GAINING A BETTER UNDERSTANDING OF TEACHER ABSENTEEISM By Jacqueline A. Gardner A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Educational Policy – Doctor of Philosophy 2020 GAINING A BETTER UNDERSTANDING OF TEACHER ABSENTEEISM ABSTRACT By Jacqueline A. Gardner This study explores the reasons for chronic teacher absenteeism, which is a growing concern in Michigan schools and districts because teacher absenteeism is expensive and has implications for student learning. In total, 21 elementary teachers from seven different Michigan school districts were interviewed about their experiences with and perceptions of teacher absenteeism. Using an interpretive lens from the management literature on employee absenteeism, this study finds chronically absent teachers used paid time off for job-related mental health more often than for personal reasons compared to non-chronically absent teachers. The contributors to job-related mental health include three elusive, intangible characteristics: job stress, low perceived organizational support, and job dissatisfaction. The sources of job stress, low perceived organizational support, and job dissatisfaction are negative student behaviors, large class sizes, accountability pressures, lack of building support, lack of administration support, increased workload, performing additional duties above and beyond teaching, and not having enough district financial support for classroom materials and resources. The findings suggest organizational factors, or challenges with school and district culture and climate, contribute to chronic teacher absenteeism and must be improved upon. Implications for research, policy, and practice are discussed, along with opportunities for future research on both chronic teacher absenteeism and the culture and climate of schools and districts. Copyright by JACQUELINE A. GARDNER 2020 ACKNOWLEDGEMENTS This dissertation would not have been possible without the support, feedback, and guidance from my dissertation committee, including Dr. John Yun, Dr. Chris Torres, Dr. David Arsen, and Dr. Amita Chudgar. Each individual on my committee invested a substantial amount of time and thoughtful feedback to my work, which contributed to the successful completion of this dissertation. I would especially like to thank my advisor, Dr. Yun, who was by my side throughout my entire PhD journey and always had my best interests in mind. Not only did Dr. Yun push me to become a stronger scholar and academic, but also his exemplary mentorship made my PhD experience both positive and memorable. I aspire to be the mentor Dr. Yun was to me. I also want to thank Dr. Torres, who stepped up to co-chair my dissertation. Dr. Torres’ qualitative research expertise helped to guide my research and inform my writing. His detailed and thorough feedback on my dissertation expanded my thinking and writing, which shaped my dissertation into its final academic form. I learned a great deal from each of my committee members, and I owe a huge thank you to my entire committee for investing their valuable time in my development and providing me with feedback to make me a better scholar. I am also incredibly grateful for the unconditional love and support from my husband, Alex, who provided me with a lot of grace, patience, and thought partnership throughout the entirety of my PhD, but especially during the dissertation writing process. I love you! I would also like to thank my parents, Lauren Swanson-Aprill and Randy Bodwin, who have supported me since day one and who instilled the value education and endless learning in me at a very young age; and in-laws, Richard and Jo Ann Gardner, for their understanding and for always checking in on me. My sister, Theresa Bodwin, also provided me with much needed iv cheerleading and support in the form of offering distractions and comic relief, for which I am especially grateful. Finally, I would like to thank my colleague, mentor, and friend, Dr. Neelam Kher, for her encouragement, thought partnership, and Friday coffees. Thank you, all, for your support, mentorship, investment, encouragement, and for staying by my side as I completed this goal and achieved my dream. TABLE OF CONTENTS LIST OF TABLES INTRODUCTION LITERATURE REVIEW U.S. Federal Government’s Take on Improving Student Achievement Problem Statement and Research Question Purpose of the Study Study Significance Ideological Statement Evidence of High Rates of Teacher Absenteeism Teacher Absenteeism Negatively Impacts Student Achievement Teacher Absenteeism is Expensive Not Much is Understood about Why Teachers are Absent Teacher Absenteeism is Hard to Measure Reducing Teacher Absenteeism Teacher Turnover Absenteeism and the Management Literature Organizational Factors and Employee Absenteeism Unions and Employee Absenteeism Challenges with Translating Management Frameworks to Education Summary of the Literature Review METHODS FINDINGS Epistemological Perspective Method for Data Collection Interviews Participants Participant Overview Positionality Statement Data Collection Validity and Reliability Data Analysis and Coding Summary of Methods Reported Teacher Absenteeism Participant Characteristics, District Characteristics, and Chronic Absenteeism Reasons for Teacher Absenteeism vi viii 1 1 4 6 8 10 12 13 15 17 18 19 20 22 26 29 33 37 40 47 47 49 51 60 66 70 74 75 77 82 84 86 87 90 Job-Related Reasons for Teacher Absenteeism Job Stress and Student Behaviors Job Stress and Class Size Job Stress and Accountability Perceived Organizational Support Dissatisfaction and Financial Support Approach to Using Paid Time Off CBAs, Teacher Attendance Policies, and Incentives Limitations Overall Contribution Next Steps and Opportunities for Future Research Research Policy Practice DISCUSSION CONCLUSION APPENDICES REFERENCES APPENDIX A APPENDIX B APPENDIX C 92 93 95 96 101 109 112 114 127 137 140 140 142 145 146 151 154 155 160 161 163 vii LIST OF TABLES Table 1: Characteristics of a Purposeful Sample Table 2: District Characteristics Table 3: Participant Overview Table 4: Considerations for Validity and Reliability Table 5: Details from CBAs about Paid Time Off, by District 63 67 70 77 116 viii INTRODUCTION Education researchers agree that teachers are a vital input in the education production function (Lankford, Loeb, & Wykoff, 2002; Corcoran, Evans, & Schwab, 2004; Chetty, Friedman, & Rockoff, 2012) and may be the single most important input to improving student learning outcomes. If teachers are the largest and perhaps most important school resource, then understanding teacher behaviors may help explain variation in student achievement. One such behavior, teacher attendance, is not well understood in the United States (U.S.) context because teacher attendance policies are created locally through a collective bargaining agreement (CBA), and, as such, data on teacher attendance are limited. Similarly, teacher attendance is not factored into state or national accountability requirements, leaving it largely overlooked. This is problematic because for teachers to impact student achievement they must first be present. Thus, gaining a better understanding around why teachers are absent from their classrooms, including but not limited to how teachers decide to be absent, could provide further insights to variation in student achievement. U.S. Federal Government’s Take on Improving Student Achievement The question of how to improve student achievement in the U.S. has perplexed researchers, policy makers, and educators for decades. The driving forces behind this question come from the desires to be internationally academically competitive and also to provide equitable education to all students, among others. When Russia launched Sputnik in 1957, the concern surfaced that the academic performance of U.S. students was below that of Russian students, especially in relation to science and math performance (Mitra, 2017). This international competition with Russia sparked U.S. educational policy reform that targeted improvement in math and science achievement in U.S. students. 1 Not long after, providing equitable education to all students, especially to economically disadvantaged students, became a top priority in the 1960s following the Civil Rights Movement and a heightened awareness of poverty. This, coupled with the desire to be internationally academically competitive, resulted in major education policy reform through the Elementary and Secondary Education Act (ESEA) (Manna, 2007). ESEA provided resources to disadvantaged populations, but it also kicked off the standards-based education reform movement and increased accountability, with the ultimate goal of improved academic achievement for all students (Manna, 2007). While ESEA had the goal of improving academic achievement for all students, there was little evidence to support the realization of such gains. Additionally, Ronald Reagan’s A Nation at Risk report renewed the urgency around the U.S. needing to be internationally academically competitive (Mitra, 2017). Education policy changes were made again in 2001 through the No Child Left Behind Act (NCLB), still with the hope of improving student achievement and being internationally academically competitive (Mitra, 2017). This time around, high-stakes, school- based accountability for districts, schools, and teachers was put in place, which called for evidence of improved performance through standardized tests that could be compared on an international scale (McDermott, 2011). The lofty goal of 100% of U.S. students being proficient by 2014 as outlined in NCLB was not fully met, sending education policy makers back to the drawing boards. The reauthorization of ESEA through the Every Student Succeeds Act of 2015 (ESSA) responds to the recent and renewed push to be internationally academically competitive, which is driven by the desire to be globally competitive in the economic market, and the idea that the U.S.’s ability to compete globally will be improved through education. While ESSA allows for flexibility in the fulfillment of accountability requirements, school-based and teacher 2 accountability are still a heavy component with an added emphasis on preparing all students to be college and career ready; however, the bulk of these school and teacher accountability measures still rely on student performance (Mitra, 2017). Through federal policy there have been many efforts to improve the academic achievement of U.S. students. Many governmental financial and human resources have been spent on this topic, and yet we are still asking the question of how to improve student achievement in the U.S. through the goals of providing equitable education to all students and being internationally academically competitive. These federal policies include heavy school and teacher accountability components, but much of the accountability is determined by student performance (Carlson, 2018). As such, throughout these changes to U.S. federal education policy, academic researchers delved into a myriad of teacher and student characteristics, and local policies and interventions thought to impact student academic performance, such as early childhood education, educational partnerships, education governance, education finance, access, equity, educational leadership, teacher quality, evaluation, school resources, and more (Jochim, 2018). While each of these areas yielded informative findings, they pose challenges for broad implementation in their own right. Again, we are left with the question of how to improve student achievement in the U.S. The bottom line is that there is no silver bullet (Jochim, 2018); however, researchers can continue to explore research streams that might yield additional pieces of information in answering the question of how to improve student achievement in the U.S. One such research stream is around teacher behaviors that might have an impact on student performance. More recently, federal policy has included increased teacher accountability (i.e. teacher effectiveness) in accountability requirements (e.g. Cochran-Smith, 2003; Ingersoll, 3 2003; Kraft, 2018). Interestingly, much of teacher accountability rests on the shoulders of student performance. This makes sense to some extent, given student performance is a reasonable outcome expected of teachers. However, there are many factors related to student performance that teachers have no control over, such as a student’s demographic and family background. In fact, researchers agree that while holding teachers accountable for student learning is important, it is equally, if not more important, that teacher accountability includes measures other than student performance, directly or indirectly (e.g. Mendro, 1998; Dinham & Scott, 2000; Cochran- Smith, 2003; Valli, Croninger, & Walter, 2007; Ballard & Bates, 2008). Teachers alone are not the only influence on student performance (Valli, Croninger, & Walters, 2007), and thus “[t]eachers and schools cannot reasonably be expected to solve problems over which they have little control nor capacity to deal with” (Dinham & Scott, 2000, p. 393). This might suggest that teachers should be held accountable for some factors that they are directly in control of, such as teacher attendance. Problem Statement and Research Question Recently, data collected through the Civil Rights Data Collection revealed U.S. teachers are absent from the classroom in a given school year at chronic rates, or 10 days or more (e.g. Thomas B. Fordham Institute, 2017; Kraft, 2018). This claim is further substantiated through anecdotal evidence in the news, additional reports, and academic studies, suggesting chronic teacher absenteeism is a real problem in U.S. schools (e.g. Finlayson, 2009; Hanover Research, 2012; Hui, 2017; Kraft; 2018; Miller, 2012; Miller, 2017; National Council on Teacher Quality, 2014; Thomas B. Fordham Institute, 2017; Wolcott, 2017). This problem is accentuated when the impact of teacher absenteeism on student performance is taken into consideration. Researchers agree that teacher absenteeism significantly negatively impacts student achievement 4 (e.g. Clotfelter, Ladd, & Vigdor, 2007; Miller, Murnane, & Willett, 2008; Finlayson, 2009; Goldhaber & Hansen, 2010; Roby, 2013), albeit this relationship is non-causal. If the U.S. strives to improve student achievement and if teacher absenteeism is such a well-documented problem with negative outcomes on U.S. student learning, then perhaps we should look deeper into why teachers are absent from the classroom. Furthermore, chronic teacher absenteeism is a national concern, but it is particularly concerning in Michigan, where a 2017 study found 24.7% of teachers in Michigan public schools are chronically absent (Thomas B. Fordham Institute, 2017). In other words, one quarter of Michigan public school teachers miss 10 or more days of work in a school year, meaning at least one quarter of Michigan’s public-school students are impacted by chronic teacher absenteeism. The challenges associated with teacher absenteeism are exacerbated when Michigan’s substitute teacher labor market is already severely strained, with one recent study finding the majority of Michigan school districts “unable to find enough substitutes multiple times a week” (Burroughs, Gardner, & Zuschlag, 2019, p. 2). This means that when teachers are absent and districts struggle to find subs, creative solutions are utilized, such as pulling school leaders and interventionists to oversee a class or combining classrooms (e.g. Burroughs, et al., 2019; Strunk, Cowen, Torres, Burns, Waldron, & Auletto, 2019). However, these solutions have echoing consequences because teachers and classrooms are overloaded, consistent instructional leadership is disrupted, and multiple important school functions are compromised (e.g. Burroughs, et al., 2019; Strunk, et al., 2019). When this sort of disruption of organizational resources occurs, it is possible that student achievement is impacted, such as through disruption to school improvement efforts. Local Michigan news has also drawn attention to the problem of teacher absenteeism (e.g. McVicar, 2017; The Detroit News, 2017; Wolcott, 2017), with an article as recent as March 5 2019 suggesting teacher absenteeism in Michigan is a problem worthy of attention (Renk, 2019). Given Michigan’s goal of becoming a top ten education state within ten years (Ackley, 2016), it might serve Michigan well to look further into teacher absenteeism. What these reports, news articles, and academic publications do not tell us is for what reasons teachers are absent. In order for Michigan to take any actions to address teacher absenteeism, it is necessary to first understand what drives teacher absenteeism behaviors. The current study seeks to address the dearth of understanding around teacher absenteeism through utilizing the following research question: • Research Question: What reasons do teachers provide for their chronic absenteeism? Purpose of the Study The above research question will provide insights to the timely topic of why Michigan teachers are absent at chronic rates and how teachers decide to be absent through using interviews to understand teachers’ realities. To make sense of the data collected as a result of this research question, it is important to use an existing theoretical framework grounded in research. The literature on teacher turnover in the education field and on organizational factors that impact employee absenteeism in the management field, such as organizational support, job stress, organizational commitment, and job satisfaction (and the influence of union membership), offer appropriate lenses to use to start making sense of teacher absenteeism and will aid in the interpretation and understanding of collected data. After all, teachers are employees in an organization where the school and district serve as the organization. Ultimately, this dissertation will call for a research stream that focuses on understanding teacher attendance behaviors in the U.S., which will provide insights into how to improve teacher attendance and the school contexts in which they teach. It is well documented that 6 teachers are an important school resource (e.g. Lankford, Loeb, & Wykoff, 2002; Corcoran, Evans, & Schwab, 2004; Chetty, Friedman, & Rockoff, 2012), but the question of what about teachers makes them so important is frequently debated. While some scholars argue that teacher preparation explains teacher effectiveness, as measured through student achievement (Darling- Hammond, Holtzman, Gatlin, & Vasquez Heilig, 2005), others argue the academic skills of teachers explain teacher effectiveness (Corcoran, Evans, & Schwab, 2004). Even others find that teacher certification impacts student achievement (Goldhaber & Brewer, 2000), only to be contradicted by scholars who argue that teacher certification matters little in teacher effectiveness (Kane, Rockoff, & Staiger, 2008). There are many additional ideas about which characteristics of teachers make them so important, and most of these ideas center around easy- to-measure qualities, such as teacher experience and certification (Goldhaber, Grout, & Huntington-Klein, 2014). Using easily measurable qualities is limiting because they exclude intangible characteristics (Goldhaber, 2002), such as support, stress, commitment, and satisfaction, that may be equally, if not more, important to explaining and understanding teacher behaviors. Additionally, these “elusive” (Goldhaber, 2002), intangible characteristics may have a relationship with both teacher absenteeism and teacher experience, for example. If this is the case, then the relationship between teacher absenteeism and teacher experience may be overstated. For example, a new teacher with little experience may have a high rate of absenteeism, but this same inexperienced teacher may also have very low job satisfaction. This might suggest that teacher experience is not the only contributing factor to the teacher’s absenteeism, but rather job satisfaction also plays a role. When job satisfaction is excluded from the equation, the impact of teacher experience on teacher absenteeism is overstated. These academic disagreements fuel the conversation around desirable teacher characteristics and offer 7 breadcrumbs in solving the mystery of how to support positive teacher behaviors and experiences that may improve teacher effectiveness. It is quite possible that this complex conversation around what characteristics make teachers so important cannot be solved with a singular answer. And, perhaps the conversation needs to consider the importance of teacher experiences in tandem with teacher characteristics. However, the disagreement in the educational academic community around the impact of school resources on student achievement and which teacher characteristics matter leaves the door open for additional research. This dissertation seeks to fill some of these voids by exploring contributing factors to teacher absenteeism, such as satisfaction and commitment, as self- reported by teachers. Given teachers must first be present in a classroom before they can impact student achievement, gaining a better understanding around why teachers are absent from their classrooms may provide insights to some of these elusive characteristics. Understanding teacher attendance behaviors certainly is not the only answer to the teacher-based mystery but having a better understanding of teacher attendance behaviors may help move the conversation forward. Study Significance Once the academic community gains a better understanding of the underlying reasons for teacher absenteeism, then local, state, and/or federal practices can be created and implemented to support or discourage teacher behaviors and experiences related to teacher attendance. Further, adding to the research on this topic through examining why teachers are absent may move this issue up the policy agenda, which could change how teacher attendance is emphasized in state and federal accountability and how teacher attendance data are collected. In theory, if teachers are the single most important input that impacts student achievement, then understanding what it is about teachers that makes them so important, such as their easy-to-measure and elusive 8 characteristics, and how to promote behaviors and experiences to improve teacher attendance, should be education policy research priorities. After a brief ideological statement, this study begins with a literature review that focuses on three themes. First, the literature review examines what is currently known about teacher absenteeism and reveals that pursuing additional research on teacher attendance behaviors is worthwhile, especially in the U.S. context and specifically around understanding the elusive, intangible characteristics that may have an impact on teacher absenteeism. Second, teacher turnover literature from the education field is reviewed and connections between teacher turnover behaviors and teacher absenteeism behaviors are evidenced. Finally, the literature review turns to research in the management field around employee absenteeism and focuses on the relationship between employee absenteeism and perceived organizational support, job stress, organizational commitment, and job satisfaction. Teacher turnover and employee absenteeism have been explored in multiple ways. For this dissertation I explore teacher absenteeism with a qualitative approach that uses interviews to facilitate the collection of data. This approach is further detailed in the methods section, which first dives into the epistemological perspective and research strategy. Then, the methods section provides details about the sample selection, participant overview, a statement about my positionality, steps for data collection, reliability and validity considerations, and my approach to data analysis and coding. After the methods section, I detail the findings that emerged from the data collected as a result of the guiding research question. The central findings of this study are that in situations of high job stress and/or low perceived organizational support, teachers’ mental health suffers and reasons for higher self-reported teacher absenteeism tend to be more job related than due to 9 personal circumstance. The antecedents to high job stress and/or low perceived organizational support are largely related to negative student behaviors, large class sizes, accountability pressures, lack of building support (e.g. principal), and lack of administration support (e.g. central office, district). Additionally, the secondary finding of this study is that chronically absent teachers’ view of and approach to using paid time off differs from non-chronically absent teachers, in that chronically absent teachers view paid time off as part of their total compensation and have a “use them or lose them” approach to using their paid days off because of feelings of being underpaid. This secondary finding relates to job satisfaction through feelings of satisfaction related to salary and compensation. The drivers behind salary and compensation dissatisfaction are similar to the drivers behind high job stress and low perceived organizational support and include feelings of increased workload, performing additional duties above and beyond teaching, and not having enough district financial support for classroom materials and resources. In short, organizational culture and climate issues impact chronic teacher absenteeism because of a combination of increased job stress, low organizational support, and salary dissatisfaction. Once the findings are illustrated, I offer a conclusion that includes a discussion, limitations, overall contribution, next steps, and opportunities for future research on teacher absenteeism. The concluding section also includes a brief discussion about the implications of this study for research, policy, and practice. Ideological Statement Before diving into the literature review, I want to briefly describe my ideology that is implicit to this study, so readers do not take my research out of context. Specifically, I believe teacher absenteeism is a problem in Michigan schools and I think there may be culture and climate issues in schools (i.e. related to organizational factors) that contribute to chronic teacher 10 absenteeism. For example, if there is a negative school climate, teachers may be more inclined to be absent. I want to learn why teachers are chronically absent so contributing factors can be better addressed. If, through this study, teachers suggest organizational factors contribute to their absenteeism, then perhaps efforts can be made to improve upon the organizational factors. In future endeavors, I would like to support the improvement of working environments for teachers, so they are more satisfied with their jobs and less tempted to be absent. In order to do this, I need to understand and be cognizant of why teachers are missing school in the first place. I provide a more detailed account of my ideology in a later section titled “Positionality Statement,” but it is important to mention my ideology at the start of this dissertation so the reader has a better understanding of my framing and motivations for this research, which certainly includes some bias. 11 LITERATURE REVIEW To better grasp what is and is not currently understood about teacher attendance, more specifically teacher absenteeism, a review of the existing literature on teacher absenteeism is warranted. Specifically, this literature review focuses on the prevalence of teacher absenteeism, its impact on student achievement, what is understood in the current literature around why teachers are absent, and what practices might reduce teacher absenteeism. Furthermore, this literature review will demonstrate that very little is known about teacher attendance behaviors, and thus a framework drawn from the management literature on employee absenteeism will be established through which teacher attendance can be viewed. Existing literature on teacher turnover will also inform the framework through which teacher absenteeism is examined. This review begins with an examination of national and international literature that focuses on teacher attendance behaviors and is divided into six sections based on emergent themes. Next, a discussion of the empirical patterns helps to highlight areas for additional contribution to the teacher absenteeism literature. Finally, a brief conclusion of the literature is offered. This is followed by a review of the education research on teacher turnover, which will provide some cues as to how to make sense of teacher absenteeism. After a review of the teacher turnover literature, I turn to the extensive research on employee absenteeism from the management literature in connection with teacher absenteeism. A discussion of union membership and employee absenteeism follows the review of the management literature. Then, a brief summary of the literature is offered and leads into a discussion around how the main findings of the literature review will inform research on teacher absenteeism. Taken together, this review will demonstrate that very little is known about teacher attendance behaviors and offers a lens through which teacher absenteeism can be examined. 12 Evidence of High Rates of Teacher Absenteeism Recent news headlines draw attention to the teacher attendance problems faced by schools and districts in the United States (U.S.), with article titles such as “Does your child’s teacher miss too much school” (Wolcott, 2017), “Being there matters: Tracking student & teacher attendance” (Miller, 2017), and “More than one-fifth of NC teachers are chronically absent. How does it affect your child?” (Hui, 2018). These headlines are substantiated by frequent needs for substitute teachers, as exemplified by the recurring call out for substitute teachers in all content areas, all grade levels, and at multiple districts via the Michigan Department of Education’s Critical Shortage list (MDE, 2017), for example. More importantly, these headlines are also corroborated by recent studies that focus on teacher absenteeism, as evidenced below. Research supports that there are high rates of teacher absenteeism domestically (e.g. Finlayson, 2009; Hanover Research, 2012; Kraft; 2018; Miller, 2012; National Council on Teacher Quality, 2014; Thomas B. Fordham Institute, 2017) and internationally (e.g. World Bank, 2004; Patel, Shah, & Lakhey, 2009; Rogers & Vegas, 2009; Patrinos, 2013; World Bank, 2013). In a study using national data from the 2009 Civil Rights Data Collection Survey, Miller (2012) found that over one third of U.S. teachers were absent 10 or more days during the 2009- 10 school year. Ten days absent is used as a comparison point because, on average, most U.S. teachers are allotted 10 days off throughout the approximately 180-day school year as part of their teacher benefits. Miller’s (2012) study indicates that teachers miss more than their allotted 10 days. Two other U.S. studies found similar results (National Council on Teacher Quality, 2014; Thomas B. Fordham Institute, 2017). In the National Council on Teacher Quality (2014) study, teacher attendance data from forty districts in some of the largest metropolitan areas in the 13 U.S. were collected for the 2012-13 school year. The study found that on average, teachers missed 10 or more days throughout the school year. Furthermore, the same study found that 16% of teachers missed 18 or more days (National Council on Teacher Quality, 2014). These findings were later confirmed in a study conducted by the Thomas B. Fordham Institute (2017), which found one in four U.S. teachers were absent from the classroom for 10 or more days throughout a typical school year. In fact, Hanover Research (2012) found rates of teacher absenteeism in the U.S. to be higher than employee absenteeism in other fields. These teacher absenteeism findings are not exclusive to the U.S. context, but rather similar results are found internationally. A World Bank survey on teacher absenteeism in developing countries found teacher absenteeism to be “a significant problem in many countries” (Patrinos, 2013, p. 70). The World Bank has conducted multiple studies on teacher absenteeism in developing counties over time (e.g. World Bank, 2004; World Bank, 2013), all coming to the same conclusion that there were high rates of teacher absenteeism in the countries being studied (Rogers & Vegas, 2009). There have been some efforts to gain a better understanding around where high rates of teacher absenteeism are more prevalent in both international and domestic contexts, but these studies have mixed findings (e.g. World Bank, 2004; Finlayson, 2009; Miller, 2012; National Council on Teacher Quality). For example, some studies found higher rates of teacher absenteeism at the elementary school level (e.g. World Bank, 2004; Miller, 2012), while others did not find this difference (National Council on Teacher Quality, 2014). Similarly, some domestic studies reported higher rates of teacher absenteeism in low-socioeconomic and urban areas (Finlayson, 2009), but others did not find this distinction (National Council on Teacher Quality, 2014). 14 While there have been numerous national and international studies focused on teacher absenteeism produced by research groups, policy groups, think tanks, and financial institutions, most of these studies are not peer reviewed because their target audience largely consists of non- academic professionals and practitioners. Further, these studies may not be peer reviewed because the publishers might have a stance or platform they are trying to advance. The peer review process can also be slow and cumbersome, which can interfere with timelines required by funders. The peer review process is important because it allows for collaboration and constructive feedback from other experts in similar research fields, which can result in sounder research and reduced bias. It also helps hold researchers accountable for strong methodological approaches. The findings on the prevalence of teacher absenteeism could be bolstered by additional peer-reviewed research on the topic; however, at minimum, the above literature suggests that teachers are absent more than expected, which is problematic internationally and in the U.S. context. The next section reviews literature that focuses on why teacher absenteeism is problematic. Teacher Absenteeism Negatively Impacts Student Achievement Teacher absenteeism is problematic because it significantly negatively impacts student achievement, although this relationship is not causal. Research focused at the U.S. district and state level (e.g. Clotfelter, Ladd, & Vigdor, 2007; Miller, Murnane, & Willett, 2008; Finlayson, 2009; Goldhaber & Hansen, 2010; Roby, 2013; National Council on Teacher Quality, 2014) and at the international level (e.g. Patrinos, 2013) all found that increased rates of teacher absenteeism were related to lower student achievement. The National Council on Teacher Quality (2014) found a “significant negative impact on student achievement in classrooms where the teacher is absent for ten days” (p. 14). Another U.S. study used a value-added model to 15 determine the impact of teacher absenteeism on teacher effectiveness (as measured through student performance, largely), and found a similar significant negative relationship (Goldhaber & Hansen, 2010). Specific to mathematics achievement, Miller, Murnane, and Willet (2008) estimated for every 10 days a teacher was absent, students’ mathematics performance was reduced by 3.3 percent of a standard deviation, and this was statistically significant. Yet another international study found that when a teacher was absent for 10 or more days, a primary school student could potentially lose up to one-quarter of a year of learning (Patrinos, 2013). All of the above studies only reported an association between teacher absenteeism and student achievement, rather than causation (e.g. Clotfelter, Ladd, & Vigdor, 2007; Miller, Murnane, & Willett, 2008; Finlayson, 2009; Goldhaber & Hansen, 2010; Roby, 2013; Patrinos, 2013; National Council on Teacher Quality, 2014), meaning that while there is evidence of a statistically significant negative relationship between teacher absenteeism and student achievement, there is no evidence to suggest that teacher absenteeism causes lower student achievement. For a causal relationship to be evidenced, a controlled study is necessary where two comparable groups are randomly assigned to a treatment and control. None of the studies mentioned in this review present causal evidence, and thus the present literature only evidences a statistically significant negative association between teacher absenteeism and student achievement. Also noteworthy is that some of these studies did not control for contextual variables, such school characteristics (e.g. urban, rural, suburban, size, etc.) and student characteristics (e.g. race/ethnicity, socioeconomic status, etc.) (Finlayson, 2009; Patrinos, 2013; National Council on Teacher Quality, 2014), while others did (Clotfelter, Ladd, & Vigdor, 2007; Miller, Murnane, & Willett, 2008; Goldhaber & Hansen, 2010; Roby, 2013). Contextual variables may be related to both teacher absenteeism and student achievement and cannot be 16 overlooked. The studies that controlled for contextual variables still found a negative relationship between teacher absenteeism and student achievement, but the studies that did not factor in contextual variables could be improved upon by including such controls. The significant negative relationship between teacher absenteeism and student achievement is well documented through peer-reviewed research, although this is not a causal relationship. If teacher absenteeism has the detrimental effect on student achievement as evidenced through these national and international studies, then further exploration of teacher attendance behaviors is warranted. The next section explores another theme that surfaced during the literature review, related to the costs of teacher absenteeism. Teacher Absenteeism is Expensive In addition to teacher absenteeism having a negative impact on student achievement, another consequence of teacher absenteeism is that it is expensive, and when teachers are absent more than expected, it is exponentially expensive (e.g. Ferris, Bergin, & Wayne, 1988; Miller, 2012). When teachers are absent, schools and districts incur significant economic costs to staff substitute teachers for the duration of their absence. While costs vary by schools and districts, individual districts can spend millions of dollars on costs associated with substitute teachers (Finlayson, 2009). The study conducted by the National Council on Teacher Quality (2014) on forty districts in large U.S. metropolitan areas found the combined forty districts spent $424 million on substitute teachers during the 2012-13 school year. When averaged out, that means a single district spent over ten million dollars on substitute teacher expenses in one year. Moreover, a Hanover Research (2012) study estimates the national costs associated with teacher absenteeism to be in the billions of dollars. Of course, U.S. schools and districts build this cost 17 into their budgets, but their budgets do not account for substitute teacher costs when teachers are absent more than expected. The high economic cost of teacher absenteeism is well documented in the literature, especially related to the economic costs in the U.S. context. Teachers are already considered one of the most expensive inputs in education, and their cost is only amplified when they are absent at high rates. Since teacher absenteeism is a great expense to districts and schools, perhaps better understanding underlying reasons that contribute to teacher absenteeism is important in order to target ways to reduce these costs. Not Much is Understood about Why Teachers are Absent Up until this point, research on the prevalence, problems, and costs of teacher absenteeism is well documented. Trying to understand why teachers are absent is more complicated because the research is limited and often focuses on easy-to-measure characteristics, such as school type (private/public) and teacher experience. For example, multiple studies find that private school teachers are absent less than their traditional public-school counterparts (e.g. Miller, 2012; Patrinos, 2013; Thomas B. Fordham Institute, 2017). This finding spans national and international borders. However, this finding does not speak to the underlying reasons as to why traditional public-school teachers are absent more often than private school teachers. Missing is the answer to the question: what about private school teachers make them less likely to be absent? One domestic study of a singular school district in Washington included teacher attendance as a variable in their exploration of teacher selection tools and found that more experienced teachers were less likely to be absent (Goldhaber, Grout, & Huntington-Klein, 2014). However, this study used an easy-to-measure characteristic, teacher experience, to explain 18 teacher attendance. This is problematic because there may be other underlying variables that are related to both teacher attendance and teacher experience that would bias this relationship. Put simply, easy-to-measure characteristics may not fully explain teacher attendance behaviors. More literature exists in the international context that speaks to why teachers are absent. For example, a study on teacher absenteeism in Andhra Pardesh, India found that teachers lacked accountability from parents and the community, therefore they did not consider their attendance as important (Patel, Shah, & Lakhey, 2009). Similar findings around how lack of accountability contributes to teacher absenteeism were evidenced in multiple additional international studies (World Bank, 2004; Rogers & Vegas, 2009; Word Bank, 2013). Other international studies found teacher management (Rogers & Vegas, 2009) and moonlighting (i.e. working other jobs) (Patrinos, 2013) as contributing factors to increased teacher absenteeism. Although there are some international findings related to why teachers are absent, the World Bank (2013) posits there is a need to better understand the causes of teacher absenteeism. The international findings, which wholly come from developing countries, cannot be generalized to the U.S. context because there are too many mediating factors that need to be considered. The domestic research on the causes of teacher absenteeism is few and far between, with most of the studies focusing on individual, special case schools and districts. Furthermore, the domestic studies tend to focus on easy-to-measure