Abstract
Multiple fields including social work, public health, psychology, and education have been researching school safety issues for the past 40 years. Findings from research and intervention have been a driving force in school safety policy change. Likewise, policy and funding patterns created have dramatically altered how and what school safety researchers study. Yet school safety policy has not been the focus of empirical studies in social work or other research fields as either independent or dependent variables. Based on prior work and current school safety policy issues, the authors created a new theoretical ecological model to generate a conceptual policy-focused model of school safety. They identified thematic, conceptual, methodological, and analytic gaps in the school safety literature using the new model as a guide. Examples from policy and empirical work are given to elucidate the conceptual model. Suggestions are made for researchers and universities to close the gap between school safety research and policy.
During the past 40 years, school safety policy and real-life events have shaped the development of research, interventions, and theory (American Educational Research Association, 2013; Astor & Benbenishty, 2019a; Astor et al., 2013; Interdisciplinary Group on Preventing School and Community Violence, 2013). Likewise, research and theory in psychology have strongly affected global school safety policy (Cohen & Espelage, 2020). Contributions to the body of school safety scholarly work have come from multiple subfields of social work, public health, psychology, and education (Mayer & Jimerson, 2019). This article focuses on both how ecologically driven theory can help better integrate policy into theory and how to enhance the impact of ongoing school safety policies. The authors reexamine socioecological school safety theory and adapt it from earlier social work ecological perspectives to a new policy-oriented heuristic conceptual model (e.g., socioecological theories focused on school safety, see Astor & Benbenishty, 2019a; Astor, Benbenishty, et al., 2021; Benbenishty & Astor, 2005; Cohen & Espelage, 2020).
The article highlights multiple ways a policy-oriented heuristic theoretical model and its potential methodological and research implications are relevant for understanding and changing school safety policies. Although the current focus of this article is school safety, many ideas expressed in this article are relevant to other policy areas in social work, education, and psychology research. Improved policy-oriented theory and methodology could increase the value of theory and research for school safety policy (e.g., policies regarding other pressing concerns, such as addressing issues of opportunity structure and recovering from the pandemic; see Astor, Benbenishty, et al., 2021; Kelly et al., 2020a, 2020b; McMahon et al., 2022a, 2022b).
School safety theoretical and empirical models do not explicitly address policy (Astor & Benbenishty, 2019a; Benbenishty & Astor, 2005). Many safety studies address policy issues only in the implications section of academic manuscripts. When researchers participate in congressional meetings, parliament briefings, state-level gatherings, and school board meetings, they often use findings from studies mainly to support policy recommendations empirically. School safety researchers often don't intentionally include policy as an intervention and driving force in research questions or theories of school safety. Policy is central to the theory and has a direct impact on scholarship. Hence, this article includes policy more intentionally as an independent and dependent variables in a new conceptual framework. Earlier work that focused on school policy change is summarized and reviewed as examples that highlight aspects of the new policy-oriented school safety conceptual model presented in this article.
This manuscript explores the following questions: How do policies affect school safety? How does school safety influence policy? How do school outcomes and contextual factors interact with policies on multiple socioecological levels? This conceptual reframing of questions surrounding school safety and policy moves from seeing policy as an independent variable that affects school safety to a more integrated reciprocal view of how policies are also dependent variables influenced by contextual factors (e.g., global events such as the pandemic, school safety watershed events like highly publicized mass shootings in schools, the impact of global wars and migrations on school safety and trauma, and the effects of multiple types of media on school safety policies). In the following sections, the main components of the model are briefly described, with a focus on the role of policy. The remainder of the article explicates in more detail real-world examples and implications related to specific methods, analyses, and samples that are relevant to social work, public health, education, and psychology researchers.
