As AI becomes increasingly integrated into K–12 education, states are moving beyond high-level principles toward curriculum- and grade-level implementation. Framed by the Multiple Streams Framework, this study compares North Carolina and California to examine how state AI policies translate governance priorities into instructional guidance. Using qualitative comparative policy analysis, in vivo coding, and thematic analysis, the study identifies key policy priorities and themes. The analysis suggests that both states present structured approaches to AI guidance but differ in emphasis. North Carolina frames AI guidance through developmental recommendations across PK–13 levels, emphasizing teacher capacity, digital citizenship, and instructional integration. In contrast, California focuses ethics, equity, and responsible AI use while aligning guidance with existing curriculum standards. The comparison illustrates distinct ways the two states organize and prioritize AI guidance within their policy documents rather than differences in policy implementation or effectiveness.
This study examines the implementation of Assembly Bill 1460, California State University’s Ethnic Studies graduation requirement. The study draws on interviews with 10 Ethnic Studies leaders across three institutional types to explore how race, governance structures, and leadership dynamics shaped policy enactment. Findings reveal that leaders’ agency was both constrained by racialized bureaucratic norms and enhanced through supportive formal leadership, strategic hiring, and distributed decision-making. The study contributes to scholarship on institutionalizing anti-racist curriculum by illuminating how racially situated leaders navigate structural inequities to advance sustainable organizational change.
This study investigates whether rigorous coursework benefits all students or mainly those with strong preparation and motivation. Using nationally representative HSLS:09 data, we find that coursework intensity is associated with greater college enrollment, selective college attendance, and STEM major selection. These gains extend to students with lower academic performance, lower motivation, and disabilities. Low-SES students experience larger college attendance gains, while high-SES students benefit more in selective admissions and STEM fields. Gender differences show higher selective college enrollment for females and greater STEM participation for males. Racial disparities persist, but coursework intensity is positively linked to postsecondary success across groups.
This study explores U.S. congressional proposals on artificial intelligence in education (AIED) from 2016 through early 2026, treating them as an underexamined site of federal policy design. The study analyzes AIED bills across the P-16 continuum to examine how Congress frames AI education through capacity building, coordination, and workforce-oriented logics rather than instructional governance. Drawing on a policy design framework, it identifies four legislative clusters and demonstrates how agenda-setting delineates the institutional boundaries of federal AIED governance. The findings suggest that federal authority is shifting from education policy venues toward science and innovation domains amid technological uncertainty and educational federalism.
Multiyear teacher–student assignments, often known as “looping”, extend relationships among teachers, students, and families beyond a single school year. Using administrative data from a large urban district spanning more than a decade, we develop an empirical approach to identifying intentional looping and estimate its associations with student achievement in grades 4–8, distinguishing looping from incidental repeat teacher–student matches and separating intended from realized continuity. Within a fixed-effects framework that accounts for selection into and out of looping classes, results show that looping is uncommon but associated with gains of roughly 0.04–0.05 standard deviations in ELA and 0.025–0.04 in math—meaningfully larger than those associated with incidental repeat matches. Estimated associations are larger in middle-grade ELA and elementary math and remain robust to alternative model specifications and definitions.
New Zealand (NZ) schools are mandated by the Education and Training Act (2020) to ensure students’ safety, but are not required to have a bullying prevention policy. This study analysed 745 NZ school websites to determine whether bullying policies were available online and applied an accessibility framework to assess ease of access. Only 388 schools (52%) had an accessible policy. Of these, a substantial proportion required multiple navigation steps (e.g. ≥3 clicks from the homepage), were not visible through standard menu structures, keyword searches (e.g. “bullying” or “policy”) did not link to relevant pages, or were embedded within password-protected platforms, presenting a level of access difficulty. Furthermore, 86% relied on a modular template supplied by SchoolDocs, a private subscription service, in which policy content is divided across multiple linked documents (e.g. bullying, cyberbullying, and complaints procedures). As a result, relevant information is not presented in a single location, requiring parents to navigate between separate sections to understand the full process. Policies were evaluated using a 42-item framework developed by Kidwai and Smith to assess policy content across four areas: definitions, reporting procedures, communication, and prevention. On average, the 388 policies scored 20/42 items, with scores ranging from 1 to 35, reflecting partial coverage. While (95%) policies defined bullying and acknowledged governance requirements, major gaps included prevention strategies, recognition of identity-based bullying, parent notification, and systems for recording and reporting incidents. Despite a national commitment to student safety, policies provided limited actionable guidance for staff. While most policies advised students or parents to report bullying, such as telling a teacher (95%), there was an absence of guidance on how staff should respond, what steps to follow, or how incidents would be managed. These findings highlight the need for stronger national guidance on bullying prevention, alongside clearer expectations for policy content, design, and online accessibility for all users.