qualities, which do not offer many insights to the underlying reasons that teachers are absent. In general, better understanding why teachers are absent, especially in the U.S. context, is an area that needs further research. Teacher Absenteeism is Hard to Measure The main reason there is limited research on teacher attendance behaviors is because teacher absenteeism is hard to measure and collect data on (Hanover Research, 2012; Rogers & 19 Vegas, 2009). As explained in the National Council on Teacher Quality (2014) study, teacher attendance policies are created locally in the U.S. through a CBA, which means each policy is different. While there may be commonalities among the policies, implementation of these policies also varies. Furthermore, unlike student attendance, teacher attendance is not regulated in the U.S. This means that schools and districts are not required to submit teacher attendance data to a centralized authority, further complicating the measurement of teacher attendance and collection of data. Internationally, researchers face similar measurement and data collection challenges with understanding the underlying reasons around why teachers are absent (Rogers & Vegas, 2009). Available data have issues of reliability and validity because of local policy creation and implementation (Rogers & Vegas, 2009). Additionally, the World Bank (2013) found teacher attendance records to be misleading and misrepresented due to corruption. These challenges in measuring and collecting data on teacher absenteeism complicate the study of teacher attendance behaviors. Reducing Teacher Absenteeism Although the existing research on teacher absenteeism makes it clear that teacher absenteeism is problematic due to its negative association with student achievement and high financial costs, the research on reducing teacher absenteeism is less clear. Using regression analysis, Clotfelter, Ladd, and Vigdor (2007) found that associating a negative incentive (a $50 penalty per day) with sick days significantly reduced the additional sick days taken by teachers in North Carolina. These findings were later substantiated by a similar study (Ahn & Vigdor, 2011). Furthermore, Ehrenberg, Ehrenberg, Rees, & Ehrenberg’s (1989) study of school districts in New York found when teachers were given more sick days in their teacher contracts, they 20 were apt to use them up, meaning they were absent more. Hypothetically speaking, if teachers are offered fewer days off then they will be absent less, but this hypothetical is only relevant in conditions identical to those in the Ehrenberg, et al. (1989) study. In the same study, Ehrenberg, et al. (1989) found when teachers were offered the incentive of receiving a bonus for unused sick days (i.e. a buy-back policy), they were absent less. Other studies found significant reductions in teacher absenteeism when teachers were incentivized with additional retirement contributions (Hanover Research, 2012) and when job security improved (Jacob, 2013). While there are a handful of studies finding significant relationships between financial incentives and reducing teacher absenteeism, there are also studies that do not find this relationship (Hanover Research, 2012). For example, the National Council on Teacher Quality’s (2014) comprehensive study found that “[a]ttendance rates did not differ among districts with or without formal policies designed to encourage attendance” (p. 11). Similarly, the mainstream review of a buy-back policy implemented in a Florida school district found inconclusive results related to reducing teacher absenteeism, and the policy was ultimately abandoned (DeNardo, 2007). These mixed results suggest that the current literature on reducing teacher absenteeism is not exhaustive. One reason for these mixed results might be that little is understood about why teachers are absent in the first place, so policies and practices targeting the reduction of teacher absenteeism are misguided and misinformed. For example, the studies mentioned above tie teacher absenteeism to financial gain or loss, but perhaps teacher absenteeism is not as financially motivated as one might think. There may be reasons for teacher absenteeism that cannot be finically incentivized, and thus learning more about the underlying reasons for teacher 21 absenteeism is crucial before policies and practices can be designed and implemented to reduce teacher absenteeism. Teacher Turnover As demonstrated in the previous pages, teacher absenteeism is problematic because it negatively impacts student achievement and is expensive, and yet, very little is known about the underlying reasons for teacher absenteeism. Teacher turnover is also problematic because it similarly has a negative association with student achievement and is expensive. Unlike teacher absenteeism, research on teacher turnover is thoroughly executed in the education literature, which positions it to possibly inform the present study on teacher absenteeism. The research on teacher turnover is comprehensive because of previous urgent claims of teacher shortages and staffing issues (e.g. Ingersoll, 2002; Ingersoll, 2003; Ingersoll & Smith, 2003; Ingersoll, 2011) and because of the strong evidence supporting that “teacher turnover has a significant and negative impact on student achievement in both math and ELA” (p. 30), even when teacher quality is controlled for (e.g. Ronfeldt, Loeb, & Wyckoff, 2013; Hanushek, Rivkin, & Schiman, 2016). The negative impact of teacher turnover on student achievement is especially salient for already disadvantaged students, such as high poverty students and minority students (e.g. Ronfeldt, Loeb, & Wyckoff, 2013; Hanushek, Rivkin, & Schiman, 2016). This relationship between teacher turnover and student achievement is similar to the negative relationship between teacher absenteeism and student achievement, which makes the literature on teacher turnover relevant to the study of teacher absenteeism. Similar to the demonstrated relationship between teacher absenteeism and student achievement, the relationship between teacher turnover and student achievement is not causal because the existing studies are not controlled studies where two comparable groups are 22 randomly assigned to a treatment and control. While there is evidence of a statistically significant negative relationship between teacher turnover and student achievement, there is no evidence to suggest that teacher turnover causes lower student achievement. The existing literature only evidences a statistically significant negative association between teacher turnover and student achievement. Even though this relationship is not causal, turning to research on teacher turnover may provide insights for researching teacher absenteeism because they share the goal of trying to understand how to keep teachers in classrooms. When increased rates of teacher turnover and teacher shortages became a real-life problem in the U.S., Richard Ingersoll, an education researcher, set out to investigate whether there was a teacher shortage and, if there was, what contributed to school staffing issues (e.g. Ingersoll, 2002; Ingersoll, 2003; Ingersoll & Smith, 2003; Ingersoll, 2011). This is similar to the purpose of the study at hand, which seeks to understand what contributes to issues with teacher attendance. Using data from the U.S. Department of Education’s National Center for Education Statistics (NCES)’s Schools and Staffing Survey (SASS), Ingersoll (2011) finds that while easy- to-measure teacher characteristics, such as age and specialty area, are related to teacher turnover, these characteristics do not fully explain teacher turnover behaviors. Ingersoll’s (2011) study goes on to find that “net of the effects of these teacher characteristics, there are also significant effects of school and organizational characteristics on turnover which have largely been overlooked by previous theory and research” (p. 524). Namely, job satisfaction, or job dissatisfaction, and organizational climate are two of the main contributing factors to teacher turnover (e.g. Ingersoll & Smith, 2003; Ingersoll, 2011), although these relationships are not causal. When probed further, the majority of the participants who indicated their turnover was 23 attributed to job dissatisfaction also specified that they were dissatisfied at their job because of characteristics related to organizational climate, such as an unsupportive school administration and lack of autonomy in decision making (Ingersoll, 2011). Although some participants attributed their job dissatisfaction to low compensation, Ingersoll (2011) is quick to point out that compensation alone was not the only antecedent to job dissatisfaction and subsequent turnover. Earlier studies by Ingersoll similarly found teacher turnover to be significantly linked to job dissatisfaction, resulting from organizational characteristics, along with salary concerns (e.g. Ingersoll, 2002; Ingersoll, 2003; Ingersoll & Smith, 2003). This is akin to an idea presented earlier about teacher absenteeism behaviors not easily motivated by financial incentives. Similar to how teacher turnover is related to organizational climate and job satisfaction, perhaps teacher absenteeism is motivated by elusive characteristics. Ingersoll is not the only education researcher to explore the antecedents of teacher turnover. International (e.g. Dinham & Scott, 1997; Dinham & Scott, 2000) and domestic (e.g. Anderman, Belzer, & Smith, 1991; Grissom, 2011; Dou, Devos, & Valcke, 2017; Dunn, 2018) studies alike consistently find school culture and climate to be directly related, although not causally related, to job satisfaction and thus teacher turnover. Dunn’s (2018) study of teachers’ public resignation letters revealed teacher turnover to be a form of resistance in response to a culture and climate that suppressed teacher agency. In other words, teachers left their jobs because they were dissatisfied with the culture and climate of the school and/or district. Some of the other domestic studies also provide significant evidence that teachers are less satisfied and more likely to turnover in high poverty and high minority schools (e.g. Anderman, Belzer, & Smith, 1991; Grissom, 2011). Another finding from the domestic studies is that principals are drivers of a positive and supportive school organizational culture, and thus positive principal 24 leadership appears to be linked to increased teacher satisfaction, increased teacher commitment, and reduced teacher turnover (e.g. Anderman, Belzer, & Smith, 1991; Grissom, 2011; Dou, Devos, & Valcke, 2017). In sum, the “the working and organizational conditions of teaching are, together, the most prominent source of turnover” (Ingersoll, 2002, p. 26) because poor organizational conditions result in job dissatisfaction (e.g. Anderman, Belzer, & Smith, 1991; Dinham & Scott, 1997; Dinham & Scott, 2000; Ingersoll, 2002; Ingersoll, 2003; Ingersoll & Smith, 2003; Grissom, 2011; Ingersoll, 2011; Dou, Devos, & Valcke, 2017; Dunn, 2018). The organizational qualities that contribute to job satisfaction may be considered elusive qualities that are challenging to measure. Current policy that targets teacher turnover is mainly concerned with teacher supply and largely overlooks organizational factors that contribute to job satisfaction, which suggests that teacher turnover problems will not be resolved until these elusive characteristics are considered through policy targeted at improving the quality of teacher’s working and organizational conditions (e.g. Ingersoll, 2002; Ingersoll, 2003; Ingersoll & Smith, 2003; Ingersoll, 2011). Organizational characteristics that are associated with job satisfaction in relation to teacher turnover may also have a connection to teacher absenteeism because the problems of teacher turnover and teacher absenteeism are both concerned with having teachers be present in the classroom (i.e. withdrawal behaviors). Accordingly, considering organizational characteristics and job satisfaction in the current study of teacher absenteeism may offer acumen to the problem of teacher absenteeism. However, these two research streams diverge and cannot be considered one in the same. The problems of teacher turnover and teacher absenteeism are related in that they have a similar negative association with student performance as evidenced 25 through academic research and the research streams share the goal of trying to understand how to keep teachers in classrooms. Yet, teacher turnover and teacher absenteeism are distinct behaviors because the ultimate intention of the teacher is different. With teacher turnover, the teacher’s intent is to leave the school permanently; with teacher absenteeism, the teacher’s intent is to return to their same classroom. In other words, turnover is final, whereas absenteeism temporary. For this reason, relying solely on a lens from the teacher turnover literature to make meaning of teacher absenteeism is too narrow of an approach because important factors that are unique to teacher absenteeism may be overlooked. Therefore, it is necessary to draw from another discipline to further inform the study of teacher absenteeism. Specifically, literature on employee absenteeism from the management field will offer additional angles to consider when exploring teacher absenteeism, as demonstrated in the next sections. Absenteeism and the Management Literature Although little is known about teacher absenteeism behaviors specifically, it is possible that a framework from the management literature could contribute to the understanding of teacher absenteeism, such as one that examines employee absenteeism, more broadly. After all, teachers are a particular type of employee. The prevalence of employee absenteeism has made it an important and well-researched topic in the management literature (Ferris, Bergin, & Wayne, 1988). As such, the remainder of this literature review sets out to contribute to the research on teacher absenteeism by using an interpretive lens from the management literature on employee absenteeism. This management lens offers a possible explanatory framework through which we can better understand why teachers are absent. While these management frameworks may help to understand teacher absenteeism, the fit is not necessarily perfect because it can be challenging to assimilate between two distinct disciplines, which is another idea explored below. Although the 26 management frameworks do not map perfectly onto the education field, there are enough common threads between the two to deem using the management lens appropriate for interpreting teacher absenteeism. Ultimately, this section will offer a lens through which teacher absenteeism can be examined. “Employee absenteeism represents one of the most persistent and costly problems that managers face in organizations” (Ferris, Bergin, & Wayne, 1988, p. 561), and thus the management literature offers many peer-reviewed studies on the absenteeism behaviors of employees in an organization (e.g. Mowday, Porter, & Steers, 1982; Allen, 1984; Leigh, 1986; George, 1989; Eisenberger, Fasolo, & Davis-LaMastro, 1990; Leigh, 1991; Gellatly, 1995; Sagie, 1998). Teachers are employees in an organization where the school district operates as their organization, so using a framework that examines employee absenteeism can easily be applied to a particular type of employee, teachers. Generally, the management literature examines employee absenteeism behaviors from a variety of institutions (e.g. private, public, manufacturing, government, health, financial, education, etc.) and from individuals who work different types of jobs (e.g. clerical, technical, sales, management, medical, professional, etc.) (e.g. Allen, 1984; Leigh, 1986; George, 1989; Eisenberger, Fasolo, & Davis-LaMastro, 1990; Leigh, 1991; Gellatly, 1995; Sagie, 1998), but a focused look at teacher absenteeism through the management lens has not been done in either the management literature or the education literature. To clarify, teachers have been included in previous management studies that examine employee absenteeism (e.g. Leigh, 1991; Eisenberger, Huntington, Hutchison, & Sowa, 1986), but the study and/or sample focus was not exclusively teachers. In other words, the emphasis of these studies was employee absenteeism, not teacher absenteeism. Employee absenteeism is well researched and peer-reviewed in the management discipline, and the research seems to fall into 27 three loosely defined categories: examining how demographic characteristics relate to employee absenteeism, examining how psychological factors relate to employee absenteeism, and examining how organizational factors relate to employee absenteeism. The non-causal relationship between employee absenteeism and demographic characteristics, such as gender, age, marital status, race/ethnicity, and education, is widely supported in the management literature (e.g. Paringer, 1983; Leigh, 1986; Garcia, 1987; Leigh, 1991; Gellatly, 1995). Similarly, the non-causal relationship between employee absenteeism and psychological factors, such as self-efficacy, mood, and depression, is widely supported in the management literature (e.g. Frayne & Latham, 1987; George, 1989; Avey, Patera, & West, 2006). While important to the overall study of employee absenteeism, demographic and psychological characteristics are not the focus of this study because I believe these factors are less important than the organizational factors that motivate teachers to be absent. Moreover, there is likely an interaction between demographic, psychological, and organizational characteristics and their impact on teacher absenteeism as evidenced through the management literature (e.g. Garcia, 1987; Leigh, 1986; Leigh, 1991; Gellatly, 1995), but the purpose of this study is specifically to better understand the relationship between organizational factors and teacher absenteeism. While I do not have any concrete personal experiences to justify this interpretation, my experiences in working with many teachers over the past seven years and my knowledge and understanding from extensive reading, research, and coursework suggest organizational factors play a large, influencing role in a teacher’s decision to be absent, and thus should be empirically studied. This speaks to my ideology in that I believe teacher absenteeism is a problem in Michigan schools and I think there may be culture and climate issues in schools 28 (i.e. related to organizational factors) that contribute to chronic teacher absenteeism, which is what motivates the study at hand. Organizational Factors and Employee Absenteeism The management literature focuses on four connected influences to explain the non- causal relationship between organizational factors and employee absenteeism: perceived organizational support, job stress, organizational commitment, and job satisfaction (e.g. Freeman 1977; Muchinsky, 1977; Beehr & Newman 1978; Steers & Rhodes, 1978; Allen, 1981; Scott & Taylor, 1985; Eisenberger, et al., 1986; Hackett, 1989; Eisenberger, et al., 1990; Mathieu & Zajac, 1990; Chaudhury & Ng, 1992; Gellatly, 1995; Sagie, 1998; Meyer, Stanley, Herscovitch, & Topolnytsky, 2002). These four organizational factors may offer a lens through which teacher absenteeism can be explained. First, perceived organizational support is the “extent to which employees perceived that the organization valued their contribution and cared about their well- being” (Eisenberger, et al., 1990, p. 52), and has been empirically linked to employee absenteeism (Eisenberger, et al., 1986; Eisenberger, et al., 1990), although these studies do not indicate causation. For example, using a survey measuring perceived organizational support, Eisenberger, et al. (1986) found that “perceived organizational support reduces absenteeism” (p. 504). In a later study of brokerage firm clerks, manufacturing hourly workers, insurance representatives, high school teachers, university resident assistants, and police officers, Eisenberger, et al. (1990) found employees with higher scores on a perceived organizational support survey also had significantly higher rates of attendance. In one of the occupations, employees with lowest scores on the perceived organizational support survey had rates of absenteeism double that of those with the highest scores on the perceived organizational support survey (Eisenberger, et al., 1990). As such, one possible explanation for why there are increased 29 rates of teacher absenteeism is because teachers do not feel like they are supported and cared for by their school and/or district. This idea warrants further exploration. Second, job stress has been non-causally linked to employee absenteeism in the management literature, which finds that an increase in job stress is associated with an increase in employee absenteeism (e.g. Beehr & Newman, 1978; Gupta & Beehr, 1979). Gupta and Beehr (1979) empirically studied the impact of job stress on withdrawal behaviors in over 600 employees from Midwest organizations. Although this was not a controlled study with two comparable groups randomly assigned to a treatment and control, they still found a significant positive relationship between job stress and withdrawal behaviors, such as employee absenteeism (Gupta & Beehr, 1979). This suggests that increased rates of teacher absenteeism might be attributable to increased job stress. Third, the management literature is saturated with research that examines the relationship between organizational commitment and withdrawal behaviors, such as employee absenteeism (e.g. Mowday, Porter, & Steers, 1982; Clegg, 1983; Eisenberger, et al., 1990; Mathieu & Zajac, 1990; Gellatly, 1995; Sagie, 1998; Meyer, et al., 2002; Klein, Cooper, Molloy, & Swanson, 2014). Sagie (1998) used an organizational commitment questionnaire to study the differences between voluntary (such as certified illness) and involuntary (such as vacation, uncertified illness) absence from work, where “voluntary rather than involuntary absences from work may reflect job dissatisfaction and lack of commitment to the organization” (p. 157). Using regression analysis, Sagie (1998) found voluntary absence was strongly related to organizational commitment, in that employees with lower scores on the organizational commitment questionnaire had more instances of voluntary absence. Gellatly’s (1995) research echoes these findings. By parsing out organizational commitment into affective (emotional attachment, 30 identification, and involvement with the organization) and continuance (perceived costs associated with leaving an organization) commitment, he found a significant negative, non- causal relationship between affective commitment and employee absenteeism (Gellatly, 1995). Similarly, Meyer, et al.’s (2002) pivotal study on organizational commitment found a negative correlation between affective commitment and employee absenteeism, thus reaffirming that employees with higher commitment to the organization are less likely to be absent from work. Furthermore, in Mathieu and Zajac’s (1990) meta-analysis of organizational commitment, they found substantial support for organizational commitment being an antecedent of employee absenteeism. Organizational commitment clearly has a non-causal negative relationship with employee absenteeism, which might suggest that teachers with low organizational commitment to their district and/or school will report higher rates of absenteeism. At minimum, the literature on organizational commitment may contribute to the framework in which teacher absenteeism behaviors are explained. Finally, job satisfaction is another organizational factor that is widely studied in the management literature, which finds employee absenteeism to have an inverse, yet non-casual, relationship with job satisfaction (e.g. Muchinsky, 1977; Steers & Rhodes, 1978; Clegg, 1983; Scott & Taylor, 1985; Garcia, 1987; Hackett, 1989; Chaudhury & Ng, 1992; Sagie, 1998). Job satisfaction with relation to employee absenteeism is so widely studied that multiple literature reviews and meta-analyses exist to organize the literature. Muchinsky’s (1977) literature review on employee absenteeism specifically focused on the relationship between employee absenteeism and job satisfaction and concluded that there is empirical support for a negative relationship between the two variables. Of the fifteen studies he reviewed that have a specific focus on job satisfaction and employee absenteeism, twelve conclusively reported a negative 31 relationship between job satisfaction and employee absenteeism (Muchinsky, 1977). The remaining three studies had mixed results, which Muchinsky (1977) attributes to validity and reliability concerns with the index used to measure job satisfaction. Muchinsky (1977) concludes “satisfaction with work has been found to have a consistent negative relationship to absenteeism” (p. 326). Scott and Taylor (1985) also conducted a meta-analysis to determine if there is a relationship between job satisfaction and absenteeism. After a review of the exiting literature, the authors found a non-causal, significant negative relationship between absenteeism and job satisfaction (Scott & Taylor 1985). They indicated that studies in which a negative relationship was not found also had small sample sizes, and thus compromised the studies’ validity (Scott & Taylor, 1985). Finally, Hackett (1989) also conducted a review of literature on employee absenteeism and job satisfaction and found similar results to Muchinsky (1977) and Scott and Taylor (1985). He reviewed three meta-analyses, each with slightly different findings, and concluded that the empirical literature widely suggests a negative association between employee absenteeism and job satisfaction (Hackett, 1989). Turning to empirical studies, in Garcia’s (1987) study of employee absenteeism and job satisfaction in sample of over 600 employees, he used cross tabulations and Chi Square tests to identify a significant negative relationship between employee absenteeism and job satisfaction, although this relationship was not causal. Similarly, in Sagie’s (1998) empirical study of clerical workers, he found that employees with higher reported job satisfaction had better attendance rates compared to those employees who were not satisfied with their work. Taken together, the management literature supports there is a non-causal negative relationship between job satisfaction and employee absenteeism, and thus, using a job satisfaction lens may help explain teacher absenteeism behaviors. 32 Although perceived organizational support, job stress, organizational commitment, and job satisfaction are discussed separately in the above review, these four organizational factors are certainly related and interact with one another (e.g. Clegg, 1983; Sagie, 1998). In sum, the management literature indicates an inverse association between employee absenteeism and perceived organizational support, organizational commitment, and job satisfaction; and a positive association between employee absenteeism and job stress. However, the studies from the management literature reviewed above do not exclusively focus on teachers, but rather focus on employees in general. This reveals an opportunity to study a subset of employees to determine how these organizational factors impact a group of employees that share similar job characteristics. Simultaneously, these organizational factors may offer a lens through which teacher attendance behaviors can be better understood. Coupling these ideas around the impact of organizational factors on employee absenteeism with the findings from the teacher turnover literature that suggest organizational factors are related to teacher turnover makes this lens from the management literature relevant to educational settings and the study of teacher absenteeism. Unions and Employee Absenteeism So far, the strongest connection between the study of employee absenteeism and the study of teacher absenteeism is that teachers are a particular type of employee. Therefore, it seems appropriate to use a management lens to examine teacher absenteeism, specifically. While the management literature examines employee absenteeism broadly, there are many studies that include another important factor: unions (e.g. Freeman, 1977; Borjas, 1979; Smith & Hopkins, 1979; Allen, 1981; Gordon & Long, 1981; Allen, 1984; Leigh, 1986; Chaudhury & Ng, 1992). Albeit dated, these studies are important to consider because they offer an additional connection between the management literature and education field. This is important in the present study 33 because the data are collected in Michigan, which has a strong teacher union presence (Winkler, Scull, & Zeehandelaar, 2012). Using a framework that has been applied with a specific focus on employee unions makes the framework especially transferrable to education because of the strong presence of unions in education, which may influence teacher absenteeism. During the late 1970s through early 1990s, many studies in the management literature not only focused on unions, but also focused on the relationship between unions and employee absenteeism (e.g. Allen, 1981; Allen, 1984; Leigh, 1986; Chaudhury & Ng, 1992) and the relationship between unions and job satisfaction (e.g. Freeman, 1977; Borjas, 1979; Smith & Hopkins, 1979; Gordon & Long, 1981; Allen, 1984). The findings from these studies may offer additional insights as to how to approach research on teacher absenteeism. For example, Allen (1981) tested an empirical model of attendance behaviors on a sample of individuals who completed the Quality of Employment Survey and he found a significantly higher rate of absenteeism among union members compared to non-union members. In a later study, Allen (1984) used three existing data sets with large sample sizes to “examine the effect of union membership on absenteeism” (p. 331). Even after controlling for relevant variables, such as age, gender, and marital status, Allen (1984) found that absence rates were significantly higher in union workers, which is problematic because absenteeism “can reduce productivity because it disrupts production plans and requires the increased use of substitute workers, who are less efficient than the workers they replace” (p. 331). Teacher absenteeism can have this same negative impact because “[s]ubstitute teachers cannot be expected to equal the quality of instruction provided by the regular classroom teacher” (Ferris, Bergin, & Wayne, 1988, p. 561). This relationship is echoed again in the management field by Leigh (1986), who found a non- causal positive relationship between having a union contract and employee absenteeism. 34 Similarly, Chaudhury and Ng (1992) used Ordinary Least Squares regression to determine the relationship between unionization and employee absenteeism as captured through a survey administered to organizations with both union and non-union membership. They found the relationship between unionization and absenteeism was not causal but was significant and positive, and they speculated this was the case because union members generally have more protection and fewer/less serious sanctions for being absent (Chaudhury & Ng, 1992). None of these studies focus on teachers as part of their sample population; however, since many teachers belong to unions, the research in the management literature on unionization and employee absenteeism is relevant to the present study on teacher absenteeism. It is important to note that all teachers in Michigan public schools work under the agreed upon union contract, regardless of if they are a union member. This is not the case of the employees mentioned in the above studies, which suggests there may be inherent organizational differences in schools compared to the organizations studied above. These differences could influence the findings specific to teacher absenteeism. In addition, there is evidence in the management literature that union membership is inversely related to job satisfaction (e.g. Freeman, 1977; Borjas, 1979; Smith & Hopkins, 1979; Gordon & Long, 1981; Allen, 1984). Earlier in this literature review, I established that the existing literature on job satisfaction may contribute to a better understanding of teacher absenteeism because of the negative relationship between job satisfaction and employee absenteeism, even though this relationship is not causal. Since many teachers belong to unions, it is worthwhile to examine the relationship between unionization and job satisfaction, which, in turn, impacts absenteeism. One of the first studies on union membership and job satisfaction comes from Freeman (1977), who found that job satisfaction was significantly lower among 35 union members. Borjas (1979) had similar findings when his empirical analysis uncovered that “on average, union members report significantly lower levels of job satisfaction” (original emphasis, p. 38). Likewise, in Chaudhury and Ng’s (1992) aforementioned study, they concluded that the reason for a higher rate of absenteeism in union members may be due to their low job satisfaction, as compared to non-union members, although these findings do not suggest causation. Again, there may be inherent organizational differences in schools compared to the organizations studied above. For example, these studies are quite dated and took place during a time when hostile work environments were more prevalent and demanded more protection for employees, especially in certain industries such as manufacturing. For this reason, the organizational environments in the dated studies above could be quite different from today’s working environments for teachers, so these findings need to be referenced with caution. While I need to heed these cautions, it is noteworthy that many of the samples of the studies mentioned in the section above titled “Organizational Factors and Employee Absenteeism” included unionized employees as part of their sample (e.g. Beehr & Newman 1978; Allen, 1981; Scott & Taylor, 1985; Eisenberger, et al., 1986; Eisenberger, et al., 1990; Mathieu & Zajac, 1990; Chaudhury & Ng, 1992; Gellatly, 1995; Sagie, 1998; Meyer, Stanley, Herscovitch, & Topolnytsky, 2002), further supporting the common thread of union-membership between the management and education disciplines. Research on unions has waned since the early 1990s, but unions are still present and strong in certain professions, such as teaching. The above studies suggest there is an association between unions and job satisfaction and an association between unions and employee absenteeism. Since these non-causal associations exist in the literature and many teachers are union members, then perhaps an interpretive lens that factors in union membership should be considered at minimum for a study on teacher 36 absenteeism, even though there are potential inherent differences between the organizational settings of the research above and the current organizational environments for teachers. Challenges with Translating Management Frameworks to Education While it appears that frameworks from the management literature transfer well to education, it is not a perfect fit. One reason the translation between education and management might be poor is because many of the management studies mentioned above are quite old, especially the studies that hone in on unions (e.g. Freeman, 1977; Beehr & Newman 1978; Borjas, 1979; Smith & Hopkins, 1979; Allen, 1981; Gordon & Long, 1981; Allen, 1984; Scott & Taylor, 1985; Eisenberger, et al., 1986; Leigh, 1986; Chaudhury & Ng, 1992). This poses a potential problem because unions may have evolved over the past thirty years, meaning the findings from the management studies may be no longer valid. As mentioned in the previous section, the studies that focus on unions took place during a time that emphasized productivity at all costs, meaning employees might not have had favorable working conditions, especially in certain industries. Working conditions have since progressed, which might invalidate the applicability of the studies on unions and employee absenteeism in understanding teacher attendance behaviors. Similarly, many of the union studies focused on the manufacturing (e.g. Borjas, 1979; Allen, 1984; Eisenberger, et al., 1986) and healthcare industries (e.g. Chaudhury & Ng, 1992), which could have integral differences from the education industry. At the very least, the management frameworks mentioned above can serve as a starting point for examining teacher absenteeism. Additionally, the time gap between when many of the studies were published and present day highlights a hole in the research on employee absenteeism and union members, in general, one which the present study can take steps to fill in. 37 A second reason a framework from the management literature might pose a challenge in its transfer to an education phenomenon is because education is part of the public sector (largely) and many of the studies mentioned above took place in the private sector (e.g. Gupta & Beehr, 1979; Fitzgibbons & Moch, 1980; Clegg, 1983; Allen, 1984; George, 1989). Individuals in the public sector compared to the private sector are subject to different forms of governance, laws, regulations, etc. For example, the teaching profession (i.e. public sector employees) is subject to many laws and regulations set forth by federal, state, and local governments. Conversely, private sector employees are subject to the rules and regulations determined by the managing organization, which may not be as strict as laws and regulations that public employees must adhere to. This means that the strictness of laws and regulations that teachers face might not be a variable considered in the management literature, but they certainly could contribute to teacher absenteeism. These challenges related to why frameworks from the management literature may not transfer well to education are allayed through one main strategy: calling attention to these potential biases and documenting them in my write up/report (e.g. Merriam & Tisdell, 2016), as further addressed in the methods section. Another point worthy of consideration is related to the methods used in many of the management studies mentioned above. Many of the above studies are quantitative and use survey methods (e.g. Allen, 1981; Eisenberger, et al., 1986; Eisenberger, et al., 1990; Chaudhury & Ng, 1992; Gellatly, 1995; Meyer, et al., 2002). In the methods section that follows the literature review, I detail my plan to conduct a basic qualitative study using interviews to gain a better understanding of the underlying reasons for teacher absenteeism. This approach is dissimilar from the bulk of the research in the management literature on employee absenteeism. However, since there are so few studies on the underlying reasons for teacher 38 absenteeism, it is first necessary to understand why teachers are absent. When the goal of a study is to understand a particular phenomenon through the collection of rich, detailed, and descriptive data, a qualitative research design is best (Maxwell, 2013). The management literature can then offer a lens through which teacher absenteeism is interpreted. Using survey methods that already have a lens built in narrows the focus of the data collected, which may result in an incomplete understanding. Once there is a clearer understanding of teacher absenteeism (aided by teacher turnover and employee absenteeism interpretive lenses), then perhaps a study that uses methods similar to the studies in the management literature (i.e. survey methods) can be conducted in the future. For the time being, this study on teacher absenteeism will use interview methods for the purpose of gaining a better understanding of teacher absenteeism. In sum, the literature on organizational factors that impact employee absenteeism in the management field offers an appropriate lens to use to start making sense of teacher absenteeism. Given that employee absenteeism is extensively studied in the management literature and teachers are a type of employee, using a framework from the management literature on employee absenteeism is suitable. While there are multiple frameworks available in the management literature on employee absenteeism, the ones of particular interest concern the organizational factors that motivate the decisions of employees to be absent from work, including perceived organizational support, job stress, organizational commitment, and job satisfaction. In turn, these four organizational factors may help explain teacher absenteeism behaviors. Additionally, it would be an omission to not include literature on the relationship between union membership and employee absenteeism because teacher unions have a strong presence in the state where this study takes place, which is why this area is drawn from to help interpret teacher absenteeism. While the frameworks in the management field do not perfectly transfer to the education field 39 and the study of teacher absenteeism, they can at least serve as a starting place for understanding the underlying behaviors that motivate teachers to be absent. Summary of the Literature Review The purpose of this literature review was two-fold. First, this review set out to assess the existing literature around teacher absenteeism and identify any themes. Second, this review examined the existing literature on teacher turnover in the education discipline and employee absenteeism in the management discipline to identify possible frameworks that can be applied to help understand teacher attendance behaviors. With regards to the first point, this literature review revealed six main themes from the exiting literature on teacher absenteeism. First, there is ample evidence to support that there are high rates of teacher absenteeism nationally and internationally. Existing research firmly purports that teachers are absent more than they are expected to be; however, much of this research is not peer reviewed. Second, teacher absenteeism is problematic because existing national and international research shows that teacher absenteeism negatively impacts student achievement, although this relationship is not causal. Third, teacher absenteeism is also problematic because it is expensive, as evidenced through numerous national and international studies. When teachers are absent more than expected, schools and districts incur high costs associated with staffing substitute teachers. Fourth, exiting research that focuses on the underlying reasons that explain teacher absenteeism is limited in scope and quantity, especially in the U.S. context. Fifth, current research acknowledges the need for understanding teacher attendance behaviors but also identifies the challenges with measuring and collecting data on teacher absenteeism. Namely, the challenges associated with measuring and collecting data on teacher attendance behaviors stem from local creation and implementation of teacher attendance policies domestically (through a CBA), and 40 corruption internationally. Finally, there are inconclusive research findings related to what practices might reduce teacher absenteeism, which suggests a general lack of understanding around why teachers are absent. Two of these six themes warrant further elaboration. First, it was important to establish that teacher absenteeism is problematic. This is well addressed through existing research, with many national and international studies finding a non- causal significant negative relationship between teacher absences and student achievement. Some findings related to this question were specific to different contexts, such as empirical findings that students’ math achievement was inversely associated with an increase in teacher absences increased (Miller, Murnane, & Willet, 2008); or results concerning the negative association on primary students’ academic achievement with an increase in teacher absences (Patrinos, 2013). However, most of the studies found a non-causal, negative association with student achievement specifically when teachers were absent for 10 or more days (e.g. Miller, Murnane, & Willet, 2008; Patrinos, 2013; National Council on Teacher Quality, 2014). These unambiguous findings confirm that teacher absenteeism is problematic because it is negatively associated with student achievement. Second, it was important to understand what the current literature says about why teachers are absent. In short, there are few studies that examine the underlying reasons why teachers are absent, and the studies that do exist provide little insight to answering this question. The handful of existing international studies focused on this topic suggest teachers are not held accountable for their attendance, and therefore attendance is not a priority (e.g. World Bank, 2004; Patel, Shah, & Lakhey, 2009; Rogers & Vegas, 2009; Word Bank, 2013). There are even fewer domestic studies that examine the underlying reasons for teacher absenteeism. The empirical pattern of the existing studies is that they focus on easily measurable teacher 41 characteristics (i.e. school type, teacher experience), such as Goldhaber, Grout, and Huntington- Klein’s (2014) study that found more experienced teachers were associated with lower absenteeism rates. Taken together, not only is it clear that teacher absenteeism is problematic as evidenced through substantial existing literature, but the underlying behaviors that explain teacher absenteeism are not well understood because existing literature is limited in quantity and scope. If teacher absenteeism has a significant negative association with student achievement and if teacher absenteeism is related to significant financial costs, then perhaps gaining a better understanding of the reasons why teachers are absent will eventually lead to solutions to address teacher absenteeism. As such, this review of literature also informs the literature on teacher absenteeism by identifying what is missing from the current research on this topic. While some international research suggests lack of teacher accountability accounts for increased teacher absenteeism (e.g. World Bank, 2004; Patel, Shah, & Lakhey, 2009; Rogers & Vegas, 2009; Word Bank, 2013), these findings cannot be generalized to the U.S. context. The findings from the international studies included in this literature review are specific to developing countries. There are too many contextual differences, such as economic, political, and cultural differences, between the U.S. and developing countries, which limits the applicability of the international findings. However, examining the role of accountability in understanding teacher attendance behaviors in the U.S. is certainly an area of interest that could be explored by future research. In general, better understanding teacher attendance behaviors in the U.S. is a research stream demanding further attention. 