School Turnaround Lessons for Policy and Practice: A Systematic Review of Research and Evaluation

: While there currently exists a vast body of research around school turnaround policies, few studies speak to why they have or have not worked. This paper undertakes a systematic review of research literature on school turnaround policies to explore why this lack of understanding prevails in the field. We find key disconnects between “policy” and “practice” research on the topic of school turnaround. We contend that this divide negates the potential to learn from school leadership policy and practice in turnaround settings. We offer implications of these findings within turnaround research and policy while pointing out that this apparent divide seems to extend beyond the example of turnaround. y prácticas: Un análisis de y evaluación recuperación mientras que división evidente Abstract: Embora exista atualmente um vasto corpo de pesquisa sobre políticas de recuperação escolar, poucos estudos examinam por que elas funcionaram ou não. Este artigo realiza uma revisão sistemática da literatura sobre políticas de recuperação escolar para explorar por que essa falta de compreensão prevalece no campo. Encontramos importantes desconexões entre a pesquisa “política” e “prática” sobre o tema da recuperação escolar. Mantemos que essa divisão nega a possibilidade de aprender da política e da prática de liderança escolar em ambientes de recuperação. Oferecemos implicações dessas desconexões descobertas na pesquisa e nas políticas de recuperação escolar, e apontamos que essa aparente divisão parece se estender além do exemplo de recuperação. on understanding the role of school system leadership, especially in the context of improving the nation’s lowest-performing schools.


School Turnaround Lessons for Policy and Practice: A Systematic Review of Research and Evaluation
The concept of school turnaround has emerged over the last twenty years as a call to rapidly improve the nation's lowest-performing schools (Calkins et al., 2007;Herman et al., 2008). The intensity of school turnaround exceeds notions of school improvement or school reform, and typically relies on considerable disruption to who leads and how they lead underperforming schools (Meyers & Smiley, 2017;Redding & Nguyen, 2020). Federal (e.g., School Improvement Grants) and local (e.g., Innovation Zones) turnaround policies have typically prescribed what districts and schools should do to increase student achievement (Pham et al., 2020). For the purposes of this paper, we conceive of turnaround policies as formal attempts by a governing body to outline how districts and schools should achieve rapid improvement. Over the last 15 years the evidence base for the efficacy of these initiatives has grown, with mixed findings emerging. For example, turnaround programs led to significant increases in achievement in several contexts (Carlson & Lavertu, 2018;Player & Katz, 2016;Schueler et al., 2017;Sun et al., 2017;Thompson et al., 2016) but not in others (Heissel & Ladd, 2018;Hochbein, 2012;Meyers et al., 2012;Stuit, 2012). While the research on the impact of turnaround policy makes broad claims about policy efficacy (or lack thereof), few studies are able to describe why these policy initiatives have or have not worked.
In part, the answer to why these policies succeed or fail is unknown because studies simply have not asked this question. Indeed, many of the conclusions of studies in the school turnaround policy arena suggest that more research is needed to understand these contextual factors. Even while policies and policy evaluations in this space continue to grow, no evidence indicates that this call is being taken up. Despite differences in their context and objectives, turnaround policies routinely identify principals as key to leading rapid school improvement (Herman et al., 2008). What is known suggests that school principals, ostensibly those most responsible for leading turnaround, feel uncertainty about which aspects of turnaround policies are the ones leading to the desired change (Welsh, 2019); a lack of district support (Meyers & Sadler, 2018); and a psychological toll on their wellbeing (Yoon & Barton, 2019). Thus, while the ambitious goal of school turnaround is to rapidly improve outcomes for the most vulnerable students, progress may come with a price for those tasked with turnaround.
Without more interplay between-or blurring of (Hargreaves, 1996)-research focused on policy implementation and outcomes and the on-the-ground experiences of those doing the work, it is unlikely that we will fully understand how or why turnaround policy and research have or have not been effective. For these reasons, we conducted a systematic review of school turnaround policy evaluation and research to inform policymakers, researchers, and others about the ways in which related bodies of literature might not fully synchronize. Through shedding light on this issue, we hope to elucidate not only why policy and practice research needs to overlap in order to better serve marginalized students, but also why we believe these disconnects in research exist and how the field can overcome them moving forward.
In this systematic review, we document the gap between what is known about the efficacy of school turnaround policies and the contextual factors that may lead to these outcomes. To do so, we pose two research questions. First, how has school turnaround been conceptualized within policy and practice-primarily leadership-research? Second, how do discrete features of the research articles-including research questions, units of analysis, findings, limitations, overlaps, and implications-differ across the policy and practice literature? To answer these questions, we systematically review research literature on school turnaround policies between the years 2000 and 2020, spanning the early years of the No Child Left Behind ([NCLB], 2002) policy in the United States and concluding with the most recent research on School Improvement Grants (SIGs).
As many schools nationwide continue to require turnaround interventions (Meyers et al., 2022), simply exploring whether or not such policy initiatives work is no longer adequate, not only because such efforts require high levels of funding from federal and state governments, but also because without understanding more about the factors that impact success or failure, schools and districts operate with limited knowledge about how to utilize such funding. We thus argue that not only is more research needed to understand turnaround efficacy, but also a specific type of research is needed, one that is able to focus on the explanatory factors related to turnaround across contexts. This systematic review points the way for such research, indicating both where such policies may be most impactful as well as the types of studies that are needed and why.

