51%) who were hired to teach in district schools with 70% to 100% minority student population as compared to 28% of the sample White teachers. In contrast to these findings, other researchers have shown that race does impact teacher retention when the race of the teacher differs from that of the student population.
108 Boyd et al. (2005) found that White teachers were nearly t wice as likely to leave or transfer from schools with a large minority population. Black teachers were found to stay longer in schools with higher rates of minority enrollment (Hanushek et al., 2002; Scafidi et al., 2003). The findings of this study differ considerably from the literature. Therefore, additional research is needed to find out why the sample Black teachers left teaching in schools where there were high levels of minority students, if indeed they should be more likely to stay. Alternative route programs attract a more diverse pool of prospective teachers in terms of age, gender, and ethnicity than traditional teacher preparation programs (Feistritzer, 2005, 2008; Zeichner & Schulte, 2001). This is an important contribution given the differences between the racial makeup of the student population and that of the teaching workforce. This study provides evidence that alternative route opportunities can attract a more diverse pool of candidates, 70% of the 139 male sample teachers, 69% of the 249 Black sample teachers, and nearly half (47%) of the 32 teachers who were Asian, Hispanic, American Indian, had incomplete or no prior teacher preparation. These teachers were transitioning to the field of education through alternative certification routes Although no relationship was found between the sample teachers level or type of teacher preparation and the probability of leaving teaching, as discussed in research question 2, further investigation of the interactions among teacher preparation, teacher race, and teacher retention in high minority schools would provide important information. Additional research is also needed to determine if there were other factors such as certification problems or working conditions involved in the attrition rate of t he sample Black teachers.
109 Consideration of job choice and organization fit theory are also instructive. The descriptive statistics in table 44 show that a disproportionate number of sample Black teachers were hired in high minority population schools. It is beyond the scope of this study to determine the rationale for the percentage of Black teachers hired in high minority schools and why they were not retained at higher rates as expected (Boyd et al., 2005; Hanushek et al., 2002; Scafidi et al., 2003). H owever, subjective job choice theory, which is discussed later may provide some insight into why the sample Black special teachers may have chosen positions in high minority schools, and why school administrators may have actively recruited these individuals. There are conflicting studies on whether gender impacts retention (Boyd et al., 2005; Ingersoll, 2001; Strunk & Robinson, 2006). There is also evidence that gender may interact with age and race as it relates to retention (Imazeki, 2005). The findings of this study did not show that the risk of leaving teaching differed with regard to gender. However, age and prior teaching experience, which typically coincide, were initially shown to have a statistically significant relationship with the newly hired t eachers risk of leaving teaching after the first year. Although prior experience remained significant through the survival analysis, age did not. Younger teachers did tend to leave teaching at a higher rate but not a statistically significant level. These findings are similar to those found in the literature (Harris & Adams, 2007). Multiple factors exist that contribute to the ushaped relationship that age and teaching experience have with attrition. Life cycle events such as child rearing, the age at whi ch a teacher begins his/her career, and retirement are all factors in attrition (Imazeki, 2005; Johnson, 2006; Miller, Brownell, & Smith, 1999; Strunk & Robinson,
110 2006). Special education teachers show the same pattern of early attrition as their general educator counterparts. Younger special educators are also more likely to transfer than older special education teachers (Billingsley, 2004). It is important to note that these teachers are not necessarily lost from the field of teaching; many of these teac hers contribute to the reserve pool of experienced teachers who later return to the field (Cook & Boe, 2007; Harris & Adams, 2007). The relationship between teacher characteristics such as age, gender, and retention has implications for programs that spec ifically recruit older, second career individuals (Dai, et al., 2007; Lui & Johnson, 2006). The large number of teachers (64% in this study) who were over the age of 36 demonstrates the importance of taking characteristics such as age into consideration when planning recruitment and selection strategies for hiring. The second career candidates are older, and have work and life experiences that can be valuable in the classroom. However, these individuals may also have considerable opportunity costs as they transition from their prior careers. Alternative route programs can help mitigate the cost of transitioning and allow the individual to begin teaching more quickly (Dai, et al., 2007). It is important for district induction and support program representati ves to recognize the unique needs of second career teachers. Provisions for certification counseling and skilled mentoring will help to provide the necessary supports while validating the knowledge and experience of older novice teachers. Retaining new t eachers so they can gain experience, and keeping experienced teachers in high need schools has taken on added importance. Evidence has shown
111 that a teachers ability increases within the first years of experience as measured by his or her contribution to s tudent learning (Clotfelter et al., 2006; Goldhaber, 2007; Hanushek, Kain, OBrien, & Rivkin, 2005; Harris & Sass, 2007; Kane, et al., 2006; Rockoff, 2004). Therefore, it becomes more important to provide the necessary supports for new teachers to persever e through the critical first years. Research question 2: Does the risk of leaving teaching differ with respect to a special education teachers level or type of teacher preparation? There were three types of teacher preparation among the sample special education teachers. The majority of the sample teachers (54%) had incomplete or no prior teacher preparation. These teachers were hired based on passing subject area certification tests for special education, and were allowed to complete their certificati on requirements while teaching. Over a third of the sample teachers (34%) had more extensive preparation in special education. The remaining 12% of the teachers were trained in general education. An analysis of the level and type of teacher preparation re vealed that the teachers who stayed in teaching and the teachers who left teaching did not differ significantly with regard to their teacher preparation. The finding that the risk for leaving teaching did not differ significantly for teachers with special education training is not consistent with other research indicating that more extensive preparation in special education is associated with higher retention (Boe, et al., 2007, Nougaret, et al., 2005). There are other factors that would provide insight int o why the sample teachers with more extensive preparation were not retained at a significantly higher rate than the alternatively certified teachers. This study did not collect data on the amount of professional development or support received at the