42 Furthermore, the few U.S. studies that do exist related to understanding teacher absenteeism focus on easy-to-measure characteristics, such as teacher experience (Goldhaber, Grout, & Huntington-Klein, 2014). Using easily measurable characteristics is limiting because they exclude intangible characteristics (Goldhaber, 2002), such as support, stress, commitment, and satisfaction, which may be equally, if not more, important to explaining and understanding teacher behaviors. Additionally, these elusive, intangible characteristics may have a relationship with both teacher absenteeism and teacher experience, for example. If this is the case, then the relationship between teacher absenteeism and teacher experience (for example) may be overstated. Since teachers must first be present in a classroom before they can impact student achievement, it is important to study what elusive teacher characteristics might also contribute to teacher absenteeism in order to identify behaviors or environments to support or discourage through policy or otherwise, to ultimately keep teachers in the classroom. For this reason, it is necessary to rely on existing frameworks in the academic literature to help sift through why teachers are absent from work. Moving forward, there are three areas in which research on teacher absenteeism can be improved. First, instead of asking the question: What is the impact of the teacher absenteeism on student achievement?; the new question should be: What are the underlying reasons for teacher absenteeism? The impact of teacher absenteeism on student achievement is already clear; however, the reasons and characteristics that explain teacher attendance behaviors are not as clear. Although teacher absenteeism has been demonstrated to have a negative influence on student achievement, even this line of research could benefit from the present study. There may be organizational factors that impact teacher attendance, and, in turn have an impact on organizational resources and school improvement efforts. Consistent instructional leadership is 43 needed to drive school improvement efforts, but if school leaders are regularly dealing with or filling in for absent teachers then school improvement efforts may be disrupted. For this reason, understanding what the underlying reasons are for teacher absenteeism is crucial so factors contributing to teacher absenteeism can be addressed. Second, more qualitative research should be included in this research stream. This literature review only found one qualitative study and two mixed methods studies when examining the teacher attendance literature. Otherwise, the many studies referenced here were purely qualitative and often used regression analysis. In order to better understand teacher attendance behaviors, more qualitative-oriented research may need to be conducted, such as through interviewing teachers. The intangible, elusive behaviors and characteristics mentioned above are hard to capture through purely quantitative data. Finally, this research stream can be improved upon by conducting more peer-reviewed research. Half of the articles discussed here were not peer-reviewed, and the peer-review process adds valuable insights, questions, and thinking around research topics. Research on teacher absenteeism would really benefit from the peer-review process because it allows like-minded researchers to bounce ideas around, which will lead to a more robust contribution to this research stream. In addition to reviewing the literature on teacher absenteeism (or lack thereof), this literature review also sought to identify an interpretive lens through which teacher absenteeism can be examined. The existing literature on teacher turnover and employee absenteeism provides potential interpretive lenses to view teacher attendance behaviors through. More specifically, viewing teacher absenteeism through a lens that considers organizational factors such as perceived organizational support, job stress, organizational commitment, and job satisfaction 44 may help make meaning of teacher absenteeism. Additionally, it is important to factor in the relationship between union membership and employee absenteeism since teacher unions have a strong presence in the state where the present study takes place. Through these different lenses, a more comprehensive understanding around why teachers are absent will emerge. In sum, this literature review reveals that pursuing additional research on teacher attendance behaviors is worthwhile, especially in the U.S. context. It contributes to the overall literature on teacher attendance/teacher absenteeism by systematically analyzing the existing literature and highlighting what is currently known about teacher attendance/teacher absenteeism. While there will still be challenges with measuring and collecting teacher absenteeism data, especially in the U.S. context because it is not regulated, it is definitely a worthwhile area for further exploration given the relationship it has with student achievement. Once the academic community gains a better understanding of the underlying reasons for teacher absenteeism, then local, state, and/or federal practices can be created and implemented to target wanted or unwanted organizational characteristics and teacher behaviors related to teacher attendance. Further, adding to the research on this topic through examining why teachers are absent may move this issue up the policy agenda, which could change how teacher attendance is factored into high-stakes accountability and how teacher attendance data are collected. In theory, if teachers are the single most important input that impacts student achievement, then understanding what it is about teachers that makes them so important and how to promote these desirable characteristics, such as their easy-to-measure and elusive characteristics, should be an education policy research priority. The next section dives into methodological considerations for this study on teacher absenteeism. The methods section provides information about my research strategy and design, 45 as well an overview of my sample and participants. I also discuss my positionality and the steps taken to enhance the validity and reliability of the study. Finally, the methods section details the study’s data collection, data analysis, and coding. 46 METHODS This section lays out the methodological approaches taken to better understand teacher absenteeism and begins with my epistemological perspective. Next, I provide support for my research strategy of qualitative design through interview methodology. I then discuss my research design, including my sampling strategy, participant overview, and data collection through interview methods. I also address threats to validity and reliability and the steps I take to bolster them, including triangulation and building rapport with my participants. The concluding section reviews my data analysis and coding process. Epistemological Perspective According to Maxwell (2013), when the goal of a study is to understand meaning, process, or context, qualitative research is best suited. Thus, a qualitative research study facilitates the understanding of teacher absenteeism behaviors. Specifically, the study at hand takes an interpretive/constructivist orientation. An interpretive/constructivist orientation is appropriate when the qualitative study seeks to understand, describe, and interpret a given phenomenon (Merriam & Tisdell, 2016). In other words, the interpretive/constructivist orientation helps to make meaning of a certain behavior (Creswell, 2013), and one of the aims of the present study is to learn about the meaning teachers make of their workplace environment that leads to teacher absenteeism. Also, this orientation supports that there can be multiple interpretations of the same event or behavior, which depends on how the event or behavior is socially constructed (Lather, 2006; Merriam & Tisdell, 2016). For this reason, the interpretive/constructivist orientation helps make meaning of the complex experiences and perspectives of the participants around teacher absenteeism behaviors. 47 The interpretive/constructivist orientation gives way to a basic qualitative study on teacher absenteeism because I am “interested in understanding the meaning a phenomenon has for those involved” (Merriam & Tisdell, 2016, p. 23). Likewise, I start with a deductive approach, where existing theories are used to hypothesize about the behaviors of collected data (Merriam & Tisdell, 2016). From the literature review, it is clear that there are some semi- suitable frameworks in the education and management literatures that can be used to form hypotheses around teacher absenteeism behaviors, and thus starting with a deductive approach is appropriate for the study of the phenomenon of teacher absenteeism. However, I complement the deductive approach with an inductive approach to be open to other theories. Inductive elements are included here because the existing theories may not adequately explain teacher absenteeism, and thus the data collected from this study builds toward a more comprehensive theory to explain teacher absenteeism (Merriam & Tisdell, 2016). The research question mentioned earlier guides this basic qualitative study and is constructed in a comprehensive way in order to provide the participants with latitude in constructing meaning of their attendance behaviors. Additionally, the interview protocol in Appendix A details the guided, open-ended interview questions and includes prompts, pivot language, and probing questions to provide many opportunities for the participants to speak about their perspectives and experiences that impact their attendance behaviors from different angles. One of the primary data collection methods for a basic qualitative study is through interviews (Merriam & Tisdell, 2016) because interviews help the researcher to understand individuals’ realities (e.g. Crouch & McKenzie, 2006; Seidman, 2006; Seidman, 2013; Lamont & Swidler, 2014). The interpretive/constructivist orientation supports the use of open-ended 48 questions during interviews because they enable participants in the present study to detail teacher absenteeism behaviors from their own perspective. Method for Data Collection The three most common ways to collect data in a basic qualitative study are through interviews, document analysis, or observation (Merriam & Tisdell, 2016). The study at hand collected data mainly through interviews, but document analysis was also used for the purposes of triangulation (more on that later). However, these behaviors are not directly observable and thus observations were not included in this study. There are many benefits to using interviews as the primary method for collecting data in a basic qualitative study. Aside from interviews yielding high response rates because participants are usually willing and interested in sharing their experiences (e.g. Appleton, 1995; Griffee, 2005), interviewing is well suited for collecting data on teacher absenteeism because it allows for flexibility in gathering the data – for example, if the participant seems puzzled by a question, either through verbal or non-verbal cues, the interviewer can re-word the question or offer clarifying information (e.g. Appleton, 1995; Schensul, Schensul, & LeCompte, 1999; Creswell & Miller, 2000; DiCicco‐Bloom, & Crabtree, 2006; Stephens, 2007; Marvasti & Freie, 2017). Additionally, teachers may be familiar with different terms and phrases that describe their paid time off (e.g. paid leave, vacation days, etc.), so if the interviewer uses a phrase that the interviewee seems confused about, the interviewer can respond by providing more information to the interviewee. Similarly, the interview method allows the interviewer to probe for additional information (i.e. ask a follow up question) if the interviewee says something interesting but potentially incomplete. For example, if an interviewee mentions that there are many reasons for 49 which he/she uses his/her paid time off, the interviewer can ask him/her for examples or to describe the reasons that account for most of the days used. In addition, the flexibility of interviews allows the researcher to ask sensitive questions but in a delicate manner (Lamont & Swidler, 2014). There may be questions related to teacher absenteeism that are sensitive, such as a question about how a participant handles a situation when their paid time off days are exhausted, but they need a day off. The interviewer can word the question in a gentle way so as not to make the teacher feel uncomfortable talking about taking more days off than allotted. For example, when asking about taking time off beyond the allotment of paid time off days the interviewer might say, “Sometimes unexpected things happen, and we all find ourselves in situations where we need to take more days off than expected. How often would you say you find yourself in a situation where you need to take more time off than what your district allots?” Framing this question with an empathetic approach might make the participant more comfortable with providing an unguarded response. This point about asking sensitive questions is pertinent to the study of teacher absenteeism because asking teachers why they are absent may feel like a threatening or accusatory question. Teachers may feel like they are being accused of missing too much school or being absent for inappropriate reasons. If participants feel threatened, then they might alter their responses or provide incomplete information, which can influence the findings (e.g. Holbrook, Green, & Krosnick, 2003; Griffee, 2005; Maxwell, 2013). There are a few ways in which I shaped the interviews to help make the participants feel comfortable with answering sensitive questions. First, I ensured confidentiality by having the participants pick pseudonyms, never revealing any identifying information about the participant, and never revealing any identifying information about the school or district the participant worked for. Second, my 50 interview protocol (Appendix A) was designed in such a way to ask factual and sociodemographic questions at the beginning of the interview and more sensitive questions toward the end of the interview. This allowed the participant to invest in the interview and also provided me with time to build rapport with the participant (Merriam & Tisdell, 2016). Third, I made a conscious effort to build rapport with the participant through my tone, dress, and body language. Interview methodology is particularly conducive to building rapport with the participant and is further discussed in a later section. Finally, I tried to frame the interview and the sensitive questions, specifically, in a way that conveyed the interview was an opportunity for the participant tell their story. In Dunn’s (2018) study of teachers’ public resignation letters, she found that teachers had a strong desire to share their narratives about why they decided to resign, and they did so through writing letters. While these interviews were confidential, it was my goal to learn about each participant’s narrative around their attendance behaviors in a non-penalizing way. In sum, interviews not only facilitate the collection of rich, detailed, and descriptive data on teacher absenteeism, but the flexibility afforded through interview methods allows for sensitive questions to be asked. A deeper look at some specifics around interview methods are discussed below. Interviews The interpretive/constructivist orientation supports the use of interviews because they enable participants in the present study to detail teacher absenteeism behaviors from their own perspectives. It is necessary to provide additional, thoughtful considerations to this approach in order to demonstrate the applicability of interviewing to the study of teacher absenteeism. One 51 way to do this is through considering some of the challenges with interviews and how I aim to overcome these challenges. In general, interviewing can be both costly (e.g. Appleton, 1995; Hoepfl, 1997; Seidman, 2013) and time consuming (e.g. Appleton, 1995; Seidman, 2013). However, these minor drawbacks are easily overcome with organization and preparation. A drawback worthy of more consideration is that the collection of rich interview data depends largely on the skills and experience of the interviewer (Appleton, 1995; Brinkmann, 2014). This is problematic because it could pose threats to the validity (i.e. does my study/interview protocol measure what it intends to measure?) and reliability (i.e. does my study/interview protocol produce consistent results over place and time?) of the study. For example, a novice interviewer may bias the data by using leading questions. To overcome this challenge, I not only piloted my interview protocol, but I also conducted multiple cognitive interviews (e.g. Appleton, 1995; Desimone & Le Floch, 2004; Maxwell, 2013; Merriam & Tisdell, 2016). These opportunities gave me feedback regarding my interview instrument and interviewing approach, provided me with opportunities to practice the interviewing technique, and allowed me to re-think and re-word questions as needed. Using these two approaches (piloting and cognitive interviews) improves the reliability and validity of my interviews (e.g. Appleton, 1995; Desimone & Le Floch, 2004; Maxwell, 2013; Merriam & Tisdell, 2016), and “the result can be vastly improved quality of survey questions and subsequent data” (Desimone & Le Floch, 2004, p. 19-20). As a novice interviewer, I also used an interview protocol that included pivot language and transition questions to improve the reliability and validity of my interviews. Using an interview protocol generally, but also one that includes pivot language and transition questions aids a novice interviewer in keeping the interviews focused and on track (Rubin & Rubin, 2012; 52 Creswell, 2013; Castillo-Montoya, 2016). Pivot language and transition questions are important because they help keep the interview conversational between topics while also steering the interview and/or refocusing the interview in the desired direction toward the central question (Rubin & Rubin, 2012; Castillo-Montoya, 2016). For example, when transitioning from questions related to teacher attendance policies to questions about organizational factors that may influence teacher absenteeism, I used pivot language such as, “My review of the academic literature suggests organizational support, organizational commitment, job stress, and job satisfaction may be a few factors related to employee absenteeism. I’m going to ask you a few specific questions regarding these factors and there will also be opportunities for you to elaborate or provide additional information and context.” This type of pivot language helped frame upcoming questions to the participant, too. Another challenge with conducting interviews is that interviewee responses may reflect social desirability (Griffee, 2005) because of the self-report nature of interview data (Appleton, 1995). While interviews allow participants to provide first-person accounts of their experiences (e.g. Crouch & McKenzie, 2006; Brinkmann, 2014), participants may feel compelled to respond in such a way that is viewed as favorable, which introduces bias to the data and compromises the reliability and validity of the interview instrument (especially the internal validity). This challenge is one that I needed to be especially attentive to because of the sensitive nature of some of the questions asked. In addition to ensuring confidentiality, being mindful of question order, and question framing, there are three main ways to overcome collecting data that are laden with social desirability. First, the majority of my interviews were face-to-face and one-on-one, as opposed to virtually, via phone, or through a focus group. While virtual or phone interviews may be appealing because of their logistical ease and cost efficiencies, interviewees may be more 53 forthright with their answers when they are face-to-face with the interviewer and have less time to construct a socially desirable response (e.g. Holbrook, Green, & Krosnick, 2003; James and Busher, 2006; Knox & Burkard, 2009). Furthermore, socially desirable responses are more common in focus groups, especially when the interview topic is sensitive or personal (Merriam & Tisdell, 2016), which is why the present study on teacher absenteeism mainly used face-to- face, one-on-one interviews. Though my goal was for all interviews to be conducted face-to-face and in person, eight of my interviews were conducted over Zoom (virtual/video conferencing) or phone. There are two reasons why this happened. First, the teachers I interviewed lived across the state of Michigan, and in some instances, scheduling was challenging, especially for those participants who lived more than an hour away from me. Second, my data collection took place over the summer, during which time many participants were on vacation. In these instances, I tried to use Zoom so the participant and I could see each other; however, there were a few instances where phone interviews took place. Although some of my interviews were virtual or over the phone, my perception is that the interview medium did not affect interview responses. Interviewees seemed to be frank with their answers and I did not notice any more hesitation or apprehension on their part compared to face-to-face interviewees. Some of these virtual and phone interviews were with chronically absent teachers, while others were not chronically absent. With that said, having to use Zoom and the phone for some interviews is definitely a limitation because it could result in responses laden with social desirability, which is why this limitation is further addressed in the conclusion. The second way to overcome the challenge of socially desirable responses is to build rapport with the interviewee. Building rapport can be a challenge in itself (e.g. DiCicco‐Bloom, 54 & Crabtree, 2006; Seidman, 2013; Spradley, 2016; Marvasti & Freie, 2017), but “[t]he strength of the interviewer-participant relationship is perhaps the single most important aspect of a qualitative research project” (Knox & Burkard, 2009, p. 569). If the interviewer is able to accomplish rapport, then the interviewee is more apt to be open and honest in his/her responses. Interviewers are more likely to build rapport with interviewees in face-to-face interview settings as compared to virtual or phone interviews (e.g. Holbrook, Green, & Krosnick, 2003; James and Busher, 2006; Opdenakker, 2006; Stephens, 2007; Knox & Burkard, 2009; Irvine, Drew, & Sainsbury, 2012; Merriam & Tisdell, 2016). Therefore, the use of mainly face-to-face interviews for this study on teacher absenteeism is further supported by the goal of building rapport between the interviewer and interviewee. Additionally, I was intentional about having a warm and friendly tone and demeanor, tried to dress in a similar level of professionalism as the participant (e.g. business casual, most likely), and conducted each interview at a mutually agreed upon time and in a mutually agreed upon, comfortable setting to create an environment conducive to building rapport with the participant (Merriam & Tisdell, 2016). Although it is easier to build rapport in face-to-face settings, I believe I was also successful at building rapport in the interviews conducted over Zoom or over the phone. I facilitated rapport in these instances by being as accommodating and flexible as possible to the interviewee’s schedule, expressing gratitude for their time, having a friendly tone and demeanor, and spending additional time at the beginning of the interviews with pleasantries. The third way to combat response bias, and thus improve internal validity, is through triangulation (e.g. Creswell & Miller, 2000; Merriam & Tisdell, 2016). One way to triangulate interview data is through document analysis (e.g. Bowen, 2009; Maxwell, 2013; Merriam & Tisdell, 2016). Document analysis can confirm (or disconfirm) interview responses (Bowen, 55 2009; Merriam & Tisdell, 2016). Document analysis also can provide context, inform important questions that should be asked of the interview participants, provide supplementary data, serve as a way to track contextual changes over time, and verify other data (Bowen, 2009). In addition, document analysis has logistical benefits, such as it being efficient, available, inexpensive, unobtrusive, and stable (Bowen, 2009). This enhances validity because either interview responses related to district attendance policies are confirmed through a secondary source, or the dissonance between interview responses and document analysis serves as a flag to the researcher that he/she needs to look further into something before publishing results. I examined teacher collective bargaining agreements and specifically reviewed official policies on teacher attendance and teacher absenteeism as outlined in the collective bargaining agreements. Examining collective bargaining agreements did not provide insights as to why teachers were absent from classrooms, but this process did provide important context around existing attendance policies that teachers operated under. Collective bargaining agreements are primary sources, public records, easily accessible, and free, and provide more insights to the existing policies about teacher absenteeism. I was able to examine these documents and consider them in tandem with interview data, which triangulated my findings. Taken together, using mainly face-to-face interviews, one-on-one interviews, taking the appropriate steps to build rapport, and using document analysis not only helped me overcome the challenge of having socially desirable responses, but also bolstered the validity and reliability of the present study on teacher absenteeism. A final challenge in using interviews is that bias may be introduced by the researcher’s analysis and interpretation of the data (e.g. Hoepfl, 1997; Crouch & McKenzie, 2006; Lamont & Swidler, 2014). This bias may come from the lens through which the researcher views the data, 56 the researcher’s experience with interview analysis and interpretation, or other sources. While using these lenses is necessary to focus on critical ideas, it is sometimes problematic because it may cause the researcher to focus on certain aspects of the interview while ignoring other, equally important aspects. For example, a novice researcher who is a strong supporter of teacher attendance financial incentives may overlook interview data that suggest teachers have safety and/or cultural concerns about their school, thus leading to their absences. This may lead the researcher to present results that are biased in support of offering teacher attendance financial incentives. Therefore, the researcher must balance using a critical lens with being objective. One way to address this is for the researcher to be aware of and upfront about his/her biases to minimize their occurrence in the data analysis and interpretation (Maxwell, 2013), which is done below in my positionality statement. Another way to address this bias is through the process of inductively coding because the researcher will view the data in a way that may not be captured by existing frameworks. A third way to address bias in researcher analysis and interpretation is to use triangulation, such as in the example of document analysis mentioned above (e.g. Creswell & Miller, 2000; Merriam & Tisdell, 2016). Finally, this issue is also addressed through member checking (e.g. Maxwell, 2013; Merriam & Tisdell, 2016), where the researcher can “solicit feedback on [their] preliminary or emerging findings from some of the people that [they] interviewed” (Merriam & Tisdell, 2016, p. 246). This allows the researcher to confirm if his/her analysis and interpretation resonate with the interview participants as being accurate. In the present study of teacher absenteeism, I used member checking as a way to enhance the validity and reliability of the study. Within interviews, there are numerous structures to consider: highly structured/standardized, semistructured, and unstructured/informal (Merriam & Tisdell, 2016). 57 For the study on teacher absenteeism, I decided to use semistructured interviews because of the guided flexibility they offer. Highly structured interviews do not allow for flexibility in question order or wording, making the interview more survey-like than conversational (Merriam & Tisdell, 2016). This may pose a challenge in building rapport with each participant, which is an important aspect when the topic is somewhat sensitive (i.e. teacher absenteeism). Unstructured interviews are useful when the researcher hopes to gain further information about a topic or phenomenon and tend to be exploratory (Merriam & Tisdell, 2016). Since I already had some knowledge of teacher absenteeism, unstructured interviews were not the best fit for the present study. Semistructured interviews allow for flexibility in question order and wording but are also guided from a semi-formal interview protocol/list of questions (e.g. Appleton, 1995; Schensul, Schensul, & LeCompte, 1999; DiCicco‐Bloom, & Crabtree, 2006; Seidman, 2006; Knox & Burkard, 2009; Maxwell, 2013; Patton, 2015; Merriam & Tisdell, 2016; Spradley, 2016). This format is helpful for building rapport and being responsive to a participant’s responses, facial expressions, and body language, which makes it the best interview structure for the present study on teacher absenteeism. An important caution to using semistructured interviews is that the flexibility afforded by them may result in inadvertent omission of important questions or topics (Patton, 2015). In order to attempt at avoiding this challenge, I used an interview protocol (Appendix A), which assisted in ensuring the necessary questions and topics were covered (Merriam & Tisdell, 2016). My interview protocol included pivot language and transition questions to help keep the interviews on track and so I did not accidentally omit important questions or topics. The interview protocol that I followed (Appendix A) included all the questions I asked the participants. My interview protocol provided a bit of structure to the semistructured interview 58 process and organized the questions so as to elicit responses in a systematic manner (Merriam & Tisdell, 2016). The interview protocol included specific questions central to my research question, but also included sub questions and possible probing questions to collect more detail, description, clarification, or examples (Merriam & Tisdell, 2016). These questions were largely open-ended to provide a conversation-like space for the participant to tell their experiences with and perceptions of teacher absenteeism. To keep the interview conversational, I also included pivot language and transition questions in the interview protocol (Rubin & Rubin, 2012; Castillo- Montoya, 2016). Furthermore, using an interview protocol not only helped me stay on track with regards topic coverage, but it also helped me stay on track with regards to time. My aim was for the interviews to last 45 minutes or less to minimize response fatigue. I did my best to respect the time of my participants, regardless of if all the interview questions were asked. The interview protocol helped the interviews move along so as to cover the main topics within the suggested time frame. In practice, the interviews lasted between 38 and 65 minutes, with the average interview length being 46 minutes. Using an interview protocol aided with the collection of data, but it only got me so far. In order for me to revisit the data and later analyze it, I recorded each interview but only with the permission of each participant. This practice allowed me to “preserve” the data for analysis and also review and improve my interviewing techniques (Merriam & Tisdell, 2016). Subsequently, the recorded interviews were transcribed verbatim by a paid transcription service offered through NVivo, a qualitative data analysis software. Before conducting any data analysis, I listened to each recording while simultaneously reading the transcript to “correct errors and fill in blanks” (Merriam & Tisdell, 2016, p. 132). While NVivo transcription captured most of the conversation in each of the interviews, there were some nuances that needed to be corrected by hand. Going 59 through this exercise kicked off my data analysis through thinking about potential themes. This point related to data analysis is further elaborated on in a later section. Participants The target sample for this study includes teachers from the state of Michigan. Broadly speaking, Michigan is a good area for studying this phenomenon because, while chronic teacher absenteeism is a national concern, it is particularly concerning in Michigan, where 24.7% of teachers in public schools were found to be chronically absent in a 2017 study conducted by the Thomas B. Fordham Institute. Additionally, chronic teacher absenteeism in Michigan is recently gaining traction in the local news (McVicar, 2017; The Detroit News, 2017; Wolcott, 2017; Renk, 2019). This study used purposeful sampling. Since the goal of this study is to understand and learn about teacher attendance behaviors and not to generalize to a broader population, purposeful sampling was appropriate (e.g. Maxwell, 2013; Patton, 2015; Merriam & Tisdell, 2016). In addition, since I am studying a particular phenomenon, teacher absenteeism, purposeful sampling allowed me to select contexts that had specific characteristics to facilitate the study of the given phenomenon (Merriam & Tisdell, 2016), such as districts that reported having high rates of teacher absenteeism. An important step in purposeful sampling is criterion-based selection, in which I decide what “what attributes of [my] sample are crucial to [my] study” (Merriam & Tisdell, 2016, p. 96). To gain a diverse perspective of teacher absenteeism, my participants needed to be permanent teachers. With regards to the teachers in the sample, temporary or short-term teachers may be held to different attendance requirements and therefore their experiences could be quite different than permanent teachers. In this case, I was interested gaining an understanding of the attendance behaviors of permanent teachers. Additionally, the 60 teachers were from elementary schools because the structure of elementary schools is inherently different from middle schools and high schools, which could result in variation that is not a primary focus of the present study. The last characteristic I aimed for in my purposeful sample was for participants to come from districts that reported high rates of teacher absenteeism, which is a contextual factor related to the study of teacher absenteeism behaviors. These locations were identified anecdotally and through word of mouth. Since my ideal sample consists mainly of teachers who are absent 10 or more days in a school year, I tried to identify chronically absent teachers by asking school administrators from schools with reported high rates of teacher absenteeism to connect me with permanent teachers who were absent 10 or more days in the 2018-19 school year. In practice, this was more challenging than originally anticipated and I was only successful in finding chronically absent teachers through this method in two instances. Administrators were not comfortable identifying teachers in this way and/or administrators could not surely say if the teachers were absent 10 or more days. Administrators seemed to know if a teacher had been absent “a lot” but not the exact number. Plus, since I used snowball sampling (elaborated on later), I asked teachers to refer me to other teachers who had been absent, but the referring teachers did not know the number of days their colleagues had been absent. As a result, my sample consists of 12 teachers who were absent 10 days or more, but it also includes nine teachers who were absent less than 10 days. However, having a sample with both chronically absent teachers and not chronically absent teachers allowed me to identify patterns that I may not have been able to identify with a sample of just chronically absent teachers. For example, both chronically absent teachers and non-chronically absent teachers reported using paid time off for job-related mental health, but chronically absent teachers used six or more days of paid time off 61 for job-related mental health compared to non-chronically absented teachers who only missed one to two days for job-related mental health. Including both groups in my sample helped uncover that chronically absent teachers missed work for job-related mental health at least three times more than non-chronically absent teachers. A more detailed description of my sample is provided in the Participant Overview section. Having a demographically diverse sample (e.g. gender, race, etc.) is not central to the current study, but perhaps worthy of consideration in future studies. Similarly, because the main goal is to have a sample that consists of teachers who were absent 10 or more days in the 2018- 19 school year, years of teaching experience was not a deciding factor for teachers included in my study. Moreover, since there is circumstantial evidence that chronic teacher absenteeism exits in schools regardless of the percentage of low-income and/or minority students in their student body, teachers from schools that are considered both high-poverty/high-minority and low- poverty/low-minority were included. Further, I included teachers from large districts in addition to teachers from small and medium districts, and teachers from districts with different locales and from different regions of the state. Table 1 provides a summary of the characteristics of my purposeful sample. 