Revised Policy-Focused School Safety Model
Overview of a Policy-Oriented School Safety Model
In Figure 1, the school is in the center of the model. The school is embedded in multiple contexts, including the characteristics of the students (e.g., special needs, ages); families (e.g., poverty); neighborhoods; communities (e.g., crime, cultures, and ideologies that influence the school); the effects of the media on the school; and universities that train the school staff, develop school safety programs, and conduct research on school outcomes. Furthermore, as seen in Figure 1, the school is embedded in larger contexts such as developments in science and technology (e.g., cellphone technology, social media innovations, distance-learning applications) and global events (e.g., pandemics, school shootings, cultural and political strife, wars, and economic downtrends).
Revised policy focused Astor & Benbenishty heuristic school safety model.
Policies are an important factor nested in each context that affects the school. Multiple policies originate at the federal, state, county, and district levels. In the United States, every school district has a policy governing body, typically a school board, which issues directives, some of them in accordance with national and state policies and others in direct conflict with macro-level policies. The United States has recently witnessed this phenomenon of discordant school safety policy layers (national, state, regional, district, and school). Some examples include COVID-19 school safety policies regarding immunization, masking, and social distancing. There is wide-scale strife about cultural issues that include the banning of critical race theory in the curriculum (Schwartz, 2021), restricting educators” autonomy to talk about LGBTQ issues (Sawchuk, 2022), and whether to allow corporal punishment in schools (Gershoff & Font, 2016). These are only a few of the many policies affecting school safety that are in blatant disagreement across the United States.
Given that schools are influenced by policies emanating from many different contextual levels, it is essential to examine policies on all levels, not only federal and state policies as it is often done today. Policy making at the school board and school level on issues of school safety have rarely been explored in empirical studies (Reynolds & Astor, 2018, in press; Reynolds et al., in press), even though school board deliberations and subsequent policies may be the core sources of intervention for many U.S. schools. This is likely due, in part, to the absence of policy in theoretical conceptual models.
In Figure 1, the school does not merely mirror the influences of outside contexts. Schools also create internal contexts, such as school climate, that could mediate external influences, buffering, and moderating them in multiple ways. Local school safety policy can also exacerbate external conditions or maintain the status quo. This becomes very important when conceptualizing systemic racism (Zimmerman & Astor, 2021) or opportunity gaps that affect specific groups over time (Astor, Benbenishty, et al., 2021). Hence, in this model, schools embedded in the same communities and neighborhoods with similar family demographics may differ widely in their school-level policy and levels of school safety. This brings theory into alignment with empirical findings over decades indicating wide variations in school safety outcomes in schools in the same communities, with similar risk and resilience factors (Astor & Benbenishty, 2019a; Benbenishty & Astor, 2005). These internal contexts include policies related to school-level academic, socioemotional, social climate, and organizational procedure contexts (CASEL, 2022; Cohen, 2014). A school’s internal policies may affect multiple school climate areas that include not only students but also the educational staff, parents, and community-school relationships (e.g., policies regarding discipline, suspension, expulsion, dealing with gangs surrounding schools, or college-readiness opportunities). Furthermore, the ways the school organization implements these policies (e.g., in fair or discriminatory ways) are part of the school's internal context (CASEL, 2021).
The school and the multiple external contexts that affect it evolve over time. Some changes in external contexts are due to gradual and long-term processes, such as changes in neighborhood demographics or normative values regarding the acceptability of specific violent behaviors. These include norm changes that are widely debated in the public as acceptable and not acceptable in schools, such as the long trajectory of women's rights culminating with the watershed #MeToo movement and policies of inclusivity regarding gender (e.g., Title IX), gender identity, LGBTQ support, and antiracist education. These movements are examples of wide normative shifts in what is considered acceptable by our culture that have a direct impact on school safety policy and interventions.