School Turnaround: An Overview
In fewer than 15 years, school turnaround has been conceptualized in several distinct ways. In 2008, a federally sponsored practice guide (Herman et al., 2008) legitimized the term, framing it as an effort to dramatically increase student achievement in three years and emphasizing the need to signal change through the school principal. Others (Murphy & Meyers, 2008) conceptualized the term more broadly, focusing on any strategic initiative prioritizing the rapid improvement of the lowest-performing schools. Shortly after, researchers (e.g., Stuit, 2012) developed statistical methods to identify the frequency with which schools achieved turnaround. Simultaneously, the federal government established SIG funds with models of turnaround prioritizing changing school leadership (Le Floch et al., 2014). Researchers and policymakers have also identified additional aggressive policy initiatives as turnaround policies that, unlike many improvement and reform initiatives, are externally imposed (Schueler, 2019). As a result of the aforementioned undertakings, our understanding of the concept of turnaround as a policy mechanism and school leadership framework has been interrupted and obfuscated. In this systematic literature review, we step back to analyze the concept of school turnaround and its derivations to extract lessons for both policy and practice.

Turnaround Policy
Since the advent of No Child Left Behind, the federal government of the United States has taken a more deliberate approach toward school improvement, focusing on a combination of incentives, additional funds, and consequences (e.g., status labeling) as vehicles for improvement and reform. In a recently published meta-analysis, Redding and Nguyen (2020) describe the four mechanisms by which school turnaround is hypothesized to affect student outcomes: improved organizational performance, changes in human capital, changes in governance/management, and incentives/sanctions. Through a federal approach of increased support and surveillance, the underlying assumption is that schools, districts, and/or those who work within them are either unwilling or unable to do the work of improving outcomes. Another way of saying this is that a prevailing assumption contends that school leaders do not have the expertise to turn around schools but instead must rely upon outside experts to guide their decision-making. Indeed, much of the research is based on the notion that "All failing schools, especially those that persistently fail, need guidance on what will work quickly to improve student outcomes" (Herman et al., 2008, p. 1).
Despite increased federal spending on school turnaround policies and initiatives, the metrics for being designated a school in need of turnaround have been relatively unclear (Schmidt-Davis & Bottoms, 2012). One constant has been a focus on dramatic increases in student test scores, primarily in language arts and mathematics (e.g., Ylimaki et al., 2014). The more than $7 billion invested in SIG funding exemplifies this intensity to make substantial achievement gains quickly (Emma, 2015). No Child Left Behind (NCLB, 2002) established the School Improvement Grants program to support schools in this effort. Congressional funding for SIGs began in 2007 and additional funds were provided through the American Recovery and Reinvestment Act. To be specific, SIGs outlined the strategies that states could employ to turn schools around. Three of these four models required changing the principal, with school closure being the only exception (Le Floch et al., 2014).
The Every Student Succeeds Act ([ESSA], 2015)-a federal law replacing NCLB and giving states more autonomy in how they hold schools accountable for student achievement-eliminated the SIGs but allowed for funding for struggling schools through increased Title 1 dollars (i.e., federal funding to supplement local funding for low-income students), resulting in similar financial support for turnaround initiatives in schools deemed to be priority (VanGronigen & Meyers, 2019). Following NCLB, ESSA encouraged states to take up tiered accountability systems targeting schools performing at the lowest levels for intensive interventions (Schueler et al., 2017). This approach provides some flexibility to states through defining these tiers of evidence for consideration and use by policymakers and educators. This flexibility is the key difference between earlier NCLB/SIG policy and current ESSA policy; NCLB/SIG policy prescribed particular interventions whereas ESSA supports a wide range of improvement initiatives (Herman et al., 2017). Despite increased flexibility in turnaround policy in recent years, in our review of the literature we see little discernible change in how turnaround research and policy evaluation are conceptualized or conducted.