112 school, both of which could be important to retention. Findings regarding job change are discussed in research question five and may also shed light on why some teachers stay and others do not. Qualitative research on working conditions is needed to explore the retention behavior of teachers with regard to their preparation. This study did not track student achievement. However, recent research has demonstrated that more extensive preparation in special education is positively related to special education student learning gains (Feng & Sass, 2009). Student achievement has also been linked to teacher retention (Boyd, Grossman, Lankford, et al., 2007). Considering the insufficient number of new special education teachers prepared each year, and the prevalence and continued growth of alternative certification routes, it is important that research linking student achievement and teacher preparation is continued. Research question 3: Does the risk of leaving teaching differ with respect to a special education teachers school assignment? The descriptive statistics show that well over half of the sample teachers hired to teach in the districts high need schools left teaching after their first year. However, an analysis of the school and job assignment characteristi cs in this study indicated that the occurrence of higher percentages of minority students or low socioeconomic student populations did not have a statistically significant impact on the risk of the sample teachers leaving. Other studies have related the characteristics of a schools student population and socioeconomic status with its ability to recruit and retain teachers (Elfers, et al., 2006; Hanushek, et al., 2004; Imazeki, 2005; Scafidi, et al., 2006). New teachers have a
113 tendency to seek out positions in schools that are close to their hometowns or in demographically similar communities (Boyd, et al., 2005). The teachers race also plays a role. White teachers are more likely to turnover than their Black counterparts if the schools they are teaching in have a high proportion of minority students. Whether this is a function of racial composition, or the high poverty conditions that typically are characteristic of urban schools, is unclear (Hanushek, et al., 2004; Strunk & Robinson, 2006). Research question 4: Does the risk of leaving teaching differ with respect to the special educators teaching assignment? The majority of the sample special education teachers (71%) was hired for teaching assignments that included working with students with specific l earning disabilities, varying exceptionalities, and the prekindergarten special education population. These students were served in either self contained settings with the same teacher or resource pull out settings for a period of the day. Twenty two percent of the teachers were hired for teaching assignments serving students with intellectual and developmental disabilities previously classified as EMH, TMH, and PMH, and students with Autism. The smallest group (7%) was hired for teaching assignments serving the population of students with emotional and behavioral disorders. Special education teachers in the latter assignment typically had at least one paraprofessional teaching assistant. Each of these assignments provides its own set of challenges for the teacher. However, analysis of the special education teaching assignment did not show a statistically significant relationship between the special educators teaching assignment
114 and the risk of the teacher leaving teaching after the first year. It is import ant to note that over half of the teachers hired in each special education assignment left teaching during the six year data collection period. A variable of interest that arose during the study was job change. The teachers who changed from one special education assignment to another assignment left teaching at a much lower rate than teachers who stayed in the same teaching assignment. The descriptive data in table 413 show that almost a quarter of the 218 teachers who stayed in teaching for the duration of the study left special education teaching assignments for general education teaching assignments. This is not unexpected. Boe and Cook (2006) found that many new and returning special education teachers were prepared as general education teachers. Therefore, it is not surprising that they would return to general education teaching when a position became available. The remaining sample teachers either stayed in the same special education teaching assignment (47%), changed special educations assignments (26%), or moved to a district position (10%). Although there is a rationale for why some sample teachers changed from special education to general education assignments, there are still unanswered questions regarding the impact of the job change variable on the probability that the sample teachers would stay in teaching. Working conditions may play a role in why the sample special education teachers who changed assignments tended to stay in teaching at a higher rate. The districts transfer policy, which required that teachers remain in their initial school placement for three years before requesting a transfer may also have played a role. Special education teachers in particular have challenges that are specific
