Search

Background: Over the past several years, numerous news stories have highlighted school districts' struggles in attracting personnel to adequately staff their schools. While these stories provide important anecdotes, there is little comprehensive work that documents school districts' demand for teachers, paraprofessionals, and other essential employees required for daily school operations. Further, there are considerable reasons to believe that schools' demand for staff may be changing. For instance, the COVID-19 pandemic posed unprecedented challenges, prompting schools to potentially expand or invest in new supports aimed at aiding students' learning recovery, altering instructional modalities, and enhancing mental and physical health support. Additionally, many schools are experiencing substantial changes in student enrollment overall and shifts in the demographic composition of their student populations, potentially altering these schools' needs for educators and school support staff. This scarcity of research describing the demand side of the educator labor market can primarily be attributed to the lack of readily available data documenting schools' precise needs for particular types of educators. While school districts and state administrative data have for years captured the population of teachers and staff employed within schools, and to a lesser extent the supply of educators newly licensed within each state, to my knowledge no state collects comprehensive data on the demand for school employees in each state. Several states do however report on school vacancies, either by utilizing existing administrative data or through surveys of school and district staff (Edwards et al, 2024). Additionally to my knowledge, at least one other team of researchers has characterized the demand for educators using job postings from districts (Goldhaber et al., 2023). It is well-documented that a myriad of educators have significant impacts on teachers' instruction as well as student behavior, cognition, and learning (Chetty et al., 2014b; Gershenson, 2016; Jennings & DiPrete, 2010; Kraft, 2019). Hence, for policymakers and researchers to fully comprehend the extent of students' equitable access to a high-quality education-- encompassing not only academic instruction but also a range of additional social supports--they must understand schools' and districts' demand for employees. Research Questions: This project aims to provide an extensive description of the demand for teachers and various school district employees, by documenting the attributes of positions sought after by school districts. Moreover, I aim to document trends in how demand changes over time and investigate the extent to which districts' needs for staff vary based on their characteristics. Specifically this research focuses on the following questions: 1. Which specific roles or positions exhibit the highest demand across school districts? 2. How does the demand for teachers, educators, and other school employees change throughout the school year? 3. How does the demand for teachers, educators, and other school employees differ based on observable characteristics of districts, such as the student demographics they serve, their geographical locations, and other labor market conditions? Setting and Data Collection: To address the primary data limitations observed in previous studies characterizing educator demand, I scrape data from 115 public school districts across North Carolina, starting in October 2023. Each job posting website was visited and scraped twice each week, capturing comprehensive snapshots of available job listings within each North Carolina Public School District. Subsequently, these collected data were cleaned and categorized based on job titles into distinct position types. For analysis, these datasets were combined with publicly available information concerning students and school characteristics at the district, county, and when possible, the school level. A full list of data sources and the covariates of interest from each source can be found below: 1. Common Core of Data: Student enrollment; student enrollment by race, sex, grade, ELL, FRPL, and IDEA status; total FTE teachers and staff; Urbanicity 2. American Community Survey: Economic, housing, social, and demographic characteristics of NC counties 3. North Carolina Teacher Working Conditions Survey: teachers' reports of school climate and working conditions, as well as aggregate reports of teachers perceptions of support within their schools, and self-reports of how likely they are to continue in the teaching profession. Research Design: This paper primarily utilizes descriptive statistics to characterize school districts' demand for educators, including teachers, paraeducators, and various school personnel. I employ an inductive approach utilizing job titles to code job postings into specific categories. To standardize the number of vacancies considering district size, job postings per district are scaled by both the number of full-time equivalent employees and the total student population, separately. The analysis summarizes the number and type of job postings both overall and how trends change over time. Additionally, cross-tabulations are calculated to show the variation in job postings based on district characteristics. Finally, Ordinary Least Squares (OLS) regression is employed to predict the proportion of job postings overall, as well as within individual job categories, based on the set of covariates outlined above. Preliminary Findings, and Conclusions: Data collection began in early October 2023 and has occurred twice each week since. The initial months of data collection provided valuable insights into the demand for teachers across the state and a pilot for improving data collection in the future. Currently, the data that has been collected has been cleaned and analyzed to begin to characterize job postings. However, it is important to note that the data collected thus far represents within school year potential job postings and thus may be very different from the kind of job postings I see posted in the Summer of 2024 (which I hope to include in my eventual conference presentation). Preliminary results show that approximately 30% of job postings are for teaching positions within these North Carolina school districts. Further, these postings disproportionately advertise for teachers in traditionally in-demand subject areas such as STEM, Mathematics, and EC. However, still, teachers do not account for the majority of job postings. Transportation staff represent the next most posted job opening, followed by paraprofessionals, and administrative staff (see Table 1). Finally preliminary results point towards moderate correlations between the socioeconomic makeup of districts (such as the percent of students on FRPL and county median income) and the number of job postings per student.

Black students account for 15.1% of public school students, yet represent 38.2% of out-of-school suspensions, 28.7% of referrals to law enforcement, and 31.6% of school arrests (U.S. Department of Education, 2021). Suspended and expelled youth are less likely to complete high school or attend college (Rosenbaum, 2020), and more likely to be arrested and incarcerated as adults (Bacher-Hicks et al., 2019). Researchers have considered numerous factors as possible explanations for racial disparities in discipline (Welsh and Little, 2018). Studies tend to focus on the behaviors and characteristics of students (Skiba et al., 2002; Wallace et al., 2008; Huang and Cornell, 2017), decision-making among school teachers and administrators, including discipline referrals and sanctions (Owens and McLanahan, 2020; Skiba et al., 2014), and sorting into punitive schooling environments (Owens and McLanahan, 2020; Gopalan and Nelson, 2019). Few studies have focused on the ways that school environments shape student outcomes through everyday routines and practices. We argue that school environments are structurally influenced by carceral ideology: a propensity to solve problems using the tools of the carceral state, including surveillance, coercion, confinement, and correction. Carceral ideology is a belief system that views societal problems as the result of individual and familial failures that require correction. Carceral ideology shapes not only school discipline, but also the distribution of knowledge, curricula, and materials. An accumulation of carceral practices in schools may alter the education, beliefs, and identities of poor and racially minoritized students in harmful ways, inhibiting their successful transition to college, the workplace, and other social environments (Golann, 2015). This study investigates not only the prevalence of carceral ideology in schools, but also explores the links between carceral ideology and the local, political context. We consider two structural drivers of punitive school environments: racialized social constructions and political ideology. First, the social construction of a target population is an important driver of policy design. According to Schneider & Ingram (1993), "deviants" who both lack power and are subject to negative stereotypes tend to be assigned burdens by policymakers. Deviant social constructions are particularly dangerous for Black people, as they can, "jeopardize the formal standing of already marginal individuals in relation to the state" (Cohen, 2004). Race is a salient aspect of social constructions, as punitive disciplinary measures are more prevalent in school districts with greater levels of segregation and racial bias (Roch and Edwards, 2017; Riddle and Sinclair, 2019). Additionally, a dependent social construction provides policy-makers and other political players with a rationale for paternalistic social control. As low-income Black people can be constructed as both dependent and deviant, institutions lean on prison-like methods of social control to constrain the flow of resources to race-class subjugated communities (Wacquant, 2009). Second, we argue that conservative political ideology mirrors carceral ideology in its goals for education. Neoconservatism in education policy decision-making promotes social control and emphasizes individual responsibility over systemic issues (Apple, 2004). Conservative belief empowers individual actors to make rational decisions, emphasizing self-reliance and freedom from state control. Mirroring carceral ideology, neoconservatives uphold individualism by blaming students and their families for inequities in student achievement, absenteeism and disciplinary disparities. Conservative policy makers favor policies that control student behavior and aim to establish a universal standard for education. This universal standard is achieved through regulation of content (curriculum) and control of student behavior. We outline two key expectations. First, as policy design is a function of the social construction of the school population, we expect that when the student population includes large numbers of Black and low-income students--who have been historically constructed as deviant or dependent--schools will be characterized by more carceral policies. Second, policy design is a reflection of political ideologies within the local community. Specifically, we expect conservative ideology within the community to correspond to more punitive school policies. To test these hypotheses, we examine the extent to which school policies reflect the tools of the carceral state-- including surveillance, coercion, confinement, and correction. To measure carceral ideology, we constructed a school-level measure of carceral ideology based on computational analyses of text from 712 North Carolina school handbooks. As formal policy documents, handbooks are widely accessible, contain relatively standardized content, and convey concise information about schools' values, goals, structure, practices, and intentions. Using a custom dictionary of carceral terms, we constructed a Carceral Ideology (CI) Score as the proportion of handbook sentences containing at least one term related to surveillance, coercion, confinement, or correction to create (See Table 1 in Appendix). Using political party affiliation as a proxy for conservative ideology, we constructed a binary variable equal to one if a plurality of voters within each school's attendance zone registered as Republican (see Figure 1). We find that carceral ideology varies across school characteristics, including charter school status, geographic locale, and population demographics (See Table 3 in Appendix). First, we find a positive, significant relationship between Carceral Ideology (CI) score and the school demographics, including the proportion of students receiving free or reduced lunch and the proportion of Black students (See Table 3 in Appendix). Second, we find that schools located in school attendance zones with a Republican Plurality of voters use more carceral language (See Table 3 in Appendix). Holding other factors constant, being located in a Republican-dominated zone increases the use of carceral language by 8.6%. For schools in conservative regions of the state, discipline rates are higher, but black-white disparities are smaller (See Tables 4 & 5 in Appendix). Our findings denote a need for education policy researchers to further explore and expand on not only school discipline, but also the broader characteristics of school environments. Additionally, surrounding community contexts play a significant role in the development of policies impacting students, specifically, students of color, which can have a detrimental impact on their later life outcomes.