There are, however, at least a couple of constants in research on turnaround. First, despite policy evaluations that have shown impact on student achievement outcomes (Dee, 2012;Schueler et al., 2017), many policymakers, researchers, and practitioners have come to believe that the overall results of such federal programs have been disappointing (Murphy & Bleiburg, 2019). Second, despite policy and scholarly insistence (e.g., Herman et al., 2008;Murphy & Meyers, 2008) on the importance of principals in leading turnaround, policy evaluations have mostly neglected the school leader when exploring the success or failure of particular initiatives. These findings appear to persist across the two waves of policies: NCLB and ESSA. Instead, studies of school turnaround can largely be considered distinctly as either policy evaluations (i.e., did school turnaround policy impact student achievement outcomes?) or case studies of principal action (i.e., what did a principal do in trying to lead a turnaround effort?). Policy evaluations do not offer insights into the potential heterogeneity of how these policies were enacted on the ground, while findings from the case studies cannot be generalized beyond the one particular school leader on whom the article focuses. The apparent disconnect between the way researchers and evaluators have analyzed school turnaround seems potentially inhibitive to learning what did (or did not) work when, for whom, and why (or why not).

Principal Leadership and Turnaround Schools
Much of the research on turnaround policy has focused on the effect of improvement efforts on student outcomes. There is some evidence, however, about the role that school leaders play in turnaround initiatives' success or failure and their experiences working in turnaround schools. As research in educational administration repeatedly emphasizes, school leaders play an essential role in student outcomes generally (Louis et al., 2010) and improvement efforts more specifically (Herman et al., 2017). Indeed, new research indicates that principals have the greatest effect on student outcomes (Grissom et al., 2021), even more than teachers, although their work is largely indirect. Even so, little is known about these mechanisms of improvement. The limited evidence on principals and turnaround suggests three key areas for consideration. First, much of the work of leading turnaround schools is primarily personnel focused rather than policy focused. That is to say, leaders spend much of their time helping teachers through the reform efforts. Schools that improved culture often fared better than schools that did not (Brown et al., 2017;Cucchiara et al., 2015;Pulliam et al., 2014). Second, the work is challenging and can have detrimental effects on principals' work and career satisfaction. Yoon and Barton (2019), as well as others (e.g., Peck & Reitzug, 2014), describe the work as taking a psychological toll on principals as they engage in the demanding and, at times, relentless work of school turnaround. Finally, of the few studies that exist on this important topic, many, if not most, are single case studies of leaders or districts engaged in the process of improvement. As we describe later in this paper, it is evident that more research is needed to understand turnaround work qualitatively at a larger scale.
Thus, answering our research questions remains critically important because such an endeavor will help shed light on how the concept of school turnaround has been understood and enacted, and what lessons there are for policy and practice going forward given the field's continued commitment to improve low-performing schools, regardless of terminology. For instance, we still do not know what principals do or need to do to actually lead turnaround. If states continue to promulgate policies that promote turnaround as a change mechanism, we need to know what principals need to do and, relatedly, how to prepare them to do it effectively.