115 to their field and working conditions, which c an affect their decision to stay, move to another school, or leave teaching altogether (Billingsley, 2003; Gersten, et al., 2001; Ingersoll, 2003; Johnson & Birkeland, 2003; Johnson, 2006). School climate and collegial relationships have also been shown to be more important to special educators than to general educators, (Carlson, et al., 2002). Workplace conditions such as reasonable teaching assignments, manageable paperwork requirements, collaborative colleagues, meaningful professional development, and safe facilities have been identified as important factors for both recruitment and retention (Billingsley, et al., 2004; DeAngelis & Presley, 2007; Gersten, et al., 2001; Johnson, 2006). Administrative support has also been recognized as one of the more im portant factors in teacher retention and sense of efficacy, in addition to the need for mentoring and collegial support for new teachers (Billingsley, 2004; Boe, et al., 2008b; Miller, et al., 1999). Qualitative data regarding working conditions and job sa tisfaction could provide important insight into why the sample teachers were more likely to stay in teaching if they changed assignments. Research question 5: How many years on average do newly hired special education teachers continue to teach in a large urban school district? The survival analysis methodology allowed for the examination of both timevarying predictor variables and missing or censored data (teachers who did not leave teaching during data collection). Instead of taking teachers who left teaching out of the risk set (those teachers eligible to leave or return to teaching), the model allowed continued gathering of information. The spell one survival analysis provided information regarding how long we can expect special education teachers to s urvive or continue to
116 teach. The findings indicated a 71% probability that the teachers would survive through the first year and a 44% probability of surviving through the third year. By the sixth year, a little more than a third of the teachers were expec ted to still be in teaching. The hazard probability demonstrated the special education teachers risk of leaving teaching. The analysis showed there was a 29% risk that the teacher would not survive after the 1st year, a 56% risk that the teacher would leave teaching in the third year, and by the 6th year there was a 66% risk that the teacher would leave teaching. The average length of time that the sample special education teachers continued to teach was about three years. However, this statistic underest imates the length of time a teacher continues to teach as the result of a skewed distribution. The median survival time has a lower bound but not an upper bound, indicating that the average teaching time would be longer. Although teacher turnover varies c onsiderably within and across school districts, urban districts typically lose teachers at a higher rate than their suburban neighbors. The findings in this study comport with other literature on urban districts. In a study of New York City schools, approx imately 44% of elementary and 55% of middle school teachers leave or migrate from their initial placement within two years. Many of these teachers tend to migrate to higher achieving schools with fewer minority students (Boyd, Grossman, Lankford, et al., 2007). In Illinois school districts about 44% of the new teachers leave their initial school within their first two years, and 67% leave within five years. Roughly onethird of the teachers who leave during their first five years return to teachbut not in the most disadvantaged schools (DeAngelis & Presley, 2007).
117 Research question 6: What is the probability of a special education teacher returning to teaching after leaving? About 40% of the 603 sample special education teachers observed were returning to teaching with at least two years of experience at the beginning of data collection. These statistics support the research indicating that many of the teachers who fill vacancies are returning teachers, or part of what is referred to as the reserve pool. The pool is identified as a group composed of experienced teachers who left teaching and those who were prepared to teach but delayed entering the field. Approximately half of the vacancies in special and general education are filled from the reserve pool (Cook & Boe, 2007; Harris & Adams, 2007). The sample data show that of the 603 special education teachers studied, 385 teachers left teaching. A total of 126 or 33% of these teachers did return for a second spell in teaching. The spell two survival analys is showed a 30% probability that a teacher would return in the first year after leaving, a 31% probability that a teacher would return to teaching by the second year, and a 33% probability that a teacher would return by the third year out of teaching. A r elationship between the teachers race and a return to teaching was shown to be statistically significant. The survival analysis provided further data showing that Black teachers tended to return to teaching after a period out of teaching at a much higher rate than their White counterparts. Forty four percent (72 of 162) of the Black teachers and 25% (51 of 202) of the White teachers returned for a second spell in teaching. Singer (1993) also found that Black special education teachers returned to teaching at a much higher rate than White teachers.
118 The pattern of hiring Black teachers in schools with very high percentages of minority students, which was demonstrated among the newly hired sample teachers, continues with the returning teachers. In the school s where 70% or more of the student population was minority, 25% (13 of 51) of the returning White teachers were hired as compared to 42% (30 of 72) of the Black teachers who were returning to teaching ( DeAngelis & Presley, 2007; Hanushek, et al., 2004; Imazeki, 2005; Scafidi, et al., 2007). The pattern of hiring may have been related to the Black teachers desire to teach at schools where there were higher levels of minority students (Imazeki, 2005). However, other factors could be involved. At the time of the study, the district required screening interviews at the central office. Once cleared for hire, the candidates were referred to schools where vacancies existed and the school principal was responsible for final hiring decisions. Therefore, it is possib le that Black teachers may have been disproportionately referred to high minority population schools. Job Choice Theory Although working conditions are critical to retaining teachers, a good fit between the teacher and the school organization is an import ant first step (Cable & Judge, 1996; Kristof, 1996) It is also important to bear in mind that how employers recruit and hire new teachers has an influence on post hire outcomes such as job satisfaction, performance, and retention. The organizational behav ior and management literature provides insight into how job choice and personorganization fit can inform hiring practices. There are three theories of job choice: objective theory, critical contact theory, and subjective theory. Objective theory proposes that candidates make job decisions based