Background: Research has consistently shown that disadvantaged students have inequitable access to high-quality teaching (e.g., Isenberg et al., 2013; Goldhaber et al. 2018; Hanselman, 2019). This non-random sorting of students to teachers has been documented at multiple organizational levels, including districts, schools and classrooms (e.g., Goldhaber et al., 2015; Clotfelter et al., 2023). Studies that have compared sorting across the different levels have shown that classroom-level inequalities tend to be the smallest in magnitude. However, classroom-level sorting remains a potentially high-impact target for policy intervention for several reasons. First, the mechanism through which classroom-level sorting occurs (i.e., classroom rostering) is under the direct purview of school staff and therefore highly malleable. Second, schools already have the resources required to implement classroom rostering, so changes to rostering policies have the potential to be cost-neutral or even cost-efficient. Third, because every student is rostered each year, rostering policies represent a "high volume, low margin" approach to producing net gains in equitable access and student learning. Purpose and Research Questions: The present study is a secondary analysis of administrative data obtained from the North Carolina Department of Public Instruction (NC DPI). Using this data source, we seek to more fully understand the potential of classroom rostering to reduce inequitable access to high-quality teaching and improve student learning. We approach this overall goal in two main steps. In the first step, we descriptively analyze the assignment of students to teachers within school- grade-year cells. Here we are interested to better understand the student characteristics that are associated with differential access to high-quality teaching. We categorize teaching quality using the tertiles of the distribution of teachers' value-added (VA) in the previous school year, and treat teachers without prior VA as a fourth category. We address the following descriptive questions: 1a) What proportion of school-grade-year cells have sufficient variability in effective teaching for rostering to be a potential mechanism for inequitable access to high-quality teaching? 1b) Which, if any, student characteristics are associated with differential assignment to teachers of different VA categories? Step 2 of the study builds on Step 1 by using the results of Research Question 1b to identify two contrasting subgroups of schools, one group with relatively equitable rostering practices and the other with relatively inequitable rostering practices. In these two contrasting cases, we address two additional questions about the effects of each of three counterfactual rostering policies. The three policies are: Counterfactual Policy 1: Balanced classrooms (no tracking). Counterfactual Policy 2: Assign stronger teachers to students with lower prior achievement. Counterfactual Policy 3: Assign students to teachers to maximize net predicted learning gains while maintaining balance. Clotfelter et al. (2023) recently suggested the first two policies and the third is novel to this study. We study the effects of each policy on two outcomes of interest: 2a) access to high-quality teaching and 2b) student learning. Data Collection and Analysis. We conduct a secondary analysis of administrative data from NC DPI over the 2011-2012 to 2018-2019 school years. We focus on ELA and Math instruction for students in 4th and 5th grades. The panel includes data on 1.2 million students and 57,854 teachers across 1730 schools from all 110 districts in NC. Data cleaning was completed in early Spring 2024, with a working paper expected to be completed in Summer 2024. Step 1. To address Research Question 1a we tabulate the number of teachers in each VA category within school-grade-year cells. To address Research Question 1b we use Poisson regression with fixed-effects for school-grade-year cells (see Springer et al., 2022). We consider a standard fixed effects specification and a specification that additionally incorporates random slopes on coefficients describing differential assignment. This is an extension of Mundlak's (1978) specification for panel data that allows us to study how rostering practices differ over school-grade-year cells, while still focusing on sources of variation that arise within cells. Step 2. We use the predicted random slopes from Research Question 1b to create two groups of schools, those in which there is strong evidence of highly equitable rostering, and those in which there is strong evidence of highly inequitable rostering. The goal is to identify 50-100 schools that unambiguously fall into each group and contrast results across groups. The three counterfactual rostering policies can be implemented by solving standard mixed- integer programs (MIPs) within each school-grade-year cell. Details of the objectives and constraints of these MIPs are currently under investigation and will be reported. The MIP solutions are a set of binary decision variables indicating whether each student is assigned to each teacher -- i.e., a set of counterfactual rosters. Having the counterfactual rosters in hand, Research Question 2a can be addressed by applying the same Poisson regression model described for Research Question 1b, but this time using the counterfactual rosters. It is important to note that Research Question 2b and Counterfactual Policy 3 both require specifying a predictive model for student learning that incorporates complementarity between students and teachers. In the present context, this predictive model is referred to as a match- output function (Graham, 2011). We have previously researched match-output functions for classroom rostering (Halpin et al. 2022) and arrived at a specification in which teachers' leave- year-out VA is estimated separately for each tertile of the within-school distribution of student prior achievement. This is a special case of Delgado's (2022) approach to teacher comparative advantage. We discuss the strengths and limitations of this specification, the latter mainly having to do with imprecise estimation of VA within prior achievement tertiles. We also discuss the conditions under which the match-output function is causally identified (Graham et al., 2016). Results are summarized in terms of the average re-allocation effects on student learning for each of the three counterfactual policies, computed within each school-grade-year cell. Research Question 2b will be addressed by comparing the distribution re-allocation effects in the two contrasting subgroups of schools.

Policymakers argue that only by enlarging the science, technology, engineering, and mathematics (STEM) pipeline in a way that attracts, supports, and sustains the participation of students from all kinds of backgrounds can the United States meet its needs for science and technology innovation, economic prosperity, and social well-being (National Academies, 2005). To meet this need, inclusive STEM high schools (ISHSs) combine rich STEM course offerings and experiences with an explicit mission to serve students from under-represented groups accepted on the basis of interest rather than competitive examination. One of the distinctive features of the present ISHS study is that it provides a comprehensive picture of the impact of ISHSs by using two sets of comparison groups: schools in the same districts as the ISHSs to control for local context; and comparable schools in districts with no access to STEM schools to alleviate potential bias caused by student self-selection into ISHSs. The two comparisons validate each other in providing solid evidence regarding the impact of ISHSs. This study addresses the following research questions: (1) Do students attending ISHSs differ from students in other same-district high schools in terms of demographic characteristics and middle school achievement? and (2) Is there evidence of an impact of ISHS attendance on students' persistence to 12th grade, high school graduation, and college readiness and aspirations? The data indicates that North Carolina ISHSs served a diverse set of students. Compared with students in the same districts, ISHS students had slightly lower incoming academic achievement and were more likely to be African American and to come from low-income households. Within- and out-of-district comparisons provide consistent findings on the impact of ISHS attendance. ISHS attendance appears to have a positive impact on students' persistence to 12th grade, high school graduation, and college readiness and aspirations. [SREE documents are structured abstracts of SREE conference symposium, panel, and paper or poster submissions.]

Context: The use of educational technology for improving K-12 math education has expanded dramatically recently. ASSISTments (Heffernan & Heffernan, 2014), a widely used digital platform, has seen its use in schools increase significantly from 800 to 20,000 teachers over the past two years. A review of evidence-based online programs by the What Works Clearinghouse named ASSISTments as one of the few digital learning programs recommended for use in response to the COVID pandemic (Sahni et al., 2021). An efficacy study conducted in Maine (Roschelle et al., 2016; Murphy et al., 2020) found that using ASSISTments significantly increased student scores on the TerraNova test (Hedges' g = 0.22, p < 0.001). From 2018-2020 we conducted a replication study to test the replicability of the findings in a heterogeneous population that more closely matches national demographics. A follow-up study was designed to measure the long-term impact of ASSISTments on student learning in 2021, one year after the intervention was over. This paper reports on the findings from the preliminary analysis of the follow-up study. Research Questions: Two research questions motivate the follow-up study: RQ1: What is the long-term impact of ASSISTments on student math outcomes at the end of Grade 8, one year after the completion of the intervention? RQ2: Do the long-term effects of ASSISTments, if exist, vary for students of different socio-economic status, race/ethnicity, or with other policy-relevant characteristics? Sample: The replication study took place in North Carolina (NC). Sixty-three schools were recruited and randomly assigned to condition (32 intervention and 31 comparison). The sample included 48 Title I schools and distributed across rural, town, suburban, and city communities. The study enrolled 102 7th grade math teachers and their classrooms. Table 1 compares demographics of participating schools to NC and US populations, and the Maine study sample. A generalizability analysis suggested the replication sample is highly representative of NC schools (generalizability index = 0.958) and schools across the nation (generalizability index = 0.899). Intervention: ASSISTments is a web-based platform that supports students' math problem-solving through immediate feedback and hints and provides teachers with reports summarizing student work to inform instruction adaptation. The intervention aligns with theory- and empirically-based instructional practices of formative assessment (Heritage & Popham, 2013) and skill development (Koedinger et al., 2013). During the study, the intervention was positioned as online support for math homework, a key opportunity for student independent practice, and for teachers to use formative data to adapt instruction (Figure 1). Research Design: The replication study used a randomized experimental design. Schools were randomly assigned to either intervention or business-as-usual comparison condition. ASSISTments was implemented by all Grade 7 teachers in intervention schools over two consecutive years--teachers learned to use ASSISTments for a year (2018-19) and then we measured the immediate outcome for students at the end of teachers' second year of experience (2019-20). Students who were in Grade 7 in 2019-20 comprised the analytic sample for the research questions. The students maintained their conditions and were followed longitudinally for another year when their Grade 8 performance (long-term outcome) was measured in spring 2021 (see the study overview in Figure 2). During the follow-up year in 2020-21, no interventions were provided to 8th grade teachers or students. Data: We obtained demographic, enrollment, and state assessment performance data from the state-wide database for all 63 schools. Student's long-term learning outcome was measured by Grade 8 state standardized End of Grade (EoG) math test. Grade 6 EoG math test scale scores from spring 2019 served as the baseline measure. Individual demographic covariates included gender, race/ethnicity, and economically disadvantaged status (EDS). School-level covariates included average 6th grade EoG score, 7th grade enrollment size, Title 1 eligibility, percentage of students with EDS, and percentage of ethnic groups. There were 9,073 students who (a) have 6th grade EoG score, and (b) enrolled in a participating school either as a 6th-grader in 2018-19, or as a 7th-grader in 2019-20 if the school didn't have 6th grade in 2018-19. Among these students, 5,991 have 8th-grade EoG test scores (2,961 intervention, 3,030 comparison) and were included in the analysis. 34% students (n=3,082) took an accelerated or alternative math course and did not take the Grade 8 EoG test. Their performance was analyzed separately and were not included in the paper. Analysis and Findings: There was no school-level attrition. Overall student-level attrition on Grade 8 EoG test was 33.97% (mostly due to the accelerated alternative math course enrollment) and the differential attrition was 0.31%. Counting all students with non-missing outcome data, the overall and differential attritions were 12.83% and 1.81%, respectively. We examined baseline equivalence on students' 6th grade EoG scores, gender, and ethnicity and found no significant differences between conditions. To address RQ 1, we conduct an intent-to-treat (ITT) analysis using two-level hierarchical linear regression models (students nested within schools), controlling for student's Grade 6 scores and other student- and school-level covariates. To address RQ 2, we added a cross-level interaction term of the school-level treatment variable and student subgroup indicator and conducted a series of moderator analysis to examine whether ASSISTments had differential impacts on the students with policy-relevant background characteristics. Results showed that students at intervention schools performed significantly better than those in the comparison group on 8th grade EoG (p = 0.011, g = 0.10) (Table 2). The moderator analyses showed that the intervention benefited minority (non-White) students significantly more than majority (White) students (p = 0.003, g = 0.14) and that the impact was more profound for Hispanic students than non-Hispanic students (p = 0.014, g = 0.13) (Table 3 and Figure 3). Conclusions: The follow-up study was motivated by the need to explore the long-term impact of ASSISTments use with diverse populations. The results suggested that it had a sustained long-term impact on students' math learning, after one year of intervention. The program also helped close the achievement gap among students of different ethnicities. Future works include additional moderator or mediator analyses on ASSISTments usage in 8th grade, benchmarking the effects, and estimating the cost-effectiveness.