Methods
To respond to our research questions, we concurrently conducted a systematic review of existing literature on school turnaround from 2000 to 2020, searching the term "school turnaround" in the database EBSCO, which houses several significant education databases, and was accessible to all researchers on the team. We searched the term in quotations to identify literature addressing the concept specifically while not adding any other limiting search terms in order to retrieve all studies of school turnaround regardless of the various ways researchers and evaluators might conceptualize or define the term. We did, however, restrict our search to the option "find all of my search terms" instead of Boolean phrases, and limited results to scholarly (peer-reviewed) journals and then academic journals. Our initial search resulted in 573 article abstracts. We then divided the abstracts among the three of us to review for fit. Our determination of fit included whether (a) the article used the words "school turnaround"; (b) the study was set in the United States; and (c) the study context was focused on K-12 public schools. Each of us designated articles as relevant or irrelevant. At this stage, we retained articles unless we were completely confident from their abstracts that they failed to meet our criteria. We then redistributed the remaining abstracts among us to review for empiricism. To be included, the study needed to be an empirical examination of school turnaround or its effects and not a theoretical one. We excluded reviews and reports not in journals. After this review, we determined that 75 peer-reviewed articles met the criteria for empiricism.
Subsequently, we separated articles into two groups based on their primary focus area: practice or policy. We came to this separation based on our prior reviews and conceptual considerations of turnaround research Murphy & Meyers, 2008). Within this sphere, we observed differences in type of journal (based on journal aims) as well as general research questions. Indeed, this distinction is what anchored our overarching aim for this study (i.e., the specific ways in which these two bodies of work lack overlap). We then chose a subset of 10 articles, five from practice and five from policy, to assess the extent to which we agreed on whether articles were fit for full review. This was our attempt to evaluate inter-rater reliability. We agreed on 90% of the articles and discussed the one disagreement to understand our difference and resolve its inclusion status. With such agreement, we separated the articles for review again, reviewing the remaining 65. We followed the same basic process to analyze complete article texts with the same topical and empirical expectations for fit, before further reducing the total by excluding articles focused on school improvement or reform more broadly or partner organizations or charter organizations, for example, specifically. We excluded studies on school improvement and reform identified by the search engine but not specifically or intentionally referencing school turnaround because the concepts of improvement and reform typically do not frame change with the same level of urgency as that found in school turnaround. Forty-three of the remaining articles were classified as empirical research and evaluation of school turnaround in traditional public-school contexts, of which 27 focused on policy and 16 on practice. At each stage of the review process, we met as a team and discussed decisions. Figure 1 provides a decision tree that illustrates this process.

Data Source and Analysis
The 43 studies identified through the systematic review process described above comprise the data source for this study (see References for a list of all of the studies reviewed). To answer each question, we created a shared document for reporting and then divided categories evenly across the research team for analysis. The categories for analysis included the basic components of research papers: (1) How is turnaround defined; (2) What are the research questions; (3) What is the unit of analysis; (4) What did the study find; (5) What did the study identify as limitations; (6) What does the study say about leadership/policy; and (7) What does the study say is needed for follow up.
Upon review of each article using this format, the research team then reanalyzed each section, looking for patterns of similarity and categorizing accordingly. For example, under defining turnaround, the articles were grouped according to no definition given, not clearly stated but generally about rapid improvement of test scores, naming of policy, and articulation of measurement rules. After each analysis, the team met for discussion and refinement of findings.