119 on economic factors such as pay, benefits, and other factors that are objective and measurable. Critical contact theory explains that candidates usually have limited knowledge or contact with the hiring organization. Therefore, job choice decisions are based on the relationship developed through contact with the recruiter. Subjective theory explains that job choice is determined based on the candidates perception of whether the organization will meet his or her psy chological needs (Behling, et al.,1968; Liu & Johnson, 2006; Young, et al.,1989). Urban Districts Subjective job choice theory may have the most applicability when staffing challenging schools and subject areas. Many individuals who are seeking ways to ent er the field of teaching, particularly in urban settings, are doing so through a sense of mission. They want to make a difference for children and communities. Clearly this type of motivation is important when considering candidates for highneed schools. Consequently, the schools culture, climate, and values are important factors to take into consideration when determining whether the candidate is a good match (Pounder & Merrill, 2001; Schein, 2004). It is essential that these candidates have realistic ex pectations, otherwise they can quickly become disillusioned when faced with the day to day challenges of teaching in a high need school. Many new teachers are relocating and working in communities with racial, ethnic, and socioeconomic backgrounds that differ from their own. Therefore, the extent to which there are opportunities available for the interviewers and interviewees to learn enough about the other to make informed decisions has important consequences (Pounder & Merrill, 2001). Principals who consider fit when hiring teachers increase the likelihood of maintaining a stable faculty, and fulfilling the mission of their school. An
120 individual is more likely to stay if the environment meets his or her needs. Conversely, individuals with low levels of c ongruence with the needs, values and culture of the school are more likely to leave (Judge, et al., 2000; Kristof, 1996; Schein, 2004). A good match or fit takes on additional importance when recruiting and retaining second career individuals who may have other career options available to them. These are all considerations when recruiting teachers for highneed schools and have implications for retention and ultimately student achievement (Dai, et al. 2007; Liu & Johnson, 2006). Implications for Practice I nformative hiring practices make a difference in the quality of job choice decisions (Liu & Johnson, 2006). Principals should involve their teacher leaders and other faculty in the selection and hiring process. Candidates should be invited to visit the school and meet and talk with other teachers. Whenever possible candidates should be given the opportunity to demonstrate their ability to teach and connect with students through demonstration lessons. These types of activities provide the candidate, principal, and potential peers with adequate information to make informed hiring decisions. Unfortunately, many teachers are hired after a review of their credentials and a brief interview at a recruitment event or school visit. The outcome is less informed appli cants and administrators and missed opportunities to make the best job matches. District Human Resource and staffing representatives should ensure that school administrators and other hiring agents are trained in effective hiring practices and have an understanding of job choice and organizationfit principles.
121 Recommendations for Future Research Teacher selection and hiring practices are critically important for schools and their stakeholders, especially in shortage areas such as special education. Additio nal research that examines effective recruitment, hiring practices, and most importantly, post hire outcomes of special educators would provide valuable information for the K 12 education community and those who prepare teachers. Ongoing study should also continue regarding how special educators are prepared to teach. Special education teachers have specific challenges with regard to student needs, paperwork requirements, and collaborating with colleagues. Research should inform professional development c ontent and induction support needs for special educators in general and second career individuals transitioning into special education in particular. As alternative certification programs expand and colleges of education examine their preparation programs, district administrators are looking for candidates who can positively impact student learning. Research that can link teacher characteristics including preparation, and student learning is needed to inform teacher educators, program developers, and distri ct administrators.
122 LIST OF REFERENCES Alliance for Excellent Education. (2005). Teacher attrition: A costly loss to the nation and to the states. Washington, DC: Author. Retrieved May 20, 2008 from http://www.all4ed.org/publications/TeacherAttrition.pdf Allison, P. D. (1982). Discrete time methods for the analysis of event histories. In S. Leinhardt (Ed.), Sociological methodology (pp. 6198). San Francisco: Jossey Bass. America n Association for Employment in Education, (2006) Educator supply and demand in the United States. Retrieved November 12, 2008, from http://www.aaee.org/cwt/external/wcpages/fi les/2006execsummary.pdf Behling, O., Laboritz, G., & Gainer, M. (1968). College recruiting: A theoretical base. Personnel Journal, 47 1319. Billingsly, B. (2003). Special education teacher retention and attrition: A critical analysis of the research l iterature (COPSSE Document No. RS2). Gainesville, FL: University of Florida, Center on Personnel Studies in Special Education. Billingsly, B. (2004). Special education teacher retention and attrition: A critical analysis of the research literature. The Journal of Special Education, 38, 3955. Boe, E. E. & Cook, L. H. (2006). The chronic and increasing shortage of fully certified teachers in special and general education. Exceptional Children, 72(4), 443460. Boe, E. E., Cook, L. H., & Sunderland, R. J. (2007). Teacher turnover in special and general education: Attrition, teaching area transfer, and school migration ( Research Report No. 2006TSDQ3). Center for Research and Evaluation in Social Policy. University of Pennsylvania: PA. Boe, E. E., Cook, L. H., & Sunderland, R. J. (2008a). Teacher qualifications and turnover: Bivariate associations with various aspects of teacher preparation, induction, mentoring, extra support, professional development, and workload factors for early career teachers in special a nd general education (Data Analysis Report No. 2008DAR1). Center for Research and Evaluation in Social Policy. University of Pennsylvania: PA. Boe, E. E., Cook, L. H., & Sunderland, R. J. (2008b). Teacher turnover: Examining exit attrition, teaching area transfer, and school migration. Exceptional Children, 75, 7 31. Boe, E. E., Shin, S., & Cook, L. H. (2007). Does teacher preparation matter for beginning teachers in either special or general education? The Journal of Special Education, 41(3) 158170.