Other changes are more rapid, sometimes in response to developing technology (e.g., quickly changing software, hardware, and applications of social networks). Some socioecological changes are in response to horrific worldwide watershed events, such as school shootings (e.g., Columbine, Sandy Hook, and Parkland), war or terrorism (e.g., 9/11 and the Iraq and Afghanistan wars), famine, and mass migrations. The timing of historical events and linked norm changes likely interact with school contexts and even affect local school-level policies (Astor & Benbenishty, 2019a). This likely happens after horrific events are widely amplified by the media. Despite their massive impact on policies, historical watershed events are not commonly included in school safety research or as time anchors to understand policy shifts. This is particularly notable regarding school shootings. Many mass shootings in schools had strong and direct impacts on all levels of policy, including new laws, program development, and educators’ personal attention to issues of school safety (Benbenishty & Astor, 2020). National professional and research organizations, including many professional psychology organizations, advocated for many years at the national and state levels to bring about these policy changes (Astor, 2018; Astor et al., 2013; Flannery et al., 2019; Interdisciplinary Group on Preventing School and Community Violence, 2013, 2018). Yet, few of these efforts are documented or included in school safety research.
Finally, the model depicted in Figure 1 includes a feedback loop that shows the impact of school outcomes on external contexts. For instance, schools that create a peaceful climate may also have a positive effect on levels of crime in their neighborhood and over time, may attract more families. A school's positive reputation academically and with safety could affect real estate prices and rents (Astor & Benbenishty, 2019a; Chen, 2021; Colley, 2017). Tragic school safety outcomes can affect safety policies on more macro levels. For example, horrific events even in one school (e.g., Parkland, Columbine, Sandy Hook), with the amplification of media, have changed the course of policies, practice, and action for schools across the world (Astor et al., 2002; Everytown for Gun Safety Support Fund, 2021; Sandy Hook Promise, n.d.). Practical and policy solutions to these global problems are generated by politicians, law enforcement, and university researchers. The intensified coverage of the media likely plays a major role in creating ground support for further policies and interventions (Meindl & Ivy, 2017; Ordway, 2019). Yet the role of widespread media or social media (e.g., viral videos of violence in schools) in tragic events at school is almost never included in conceptual models or school safety research findings. Figure 1 highlights these issues as potential variables for research. Studies that include these variables could increase the value of research on school safety to media companies and lawmakers from a policy perspective.
Figure 1 includes the role of universities. University researchers have influenced policy by promoting evidence-based research on school safety and conversely, are affected by school safety policies. For example, the availability of grant funding has influenced what researchers study and the types of programs they develop and test (Astor & Benbenishty, 2019a). Most current national websites, calls for grants, and clearinghouses on school safety-related evidence-based practice emerged in response to wide media coverage of national shootings in schools, along with advocacy from academic researchers. Debates have occurred at the national level about the role of mental health and socioemotional learning to prevent school shootings (Chatterjee, 2019; Varghese, 2021). Empirical research on school safety, conducted mainly by university researchers, is one way that safety approaches inform local, state, and national policies. Only a handful of studies has explored the role of university research in influencing school safety policy. Perhaps for the first time, the present school safety model includes both universities and their potential impact on policy.
University researchers are at the center of societal conflict surrounding evidence-informed research, research without evidence, or practices that the evidence shows are harmful. Based on large research literatures, almost all evidence-based approaches being promoted by university researchers are aimed at increasing positive climate, socioemotional learning, restorative justice, psychologically based mental health, or trauma-informed approaches. These have been widely included in many state, national, and local policies.
University researchers are not the only major group proposing school safety policy. With equal policy verve, law enforcement, security specialists, policymakers, and military experts have promoted measures that “harden” schools. In some school safety literature, this is called a “zero-tolerance” approach (American Psychological Association, 2008; Astor & Benbenishty, 2019b). These include metal detectors, police on school grounds, scent dogs, see-through backpacks, strict and frequent use of expulsion and suspension, arming educators, and dramatic stagings of school shooting drills. The vast body of research suggests that zero-tolerance and police-oriented measures contribute strongly to the school-to-prison pipeline and an overall “prisonlike” negative mental health environment (Paxson, 2021). Even so, in recent decades, the United States has seen mounting strong tensions between policies emerging from law enforcement and university researchers and national psychological associations (Mayer & Astor, 2018; Reynolds & Astor, in press; Reynolds et al., in press). At the school level, recent national findings suggest many schools across the United States have adopted some combination of both zero-tolerance and socioemotional measures (McMahon et al., 2022a, 2022b).