The rapid adoption of artificial intelligence (AI) in K-12 education has outpaced the development of coherent policy frameworks, raising governance challenges for policymakers. This study examines how U.S. states conceptualize AI governance through qualitative content analysis of policy guidance documents from 40 states and territorial jurisdictions issued between 2023 and 2026. Findings indicate that ethics and data privacy dominate current policy approaches, while equity considerations and educator capacity-building receive moderate attention, and stakeholder trust and ongoing monitoring remain underdeveloped. The study identifies policy gaps and implications for developing more comprehensive and responsible AI governance in K-12 education.
This study examines how students experience out-of-school suspensions using parent-reported data from a national survey of parents whose children were suspended ( n = 117). Most parents reported that their suspended child(ren) did not receive learning materials or opportunities, many lacked adult supervision, and nearly one-third got into additional trouble while suspended. Differences by student subgroups emerged, where students from lower-income households were less likely to receive academic supports, while non-White parents more often reported lack of supervision and further trouble during the suspension. Further, lack of academic supports and supervision strongly predicted additional trouble, highlighting how suspensions may exacerbate disengagement and inequities.
This content analysis examines messages P–12 school leaders received from practitioner magazines during the second year after ChatGPT’s public release. We collected 107 articles from 12 popular U.S. practitioner media outlets published in the second year after ChatGPT’s release, focusing on the intersection of AI and school leadership. Five themes spanned policy development, teacher AI literacy, equitable student AI use, administrator capacity, and stakeholder collaboration. While second-year guidance grew more prescriptive, it remained largely anecdotal, with equity frequently invoked but seldom operationalized.
This study examines whether attending a charter high school during the pivotal middle-to-high school transition yields meaningful educational benefits for Texas students who previously graduated from traditional public middle schools. To account for high school mobility, we distinguish charter stayers (who remain in charter high schools) from charter leavers (who return to traditional public schools) and match each group to observationally similar traditional public school (TPS) students (who remain in TPS). Results indicate that charter stayers accumulated more dual-credit coursework and achieved moderately higher standardized exit-exam scores, whereas charter leavers appeared broadly similar to TPS students. In postsecondary outcomes, charter stayers were more likely to enroll on time, attend four-year colleges, and complete degrees earlier, while charter leavers showed little difference in longer-run attainment. These positive associations were concentrated among students attending campus charter high schools, with relatively limited differences for students attending open-enrollment charters.
We explore North Carolina’s teaching assistant (TA)-to-teacher pipeline by assessing outcomes for early-career teachers with and without prior TA experience. We find that those with prior TA experience make up a meaningful percentage of beginning teachers and possess characteristics that are needed in the teaching workforce (e.g. more likely to be a person of color and teach special education students). Former TAs are performing comparably to peers without TA experience and are much more likely to stay in teaching. Future research should assess ways to best recruit and support individuals transitioning from TA to teacher roles.
This study was designed to investigate the impact of a student success team intervention on the ninth grade on-track to graduation (9G-OTG) rates of students in Oregon. Interrupted time series (ITS) models were used to estimate the baseline change in 9G-OTG rates during the four academic years prior to the implementation of the success team intervention (2013–2014 to 2016–2017) and in the 5 years that followed (excluding 2019–2020). Results demonstrated a slowing of the 9G-OTG growth rate, but a closing of the gap between schools with different take up levels following the onset of the intervention. During the COVID-19 era, a steep decline followed by a sharp rebound in 9G-OTG rates was observed for all implementation levels. Across the intervention years, schools that implemented the success team model experienced a modest gain in 9G-OTG rates while non-implementing schools recorded a slight loss. Implications for education policy and practice are discussed.
As generative AI reshapes K-12 education, U.S. states are rapidly developing guidance to navigate its implementation. This study investigates the structural composition of AI education guidance across thirty-five states through qualitative document analysis. Utilizing the CAPE framework and the concepts of policy resilience versus fragility, the research examines how state-level signals influence systemic preparedness for AI education. Findings reveal a landscape of structural fragmentation, where guidance often prioritizes risk mitigation over pedagogical innovation. These early policy choices have potential to shape teacher capacity and sustainability, determining whether AI fulfills its promise of empowerment or widens existing education opportunity gaps.