Results
When analyzing our identified articles, we examined various sections of each paper to examine any differences between the pieces we classified as policy or practice. Specifically, we looked at turnaround definitions, research questions, unit of analyses, findings, limitations, the extent to which policy articles spoke to practice research and vice versa, and directions for future research for each paper. We decided on these categories to guide our analyses as these are typically the sections by which empirical, peer-reviewed journal articles are organized. As such, delineation by these categories allowed us to conduct appropriate comparisons across articles while ensuring that we did not overlook important details. Below we outline our findings around each of these sections, emphasizing how the two groups of studies compared to each other within each section.

Defining Turnaround
We began our analysis by examining the definition of turnaround that each article adopted. Table 1 shows our resulting categorization. Seven of the 43 total articles provided general definitions of turnaround only. Specifically, these articles mentioned the rapid improvement of test scores and/or the dramatic improvement of low-performing schools as evidence of turnaround occurring. Of these seven, we classified one as a policy article and six as practice pieces. In addition to offering broad definitions of turnaround, several articles also discussed policies that were aimed at turning schools around, such as the School Improvement Grants. Twenty-one of these were categorized as policy pieces (of the 27 total policy pieces) and 10 were deemed to be practice articles (of the 16 total practice articles). As is evident, a majority of articles in both policy and practice realms offered broad definitions of turnaround, which were accompanied by an identification of policy prescriptions on the topic. There was also one policy piece that only named a turnaround policy without offering any definition of turnaround (Abdulkadiroğlu et al., 2016). As a final designation, four articles articulated specific measurement rules-for example, student test scores must increase by x amount in y number of years-when stating their definition of turnaround. These were all policy pieces. Of these, one article, Hochbein (2012), presented both a broad and a specific definition of turnaround. The remaining three (Heissel & Ladd, 2018;Hochbein et al., 2013;Stuit, 2012) shared general definitions of turnaround, specific definitions of turnaround, and related turnaround policy. Table 1 Categorization of School Turnaround Definition

Policy
General statements about rapid improvement of test scores / improvement of low-performing / chronically low-performing schools The articles we identified as policy pieces had more heterogeneity across them when considering how the topic of turnaround was introduced, relative to the practice pieces. Despite having successfully categorized all articles by how they defined turnaround, we found that many articles did not explicitly state a definition of turnaround, either mentioning in passing that school improvement through the lens of academic achievement was being explored or that leaders who had "turned around" schools were being examined. Very few pieces explicitly stated what it meant for a school to have turned around and how that benchmark was established and measured.

Research Questions
When considering research questions (see Table 2), all 16 of the practice articles posed questions that were descriptive in nature. Of these, four also asked relational questions (Cucchiara et al., 2015;Hitt et al., 2019;May & Sanders, 2013;Pulliam et al., 2014) and one also asked a perceptual question (Yoon & Barton, 2019). Unlike the practice pieces, several of the policy studies asked causal questions. Specifically, eight studies posed causal questions, an additional four posed both causal and descriptive questions, and one posed causal and relational questions. Ten of the policy studies asked only descriptive questions, while one posed descriptive and perceptual questions (Schueler, 2019), two asked descriptive and relational questions (Hochbein & Carpenter, 2017;, and one addressed descriptive, relational, and psychometric questions (Hochbein et al., 2013). Similar to the findings when examining turnaround definitions, policy articles had a lot more variation in the types of research questions that they were tackling, relative to the practice pieces.