As another current school safety example, some school districts passed recent policies on “de-policing” schools as a response to the Black Lives Matter movement—stemming from the continued videotaped and publicized murders of unarmed Black citizens by police officers. Many have argued that police, school resource officers (SROs), and security guards, by definition, create an unsafe school climate for all students but especially students of color, who may have more negative personal and historical experiences with law enforcement and attribute a different meaning to weapons and police (Heidelburg et al., 2022). Researchers have argued that a large body of evidence suggests a law enforcement orientation toward school safety sends a structurally negative message to students of color about schooling and possibly even primes students for the school-to-prison pipeline (Barnum, 2020; Johns Hopkins Medicine, 2021).
Most research discussions of these important empirical and philosophical discussions call into question the compatibility of employing both hardening and softening measures in the same schools. School safety theory and research have not commonly directly evaluated how the multiplicity of different programs with evidence, without evidence, or with evidence showing harm affects school safety (Astor & Benbenishty, 2019a). Figure 1 offers conceptual opportunities and suggested paths for researchers in social work and other subfields to find ways to examine the confluence of these school safety policies and urges researchers to include historical anchors that trigger policies as both independent and dependent variables in research (depending on the questions being asked). The following sections expand and highlight some promising areas for further research based on this model.
Implications of the Model
School in the Center: School-Level Opportunities for Policy Research
The center of the model is the whole school (rather than an individual student) embedded in multiple external contexts. As such, it provides many opportunities to understand policy as it relates to school safety research. First, although school safety policies are developed to increase the safety of individual members of the school community, they are implemented, mediated, and modified by each school unit and the school districts in which they are embedded. When the focus is on the individual student, studies tend to center on an individual's characteristics and interactions with others (peers and adults), rather than the relevant organizational or capacity variables that affect the school or school district context. These organizational variables, although commonly known by most educators, often remain in the distant background and are not an integral part of school safety theory or research.
By contrast, the conceptual emphasis on centering the school in the theoretical model brings the policies that influence the school to the foreground. Policies can more easily be examined as one of the important contextual factors that affect schools. As an example, to understand staff victimization of students, it may not be enough to examine the interpersonal dynamics between the staff and students. It is essential to understand the policy context, both formal and informal, regarding corporal punishment.
One would expect major differences in the prevalence of student victimization by staff members in schools that operate under policies that allow or disallow corporal punishment. It is possible that teachers in any of the 19 U.S. states that allow corporal punishment as a form of school discipline behave differently than teachers in states that ban corporal punishment, irrespective of their personal characteristics.
Corporal punishment also has philosophical empirical underpinnings that drive tensions between policies allowing and disallowing it in states and school districts. As an example, for many decades now, corporal punishment in schools has not been taught as a legitimate form of school discipline in any U.S. university educator training program and is perceived as harmful to students by most university educators (Font & Gershoff, 2017; Global Initiative to End All Corporal Punishment of Children, 2019). As previously mentioned, university researchers and educators are strongly in favor of a positive school climate, socioemotional programs, and evidence-based practices to curtail school violence and increase school safety.
Does it make philosophical sense to have socioemotional learning or anti-bullying programs in schools that actively use corporal punishment? Interestingly, despite many discussions on hardening schools or the school-to-prison pipeline, the school safety literature has not empirically examined how students, faculty members, or parents perceive school safety as part of the corporal punishment policies present or absent in their schools. A recent study suggested that schools in many states that allow corporal punishment also invest in socioemotional learning and positive school climate interventions (McMahon et al., 2022a, 2022b). Most academicians see softening approaches that aim to create a positive climate as philosophically incompatible with hardening and zero-tolerance policies such as teachers carrying guns, police on campus, and dog inspections of personal belongings (Interdisciplinary Group on Preventing School and Community Violence, 2018; Mayer & Astor, 2018).