123 B oyd, D., Grossman, P., Hammerness, K., Lankford, R. H., Loeb, S., McDonald, M., Reininger, M., Ronfeldt, M. & Wyckoff, M. (2007, August). Surveying the landscape of teacher education in New York City: Constrained variation and the challenge of innovation. Retrieved May 10, 2008 from http://www.teacherpolicyresearch.org/ResearchPapers/tabid/103/Default.aspx Boyd, D., Grossman, P., Lankford, R. H., Loeb, S., & Wyckoff, J (2007). Who leaves? Teacher attrition and student achievement. Retrieved May 10, 2008, from http://www.teacherpolicyresearch.org/ResearchPapers/tabid/103/Default.aspx Boyd, D., Lankford, H., Loeb, S., Rockoff, J., & Wyckoff, J. (2007). The narrowing gap in New York City teacher qualifications and its implications for student achievement in highpoverty schools. Retrieved May 10, 2008, from http://www.urban.org/url.cfm?ID=1001103 Boyd, D., Lankford, H., Loeb, S., & Wyckoff, J. (2002). Initial matches, transfers, and quits: Career decisions and the disparities in average teacher qualifications across schools. Retriev ed May 12, 2008, from http://www.teacherpolicyresearch.org/ResearchPapers/tabid/103/Default.aspx Boyd, D., Lankford, H., Loeb, S., & Wyckoff, J. (2005). The draw of home: How teachers preferences for proximity disadvantage urban schools. Journal of Policy Analysis and Management, 24, 113132. Breaugh, J. A. & Starke, M. (2000). Research on employee recruitment: So many studies, so many remaining questions. Journal of Management, 26 (3), 405434. Brownell, M. T., Hirsch, E., & Seo, S. (2004). Meeting the demand for highly qualified special education teachers during severe shortages: What should policy makers consider? Journal of Special Education, 38, 5661. Brownell, M. T., Ross, D. D., Colon, E. P. & McCallum, C. L. (2005). Critical features of special education teacher preparation: A comparison with general teacher education. The Journal of Special Education, 38(4), 242 252. Cable, D. M. & Judge, T. A. (1996). Personorganization fit, job choice decisions, and organizational entry. Organizational Behavior and Human Decision Processes, 67(3), 294311. Carlson, E., Brauen, M., Klein, S., Schroll, K., & Willig, S. (2002). Spence: Study of personnel needs in special educa tion. Retrieved January 20, 2007, from http://www.spense.org Clotfelter, C. T., Ladd, H. F., & Vigdor, J. L. (2006). Teacher student matching and the assessment of teacher effectiveness. Journal of Human Resources, 41(4), 778 820.
124 CochranSmith, M., & Fries, K. (2005). Researching teacher education in changing times: Politics and paradigms. In M. CochranSmith & K. M. Zeichner (Eds.), Studying teacher education: The report of the AERA panel on research and teacher education (pp. 69109). Mahwah, NJ: Erlbaum. CochranSmith, M., & Zeichner, K. M. (Eds.). (2005). Studying teacher education: The report of the AERA panel on research and teacher education. Mahwah, NJ: Erlbaum. Cook, L. H., & Boe, E. E. (2007). National trends in the sources of supply of teachers in special and general education. Teacher Education and Special Education, 30(4), 217232. Council of the Great City Schools. (2000, January). The urban teacher challenge: Teacher demand and supply in the great city s chools. Retrieved February 7, 2008, from http://www.cgcs.org/pdfs/utc.pdf Dai, C., Sindelar, P. T., Denslow, D., Dewey, J., & Rosenberg, M.S. (2007). Economic analysis and the design of alternative route teacher education programs. Journal of Teacher Education, 58(5), 422439. Darling Hammond, L. (1999). Teacher quality and student achievement: A review of state policy evidence. Seattle: University of Washington, Center for the Study of Teaching and Policy. Dar ling Hammond, L., Holtzman, D. J., Gatlin, S. J., & Heilig, J. V. (2005). Does teacher preparation matter? Evidence about teacher certification, teach for America, and teacher effectiveness. Education Policy Analysis Archives, 13(42). DeAngelis, K. J., & Presley, J. B. (2007). Leaving schools or leaving the profession: Setting Illinois record straight on new teacher attrition. Edwardsville, IL: Illinois Education Research Council Policy Research Report: IERC 2007 1. (ERIC Document Reproduction Service No. ED 497703) Decker, P. T., Mayer, D. P., & Glazerman, S. (2004). The effects of teach for America on students: Findings from a national evaluation. Princeton, NJ: Mathematica Policy Research. Eide, E., Goldhaber, D. & Brewer, D. (2004). The teacher labour m arket and teacher quality. Oxford Review of Economic Policy, 20(2), 230244. Elfers, A. M., Plecki, M. L., & Knapp, M. S. (2006). Teacher mobility: Looking more closely at the movers within a state system. Peabody Journal of Education, 81, 94127. Feist ritzer, C. E. (2005). Profile of alternate route teachers. Washington, D.C.: National Center for Education Information.