The pandemic has been an unprecedented disruption to the lives of college students and the operation of higher education institutions. Students' performance during the pandemic was likely influenced by the hardship they personally experienced, how easily they adapted to remote teaching (including technology access), and their professors' ability to support students. Generally, college students who can better adjust to stressors are those with higher perceived levels of social support, a greater connection with campus, and lower levels of psychological distress (Pidgeon et al., 2014), and these factors vary along race and class lines (Turner & Avinson, 2003). Faculty could provide support and flexibility to struggling students, but they had to identify which students needed help. Students' backgrounds likely shaped the frequency and content of their interactions with faculty, as lower-income and first-generation students are less likely to reach out to professors (Barry, et al., 2009; Jack, 2016; Kim & Sax, 2009). To alleviate student stress, many colleges implemented COVID-specific grading policies to allow students increased flexibility in choosing to take courses pass/fail rather than for a letter grade (Cheng, 2020). COVID adjustments also often extended the window during which students could withdraw from a course and were generous in allowing students additional time to convert incompletes into grades. It is not surprising that universities relied on changes to grading policies to reduce the potential negative effects of the pandemic on students' academic performance since these policies are a common recourse for individuals facing difficulties in any given semester. Despite their ubiquity, we know little about student use of various aspects of grading policies or how student uptake of grade flexibility is associated with future success. The pandemic provides a unique opportunity to study grading policies and how students use them. Given the ongoing uncertainty of COVID-19 and the likelihood of future disruptions (illnesses, natural disasters, pandemics, etc.), universities need to have sound policies for grading accommodations that they can implement in the event of crises. Thus, it is important to understand whether and how accommodations like flexible grading policies differentially influence students from varying backgrounds, particularly since the pandemic's hardships have fallen disproportionately on non-white, low-income individuals (Garg, et al., 2020; Karpman et al., 2020). The context of this paper is the University of North Carolina (UNC) System's 16 diverse 4-year public institutions, which enrolled approximately 240,000 undergraduate students in Spring 2020. We answer the following research questions: 1. What were the pandemic-induced changes to university grading policies? How did they compare to pre-pandemic policies? Do they vary across institutions? 2. Did student course performance, take up of pass/fail options, and use of withdrawals in Spring 2020 differ from prior spring semesters? Was there variation by student characteristics? 3. Were faculty more lenient in Spring 2020 relative to past spring semesters? Did deviations from prior grading patterns vary by major/program or university? 4. Did student changes in course performance, reliance on pass/fail, or use of withdrawals in Spring 2020 predict postsecondary enrollment in subsequent semesters? We start by analyzing historical documents and policies available from the UNC System's institutions. Across 16 institutions' policies, we contrast differences and similarities in how universities adjusted how students could use pass/fail options and withdrawals. We then rely on UNC System administrative data to conduct the remaining analyses. The main dataset of interest is student transcripts showing UNC System course enrollment and course outcome (i.e., grade, pass/fail, withdrawal). We are specifically interested in students' course outcomes: letter grade, taking a class pass fail, or withdrawals, as well as university enrollment. We explore heterogeneity in outcomes by student demographic information (gender, race/ethnicity, and home zip code), socio-economic status from financial aid data, and high school test scores and GPA. To investigate differences in course outcomes, we compare the distributions of student grades, pass/fail uptake, and withdrawals in Spring 2018/2019 to the Spring 2020 semester. We also model the association between Spring 2020 and course outcomes with OLS and fixed effects regression approaches that account for student characteristics (or student fixed effects), prior academic performance, and major. Finally, we estimate these models for different groups of students (e.g., Pell eligibility, prior academic performance, race/ethnicity) to explore heterogeneity in course performance and reliance on COVID grading policies. To understand whether faculty members adjusted their grading, we estimate faculty fixed effects regression models. Here, the outcome is the average grade faculty members assign, and we look at how grades in Spring 2020 deviate from grades awarded in Spring 2018/2019. We also explore how variation in faculty grading differs across majors/programs and institutions. Finally, we show how Spring 2020 course outcomes are correlated with re-enrollment in a UNC System university in the following fall. This analysis allows us to see if course grades or the reliance of pass/fail options or withdraws could be used by universities as an early indicator that a student is disengaging from their post-secondary education. Our preliminary results (see supplemental tables) indicate a rise in the distribution of grades in Spring 2020 (regardless of student background or course department). Overall, 47 percent of grades given in Spring 2020 were A's (compared to 44 percent in Spring 2018/2019), and the share of course failures also declined (2.1 percent of courses ended in a failure in Spring 2020 versus 4.3 percent the prior two springs). However, students were about 22 percentage points more likely to take a course pass/fail in Spring 2020. These preliminary findings suggest that students were taking advantage of changes to university grading policies in Spring 2020, and that faculty were generally more lenient in their grading when students did take a course for a grade. Lastly, pass/fail uptake in Spring 2020 was associated with higher likelihood Fall 2020 re-enrollment for all students.

This paper reports preliminary results from a randomized controlled trial (RCT) of a multi-component nurturing program designed to increase gifted identification among minority students. In 2014-15, the Wake County Public School System (WCPSS), in partnership with Duke University, implemented "Nurturing for a Bright Tomorrow" ("Nurturing") in response to chronically-low gifted identification among Black and Hispanic students in two-thirds of the district's elementary schools. The purpose of "Nurturing" is to train K-2 teachers to develop the skills and expectations required to help children attain gifted status. [SREE documents are structured abstracts of SREE conference symposium, panel, and paper or poster submissions.]

School districts across the United States have increasingly adopted information and communications technology (ICT) in an attempt to improve student achievement across a wide range of educational settings. This paper examines the impacts of Achieve3000, an early literacy program that differentiates non-fiction reading passages based on individual students' Lexile scores. To estimate the causal impact of Achieve3000, the authors randomly assigned the program in 2013-14 in grades 2-5 among 32 elementary schools in the Wake County Public School System (WCPSS). The analytic sample of 32 representative schools includes those that expressed a high level of interest in the program. These schools were sorted on the percentage of students proficient in reading based on 2011-12 End-of-Grade (EOG) data and Achieve3000 was randomly assigned to one school within each ranked pair. Results suggest that Achieve3000 had mixed impacts across the range of outcomes and models. [SREE documents are structured abstracts of SREE conference symposium, panel, and paper or poster submissions.]

The purpose of this paper is to showcase new research that seeks to provide guidance on the heterogeneity of treatment effects by utilizing the variance of demographic differences in state assessments. This study is focused on a simple randomized block design where students are nested within schools, and within each school students are randomized into treatment or control conditions. The data utilized come from several states: Arizona, Arkansas, Colorado, Illinois, Kentucky, Louisiana, Massachusetts, Minnesota, North Carolina, West Virginia, and Wisconsin. From these states, the authors examined the difference between demographic groups in grades 3-8 in both math and reading scores. While the online resource will include all these grades and subjects, the authors present tables for 4th and 8th grade only for brevity. [SREE documents are structured abstracts of SREE conference symposium, panel, and paper or poster submissions.]

Given the positive impacts of the small, stand-alone early college model and the desire to provide those benefits to more students, organizations have begun efforts to scale up the early college model in a variety of settings. These efforts have been supported by the federal government, particularly by the Investing in Innovation (i3) program. This paper will describe current efforts to take the early college model and apply it in multiple settings and It will describe programs being supported by three i3 grants, focusing on the commonalities of these three efforts. The specific questions addressed by this paper include: 1) How are providers conceptualizing the scaling up of the early college model? 2) What adaptations are expected as the model is being scaled up? 3) What are early lessons learned from the scale-up efforts? The results report that one of the primary lessons learned by project and school staff was the importance of early and frequent communication about the goals of the project and the importance of working with leadership at both the district and school levels.

School districts are increasingly adopting technology-based resources in an attempt to improve student achievement. This paper reports the two-year results from randomized control trial of Achieve3000 in the Wake County Public School System (WCPSS) in Raleigh, North Carolina. Achieve3000 is an early literacy program that differentiates non-fiction reading passages based on individual students' Lexile scores. Two key questions and focal areas for research are addressed: (1) Do students who use Achieve3000 outperform students who do not use Achieve3000 on a host of reading achievement measures?; and (2) Does the performance of students who use Achieve3000 differ across student subgroups? An experimental research design in the form of a cluster-randomized trial was employed. Data for this investigation come from the school system's administrative and testing records; Amplify, Inc.'s mClass reporting system; and Achieve3000's activity completion and LevelSet Lexile pre- and post-test assessments. Results from a two-year randomized control trial of Achieve3000 suggest that the impacts, implementation, and overall promise of a technology literacy solution fell short of expectations. Tables and figures are appended. [SREE documents are structured abstracts of SREE conference symposium, panel, and paper or poster submissions.]

There is growing consensus among researchers and practitioners that children's socialemotional readiness makes unique contributions to their successful transition to and progress through school. However, many children still begin school ill-prepared for the behavioral demands they will encounter in the classroom. This study examines the joint contributions of social-emotional readiness and classroom experiences to children's social-emotional trajectories in elementary school. Specifically, using a sample of 1292 rural, mostly poor children in Pennsylvania and North Carolina, this study examines: (1) the contributions of children's prosocial behavior, inattention, and conduct problems to their developmental trajectories in these same domains; (2) the effects of classroom management, emotional support, and instructional support in kindergarten through third grade classrooms on children's social-emotional functioning; and (3) the interactive effects of social-emotional readiness and classroom experiences on social-emotional trajectories from kindergarten to grade 3. Data were drawn from a large-scale longitudinal study of children and families living in high-poverty, rural counties in North Carolina and Pennsylvania. This study utilized a longitudinal, non-experimental design. Children's prosocial behavior, inattention, and conduct problems were assessed in the year prior to kindergarten by parent and/or preschool teacher report, and they were assessed again each spring in kindergarten through grade 3 by teacher report. Children's prosocial behavior and conduct problems were assessed by the Strengths and Difficulties Questionnaire, and children's inattention was assessed with items from the ADHD subscale of the Disruptive Behavior Disorders Rating Scale, which was based on Diagnostic and Statistical Manual of Mental Disorders (DSM) diagnostic criteria. Fostering early social-emotional functioning and improving classroom experiences for children are two promising policy levers through which low-income children may be supported as they progress through school. Research suggests that achievement and behavioral gaps between higher- and lower-income children are present at school entry and in fact increase over time. It is thus important to identify the child-level and classroom-level factors that might counter this trend. Although this study is non-experimental, results contribute to the literature on the extent to which social-emotional readiness and certain classroom experiences may reduce skills gaps (i.e., encourage catch-up) among rural, mostly low-income children. Two tables and a figure are appended.

This paper presents results from the longitudinal experimental study conducted on North Carolina's early college model described in an earlier paper. The primary purpose of this paper is to present the impact of the early college model on outcomes related to postsecondary enrollment. The specific research questions driving this study include: (1) What is the impact of the early college model on students' enrollment in postsecondary education? (2) What is the impact on students' attainment of postsecondary credentials? One of the main ways the early college model accomplished its goal was by enrolling high school students into college classes as early as ninth grade. By the time students were in their junior or senior years of high school, they were taking mostly college classes with other college students. The early college also provided explicit instruction in college readiness skills and provided extensive supports to students. The study results showed early college is increasing students' access to postsecondary education, driven primarily by their experience in high school. The increase in enrollment is primarily in the area of two-year colleges because most of the early colleges are located on community college campuses. This increase in two-year enrollment is not happening at the expense of four-year enrollment, given that there is a slight increase in that arena as well. Tables and figures are appended. [For the related article: "The Small, Stand-Alone Early College: Impact on High School Outcomes," see ED567039.]

North Carolina's Early College model is the subject of an IES-funded eleven-year longitudinal experimental study that utilized a lottery process to assign early college applicants to either treatment or control groups. This paper presents findings related to high school outcomes. The primary goal of the early college model is to increase the number of students who graduate from high school and who continue on to and succeed in college. The specific research questions addressed in this paper include the following: (1) What is the impact of the early college model on the percentage of students successfully completing a college preparatory course of study in high school? (2) What is the impact of the early college model on the number of college credits earned while in high school? (3) What is the impact of the early college model on students' graduation from high school? (4) How does the impact vary for specific sub-groups including those who are low-income, first in their family to go to college, members of underrepresented minority groups, and those who entered high school below grade level? Results from this study show that the Early College High School model is having positive impacts on students' college preparatory course-taking, their attainment of college credits, and their graduation from high school. Tables and figures are appended. [For the related article: "The Small, Stand-Alone Early College: Impact on Postsecondary Outcomes," see ED567041.]

Funded by an Investing in Innovation (i3) Validation grant, the Rural Innovative Schools (RIS) Project is the first widespread effort to scale up the early college model by implementing it in comprehensive high schools. This paper will present preliminary findings from the evaluation of this project. The impact study uses a quasi-experimental design to assess the impact of the RIS Project on a core set of student outcomes. The treatment schools were matched to schools that were equivalent on a core set of baseline characteristics. The specific outcomes examined in this study include: (1) Percent of students who have enrolled in at least 1 College-Credit Bearing- Course by the end of 11th grade; (2) Average number of college-credit bearing courses students have taken and passed by the end of 12th grade; (3) Cohort graduation rate; (4) Attendance; (5) Dropout rate; (6) College preparatory course-taking; and (7) College preparatory course-taking and success. The results from this study will show the extent to which implementation of early college strategies in comprehensive high schools can have the same impact as the stand-alone early college model. Tables and figures are appended.