Change in the School as a Unit vs. Change in Individual Students’ Development
The focus on the school in the center is relevant to understanding changes in and surrounding the school on issues of safety over time. Most bullying and safety research has incorporated time only as it relates to individual students (Cohen & Espelage, 2020). For instance, some studies focused on students’ developmental trajectories regarding bullying involvement, whereas others compared individual students’ bullying involvement before and after participating in anti-bullying programs. Only a few researchers have addressed changes over time with schools as the unit of analysis (Benbenishty et al., 2016).
Examining the safety of the school as a unit of analysis is an important policy-oriented approach because schools change over time and create different social contexts for students as they improve, stay the same, or decline in organizational efficiency or climate. To illustrate, some students may attend a school that has invested time and human resources to create a safer and superior organizational climate for a few years. Most school safety researchers would predict that safety should then improve for students in this school over time, whereas students from other schools that are disorganized with a poor school climate—even if they are in the same communities—may not fare as well (Astor et al., 2009). Students, faculty members, and families in those disorganized schools would likely have fewer positive safety outcomes during the same period than those affiliated with the more organized school. This is such a simple proposition that many may wonder why this is a research prediction at all! Yet few school safety studies have examined this critical organizational climate policy question. If the school social and organizational unit is not included in the model, understanding the nexus between safety changes at the organizational level and individual developmental trajectories will likely have limited value.
Implementation of Programs and Policies
Another example of the value of having the school as a unit in theory and research is the examination or implementation of school safety policies and programs at the school or district level. In most studies, programs are tested for efficacy in a relatively small set of schools. Generally, researchers collaborate with schools where the superintendent and principals agree to participate in the study. Researchers also tend to select districts where the superintendent and principals are cooperative and excited about trying a program. Variability among schools by districts, counties, and regions is studied mostly as an ex post facto explanatory variable for variations in the implementation of programs and subsequent outcomes. This theoretical approach expects researchers to consider variability among schools as an important determinant of the relevance of the program or policy to each school and the probability that it would produce positive outcomes.
A selection process that does not consider the school unit characteristics may be the reason for so many failures to implement programs at scale. One size does not fit all, and if the school characteristics are not part of the program development and implementation process (and the focus is only on the characteristics of the individual students targeted by the intervention group), it should not be surprising that many schools would find an evidence-based program developed with other schools as irrelevant to their context.
However, this kind of selection process created by funding policies may be a factor explaining why many evidence-based programs supported by federal, state, and foundation sources have reported failures when scaling up to large regions (Astor & Benbenishty, 2019a). The exact variables that excluded a school or district from being in a study may be responsible for the failure of the implemented program. In the model presented here, program development and testing need to be sensitive to school-level characteristics and ensure that effectiveness studies include representative samples of schools.
The selection of specific schools and school districts for testing intervention programs is often primed for the success of the studies and interventions in optimal contextual conditions. Schools that are disorganized, punitive, or don't connect with policies or programs may not invest time and human resources to scale a program or intervention effectively.
As an alternative to current practices, consider, for instance, the development of programs that arm teachers or include armed SROs on campus to enhance school safety. In the model presented here, schools are the unit of analysis, and it is important to conceptualize, in advance, external and internal contextual factors that could determine the effectiveness or ineffectiveness of such policies or programs. By including such school-level consideration, it would be possible to tailor programs to specific contexts and bring them to scale only among schools that have similar characteristics. Recent studies in Virginia have shown that many faculty members and students, including faculty members and students of color, have favorable views toward SROs (Breen, 2020). This may not be the case in California or other states. Therefore, research exploring the school as a unit should first consider the cultural values and ideology of parents and staff members as a first step to examining whether such a program should be implemented in a school and consider these issues in assessing program implementation and outcomes (Astor & Benbenishty, 2018; Astor et al., 2018; Astor, Noguera, et al., 2021).