62 Table 1: Characteristics of Purposeful Sample Characteristic Position (Permanent Teachers) Elementary School Level Districts with High Rates of Teacher Absenteeism Years of Teaching Experience District Size District Student Body District Locale District Region Rationale Gain the perspective of teacher attendance behaviors from teachers. Permanent teachers are necessary because temporary or short-term staff may be subject to different attendance requirements. Focusing data collection at the elementary school level facilitates saturation of data with regards to variation; middle and high schools have structural and organizational differences from elementary schools, which are not main focuses of this study. Recruiting participants from districts with high rates of teacher absenteeism better facilitates the collection of data around the phenomenon under study, teacher absenteeism. Capture perspectives of teachers with different amounts of teaching experience to learn about differences in absenteeism experiences based on teaching experience. Capture perspectives of teachers from districts of different size (small and medium) to learn about differences in teacher absenteeism experiences based on district size. Recruiting participants from districts with different student body compositions assists with saturation of data and allows for comparison of teacher absenteeism experiences based on different student bodies. Capture perspectives of teachers from different district locales (e.g. town, city, suburb) to learn about differences in teacher absenteeism experiences based on district locale. Capture perspectives of teachers from different regions of the state (e.g. Central, Eastern, Western) to learn about differences in teacher absenteeism experiences based on district region. Convenience sampling, which has the benefit of being time and cost efficient (Merriam & Tisdell, 2016), was used to initially gather a pool of teacher participants. I continued my purposeful sampling with snowball sampling, where my initial, criterion-based participants referred me to other teachers to interview to facilitate the collection of diverse accounts and 63 experiences around teacher absenteeism (Merriam & Tisdell, 2016). Initial contact was made with the convenience sample via a recruitment email (Appendix C). Of those in the convenience sample who participated, I asked them if they knew of any other teachers who I might be able to talk with to get additional perspectives and rich information on teacher absenteeism. Specifically, I asked participants if they knew of a colleague(s) that they could put me in touch with who may have missed school for a different reason(s) and might be willing to talk to me about their experiences related to teacher absenteeism. All participants were given a ten-dollar Starbucks gift card as an incentive for their participation. Given this sampling approach, it was imperative that I was attentive to both selection bias and sampling bias. Selection bias is relevant to this study because those who agreed to talk with me may have wanted to tell me about their experiences, but their experiences might be quite different from someone who did not want to talk with me. This poses a threat to the internal validity of my study. Selection bias “results from using nonrandomly selected samples to estimate behavioral relationships” (Heckman, 1979). In this case, bias is introduced when my participants are only those who want to tell me about their experiences, but this excludes the experiences and perspectives of those teachers who may not want to initially talk with me. To mitigate selection bias, I expanded my convenience sample through snowball sampling, in which initial participants (who wanted to talk to me) connected me to others who may or may not have been interested in talking with me. I did this through using language such as, “Do you know of any colleague(s) who may have missed school for a different reason(s)? Would you be willing to put me in contact with them?” Of course, anyone was free to decline participation. Using this method introduced me to more cautious teachers who were apprehensive to share their experiences related to chronic absenteeism. Once I was connected with new teachers through the 64 snowball sampling method, I was intentional about mentioning that all responses were confidential and anonymous, which seemed ease any apprehensions about participating. These steps minimized selection bias because my sample consists of teachers who wanted to initially participate and those who may not have initially self-selected into my study. The other type of bias, sampling bias, is discussed in a later section titled Participant Overview. My target sample size was 15 teachers. However, I allowed room for fluctuation in either direction, so saturation or redundancy of information was reached (e.g. Maxwell, 2013; Merriam & Tisdell, 2016), meaning I sampled to the point where I started hearing similar responses and experiences across participants. When I started hearing similar accounts, it suggests I may have collected all variations of information on the phenomenon under study and little new information will surface. My sampling goal was to maximize the variation of responses collected, which was evident when I started hearing similarities in responses (e.g. Maxwell, 2013; Merriam & Tisdell, 2016). For this reason, I started analyzing data alongside data collection, so I was able to stay informed and knowledgeable about already collected data to be able to identify when redundancy started to take place. While my target sample size was 15 teachers, I ended up interviewing 21 teachers. I started hearing similar responses once I had interviewed 12 teachers but decided to continue with additional interviews to ensure redundancy was reached. Of course, before participants were identified, contacted, or interviewed, I obtained approval from Michigan State University’s Institutional Review Board (IRB). IRB approval is necessary to protect the rights of my research participants. IRB also protects me, as a researcher, by vetting my study and ensuring I do not engage in any research that could compromise my overseeing institution. Specifically, the IRB ensures ethical standards are followed in research that involves human subjects. The approved IRB application for this study can be obtained from 65 my academic advisor, Dr. John Yun, and the informed consent form can be found in Appendix B. A more detailed look at the participants in my study follows in the next section. Participant Overview Snowball sampling resulted in 21 participants from seven districts in Michigan. As mentioned above, these districts had self-reported high rates of teacher absenteeism as indicated by teachers and/or administrators. There was at least one chronically absent teacher from each district interviewed for this study. The districts in my sample varied both in size, as determined by student population, and locale, as determined by location. For example, District 5 had just over 1,600 students in the entire district in the 2018-19 school year, while District 6 had over 13,000 students in the entire district for the same school year. Similarly, District 5 was classified as a midsized suburban district while District 6 was classified as a small city district. The districts also represented different regions of the state, such as central Michigan, eastern Michigan, and western Michigan. Additionally, each represented district had a slightly different student body. For example, the student body of District 1 was mostly white and just over one third of the students were economically disadvantaged. Alternatively, the student body of District 2 was mostly non-white, and the majority of students were economically disadvantaged. Information about each district’s locale and region was obtained through the Michigan’s Educational Entity Master (Michigan’s Center for Educational Performance and Information, 2019). Information about each district’s student body was obtained through Michigan’s official public portal for education data, MISchoolData (Michigan’s Center for Educational Performance and Information, 2019). More details about these districts are found below in Table 2. 66 Table 2: District Characteristics Number of Teachers (N) Number of Teachers Chronically Absent (N) 4 4 2 2 2 3 4 1 4 1 2 1 2 1 2018-19 Student Body Entity Region Total Student Enrollment % % (N) White Female % Economically Disadvantaged % % English Language Learners Students with Disability Central 2,799 87.14 48.41 34.73 <5.00 10.54 Eastern 4,245 14.98 44.41 90.41 <5.00 20.31 Eastern 10,898 52.25 48.67 53.22 <5.00 17.40 Western 3,364 36.18 48.93 64.48 13.20 15.78 Western 1,603 70.74 47.97 54.15 <5.00 12.48 Eastern 13,150 51.00 48.37 15.37 16.24 9.71 Eastern 4,375 75.02 50.51 42.35 <5.00 11.43 Entity Locale Name Town: Fringe City: Small Suburb: Large City: Small Suburb: Midsize City: Small Suburb: Large 67 District Code District 1 District 2 District 3 District 4 District 5 District 6 District 7 Of the 21 total participants, 12 were chronically absent, meaning they reported missing 10 or more days during the 2018-19 school year. While both males and females were represented in my sample and the majority were females, this roughly mirrored the gender composition of all teachers in Michigan (Michigan’s Center for Educational Performance and Information, 2019). Unlike the population of teachers in Michigan, the teachers in my sample were all white. This points to one of the challenges with snowball sampling in that representiveness is compromised (Atkinson & Flint, 2001; Cohen, Manion, & Morrison, 2011; Patton, 2015), which is a threat to external validity. When snowball sampling is used, initial participants are asked to refer the researcher to other participants (Patton, 2015; Merriam & Tisdell, 2016), as described in the previous section. While this method can result in additional participants, often times relational demography plays a role in that the initial participants’ social network consists of individuals who are demographically similar to them (Riordan, 2000), which may lead the participant to refer the researcher to others who are demographically similar to the initial participant. This is an example of sampling bias, where the sample does not accurately represent the population under study (Atkinson & Flint, 2001; Cohen, Manion, & Morrison, 2011; Patton, 2015). One of the best ways to mitigate sampling bias is to have a large sample size and use a method for randomization; however, with qualitative research that uses interview methodology, this is challenging because interviews are expensive and time consuming (Crouch & McKenzie, 2006). For the study at hand, the critical characteristic of my sample was to have a sample that consisted mainly of teachers who were absent 10 or more days in a school year. As such, having a demographically diverse sample (e.g. gender, race, etc.) was not central to the current study, but perhaps worthy of consideration in future studies. Even so, the fact that my sample consisted 68 only of white teachers must be considered when thinking about findings and implications. These cautions are heeded in later sections. When looking at the spread of teacher experience, education level, age, and teaching grade, my sample had more variation than it did with race. Only two teachers in the sample were in their first five years of teaching, while the rest had been teaching for at least six years, with the longest tenured teacher being in the teaching profession for 36 years. All interviewed teachers had a bachelor’s degree at minimum, and the majority of the sample also had at least one master’s degree. While the majority of teachers in the sample were under the age 50, there was still quite a bit of spread with regards to age of participants in the sample. As I mentioned earlier, I focused data collection at elementary school teachers. Teachers in my sample were well distributed between first grade, second grade, third grade, fourth grade, and fifth grade, with third grade having the largest teacher representation in the sample. The final characteristic reported on is whether or not the teacher was a union member. The vast majority of teachers in this sample were union members, which is not surprising because there is a strong teacher union presence in Michigan (Winkler, Scull, & Zeehandelaar, 2012). Again, this introduces a bit of sampling bias because my sample had very few perspectives from teachers who were not union members. Moreover, neither of the two non-union teachers were chronically absent, meaning my sample did not include perspective from a non-union, chronically absent teacher. This limitation compromises representiveness of the sample, poses a threat to external validity, and will be considered further in the Discussion section. More details about participant characteristics are found below in Table 3. 69 Table 3: Participant Overview Demographic Category Demographic Subcategory Gender Race/Ethnicity Years of Teaching Experience Highest Education Level Age Range Teaching Grade Union Member Positionality Statement Female Male White 0-5 Years 6-10 Years 10-20 Years 20+ Years Bachelor's Degree Master's Degree 20-29 30-39 40-49 50-59 60-69 1 2 3 4 5 Yes No Count (N) 17 4 21 2 8 5 6 6 15 5 7 5 2 2 2 4 7 4 4 19 2 Count of Chronically Absent (N) Total Chronically Absent (N) / Total Participants (N) 10 2 12 1 5 4 2 3 9 2 5 4 1 0 1 1 5 2 3 12 0 12 / 21 Calling attention to my positionality, as the researcher, with respect to the study of teacher attendance behaviors is important because my positionality can impact data access, collection, and analysis (e.g. Rossman & Rallis, 1998; Maxwell, 2013; Merriam & Tisdell, 2016). As mentioned earlier in the methods section, a researcher’s background and ideologies can introduce bias to the study. One way to address this is for me to own and account for my 70 positionality to minimize the potential for bias in the data analysis and interpretation (e.g. Rossman & Rallis, 1998; Maxwell, 2013; Merriam & Tisdell, 2016). I approached this study as an Education Policy Ph.D. candidate with a background in working with Michigan K-12 districts, schools, teachers, principals, and central office administrators to support their school improvement efforts. However, I have never been a teacher, principal, or central office administrator. This point is noteworthy because it makes me an outsider to this study on teacher attendance behaviors. This may impact access to participants as well as what the participants are comfortable sharing with me about teacher absenteeism. For this reason, the first tier of my sampling utilized a convenience sample of teachers that I had existing connections with. Because I am an outsider, I also had to make a very concerted effort to build rapport with my sample through eye contact, body language, dress and appearance, demeanor, and communication. For example, one step I took to build rapport with my participants was to dress like them. I was sensitive to walking into a meeting with a teacher dressed as a business professional because that might have reminded them of their principal, superintendent, or state officer coming in to evaluate them. I wanted to avoid this feeling, so I dressed more casually while still maintaining a level of professionalism. While race and gender are not central variables to this study, it is noteworthy to my positionality that I am a white female, which may shape the responses of those who are similar or dissimilar to my race and/or gender. All teacher participants are white, which minimizes race-of- interviewer effects (Schaeffer, 1980; Rhodes, 1994). If there were participants with a different race than myself, I would need to be attentive to the potential for introducing bias; however, since my topic and questions do not include racial content, race-of-interviewer effects may not be as serious of a concern (Schaeffer, 1980; Rhodes, 1994). Age and socio-economic status are 71 also not central to this study; however, race and socio-economic status should still be accounted for and may be considered in future studies. As such, I am in my early thirties, come from a middle-class background, and reside in a medium-sized community. In the interviews, I was especially sensitive to establishing rapport with participants with different backgrounds than me, such as those teachers who were older/had more tenure than myself. Perhaps of highest importance related to my positionality is how I came about this study and the biases that I might have brought to it. My interests in better understanding teacher attendance behaviors developed in my professional career as a data and evaluation specialist while supporting Michigan K-12 districts, schools, teachers, principals, and central office administrators with their school improvement efforts. In my work, I continued to hear anecdotal accounts of teacher absenteeism being a problem in the schools I was supporting, but these claims from teachers and administrators were never substantiated by any data or research. I began some primary research on the topic, which suggested teacher absenteeism was a problem faced by many schools and districts but the research literature on the topic was very limited. When I looked into how teacher attendance data were used, I quickly found that teacher attendance data were not used at the state level in Michigan, which might suggest these data are not used for decision making at the school or district level, either. I believe that teacher attendance is important for student learning and I find it problematic that teachers are not showing up to teach. This made me wonder why teachers are not showing up to teach. The culture and climates of the schools I worked in were sometimes challenging environments because they were identified as chronically underperforming (i.e. Priority schools or Comprehensive Support and Improvement schools) by the Michigan Department of Education. This made me curious if teacher absenteeism was associated with these challenging 72 environments. In other words, I wanted to know if declining organizational (i.e. school) environment was related to “exiting” behaviors by teachers as manifested through absenteeism (Hirschman, 1970). Given the anecdotal accounts I heard, I believe teacher absenteeism is a problem in Michigan schools and I think there may be culture and climate issues in schools (i.e. organizational factors) that contribute to chronic teacher absenteeism. For example, if there is a negative school climate, teachers may be more inclined to be absent. I want to learn why teachers are chronically absent so contributing factors can be better addressed. If, through this study, teachers suggest organizational factors contribute to their absenteeism, then perhaps efforts can be made to improve upon the organizational factors. In future endeavors, I would like to support the improvement of working environments for teachers, so they are more satisfied with their jobs and less tempted to be absent. In order to do this, I need to understand and be cognizant of why teachers are missing school in the first place. Therefore, I decided to delve into this topic for my dissertation research. This narrative points to some of the biases I bring with me to this study, namely I believe organizational environments have a relationship with teacher absenteeism. Furthermore, my interest in education policy, teacher accountability, teacher working conditions as they relate to culture and climate, and reforming the teaching profession also contribute to positionality in relation to this study. Through intentional education policy, I believe important changes can be made to elevate the teaching profession and make it a highly desirable, respected, and prestigious career for college-aged students. In order to do this, it is necessary to understand some of the challenges with the teaching profession, which may be related to teacher accountability and workplace culture and climate. Instead of speculating on why teachers are/are not showing up to work, I want to give teachers the opportunity to tell me what contributes to their decision to be absent so I can better understand these behaviors. Perhaps 73 this will inform policy that targets wanted or unwanted practices, environments, etc. related to supporting teacher attendance. These ideas speak to my positionality and rest on the assumption that organizational factors impact teacher absenteeism. Therefore, I approached the interviews with a potentially narrow lens and thus these biases must be noted and addressed in my data collection and analysis. Being aware of and bringing attention to these biases is a critical step in assuaging them (e.g. Merriam & Tisdell, 2016). However, I welcomed the participants to challenge my assumptions and am hopeful that my interview questions created a space for the participants to do so. Data Collection Participants were initially contacted through recruitment email (Appendix C). I then made individual arrangements with those that voluntarily agreed to participate to meet at a mutually agreed upon, comfortable setting at a mutually agreed upon date and time. For those interviews conducted over the phone or Zoom, we decided on a mutually agreed upon date and time for the phone or video call. I also indicated that the interview would take between forty-five minutes and one hour. Once an interview was scheduled, I provided the participant with a consent form to review (Appendix B). Upon meeting for the interview but before the actual interview took place, I presented the consent form again, reviewed it with the participant, and asked the participant to sign the form if they still agreed to participate. For the in-person interviews, the participant signed the consent form on the spot. For the phone and Zoom interviews, the consent form was either emailed or snail mailed to the participants ahead of time and returned to me with their signature over email or snail mail. I also presented the participant with a ten-dollar Starbucks gift card as an incentive for their participation (either as a physical gift card or virtual gift card). Each participant was asked to sign in receipt of the gift card so I 74 could keep an audit trail. Then, I read the interview protocol (Appendix A) to each participant, which also asked the participant for permission to record the interview. After obtaining written and verbal consent that I could record the interview, I turned the recorder on and proceeded with the interview questions as outlined in the protocol. These recordings were later transcribed for data analysis. At the close of the interview, I thanked the participant for their time and asked for permission to contact them with any follow up questions. Additionally, I indicated that I would like the participant to review my analysis of their interview, so the participant would have the opportunity to provide feedback around accuracy of my interpretation (i.e. this is called member checking). I used member checking in 17 of the 21 interviews. Specifically, I sent individual emails to each of the participants asking them to review the included analysis of and quotations from their interviews. Ten participants responded they were satisfied with my analysis and selected quotes, and the other seven had minor tweaks. During the interviews, I also made observations and took field notes; however, this was not the primary method of data collection. Finally, I obtained the CBAs from each of the school districts represented by the participants in order to analyze teacher attendance policies as written in the CBAs. Teacher CBAs are publicly available, usually through a school or district website, and provided me with supplementary information around the exiting teacher attendance/teacher absenteeism policies in a given district. Validity and Reliability Certain validity and reliability concerns have already been addressed in the above discussion. Namely, pilot interviews, cognitive interviews, document analysis (i.e. triangulation), member checking, sampling to the point of redundancy, inductive and deductive analysis, and acknowledging my positionality were all used to enhance reliability and validity. There are a few 75 other points related enhancing the validity and reliability of this study worth mentioning. First, I kept an audit trail to readily revisit steps taken, interviews conducted, and decisions made. An audit trail is especially useful so myself and others can follow the steps I took and processes I used to arrive at my findings (Creswell & Miller, 2000; Merriam & Tisdell, 2016). Second, as briefly mentioned earlier, I aimed to have a large enough sample size to the point where I saw repetition in the interview responses (Crouch & McKenzie, 2006; Merriam & Tisdell, 2016). As explained by Merriam & Tisdell (2016), “[r]eaching a point of saturation or redundancy means that you begin hearing the same responses to your interview questions or seeing the same behaviors in observations; no new insights are forthcoming” (p. 101). Finally, throughout this dissertation I try to provide “rich, thick description” of my study and of my findings so the findings can be transferable to other contexts and settings (Merriam & Tisdell, 2016). Taken together, recording an audit trail, sampling to the point of redundancy, and providing detailed descriptions of my study and findings are three additional practices I employed to improve the reliability and validity of my study on chronic teacher absenteeism. Table 4 provides a summary of the considerations for bolstering validity and reliability. 76 Table 4: Considerations for Validity and Reliability Criteria Internal Validity (credibility) Strategy Pilot interviews, cognitive interviews, ensuring confidentiality, question order, question framing, face-to- face and one-on-one interviews, referral from initial participants to others who may not initially be interested in sharing their experiences, build rapport, document analysis, member checking, inductive and deductive analysis External Validity (transferability) Addressing researcher bias and positionality, providing rich and detailed descriptions, keeping an audit trail Sampling to the point of redundancy, document analysis, positionality statement, audit trail Document analysis; member checking, audit trail, inductive and deductive analysis Reliability (dependability) Objectivity (confirmability) Data Analysis and Coding Upon completion of each interview, the recording was transcribed. Given that qualitative research is an iterative process (Merriam & Tisdell, 2016), transcription occurred alongside new interviews so I could learn and make meaning as the data collection continued. The recorded interviews were transcribed verbatim by a paid transcription service offered through NVivo, a qualitative data analysis software. Before conducting any data analysis, I listened to each recording while simultaneously reading the transcript to “correct errors and fill in blanks” (Merriam & Tisdell, 2016, p. 132). While NVivo transcription captured most of the conversation in each of the interviews, there were some nuances that needed to be corrected by hand. Going through this exercise kicked off my data analysis through thinking about potential themes. The first step in my data analysis was to create an inventory of the different pieces of collected data (e.g. interview transcripts, documents, field notes, observations). This inventory helped to keep my data organized and also indicated connections between the different data, such as if the data 77 came from the same school and/or district. My next step was to read through all the interview transcripts while taking notes as I went in order to document thoughts and observed patterns as they related to my research question and framework. This step allowed me to reflect on the data to start identifying common ideas and themes. Another step I took was to listen to each interview recording without reading along with the transcription. I listened to each interview from beginning to end at least four times. This step forced me to really listen to the participants’ responses, rather than focusing on the accuracy of the transcription. During this step, I jotted down notes and noted timestamps I wanted to revisit. I continued this practice throughout the entirety of my data analysis and it even continued into some of my writing. This was one of the most productive data analysis steps I took because it allowed me to immerse myself in the participants’ realities and really get to know my data. At this point, I started deductive data analysis through a first round of coding. My first round of coding relied on codes derived from the management and teacher turnover framework, such as job stress, organizational support, commitment, and job satisfaction. In the first round of coding, I also included codes more generally related to absenteeism, such as personal illness and vacation. These codes fell into two categories: personal reasons for using paid time off (e.g. illness, vacation) and job-related reasons for using paid time off (e.g. job stress, organizational support). The first round of coding did not reveal the antecedents to job stress and organizational support, for example, so I moved to an inductive approach for my second round of coding. The second round of coding focused on codes related to the drivers behind job stress and organizational support, such as student behaviors, class size, accountability, building support, and administration support. I also engaged in a third round of coding because through my review of the transcripts and recordings, I noticed unexpected commonalities in responses that were 78 unrelated to my framework and wanted to be sure to capture them. This round of coding also used an inductive approach and captured codes that fell into a financial category, such as salary, compensation, and incentives. By coupling deductive data analysis with inductive data analysis, comparative data analysis commenced. This approach is commonly used in qualitative research because it allows the researcher to learn and make meaning of the collected data while considering a lens that might help explain some of the behaviors (Merriam & Tisdell, 2016). As mentioned earlier, my goal was to make meaning of the interview data collected on teacher attendance behaviors, and a teacher turnover and employee absenteeism lens supported my understanding. However, this deductive approach may overlook certain nuances related to teacher attendance behaviors because the existing theories are not perfect fits; therefore, I also used some inductive analysis to capture anything these lenses might miss. To begin the inductive process, I reviewed my research question and then systematically went through every transcript to identify responses, phrases, words, etc. that spoke directly to my research question (Merriam & Tisdell, 2016). Part of my process was to use a qualitative analysis program (NVivo) to identify certain words and phrases that were repetitious across interviews or seemed to speak directly to my research question. I indicated these codes in NVivo to represent categories and themes, and the code words used were researcher-generated or participant-generated (Saldaña, 2015). The participant-generated codes were indicated through the use of quotation marks. Next, I used NVivo to examine and group similar codes together. As I did this, I kept in mind my interpretive/constructivist orientation to make meaning of the data. Simultaneously, I revisited the potential biases that I brought to the analysis, especially with regards to my positionality (Merriam & Tisdell, 2016). I continued to use NVivo to refine the different codes, so they were most representative of the data 79 and the purpose of the study on chronic teacher absenteeism. Coincidently, this type of coding is called In Vivo coding because it seeks to make meaning from spoken word, perspectives, and experiences (Saldaña, 2015). One of the main benefits to using In Vivo coding is that it “prioritizes[s] and honor[s] the participant’s voice” (Saldaña, 2015, p. 106), which is important to the meaning making process of the study at hand. Moreover, Saldaña (2015) emphasizes the importance of the alignment between the coding method and the research question asked. Since the research question guiding this study is an ontological question because it seeks to understand participants’ realities, In Vivo coding is aligned with this ontological characterization in that it helps make meaning through and preserves participants’ voice and spoken word (Saldaña, 2015). Once my first pass at In Vivo coding was completed, I read through the transcripts again to check that I identified all the significant words and phrases as they related to my research question. Any newly identified codes, notes, etc. were indicated with a new color. Following the In Vivo coding, I themed my data by looking across both deductive and inductive codes and grouping them into categories that represented potential patterns or themes within and across interviews. For example, the codes “job stress” and “organizational support” were grouped together into the category “job-related reasons for using paid time off.” Eventually, this category gave way to a theme around differences in using paid time off for job- related reasons between chronically absent teachers and non-chronically absent teachers. Themes provide a profile of my data and help summarize some of the ideas presented by the participants (Saldaña, 2015). I explored both the data and the codes that resulted from In Vivo coding to theme my data. The spreadsheet that contains all the codes allowed me to easily group similar codes together and generate themes. Theming also provided space to reorganize and reanalyze 80 my data. This process was an additional step in helping me make sense of the data and to start exploring some of the lenses mentioned in the literature review to organize the data. While I believe In Vivo coding, including theming, adequately captured the categorization of the collected interview data on teacher absenteeism, I decided to engage in multiple cycles of coding to “develop a sense of categorical, thematic, conceptual, and/or theoretical organization from [my] array of first cycle codes” (Saldaña, 2015, p. 234). For example, I used pattern coding to group some of my initial codes together (Saldaña, 2015). This step is not always necessary, but I felt that my first few passes at In Vivo coding resulted in too many codes or too broad of summaries of the data and thus I engaged in another cycle of coding. Upon completion of the coding, I revisited the lenses proposed from the teacher turnover and employee absenteeism literature and attempted to use them to understand the codes and behaviors reported by my participants. These understandings were written up and inform the findings in a later section. In addition to In Vivo coding, I conducted document analysis with the collective bargaining agreements (CBAs) from the represented school districts. It is common for document analysis to be used alongside other methodologies, such as interviewing, in qualitative research for the purposes of triangulation (Bowen, 2009). A primary purpose of document analysis is to “provide data on the context within which research participants operate” (Bowen, 2009, p. 30). For this study, I used document analysis of CBAs to provide context around the existing teacher attendance policies as officially written in the CBAs. Specifically, analyzing the CBAs of represented districts provided additional context for teachers’ perceptions, beliefs, and decisions related to their absenteeism. Using document analysis alongside analysis of the interview data allowed me to compare actual attendance policies with teachers’ perceptions of them and identify any disconnects between the two that may have had an influence on the teachers’ 81 decisions to be absent. My process for analyzing the collective bargaining agreements started by skimming each agreement to get a general understanding of the content and layout. Then, I read each agreement more carefully, making sure to take notes in the margins and on a separate sheet as I went along. Next, I read the documents again and identified themes, while being mindful of the intended purpose and target audience of each agreement. I organized my notes in a document so I could think about grouping similar notes together and generating codes. More often than not, these codes came directly from the text itself, such as “vacation,” “illness,” and “incentive policy.” At this point, I could evaluate the themes and codes as they related to the problem, purpose, and research question around teacher absenteeism. The document analysis process allowed me to start understanding the context within which teachers operated, as it related to teacher attendance policies as officially defined in CBAs. The information from the document analysis of CBAs was used for triangulation and to inform my findings. Summary of Methods In review, this section provides details around the methods used to better understand teacher absenteeism behaviors. After much consideration, a basic qualitative design using interviews was selected because it is supported by the interpretive/constructivist orientation of this study, which seeks to understand, describe, and interpret chronic teacher absenteeism. The flexibility of interviews was especially helpful to the study of chronic teacher absenteeism because it provided space for me to build rapport with participants, which is an important step when sensitive questions are asked. With permission from the participants, the interviews were recorded and transcribed. The participants were teachers from Michigan schools and districts, and a convenience sampling strategy was used to initially recruit participants. This was followed by a snowball sampling approach to expand my sample by asking initial participants to connect 82 me with other teachers who may or may not have been initially interested in talking with me. Throughout my data collection, I tried to be sensitive to my positionality as a Ph.D. student, white, middle-class, female, who has an interest in education policy, teacher accountability, teacher working conditions as they relate to culture and climate, and reforming the teaching profession. Data collection was guided by an interview protocol that included both primary questions and probing questions. Validity and reliability were enhanced through using pilot interviews, cognitive interviews, triangulation, member checking, sampling to the point of redundancy, participant referral, acknowledging my positionality, keeping an audit trail, inductive and deductive analysis, and providing rich descriptions in my findings. In Vivo coding assisted in making meaning of the collected data. Additionally, document analysis of teacher attendance policies as written in CBAs was used to triangulate my interview data. Ultimately, these various steps facilitated the collection of rich data on teacher attendance behaviors in order to better understand the drivers behind a teacher’s decision to be absent. The findings as a result of these methods are described in the next section. 83 FINDINGS The research question for this study, “what reasons do teachers provide for their chronic absenteeism?,” led to the collection of rich, descriptive interview and document analysis data. Before diving into why teachers were absent, I want to provide a bit more context around what teachers in my sample thought about teacher absenteeism and its impacts. In the majority of the interviews I conducted, regardless of whether or not they were chronically absent, teachers alluded to or blatantly mentioned a perceived “problem” with chronic teacher absenteeism in their buildings and districts. To give a sense for challenges with chronic teacher absenteeism, one teacher’s perspective is captured by the following quote: In my building they pull us from our planning period. […] But literally last year students would go to a classroom and there would be a sign on the door that said if there is no teacher here don't report to this classroom. They would then report to another classroom and there would be a sign on that door that would say if there is no teacher go to this classroom. So much so that the last place they landed was the auditorium. You could have three or four classes in the auditorium. With one [teacher]. Repeatedly, regularly that will happen. This teacher’s account of having to corral multiple classrooms into the auditorium because there were so many absent teachers may seem extreme, but similar accounts were provided by other teachers. What is even more concerning is that her perception is that having to combine classrooms due to chronic teacher absenteeism is regular occurrence in her building – not just occasional. When I followed up and asked this teacher how often classrooms in her building were combined because of chronic teacher absenteeism, she told me it was daily. She also explained that more often than not, the combined classrooms were not same grade level classrooms. For example, a fifth-grade class was combined with a first-grade class. When combinations like these happen, it begs the question, “How is student learning impacted when students from different 84 grade levels are combined into one classroom?” Not to mention, now one teacher has responsibility of approximately double the number of students when classrooms are combined due to chronic teacher absenteeism. The perception of the teachers in this sample is that chronic teacher absenteeism is a very real problem in their schools and districts, and yet many of those same teachers were chronically absent. Likewise, chronic teacher absenteeism touches all teachers – the chronically absent and the not chronically absent. Moreover, chronic teacher absenteeism is expensive and could have implications for student learning, as already demonstrated in the literature review (e.g. Clotfelter, Ladd, & Vigdor, 2007; Miller, Murnane, & Willett, 2008; Finlayson, 2009; Goldhaber & Hansen, 2010; Roby, 2013). Despite these understandings about the impact of teacher absenteeism and the importance of teacher attendance, most teachers in my sample were chronically absent. The findings resulting from the interviews and document analysis address the research question and take a detailed look into why teachers were absent, and are further discussed below. After a brief section that describes the absenteeism behaviors reported by participants, the central findings from this study are broadly captured by five sections. First, I will draw loose connections between chronic teacher absenteeism and some of the characteristics of the participants and districts they represent. Second, I will discuss emergent themes about job stress that relate to reported teacher absenteeism. Third, I will discuss emergent themes about perceived organizational support that relate to reported teacher absenteeism. Fourth, I will discuss an inductive emergent theme about the connection between financial support and reported teacher absenteeism. Finally, I will explore themes related to chronic teacher absenteeism that emerged while I was triangulating findings with document analysis of teacher attendance policies in each represented district’s collective bargaining agreement (CBA). 