125 Feistritzer, C. E. (2008). Alternative Teacher Certification: A state by state analysis. Washington, D.C.: National Center for Education Information. Florida Department of Education Information and Accountability Services. (2009, November). Florida schools indicator report: Student membership by category. Retrieved November 10, 2010, from http://www.fldoe.org/eias/eiaspubs/fsir.asp Florida Legislature Office of Program Policy Analysis and Government Accountability. (2003, July). Exceptional student education population grows dramatically: More accountability and better training needed to implement funding matrix (Special Report No. 0304). Retrieved January 15, 2006, from http://www.oppaga.state.fl.us/reports/pdf/0340rpt.pdf Florida TaxWatch. (2002, October). Amendm ent 9 on class size reduction is structurally unsound: Putting at great risk revenues, costs, quality teachers and Florida's future. Retrieved December 2, 2005, from http ://www.floridataxwatch.org/resources/pdf/102002amendment9Unsound.pdf Gersten, R., Keating, T., Yovanoff, P., & Harniss, M. K. (2001). Working in special education: Factors that enhance special educators intent to stay. Exceptional Children, 67(4), 54956 7. Goe, L. (2007). The link between teacher quality and student outcomes: A research synthesis. National Comprehensive Center for Teacher Quality. Retrieved November 15, 2008, from http://www.tqsource.org/link .php Goldhaber, D. D. (2007, April). Everyones doing it, but what does teacher testing tell us about teacher effectiveness? (Working Paper No. 9). National Center for Analysis of Longitudinal Data in Education Research. Retrieved February 18, 2009, from www.caldercenter.org/PDF/1001072_everyones_doing.PDF Goldhaber, D. D., & Anthony, E. (2003). Teacher quality and student achievement Urban Diversity Series No. 115. Institute for U rban and Minority Education, ERIC Clearinghouse on Urban Education. New York: Teachers College Goldhaber, D., & Brewer, D. (2000). Does teacher certification matter? High school teacher certification status and student achievement. Educational Evaluation and Policy Analysis 22(2), 129145. Guarino, C. M., Santibanez, L., & Daley, G. A. (2006). Teacher recruitment and retention: A review of the recent empirical literature. Review of Educational Research, 76(2) 173208. Hanushek, E. A., Kain, J. F., OBrien D. M., & Rivkin, S. G. (2005). The market for teacher quality ( Working Paper). National Bureau of Economic Research. Cambridge, MA. Retrieved September 10, 2008, from http://edpro.s tanford.edu/Hanushek/files_det.asp?FileId=136
126 Hanushek, E. A., Kain, J. F., & Rivkin, S. G. (2004). The revolving door: A pathbreaking study of teachers in Texas reveals that working conditions matter more than salary. Education Next, 1, 7782. Harris, D N., & Adams, S. J. (2007). Understanding the level and causes of teacher turnover: A comparison with other professions. Economics of Education Review 26, 325337. Harris, D. N., & Sass, T. R. (2007, March). Teacher training, teacher quality, and student achievement (Working Paper No. 3). National Center for Analysis of Longitudinal Data in Education Research. H. R. Rep. No. 108779, (2004). Humphrey, D. C., Wechsler, M. E. (2007). Insights into alternative certification: Initial findings from a national study, Teachers College Record, 109(3), 483 530. Humphrey, D. C., Wechsler, M. E., Hough, H. J. (2008). Characteristics of effective alternative teacher certification programs. Teachers College Record, 110, 1 63. Imazeki, J. (2005). Teacher salaries and t eacher attrition. Economics of Education Review, 24, 431449. Ingersoll, R.M. (2001). Teacher turnover, teacher shortages and organization of schools. Seattle: University of Washington, Center for the Study of Teaching and Policy. Ingersoll, R.M. (2003). Is there really a teacher shortage? Seattle: University of Washington, Center for the Study of Teaching and Policy. Johnson, S. M. (2004). Finders and keepers: Helping new teachers survive and thrive in our schools San Francisco: Jossey Bass. Johnson, S. M. (2006). The workplace matters: Teacher quality, retention, and effectiveness. Washington, DC: National Education Association. (ERIC Document Reproduction Service No. ED495822) Johnson, S. M., Berg, J. H., & Donaldson, M. L. (2005). Who stays in teaching and why: A review of the literature on teacher retention. Boston, MA: Harvard Graduate School of Education. Available at http://assets.aarp.org/www.aarp.org_/articles/NRTA /Harvard_report.pdf Johnson, S. M., & Birkeland, S. E. (2003). Pursuing a sense of success: New teachers explain their career decisions. American Educational Research Journal, 40(3), 581617.