As artificial intelligence (AI) becomes increasingly integrated into K–12 education, hybrid and microschooling environments present both opportunities and challenges for responsible implementation. This paper examines stakeholder preferences for AI use policies through a nationally distributed survey and a conjoint experiment targeting school leaders, teachers, and parents from hybrid schools in the United States. We find broad support for tightly regulated student-facing AI applications, particularly for low-stakes tasks like tutoring or brainstorming, and cautious openness to teacher-facing tools. However, significant misalignment exists across stakeholder roles: while school leaders express greater enthusiasm for AI integration, teachers and parents adopt more guarded stances. Our findings underscore the importance of co-designing AI policy by incorporating perspectives from all key stakeholders and by tailoring rules to specific educational tasks. The study further offers practical insights to guide school leaders in developing responsible, stakeholder-aligned AI policies within hybrid learning contexts.
A substantial number of U.S. students experience homelessness, yet our understanding of how homelessness shapes student outcomes is limited. We use 7 years of longitudinal data on Indiana students in kindergarten through eighth grade, including more than 40,000 students who experienced homelessness, to examine the associations between homelessness and academic and behavioral outcomes. Our data indicate that Black and low-income students are more likely to experience homelessness and for longer periods. Student fixed-effects models indicate that cumulative years of homelessness are associated with lower math and ELA achievement, as well as increased disciplinary incidents and absences. These associations are particularly pronounced for male, White, and middle school students, suggesting that in the context of homelessness, being male, White, and in older grades may function as risk factors for more adverse outcomes. Together, our findings underscore heterogeneous links between homelessness and adverse outcomes across student subgroups.
School Resource Officers (SROs), or sworn police officers in schools with arrest powers, are increasingly common in schools. Despite concerns surrounding SROs broadly and their use of arrest, in particular, little is known about how SROs understand their role in arrest. Drawing on a case study of SROs’ perceptions of arrest across two school districts and using a niceness as Whiteness framework, we found that SROs generally expressed trying to avoid arrest. Yet, when they recounted making arrests, they used mechanisms of niceness as Whiteness to justify arresting young people. We examine implications for school safety, care, and future research, policy, and practice.
Emergency preparedness is a critical priority in schools as institutions seek to ensure the safety of students, including those with disabilities amid rising concerns over natural hazards, community violence, and public health crises. This mixed-methods study explores the relations between school personnel’s familiarity with Emergency Operations Plans (EOPs), often referred to as Emergency Response Plans (ERPs), and their perceived preparedness to support students across these diverse emergency contexts. Results indicate that EOP familiarity significantly predicts self-reported preparedness to support both general education students and students with disabilities. Qualitative findings highlight systemic barriers in implementation, including insufficient training, evacuation challenges for students with mobility needs, behavioral concerns, and communication gaps. Findings suggest that inclusive emergency planning serves as a lever for whole-school improvement by forcing institutions to address interprofessional collaboration, communication protocols, and accessibility, thereby strengthening the educational environment for all students.
Professional learning partnerships (PLPs) are an increasingly popular policy approach for improving instruction. With data from 12 PLPs serving major urban districts, we leverage teacher surveys using structural equation modeling to examine the policy attributes and school context factors contributing to improved instruction. We found that teachers with higher buy-in to the partnership and who rated their PL as relevant and useful were more likely to use ambitious and culturally responsive instruction and less likely to use traditional approaches. We discuss implications for understanding the supports necessary to build successful partnerships as an effective policy mechanism for school improvement.
Generative artificial intelligence has forced school systems to rapidly pivot from reactive bans to proactive management. Yet, research on local policy responses remains scarce. This study analyzes how the twelve largest US school districts have responded to the introduction of GenAI. Through document and policy analyses, we identify three interrelated policy moves: redefinition of foundational concepts like academic integrity; regulation by creating standards for both users and vendors; and innovation through intentional experimentation. We offer a typology for district-level AI governance, highlight domains of regulatory changes, and suggest a baseline for future causal, evaluative, and implementation research.
This study presents a comparative policy mapping analysis of publicly available generative AI policies and guidance across America’s flagship public universities, based on documents collected in 2025 and analyzed through thematic coding informed by Clark’s Triangle of Coordination. Findings indicate a distributed governance architecture in which instructional decisions are typically delegated to the course level and supported through instructor-facing resources, while institution-level guidance emphasizes data protection, approved tools, and risk management. Although institutions vary in posture, codification, and the depth of guidance provided, the findings point to structural convergence around this shared governance model, suggesting that flagship universities are responding to generative AI in ways that reinforce established distributions of authority within public higher education systems.