85 Reported Teacher Absenteeism While the purpose of this study is to understand what contributes to issues with teacher attendance, it was also necessary to ask for personal accounts of teacher absenteeism to better understand the absenteeism behaviors specific to participants in this study. As demonstrated earlier, chronic teacher absenteeism is defined by teachers missing 10 days or more during the school year, and previous research supports that there are high rates of teacher absenteeism in the United States. This definition of chronic absenteeism only considers missing 10 or more days in addition to the holidays and vacations already built into the school year and does not include missing days for professional development (that is a separate research stream). Of the 21 teachers interviewed, twelve teachers were absent for 10 or more days in the 2018-19 academic year, seven teachers were absent for more than five but less than 10 days, and two teachers were absent for less than five days in the 2018-19 academic year. There were not any teachers in the sample who did not miss any days of school. When the interviewees were asked if they exhausted their days off, meaning they took all days off allotted through their CBAs, five teachers reported taking all their days off. Three of those five teachers also reported having to take days off beyond their allotted days, which were unpaid in one instance, rollover days (explained later) were used in two other instances, and the remaining two teachers who exhausted their days off reported that the additional days off seemed to go unnoticed. In other words, their pay was not impacted even though they took off more days than allowed in their CBA. When interviewees were asked if they knew of teacher colleagues who had exhausted all their allotted days off in the 2018-19 school year, all 21 interviewees reported to know a co- worker who had. 86 Before diving into the reported reasons for teacher absenteeism, it is important to gain an understanding of how participant and represented district characteristics might be connected to chronic teacher absenteeism. The next section does just that and examines descriptive characteristics of the participants and represented districts to establish some context within which chronic teacher absenteeism occurs. Participant Characteristics, District Characteristics, and Chronic Absenteeism In an earlier section titled “Participant Overview,” I provided two tables that described characteristics of the represented districts in my sample (Table 2) and characteristics of the participants in my sample (Table 3). While connecting chronic teacher absenteeism to characteristics of my participants and represented districts was not a central focus of this study, I would be remiss to not discuss possible emergent patterns, especially because they might provide some context around circumstances within which chronic teacher absenteeism occurs. However, the discussion that follows in this section is by no means evidence of causation or even concrete patterns because the sample is very small and certainly not representative of the broader teacher population, especially because snowball sampling was used. For example, as noted in Table 3, all participants are white meaning all chronically absent teachers in my sample were white, but this is not to say that white teachers are more likely to be chronically absent. That statement is simply something this study cannot speak to. Rather, this section reviews observational findings specific to this study as they pertain to participant and represented district characteristics and potential connections to chronic teacher absenteeism. With regards to participant characteristics, chronic teacher absenteeism seemed to spread across gender, years of teaching experience, education level, age range, and teaching grade. There appeared to be an ever-so-slight pattern of chronic absenteeism being more common 87 among teachers who had six-to-ten or ten-to-twenty years of teaching experience and, similarly, teachers between the ages of 30 and 49. However, this was just an observational finding and warrants further study. Since the snowball sampling method resulted in a sample of only white teachers and the majority of the sample being union members, I cannot make statements about whether or not race/ethnicity or union membership are reflective of the types of teachers who may be likely to be chronically absent. The story is slightly different when examining the association between chronic teacher absenteeism and the characteristics of represented districts. As a reminder, the districts in this study were used because they were anecdotally identified as districts with high rates of teacher absenteeism. While chronically absent teachers were spread across all seven represented districts, there were two districts in which all interviewed teachers were chronically absent – District 2 and District 4. This may have been due to coincidence or it could be evidence of a pattern, but I was unable to obtain broader data (such as administrative/human resources data) to substantiate either of these possibilities. These two districts came from different regions of Michigan, but they were both classified as small cities. Of the seven districts represented by the sample, these two districts also had the lowest percentage of white students (in other words, the highest percentage of minority students) and also the highest percentage of economically disadvantaged students. This might suggest districts with a higher percentage of students who are minorities and/or districts with a higher percentage of students who are economically disadvantaged more commonly have chronically absent teachers. One chronically absent teacher from District 2, who brought attention to the fact that most of the students in her district were minorities and “free and reduced lunch” (i.e. economically disadvantaged), explained the source 88 of some of her experienced job stress and needing a mental health day as stemming from the challenges faced by the students she taught: I think just the ratcheting up of the needs of our kids. I mean, we have a lot of kids living in trauma. It’s just, it’s amazing. Just. It’s staggering. You know, some of the circumstances that these children live in and in school we just don’t have the right kind of staff or the money to support them. I mean, we put a lot of supports into behavioral supports for our kids, […] but it’s just a drop in the bucket compared to what they need. And I mean, they need therapists and we can’t, you know, we don’t have the funds to hire as many as they need. So, I’m expected to be their teacher and their therapist and it’s just too much. Earlier in the interview, this teacher identified challenges with student behavior as contributing to increased job stress and using paid time off for job-related mental health. In the above quote, the teacher makes further connections between how the high needs of her students connect to behavioral issues, which connects to supports for students (or lack thereof) and what kind of toll it takes on the teacher when she is expected to fill in voids. This evidence supports a possible connection between districts with a higher percentage of students who are minorities and/or districts with a higher percentage of students who are economically disadvantaged and chronic teacher absenteeism. However, the very small sample size and snowball sampling method means this potential connection between district characteristics and chronic teacher absenteeism does not indicate causation. Rather, this finding provides a snapshot of the district context within which chronic teacher absenteeism occurs. Further empirical exploration is needed to substantiate this finding. These loose connections between chronic teacher absenteeism and some of the characteristics of the participants and districts they represent might suggest chronic teacher absenteeism is more common among teachers who have six-to-twenty years of teaching experience, teachers who are between the ages of 30 and 49, or districts with a higher percentage of students who are minorities and/or students who are economically disadvantaged. To make 89 this connection more granular, it might be the case that the underlying reasons for chronic teacher absenteeism (discussed next) are more common among teachers and districts with these specific characteristics. While these findings are specific to this study and require additional empirical research, at minimum they provide some context for which chronic teacher absenteeism occurs for the participants of this study. The next section continues to examine teacher absenteeism by taking a deeper look into the reported reasons for teacher absenteeism. Reasons for Teacher Absenteeism All interviewees were asked what they used their days off for and they mentioned multiple reasons that can be divided into two groups: personal reasons and job-related reasons. The personal reasons for teacher absence included personal illness, vacation, family obligations, and continuing education. All 21 interviewees responded they used some of their paid time off for personal illness. When probed further about type of illness, their responses were specific to physical illness like the flu, a viral cold, pneumonia, etc. Eighteen interviewees responded another reason they used their paid time off was for vacation. Another personal reason for using paid time off that emerged was related to family obligations. Eleven respondents mentioned having to miss work for responsibilities related to children, such as childcare, school events for a child, or a child’s illness. Similarly, five interviewees indicated they used some of their paid time off to care for other adult family members, such as spouses and parents. The final personal reason for missing work that emerged was for completing graduate coursework, as mentioned by three interviewees. These personal reasons for teacher absence are certainly not exhaustive of all personal reasons for teachers missing work, but they do offer a snapshot into the some of the personal-related reasons teachers may be absent. 90 The second group responses about reasons for teacher absenteeism fell into was job- related reasons. All 21 interviewees responded that they used paid days off for job-related mental health, as evidenced by one teacher who said, “There’s been years that are more challenging than others and this was one of those years. I had a really tough group of kids. There were some days I just took. I planned ahead and I just needed a day off.” This account was in response to a question that asked for further explanation after the teacher had identified mental health as a reason for being absent. In this case, the teacher felt the need to miss a day of work for a mental health day because she was having a particularly challenging year. After identifying mental health as a reason for being absent, another teacher who exhausted all of her paid days off and took additional unpaid days further explained, “Every once in a while, I’ll just need a day and I’ll take a personal day and I’ll schedule to call in the morning or the night before. It’s just kind of my way of taking a break. I know there are other teachers, too, who feel like I just need a break. I’m okay not getting paid this one day and I’m going to take it.” This teacher indicated needing a mental health break and choosing to miss work even though she did not have any remaining paid time off days and knew her pay would be docked. Similar responses about using paid time off for job-related mental health reasons were echoed by all 21 teachers interviewed. To note, the line between personal-related mental health and job-related mental health is blurry because stressors can have a cumulative effect. However, the teachers interviewed in this study clearly pointed to job-related mental health as a concrete reason for using paid time off. In fact, not one participant interviewed listed personal-related mental health as a reason for using paid time off. The other job-related reason interviewees identified for using their paid time off was to catch up on work, which was reported by 11 teachers. When further probed, respondents listed 91 grading and preparing for teacher evaluations as reasons for needing extra time to catch up on work and using one of their paid days off. One teacher explained: I’ve used [days off] to catch up with my workload. […] It had to do with report cards. Our district still uses report cards, and you often have like a lot of writing samples to grade student work and to grade those are very time consuming on top of then doing the grades and then writing the comments. […] And it’s quite time consuming. So that’s why I’ve taken [days off]. When there’s a lot to score, I’m like, it’s okay taking a day off and then like doing grading all day. This teacher later explained that she had colleagues who engaged in the same practice of using paid time off to complete grading. Her quote suggests that her districts’ laborious grading and report card process resulted in having to use paid time off to complete the process. Taken together, teachers in this sample definitely used their paid time off and they used it for a variety of reasons. While just over half of the participants in this study were chronically absent, all participants reported to have colleagues who exhausted all their days off. Also noteworthy is that all participants indicated they used some of their paid time off for job-related mental health days. However, chronically absent teachers reported using more of their paid time off for job-related mental health than for personal reasons. The opposite was true for teachers who were not chronically absent. A deeper look at the job-related reasons for chronic teacher absenteeism is further explored next. Job-Related Reasons for Teacher Absenteeism The personal reasons for missing work mentioned above, such as personal illness, vacation, family obligations, and continuing education, are fairly self-explanatory. However, the job-related reasons for missing work, specifically missing work for job-related mental health, are not as clear-cut. For example, it is not clear what job-related factors contribute to teachers feeling the need to take a day off for mental health, especially for those teachers who were chronically absent. All 21 participants indicated using paid time off for job-related mental health, which 92 signals a real pattern with teacher absenteeism. However, the difference between teachers who were chronically absent and those who were not was that chronically absent teachers took more days off for job-related mental health than those not chronically absent. Across the sample, interviewees who missed less than 10 days of work in the 2018-19 school year (i.e. non- chronically absent teachers) reported only taking one to two days of paid time off for job-related mental health, while chronically absent interviewees reported taking six or more days of paid time off for job-related mental health. In one alarming account, in response to me asking a teacher what she used most of her paid time off for, she responded: It is not sickness and it usually isn't even my own kids. It is almost always a mental health day. Here's what I’ll say. There are cases where it is a doctor's appointment but I’m intentionally scheduling doctor's appointments during the school day so that I can take it off. Because it is so stressful every day going in, I need a day. But, for the most part, when I'm sick, if I'm a little sick, I come in because I need to save the day for when I'm so stressed I have to have it. This response came from a chronically absent teacher who indicated using paid time off for job- related mental health was more important than using paid time off for physical illness. Job-related factors contributing to using paid time off for a mental health day fell into two categories. Some responses indicated increased job stress driven by negative student behavior, large class sizes, and accountability pressures, as contributing to needing a day off for mental health. Other responses suggested low perceived organizational support sourced from minimal principal and administrative support as contributing to both job stress and needing a day off for mental health. These categories are discussed sequentially below. Job stress and student behaviors. Many teachers (12) expressed high job stress because of student behavior issues, which had a profound impact on their absenteeism. These 12 teachers came from different school contexts and had a range of teaching experience. After one more 93 experienced, chronically absent teacher identified an increase in job stress as a reason for taking mental health days, I asked what contributed to her job stress. She responded: Students. Student behaviors are getting significantly worse. I had major behavior issues this year. I’m really having problems. They’re really volatile. Students yelling and screaming and behaviors that are just really tough. And I had multiple kids like that in my classroom. This teacher’s account of issues with student behaviors is filled with negatively charged words, such as “tough” and “volatile,” suggesting the teacher was feeling overwhelmed by the negative student behaviors. According to the teacher, these negative student behaviors contributed to feelings of stress, leading her to take days off for mental health. Another teacher, who mentioned she never took mental health days in the past but had taken them more recently because of increased job stress, gave a similar account about job stress and student behavior: So, I think my stress is coming from the students. And like I said this is the only school I've ever worked in, so I don't know if it's the same for everywhere but the behaviors of the students in our school seem to be worse than behaviors from what I remember going through school. The students have a fairy nonchalant attitude about teachers and their authority and giving them respect. They just kind of, you know, “I don't have to listen to you. You're not my you're not my parent. I'm going to do what I want to do.” So, it's like they don't have respect for the teachers. So, it's very hard to manage behaviors from students like that. The perspective of this newer teacher was that negative student behaviors, specifically students disrespecting teachers, had increased from her frame of reference of being in school herself. While this teacher was not chronically absent, the disrespect she experienced increased her job stress, and thus resulted in her being absent from work, although her absences totaled to less than 10 days. This teacher’s account is important because, although she was not chronically absent, she indicated a recent increase in using more paid time off for mental health days because of job stress. This pattern, demonstrated by both experienced and newer teachers, deserves attention so 94 it can be reversed and so the non-chronically absent teacher does not eventually become chronically absent. Yet another experienced teacher gave a very detailed example of the negative student behaviors he’s experienced that have contributed to increased job stress: I think there's just more disruptive behavior. There's more kids in trauma. You know I’m seeing more and more anger. I’ve had a situation where I had to empty out my classroom because a student is going off and throwing a fit and you know throwing desks and doing all that kind of stuff. A first grader! In this particular situation, the chronically absent teacher had to not only think about the disruptive student, but also carried the stress of worrying about the safety of the other students in his classroom. This interviewee mentioned this was not an isolated incident and he was regularly on edge about the behavior of this particular student, which took its toll on his mental health. As discussed in a later section, school and district supports for student behavior were lacking for this teacher, too, which exacerbated the negative impact on his job-related mental health. Because of the increased job stress induced by worsening student behaviors, he used some of his paid time off to mentally recover from a particularly stressful day. Job stress and class size. Increased job stress driven by negative student behaviors was exacerbated by large class size, as mentioned by five of the 14 teachers who discussed student behaviors having an impact on their absenteeism. One chronically absent third grade teacher said, “My school has really large classes. I have 32. When you look at class size and if you have difficult to manage students and a large class, that’s very stressful.” Negative student behaviors alone are stressful, but when a teacher has a large class coupled with students who have behavioral issues job stress can be even greater, leading to teachers taking days off for mental health. 95 Four additional interviewees mentioned large class size contributed to job stress independent of negative student behaviors, suggesting class size inadvertently impacted teacher absenteeism. When I asked one teacher what it was about large classes that made her job stressful, she responded: I can’t give my students individual attention and address like individual needs. I want to do more for them and want to like personalize what each student needs. I have like 36 kids my room and it's so, so hard to get to what every kid needs and to hear each of their voices. This chronically absent teacher experienced increased job stress because she felt unable to provide adequate attention to each of the students in her large class. Her response suggests the large class size made her feel overextended and had a cumulative impact on her job stress, which led to her taking paid time off. The remaining three interviewees who mentioned large class size independent of student behavior as contributing to job stress and thus absenteeism, all shared similar sentiments of feeling overextended because they were not able to meet the individual needs of each of their students. Issues with student behavior and large class sizes clearly had an impact on teachers’ job stress, leading them to using paid time off for job-related mental health. Of the 18 teachers discussed above, 10 of them were chronically absent. Again, while 18 teachers are discussed above, the 10 chronically absent teachers missed six or more days of work due to job-related mental health, which they attributed to job stress driven by negative student behaviors and/or large class size. Job stress and accountability. In addition to negative student behaviors and large class size being sources of increased job stress, 16 teachers reported accountability pressures, such as demands related to standardized testing and teacher evaluations, as substantially contributing to their job stress and thus impacting their absenteeism. Of these 16 teachers, 11 reported using paid time off to catch up on work and 12 were chronically absent. Generally, these teachers said 96 accountability pressures resulted in high job stress because of the increased workload and constant pressure to perform. One more experienced chronically absent teacher explained: Changes to accountability have absolutely impacted my stress. How I teach and how I approach my work. When I look at where I started 20 something years ago to now. I thought I worked hard then. I definitely work a lot harder now. I just feel like there is more stress. There’s so much more paperwork. There’s so much more out-of-the-classroom responsibilities. This teacher’s perception was that accountability has changed over time and resulted in more work for teachers, which increased job stress and impacted absenteeism. In a few instances, teachers indicated the increase in paperwork resulted from preparing for teacher evaluations and more students needing individual education plans or behavioral intervention plans, while their additional classroom responsibilities resulted from more required professional development. Related to accountability, another experienced chronically absent teacher specifically said about job stress and its influence on her absenteeism, “One of the sources of stress is the constant pressure of test scores.” When I asked her to further explain why testing was stressful and how it impacted her absenteeism, this teacher went on to explain that standardized testing and high- stakes accountability made her job incredibly prescriptive. She indicated that this left her feeling like she could not practice the art of teaching because she was constantly worried about checking curriculum boxes to make sure she covered what the students were going to be tested on, which had a cumulative effect on her feelings of job stress and eventually made her need to use paid time off. Another chronically absent teacher with fewer years of teaching experience shed more light on the job stress associated with accountability and standardized testing, when she explained: So, there is a lot of stress just in terms of test scores, like some of our pay is tied to how our kids perform because it is included in our evaluations. So, I feel like there is a lot of job stress in terms of like getting your kids to succeed because it affects our bonuses. 97 This teacher felt recent accountability changes to teacher evaluations, where student performance factors into teacher evaluation ratings, contributed to her feelings of job stress because of the pressure to perform and the high-stakes financial implications. This teacher indicated she missed multiple days of work for mental health, stemming from job stress. Similarly, recent accountability changes to teacher evaluations made another more experienced chronically absent teacher feel immense job stress, which influenced her absenteeism behaviors. After she indicated that she had used paid time off due to job stress, I asked this teacher what the primary source of her job stress was, and she responded: Oh, I would say I’m so overloaded. The amount of work we have to do for our evaluation. The amount of data and evidence that we have to provide is immense. We just have to prove everything that we do. You know, I know we have to prove ourselves. But the amount we have to provide and the time it takes to gather and upload and organize and submit and, you know, give a little caption or a blurb about what we learned from it and what we did that time. […]I’ve devoted this past year to my evaluation, and it could have been better spent planning more quality lessons for my students and grading and analyzing my student data more. It’s tedious. So, I would say my number one stressor was making sure I had that all up to date. This teacher’s perspective provided a detailed account of the work that went into preparing for evaluations from a teacher’s perspective. Not only was the workload related to preparing for the evaluation stressful, but it was also stressful because she felt it impacted her planning and preparation time. In the case of this teacher, regular job stress because of feeling overloaded with work factored into her use of paid time off, which contributed to her chronic teacher absenteeism. Also noteworthy and related to teacher evaluation, of the seven districts represented by participants, two districts punitively tied teacher attendance to teacher evaluations, as reported by all participants (six teachers in total) from each of these districts. All six of these teachers were chronically absent and had varying levels of experience. In other words, all six participants from 98 the two districts that strictly tied teacher attendance to teacher evaluations were absent for 10 or more days in 2018-19. While teacher evaluations are required as part of accountability reporting and there are specific guidelines districts must follow in their teacher evaluation formula, districts also have autonomy in determining some of the pieces in teacher evaluations. According to these six teachers from two districts, the districts created local policies that made it so teachers could not receive an evaluation rating of highly effective if they missed a certain number of days, even though the days were provided in the official attendance policy through the district’s CBA. All four teachers from one district similarly explained that their teacher evaluation policy did not allow them to receive the highest evaluation rating if they missed more than seven days and they could not receive the second highest evaluation rating if they missed more than 10 days, even though their CBA allotted them 12 days of paid time off without stipulations. So, if teachers used all 12 (or even 11) of their allotted paid days off, they could only receive minimally effective or ineffective evaluation ratings. In the other district that tied teacher attendance to teacher evaluation, the policy was described by one teacher when he said, “I cannot miss more than two days. If you miss more than the two days of work, you are taken out of the highly effective range automatically.” This policy was similarly described by the other teacher in the district and points to a common understanding about how absenteeism could punitively impact teacher evaluation ratings. Teacher evaluation policies are not negotiated and documented in a public CBA, but rather they are determined by district administration and thus these accounts of teacher evaluation policies are only perceptions. However, the perceptions of these teachers demonstrated the understanding under which they operated, including related to using paid time off, which is relevant to the context of this study about teacher absenteeism. According to the 99 teachers, their districts were motivated to tie teacher evaluations to teacher attendance because the districts were dealing with high rates of teacher absenteeism. The districts thought a negative incentive would deter teachers from using paid time off. However, one teacher explained the policy had an opposite effect when she said: It seems very unfair to me because you're given [paid days off]. I think it's made teacher attendance worse. Teachers seem to really resent that. The factors that were contributing to the attendance are still there and now worse. And so, I don't think that made any difference at all. The factors contributing to teacher absenteeism alluded to by this teacher included job stress driven by student behaviors and increased workload, as she explained when I asked for clarity. All six of the teachers from the two districts with these teacher evaluation policies were chronically absent and indicated that the practice of punitively linking attendance to teacher evaluations made them feel more stressed. Taken together, job stress, as a result of negative student behaviors, large class sizes, and increased workload and pressure to perform stemming from accountability, had a profound impact on teacher absenteeism for the participants in this study. While not all teachers who identified student behavior issues, large class sizes, and accountability pressures as contributing to job stress were chronically absent, all chronically absent teachers in this study identified these reasons as being main contributors to job stress, and influencing their decisions to take multiple days off (i.e. six or more days) for job-related mental health. Moreover, all participants in this study indicated they missed work due to high job stress resulting from student behavior issues, large class sizes, and/or accountability pressures, although some participants only missed one or two days because of job stress. One chronically absent teacher cleanly summed up her feelings about job stress and absenteeism when she said: 100 So, I mean overall, I wish we could get through to legislators and superintendents like if you gave us levels that we could handle as far as kids and workload and class size, then we would be there more often and we would be happier because we could handle it more efficiently. Her basic sentiment was that if she experienced less job stress, then she would be less likely to be chronically absent. Job stress, an elusive and intangible characteristic, clearly influenced teacher absenteeism to the point where some teachers used multiple days of paid time off for job stress, which contributed to their chronic absenteeism. Managing job stress is also important because, as one chronically absent teacher put it, “If it wasn’t so stressful I wouldn’t get sick as much and have to miss as much.” Three other teachers echoed concerns over the connection between job stress, physical health, and missing work. In other words, mental health can bleed into physical health and lead to increased absenteeism for illness, which actually arises from job-related mental health. If the sources of job stress are not addressed, there is a chance that more teachers could become chronically absent due to job-related mental health. While personal characteristics may influence how teachers deal with job stress and whether or not they need to use paid time off for job-related mental health, the teachers who were not chronically absent more frequently shared positive accounts of school and district support, which is discussed in greater detail in the next section about the influence of perceived organizational support on chronic teacher absenteeism. Perceived organizational support. In addition to high job stress having an impact on job-related mental health and thus chronic teacher absenteeism, many teachers (13) also expressed feelings of low perceived organizational support had a substantial impact on their absenteeism, 12 of whom were chronically absent. All 13 teachers who described feelings related to low perceived organizational support also described feelings of job stress, which suggests these two organizational constructs are related and low perceived organizational support might 101 mediate job stress. In fact, the management literature supports that perceived organizational support and job stress are related and interact with one another (e.g. Clegg, 1983; Sagie, 1998); however, they are treated as unique but related constructs in the management literature, which is why they are discussed distinctly here but with some reference back to job stress. Perceived organizational support is another elusive characteristic and is the “extent to which employees perceived that the organization valued their contribution and cared about their well-being” (Eisenberger, et al., 1990, p. 52). Responses related to perceived organizational support fell into two categories: support from the building (including the principal) and support from the administration (or district/central office). Responses related to perceived organizational support actually emerged in 20 interviews, and 13 interviews indicated an association between low perceived organizational support and teacher absenteeism. Of the 13 interviews in which perceived organizational support was said to impact teacher absenteeism, eight discussed both building and administration support, while five interviews mentioned only administration support. For example, when explaining why she used paid time off for job-related mental health, one chronically absent teacher said: I thought my principal was gonna be good but I ended up being very wrong. I was very disappointed in her. She became what we called was like a bobblehead. She just went along with everything from the superintendent. I was very disappointed in her and her decisions. […] She was not supportive of me and some decisions I wanted to make around student behavior. It was frustrating. So, yea, having a principal that has my back would make me want to come to work. In this case, the teacher had low perceived organizational support because she was not feeling supported by her building via her principal. Her quote demonstrates the relatedness between low perceived organizational support and job stress, as student behavior was identified as a stressor in the previous section. This teacher’s experience of job stress driven by student behaviors may have been amplified because she did not feel supported by her principal in her decisions related 102 to negative student behaviors. While she independently identified lack of principal support as contributing to her job-related mental health and impacting her attendance, a close review of her quotes gives the impression low perceived organizational support may have increased her feelings of job stress. Of the 20 total responses that had responses related to perceived organizational support, many (12) of the interviewees who mentioned building support spoke favorably of the support they received from their buildings, as evidenced by one participant who said, “I feel very supported by my principal. I feel like she really enjoys having me as a teacher in her building, and I think that really helps me feel valued.” This response came from a teacher who was not chronically absent and was in response to a question that asked if the teacher’s building or principal could do anything differently to positively impact teacher attendance. This teacher’s response suggested she had high perceived organizational support from her building, which contributed to her enthusiasm to show up to work and thus positively impacted her attendance. Similarly, in response to the same question about how building or principal support impacts teacher absenteeism, another participant said about her principal: I think he’s very supportive. I think he listens. He respects people and he looks for chances to share what people are doing that could help other people. He’s in the classrooms all the time. I think I feel very respected and others do, too. This was another teacher who was not chronically absent, and her response demonstrated feelings of high perceived organizational support. Nine of the 12 interviewees who discussed positive building support were not chronically absent; however, there were three interviewees who had perceptions of positive building support, but they were still chronically absent. All three of these interviewees cited job stress stemming from negative student behaviors, class size, and/or accountability pressures as influencing their 103 attendance behaviors without making a connection to organizational support. This evidence supports that, while related, job stress and perceived low organizational support are idiosyncratic constructs. In other words, high perceived organizational support does not always mitigate feelings of job stress. Alternatively, all eight of the interviewees who discussed negative building support were chronically absent. This suggests that in situations of positive building and principal support, high perceived organizational support is fostered, and teachers are less likely to be chronically absent for job-related reasons. This finding is aligned with previous empirical findings in the management literature that found employees with higher scores on a perceived organizational support survey also had significantly higher rates of attendance (Eisenberger, et al., 1990). Conversely, in situations of negative building and principal support, there is evidence of low perceived organizational support and self-reported absenteeism tends to be more job related than due to personal circumstance. In addition to lack of principal and building support being a source of low perceived organizational support, 13 teachers reported a lack of support from district administration impacted their mental health and contributed to their absenteeism. Twelve of the 13 participants who gave responses related to low perceived organizational support because of lack of support from district administration were chronically absent. When asked about how support from the district administration impacted their absenteeism behaviors, participants noted that they felt the administration took away a lot of their planning time, so they sometimes took days off work just to catch up on work, as reported by 11 participants. One chronically absent teacher explained: Right now, we get 90 minutes a week to do all of our planning. It’s short. And in two of those a month are taken by district [professional development]. I had 50 minutes every single day to do my planning in my old district. […] So, I definitely 104 take time off a couple of times per year to catch up on my grading. Even then I’m just barely caught up. This experienced teacher explained her frustration that over her district tenure, the district had slowly chipped away at teachers’ planning time. Not only had her planning time been reduced over time, but the planning time she was allotted was often consumed by professional development required by the district. She implied that she felt like she did not have enough time to get everything done in preparation for her teaching, which resulted in her using paid time off to catch up. Another chronically absent teacher, who was newer to the teaching profession, gave a similar example of her frustrations with her districts’ support when she explained the district decided to divide teachers’ allotted 75-minute planning into ten-to-fifteen-minute chunks throughout the week rather than giving them an uninterrupted block of time. The teacher insinuated this practice left the teacher feeling like she could never accomplish her planning because it was disruptive and inefficient to have her planning time divided into small spurts. She went on to explain that the majority of her planning happened at home and she periodically used paid time off so she could adequately and efficiently plan her lessons. Ultimately, she decided that using paid time off was worth the tradeoff so she could feel prepared with well-planned lessons for her students. Her district did not support her need for concentrated planning time, which led to low perceived organizational support and influenced her absenteeism. Other teachers displayed low perceived organizational support when they noted administration did not address worsening student behaviors, which led to the staff feeling unsupported, stressed, and needing to take a day off, as expressed by the following account: There needs to be more immediate relief for difficult to manage students. Like remove the student from the situation if needed but that’s just temporary. […] I just think that there overall needs to be a better support system for classroom teachers 105 and included with that perhaps is more management strategies for teachers within the classroom. Since we don’t have that the stress is higher, and I need more of a break. This chronically absent teacher’s response demonstrates that she felt unsupported by her district with regards to how student behaviors were addressed and managed. Again, this quote demonstrates the connection between low perceived organizational support, job stress, job- related mental health, and teacher absenteeism. In this particular instance, low perceived organizational support stemming from district administration heightened the teacher’s feelings of job stress and influenced her withdrawal behaviors. Dissatisfaction with districts’ responses to and management of student disciplinary issues was echoed by many other participants, including by one chronically absent teacher who explained her district’s policy around student’s with behavioral issues: I think in our district we have so many behavioral issues. And staff resources to help you with that. You know you have to keep these kids in the classroom, and we don't you know send the kids home. […] We try not to use that option to send them home. They need to be in school. A lot of your behaviors will leave the classroom for maybe a few hours and then they're sent back and so you're continually teaching and teaching, you’re teaching as much as you can with those behaviors. And after a while you know every day of the same behavior is just coming right back to your classroom. That’s the policy. It stops your day is it. It's extremely stressful. The low perceived organizational support and increase in job stress demonstrated through this quote stemmed from both the teacher not feeling supported both by the district’s policy around addressing student behavioral problems and also from the teacher feeling like staffing resources were lacking to support her in dealing with student behavior. In short, in situations of poor district support for student behavior management, teachers’ perceived organizational support suffers and reasons for higher self-reported absenteeism tend to be more job related than due to personal circumstance. Furthermore, in some instances, situations of poor district support for 106 student behavior management impact teachers’ perceived organizational support and job stress, which duly influence teacher attendance. There also seemed to be a general frustration with and separation between district administration and teachers, especially for chronically absent teachers. One chronically absent teacher explained that she “wished my administration were to actually come in and observe me” so they could better understand the work she did and the diverse challenges she faced on a day- to-day basis. Another participant felt “overlooked” by her district because she felt like they did not provide her with adequate support and resources, which also led to perceptions of low organizational support. These two teachers explicitly mentioned lack of district support as influencing their need for taking job-related mental health days. Many other chronically absent teachers expressed dissatisfaction with their districts’ support because they felt they were not provided with adequate resources to perform their job, which had an effect on their job-related mental health and attendance behaviors. Responses about limited resources ranged from not having access to supplies (e.g. paper, books), to not having access to resources for new and constantly changing initiatives, and to not having adequate support staff, such as instructional aides, interventionists, and school counselors. With regards to needing to take a mental health day because of lacking support staff, one chronically absent teacher lamented: Overall, there needs to be more interventionists for the children that need it and it is difficult. We have one counselor for our four elementary buildings. Four. And you know she works her tail off, but we could easily have one counselor in our building. She tries to support everyone but she’s doing more putting out fires with like your highest fliers. So, I just think that there overall needs to be a better support system for classroom teachers. […] You need some realistic support that makes your job easier, not harder. When you ask for support, you walk away from that meeting with ten things that they want you to do that you know you just can’t do. Now your stress is higher than before [you asked for support]. So sometimes I think teachers feel like I’m not you know I’m not going there because they’re just going to give me more work to do and I can’t do what I have already. 