127 Johnson, S. M., Birkeland, S. E., & Peske, H. G. (2005). Life i n the fast track: How states seek to balance incentives and quality in alternative teacher certification programs. Educational Policy, 19(1), 6389. Johnson, S. M., Kardos, S. M., Kauffman, D., Liu, E., & Donaldson, M. L. (2004). The support gap: New teachers early experiences in highincome and low income schools. Education Policy Analysis Archives, 12(61). Retrieved August 14, 2008, from http://epaa.asu.edu/epaa/v12n61/ Judge, T. A., Higgins, C. A., & Cable, D. M. (2000). The employment interview: A review of recent research and recommendations for future research. Human Resource Management Review, 10(4), 383406. Kane, T. J., Rockoff, J. E., & Staiger, D. O. (2006). What does certification tell us about teacher effectiveness? Evidence from New York City (Working paper No. 12155) Cambridge, MA: National Bureau of Economic Research. Katsiyannis, A., Zhang, D., & Conroy, M. A. (2003). Remedial and Special Education, 24(4), 246253. Kristof, A. L. (1996). Perso n organization fit: An integrative review of its conceptualizations, measurement, and implications. Personnel Psychology, 49, 1 49. Lankford, H., Loeb, S., & Wyckoff, J. (2002). Teacher sorting and the plight of urban schools: A descriptive analysis. Educ ational Evaluation and Policy Analysis, 24, 3762. Levine, A. (2006). Educating school teachers Retrieved December 18, 2007, from http://www.edschools.org/teacher_report.htm Levin, J., & Quinn, M. (2003). Missed opportunities: How we keep high quality teachers out of urban classrooms. Washington, DC: New Teacher Project. Liu, E. & Johnson, S. M. (2006). New teachers experiences of hiring: Late, rushed, and information poor. Educational Administrat ion Quarterly, 42(3), 324360. Loeb, S., DarlingHammond, L., & Luczack, J. (2005). How teaching conditions predict teacher turnover in California schools. Peabody Journal of Education, 80(3), 4470. Masyn, K. (2003). Discretetime survival mixture analys is for single and recurrent events using latent variables. Unpublished doctoral dissertation, University of California, Los Angeles. Mcleskey, J., Tyler, N. C., & Flippin, S. S. (2004). The supply of and demand for special education teachers: A review of research regarding the chronic shortage of special education teachers. The Journal of Special Education, 38(1), 5 21.
128 Meglino, B. M., Ravlin, E. C., & DeNisi, A. S. (2000). A meta analytic examination of realistic job preview effectiveness: A test of three counterintuitive propositions. Human Resource Management Review, 10(4), 407434. Mikulecky, M., Shkodriani, G., & Wilner, A. (2004). Alternative certification: A growing trend to address the teacher shortage. Education Commission of the States. Retrieved on May 20, 2008 from http://www.ecs.org/clearinghouse/57/12/5712.pdf Miller, M. D., Brownell, M. T., & Smith S.W. (1999). Factors that predict teachers staying in, leaving, or transferring fro m the special education classroom. Exceptional Children 65 (2), 20118. Murphy, P., DeArmond, M., Guin, K. (July 31, 2003). A national crisis or localized problems? Getting perspective on the scope and scale of the teacher shortage. Education Policy Analysi s Archives, 11(23). Retrieved August 22, 2008, from http://epaa.asu.edu/epaa/v11n23/ National Commission on Teaching and Americas Future. (2003). No dream denied: A pledge to Americas children. Retrieved Dec ember 6, 2007, from http://www.nctaf.org/resources/research_and_reports/nctaf_research_reports/ National Education Association. (2006, July). The workplace matters: Teacher quality, retention, and effectiveness. (Working Paper) Washington D.C: Johnson, S. M. Nougaret, A. A., Scruggs, T. E., & Mastropieri, M. A. (2005). Does teacher education produce better special education teachers? Exceptional Children, 71, 217229. Peske, H. G. (2003). Faculty are the backbone: Quality control in Connecticuts alternative route to certification. Retrieved July 14, 2005, from Harvard Graduate School of Education: http://www.gse.harvard.edu/~ngt/Peske%20AERA2003%20CT%20ARC.pdf Popovich, P., & Wanous, J. P. (1982). The realistic job preview as persuasive communication. Academy of Management Review, 7, 570578. Pounder, D. G., & Merrill, R. J. (2001). Job desir ability of the high school principal: A job choice theory perspective. Educational Administration Quarterly, 37 (1), 27 57. Rivkin, S., Hanushek, E., & Kain, J. (2005) Teachers, schools, and academic achievement Econometrica, 73(2), 417458. Rockoff, J. (2 004). The impact of individual teachers on student achievement: Evidence from panel data. American Economic Review 94(2), 247 252. Rosenberg, M. S., Boyer, K. L., Sindelar, P. T., & Misra, S. K. (2007). Alternative route programs for special education: Program infrastructure, instructional delivery, and participant characteristics Exceptional Children, 73(2), 224 241.