A growing body of research has shown that high quality early care and education (ECE) is positively related to the development of children's social and academic skills (e.g., Barnett, 2011; Lamb & Ahnert, 2006; NICHD ECCRN, 2006). There is evidence that high quality ECE experiences can improve children's levels of social adjustment (Bierman et al., 2014; Landry et al., 2014) and early academic skills (e.g., Burchinal, Kainz, & Yaping, 2014; Yoshikawa et al., 2013). However, the results from large-scale program evaluations (e.g., Administration for Children and Families, 2010; Lipsey, Hofer, Dong, Farran, & Bilbrey, 2013) and observational studies (e.g., Belsky et al., 2007; Peisner-Feinberg et al., 2001) suggest that the positive effects of high quality ECE tend to diminish in long term. One possible explanation that is given for these fade out effects is that children who are in most need of high quality ECE support often transition from emotionally supportive and structured classrooms into lower quality elementary school environments (e.g., Lee & Loeb, 1995), which may override the positive effects of high quality ECE. What is less known is how the positive effects of high quality ECE may manifest in longer term if they were followed by a high quality kindergarten environment? Could two consecutive years of high quality ECE, experienced during pre-kindergarten and kindergarten years, serve as a buffer that protects children from the fade-out effects that are often reported in the literature? The primary goal of the present study was to explore whether two years of emotionally supportive and structured classroom environments during pre-kindergarten and kindergarten years can predict the initial level and growth of social and academic skills during first, second, and third grades. Based on the results of the current study, it appears that in terms of children's social skills, there is no evidence of fade-out effects associated with emotionally supportive and structured classroom environment and that children who experienced high quality ECE in pre-kindergarten and kindergarten years had the highest level of social skills during first, second, and third grades comparing to children who did not have high quality ECE in pre-kindergarten and kindergarten or had high quality ECE in kindergarten only. Overall, the results suggest that high levels of emotional support and classroom organization in both pre-kindergarten and kindergarten years appear to have long term impacts on children's social skills, but not on academic skills. Tables and figures are appended. [The folllowing Family Life Project Key Investigators contributed to this document: Lynne Vernon-Feagans, Mark Greenberg, Martha Cox, Clancy Blair, Margaret Burchinal, Michael Willoughby, Patricia Garrett-Peters, Roger Mills-Koonce, and Maureen Ittig.]

Cluster randomized experiments are now widely used to examine intervention effects in prevention science. It is meaningful to use empirical benchmarks for interpreting effect size in prevention science. The effect size (i.e., the standardized mean difference, calculated by the difference of the means between the treatment and control groups, divided by the pooled standard deviation of the two groups) provides one way to interpret the substantive or practical significance as compared to the statistical significance of interventions. The purpose of this study is to provide reference values of design parameters by analyzing data from four large scaled intervention studies. In particular, to estimate: (1) the empirical benchmark of meaningful effect sizes regarding the developmental annual growth, and policy-relevant demographic performance gaps (race/ethnicity, gender, and SES); (2) ICC for schools or/and teacher/class; and (3) R-squared at different levels (school, teacher/class, and student). The authors will focus on 3-level cluster randomized experiments. The data structure of a 3-level cluster randomized experiments can be students (level 1) nested within classrooms (level 2), and classrooms nested within schools (level 3). Data for this study came from four separate cluster randomized trials (RCTs) evaluating the effectiveness of school-based prevention interventions. Two of the cluster RCTs evaluated the efficacy of the Incredible Years Teacher Classroom Management (IY TCM) program. Data from two additional cluster RCTs investigated the impact of a school-level prevention intervention called Positive Behavior Intervention and Support (PBIS). Findings included: (1) Normative expectations for change; (2) Demographic performance gaps among subgroups; and (3) Intraclass correlations (ICCs) and proportions of variance explained by the covariates (R-squared). Tables are appended.

The changing nature of the U.S. economy has fostered concerns that too few students are successfully completing postsecondary education (Achieve, 2004). Three quarters of those who enter high school graduate within four years, with approximately 70 percent of those graduates enrolling immediately in some form of postsecondary education (Ross et al., 2012). Of those who do attend, insufficient numbers complete a degree with only a little less than half (49 %) of beginning postsecondary students attaining some sort of a postsecondary credential within six years of enrolling (Ross, et al., 2012). As a result, there have been numerous initiatives to increase the number of students who graduate from high school prepared to enroll and progress in postsecondary education. One approach is the Early College High School (ECHS or early college) model, small schools that blur the line between high school and college. The primary goal of the early college model is to increase the number of students who graduate from high school and who continue on to and succeed in college. This paper presents results from a longitudinal experimental study that is examining the impact of early colleges on students' outcomes in high school and in postsecondary (PS) education, specifically on postsecondary enrollment and college credit accrual during and after high school. Early results from this study show that the Early College High School model is increasing students' enrollment in postsecondary education, primarily by the required exposure in high school. The results show that, as is inherent in its design, the program is successful in providing early access to college. This paper also reports results from a detailed cost and benefit analysis of the early college model. The following are appended: (1) References; and (2) Tables and Figures.

Background: There is substantial descriptive and correlational evidence that the COVID-19 pandemic has exacerbated achievement gaps, as historically underserved students have fallen farther behind in reading. Furthermore, contextual factors like concentrated poverty have systematically amplified the negative effects of the pandemic. Recent data from the Grade 3 NWEA MAP reading test (Lewis et al., 2021) revealed larger declines from 2019 to 2021 in reading comprehension performance in high-poverty schools than low-poverty schools. Purpose: We report findings from an effectiveness study that was focused on empowering teachers to make structured adaptations, in which teachers were provided opportunities to adapt program components, to support third-graders' reading comprehension outcomes during the COVID-19 pandemic school year when most US public school districts provided instruction remotely. Setting: The setting for this study is a large urban school district in the southeastern United States serving a diverse student population serving a high proportion of historically underserved and high poverty students (see table 1 for sample characteristics). Intervention: The randomized controlled trial aims to test the efficacy of allowing teachers to make structured adaptations to an evidence-based online science content literacy intervention, Model of Read Engagement (MORE). To prepare teachers to make these adaptations as well as increase their knowledge of the science of reading, the adaptive MORE teachers participated in a series of 4 asynchronous modules in November. The modules were based on Team-Based Learning (TBL) which aims to tightly couple knowledge acquisition with knowledge application (Michaelsen & Sweet, 2008). The knowledge application phase occurred in December 2020 as teachers applied their learnings to make structured and productive adaptations to MORE for their contexts. Teachers were given the flexibility to make both procedural-based and content-based changes to implementation, while keeping the core components of the intervention unchanged. Then in early January all teachers were provided training for the Online MORE curriculum with a focus on MORE's theory of change, the details of the lessons, and best practices based upon previous years in-person results. Classroom implementation of Online MORE curriculum took place from February to March 2021, and posttest assessments were administered in-person in April 2021 when the school district brought students back to their schools for face-to-face instruction. The Online MORE curriculum provides teachers with five weeks of science lessons including 10 synchronous lessons (2x per week) as well as 3 days per week of 30-minute asynchronous lessons that the students are expected to complete on their own. Science lessons focused on human body systems (muscular, skeletal, nervous systems). The Online version of the curriculum adapts the in-person version of the science MORE curriculum. Research Design: A total of 98 third-grade classroom teachers (N = 1,914 third-grade students in the analytic sample) within 26 elementary schools were randomly assigned to (1) a core treatment condition that replicated implementation procedures used in previous studies or a (2) core treatment with structured adaptations (Adaptive MORE). Figure 1 shows the consort flow diagram. Data Collection and Analysis: Our analysis explores three main research questions: 1. First, compared with core MORE, what is the effect of structured teacher adaptations on student and family engagement in MORE activities (online curriculum, and comprehension activities)? 2. Second, what is the effect of structured teacher adaptations on student reading comprehension scores in the Spring? 3. Third, compared with core MORE, what is the effect of structured teacher adaptations on teachers' pedagogical content knowledge and the quality of implementation of the synchronous instruction? To investigate these questions, we employed a two-level hierarchical linear model (HLM) with students at level 1, teachers at level 2, and fixed effects for school. Findings/Results: First, our results revealed no differences in student engagement with the online curriculum or mailed comprehension activities. However, on the comprehension activities that were part of the asynchronous App activities, students in the Adaptive MORE condition scored higher than CORE students on measures of the number of books completed and exposure to target vocabulary words. Second, HLMs indicated that students taught by teachers in the Adaptive MORE condition significantly outperformed students taught by teachers in the core treatment on a domain specific measure of science content knowledge (ES = 0.096, s.e. = 0.038) and science reading comprehension (ES = 0.074, se = 0.033). There were no significant effects on domain general measures of reading comprehension on the NWEA MAP (ES = 0.005, s.e. = 0.032) or the North Carolina End of Grade test (ES = -0.017, s.e. = 0.033) Third, there were no differences on a teachers' pedagogical content knowledge. However, fidelity analyses of the lessons using a quality of teacher-student interaction rating scheme applied to the Zoom recordings indicated that adaptive MORE teachers had higher quality of interactions with their students on indicators of engagement, questioning, and feedback (See Figure 2). Conclusions: The Online MORE intervention was delivered remotely during the COVID-19 pandemic school year (2020-21) and demonstrates how structured teacher adaptations may enhance both the quality of teacher and student engagement and support improvement in science content reading comprehension outcomes.

Background: Over 15% of kindergarten and first grade students in U.S. public schools are English learners (ELs) (National Center for Education Statistics [NCES], 2021). Many classroom teachers, however, have not had specialized training in providing instruction for ELs. The BELLA (Bridging English Language Learning and Academics) professional development program was created to support structured collaboration between English-as-a-Second Language (ESL) and K-1 classroom teachers as they align implementation of high-impact instructional strategies and incorporate families' cultural wealth into the classroom to benefit EL students' language development and literacy achievement. Purpose and Research Questions: The purpose of this study was to examine the efficacy of the BELLA program for improving K-1 teachers' literacy instructional practices for ELs and improving ELs' outcomes in language and literacy. The research questions included: (1) What impact does the professional development program have on kindergarten and first grade classroom teachers' use of high-impact instructional strategies, incorporation of cultural wealth in the classroom, and collaboration processes and frequency as compared to control teachers?; and (2) What impact does the professional development program have on EL students' language and literacy outcomes as compared to ELs in the control condition? Setting and Participants: This study was conducted in 13 elementary schools in two school districts in North Carolina. The percentage of students from economically disadvantaged backgrounds ranged from 26.6% to 70.6%. Forty classroom teachers (18 kindergarten and 22 first grade) along with 13 ESL teachers participated in the RCT during the 2018-2019 school year. School teams were randomly assigned to either the intervention group or the control group. One hundred eight students from participating teachers' classrooms who were eligible for ESL services and spoke Spanish at home participated in the study. Intervention The BELLA professional development program included four days of in-person workshops that focused on: (1) high-impact instructional strategies for promoting language and literacy development for young ELs; (2) a structured framework to support collaboration among classroom and ESL teachers, and (3) approaches for building on students' cultural wealth in the classroom. School teams collaborated each week to align their instructional strategies and content instruction for their ELs. The implementation coach from the research team met with each school team about once per month. Research Design and Data Collection: This study is a cluster randomized controlled trial with randomization at the school level. Teacher outcomes included classroom observations of classroom literacy instruction, observations of cultural wealth in the classroom, and teachers' reports of their collaboration process and frequency. Research staff who did not know the intervention status of the teachers observed classroom literacy instruction using the BELLA Observational Tool and a cultural wealth checklist prior to the professional development (baseline) and three additional times during the school year. The BELLA Observational Tool included 38 indictors of 8 instructional strategies that target both word level and text level skills. Inter-rater reliability was in the moderate range with Cohen's weighted kappa ranging from 0.66 to 0.88. The student outcome measure was the MAP Growth K-2 (MAP) reading assessment (Northwest Evaluation Association, 2019) administered at the beginning, middle, and end of the school year. The total score, called the RIT score, was used as the primary student outcome measure. We also examined students' scores on the four goal areas (or subtests): Foundational Skills (FS), Language and Writing (LW), Literature and Information (LIT), and Vocabulary Use and Functions (VUF). Data Analysis Study hypotheses were tested using repeated measures ANCOVA models that incorporated the following covariates: at the student level, gender; at the school level, proportion of students who identify as female, proportion of teachers with greater than five years of teaching experience, proportion of teachers with a master's degree, proportion of teachers who indicated they speak Spanish, percentage of students at the school identified as economically disadvantaged, and proportion of third-, fourth-, and fifth-grade students on grade level for reading. Focal allowed-for factors included Time (Time 0, Time 1, or Time 2), Grade (Kindergarten vs. First Grade), Condition (Control vs. Intervention), and their interactions. Our data reflected a fully nested, hierarchical design. A mixed modeling approach with restricted maximum likelihood estimation was employed using the SAS procedure MIXED (SAS/STAT software version 9.4). For each outcome, a random intercept was estimated to account for the clustering of students within schools and districts. District was included in all models as a fixed effect. Covariates were group-mean-centered, such that the observed variable score captured the deviation of the original score from the mean score of the cluster to which the sampled unit belonged. To model within-student (R-side) covariation, a series of three models was tested for each outcome, with each model specifying a different candidate covariance structure for the student-level residuals. Information criteria were used to choose the optimal structure for each outcome. After selection of a final model, fixed effect parameters were estimated and interpreted. Findings: A statistically significant Time x Condition interaction effect was observed for the overall RIT score outcome (p = 0.02) and the LIT subtest outcome (p = 0.002). For both outcomes, there was a positive and statistically significant effect of intervention at both Time = 1 and Time = 2 (but not at baseline). The Time x Condition effect was not statistically significant for the FS, LW, or VUF subtests. Conclusions: In this RCT to evaluate the efficacy of the BELLA professional development program, we found a significant impact of the intervention on students' progress in language and literacy skills. In our presentation, we will also describe the impact of the intervention on teacher outcomes. This study provides evidence of efficacy for a teacher PD program for improving instruction and accelerating student growth. Additional research is needed to test mediation hypotheses that address the extent to which students' growth in language and literacy can be explained by changes in teacher outcomes.