107 This teacher’s feelings of being overwhelmed because of lack of staffing support from her district administration is palpable. She pointed to needing additional human capital to not only support students but also to support her so she could perform her job. Additionally, she felt the culture in her district discouraged her from asking for support when she needed it because she was concerned more would be added to her plate when she was already feeling overloaded. As indicated in the quote, her district’s culture and climate, which facilitated her low perceived organizational support, impacted her job stress levels. The quote from this teacher came in response to me asking her to explain why she used paid time off due to not feeling supported by her district. A similar account came from a chronically absent teacher who explained that she was no longer just a teacher, but rather was a teacher, disciplinarian, counselor, interventionist, and more because her district also did not have the human capital to support the diverse needs of the students in her district – it all fell on the shoulders of the teachers. Having to wear these many hats meant teachers spent less time practicing their practice of teaching and impacted their perceived organizational support, job stress, job-related mental, and thus their attendance behaviors. In sum, lack of building (e.g. principal) and administration (e.g. central office, district) support resulted in low perceived organizational support, which also affected job stress, and evidence of low perceived organizational support was a common pattern among chronically absent teachers in my sample. As demonstrated above, the management constructs of job stress and low perceived organizational support are related, as evidenced by this sample where many (but not all) of the chronically absent teachers had responses related to both concepts. The difference, however, is in how respondents constructed their responses with regards to the reasons for taking the paid time off. For example, responses related to perceived organizational 108 support specifically referenced to building and/or administration support. In some of these cases, respondents tied perceived organizational support to job stress, but in other cases they did not. The presence of one or both of these factors influenced teachers’ job-related mental health and, more likely than not, when one or both of these factors was present, there was higher self- reported teacher absenteeism. The findings about chronic teacher absenteeism being related to job stress and low perceived organizational support largely resulted from deductive analysis, with some support from inductive analysis. Another theme emerged through inductive analysis, which is the connection between job satisfaction as manifested through financial support and reported teacher absenteeism. This emergent theme is explored in the next section. Dissatisfaction and financial support. An unexpected pattern emerged during my data analysis process. Eleven of the 12 chronically absent teachers interviewed had responses about job dissatisfaction stemming from their salary and compensation, which influenced their absenteeism because either they worked another job to compensate for their salary dissatisfaction or they felt they were entitled to their paid days off to balance out their perceived inadequate salary. The drivers behind salary and compensation dissatisfaction are similar to the drivers behind high job stress and low perceived organizational support and include feelings of increased workload, performing additional duties above and beyond teaching, and not having enough district financial support for classroom materials and resources. In other words, teachers expressed feelings of dissatisfaction with regards to their salary and compensation because their teaching circumstances demanded more of them, but their salary and compensation did not recompense for these demands. This pattern came up in one of two ways. First, eight chronically absent teachers flat out expressed dissatisfaction with their salary. To note, a question about 109 salary and compensation was not asked of the participants. But rather, eight chronically absent interviewees independently introduced the topic of salary dissatisfaction. When one chronically absent teacher was explaining that she used her paid time off because she was “owed” more, she described the duties teachers were expected to perform “these days” in addition to teaching when she said, “I think more money should go to the teachers.” She was implying that teachers are expected to do more than teach and, as such, they should be compensated accordingly. Her perspective was that her exhaustion of paid time off made up the difference for what she was owed to perform additional duties. When explaining why he was absent, another chronically absent teacher explained, “I haven't felt very valued and I haven't felt financially compensated for the work that I do.” This teacher’s quote demonstrates low morale and low satisfaction with regards to his job because of lack of general appreciation and lack of financial appreciation, which is why he decided to use the majority of his paid days off. In fact, six of the eight chronically absent teachers mentioned they worked another job throughout the year in addition to teaching. This is often called “moonlighting” and serves as an additional source of income. Interviewees indicated needing additional income to not only keep up with financial responsibilities, but also to fill the perceived void resulting from feelings of underappreciation as demonstrated through unsatisfactory salary and compensation. Some interviewees even mentioned using paid time off from their teaching job to complete moonlighting-related work. Interestingly, salary, compensation, and moonlighting did not come up in the other 13 interviews, the majority of which were with teachers who were not chronically absent. As such, this suggests chronic teacher absenteeism may be more likely to occur when teachers are dissatisfied with their financial compensation. 110 The second way the pattern of dissatisfaction with regards to salary and compensation emerged was when chronically absent teachers explained they viewed their paid time off as part of their total compensation package. Eight chronically absent teachers explained they viewed paid time off as part of their salary, so they tried to maximize their use of paid time off by using most, if not all of their allotted paid time off. This included using all their paid sick time, regardless of if they were actually ill. One chronically absent teacher explained: I look at those leave days, which I know are intended for sake of illness or personal use or whatever, as just simply those are my days to use and I’m going to use them because I don’t get paid as well as I should. His account clearly demonstrates that not only was he dissatisfied with his financial compensation, but because of it he viewed his paid time off as part of his total compensation package. This particular chronically absent teacher admitted he approached each school year with the mindset and plan to use most or all of his paid time off by the end of the school year, whether or not he needed it. He also explained that he started the school year by using his paid time off only on an as needed basis, such as for illness, but by the end of the school year, if he had remaining paid time off, he would take off days here and there “just for fun.” This teacher’s district even had a payout incentive where teachers could receive monetary compensation for unused paid time off upon retirement, yet he still decided to use all his paid time off by the end of the school year (greater detail about payout incentives is provided in the next section). Other chronically absent teachers relayed similar views around their paid time off because they “assume that these sick days or these personal days are part of my benefit.” Some chronically absent teachers mentioned both general dissatisfaction with their salary and viewing paid time off as part of their total compensation package, while others just mentioned one or the other. 111 In sum, the unexpected theme about the connection between satisfaction with financial support and reported teacher absenteeism that emerged through inductive analysis points to a connection between chronic absenteeism and dissatisfaction with compensation. Specifically, teacher attendance suffers in situations of salary dissatisfaction stemming from increased workload, performing additional duties above and beyond teaching, and not having enough district financial support for classroom materials and resources. Moreover, in situations of salary dissatisfaction, some teachers view paid time off as a part of their total compensation, and teachers rationalize their absenteeism by explaining their paid time off is owed to them as part of their compensation. In this study, salary dissatisfaction was only evidenced in interviews from chronically absent teachers. Approach to using paid time off. Looking across interviews from chronically absent teachers and non-chronically absent teachers, and reviewing the different patterns and themes thus far, revealed a difference in approach to using paid time off between chronically absent teachers and non-chronically absent teachers. Namely, all nine non-chronically absent participants in my sample implied that they tried to minimize their use of paid time off. One teacher, who missed less than five days in 2018-19, explained: I do not like to be gone. So, I will only be gone if I absolutely have to. […] Just because it's hard being gone. It takes a lot of work to get a sub ready for my job. It takes a lot of work. […] There's work you have to put in ahead of time and when you come back and it's actually much easier being at school than not being at school. So, my strategy is to not be gone. […] I just feel really dedicated and sometimes curriculum is so tight that I worry if I'm gone a day that it's going to put us behind. This teacher’s approach to using her paid time off was to only use paid time off if it was absolutely necessary. Being absent was disruptive to her students’ learning and also created more 112 work for herself because she had to prepare plans for substitute teachers and felt behind when she returned from being off. Similarly, another teacher explained, “When you leave the class, when you’re gone for a day and come back, it’s more stressful to me.” Again, this teacher, who was not chronically absent, explained that she planned to only use her paid time off when it was necessary. Finally, a teacher who was absent six days in the 2018-19 school year explained: You really have to weigh being out of the classroom if it’s worth it. It is always all that additional work that you have to do. […] It's almost not even worth being gone to begin with. It's like the workload that goes into taking days off is almost sometimes not worth it. I really have to weigh it sometimes because it's like hard work to be out of the office on top of the normal work that you do. It’s easier to just be there. For this teacher, the drawbacks of using paid time off sometimes outweighed the benefits so she tried not to be absent from work. These perspectives of non-chronically absent teachers, who tried to minimize their use of paid time off because it was more stressful for them to be gone, offer a contrast to all of the chronically absent teachers who reported needing to use paid time off because of job stress. This pattern of minimizing the use of paid time off was only present in interviews from teachers who were absent less than 10 days in the 2018-19 school. The approach these teachers took to using their paid time off was strikingly different from the chronically absent teachers who often sought to maximize their use of paid time off. In short, the teachers who were not chronically absent were very intentional about being conservative in their use of paid time off and only using paid time off for personal circumstances, such as illness, whereas this was not the case with chronically absent teachers. A final pattern related to chronic teacher absenteeism emerged through inductive analysis and document analysis of the teacher attendance policies outlined in each district’s CBA. 113 Document analysis is important because it provides contextual information around the teacher attendance policies the participants are subject to. In this case, when document analysis was paired with analysis of the interviews, a pattern inductively emerged. Namely, there appeared to be a connection between chronic teacher absenteeism and incentive policies that targeted teacher attendance. The next section further explores this emergent pattern. CBAs, teacher attendance policies, and incentives. Before diving into the connection between incentive policies and chronic teacher absenteeism, it is first necessary to provide some background and details about the CBAs that were examined during document analysis. I was surprised to find that each of the seven CBAs I analyzed had different policies around teacher attendance. While there were certainly some similarities, no teacher attendance policy was identical to another. When the seven collective bargaining agreements were examined, each agreement referred to the concept of paid time off in a unique way, including: “sick and emergency leave,” “personal leave days,” “sick leave,” “personal leave,” “leave pay,” “personal and family illness days,” and “leave days.” All of these terms referred to days teachers were paid for if they had to miss school (not including professional development days). There was overlap in the terms, but no two terms were exactly alike. Similarly, within and between districts, participants referred to their paid time off with different terms in the interviews. Even within interviews, participants switched their language when discussing paid time off, at which point I had to clarify if the terms had different meanings, which they did not. Participants referred to paid time off as “paid time off,” “vacation days,” “personal days,” “personal time,” “paid leave,” “sick days,” and “sick time.” The details of each attendance policy also differed. One CBA provided 13 “personal leave days” per year, another provided only five days of “sick and emergency leave” per year, 114 another provided 10 days of leave pay per year, while yet another provided two personal leave days and 10 to 12 sick leave days per year depending on contract duration. The remaining three CBAs also followed this pattern of having different teacher attendance policies. Some CBAs grouped personal and sick time together while others parsed them out. Furthermore, some CBAs provided a lot of detail around what constituted a sick day versus a personal day, or under what circumstance paid time off could be used, while other teacher attendance policies were as short as five lines of text and were quite vague. Even in the CBAs that provided more detail, there was still plenty of room for interpretation. In short, each paid time off policy varied by district and Table 5 provides a brief description of each district’s paid time off policy. For ease, I grouped descriptors of paid time off into two categories: personal days and sick days. For example, “vacation days” are captured under the column for personal days, whereas “sick time” is captured under the column for sick days. Table 5 also includes detail about offered incentives, which will be discussed in more detail later. 115 Table 5: Details from CBAs about Paid Time Off, by District Number of District Code Teachers (N) Number of Teachers Chronically Absent (N) Number of Number of "Personal" Days "Sick" Days Total Number of Paid Days Off Payout Incentive for Unused Accrued Paid Time off Upon Retirement or Severance Yes, paid out up to 250 days, sliding pay scale, depending on hire date ($35 to $47/day) Other Attendance Incentive Included in CBA No 13 0 13 10 to 12, depending on contract duration 12 to 14 No Yes, $200 per semester if 0 days are missed during the semester Yes, maximum incentive is $2,500 annually depending on days missed and FTEs No Yes, no limit, $30/day No Yes, paid out up to 95 days, rate is determined by base pay at severance or retirement, dependent on years of service Yes, $100 per semester if 0 days are missed during the semester Yes, paid out up to 90 days, Yes, annual bonus if four rate is determined by base pay at severance or retirement, dependent on years of service Yes, paid out up to 180 days at $60/day or less days are missed during the school year; amount varies by year Yes, $400 max incentive annually if 0 days used 1 4 1 2 1 2 1 4 4 2 2 2 3 4 District 1 District 2 District 3 District 4 District 5 District 6 District 7 2 3 10 0 14 12 10 0 5 0 0 13 10 5 14 12 116 Interview responses were also dissimilar and inaccurate when participants were asked to articulate the attendance policies as written in CBAs. Again, within and between districts, participants had different accounts of what they were allowed to use their paid time off for. While some participants said they could use it for sick and vacation time, others said it was only for sick time, yet others said it was all for personal time. The varied understandings imply that not only did teachers not fully understand the teacher attendance policy as outlined in their CBA, but within districts teachers were operating under different understandings of what they were allowed. This unanticipated but important finding indicates an organizational mismatch between teachers’ knowledge around how they are allowed to use their paid time off and how they are not. This is important because the organizational mismatch might impact how often and when teachers use their paid time off, which could influence chronic absenteeism. Likewise, if the varied understanding of teacher attendance policies impacts the practice of using paid time off, then teachers will use their paid time off for different reasons and at different rates, potentially leading to some level of perceived unfairness in the implementation of the attendance policy. This organizational mismatch and possible perception of unfairness is unhealthy to the school or district’s culture and climate, which could further contribute to teachers’ feelings of job stress and perceived organizational support. Moreover, some teacher attendance policies provided parameters around how paid time off days could and could not be used; however, interview responses indicated that districts did not enforce these parameters. For example, one district’s CBA said paid time off could not be used for routine medical and dental appointments. When I asked the teachers from this district to describe their teacher attendance policy, this caveat never came up. When I probed further about these specific parameters, one teacher was not aware of them and the other teacher was aware of 117 them but said the district never enforced them and never questioned her use of paid time off. In general, all interview responses reflected perceived apathy from the district with regards to the district’s enforcement of the teacher attendance policy and response to teachers’ use of paid time off. This response, or lack thereof, from the district was, in a sense, an “informal policy” or “informal agreement.” Such informal policies might make it easier or more appealing for teachers to use their paid time off and suggests that teachers are solely responsible for following the teacher attendance policy outlined in their CBA because administrator oversight is lacking. In fact, 10 teachers, both chronically absent and not chronically absent, from four different districts explained sometimes their building principals turned a blind eye, provided “comp days,” or filled in for teachers when they missed part or all of a day. In these scenarios, teachers did not submit an official request for time off. This is another example of an informal policy or agreement around teacher attendance. The perception was principals exercised some discretion and even favoritism with this practice, but the general sentiment was that principals understood that sometimes teachers just needed a break and they wanted to help. Whether or not district administration was aware of this informal policy and practice is unclear, but the teachers who described this practice also indicated the district never got involved. Given the common response from interviewees about how their districts responded to use of paid time off, chronic teacher absenteeism does not appear to be connected to how teacher attendance policies are enforced or monitored by the district. However, in two instances teachers exhausted their allotment of paid days off, took additional days off beyond their allotment, but did not lose pay for taking days off beyond their reserve. These teachers came from two different districts and each CBA explained that teachers who missed more time than they had accrued would be docked pay. The additional days these 118 chronically absent teachers took seemed to go unnoticed by the district as they did not lose pay. In only one instance did an interviewee report losing pay because she took days off beyond what she had available to her. This lack of attention by the district around teachers’ time off might suggest a connection between loose enforcement of teacher attendance policies and chronic absenteeism, but the present study is not equipped to substantiate this idea other than to speculate. Generally, teacher attendance policies from the seven represented districts varied, as did the districts’ implementation and enforcement of the policies. Although all seven teacher attendance policies were different, there was one similarity. Specifically, all seven districts’ teacher attendance policies allowed teachers to accrue their unused paid time off days from year to year. In other words, if a teacher was provided with 12 paid time off days per year but only used nine in a school year, the remaining three carried over to the next year. These are commonly referred to as “rollover days.” In the next year, that same teacher started the year with 15 days of paid time off instead of the standard 12. Of the 21 teachers in my sample, 16 had accrued rollover days from previous years in addition to the standard yearly allotment. The five teachers who did not have rollover days had exhausted their paid time off in previous years or were newer in their teaching careers (i.e. less tenure) and had fewer years to accrue unused days. In essence, the policy of being able to accrue unused days made it possible, even easy, for teachers to take 10 or more days off in a given school year without financial implications. Even though this policy was available to all teachers in my sample and 16 had rollover days at their disposal, only two teachers dipped into their rollover banks. The remaining teachers with rollover days saved their days for the next year. While all seven districts allowed for accrual of paid time off days from year to year, each district had different caps on how many total days teachers could accrue over time. For example, 119 one CBA allowed teachers to carry up to 20 days of paid time off year to year, another CBA allowed 60 days year to year, while another CBA allowed unlimited accrual of days. Any days accrued beyond the maximum rollover allotment were either paid out upon severance/retirement or forfeited. According to interviewees, this is a common practice across school districts. Given the commonality of this practice across districts, there does not seem to be an association between accrual of rollover days and chronic teacher absenteeism. Alternatively, some districts’ CBAs offered payout incentives around accrued days that might influence a teacher’s decision to be absent. All teachers interviewed were aware of the presence or absence of a payout policy for accumulated leave days, but none of the teachers described the policies with accuracy as outlined in the CBAs. In five districts, the CBAs offered a payout for unused accrued paid time off upon retirement or severance depending on years of service. In the five districts with this policy, the amount of payout per day and the number of days that could be paid out differed. For example, one district allowed a payout of up to 90 days at a rate determined by the base pay at the date of severance or retirement and dependent on the teacher’s years of service. Another district offered a payout of up to 180 days at $60 per day upon retirement. Yet another CBA outlined a step payout depending on the hire date of the teacher, where teachers hired before a given date could be paid out for up to 250 days at more than $40 per day. Table 5 above further summarizes these payout incentives. A potentially interesting connection with absenteeism is that in the two districts without a formal payout policy for unused leave days, all interviewed teachers were chronically absent. Of course, these districts and teachers might have had other characteristics that connected to chronic teacher absenteeism, but the presence or absence of this incentive may play an influencing role. 120 The presence of payout policies may have a double-edged influence on teacher absenteeism. On the one hand, the policies could encourage teachers to not use paid time off so they can accrue days for payout upon retirement, as demonstrated by one non-chronically absent teacher who explained how her district’s payout policy influenced her use of paid time off when she said, “I typically don’t let myself miss any more than four days if I can help it so I can save them for [the] retirement [payout]. So, if I haven’t taken any sick days, I will take my personal days and stop at that.” For this teacher, who was newer to the teaching profession, the retirement payout incentivized her to limit her use of paid time off so she could accrue days. On the other hand, the payout policies may encourage exhausting paid time off once teachers have saved up to the allotted threshold of the amount of days they can be paid out for. About her district’s leave time payout policy, one chronically absent teacher explained: It makes it so teachers want to save [paid days off] up so they can get paid out at retirement. But, since they're only paying you for 90 days, then beyond saving those up, you know, there’s that feeling of “I might as well use them.” This veteran teacher’s explanation suggests the payout policies had an incentivizing impact on absenteeism to a certain extent. Even if teachers were motivated to save 90 days to be paid out upon retirement, once the accrual leave days bank was maximized, teachers could then aim to take all additional earned paid days off because once the payout bank was filled, teachers eventually forfeited any additional accumulated days. For example, three chronically absent interviewees in this study mentioned they had maximized their bank of days that could be paid out upon retirement, so they did not feel as bad about taking more days off work when they needed them. One chronically absent veteran teacher explained, “When I was younger, I would try to strategize, and I would try never to take a day for myself no matter what because I knew I was going to get paid for those. I don’t do that now." In the past, this teacher would push through 121 an illness or extreme stress so she could save up her days to be paid out upon retirement, but now that her leave days bank was maximized she allowed herself the luxury of taking her paid days off with more regularity. Teachers’ approaches to accruing paid days off for the payout incentive seemed to be connected to years of teaching experience, in that the policy appeared to be less motivating to veteran teachers who had already maximized their payout accrual bank, as compared to more novice teachers who had not yet reached the limit. Though, when I asked teachers for what reasons they used paid time off, there were not any teachers who initially indicated they might as well use the days because otherwise they would lose them. Rather, initial responses were related to missing work for job-related mental health. Responses related to the “use them or lose them” mindset surfaced after additional probing and in response to questions about the teacher attendance policy as outlined in the CBA. The majority of responses with the “use them or lose them” mentality came from chronically absent teachers who had more years of teaching experience. More research is needed to determine if payout policies truly have a connection to chronic teacher absenteeism; however, there did appear to be a connection between chronically absent veteran teachers and their “use them or lose them” approach to using, even exhausting, paid time off. Three teachers from two districts indicated they were motivated by their district’s paid time off payout incentive to save up their days and not use paid time off. All three of these teachers were not chronically absent, were 50 years of age or older, and also mentioned being motivated by other incentive policies offered by their districts that targeted minimizing teacher absenteeism. The other fairly common teacher attendance incentive policy offered in five districts was an annual or bi-annual bonus for not using paid time off. Commonly referred to as an “attendance incentive,” the specifics of these policies varied by district. For example, to 122 “encourage employees to only use leave days when absolutely necessary,” one district’s CBA offered an annual bonus to teachers if they missed four or less days annually. The bonus was on a step-schedule, which ranged from receiving a maximum of four hundred dollars when zero leave days were used annually to two hundred dollars when no more than four leave days were used annually, with steps of 50 dollars for missing one, two, or three days. Another district’s CBA offered a bi-annual attendance incentive where teachers who missed zero days each semester received a bonus of two hundred dollars, for a total possible bonus of four hundred dollars in the school year. Table 5 above further summarizes the attendance incentives offered through each district’s CBA. To note, Table 5 only includes attendance incentives as explicitly written in the CBAs and does not include unwritten attendance incentives (positive or negative), such as a district’s decision to punitively tie teacher attendance to teacher evaluations. As mentioned above, three non-chronically absent teachers said they regularly aimed for the attendance incentive written in the CBA and tried not to use paid time off so they could receive the bonus. One teacher explained, “I’m really frugal. To get that 250 bucks kind of, you know, it motivates you. […] That little amount of money works for me.” For this teacher, the extra money was motivating, and her response suggests the incentive influenced her decision to use or not use paid time off. Alternatively, the other teachers who worked in districts with attendance incentives were aware of the attendance incentive but did not know the details of the policy and their responses suggested they were not motivated by the incentive. Nine of these 15 teachers were chronically absent. Not only were they not able to accurately articulate policy, but also, they were almost dismissive of the attendance incentive. When I asked one chronically absent teacher if her district offered any incentives around teacher attendance and to describe it, she responded: 123 Yes, cash. I don’t know what it is. It’s, I want to say, it’s like a measly 100 or 150 dollars or something. […] It’s very minimal. It clearly isn’t motivating enough for me to pay attention to it. It isn’t very much. This teacher went on to explain that it did not even cross her mind when she was deciding whether or not to use paid time off. Rather, she explained that her decision to use paid time off was influenced by her level of stress and how badly she felt she needed a day off. Many of the other chronically absent teachers felt similarly and mentioned that the incentive was too insignificant to consider changing their attendance behaviors. When I asked another teacher the same question, she responded in a similar manner by saying, “I’m never even close to [the attendance incentive], so I’m not sure exactly what it is.” In this similar account, the attendance incentive policy did not appear to influence the chronically absent teacher’s decision to use paid time off. In another response about an attendance incentive, a chronically absent teacher described her process of weighing the costs and benefits of using paid time off, when she explained: I’m not sure that anybody really cares about that extra [attendance] incentive. […] It’s not worth it for me, you know? If I want to take a day I’m gonna take a day. […] It’s not something that I base my entire decision on but I always kind of keep that in the back of my head. You know, I think, like, do I really need to take this day off or do I just want to? I don’t really care about the extra money. I try to think like, is this day worth that much to me? While this teacher did not completely dismiss the attendance incentive, in the 2018-19 school year she ultimately decided she needed to use the days off more than she needed the bonus, and she ended up missing 11 days throughout the school year. In other words, her paid time off was worth more than what the attendance incentive offered. She later went on to explain that maybe the incentive would influence her decision more if more money were offered. Even though she was chronically absent, the attendance incentive still influenced her absenteeism behaviors because the incentive was not great enough for her to consider not using paid time off. The idea 124 of the incentive not being great enough to influence attendance behaviors also came up in other interviews with chronically absent teachers, where one chronically absent teacher said, “Maybe if [the attendance incentive] was more money. Then maybe I’d think about it more seriously. They need to make it like a thousand dollars or something that’s substantial.” While these teachers suggested a greater incentive might influence their attendance behaviors, there are no findings in the present study to substantiate their speculations. Examining the relationship between attendance incentives and chronic teacher absenteeism is not a central focus of this study but given conversations around incentives came up in the majority of interviews it is an important topic to highlight. The connection between attendance incentives and chronic absenteeism is not clear; however, there did appear to be a pattern around how chronically absent teachers in my sample were only vaguely aware of and dismissive of the attendance incentives offered in their CBAs. Even if attendance incentives did influence teacher attendance behaviors, it would be remiss to not consider other factors, such as influences from job-related mental health, that might confound the influence of attendance incentives. Using document analysis of each represented district’s CBA triangulated some of my findings by confirming and disconfirming responses from interviewees. Triangulation uncovered an organizational mismatch between teacher attendance policies as outlined in CBAs and teachers’ knowledge of the policies. This organizational mismatch may have an influence on how teachers use their paid time off and for what reasons. Document analysis also provided me with detailed information about general teacher attendance policies and attendance incentive policies. These details allowed me to ask interviewees targeted questions related the policies and the reasons teachers give for being absent from work. Document analysis aided in revealing 125 some unanticipated, yet important findings about teacher absenteeism behaviors. These findings are further discussed in the discussion section. 