Perceptual
Yoon & Barton (2019) Despite that variation, research questions for policy pieces were limited to quantitative framings to show frequencies, relationships, and causality. Collectively, however, only one of the policy pieces specifically asked research questions on perception (Schueler, 2019). On the other hand, no practice piece attempted to answer causal questions on the topic of turnaround. The research questions in the descriptive practice pieces also differed in scope than those in the policy pieces. Whereas the policy articles asked descriptive questions about counts, percentages, and frequencies, research questions in practice articles were framed to learn about what principals and others were doing in schools that had successfully turned around or were attempting to turn around. In other words, practice pieces asked questions about what principals were doing in schools-whether or not there was evidence that the schools had been successful in turning around or not-whereas policy pieces asked questions about whether turnaround had been achieved or not, but not what contributed to or impeded turnaround.

Unit of Analysis
We then moved on to categorizing each study by its unit of analysis. Specifically, we searched for whether a study's outcomes pertained to states, districts, schools, principals, and/or students. Results can be found in Table 3. Nine (seven policy and two practice) of the 43 studies had outcomes relating to more than one category. Of these, two studies looked at state, district, and school outcomes (Schueler, 2019;Welsh & Williams, 2018), two at district and school outcomes (Marsh et al., 2012;Mette, 2014), five at school and student outcomes (Carlson & Lavertu, 2018;Dougherty & Weiner, 2019;Heissel & Ladd, 2018;Strunk et al., 2016a;Strunk et al., 2016b), one at school and principal outcomes (May & Sanders, 2013), and one at school, student, and principal outcomes (Pulliam et al., 2014). Perhaps the most striking revelation in this section was that no policy study had the principal as the unit of analysis and no practice study had the state as the unit of analysis. Even though some practice studies did focus on districts, the point to note here is that no practice study took a macro-level perspective on the topic of turnaround that was comparable to policy pieces. In a similar vein, policy pieces failed to offer a local, "on-the-ground" perspective that was similar to the practice studies. This distinction speaks to the divide across policy and practice pieces; policy pieces tend to overlook the role of the principal in turnaround while practice studies remain localized and unable to make broader claims.

Findings
Once we started categorizing findings (Table 4), takeaways converged into four broad, distinct groups (student achievement results, policy lessons, leadership lessons for districts, and leadership lessons for schools), with some studies falling into two groups. Sixteen (14 policy and two practice) studies examined changes in student academic outcomes. Of these, Dougherty and Weiner (2019) also looked at student mobility, Heissel and Ladd (2018) explored changes in student body composition, and Schueler et al. (2017) included non-test outcomes in their evaluation. Interestingly, 16 studies-of which 14 were policy focused-reported findings pertaining to policy design and consideration. Specifically, six of the studies in this group were also studies that discussed student achievement outcomes. All six of these studies were from the policy group. Eleven studies (six and five, respectively) made up group three, focusing their findings on district-level leadership lessons. Finally, group four-studies emphasizing lessons for school leaders-included 16 studies, of which 14 were from the practice group. Complementary studies by Strunk et al. (2016a;2016b) were the only studies addressing each of the four groups on some level.  Even though we collectively organized all studies into the four aforementioned categories, some important distinctions emerged across the policy and practice pieces. With two exceptions, all studies that had findings pertaining to student achievement outcomes were policy articles. Similarly, only two studies that offered policy-relevant lessons were not policy studies: Bonda and Mitchell (2015) and Mette (2013). Another way to look at this is that, with the two exceptions highlighted above, all practice studies offered findings relevant to district and school leaders, but not student achievement or policy. The most balanced group that emerged was group three, focusing on district leadership findings. Findings for school leadership were almost exclusively limited to practice articles. Thus, three of the groups were starkly split by study classification as policy or practice. Some of the issues highlighted here should be expected: categorizing articles as policy or practice based on their focus naturally leads to some division in findings. What is concerning, however, is the clear lack of overlap between those categories. The fact that we can so cleanly split these studies speaks to the divide in research on turnaround. Other than a handful, none of the policy-relevant studies spoke to implications for school or district leaders. Similarly, very few of the practice pieces had findings relevant to student outcomes or macro-level policy.