Background/Context: Sometimes students with the academic credentials to attend more selective colleges enroll in a less selective institution or do not enroll at all, a phenomenon known as undermatching (Bowen et al., 2009; Roderick et al., 2011; Smith et al., 2013; Rodriguez, 2015). Undermatching is associated with lower graduation rates and longer degree completion times at four-year colleges (Bowen et al., 2009). Undermatching disproportionately affects low-income students and students of color (Lowry, 2017), particularly students from these groups with high levels of achievement in high school (Cortes & Lincove, 2019). Previous research has established definitions for undermatching using high school GPA and test scores (Bowen et al., 2009; Roderick et al., 2011; Rodriguez, 2015). This study applies such a framework to the early college high school (ECHS) setting. This paper combines research about the ECHS model and postsecondary undermatching using an RCT. We find that ECHS students are less likely to undermatch at post-high school institutions than control students. These findings are robust to various model specifications and undermatch definitions. Research Questions: This paper analyzes the impact of attending an ECHS on undermatching outcomes, answering the following research questions: What is the impact of attending an ECHS on: (1) Undermatching at a student's first post-high school postsecondary institution; (2) Extent of undermatching at a student's first post-high school postsecondary institution (3) Undermatching for all post-high school semesters of enrollment; and (4) Proportion of semesters enrolled at undermatched institutions? (Completed by May 2022) To what extent do these outcomes differ for: (1) Economically disadvantaged; (2) First-generation; and (3) Students of color? Setting/Population: We use data from a longitudinal North Carolina RCT study (treatment students attended ECHS, control students attended business-as-usual high school). The analytic sample contains 1,993 students (48% economically disadvantaged, 35% students of color, 38% first-generation college) entering ninth grade in 2009 and 2010. Intervention: The ECHS merges high school and college experiences, allowing students to earn an associate degree concurrently with a high school diploma. Students receive dual enrollment opportunities and supports navigating the postsecondary environment, focused on underserved populations (first-generation, economically disadvantaged, students of color) (Edmunds et al., 2017). Analyses have shown positive impacts of ECHS on earning college credits, enrolling in postsecondary education (Edmunds et al., 2017), and earning postsecondary credentials (Edmunds et al., 2020). Research Design: To determine undermatching incidences, we compared the postsecondary institutions students attended and the highest level of institution a student could likely attend predicted by their high school performance (GPA and SAT/ACT score). Consistent with previous studies on undermatching (Bowen et al., 2009; Roderick et al., 2011; Smith et al., 2013; Rodriguez, 2015), we used Barron's selectivity codes to determine the "level" of each institution. We then used National Student Clearinghouse data to categorize the selectivity level of each institution attended by sample students. To classify students as undermatched, we compared the level of the institution of enrollment to the level of matched institution determined by the "predicted probability" taxonomy in Figure 3 (Rodriguez, 2015). Students whose institutions of attendance were at lower levels than their "matched" level or who did not enroll in college after high school were undermatched. All other students were "not" undermatched. We used a regression model for each outcome with treatment status, demographic covariates, lottery indicators, and weights determined by the probability of treatment assignment within each lottery. We estimated clustered standard errors by high school. Data Collection/Analysis: We conducted lotteries at 14 oversubscribed ECHS for the ninth-grade classes of 2009 and 2010 (26 cohorts). Treatment students were offered admission at an ECHS; control students attended a business-as-usual high school. Additional data include unweighted high school GPA and ACT/SAT scores, National Student Clearinghouse (NSC) postsecondary enrollment by semester, and college selectivity (Barron's Selectivity Index). Our total attrition of 16.5% and differential attrition of 3.5% allow us to maintain an experimental contrast. Findings: More ECHS students enrolled in selective institutions in their first post-HS semester than control students. Conversely, more control students enrolled in a two-year college than ECHS students (Figure 1). Treatment and control students "matched" to postsecondary institutions at similar levels (Figure 2). Treatment students were slightly more likely than control students to be matched to 'selective' vs. 'very selective' institutions. Our regression analysis (Table 1) determined found that ECHS students were: (1) 1 percentage points more likely to attend a 'Selective' or higher institution; (2) 0 percentage points less likely to undermatch at their first post-high school; (3) 4 percentage points less likely to severely undermatch (two or more levels below); (4) 2 percentage points less likely to 'always undermatch'; and (5) Undermatched for 9.4 percentage points fewer semesters. We sought to estimate the treatment-control contrast rather than focus on undermatching point estimates. However, as other scholars note (Bastedo & Flaster, 2014), undermatching analysis requires many assumptions. To test the robustness of our results, we ran sensitivity analyses using different estimations of college match and moving students within an error margin between two matching classifications (Figure 4) to the other classification group, summarized in Table 2. Conclusions: Early college students were less likely to undermatch at colleges they attended after high school, driven by differences in selective vs. two-year institution enrollment. These results were robust to various definitions of undermatch and sensitivity analysis at the margins of our classification model. One limitation is the method we used to categorize the "matched" institution for each student. We relied on the outputs of such models to classify students into discrete categories of high school GPA and SAT/ACT score. Future research could combine our sample with a national longitudinal data set to include a more robust predictive model of college match. However, we believe our study confronts this limitation through our sensitivity analysis and by answering questions of experimental contrast rather than point estimates. The results provide several opportunities for additional research on undermatching for students attending early college high schools. We plan the following additional analyses: (1) Exploring underlying mechanisms to explain why ECHS undermatch at lower rates; (2) Linking undermatch status with four-year degree outcomes; and (3) Using NSC data to understand transfer patterns for sample students attending two or more colleges.

The Wake County Public School System (WCPSS) in North Carolina recently addressed the issue of advancement in and equitable access to advanced mathematics. Under a recent policy, WCPSS uses a SAS-generated predicted probabilities of students' success in obtaining a passing score on the NC Algebra I End-of-Course (EOC) exam, to determine recommended course placement in grades 6 through 8. The authors examine the question: do students who are just eligible for advancement in mathematics in middle school and who participate in advanced mathematics lead to improved academic outcomes compared to students who just miss being eligible for advancement and who do not enroll in advanced mathematics? By focusing on the students on the margin of eligibility under this policy the authors use a regression-discontinuity design to estimate the causal impact of this mathematics acceleration policy on subsequent test scores, grades, and course-taking behavior of middle-school students in Wake County who were exposed to this policy. The results indicate that students on the margin of eligibility and who participate in advanced mathematics in middle school may experience negative effects as a result of being in a peer group of more advanced ability, on average. Tables and figures are appended.

In 2008, the California State Board of Education voted to require all students to enroll in algebra by 8th grade. This policy initiative, yet to be actually implemented, represents the culmination of a decades-long movement toward offering algebra instruction before the traditional high school years. Nationally, the proportion of 8th grade students enrolled in algebra doubled between 1988 and 2007 (Perie, Moran and Lutkus, 2005; Walston and McCarroll 2010), reaching rates over 50% in three states and the District of Columbia. The movement to offer algebra instruction before high school has been inspired in large part by correlational research documenting significant differences in later-life outcomes between those students who enroll in algebra by 8th grade and those who do not. But correlation need not imply causation, and it is unclear whether accelerated algebra enrollment--particularly when not accompanied by complementary curriculum reform in earlier grades--yields positive or negative effects (Loveless, 2008). This paper provides a quasi-experimental estimate of the causal impact of accelerating the introduction of algebra coursework. The research is based on explicit policy initiatives in two large North Carolina school districts in 2002/03 and it is supplemented with observations from eight additional large North Carolina Districts. Two tables and two figures are appended.

READS (Reading Enhances Achievement During Summer) has been implemented in North Carolina schools for the past several years as part of an Investing in Innovation (i3) validation grant that began in fall 2010. This paper focuses on the Year 3 study and compares the Year 3 results to the Year 1 and Year 2 results, and the results of earlier studies. The purpose of the Year 3 study was to test the effectiveness of READS-PIF (Reads with Parent Involvement and Follow-up) in a larger sample of school districts and schools. Researchers addressed three main research questions: (1) What was the overall impact of READS-PIF on students' reading comprehension?; (2) Was the impact of READS-PIF greater in high poverty schools than in moderate poverty schools?; and (3) Was the impact of READS-PIF greater for students receiving free or reduced price lunch (FRL)? This study took place in 59 elementary schools from 7 school districts in North Carolina. Of the 59 schools, 44 were high-poverty schools, and 15 were moderate-poverty schools. The final analytic sample included 5,284 students (2,798 2nd graders, and 2,486 3rd graders). Nearly 80% of the students in the study were eligible for FRL. On the spring pretest, students scored somewhat below national norms on the Iowa Tests of Basic Skills (ITBS) reading comprehension test. The sample mean normal-curve equivalent score was 46. Intervention methods included: (1) end-of-year lessons; (2) family nights; (3) tri-folds; (4) follow-up calls; and (5) book matching and summer book distribution. Within each grade and school, students were randomly assigned to the treatment or control condition and pre- and post-tested on a standardized reading comprehension test, the Iowa Tests of Basic Skills (ITBS). OLS regression was used to estimate the main effect of the treatment condition, both overall and separately by grade. To address the question of whether school poverty moderated the treatment effect, researchers included a cross-level interaction involving the student-level treatment effect and school poverty. The results suggest that the scaffolded summer reading program may be effective for third grade girls but not third grade boys. Unlike the Year 1 and Year 2 studies, the results of this study do not indicate that treatment effects are moderated by poverty measured at the school level or individual student level. Tables are appended.