126 DISCUSSION The research question for this study, “what reasons do teachers provide for their chronic absenteeism?” guided the collection of rich, descriptive data from teachers in school districts that had an anecdotal problem with chronic teacher absenteeism. The collected data provide insights into some of the elusive, intangible characteristics that may contribute to teacher absenteeism. Findings emerged both deductively and inductively and there were both expected and unexpected findings. As expected, the teachers in this study used their paid time off and they used it for a variety of reasons. More than half of my sample missed 10 days or more of work in the 2018-19 school year, making them chronically absent. Even those teachers who missed less than 10 days still missed many days of school. This provides some evidence that chronic teacher absenteeism exists. Additionally, while participants provided both personal (e.g. personal illness, vacation, family obligations, and continuing education) and job-related (job-related mental health) reasons for using paid time off, chronically absent teachers referenced using paid time off for job-related mental health more often than for personal reasons compared to non-chronically absent teachers. Specifically, chronically absent teachers in the sample reported taking six or more days of paid time off for job-related mental health, compared to teachers not chronically absent who only missed one to two days in the 2018-19 school year for job-related mental health. Further probing into the job-related mental health reasons for using paid time off through deductive analysis revealed two central themes centered on elusive, intangible characteristics. One theme around why teachers used paid time off for job-related mental health was because of increased job stress driven by negative student behaviors, large class sizes, and accountability pressures. The other theme around why teachers used paid time off for job-related mental health was because of low 127 perceived organizational support sourced from minimal principal and administrative support. These two themes were especially pronounced in interviews from chronically absent teachers. The findings related to chronically absent teachers using paid time off for job-related mental health because of increased job stress and low perceived organizational support were expected because the management literature supports links between employee absenteeism and job stress (e.g. Beehr & Newman, 1978; Gupta & Beehr, 1979) and employee absenteeism and perceived organizational support (Eisenberger, et al., 1986; Eisenberger, et al. 1990). While the management literature provides evidence of these non-causal relationships, the empirical studies do not solely focus on teachers as the population under study. The present study’s focus is on teachers, and the findings support that increased job stress and low perceived organizational support influenced chronic absenteeism as self-reported by teachers. In other words, in situations of high job stress and/or low perceived organizational support, teachers’ mental health suffered and reasons for higher self-reported teacher absenteeism tended to be more job related than due to personal circumstance. Through interview data and analysis, these two elusive organizational factors were conceptualized in relationship to teacher absenteeism. These two elusive organizational factors may point to challenges with culture and climate within the schools and districts under study, and more broadly for districts with reported high rates of teacher absenteeism. Being aware of the perceived associations between these two organizational factors and chronic teacher absenteeism offers a starting place for addressing chronic teacher absenteeism. Job satisfaction is an additional elusive and intangible organizational factor discussed in my literature review and the management literature evidences a non-causal inverse relationship between job satisfaction and employee absenteeism (e.g. Muchinsky, 1977; Steers & Rhodes, 128 1978; Clegg, 1983; Scott & Taylor, 1985; Garcia, 1987; Hackett, 1989; Chaudhury & Ng, 1992; Sagie, 1998). Likewise, the teacher turnover literature suggests certain organizational factors, such as salary and compensation, are antecedents to job satisfaction and subsequent teacher turnover (e.g. Ingersoll, 2002; Ingersoll, 2003; Ingersoll & Smith, 2003). While some of the coding categories discussed earlier may be related to job satisfaction, they more clearly spoke to themes related to job stress and perceived organizational support. This is not surprising given all four organizational factors discussed in the literature review are related and interact with one another (e.g. Clegg, 1983; Sagie, 1998). Nevertheless, responses related to job satisfaction emerged when chronically absent teachers regularly referenced dissatisfaction with regards to salary and compensation. Dissatisfaction with salary and compensation stemmed from increased workload, performing additional duties above and beyond teaching, and not having enough district financial support for classroom materials and resources. My data analysis of this pattern suggests that in situations of salary and compensation dissatisfaction, teacher attendance suffered because teachers viewed paid time off as part of their total compensation and thus they tried to maximize their use of paid time off. This finding speaks to a small part of job satisfaction, in that research supports that salary and compensation satisfaction fall under the job satisfaction umbrella (e.g. Muchinsky, 1977; Ingersoll, 2002; Ingersoll, 2003; Ingersoll & Smith, 2003). The topic of teacher salary has drawn recent attention nationally (e.g. Kiersz & Perino, 2019) and within the state of Michigan (Citizens Research Council of Michigan, 2019), so while this finding was slightly unexpected it was not surprising. This finding speaks to the climate around teaching within each of the districts included in this study and suggests a connection between teacher chronic absenteeism and job dissatisfaction as evidenced through salary and compensation dissatisfaction. Moreover, this finding speaks to a 129 broader policy conversation related to teacher salary and will be briefly discussed further in a later section. The potential non-causal connection between chronic teacher absenteeism and salary and compensation dissatisfaction suggests chronically absent teachers took it upon themselves to compensate for their unsatisfactory salary by maximizing their use of paid time off, which contributed to being absent 10 or more days in a school year. In fact, across findings there seemed to be a common thread of teachers not feeling respected or valued in the climates within which they worked. This was true for most teachers in my sample, but especially for chronically absent teachers. All of the findings discussed above that have a perceived influence on chronic absenteeism, such as job stress, perceived organizational support, and job satisfaction, as driven by negative student behaviors, large class sizes, accountability pressures, poor school and district support, and unsatisfactory compensation, suggest that teachers are working in climates where they do not have the resources or support they need. These factors have a compounding effect and one way the impact is felt is through teachers being chronically absent. In short, it appears that we are not investing enough in the largest and perhaps most important school resource, teachers. This may point to a general lack of support and lack of resources allocated to local schools and districts in terms of funding from the state government, which constrains the support local schools and districts are able to provide to teachers in the first place. In effect, being absent may be a teacher’s way of exercising some control in an under-resourced and unsupportive working environment. Unfortunately, while absenteeism may provide teachers with some relief, chronic teacher absenteeism exacerbates many of the challenges faced by schools and districts and perpetuates a negative culture and climate. 130 Another unexpected finding was that chronically absent teachers were only vaguely aware of and dismissive of their district CBA’s teacher attendance incentive policy compared to non-chronically absent teachers, who were well-aware of and able to accurately describe their district CBA’s teacher attendance incentive policy. Not only did their awareness of the teacher attendance incentive policies as outlined in the district’s CBA differ, but chronically absent teachers differed in their approach to using paid time off from not chronically absent teachers. The teachers who were not chronically absent were intentionally conservative about using their paid time off and they indicated they were motivated to do so because they wanted to receive the attendance incentive offered by their district. Non-chronically absent teachers also specified that they tried to only use their paid time off for personal circumstances, such as illness. On the other hand, chronically absent teachers had a “use them or lose them” approach when deciding to use paid time off and their reasons for using paid time off were more often job-related. To be clear, this is not to say that salary dissatisfaction and/or attendance incentives cause chronic absenteeism. Rather, salary dissatisfaction and/or attendance incentives may influence chronic teacher absenteeism and could be confounded with job stress and low perceived organizational support. In fact, the findings in this study related to teacher absenteeism and incentives further contribute to the mixed empirical findings around using incentive policies to reduce teacher and/or employee absenteeism (e.g. Ehrenberg, Ehrenberg, Rees, & Ehrenberg, 1989; Clotfelter, Ladd, & Vigdor, 2007; DeNardo, 2007; Ahn & Vigdor, 2011; Hanover Research, 2012; National Council on Teacher Quality, 2014), as described in the literature review. Although there were some non-chronically absent teachers who reported their attendance was impacted by attendance incentives, there were also chronically absent teachers from the same districts, who were subject 131 to the same attendance incentives, and were still chronically absent. Therefore, it is very unclear what kind of impact attendance incentive policies had on the attendance of teachers in this sample. Furthermore, even if attendance incentives were present, they did not seem to change the attendance behaviors of teachers who were dealing with organizational issues, such as job stress and low perceived organizational support. As evidenced by the chronically absent teachers in this study, it appears that attendance incentives, which tended to be monetary, did not outweigh the cost of coming into work when a paid time off day was needed for job-related mental health. This unexpected finding adds to the mounting inconclusive empirical evidence surrounding the impact of attendance incentive policies on reducing teacher and/or employee absenteeism. Furthermore, this finding signals that teacher attendance is a complex behavior in which multiple elusive inputs are at play when a teacher is deciding to use paid time off. For example, salary dissatisfaction and/or attendance incentives may influence chronic teacher absenteeism but could also be confounded with job stress and perceived organizational support. This might suggest that attendance incentives oversimplify the problem of teacher absenteeism by assuming monetary incentives alone will solve the problem, which is inappropriately placed if there are pervasive organizational and culture and climate issues that contribute to a teacher’s decision to use paid time off. The findings related to attendance incentives were triangulated through document analysis of each represented district’s CBA. During document analysis, it also became clear that teachers did not have a comprehensive or clear understanding of their district’s attendance policy as outlined in their CBAs. This was another unanticipated but important finding and points to an organizational mismatch between districts’ teacher attendance policies and teachers’ knowledge around how they are allowed to use their paid time off and how they are not. This organizational 132 mismatch has implications for the culture and climate of districts and schools. When there are varied understandings of policies, then policies may be implemented differently. When there is varied implementation of policies, then there may be the perception of unfairness, even if it is unintentional. Additionally, the organizational mismatch may lead to teachers feeling unnecessarily constrained by an attendance policy that they have a misunderstanding of. The flip side of that is the organizational mismatch may lead to teachers feeling like they have more flexibility to use paid time off than they actually do. Either way, organizational mismatch between attendance policies as written in CBAs and teachers’ understanding of these policies can create waves within an organization that can contribute to the organization’s culture and climate. In addition to teachers within the same district having varied understandings of teacher attendance policies as outlined in their CBAs, the other secondary finding that emerged through document analysis was that there was quite a bit of variance between districts around teacher attendance policies. In other words, different districts allotted different amounts of paid time off, as demonstrated earlier in Table 5. This is important because it begs the question of how to define chronic absenteeism. The existing research largely defines chronic absenteeism as when teachers are absent from the classroom 10 days or more in a given school year (e.g. Miller, 2012; Miller, 2017; Thomas B. Fordham Institute, 2017; Hui, 2018; Kraft, 2018). However, one of the districts in my study only allowed five days of paid time off. When I interviewed the teachers in this district about chronic absenteeism, they assumed chronically absent meant missing more days from work than was allotted in their CBA. Another district in my study allotted 13 days of paid time off and the teachers from that district also assumed chronically absent meant missing more days from work than was allotted in their CBA. I clarified their understanding by providing them with the definition of chronic absenteeism as supported by the existing literature and 133 explained that for the purposes of the present study, chronic teacher absenteeism was defined by teachers missing 10 days or more during the school year. Providing this clarification facilitated common understanding around the term “chronic teacher absenteeism.” This difference in understanding leads to questions around how chronic teacher absenteeism is defined both in practice and in research. For example, should chronic teacher absenteeism be defined as teachers missing 10 days or more; or, should chronic teacher absenteeism be defined as teachers missing more days off than what is allotted in their CBAs? There are pros and cons to each of these two options, but the general takeaway is that the varied teacher attendance policies and confusion around what constitutes chronic teacher absenteeism makes it an increasingly challenging behavior and construct to nail down in both practice and research. The fourth elusive organizational factor that was included in the literature review but has not yet been discussed is commitment. While the management literature supports a relationship between organizational commitment and withdrawal behaviors, such as employee absenteeism (e.g. Mowday, Porter, & Steers, 1982; Clegg, 1983; Eisenberger, et al., 1990; Mathieu & Zajac, 1990; Gellatly, 1995; Sagie, 1998; Meyer, et al., 2002; Klein, Cooper, Molloy, & Swanson, 2014), themes related to organizational commitment (in other words commitment to the school or district) did not manifest in the data collected here. I certainly expected to find themes related to organizational commitment during teachers’ interviews about their absenteeism behaviors, but perhaps the questions I asked did not get at the management construct of organizational commitment. Alternatively, perhaps organizational commitment does not have a relationship with teacher absenteeism behaviors. This is an area that could be explored in future research. Though themes related to organizational commitment did not surface during data analysis, one 134 tangential but noteworthy theme that came through in all 21 interviews was a strong and passionate commitment to the students. In other words, through their interview responses, all interviewees in this study demonstrated a deep commitment to the learning, well-being, and success of the students they taught. Coupled with the findings from this study about the perceived connection between chronic teacher absenteeism and job-related mental health, this divergent theme is worth briefly noting because a reduction in negative job-related mental health may bolster teachers’ presence in the classroom so they can practice their profession and support the students they are committed to. One final consideration for chronic teacher absenteeism was discussed in the literature review – the relationship between union membership and teacher absenteeism. The existing management literature supports a non-causal, positive relationship between employee absenteeism and union membership (e.g. Allen, 1981; Allen, 1984; Leigh, 1986; Chaudhury & Ng, 1992), and thus I was expecting there to be evidence of a relationship between chronic teacher absenteeism and teacher union membership. However, my findings do not speak to this relationship. Teacher unions have a strong presence in Michigan (Winkler, Scull, & Zeehandelaar, 2012), so it is not surprising that almost all of the teachers in my sample belonged to a teacher union. This is an example of sampling bias because the sample did not accurately represent the population under study (Atkinson & Flint, 2001; Cohen, Manion, & Morrison, 2011; Patton, 2015). The uniformity of my sample with regards to union membership is a limitation and means I cannot make statements about whether or not union membership is reflective of the types of teachers who may be likely to be chronically absent. The teachers in my sample were also all white, which similarly means I cannot discern connections between race/ethnicity and chronic absenteeism in this study. There were, however, 135 findings related to other participant and district characteristics that might be reflective of the types of teachers and districts that are more susceptible to stressors, poor organizational support, and job dissatisfaction, which contribute to chronic teacher absenteeism. Namely, teachers who had six-to-twenty years of teaching experience or who were between the ages of 30 and 49, and/or teachers from districts with a higher percentage of students who were minorities and/or students who were economically disadvantaged may be connected to increased job stress, low perceived organizational support, salary dissatisfaction, disinterest in attendance incentives, and ultimately chronic teacher absenteeism. The non-causal connection between teachers being more susceptible to job stress and job dissatisfaction and thus chronic teacher absenteeism in high poverty and high minority schools is similarly evidenced in the teacher turnover literature, which found that teachers are less satisfied and more likely to turnover in high poverty and high minority schools (e.g. Anderman, Belzer, & Smith, 1991; Grissom, 2011). Likewise, teacher experience seemed to be connected with teachers’ approaches to accruing paid days off for attendance payout incentives. Across my findings, both novice and veteran teachers were chronically absent and provided similar reasons that spoke to why they decided to be absent. The exception to this was when veteran teachers explained that attendance payout policies did not motivate their attendance behaviors because they had already maximized their payout accrual bank, which implies teacher experience may be associated with chronic teacher absenteeism. Albeit, responses related to attendance payout incentives came second or even third to responses pointing to job-related mental health as reasons for missing work. Being aware of these teacher and district characteristics may help with early identification of teachers and districts that are more vulnerable to chronic teacher absenteeism, which could inform policy and practice to abate chronic teacher absenteeism. The demographic characteristics of the participants and districts 136 were not a central focus of this study, but they provided some contextual information that may warrant further research in future studies. Both expected and unexpected findings surfaced through data analysis of the interviews about chronic teacher absenteeism. Some of these findings extend existing research from the management literature on employee absenteeism and the education research related to teacher withdrawal behaviors (teacher turnover), other findings add to the inconclusive evidence on the impact of attendance incentive policies on reducing absenteeism, while even other findings are preliminary and can inform future studies on chronic teacher absenteeism. These findings support the use of a management and teacher turnover lens to make sense of teacher absenteeism because the job stress, perceived organizational support, and job satisfaction lenses, specifically, helped to make meaning of chronic teacher absenteeism in this study. Although the findings clearly support a non-causal connection between chronic absenteeism and job stress, perceived organizational support, and/or salary and compensation satisfaction, there are limitations to this study that must be considered. These limitations are discussed next. Limitations The present study on chronic teacher absenteeism is not without limitations, and these must be considered when thinking about the validity and reliability of the findings and how the findings can be used to explain the reasons for chronic teacher absenteeism. Earlier in this paper, I discussed the various steps I took to improve the reliability and validity of my study (see Table 4). Specifically, pilot interviews, cognitive interviews, document analysis (i.e. triangulation), member checking, sampling to the point of redundancy, acknowledging my positionality, recording an audit trail, inductive and deductive analysis, and providing detailed descriptions of 137 my study and findings were all used to enhance reliability and validity. Although these steps were taken, there are six limitations that are worthy of revisiting. First, my positionality as the researcher limits this study because it introduces bias to the data collection, analysis, and interpretation (e.g. Rossman & Rallis, 1998; Maxwell, 2013; Merriam & Tisdell, 2016). I accounted for my positionality earlier and took steps to mitigate it, but my personal background and ideas about chronic teacher absenteeism pose threats to the reliability and validity of my study. For example, I approached this study with the belief that teacher absenteeism is a problem in Michigan schools and culture and climate issues in schools (i.e. organizational factors) may contribute to chronic teacher absenteeism. Even though I bring attention to my positionality, it undoubtably influences the study design, data collection, data analysis, and interpretation. Second, there are a few limitations with regards to my sample. My sample size is fairly small, which limits the generalizability of my study to the broader population of chronically absent teachers. Generalizability is not the primary aim of qualitative research, but the small sample size introduces both selection bias and sampling bias. For example, my entire sample consists of white teachers, which is not representative of the broader population of teachers. The absenteeism experiences of white teachers could be different from the experiences of non-white teachers. Or, my sample might only consist of teachers who wanted to share their experiences with me, which would skew the data. The small sample size threatens both the reliability and validity of this study. Third, the sensitive nature of the topic of chronic absenteeism poses threats to validity and reliability through the potential for social desirability bias. Chronic absenteeism is a sensitive topic because it may feel threatening or accusatory. I took calculated steps to alleviate the 138 potential for social desirability bias, such as carefully ordering and wording questions, carefully framing the interview, ensuring confidentiality, using document analysis for triangulation, and building rapport, but the sensitive nature of the topic remains a limitation to this study. This is because participants may not have felt comfortable being fully transparent about their own absenteeism. My perception is that the interviewees were very candid in their responses about their teacher absenteeism, but this is not something I can be completely certain of. Fourth, even one of the steps I took to bolster reliability and validity, triangulation through document analysis, has limitations. I conducted document analysis of the policies on teacher attendance and teacher absenteeism as outlined in the CBAs of each represented district. The policies I examined offered insights to policies teachers operated under, but the policies did not unveil how the policies were implemented or to what degree of fidelity they were implemented with. While they certainly provided some context and helped to triangulate the interview data, they were limiting in that they only evidenced what was written in the policy and not the actual implementation. In addition, the policies were not written for research purposes, so the information included in them was limited. These limitations threaten the confirmability of the document analysis. Fifth, there may be some instances where I did not probe enough into teachers’ responses to my questions. For example, some chronically absent teachers in my sample explained they chose to miss work even though they did not have any remaining paid days off, which resulted in loss of pay. I did not collect data around teachers’ financial circumstances, such as if a teacher’s partner or spouse had financial stability, which may have influenced teachers’ thought processes when weighing the decision to take a day off against the loss of pay. Another example is the lack of responses that spoke to organizational commitment, which may be absent because my 139 questions did not hone in on this management construct. This implies there may be additional areas to explore regarding the reasons for chronic teacher absenteeism. Finally, some of my interviews occurred virtually over Zoom or the phone, which may have impacted my ability to establish rapport with the participants. This is a limitation because it increased the chances for response bias. My goal was to conduct all interviews in person, but schedules and travel complicated this. My perception is that the frankness of responses from interviews that took place over Zoom or the phone did not differ from the frankness of in person interviews, but this may not be the case in actuality. While I took many steps to address these limitations and ultimately enhance the reliability and validity of the present study on chronic teacher absenteeism, some limitations remain and must be considered when thinking about the validity and reliability of the findings and how the findings can be used to explain chronic teacher absenteeism. Future research on chronic teacher absenteeism with larger sample sizes and diverse methods may be able to address some of these limitations. At minimum, the findings from this study begin to shed light on the increasingly problematic phenomenon of chronic teacher absenteeism. Overall Contribution The present study on chronic teacher absenteeism makes several contributions to the education field, and these contributions can be divided into three sections: research, policy, and practice. Research. This study adds to the existing research in both the management field and education field. In the management field, this study adds to the mounting evidence about a relationship between organizational factors and employee absenteeism. The findings presented here suggest a non-causal positive connection between employee absenteeism and job stress, low 140 perceived organizational support, and job dissatisfaction, which has already been demonstrated in the management literature (e.g. Beehr & Newman, 1978; Gupta & Beehr, 1979; Mowday, Porter, & Steers, 1982; Clegg, 1983; Eisenberger, et al., 1990; Mathieu & Zajac, 1990; Gellatly, 1995; Sagie, 1998; Meyer, et al., 2002; Klein, Cooper, Molloy, & Swanson, 2014). This study also extends the findings of teacher turnover research by similarly finding a connection between organizational factors, job satisfaction, salary and compensation, and teacher withdrawal behaviors (absenteeism in the case of this study; turnover in the case of previous research) (e.g. Ingersoll, 2002; Ingersoll, 2003; Ingersoll & Smith, 2003). Perhaps of greater consequence are the contributions this study makes to the education field related to teacher absenteeism. The reasons for chronic teacher absenteeism have not been widely studied in the existing education research and this study offers a first empirical glimpse into the reasons for chronic teacher absenteeism as self-reported by teachers through qualitative interviews. The urgency surrounding chronic teacher absenteeism as evidenced by the high costs associated with teacher absenteeism (e.g. Ferris, Bergin, & Wayne, 1988; Miller, 2012), recent media attention (e.g. McVicar, 2017; The Detroit News, 2017; Wolcott, 2017; Renk, 2019), and the non-causal connection to student achievement (e.g. Clotfelter, Ladd, & Vigdor, 2007; Miller, Murnane, & Willett, 2008; Finlayson, 2009; Goldhaber & Hansen, 2010; Roby, 2013) demands for research to explore the reasons for chronic teacher absenteeism, and this study does just that. Not only does this study explore the reasons for chronic teacher absenteeism, but it does so in a way that examines and conceptualizes hard-to-measure elusive and intangible teacher characteristics and behaviors that contribute to teacher absenteeism, such as job stress, perceived organizational support, and job satisfaction. 141 Furthermore, the findings from this study extend the teacher turnover research by exploring a similar withdrawal behavior to turnover, absenteeism. Similar to the education research on a withdrawal behavior, teacher turnover, this study finds associations between the withdrawal behavior of teacher absenteeism and organizational climate and job dissatisfaction. Both the teacher turnover research and the research conducted through this study share the goal of trying to understand why teachers withdraw from the classroom (via turnover and absenteeism) and thus how to keep teachers in classrooms, and the present study on chronic teacher absenteeism offers some empirical insights around this shared goal. Also, the findings from this study contribute to the inconclusive empirical evidence surrounding the impact of attendance incentives on reducing teacher and/or employee absenteeism, as previously found in the education and management literature (e.g. Ehrenberg, Ehrenberg, Rees, & Ehrenberg, 1989; Clotfelter, Ladd, & Vigdor, 2007; DeNardo, 2007; Ahn & Vigdor, 2011; Hanover Research, 2012; National Council on Teacher Quality, 2014). Like the previous research on this topic, it is unclear what kind of impact attendance incentives have on teacher absenteeism as demonstrated through the findings of the present study. Taken together, the present study contributes to the current management and education research by extending and adding to findings that speak to the relationship between organizational factors, elusive characteristics, and employee/teacher absenteeism. Policy. This study also has implications for education policy at the federal level, state level, and local level. At the federal and state level, the findings of this study make a case for better collection of teacher attendance data. Currently, teacher absenteeism is hard to measure and collect data on because teacher attendance policies are created locally and districts are not required to submit teacher attendance data to a centralized authority (National Council on 142 Teacher Quality, 2014). This means that each teacher attendance policy is different and teacher attendance is not regulated, making the collection and measurement of teacher attendance very difficult. The findings presented here demonstrate evidence of chronic teacher absenteeism. Since teacher absenteeism is expensive and there is a non-causal link between teacher absenteeism and student achievement, it would be useful to know how prevalent chronic teacher absenteeism is and in what schools and districts it is most prevalent. Having teacher attendance data will facilitate a deeper understanding of the problem, allow for further research, and potentially lead to solutions to address the problem. As such, perhaps teacher absenteeism data should be included in federal and/or state level policy and accountability so there can be valid and reliable data collection on teacher attendance. Along similar lines, the findings presented here suggest that teachers do not feel valued and are working in climates where they do not have the resources or support they need. This may point to a symptom of a greater problem, being that schools and districts are under-resourced in general due to a lack of funding from the state government. In order for teachers to have an impact on student achievement, they must first be present in the classroom. Therefore, it seems like making sure our schools and districts are adequately resourced should be a state education policy priority in order to support teachers and their working environments. Investing in teachers and their working environments could improve the culture and climates within which they work, and ultimately impact teacher attendance and absenteeism. The policy conversation around adequately resourcing our schools and districts should include the topic of teacher salaries because one way to make teachers feel valued for their work is to compensate them appropriately. 143 Further, the findings from this study demonstrate the wide variance in teacher attendance policies as written in district CBAs. Not only does such variance create challenges in measuring and understanding teacher absenteeism, but it also could have an impact on teacher attraction, retention, and turnover. Perhaps federal or state education policy should include minimum requirements around teacher attendance. Interestingly, federal and state education policy include requirements and reporting around student attendance. Challenges with teacher absenteeism may benefit from similar state and federal regulations around teacher attendance. This study also offers insights to the underlying reasons for chronic teacher absenteeism, including organizational factors such as job stress, low perceived organizational support, and job dissatisfaction. If additional research substantiates these findings, perhaps state and federal policy can be created to target the culture and climate issues by encouraging or discouraging organizational characteristics and teacher behaviors in schools and districts that contribute to chronic teacher absenteeism. Alternatively, current state and federal education policy, specifically accountability-related polices, could be revisited and amended to alleviate some of the pressures teachers face that contribute to their feelings of increased job stress and thus influence their attendance behaviors. At the local level, policy can similarly be created and implemented to target wanted or unwanted organizational characteristics and teacher behaviors related to teacher attendance. Another contribution made by this study is the finding that attendance policies as written in district CBAs were implemented and enforced differently within districts. This finding might suggest a policy audit or policy implementation audit is needed at the local level to determine how the attendance policy is understood and implemented by the administrators and teachers. A 144 policy audit could result in new policy creation, current policy modification or clarification, and tighter policy implementation. In sum, since teachers are the single most important input that impacts student achievement, then promoting behaviors and experiences to improve teacher attendance should be an education policy priority. Some of the policy contributions discussed above are much larger than this dissertation is equipped to tackle, such as the conversation around teacher salaries, but this study can bolster exiting evidence and research to make a case for investing in our districts, schools, and teachers. Practice. Finally, the present study on chronic teacher absenteeism has implications for practice. The non-causal connection between teacher absenteeism and student achievement coupled with the evidence presented here of chronic teacher absenteeism in Michigan, demands for changes to practice. This demand is resounded by the recent and expensive substitute teacher shortage in Michigan (Burroughs, et al., 2019). In short, organizational changes are needed so teachers are present in the classroom. Job stress, perceived organizational support, and job satisfaction are three areas that need improvement. Administrators need to be aware of the sources of job stress for teachers so practices can be put in place to reduce job stress. Teachers indicated their sources of job stress came from negative student behaviors, large class sizes, and accountability pressures. While districts and administrators are not solely responsible for all of these, they could consider changes to their handling of negative student behaviors. Or, they could agree to class size limits and honor those agreements. Teachers also indicated feelings of low perceived organizational support at both the school and district levels contributed to their needing to use paid time off for job-related mental health. To improve perceived organizational support at the school level, perhaps districts could 145 focus efforts on leadership training, so principals are adequately trained in how to best support teachers. Lack of planning time also contributed to teachers’ feelings of low perceived organizational support. Building in adequate planning time for teachers may be a small change with a big impact for chronically absent teachers who feel low perceived organizational support and thus use paid time off for job-related mental health reasons. Another implication for practice is for districts to make sure teachers have adequate resources to perform their jobs. Improved and increased resources might include access to supplies (e.g. paper, books), support for initiatives, and having adequate support staff, such as instructional aides, interventionists, and school counselors. Making such changes to practice are not easy because many resources are needed, and they are all expensive. However, the alternative, chronic teacher absenteeism, also consumes many resources. As demonstrated in the literature review, not only is chronic teacher absenteeism expensive, but it also has non-causal impacts on student achievement. The suggested changes to practice are costly, but the benefit to investing in these changes in practice may outweigh the costs. The research, policy, and practice contributions and implications discussed above as a result of the present study on chronic teacher absenteeism make this study relevant and impactful to the education field. Next Steps and Opportunities for Future Research My hope is that, at minimum, this study makes a case for better collection of teacher attendance data and on a much larger scale. Once a better data set exists for teacher attendance, additional patterns and themes related to teacher absenteeism can be identified and studied. This may include using teacher, school, and district-level attendance data to confirm or disconfirm if 146 chronic absenteeism exists, where it is most prevalent, and for what kinds of teachers it is most likely to occur (e.g. in terms of demographic characteristics, such as teacher experience, age, union membership, etc.). I also hope this study pushes researchers, educators, and policy makers to agree on a definition of chronic teacher absenteeism. For this study, I operated with chronic teacher absenteeism defined by teachers missing 10 days or more during the school year, which is common practice in the existing research (e.g. Miller, 2012; Miller, 2017; Thomas B. Fordham Institute, 2017; Hui, 2018; Kraft, 2018). However, a case could be made for the definition of chronic teacher absenteeism to include missing more days off than what is allotted in district CBAs. A common definition is needed to move forward with research on chronic teacher absenteeism to facilitate a mutual understanding of the phenomenon under study. In the meantime, the next steps related to the present study are two-fold. First, I am interested in circulating my findings in the education policy sphere with local and state lawmakers. I hope to generate interest and buy-in from local and state lawmakers for additional support to pursue research on teacher absenteeism and to move it up the policy agenda. Having interest and support from local and state lawmakers could eventually change how teacher attendance is factored into high-stakes accountability, how teacher attendance data are collected, and refocus school improvement efforts to invest in school and district culture and climate to improve teaching conditions. One way to gain some credibility and traction for this work is through the peer-review and publishing process. For that reason, I hope to submit portions of the present study to a well-respected academic journal for peer review and potential publication. Second, I hope to continue this research by partnering with a district(s) and/or Intermediate School District(s) (ISD) to conduct both a quantitative study through analyzing local teacher attendance data and a case study. Teacher attendance data are collected locally, so 147 there may be local opportunities to dive into teacher attendance data, teacher attendance policy implementation, and teacher attendance behaviors. This opportunity might be particularly appealing to a district or ISD that perceives a chronic problem with teacher absenteeism and is interested in learning more about the problem and exploring solutions to promote teacher attendance. In addition to the next steps I might take to extend this study, there are a number of future research opportunities related to chronic teacher absenteeism that could be pursued. Research from the management literature suggests a non-causal negative connection between organizational commitment and employee absenteeism (e.g. Mowday, Porter, & Steers, 1982; Clegg, 1983; Eisenberger, et al., 1990; Mathieu & Zajac, 1990; Gellatly, 1995; Sagie, 1998; Meyer, et al., 2002; Klein, Cooper, Molloy, & Swanson, 2014), but organizational commitment did not surface in the study at hand. It is possible that the interview questions I used did not provoke responses related to the construct of organizational commitment. As such, future research might take a keen focus into organizational commitment and its relationship with chronic teacher absenteeism. Similarly, research from the management literature also suggests a non-causal negative connection between job satisfaction and employee absenteeism (e.g. Muchinsky, 1977; Steers & Rhodes, 1978; Clegg, 1983; Scott & Taylor, 1985; Garcia, 1987; Hackett, 1989; Chaudhury & Ng, 1992; Sagie, 1998). Teacher turnover research finds a similar non-causal inverse relationship between job satisfaction and teacher turnover (e.g. Ingersoll & Smith, 2003; Ingersoll, 2011). However, only one finding that is peripherally related to job satisfaction emerged in this study (i.e. dissatisfaction with regards to salary and compensation was regularly mentioned in interviews with chronically absent teachers). Again, it is possible that the interview questions I 148 used did not elicit responses related to the construct of job satisfaction. Future research could take a specific look at the connection between job satisfaction and chronic teacher absenteeism. Future research could also take a different methodological approach to examining some of the same management constructs studied here. The four management constructs covered in this study related to employee absenteeism – job stress, perceived organizational support, organizational commitment, and job satisfaction – each have measurement scales associated with them that are well-developed and grounded in theory and research. These scales could be used in a study that examines chronic teacher absenteeism using survey methods. One of the many benefits of using survey methods is that a larger sample can potentially be reached (e.g. Maxwell, 2013; Merriam & Tisdell, 2016), which would assuage reliability and validity concerns. Existing management research also evidences that union membership is inversely related to job satisfaction (e.g. Freeman, 1977; Borjas, 1979; Smith & Hopkins, 1979; Gordon & Long, 1981; Allen, 1984). This means that union membership might have a relationship with both job satisfaction and chronic teacher absenteeism, which should be explored in a future study. Exploring the connection between union membership and chronic teacher absenteeism in a future study is especially intriguing because of the strong presence of teacher unions in Michigan. Other descriptive characteristics, such as the demographic characteristics of participants (e.g. race/ethnicity, age, teaching experience, etc.) and district characteristics (e.g. student body demographics, district type, district size, etc.), could also be considered for future research on teacher absenteeism. Additionally, future research could explore the development and implementation of teacher attendance policies. This type of research might compare and contrast teacher attendance 149 policies in policy (i.e. as written) and in practice (i.e. as implemented) to identify congruence and examine the impact of congruence, or lack thereof, on teacher attendance. Similarly, future research could do a deeper examination on the impact of attendance incentives on teacher absenteeism. Both positive and negative incentives with varying magnitudes and their influence on teacher attendance behaviors could make for an intriguing study. Finally, an interesting future research stream might be to examine how accountability policy impacts teacher attendance. For example, if teacher attendance is included in federal and/or state accountability, how do teacher attendance behaviors change? To put it briefly, teacher absenteeism is a fairly untapped area of research, especially as it relates to the reasons for teacher absenteeism, and thus there are many diverse opportunities for future research. 