Limitations
Twelve policy and six practice studies did not identify any study limitations (see Table 5 for our categorization). All of the remaining articles identified limitations pertaining to the design of their methodological approach. These included issues such as generalizability/small sample sizes (see Hochbein & Carpenter, 2017;Strunk et al., 2016a for examples), the role of selection bias (see Yoon & Barton, 2019 as an example), and unobservable factors (see Gandhi et al., 2018 as an example). The level of focus on limitations varied noticeably across studies, with some noting the lack of generalizability quickly and broadly while others provided detailed explanations about how selection bias existed within the study and why it should inform our interpretation. No concerning or particularly relevant differences across the two groups of studies, however, emerged when analyzing this section of the papers. Yet, in this case, the similarities are noteworthy for at least two reasons. First, for such drastically different studies to be relatively similar calls into question how deeply researchers are considering limitations. Second, and perhaps more importantly, none of the studies regardless of designation consider the lack of the other (i.e., practice without policy or policy without practice) as a limitation.

Policy-Practice Overlap
When conducting our evaluation, we quickly realized that the ways in which the strands of literature "spoke" to each were thin. We show our results in Table 6. Nine of the policy studies did not prioritize or consider issues of practice at all. Another 17 of the policy studies included practice as part of a literature review or policy overview. In other words, terms such as school accountability, turnaround, School Improvement Grants, and others were described to varying degrees as parts of article introductions, research review sections, or context in research design. As previously described, most policy pieces were evaluations of policies in which the principal is central. Thus, the role of principal was discussed in terms of the policy (e.g., Heissel and Ladd, 2018). Six articles included practice in their analyses and eight articles included practice results. Of those, principal practice was typically a secondary or tertiary consideration and the depth of data and their analysis were reflected in limited reporting of leadership practice results. Only three policy articles (Marsh et al., 2012;Mette, 2014;Schueler, 2019) detailed practice considerations in the overview, analysis, and results sections.
The same basic pattern exists for studies designated as practice. Six of them did not prioritize or consider policy at all. Eleven practice articles included policy in literature review and/or contextual overview. That is, many of the practice articles that focused on turnaround principal/leadership set the stage for the study by noting the study was conducted in a school identified as turnaround or in need of turnaround. No practice study, however, specified how policy was included or considered in data analysis. Only three practice studies (Bonda & Mitchell, 2015;Mette, 2013;Mette & Stanoch, 2016) included policy as a result or finding.

Table 6
Categorization of Policy-Practice Overlap

Policy
Practice not prioritized or considered

Lessons and/or Directions for Future Work
Most studies across both the policy and practice groups offered ways to extend the existing research (results in Table 7). Eight studies were an exception to this and should be highlighted for clear difference. That is, only three policy studies did not identify any next steps whereas five practice studies did not. Not surprisingly, each set of studies was more likely to highlight lessons mapping back to their categorizations. Seventeen policy studies offered lessons or recommendations for policy but only seven did so for practice. Inversely, 10 practice studies offered lessons or recommendations for practice but only four did so for policy. Both sets of studies made notable research recommendations-25 studies total, 16 designated as policy studies and another nine designated as research studies. In general, policy studies that suggested research recommendations focused on issues such as more evaluations on the topic. Practice articles focused more on researching additional sites to build a knowledge base on how to lead turnaround. This was the most notable difference between policy and practice pieces on the topic of directions for future work.