Funding, Community Resources, Human Resources, and Regional Supports
To influence policy on school safety, it is essential to assess the implementation of current policies that provide additional and ongoing fiscal, organizational, community, and human resources in schools. In fact, as a policy question, how much schools and districts spend on each student and specifically how much money is spent on school safety and related issues is critical. There is little research on levels of funding needed for schools and districts to successfully sustain or implement safety-related issues over time. Different types of policies may be needed at the school level (e.g., climate, discipline, welcoming, socioemotional learning, or community oriented) that work in concert to improve overall school safety in a sustainable way. Human resources at the school level also likely matter for programs to exist. How many professionals or school staff members are dedicated to implementing and maintaining programs at the school or district levels? Are they all funded by external grants? Is there a plan to integrate these types of resources that contribute to safer schools into the school or district budget over time? These state, district, and school policies are rarely addressed in research policy studies. Fiscal and human resources are important variables to include, especially if researchers want to track changes over time that explain the sustainability and scaling of school safety programs. These policy variables could be researched as independent and dependent variables and better connect school safety research to the concerns of policy makers.
Many Schools are Needed for a Good Study Sample
Overlooking the importance of the school as a unit in school safety may reflect the difficulty of securing enough schools to have a meaningful analysis of school-level variables. This is especially obvious in school safety intervention studies that attempt to test a whole-school safety intervention but have only a few school units and conduct most of their analyses at the student level. This is a major limitation of existing studies and reduces their contribution to the development of effective policies that can scale to large regions or states. This may have led to a one-size-fits-all approach with evidence-based programs that were mandated or endorsed by state-level policies but never validated for entire regions or large school districts (Astor et al., 2016). One solution that could align state-level policy on evidence-based programs and challenges involving scaling is to involve entire states and regions in developing best practices for school-level data collection and measuring their effectiveness in promoting feelings of safety and reducing violence over time. Many regions have surveillance tools they use regularly but are not linked to local or statewide school safety intervention attempts. For school safety policies and programs to be developed and disseminated at scale, it is essential to have the school at the center of the model and as a unit of sampling and analysis (Astor & Benbenishty, 2018).
Schools Embedded in Multiple Contexts
Having the school in the center of this ecological model helps highlight the many contextual factors that affect the school, alongside policy. These include, among many others, the characteristics of students and their families, staff members, and the community (for discussions on the importance of these contextual factors to school safety, see Benbenishty & Astor, 2005; Benbenishty et al., 2016; Khoury-Kassabri et al., 2005). The model helps distinguish between the characteristics of the individual student or staff member and the school-level composition in terms of students and the staff. This provides a theoretical and empirical opportunity to identify how policies are implemented across the school. The proportion of specific types of students in a school may contribute greatly to the safety and a wide array of other resilience or risk factors at the school level. For example, a large array of studies has suggested that having a large proportion of first-generation immigrant students in a school is highly protective for various factors, including substance use, obesity, school violence, and other health variables. Conversely, a large proportion of second-generation immigrants in a school may increase the prevalence of negative outcomes for students (Astor & Benbenishty, 2019a). The proportionality of different groups of students in schools may reflect protective or risk factors. Greater explorations of this phenomenon should occur to further policy on diversity and collective support in schools.
Replicating Results Found in Evidence-Based School Safety Programs at Scale
The inclusion of context and time in the model may help address issues that have concerned the scientific community in recent years, such as difficulties in replicating important findings and disappointing outcomes when evidence-based programs are implemented at scale (Astor et al., 2010). Perhaps the fact that many findings cannot be replicated is not only because the original studies had methodological problems but also because the original findings were associated with specific micro contexts and historical points in time. To illustrate, schools in transition from one principal to another are rarely, if ever, included in randomized controlled trials testing the effectiveness of programs. Yet those of us in the field commonly hear how programs shift and change when principals leave for other schools. It is no surprise, therefore, that when these programs are disseminated at scale and implemented in a wide range of high- and low-functioning schools, including those in organizational transition, they may not show positive effects. The problem is not so much the implementation but the exclusion of issues of context and time in testing and disseminating programs. By including the organizational context and time in the school safety model, the relevance of policies, programs, and practices to other contexts and times could be critically examined (Astor et al., 1999; Benbenishty et al., 2016).