When randomized control trials (RCT) are not feasible, researchers seek other methods to make causal inference, e.g., propensity score methods. One of the underlined assumptions for the propensity score methods to obtain unbiased treatment effect estimates is the ignorability assumption, that is, conditional on the propensity score, treatment assignment is independent of the outcome. The purpose of this study is to use within-study comparisons to assess how well propensity score methods can approximate experiments under various conditions. The authors test three ways of constructing comparison groups: (1) using the sample from the states that are different from the original experiments, with pretest and demographic information; (2) using the sample from the same state and districts (local matching) with the experiments, with demographic information only; and (3) same as (2) but with pretest as well. Propensity score methods (optimal matching, propensity score weighting, and stratification) are used to estimate the treatment effects for three ways of constructing comparison groups, which are compared with the benchmark from the experiment to assess estimate bias. This study will contribute to the literature by providing empirical evidence about how well propensity score methods can approximate experiments under various conditions. This study uses data from four IES funded projects, among which three are large scaled experiments: (1) "Scaling up TRIAD: Teaching Early Mathematics for Understanding with Trajectories and Technologies"; (2) "Evaluating the Effectiveness of Tennessee's Voluntary Pre-Kindergarten Program"; and (3) "Experimental Evaluation of the Tools of the Mind Pre-K Curriculum", and one is a measurement study, "Learning-Related Cognitive Self-Regulation School Readiness Measures for Preschool Children Study". The analysis is undergoing. The authors report partial results in this report. Tables and figures are appended.

Children enter school with vastly different skill levels and formal schooling often magnifies these disparities over time. Widening achievement gaps between high- and low-income children have grown substantially in the last 50 years. Further, the opportunity gap facing most low-income students contributes to a host of academic and social challenges including: lower performance in math and reading, increased truancy and incarceration, less higher-level course taking, and lower graduation and college entrance rates than their higher-income peers, and these disparities are not new. Teach For America (TFA) was founded with the purpose of addressing these educational inequities. Early on in its existence, TFA became focused on "closing the achievement gap" for students in the schools it serves, and put a large stake in promoting, "significant gains," (defined as 1.5 to 2 grade levels of improvement in core subject areas) for its students (Teach For America, 2013). For many of the students it serves, this target aimed to get students caught up to grade level standards. The research on TFA's short term impacts is mixed, although experimental and quasi-experimental evidence finds that TFA teachers are more effective than non-TFA teachers in many settings, particularly in high school math and science, and elementary math. This study extends existing research on short-term TFA impacts to examine whether TFA has any relationship with students' long-term academic outcomes. Specifically, I examine whether having a TFA teacher in elementary, middle, or high school is associated with improved student educational outcomes at the end of high school. The author examined the relationship between TFA and long-term student attainment using administrative data from the state of North Carolina. TFA places teachers in two regions in North Carolina. Eastern North Carolina, which encompasses rural and urban areas, opened in 1990 and was one of TFA's charter locations. Charlotte, which is a primarily urban region, has had corps members since 2004. Analyses for this study concentrate on the students in grades 3-12 and their teachers. The author examined the relationship between having a TFA teacher and end of high school attainment by comparing students who had TFA teachers with students who had the possibility of being in a TFA teacher's classroom but were not. To do so, the author compared students within schools, grades, and years that have TFA teachers using fixed effects. Results suggest different patterns for students who have TFA teachers in grades 3-8 versus 9-12. Among students who ever have a TFA teacher in high school, there is a negative association with dropping out, a decline of 0.012 standard deviations. There is also a positive association of having a TFA teacher and graduating high school, an increase of 0.034 standard deviations, compared with students who were in the same schools, grades, and subjects, but did not have a TFA teacher. In contrast, in elementary and middle school, there is no evidence of a relationship between having a TFA teacher and long term outcomes. Tables are appended.

Students entering high school in 9th grade face a formidable challenge. The transition to high school from 8th grade brings with it increased risks for all students. For example, students in 9th grade are anywhere from three to five times more likely to fail a class than students in any other grade. Similarly, ninth grade retention rates are higher than in any other grade. More importantly, research indicates that 70 to 80 percent of students failing in 9th grade will eventually dropout of high school. Early College High Schools (ECHS) are a new model merging high school and college designed to increase the number of students who graduate from high school and enroll and succeed in college. This paper presents results from a longitudinal, experimental study of ECHS, tracking students from 9th grade through graduation from high school and enrollment in college. This paper will examine impacts of the ECHS model on "underprepared" students, defined as those students who did not pass the 8th grade state exams in either or both reading and math. "Prepared" students, on the other hand, are those passing both exams. In this paper, the authors report on the impact of the model on the performance gaps between underprepared and prepared students from 9th-11th grades. This paper addresses the following specific research questions: (1) Are underprepared students more likely to be from traditionally underrepresented groups such as first generation, minority and free/reduced price lunch students?; (2) What is the impact of the ECHS model on these students' academic performance?; and (3) How does that impact compare to the impact on students who are prepared? This study focuses on ECHSs as implemented in North Carolina. North Carolina has the largest concentration of ECHSs in the country, with over 70 across the state; all managed by the same entity, the North Carolina New Schools. The 19 schools that are a part of this study are geographically distributed throughout the state; they include schools in rural and urban areas, schools with predominantly white populations and schools with predominantly African American populations. Results from this experimental study show that it is possible for schools to have an impact on students entering high school at-risk for academic failure and dropout. These data suggest to us that the reduction or elimination of performance gaps in the early college is a product of a purposeful implementation of a high quality learning environment with high expectations, rigorous courses and instruction, positive relationships, extensive student support, and teachers taking responsibility for student learning. Tables and figures are appended.

Research has shown that "last hired, first fired" policies maximize the number of teachers subject to reductions in force by eliminating those teachers that are lowest on the pay scale first. Until now, advocates of effectiveness-based reduction-in-force (RIF) policies could only point to simulated policy exercises as evidence of the potential benefits of a discretionary reduction-in-force policy. This study suggests that, while reductions in force negatively affect student achievement, districts have the potential to reduce these negative effects by concentrating layoffs among the lowest-performing teachers. The analyses of this study focuses on three main objectives: (1) comparing the relative weight that North Carolina's Charlotte-Mecklenburg Schools' (CMS) administrators and principals placed on a variety of RIF criteria when implementing layoffs; (2) estimating the causal effect of discretionary layoffs in CMS on student achievement in the following academic year; and (3) documenting how these negative effects are exacerbated when highly effective teachers are laid-off teachers. Results from descriptive regression analyses suggest that CMS used multiple RIF criteria when selecting teachers for layoffs including tenure status, licensure status and type, and job performance. Although replacing seniority with performance measures can minimize effects on student achievement, exchanging one inflexible criterion for another will not provide districts with any discretion in navigating a complex process aimed at preventing a variety of negative consequences. Tables are appended.

As highlighted by the National Math Panel Report (2008), consistent results from laboratory studies have demonstrated that interleaving worked examples with problems to solve improves learning for novices. The purpose of this work is to create materials and tests that can be used flexibly in classrooms and which employ worked examples interleaved with practice problems for students to solve. A total of 42 assignments addressing some pre-Algebra and 8 major Algebra topics have been developed through an iterative process. The effects of AlgebraByExample on learning has also been explored with particular attention to the role of individual differences. The eight participating districts are all members of the Minority Student Achievement Network (MSAN) which is a self-formed alliance dedicated to addressing the achievement gap: Ann Arbor (MI), Arlington (VA), Chapel Hill--Carrboro City (NC), Evanston Township (IL), Evanston/Skokie 65 (IL), Green Bay (WI), Shaker Heights (OH), and Madison Metropolitan School District (WI). Although this paper does not cover all of the individual studies, the body of work as a whole has been conducted in 300 classrooms with over 6000 7th-10th graders, with the majority of 8th or 9th graders enrolled in "regular" Algebra. Approximately half of the students in each participating class were randomly assigned to the example-based group, in which they received the example-based assignments designed for the study. The other half was randomly assigned to the control group, in which they received an alternate version of the assignments that contained the same types of problems, but no examples or self-explanation prompts. Results demonstrate that students in high-minority populations benefit more from example-based assignments than do those in lower-minority groups in terms of an increase in their procedural performance. Results also indicate that conceptual learning occurs more readily for students in low-minority classrooms -- students in higher-minority classrooms do not tend to learn as much. However, for students in high-minority classrooms, being in the example-based condition leads to greater learning than the control condition. One table and two figures are appended.

The problem of summer reading loss among low-income children has been amply documented by researchers. Over time, income-based disparities in reading achievement appear to grow more rapidly during the summer months than the academic school year. This paper describes two experiments in which teachers scaffold summer book reading with two different approaches to improving comprehension through end of school year lessons. Because teachers implemented different comprehension lessons in the two studies, the instructional activities varied across the two studies. Both studies used the same norm-referenced reading comprehension test (Iowa Test of Basic Skills), enabling a comparison of impacts on a similar measure across two studies. The purpose of this study is to capitalize on the lessons from the two experiments. Results from the 2006 experiment, which focused on comprehension strategy instruction, improved children's performance on comprehension tasks that required inference and interpretation (d = 0.15). Results from the 2011 experiment suggest that content-oriented comprehension instruction improved children's performance on tasks that required analysis and generalization (d = 0.11) compared to strategy-oriented instruction. In study 1, strategy lessons appeared to benefit Grade 4 children's ability to make inferences and interpretation. In study 2, content lessons appeared to benefit Grade 3 children's ability to make analyses and generalizations. Tables are appended.

The aim of the "Tools of the Mind" prekindergarten curriculum is to enhance children's executive function skills within an instructional context that promotes the basic academic and social skills that prepare them for kindergarten and beyond. To investigate the effectiveness of "Tools" in achieving this aim, the authors are conducting a longitudinal randomized experiment to answer the following questions: (1) Do children in "Tools of the Mind" classrooms improve more in literacy, math, social skills, and behavior problems during the preschool year than children in "business as usual" control classrooms? Are those gains sustained through kindergarten and first grade?; (2) Do children in "Tools of the Mind" classrooms show greater gains in executive function than children in the control classrooms? Do those gains mediate the curriculum effects on literacy, math and social skills outcomes?; and (3) Are there differential effects of "Tools of the Mind" associated with characteristics of the children or the classrooms? Four school districts in Tennessee and two in North Carolina participated in the study. The authors' results show that there were no significant treatment effects on any of their outcome variables. Students in "Tools" classrooms performed about equally well on all outcome variables, including the executive function measures, after receiving a year of the curriculum as students who received the usual preschool curriculum. Similarly there were no differences between the two sets of classrooms in teacher ratings of social and behavioral competence. In addition, the "Tools" curriculum did not appear to result in significantly better outcomes for any student subgroups (i.e., ELL, ethnic groups, gender) when compared to the control condition. Gains on all outcomes were observed across the preschool year in both "Tools" and comparison classrooms. (Contains 2 tables.)

This study aims to determine whether educational accountability promotes educational triage. This study exploits a natural experiment in North Carolina in which standards increased first in math in 2006 and second in reading in 2008 to determine whether an increase in educational standards caused an increase in educational triage at the expense of low and high achievers and to the benefit of students near grade level. The authors hypothesize that changing the rigor of state academic standards will have heterogeneous effects on student achievement. Triage theory predicts that under a status based system such as NCLB, students near grade level in schools that fail AYP will have larger increases than students well below or well above grade level. They hypothesize that the disparities between average achievers and low achievers and average achievers and high achievers will widen when the standards for defining grade level increase. In other words, triage will become worse when the tests become more difficult and the proficiency cut score increases. Results indicate that in math there is strong evidence that the increase in academic standards benefitted students near grade level more than low or high achieving students. In reading, the evidence is somewhat weaker, but it is clear that high achieving students benefitted less than students near or below grade level. In the complete version of this paper the authors plan to examine whether these effects vary by the number below-grade-level students a school serves. In addition, they plan to present predicted values for low, average, and high achievers in both the pre and post period to more clearly assess the differential effects of increasing standards. (Contains 2 figures, 2 tables and 2 footnotes.)