Discussion
For this article, we systematically reviewed empirical studies of school turnaround. The results show two distinct strands, one focusing on policy and the other on practice. Approximately twice the number of studies have been conducted in or on policy than practice, yet each is significant enough to be analyzed. From our analysis, we illustrate that the two strands of literature seldom-and even then, only superficially-interact. That is, school turnaround policy evaluation advances little on the practice side and the practice research addresses school turnaround policy only as study context. The implications of this separation are substantial. We now highlight some of the most important ones.
Perhaps the most glaring implication of our analysis is how much the field's collective learning is impeded by the separation of research. Policy studies, for example, are designed primarily to evaluate whether a policy did or did not have impact, which has typically been defined as student achievement outcomes. It follows that if a policy is designed with increases in student achievement outcomes as the goal, analyzing to measure whether those impacts were achieved or not is not only practical but necessary. Yet, a substantial body of research suggests that the principal is critically important to increasing student achievement (Grissom et al., 2021;Louis et al., 2010). The federal practice guide on school turnaround (Herman et al., 2008) underscored how important school leadership is to achieving school turnaround. Moreover, the three turnaround policy options in SIG besides school closure all emphasized the importance of changing leadership practices, including in many cases terminating the current principal (Meyers, 2012). Thus, the very turnaround policies that are being evaluated hold the principal as a critical lynchpin for improvement. To evaluate the impact of the policy on student achievement outcomes alone or by including leadership as a relatively simplified-and maybe minimalized-explanatory variable negates our ability to learn critical lessons for both school leadership policy and practice in turnaround settings.
Relatedly, turnaround practice studies missed key opportunities to extend our understanding about how turnaround policy actually works on the ground. The practice studies we reviewed consistently detailed school turnaround policy in the front matter, including the introduction, literature review, and/or policy/conceptual framework. Researchers recognized that policy influences mattered for the context of their leadership studies. Few, however, made any substantial effort to write about how the policy contexts actually influenced principal knowledge, understanding, and/or behavior. Moreover, seldom did researchers return to policy considerations in their discussions, which tend to be the sections that allow for more freedom to consider implications and possibilities. In other words, researchers examining turnaround practice almost never connected their findings to policy or offered policy considerations that could further advance practice.
As a result of the previous two points, we wonder about how policies and policymakers, funders, and incentive structures merge to unintentionally undermine more robust research and policy analysis. A number of school turnaround policies identify change in leadership as critical but do not embed measuring or understanding how leadership changed or how those changes in leadership impacted other outcomes, such as student achievement. It is unsurprising, then, that many policy evaluations do not include more extensive measures of leadership or include qualitative methods of inquiry to learn more about principals in turnaround policy contexts.
Funding structures also matter on this point. Many of the policy evaluations have been financed by the federal government or state or local education agencies. These entities-in alignment with expectations embedded in a school accountability era-have funded studies to evaluate policy impacts on student achievement while seldom explicitly calling for inquiry into school leaders (Herman et al., 2017). In the articles we reviewed, we were unable to find any instance of education leadership researchers more focused on practice indicating that they had received external funding to study school principals and leaders or turnaround processes within schools.
In sum, there is considerable opportunity to rethink school turnaround and the notion of improving underperforming schools rapidly. Both policy and practice have clearly identified leadership and student achievement outcomes as critical. We contend that policymakers, funders, leaders of colleges of education, and scholars studying school turnaround should reconsider how to embed interdisciplinary work into how they operate. There are clear opportunities to share different conceptualizations and epistemologies on the front end; consider alternative methodologies for more robust learning; and prioritize policy and practice lessons regardless of the research and methodological interests at the heart of the study. For example, achieving rapid increases in student achievement and sustaining them likely requires policy not only focused on disruption, but also on engaging community. Understanding the impact of turnaround policy likely cannot be limited to a quasi-experimental design but also necessitates qualitative inquiry into understanding what changed and why. The goals of turnaround policy evaluation should go beyond whether or not fairly immediate increases in student achievement occurred, but to do so, research teams must be designed on the front end to not analyze impact and implementation data only. Ethnographers, critical race scholars, community and family engagement researchers, and others should be engaged at the outset to understand the nuance of turnaround policy when converted to practice and lived experience. Without a significantly more strategic study of school turnaround-and many other educational issues, for that matter-it seems unlikely for us to develop the robust understanding of challenges that we need to make greater, lasting improvement.