Comparing International Contexts to Enhance Understanding of Systems
Most comparative studies on school safety in psychology have focused on differences in victimization base rates and their emotional impact between students who live in different countries (Benbenishty & Astor, 2005). These studies rarely considered and compared the organization of schools, policies, programs, and safety services in these countries (Arënliu et al., 2022; Benbenishty et al., 2019, 2022). This is a significant limitation because these differences in organizational and policy context could explain differences found between student safety outcomes across many different countries. Expanding the ecological context to settings outside the United States and Western cultures could give a greater sense of the research policy interchange and definitions of safe schools in different hierarchical systems (Chen & Astor, 2010). Understanding better the hierarchical nestedness of various units in the educational system allows researchers to see the complexities involved in comparing students from different cultures and countries. Although some of these comparisons try to account for differences in student characteristics (mainly socioeconomic status), little current empirical attention is being paid to major differences in the organization of the educational systems and policies in which these students are embedded.
Compare a school in Israel with a school in the United States. During the last 25 years, Israeli public school capacity in terms of support for violence prevention and improved climate has grown tremendously to serve the nation's 5,000 schools. Close to 9,000 counselors and psychologists were retrained to define school safety and positive school climate as a primary responsibility. The centralized educational system provides materials and education to these counselors and psychologists and supports them with supervision and expert advice in a centralized and sustainable way. In addition to regular support, schools are eligible to participate in a three-year process for schools that voluntarily wish to improve their climate and reduce school victimization (for more detailed descriptions of the Israeli context, see Benbenishty et al., 2020).
The clear message from Israel's Ministry of Education about the importance of climate and prevention of violence is underscored by an accountability system that integrates academics and climate, sends feedback reports to each school, and drives school, regional, and national policies, and practices on school safety. Providing school districts with reports that combine academic achievement data and a picture of their school climate over time (compared with other similar schools) is a powerful statement of the priorities of the national educational system.
Compare this centralized system with the situation in the United States. California alone has more than 10,000 schools and 1,000 school districts. Each district has a locally elected school board whose members deliberate every year to determine priorities and resource allocations (Reynolds & Astor, in press). How federal, state, county, and city priorities and policies are interpreted by and translated into resource allocations for districts and how school boards select where to invest funds and safety programs assigned by districts to specific schools creates a very complex picture, influenced by each level and the ecological policy hierarchy. Studies looking at different units (district, county, state, and country) would be extremely helpful in filling the gaps in empirical knowledge as to how the vertical organizational structures in Figure 1 affect schools and students. It is very likely that the programs work well in some organizational systems and less well or not at all in other systems. Exploring this in empirical studies could provide empirical bases for making organizational changes in centralized and decentralized school systems that could improve safety.
Policy as an Independent and Dependent Variables Related to School Safety
The model highlights both the impact of policies on schools (policy as an independent variable) and the factors that shape policies (policy as a dependent variable). Policies, especially as they translate to resources for schools, have an important impact on schools. Here, it is important to identify the multiple levels of policies, such as federal, state, county, district, and school board policies. These policies interact, mediate, and moderate one another. Hence, to understand the impact of federal-level policy (and the resources attached to the policy) on schools, it is essential to examine how higher-level policies interact with local policies, sometimes enhancing their impact and other times diluting the original federal-level goals. This is especially important because these lower-level policies are influenced to a larger extent by more local cultural values and ideologies that may embrace, block, or modify higher-level policies.