The TALPS study aims to build on the existing research base to develop promising methodologies to identify chronically low-performing and turnaround schools, as well as to identify promising strategies for turning around chronically low-performing schools. By looking specifically at schools identified as turnaround, in comparison to nonturnaround schools, this study may help discern what lessons from prior research apply to turnaround efforts and what unique solutions--and challenges--are associated with turnaround. The TALPS study was composed of Study I and Study II. Study I focused on identifying turnaround schools, as distinct from moderately improving (MI) and not improving (NI) schools. Study I used student-level data from Florida, North Carolina, and Texas, spanning Grades 3 to 8 for the six school years between 2002-03 and 2007-08. The authors focused on student scores on state accountability assessments in reading/language arts and mathematics. They examined both achievement level (at a given grade level) and growth (progress students make as they move from one grade to the next). The Study I identified 1,042 schools as CLP, including TA, MI, and NI schools, which were the population of Study II sample. Study II sought to examine whether particular policies, programs, and practices (PPPs) occurred more often in turnaround (TA) schools than in nonturnaround (moderately improving and not improving) schools. It appears from this study that (1) accountability pressures and support from the district combined with (2) strong instructional leadership, (3) strategic staffing (i.e., strategic recruitment, assignment, and "counseling out" of ineffective staff), (4) intensive professional development, and (5) data use focused on identifying and assisting struggling students are key components of a school's turnaround process. How these components work together and are implemented should be explored further in future real-time, qualitative research efforts. Appended are: (1) References; and (2) Tables and Figures. (Contains 3 exhibits.)

Teachers are generally recognized as the schooling factor accounting for the highest proportion of student learning outcomes (Aaronson et al., 2007; Hanushek, 1986). This implies the quick and dramatic improvement in school performance observed in turnaround (TA) schools was associated with a major change in the performance of its teachers. This change could be manifest in one of two ways--either the teachers in the school dramatically improve or previously ineffective teachers are replaced with effective teachers. This paper seeks to understand how each of these potential changes in the school's human capital stock contributed to turnaround in Florida and North Carolina. Specifically, the study investigates the following research question: is the improved achievement observed in successful turnaround schools most strongly associated with workforce "development" or "turnover" (or, some combination of the two)? The results suggest TA schools improved through a combination of improving the stable teachers who had been in the school even when it was performing poorly and bringing in relatively effective talent into the school. Recommendations and limitations based on findings are also presented. Appended are: (1) References; and (2) Tables and Figures. (Contains 2 tables.)

Launched in 2008, the Strategic Data Project, housed at the Center for Education Policy Research at Harvard University, seeks to bridge the divide between educational research and practice in order to transform the use of data in education to improve student achievement. Through the project, the authors build robust research partnerships with educational agencies--currently including 32 school districts, state education agencies, charter management and other non-profit organizations--in order to grow the analytic capacity of the agencies while also fostering more strategic and informed use of data for informing policy and practice. The proposed panel on the Strategic Data Project will have two primary aims. First, the authors will share background on the genesis of the Strategic Data Project, provide an overview of its key components and highlight how these components contribute to the effectiveness of each partnership overall. After providing a broad overview of the Strategic Data Project, including a summary of selected diagnostic analyses across several partner agencies, the authors will turn to the second aim of the proposed panel: to highlight the work of data fellows placed within two of their partner agencies.

There is some evidence to indicate that chronically low-performing schools, whether improving student performance or not, often report pursuing substantially similar policies, programs, and practices. However, while chronically low-performing schools may pursue similar school improvement strategies, there is some evidence that the level and quality of implementation, as well as the coherence, alignment, and persistence of implementation, may lead to different prospects for school turnaround. This study investigated how school improvement efforts were implemented in turnaround (TA) and non-improving (NI) schools, with a focus on the external and internal conditions that were perceived to support these efforts. Findings showed that differences between the TA and NI case study schools were apparent, although the study's necessary reliance on individuals' memories and on a small convenience sample dictate caution in interpreting these findings. With respect to central improvement efforts, respondents in TA schools compared with those in NI schools more frequently cited data use (chiefly for pinpointing individual students' progress and needs), targeted student supports (during or beyond the school day), and the use of common planning time for collaboration. Conversely, more NI case study schools identified adopting new curricula or instructional approaches as central to their work during the study period.

The criteria for determining the student outcomes that define a school as having "turned around" are not well defined, and the definition of turnaround performance varies across studies. Although current policy initiatives offer guidelines for identifying CLP schools, there is no standard definition or methodology in common usage. This paper summarizes the lessons learned from this exercise of empirically identifying CLP schools and binning them into performance categories based on their trajectories. This paper provides guidance for others charged with a similar task. Specifically, the authors learned the critical importance of using student-level data (rather than school-level aggregate measures), using growth-based measures in conjunction with status-based performance metrics, the stability of these performance metrics over time, and how to empirically recognize turnaround in schools as it occurs. Also, the authors learned that low-performing schools turned around their performance more frequently than one might have presumed based on prior research. In Florida, they identified approximately 15% of chronically low-performing elementary and 14% of chronically low-performing middle schools as turnarounds. Similar rates were observed in North Carolina--13% and 16%, respectively; and even higher in Texas--29% and 31%, respectively. (Contains 3 figures and 2 tables.)

In recent decades, much attention has been focused on student achievement in the United States. Many policy initiatives have been attempted in an effort to bolster achievement, including increasing school revenue, decreasing class size, expanding early childhood programs, and introducing vouchers and charter schools, to name a few, but not all of these initiatives have had the desired impact. Research has shown that other, less tangible factors such as teacher quality and the characteristics of a student's peers may play a much greater role. This research focuses on one of these less tangible factors: principal quality. This paper has two main goals. The first is to measure principal quality by estimating principal value added to student achievement. The authors estimate a set of principal fixed effects from test score gains with data on North Carolina students between third and eighth grade. The second goal is use the estimated principal value added to determine what makes a principal better than another at raising test score value added. (Contains 1 footnote.)

Over a decade ago, Bredekamp asserted (1996) that the field of early childhood education lacked a research base to document the effectiveness of approaches to producing highly skilled and effective early educators. The National Center for Research in Early Childhood Education's (NCRECE) Course and MyTeachingPartner (MTP) Coaching were developed to fill the gap in effective PD targeted toward improving teacher-child interactions. In both cases, this PD focused on improving the interactions known to support children's development, as identified by the Classroom Assessment Scoring System (CLASS; Pianta et al., 2008). The NCRECE study was intentionally designed to use two distinct professional development approaches, targeted toward the same outcome (improved teacher-student interactions), to examine their unique individual or combined influence on enhancing interactions. As part of the randomized controlled trial, intent to treat results suggest that both the NCRECE Course and the MTP Coaching do change teachers' interactions in ways that support children's development (Hamre et al, in press; Pianta et al., 2011). This paper explores variation in teachers' responsiveness during both the Course and MTP Coaching phases of the NCRECE study. Of particular interest is the identification of classroom and teacher factors that may serve as moderators between participants' responsiveness and change in their instructional practice. A total of 433 teachers were recruited into the course phase of the study. Upon consent, they were randomized at the site location level into the course or control group for the first phase of the study so that approximately half of the teachers participated in each group. A total of 405 teachers were recruited into the consultancy phase of the study. Of teachers in the analytic sample, 41% worked in Head Start programs and a significant portion worked in public schools (37%). Teachers were experienced, with an average of 14 years of experience teaching and an average of 16 years of education. Most of the teachers were African American (47%) or White (31%) with a smaller number of Latino (15%), Asian American (4%), and multi-ethnic (4%). The poverty rate among children in the classrooms taught by teachers in the sample was quite high (mean = 0.88, sd = 0.21). The NCRECE Course and MTP coaching intent to treat findings reported elsewhere indicate there are different types of intervention designs that can improve teachers' practice in ways that matter for children's development (Pianta et al., 2011). The current treatment on the treated study takes this work a step further, looking inside main effects of interventions to begin to understand the mechanisms through which change can occur, and for whom. In this case, participants' personal and classroom characteristics as well as their responsiveness to two different types of intervention were considered, and shown to matter in some ways the same, and in some ways different, across the interventions. Consistent with previous findings (Downer et al, 2009), and across both types of interventions, teachers who have less experience in prek classrooms benefit most from the intervention experiences. This is consistent with other literature on novice teachers, pointing to both their continued need and desire to develop (Darling-Hammond, 2001). Given the persistent concern regarding new teacher quality and retention, tending to the explicit needs of new teachers with interventions known to support effective practice should be a priority. Although there were several teacher/classroom characteristics that related to change in one intervention phase or the other, the main difference of note is how participants' responsiveness in the MTP Coaching intervention, but not in the NCRECE Course, was associated with improved instructional interactions. Given the personal and interactive nature of coaching, responsiveness by the participant may be more critical to change in their practice then attending a course (Pianta et al., 2011). Further, participant responsiveness to MTP and change in practice was moderated by teacher experience and composition of children in the classroom. This points to a need for professional development opportunities that align with the same intended outputs, but use different learning formats or modalities that work best for the individual learner involved. Based on these preliminary findings, future work would benefit from examining a more comprehensive look at intervention responsiveness, as well as unexplored teacher/classroom characteristics.

In the current study characteristics of the teacher and classroom that influence teacher responsiveness to the course or consultancy are explored. If classroom quality or characteristics of teachers, such as anxiety levels or efficacy in teaching prior to intervention participation are predictive of responsiveness, then future programs will be able to provide additional supports and resources for these individuals. Further, it may be that the influences of these characteristics differ depending on the intervention setting. For example, the course may provide a safe haven for teachers who are shy and nervous, while the consultancy, which is an intensive, one-on-one relationship, may increase teacher anxiety levels and limit participation. On the other hand, teachers who have more child-centered beliefs about development and classroom interactions may be more responsive to the course and consultancy. Two hundred twenty-three teachers were randomized into the course, however only 167 teachers had at least partial data. Two hundred eight teachers were randomized into the consultancy, and 193 had at least some data and were included in analyses. Teachers were diverse in their educational attainment, experience, and in their racial/ethnic background. Overall, higher authoritarian beliefs about children limited teacher responsiveness to the course while anxiety prohibited teachers from fully benefiting from the consultancy. Other teacher characteristics (e.g., education, ethnicity) also impacted responsiveness to both intervention settings. Further, participation in the course and consultancy was differentially impacted for teachers' with literacy-rich classrooms. These findings have implications for promoting participation in varying intervention settings and shed light on identifying individuals for which more resources for participation may be allocated. (Contains 4 tables.)

MyTeachingPartner (MTP) is an interactive, web-based professional development format created at the Center for Advanced Studies in Teaching and Learning (CASTL) at the University of Virginia (Hadden & Pianta, 2006). The MTP model is based on the understanding that effective teacher professional development requires opportunities for teachers to watch high quality teaching and receive regular feedback on their own teaching practice (Pianta, Mashburn, et al., 2008). The MTP model includes many of the above identified sources of self-efficacy, and its components map on to Bandura's sources of self-efficacy. For example, the consultancy allows teachers to practice teaching in a supportive environment, moving them to greater levels of mastery. They collaboratively evaluate their teaching practice with the feedback from a qualified consultant, and gain a greater sense of control over their own outcomes, therefore reducing negative arousal, and increasing positive emotional cues. The video exemplars of high quality teaching, to which consultants often refer teachers, can serve as efficacy-enhancing vicarious experiences. In the context of an intervention study aimed at improving teacher-child interactions through the use of the MTP consultancy, the authors positioned the present study as an attempt to test the efficacy of MTP as a professional development approach for targeting teacher self-efficacy. Specifically, this study examines the extent to which treatment assignment relates to teacher self-efficacy in the domains of classroom management, student engagement, and instructional strategies. (Contains 3 tables.)

The overall purpose of this IES-funded Goal 3 study, now entering its third year, is to evaluate the efficacy of Incredible Years Teacher program (IYT) for improving K-2 students' academic achievement, attention, and social-emotional competence. The authors will also examine the extent to which classroom behavior mediates the relationship between teacher management skills and students' academic performance in the classroom. The current poster, based upon teacher data from the first two cohorts, will present preliminary data on whether participation in IYT is associated with change over time in the classroom environment and quality of teacher interactions with students for trained teachers as compared to wait-list control teachers. Data include the CLASS (Pianta & Hamre, 2005) and the Teacher coder Impressions Inventory (TCI; Webster-Stratton, Reid, & Hammond, 2001), which are based on blind observer ratings. The CLASS is a multi-dimensional standardized instrument that has been widely used in early education classrooms and has been associated with gains in students' achievement and social adjustment. The TCI was developed to align more closely with the intervention and has been used primarily in research on the efficacy of IYT. (Contains 2 tables and 3 figures.)