129 Rosenberg, M. S., & Sindelar, P. T. (2005). The proliferation of alternative routes to certification in special education: A critical review of the literature. The Journal of Special Education, 39(2), 117127. Rynes, S. L., Bretz, R. D., & Gerhart, B. (1991). The importance of recruitment in job choice: A different way of looking. Personnel Psychology, 44, 487521. Sachs, S. K. (2004). Evaluation of teacher attributes as predictors of success in urban schools. Journal of Teacher Education, 55(2), 177187. Sanders, W. L. & Rivers, J. C. (1996). Cumulative and residual effects of teachers on future student academic achievement. Research Progress Report. Knoxville, TN: University of Tennessee ValueAdded Research and Assessment Center. Scafidi, B., Sjoquist, D. L., & Stinebrickner, T. R. (2007). Race, poverty, and teacher mobility. Economics of Education Review, 26, 145159. Schein, E. (2004). Organizational culture and leadership (3rd ed.). San Francisco: JosseyBass. Sindelar, P. T., Daunic, A., & Rennells, M. S. (2004). Comparisons of traditionally and alternatively trained teachers. Exceptionality, 12(4), 209223. Singer, J. D. (1993). Onc e is not enough: Former special educators who return to teaching. Exceptional Children, 60, 5872. Singer, J. D., & Willett, J. B. (1993). Its about time: Using discrete time survival analysis to study duration and the timing of events. Journal of Educati onal Statistics, 18, 155195. Singer, J. D., & Willett, J. B. (2003). Applied longitudinal data analysis: Modeling change and event occurrence. New York: Oxford University Press. Strunk, K. O., & Robinson, J. P. (2006). Oh, wont you stay: A multilevel analysis of the difficulties in retaining qualified teachers. Peabody Journal of Education, 81(4), 65 94. Thornton, B., Peltier, G., & Medina, R. (2007). Reducing the special education teacher shortage. The Clearing House, 80(5), 233238. Tyler, N. C., Yzquierdo, Z., Lopez Reyna, N., & Flippin S. S. (2004). Cultural and linguistic diversity and the special education workforce: Critical overview. Journal of Special Education 38 (1), 22 38. U. S. Department of Education, National Center for Education Statistics Higher Education General Information Survey. (2007). Degrees and other formal awards conferred surveys, 197071 through 198586; and 199091 through 200506 (IPEDSC:91 99).
130 U. S. Department of Education, National Center for Education Statistics, Teacher Followup Survey (TFS), Current and former teacher data files, 200405; Teacher attrition and mobility: Results from the teacher follow up survey, 200001, U.S. Department of Education, National Center for Education Statistics (NCES 2004301). U.S. Depa rtment of Education, Office of Special Education and Rehabilitative Services, Office of Special Education Programs, 27th Annual (2005) Report to Congress on the Implementation of the Individuals with Disabilities Education Act, vol. 1, Washington, D.C., 2007. U.S. Department of Labor, Bureau of Labor Statistics (2008). Occupational outlook handbook, 200809. Retrieved February 8, 2008, from http://www.bls.gov/oco/ocos070.htm Villegas, A. M., & Clewell, B. C. (1998). Increasing the number of teachers of color for urban schools. Education and Urban Society, 31, 4261. Wayne, A. J. & Youngs, P. (2003). Teacher characteristics and student achievement gains: A review. Review of Educational Research, 73(1), 89122. Willett, J. B., & Singer, J. D. (1993). Investigating onset, cessation, relapse, and recovery: Why you should, and how you can, use discretetime survival analysis to examine event occurrence. Journal of Consulting and Clinical Psychology, 61, 952965. Wi llett, J. B., & Singer, J. D. (1995). Its dj vu all over again: Using multiplespell discretetime survival analysis. Journal of Educational and Behavioral Statistics, 20, 4167. Wilson, S.M., Floden, R. E., & Ferrini Mundy, J. (2001). Teacher preparati on research: Current knowledge, gaps, and recommendations. Seattle, WA: Center for the Study of Teaching and Policy. Wright, S. P., Horn, S. P., & Sanders, W. L. (1997). Teacher and classroom context effects on student achievement: Implications for teacher evaluation. Journal of Personnel Evaluation in Education, 5767. Young, I. P., Rinehart, J. S. & Place, A. W. (1989). Theories for teacher selection: Objective, subjective, and critical contact. Teaching & Teacher Education, 5(4), 329336. Zeichner, K. M. & Conklin, H. G. (2005). Teacher education programs. In M. CochranSmith & K. M. Zeichner (Eds.), Studying teacher education: The report of the AERA panel on research and teacher education (pp. 645735). Mahwah, NJ: Erlbaum.
131 Zeichner, K. M. & Schulte, A. K. (2001). What we know and dont know from peer reviewed research about alternative teacher certification programs. Journal of Teacher Education 52(4), 266282.
132 BIOGRAPHICAL SKETCH Melissa Dunn completed her undergraduate degree in e lementary e ducation at the University of West Florida in 1987. She served her internship in a Title I school where she taught fifth grade. After a move to northeast Florida, Melissa taught various grade levels in both Title I and nonTitle elementary schools. In 1998, Melissa completed her masters degree in e ducation with a program of study in Reading from the University of North Florida (UNF). She later completed additional graduate coursework in Educational Leadership at UNF in order to obtain Florida Educator C ertification in Leadership. Melissa began a new career path in education when she accepted a special assignment in Human Resources with the Clay County School Board. This position provided the opportunity to serve as a resource teacher assisting the districts begin ning teachers and also serving as a clinical instructor for field experience seminars in the UNF College of Education. Upon completion of this threeyear commitment Melissa accepted a position with Duval County Public Schools working with the Teacher Induction Program. Since 2001, she has worked in various capacities with teacher induction and alternative certification programs. Melissa is currently the Supervisor of Teacher Induction in the district She holds Florida Educator Certification in Elementary Education grades 1 6, Reading grades K 12, and Educational Leadership.