150 CONCLUSION This study takes initial steps in empirically studying the underlying reasons for chronic teacher absenteeism. The literature on organizational factors that impact teacher turnover in the education field and employee absenteeism in the management field, such as job stress, perceived organizational support, and job satisfaction, offer appropriate lenses to make sense of the teacher absenteeism interview data collected here. These lenses provided insights as to why teachers were absent by focusing in on some teacher characteristics that are elusive, intangible, and thus hard to measure, such as job stress, perceived organizational support, and job satisfaction. Succinctly, in situations of high job stress, low perceived organizational support, and/or low job satisfaction, self-reported absenteeism tends to be more job related than due to personal circumstance. Furthermore, teachers who are not chronically absent are very intentional about being conservative in their use of paid time off and trying to only use paid time off for personal circumstances, such as illness, whereas this is not the case with chronically absent teachers. These findings suggest organizational factors, or challenges with school and district culture and climate, are prominent sources of chronic teacher absenteeism. As demonstrated at the beginning of this study, teachers may be the single most important input to improving student learning outcomes. Thus, understanding what it is about teachers that makes them so important and how to promote these desirable characteristics, such as their easy-to-measure (e.g. teacher experience) and elusive characteristics (e.g. stress, support, and satisfaction), is prudent. This study sheds light on some of these elusive, intangible teacher characteristics and suggests certain organizational factors are key contributors to heightening them, in both positive and negative directions. With the current dire state of education in the U.S., a fresh approach to examining the relationship between teachers and student achievement is needed. This study does just that by 151 providing some clarity around the drivers behind a teacher’s decision to be absent from the classroom. Understanding teacher attendance behaviors certainly is not the only answer to what characteristics make teachers so important but having a better understanding of teacher attendance behaviors may help move the conversation forward. Knowing how imperative organizational factors are on teachers’ experience of job stress, organizational support, and job satisfaction signals that emphasis is needed for culture and climate reform at the organizational level (school and district) because these factors impact their job-related mental health and contribute to teachers’ need to take paid time off for job-related mental health. According to the findings from the present study, school and district organizational culture and climate need to change so teachers experience less job stress, feel supported and valued by their administrators and organizations, and are satisfied at work. Better supporting teachers in their work environments may contribute to improvements in teacher attendance. Until something is done to address the drivers behind chronic teacher absenteeism, teachers will continue to experience negative job-related mental health and will continue to exhibit withdrawal behaviors. An investment in future research, policy, and practice is needed around teacher absenteeism and teacher withdrawal behaviors, in general. There is more to learn through research about elusive, intangible characteristics that are related to teacher withdrawal behaviors, and this learning needs to happen on a larger scale. We need our teachers to be present, engaged, supported, and satisfied in order to have a positive impact on Michigan students and, more broadly, U.S. students. Policy and practice must be designed to support positive teacher behaviors and experiences, which may improve teacher effectiveness. Through intentional education policy designed to improve the conditions within which teachers work, 152 teachers’ work experiences and perceptions may improve. Such important changes are necessary to elevate the teaching profession and make it a highly desirable, respected, and prestigious career for current teachers and for high school and college-aged students who are trying to decide what their future profession will be. Having the understanding that organizational factors profoundly shape a teachers’ professional experiences and perceptions and subsequent behaviors, such as absenteeism, is a starting place for devising action and change. Perhaps, as a result of this study and future similar studies, local, state, and federal policies will begin to target wanted or unwanted practices, environments, conditions, etc. related to supporting teacher experiences and thus attendance. 153 APPENDICES 154 APPENDIX A Teacher Interview Protocol Introduction to Interviewee: Hello ___________. My name is Jacqueline A. Gardner and I want to thank you for taking the time today to speak with me. I have been studying teacher attendance behaviors and am in the process of doing interviews with Michigan teachers, which will serve as the foundation for my dissertation. The purpose of this interview is to understand your attendance behaviors at work. Specifically, I am interested in learning about underlying reasons for teacher absences and how teachers decide to be absent. I am also interested in your perceptions of the attendance behaviors of other teachers in your school and district. My goal is to better understand the contributing factors to teacher absences. The interview should last no longer than 45 minutes to one hour. I will not use your name, school name, district name, or any personal identifiers. If I believe I need to use a quote or phrase that will identify you, I will work with you on rephrasing the statement. I will also ask you to pick a pseudonym for reference. Your privacy and confidentiality will be protected to the maximum extent allowable by law. I will be making a few notes while you are talking and because I am unable to capture every word you say during the interview, I am requesting that I use a recorder. If you have any questions during the interview do not hesitate to ask me. At the close of the interview, I will present you with a $10 gift card to Starbucks in appreciation of the time you have spent with me today. Do you have any clarifying questions before we begin? … Ok. I am going to turn on the recorder now. [Consent will be obtained via signature and also on the recording device. I will read the consent form aloud to the participant and will ask for verbal and written consent.] 155 Introductory Questions 1. Tell me about your teaching career, such as why you decided to go into teaching, where you’ve taught, what you’ve taught, etc.? a. What is your educational background? i. Level of education? ii. Age? b. How many years have you been in the teaching profession? c. What grade level do you currently teach? i. Have you always taught that grade level? d. What subject do you currently teach? (where applicable) i. Have you always taught that subject? e. Why did you go into teaching in the first place? i. What motivated you to enter the teaching profession? 1. Does that still motivate you? f. Have you always taught in your current building? i. How many different physical building moves have you made within your current district? ii. How many different physical building moves have you made over the course of your career? g. How stable has your current district been? i. How do you perceive your district? 1. What are its strengths and weaknesses? ii. Why did you decide to work in this district? 1. Where are you originally from? [SEGWAY] What reasons do teachers provide for their chronic absenteeism? Research Question 1. There are a lot of different policies around sick and vacation time. I’m wondering if you could tell me a little bit about the policies in your district as outlined in your district’s 156 collective bargaining agreement? (This might be referred to as a teacher attendance policy, sick/emergency/vacation time, paid time off.) 2. How does your district enforce these policies? a. In other schools, I’ve noticed that sometimes schools enforce formal “teacher attendance policies” differently. To what degree does your school enforce the district wide teacher attendance policy as outlined in your district’s collective bargaining agreement? b. Are there any instances that you know of where the district’s formal teacher attendance policy is enforced differently? (i.e. in own school and other schools in district) i. What does it look like? c. What do you think about what the district/school does in terms of enforcing the teacher attendance policy as outlined in your district’s collective bargaining agreement? i. Are there any “loopholes” to the teacher attendance policy in the collective bargaining agreement that allow teachers to take more days off than their allotted days? 3. I’ve also noticed that sometimes there are “informal policies and agreements” around teacher time off. For example [give example]. What, if any, informal agreements exist around teacher time off in your school? a. What about in other schools? 4. Teachers use a lot of strategies when using their paid time off. Some save up all the time, some use it throughout the year as needed, some use it around vacation time. What is your approach to using your allotted paid time off? a. In general, do you exhaust your paid time off each year? Why or why not? i. How often do you use more than your allotted paid time off each year (in days)? b. Do you take your paid time off as needed? Or, do you plan it out at the beginning of each school year? Why do you take this approach? c. Are there other strategies that you’ve thought of or heard of? d. Why did you decide on the strategy you mentioned earlier? 157 5. What do you do when you want to take a day off for illness or personal reasons? a. Do you have to specify what you are using your days off for? 6. Teachers use their days off for different purposes, such as running errands, childcare, caring for elderly parents, health, job searching/interviews, or they just need a day. When you use paid time off, what reason(s) do you use it for? What reasons make up for most of your days off? a. What reasons do you report? b. How do these reported reasons differ from actual reasons for using paid time off? i. What are some reasons for using paid time off that you do not officially report? ii. How common is it for you to use a day off as a personal day? 7. In what ways do you think other teachers use their days off? 8. How do you handle a situation where your days off are exhausted, but you need a day off? a. How often do you find yourself in such a situation? Why? iii. Is it because you aren’t allotted enough days off? Job stress? Personal stress? b. How do other teachers handle this situation? 9. A lot of people find themselves in a situation where they’ve used all their days off, but they need another day. Has that ever happened to you? If yes, about how often does that happen? How do you deal with it? a. What reasons lead you to be absent from work without using a day off? iv. What about your colleagues who are absent without using paid time off (frequency and reasons)? 10. How often do you miss work due to job-related stress (i.e. not illness or vacation related)? a. In your perception, how often do your colleagues miss work due to job-related stress? 11. How do you decide whether to be absent from work? a. What work-related factors contribute to your decisions whether to be absent from work? 158 12. Are there any official school and/or district level incentives that encourage teachers not to use their days off? If yes, what are they? a. Are there any unofficial incentives? 13. Do you think these incentives impact your attendance? How? 14. Is there anything your school or district could do to change how you use your days off? (i.e. If your school was less chaotic, if you felt more supported, if you had a better leader, etc.) a. If yes, what would it look like? b. How would a greater incentive impact your attendance behavior, if at all? 15. What would have the greatest impact on your attendance behaviors? a. Incentive? b. Culture/climate? Concluding Questions 1. What else would you like to tell me? 2. Do you have any questions for me? 159 APPENDIX B Teacher Attendance Research Participant Information and Consent Form Explanation of the Research and What You Will Do ● The purpose of these interviews is to better understand underlying reasons for teacher absences and how teachers decide to be absent. These interviews will also focus on your perceptions of the attendance behaviors of other teachers in your school and district. ● Your participation will be limited to one 60-minute interview. If you agree to participate, your responses will be recorded and later transcribed. You will also be assigned a pseudonym of your choice and all personal details will be masked. ● It is possible that following the completion of your participation, one member of our research team will contact you to clarify or ask for elaboration on a certain point. Your Rights to Participate, Say No, or Withdraw ● Your participation in this study is completely free and voluntary: you may refuse to respond to any question; and you may end your participation at any time without consequence. Costs and Compensation for Being in the Study ● Your time and energy are the only costs associated with your participation. ● Risks associated with the study topic and methods are minimal. ● After your interview, you will receive a $10 Starbucks gift card. Contact Information for Questions and Concerns ● If you have any concerns or questions about this study, please contact the study’s investigator. ● Investigator: Jacqueline A. Gardner at swans126@msu.edu ● If you have questions or concerns about your role and rights as a research participant, would like to obtain information or offer input, or would like to register a complaint about this study, you may contact, anonymously if you wish, the Michigan State University’s Human Research Protection Program at 517-355-2180, Fax 517-432-4503, or e-mail irb@msu.edu or regular mail at 4000 Collins Rd, Suite 136, Lansing, MI 48910. Documentation of Informed Consent ● Please indicate your voluntary agreement to participate by signing below. ___________________________________ Name ________________________ Date ● Please indicate receipt of the gift card by signing below. ___________________________________ Name ________________________ Date 160 APPENDIX C Recruitment Material Introduction to Interviewee: Hello ___________. My name is Jacqueline A. Gardner and I want to thank you for taking the time today to speak with me. I have been studying teacher attendance behaviors and am in the process of doing interviews with Michigan teachers, which will serve as the foundation for my dissertation. The purpose of this interview is to understand teacher attendance behaviors. Specifically, I am interested in learning about underlying reasons for teacher absences and how teachers decide to be absent. I am also interested in your perceptions of the attendance behaviors of other teachers in your school and district. My goal is to better understand the contributing factors to teacher absences. The interview should last no longer than 45 minutes to one hour. I will not use your name, school name, district name, or any personal identifiers. If I believe I need to use a quote or phrase that will identify you, I will work with you on rephrasing the statement. I will also ask you to pick a pseudonym for reference. Your privacy and confidentiality will be protected to the maximum extent allowable by law. I will be making a few notes while you are talking and because I am unable to capture every word you say during the interview, I am requesting that I use a recorder. If you have any questions during the interview do not hesitate to ask me. At the close of the interview, I will present you with a $10 gift card to Starbucks in appreciation of the time you have spent with me today. Do you have any clarifying questions before we begin? … Ok. I am going to turn on the recorder now. [Consent will be obtained via signature and also on the recording device. I will read the consent form aloud to the participant and will ask for verbal and written consent] 161 Dear : Recruitment E-mail I am a doctoral candidate at Michigan State University in the Education Policy program working on data collection for my dissertation. My dissertation focuses on teacher attendance behaviors and I am in the process of doing interviews with Michigan based teachers. I am hoping that you would be willing to meet with me for 60 minutes to share your experiences and perspectives around teacher attendance behaviors. Is there a day or time that you would be available to meet? If you have any concerns or questions about this study, please contact Jacqueline A. Gardner by e-mail (swans126@msu.edu) or by phone at 517-230-3049. Thank you! Jacqueline A. Gardner, Doctoral Candidate MSU Institutional Research reminder: Participation in this evaluation is completely voluntary. You have the right to say no. You may change your mind at any time and withdraw. You may choose not to answer specific questions or stop participating at any time. The interview should take about 60 minutes of your time to complete. The collected interview data will be stored on a password protected computer. Only the researcher will have access to the data. Any identifiable information connected to you will be kept confidential to the maximum extent allowable by law. If you have any concerns or questions about this study, please contact Jacqueline A. Gardner by e-mail (swans126@msu.edu) or by phone at 517-230-3049. If you have questions or concerns about your role and rights as a research participant, would like to obtain information or offer input, or would like to register a complaint about this study, you may contact, anonymously if you wish, the Michigan State University’s Human Research Protection Program at 517-355- 2180, Fax 517-432-4503, or e-mail irb@msu.edu or regular mail at 4000 Collins Rd, Suite 136, Lansing, MI 48910. 162 REFERENCES 163 REFERENCES Ackley, M. (2016, February 9). State Supt. Brian Whiston unveils strategies to help Michigan become a top 10 education state. Michigan Department of Education. Retrieved from https://www.michigan.gov/mde/0,4615,7-140--376451--,00.html. Ahn, T. & Vigdor, J. L. (2011). Making teacher incentives work: Lessons from North Carolina’s teacher bonus program. American Enterprise Institute, 5(June). Allen, S. G. (1981). An empirical model of work attendance. The Review of Economics and Statistics, 63(1), 77-87. Allen, S. G. (1984). Trade unions, absenteeism, and exit-voice. ILR Review, 37(3), 331-345. Anderman, E. M., Belzer, S., Smith, J. (1991). Teacher commitment and job satisfaction: The role of school culture and principal leadership. Paper presented at the Meeting of American Educational Research Association. Chicago, IL. Appleton, J. V. (1995). Analysing qualitative interview data: Addressing issues of validity and reliability. Journal of Advanced Nursing, 22(5), 993-997. Atkinson, R. & Flint, J. (2001). Accessing hidden and hard-to-reach populations: Snowball research strategies. Social Research Update, 33, 1-8. Avey, J. B., Patera, J. L., & West, B. J. (2006). The implications of positive psychological capital on employee absenteeism. Journal of Leadership and Organizational Studies, 13(2), 42-60. Ballard, K. & Bates, A. (2008). Making a connection between student achievement, teacher accountability, and quality classroom instruction. The Qualitative Report, 13(4), 560- 580. Beehr, T. A. & Newman, J. E. (1978). Job stress, employee health, and organizational effectiveness: A facet analysis, model, and literature review. Personnel Psychology, 31, 665-699. Borjas, G. J. (1979). Job satisfaction, wages, and unions. The Journal of Human Resources, 14(1), 21-40. 164 Bowen, G. A. (2009). Document analysis as a qualitative research method. Qualitative Research Journal, 9(2), 27-40. Brinkmann, S. (2014). Interview. In Teo T. (eds) Encyclopedia of Critical Psychology (pp. 1008- 1010). New York: Springer. Burroughs, N., Gardner, J., & Zuschlag, D. (2019). Is Michigan facing a substitute teacher shortage? Michigan Applied Public Policy Research. East Lansing, MI: Institute for Public Policy and Social Research. Retrieved from http://ippsr.msu.edu/public- policy/michigan-wonk-blog/michigan-facing-substitute-teacher-shortage Carlson, D. (2018). Testing and accountability: What have we learned and where do we go? In F. M. Hess & M. Q. McShane (Eds.), Bush-Obama school reform: Lessons learned (13- 31). Cambridge, MA: Harvard Education Press. Castillo-Montoya, M. (2016). Preparing for interview research: The interview protocol refinement framework. The Qualitative Report, 21(5), 811-831. Chaudhury, M. & Ng, I. (1992). Absenteeism predictors: Least squares, rank regression, and model selection results. The Canadian Journal of Economics, 25, 615-635. Chetty, R., Friedman, J. N., & Rockoff, J. E. (2012). Great teaching: Measuring its effects on students’ future earnings. Education Next, 58-68. Citizens Research Council of Michigan. (2019). Stagnant teacher salaries in Michigan: Recent impact of unfunded retirement obligations. Livonia, MI. Clegg, C. W. (1983). Psychology of employee lateness, absence, and turnover: A methodological critique and an empirical study. Journal of Applied Psychology, 68(1), 88-101. Clotfelter, C. T., Ladd, H. F., & Vigdor, J. L. (2007). Are teacher absences worth worrying about in the U. S.? NBER Working Paper Series, November(13648), 1-52. Cochran-Smith, M. (2003). The unforgiving complexity of teaching: Avoiding simplicity in the age of accountability. Journal of Teacher Education, 54(1), 3-5. 165 Cohen, L, Manion, L., & Morrison, K. (2011). Research methods in education (7th ed.). New York, NY: Routledge. Corcoran, S., Evans, W., & Schwab, R. (2004). Changing labor-market opportunities for women and the quality of teachers, 1957-2000. The American Economic Review, 94(2), 230-235. Creswell, J. W. (2013). Qualitative inquiry & research design (3rd ed.). Thousand Oaks, CA: Sage. Creswell, J. W. & Miller, D. L. (2000). Determining validity in qualitative inquiry. Theory Into Practice, 39(3), 124-130. Crouch, M. & McKenzie, H. (2006). The logic of small samples in interview-based qualitative research. Social Science Information, 45(4), 483-499. Darling-Hammond, L., Holtzman, D. J., Gatlin, S. J., & Vasquez Heilig, J. (2005). Does teacher preparation matter? Evidence about teacher certification, Teach for America, and teacher effectiveness. Education Policy Analysis Archives, 13(42), 1-51. DeNardo, C. (2007, September 3). ‘A’ is for absent teachers. Palm Beach Post. Retrieved from https://www.nctq.org/nctq/research/1190911618353.pdf Desimone, L. M., & Le Floch, K. C. (2004). Are we asking the right questions? Using cognitive interviews to improve surveys in education research. Educational Evaluation and Policy Analysis, 26(1), 1-22. DiCicco‐Bloom, B. & Crabtree, B. F. (2006). The qualitative research interview. Medical Education, 40(4), 314-321. Dinham, S. & Scott, C. (1997). A three domain model of teacher and school executive career satisfaction. Journal of Educational Administration, 36(4), 362-378. Dinham, S. & Scott, C. (2000). Moving into the third, outer domain of teacher satisfaction. Journal of Educational Administration, 38(4), 379-396. Dou, D., Devos, G., & Valcke, M. (2017). The relationships between school autonomy gap, principal leadership, teachers’ job satisfaction and organizational commitment. Educational Management Administration & Leadership, 45(6), 959-977. 166 Dunn, A. H. (2018). Leaving a profession after it’s left you: Teachers’ public resignation letters as resistance amidst neoliberalism. Teachers College Record, 120(090301), 1-34. Ehrenberg, R. G., Ehrenberg, R. A., Rees, D. I., & Ehrenberg, E. L. (1989). School district leave policies, teacher absenteeism, and student achievement. NBER Working Paper Series, March(2874), 1-50. Eisenberger, R., Huntington, R., Hutchison, S., & Sowa, D. (1986). Perceived organizational support. Journal of Applied Psychology, 71(3), 500-507. Eisenberger, R., Fasolo, P., & Davis-LaMastro, V. (1990). Perceived organizational support and employee diligence, commitment, and innovation. Journal of Applied Psychology, 75(1), 51-59. Elementary and Secondary Education Act of 1965, Pub. L. No. 89-10 § 64 Stat. 1100 (1965). Every Student Succeeds Act of 2015, Pub. L. No. 114-95 § 114 Stat. 1177 (2016-2016). Ferris, G. R., Bergin, T. G., & Wayne, S. J. (1988). Personal characteristics, job performance, and absenteeism of public school teachers. Journal of Applied Social Psychology, 18(7), 552-563. Finlayson, M. (2009). The impact of teacher absenteeism on student performance: The case of the Cobb County School District. Dissertations, Theses and Capstone Projects, Paper 4. Frayne, C. A. & Latham, G. P. (1987). Application of social learning theory to employee self- management of attendance. Journal of Applied Psychology, 72(3), 387-392. Freeman, R. B. (1977). Job satisfaction as an economic variable. The National Bureau of Economic Research, 1-13. Garcia, R. L. (1987). Sick-time usage by management and professional employees in the public sector. Review of Public Personnel Administration, 7(3), 45-59. Gellatly, I. R. (1995). Individual and group determinants of employee absenteeism: Test of a causal model. Journal of Organizational Behavior, 16(5), 469-485. George, J. M. (1989). Mood and absence. Journal of Applied Psychology, 74(2), 317-324. 167 Goldhaber, D. (2002). The mystery of good teaching: Surveying the evidence on student achievement and teachers’ characteristics. Education Next, 2(1), 50-55. Goldhaber, D. D., & Brewer, D. J. (2000). Does teacher certification matter? High school teacher certification status and student achievement. Educational Evaluation and Policy Analysis, 22(2), 129-145. Goldhaber, D. & Hansen, M. (2010). Is it just a bad class? Assessing the stability of measured teacher performance (CEDR Working Paper 2010-3). Seattle, WA: The Center for Education Data & Research. Goldhaber, D., Grout, C., & Huntington-Klein, N. (2014). It’s selective, but is it effective? Exploring the predictive validity of teacher selection tools. Seattle, WA: The Center for Education Data & Research. Gordon, M. E. & Long, L. N. (1981). Demographic and attitudinal correlates of union joining. Industrial Relations, 20, 306-311. Griffee, D. T. (2005). Research tips: Interview data collection. Journal of Developmental Education, 28(3), 36-37. Grissom, J. A. (2011). Can good principals keep teachers in disadvantaged schools? Linking principal effectiveness to teacher satisfaction and turnover in hard-to-staff environments. Teachers College Record, 113(11), 2552-2585. Gupta, N. & Beehr, T. A. (1979). Job stress and employee behaviors. Organizational Behavior and Human Performance, 23, 373-387. Hackett, R. D. (1989). Work attitudes and employee absenteeism: A synthesis of the literature. Journal of Occupational Psychology, 62, 235-248. Hanover Research. (2012). Addressing teacher absenteeism and attendance. Washington, D.C. Hanushek, E. A., Rivkin, S. G., & Schiman, J. C. (2016). Dynamic effects of teacher turnover on the quality of instruction. Economics of Education Review, 55, 132-148. Heckman, J. J. (1979). Sample selection bias as a specification error. Econometrica, 41(2), 153- 161. 168 Hirschman, A. O. (1970). Exit, voice, and loyalty: Responses to decline in firms, organizations, and states. Cambridge, MA: Harvard University Press. Hoepfl, M. C. (1997). Choosing qualitative research: A primer for technology education researchers. Journal of Technology Education, 9(1), 47-63. Holbrook, A. L., Green, M. C., & Krosnick, J. A. (2003). Telephone versus face-to-face interviewing of national probability samples with long questionnaires. Public Opinion Quarterly, 67(1), 79-125. Hui, T. K. (2018, April 4). More than one-fifth of NC teachers are chronically absent. How does it affect your child? The News & Observer. Retrieved from https://www.newsobserver.com/news/local/education/article207903534.html Ingersoll, R. M. (2002). The teacher shortage: A case of wrong diagnosis and wrong prescription. NASSP Bulletin, 88(631), 16-31. Ingersoll, R. (2003). Is there really a teacher shortage? CPRE Research Reports. Retrieved from https://repository.upenn.edu/cpre_researchreports/37/ Ingersoll, R. M. (2003). Who controls teachers’ work? Power and accountability in America’s schools. Cambridge, MA: Harvard University Press. Ingersoll, R. M. & Smith, T. M. (2003). The wrong solution to the teacher shortage. Educational Leadership, 60(8), 30-33. Ingersoll, R. (2011). Teacher turnover and teacher shortages: An organizational analysis. American Educational Research Journal, 38(3), 499-534. Irvine, A., Drew, P., & Sainsbury, R. (2012). ‘Am I not answering your questions properly?’ Clarification, adequacy and responsiveness in semi-structured telephone and face-to-face interviews. Qualitative Research, 13(1), 87-106. Jacob, B. A. (2013). The effect of employment protection on teacher effort. Journal of Labor Economics, 31(4), 727-761. James, N. & Busher, H. (2006). Credibility, authenticity and voice: Dilemmas in online interviewing. Qualitative Research 6(3), 403-420. 169 Jochim, A. (2018). The limits of policy for school turnaround. In F. M. Hess & M. Q. McShane (Eds.), Bush-Obama school reform: Lessons learned (13-31). Cambridge, MA: Harvard Education Press. Kane, T. J., Rockoff, J. E., & Staiger, D. O. (2008). What does certification tell us about teacher effectiveness? Evidence from New York City. Economics of Education Review, 27, 615- 631. Kiersz, A. & Perino, M. (2019, August 21). Here’s how much every US state pays its teachers and spends on a single student. Business Insider. Retrieved from https://www.businessinsider.com/teacher-salary-in-every-state-2018-4 Klein, H. J., Cooper, J. T., Molloy, J. C., & Swanson, J. A. (2014). The assessment of commitment: Advantages of a unidimensional, target-free approach. Journal of Applied Psychology, 99(2), 222-238. Knox, S. & Burkard, A. W. (2009). Qualitative research interviews. Psychotherapy Research, 19(4-5), 566-575. Kraft, M. A. (2018). Federal efforts to improve teacher quality. In F. M. Hess & M. Q. McShane (Eds.), Bush-Obama school reform: Lessons learned (13-31). Cambridge, MA: Harvard Education Press. Lamont, M. & Swidler, A. (2014). Methodological pluralism and the possibilities and limits of interviewing. Qualitative Sociology, 37(2), 153-171. Lankford, H., Loeb, S., & Wykoff, J. (2002). Teacher sorting and the plight of urban schools: A descriptive analysis. Educational Evaluation and Policy Analysis, 24(1), 37-62. Lather, P. (2006). Paradigm proliferation as a good thing to think with: Teaching research in education as a wild profusion. International journal of qualitative studies in education, 19(1), 35-57. Leigh, J. P. (1986). Correlates of absence from work due to illness. Human Relations, 39(1), 81-100. Leigh, J. P. (1991). Employee and job attributes as predictors of absenteeism in a national sample of workers: The importance of health and dangerous working conditions. Social Science Medicine, 33(2), 127-137. 170 Manna, P. (2007). Patterns of federal interest in education. In School’s In: Federalism and the national education agenda. Washington, D.C.: Georgetown University Press. Marvasti, A. & Freie, C. (2017). Research interviews. In Wyse, D., Selwyn, N., Smith, E. & Suter, L. E. (Eds.), The BERA/SAGE Handbook of Educational Research (pp. 624-639). London: SAGE Publications Ltd. Mathieu, J. E. & Zajac, D. M. (1990). A review and meta-analysis of the antecedents, correlates, and consequences of organizational commitment. Psychological Bulletin, 108(2), 171- 194. Maxwell, J. A. (2013). Qualitative research design: An interactive approach. California: SAGE Publications, Inc. McDermott, K. (2011). High-States Reform: The Politics of Educational Accountability. Washington, D.C.: Georgetown University Press. Mendro, R. L. (1998). Student achievement and school and teacher accountability. Journal of Personnel Evaluation and Education, 12(3), 257-267. Merriam, S. B. & Tisdell, E. J. (2016). Qualitative research: A guide to design and implementation. San Francisco, CA: Jossey-Bass. Meyer, J. P., Stanley, D. J., Herscovitch, L., & Topolnytsky, L. (2002). Affective, continuance, and normative commitment to the organization: A meta-analysis of antecedents, correlates, and consequences. Journal of Vocational Behavior, 61, 20-52. Michigan’s Center for Educational Performance and Information, Educational Entity Master 2018-19. https://cepi.state.mi.us/eem/ (accessed October 12, 2019). Michigan’s Center for Educational Performance and Information, Staffing Count 2018-19. https://www.mischooldata.org/DistrictSchoolProfiles2/StaffingInformation/StaffingCount s3/StaffingCount.aspx (accessed October 12, 2019). Michigan’s Center for Educational Performance and Information, Student Counts 2018-19. https://www.mischooldata.org/DistrictSchoolProfiles2/StudentInformation/StudentCount s/StudentCount2.aspx (accessed October 12, 2019). 171 Michigan Department of Education (MDE). (2017). Academic year 2016-17 critical shortage disciplines for employment of retirees. State of Michigan. Miller, R. (2012). Teacher absence as a leading indicator of student achievement: New national data offer opportunity to examine cost of teacher absence relative to learning loss. Washington, D.C.: Center for American Progress. Miller, R. (2017, March 3). Being there matters: Tracking student and teacher attendance. FutureEd. Retrieved from https://www.future-ed.org/being-there-matters-student-and- teachers-attendance/ Miller, R. T., Murnane, R. J., & Willett, J. B. (2008). Do teacher absences impact student achievement? Longitudinal evidence from one urban school district. Educational Evaluation and Policy Analysis, 30(2), 181-200. Mitra, D. L. (2017). Educational change and the political process. New York, NY: Routledge. Mowday, R. T., Porter, L. W., & Steers, R. M. (1982). Employee-organization linkages: The psychology of commitment, absenteeism, and turnover. New York, NY: Academic Press, Inc. Muchinsky, P. M. (1977). Employee absenteeism: A review of the literature. Journal of Vocational Behavior, 10, 316-340. National Council on Teacher Quality. (2014). Roll call: The importance of teacher attendance. Washington, D. C.: Joseph, N., Waymack, N., & Zielaski, D. No Child Left Behind Act of 2001, Pub. L. No. 107-110 § 115 Stat. 1425 (2001-2002). Opdenakker, R. (2006). Advantages and disadvantages of four interview techniques in qualitative research. Forum: Qualitative Social Research, 7(4). Owen, C. (2017). Surveys. In Swain, J., (ed.) Designing Research in Education Concepts and Methodologies. (pp. 123-147). London: Sage Publications Limited. Paringer, L. (1983). Women and absenteeism: Health or economics? The American Economic Review, 73(2), 123-127. 172 Patel, D., Shah, P., Lakhey, S. (2009). Improving student enrollment and teacher absenteeism outcomes through social accountability interventions in Nalgonda and Adilabad districts, Andhra Pradesh, India. Social Accountability Series, World Bank, 8, 1-12. Patrinos, H. A. (2013). The hidden cost of corruption: Teacher absenteeism and loss in schools. In Transparency International, Global corruption report: Education (70-73). New York, NY: Routledge. Patton, Q. M. (2015). Qualitative research and evaluation methods (4th ed.). Thousand Oaks, California: Sage. Renk (2019, March 4). What is going on in Michigan and Detroit schools with absenteeism? Retrieved from https://wbckfm.com/what-is-going-on-in-michigan-and-detroit-schools- with-absenteeism/. Rhodes, P. J. (1994). Race-of-interviewer effects: A brief comment. Sociology, 28(2), 547-558. Riordan, C. M. (2000). Relational demography within groups: Past developments, contradictions, and new directions. Research in Personnel and Human Resources Management 19, 131- 173. Roby, D. (2013). Teacher attendance effects on student achievement: Research study of Ohio schools. Education, 2(Winter 2013), 201-206. Rogers, F. H. & Vegas, E. (2009). No more cutting class? Reducing teacher absence and providing incentives for performance. Policy Research Working Paper, February(4847), 1-31. Ronfeldt, M., Loeb, S., & Wyckoff, J. (2013). How teacher turnover harms student achievement. American Educational Research Journal, 50(1), 4-36. Rossman, G. B. & Rallis, S. F. (1998). Learning in the field: An introduction to qualitative research. Thousand Oaks, California: Sage. Rubin, H. J. & Rubin, I. S. (2012). Qualitative interviewing: The art of hearing data (3rd ed.). Thousand Oaks, CA: Sage. 173 Sagie, A. (1998). Employee absenteeism, organizational commitment, and job satisfaction: Another look. Journal of Vocational Behavior, 52, 156-171. Saldaña, J. (2015). The coding manual for qualitative researchers. Thousand Oaks, California: Sage. Schaeffer, N. C. (1980). Evaluating race-of-interviewer effects in a national survey. Sociological Methods & Research, 8(4), 400-419. Schensul, S. L., Schensul, J. J., & LeCompte, M. D. (1999). In-depth, open-ended interviewing. In Essential ethnographic methods: Observations, interviews, and questionnaires (pp. 121-148). Walnut Creek, Calif: AltaMira Press. Schensul, S. L., Schensul, J. J., & LeCompte, M. D. (1999). Semistructured interviewing. In Essential ethnographic methods: Observations, interviews, and questionnaires (pp. 149- 164). Walnut Creek, Calif: AltaMira Press. Scott, K. D. & Taylor, G. S. (1985). An examination of conflicting findings on the relationship between job satisfaction and absenteeism: A meta-analysis. The Academy of Management Journal, 28(3), 599-612. Seidman, I. (2006). Chapter 1, Why Interview? Interviewing as qualitative research: A guide for researchers in education and the social sciences, 3rd edition. (p. 7-14). New York: Teachers College Press. Seidman, I. (2013). Interviewing as qualitative research: A guide for researchers in education and the social sciences. New York: Teachers College Press. Smith, R. L. & Hopkins, A. H. (1979). Public employee attitudes towards unions. Industrial and Labor Relations Review, 32, 484-495. Spradley, J. P. (2016). The ethnographic interview. Long Grove, Illinois: Waveland Press. Springer, M. G., Ballou, D., Hamilton, L., Le, V., Lockwood, J. R., McCaffrey, D. F., Pepper, M., & Stecher, B. M. (2010). Teacher pay for performance: Experimental evidence from the project on incentives in teaching. National Center on Performance Incentives, 1-86. 174 Steers, R. M. & Rhodes S. R. (1978). Major influences on employee attendance: A process model. Journal of Applied Psychology, 63(4), 391-407. Stephens, N. (2007). Collecting data from elites and ultra elites: Telephone and face-to-face interviews with macroeconomists. Qualitative Research, 7(2), 203-216. Strunk, K. O., Cowen, J. M., Torres, C, Burns, J., Waldron, S., & Auletto, A. (2019). Partnership turnaround: Year one report, October 2019. Retrieved from the Education Policy Innovation Collaborative website: https://epicedpolicy.org/wp- content/uploads/2019/10/Year-One-Partnership-Report.pdf Thomas B. Fordham Institute. (2012). How strong are U.S. teacher unions? A state-by-state comparison. Washington, D.C.: Winkler, A. M., Scull, J., & Zeehandelaar, D. Thomas B. Fordham Institute. (2017). Teacher absenteeism in charter and traditional public schools. Ohio: Griffith, D. Valli, L., Croninger, R. G., & Walters, K. (2007). Who (else) is the teacher? Cautionary notes on teacher accountability systems. American Journal of Education, 113, 635-662. Wolcott, R. J. (2017, August 24). Does your child’s teacher miss too much school? Lansing State Journal. Retrieved from https://www.lansingstatejournal.com/story/news/local/2017/08/24/mid-michigan-schools- absent-teachers/577909001/ World Bank. (2004). Ecuador: Creating fiscal space for poverty for poverty reduction: A fiscal management and public expenditure review (Volume 2). Washington, D.C.: World Bank. World Bank. (2013). Provider absence surveys in education and health: A guidance note. Washington, D.C.: Rogers, H. & Koziol, M. Yin, R. K. (2018). Case study research and applications: Design and Methods. California: SAGE Publications, Inc. 175