The Recognition & Response (R&R) model was developed and is being validated by a research team at the FPG Child Development Institute at the University of North Carolina at Chapel Hill. R&R has generated widespread attention in the early childhood field as a promising RTI model for pre-k (see entire issue of NHSA Dialog, Volume 12[3], 2009) and as one of the few to include a focus on improving children's academic learning (Greenwood et al., 2011). The R&R system is a framework for linking formative assessment to tiered instruction, designed for use with children who have a wide array of learning characteristics across a number of content areas. R&R has a dual focus on improving the quality of instructional practices for all children as well as providing additional supports for some children to ensure that every child succeeds in school. The Development and Innovation (IES Goal 2) study reported in this paper builds on an existing line of research that is focused on further development and evaluation of the R&R model, including adaptations for mathematics instruction, the integration of learning and behavioral interventions/supports, and instructional practices to support dual language learners, in addition to the current emphasis on language and literacy instruction. The current study replicates an earlier small-scale quasi-experimental study that used a different sample of children and programs. The poster focuses on the recently completed Development and Innovation study, but presents findings from both studies. The purpose of the Development and Innovation study was to design and evaluate the R&R model for use by pre-k teachers to address children's language/literacy skills. The study evaluated the feasibility, acceptability, implementation fidelity, and potential for improved classroom and child outcomes of this approach. A total of 26 pre-kindergarten classrooms were included in the main study (Study 2). Results suggest that the R&R system offers evidence of promise for improving language and literacy outcomes for young children. Positive effects were found in the growth rates for target children compared to their peers on formative assessment and standardized measures. Although target children had substantially lower scores initially as well as following the intervention, their rates of growth were greater than or the same as comparison children. These results indicate that through the assessment and intervention components, teachers who used the R&R model were able to successfully determine target children for the interventions (i.e., those with significantly lower skill levels than their peers), and potentially to alter their developmental trajectory so that they began catching up to their peers in some areas and maintained pace in others. Moreover, positive effects were found across different populations of children; although Study 2 included a relatively more disadvantaged population, as evidenced by their background characteristics and fall scores, the model had positive effects for target children in both studies. Not surprisingly, stronger effects were found in critical areas of receptive and expressive language skills when R&R was implemented under the more ideal conditions present in Study 2 (i.e., a full year rather than one semester, more highly educated teachers). Although these studies did not provide the opportunity for a true control group (given the requirements for Goal 2 Development and Innovation studies), the results clearly provide empirical evidence of the promise of R&R as an educational intervention for pre-k, and suggest that further research of the efficacy of this model is warranted. (Contains 1 figure and 1 table.)

The present paper describes the recruitment and site selection process that has been underway since January 2011, with particular emphasis on the use of Mahalanobis distance score to determine matched pairs of sites prior to randomization to treatment and control groups. Through a systematic winnowing process, the authors found that they could reduce the number of potential matching variables from more than 100 to just 11 matching variables. By computing Mahalanobis Distance scores for each eligible school, and discussing the results as a team, they were able to identify a set of matched pairs that, when randomly assigned to their two treatment conditions, produced two sets of schools that were well matched on the relevant variables, i.e., represented a level playing field for the RCT in North Carolina. They will use the same process to match and randomly assign schools in their other two regions. They hope that other researchers engaged in similar work will find this account useful. (Contains 5 tables.)

This study examines the incentive effects of North Carolina's practice of awarding performance bonuses on test score achievement on the state tests. Bonuses were awarded based solely on whether a school exceeds a threshold on a continuous performance metric. The study uses a sharp regression discontinuity design, an approach with strong internal validity around the cutoff of the treatment assignment score, to examine three questions: (1) Do bonuses induce incentive effects to increase math or reading test score gains?; (2) Do bonuses promote "educational triage" based on the achievement level of the student?; and (3) Do bonuses promote a narrowing of the curriculum at the expense of science? The study is set in North Carolina public schools elementary schools (statewide) in the spring of 2008. The study finds evidence consistent with the hypothesis that educators in North Carolina respond to incentives to increase test score gains in reading and math. Those students in schools that just missed the bonus threshold in 2007 have higher test score gains in 2008. This suggests that educators expend additional effort and may implement new practices in response to the failure to receive a bonus. The author finds suggestive, but not conclusive, evidence that math gains are primarily driven by low and average achieving students. Contrary to expectations, reading gains are disproportionately driven by students with the highest within-school achievement. This suggests that either schools targeted high achieving students with reading interventions, which is unlikely, or that schools used whole-school interventions that had positive effects on high achievers and no effects on low achievers. This finding deserves future research into its generalizability across different time periods and investigation of the mechanisms through which this differential effect was produced. The author finds no evidence of a narrowing of the curriculum at the expense of science. This is in contradiction to theory and prior research on a "narrowing of the curriculum" at the expense of low-stakes and non-tested subjects. The fact that the policy is focused on test score gains, rather than levels, however, raises questions about whether incentive effects on test score levels should be expected. That North Carolina's bonus policy had no effect on test score levels may be viewed as a shortcoming of the policy if absolute, rather than relative, levels of performance are also of interest. (Contains 9 figures and 1 table.)

To address concerns about low graduation rates and a lack of workforce readiness, states and districts have been experimenting with different models of high school reform. One of the most popular models has been the Early College High School (ECHS) model, small schools that blur the line between high school and college. Since 2002, over 200 ECHSs have been created under the auspices of the Early College High School Initiative, which is primarily funded by the Bill & Melinda Gates Foundation. North Carolina has the largest concentration, with over 70 ECHSs across the state. The purpose of this study is to rigorously examine the implementation and impact of the Early College High School model in North Carolina. This study is the first to utilize a longitudinal experimental design to assess the impact of the ECHS model on student outcomes. The study has three main goals: (1) Determine the impact of the model on selected student outcomes; (2) Determine the extent to which impacts differ by student characteristics; and (3) Examine the implementation of the model by site and the extent to which variation in specific model components is associated with impacts. Nineteen ECHS in North Carolina are participating in the study. Early results from this study show that the Early College High School model is having a positive impact on many outcomes associated with remaining in school and becoming ready for college. These results suggest that the ECHS is making substantial progress towards its goal of graduating more students who are ready for college and work. Although these findings are very positive, they are restricted to the model as implemented in North Carolina. Other early colleges around the country may not follow the same design principles and may not receive the same level of assistance in implementation as the schools in North Carolina; as a result, their results may differ. (Contains 1 figure and 4 tables.)

The National Center for Research on Early Childhood Education's (NCRECE) program of research is a series of experimental studies of specific approaches to training early childhood (EC) educators to be effective in implementation of curriculum and instructional interactions focused on promoting language and literacy skills, two domains that operate as gatekeepers to later achievement. Previous NCRECE work has demonstrated that a 14-week course designed to enhance teachers' use of effective teaching practices was effective in changing teacher beliefs, knowledge, and observed practice (Hamre et al., 2010). This paper examines the extent to which teachers from a variety of backgrounds were engaged in the course and the extent to which their engagement was associated with significant changes in belief, knowledge, and practice. The authors briefly discuss the context for this work. First, the authors examine the extent to which there was variation in implementation and engagement across the 15 course sections in 9 sites. Second, they examine the extent to which teachers from a variety of backgrounds engaged equally in the course, both in terms of quantity and quality of engagement. Finally, they examine which aspects of participation are most closely associated with changes in teachers' belief, knowledge, and practice. The NCRECE course was offered in nine sites across the country: Charlotte, North Carolina; Chicago, Illinois, Columbus, Ohio; Dayton, Ohio; Hartford, Connecticut; Memphis, Tennessee; New York City; Rhode Island; and Stockton, California. Participating teachers worked in a variety of early childhood (EC) programs including Head Start, preschool, and child care. Results suggest several teacher-level factors that predict engagement. Teachers with less authoritarian beliefs and teachers with less pre-k teaching experience were more likely to be engaged in course material. It may be that the course content, which focused on the importance of teacher-child interactions and also highlighted the importance of fostering children's autonomy, was in conflict with some teachers beliefs in ways that led them to become more disengaged with course material. Teachers with more teaching experience reported the course was somewhat less useful, although they still gave generally high ratings. Interestingly, the quality of experienced teachers' engagement was somewhat lower--they performed more poorly on homework and midterm assignments. This is somewhat counterintuitive as everyone might expect these teachers to know the most about classroom interactions. However, it may be that teachers who have been in the classroom for a long time have more set ideas about what effective interactions look like, and thus have a harder time performing on tasks that ask them to shift to thinking about new ways to interact in the classroom. Teachers from a wide variety of educational backgrounds reported that the course was useful and displayed equal level of engagement. It is unusual for a course to target such a broad range of teachers but this finding suggests that the course material, focused on teacher-child interactions, was relevant to teachers across the educational spectrum. This may reflect the fact that few teachers, even those who have taken many early childhood courses, have had opportunities to learn about effective teacher-child interactions in a college course setting. Adding to efficacy evidence from the intent-to-treat analysis, this study suggests that teachers who attended more course sections and completed more of the homework assignments displayed greater changes from pre- to post-test on three of the four measures of beliefs and knowledge. Thus, dosage appears to matter. There was not evidence to support the idea that teachers' own views of usefulness, instructor-reported engagement, or the quality of teachers engagement (i.e. homework and midterm scores) were systematically associated with changes in belief and knowledge. Future analyses may examine the individual components of participation and quality of engagement in more detail to try to unpack the "active ingredients" of this course. (Contains 4 tables.)

"The No Child Left Behind Act of 2001" (PL 107-110) established the Reading First Program (Title I, Part B, Subpart 1), a major federal initiative designed to help ensure that all children can read at or above grade level by the end of third grade. Reading First (RF) builds upon research indicating that effective reading instruction on specific early reading skills, particularly in the primary grades, significantly reduces the number of students who experience reading difficulties in later years. This study was designed to answer two broad sets of questions, each focusing on a different aspect of sustainability of the program. These policy questions are critical to the ongoing conversations about the directions federal education policy should take to close achievement gaps and promote high achievement. The first set of questions is the following: "What are states' planned responses to the Reading First budget reduction?" and "Which RF program elements do state-level staff believe can be sustained beyond Reading First?" These questions address the planned and likely sustainability of the program from the perspective of state-level program administrators. The second set of questions is the following: "Does student achievement in later grades differ for students exposed to Reading First in K-3 compared with students in Title I Schoolwide Program (SWP) schools?" and "Is there evidence that the effect of RF spilled over to students in non-RF Title I SWP schools in RF districts?" These questions address the sustainability of program effects on students by examining patterns of student performance after students have completed the targeted grades. Findings are embargoed pending release of report results by the U.S. Department of Education (anticipated early 2011). (Contains 1 footnote.)

The purpose of this paper is to examine the potential for salary-based policies to generate a more equitable distribution of teachers characterized by strong qualifications across schools. The analysis is based on rich longitudinal administrative data on teachers and schools in the state of North Carolina. The authors' analysis first confirms the findings in the extent literature that teachers are reluctant to remain in schools with high proportions of nonwhite students, even after they control for the proportion of students eligible for free lunch. Second, they find that in making their decisions to leave a school, teachers are quite responsive to salary differentials. Taken together, these results suggest that salary differentials are potentially useful as a tool for retaining teachers in the schools serving disadvantaged students. Importantly, however, the authors find that teachers with the stronger qualifications are both more responsive to the racial characteristics of the students and less responsive to salary differentials than are their counterparts with weaker qualifications. As a result the salary differentials required to retain the teachers with strong qualifications are far larger than those for weaker teachers.