To illustrate, the pandemic demonstrated how school boards and districts responded differently to state or federal policies. There is a current void in safety research regarding the role of school boards or school districts in shaping school safety policies (Astor & Benbenishty, 2019a; also see Reynolds & Astor, in press). Although school researchers are aware that they need to seek the permission of the principal, school district, or school board to conduct a school safety study, they do not methodologically or analytically include the administration or board in their studies, despite their important role in formulating school safety and bullying policies (Reynolds & Astor, in press). Consequently, from a methodological point of view, analyses of the impact of certain policies, such as zero tolerance, need to consider local policies as important sources of variation.
Figure 1 highlights the fact that although external contexts affect school safety, they are buffered and moderated by internal school contexts. These include the socioemotional, academic, and organizational climate of the school. External policies, including guidelines and resources, affect students only after they are processed by the school internal climate. As such, to understand the impact of policies, it is essential to understand the internal school environment and school-level policies. For instance, schools that receive the same resources under a federal or local policy may differ significantly in how effectively they take advantage of these resources to improve school safety.
Recent studies on military students in public schools highlight specific aspects of Figure 1. Studies showed that a significant number of students from military families identified major variations in the extent to which civilian public schools took advantage of federal policies intended to support military students. These policies set a threshold for the number of military students that makes a school eligible for significant federal funding (Astor et al., 2012a, 2012b, 2012c, 2012d; De Pedro et al., 2018; Esqueda et al., 2012). This included annual subsidies (Military Impact Aid) and Department of Defense Educational Activity partnership grants (Astor et al., 2012a, 2012b, 2012c, 2012d; De Pedro et al., 2018). Public schools could not apply for these funds unless they knew how many students they had from military families. Whereas some schools in each district had an annual count of their military students and submitted a request for funds and received them regularly, other schools, sometimes near military bases, were not even aware of the fiscal opportunities that the policy offered. Some schools were aware but disorganized to the extent they could not find human resources to count the number of military students (Astor et al., 2013; De Pedro et al., 2014; Gilreath et al., 2014). As part of a university-led intervention, schools and districts became aware of the federal policies and obtained hundreds of thousands of additional dollars each year to support and sustain safety and other programs (Astor, Benbenishty, et al., 2021; Benbenishty & Astor, 2014). Similar issues were noted with many schools not aware of policies that provide funding for students experiencing homelessness or in foster care (Benbenishty et al., 2018; Moore et al., 2020). In California, some schools were not aware they could obtain local accountability plan support for their school safety efforts (California Department of Education, 2022). When school-level policies were modified, the funds provided by state-level policies could be harnessed to transform climates of safety for multiple groups of students. It is important to study systematically and comprehensively how school policies interact with higher-level policies to affect students’ safety.
A theory of school safety needs to consider how policies affect school safety but also how these policies are developed and implemented. Understanding policies is critical to identifying how they may have a negative impact on school outcomes and how to leverage policies to help improve school safety. The model in Figure 1 demonstrates some of the many contextual factors that influence policies. Some influences are more local and have more effect on local-level policies. These include, for instance, the characteristics of students’ families. Their culture, religion, religiosity, and political ideology may influence the district and school boards in shaping local policies and modifying or blocking higher-level policies.
Other larger contextual factors may influence policy formation on all levels. Some influences are not directly related to school safety but have implications for safety. For instance, the advent of smartphones and social networks required policymakers on all levels to develop policies and programs to protect students from safety-related threats (cyberbullying, sexting, racial slurs, posting of inappropriate videos, etc.) of these technologies (Astor & Benbenishty, 2019a). Another example is policies enacted during the COVID-19 pandemic. Although the issues addressed were health-related, they affected students, educators, staff members, and parents in multiple ways—including reduced face-to-face bullying, increased internet-based victimization, and increased confrontations between the school staff and parents (Reynolds et al., in press; Watson et al., 2022). School safety researchers using an ecological model need to be vigilant in identifying how emerging contextual developments may affect students’ safety and well-being so that they can advocate for policies that may help prevent negative outcomes and promote the safety of students and schools.
What Academicians Can Do to Include Policy Research at the University Level
An important first step to promoting this approach is to be intentional. This means universities need to include policy impact and research in master's and doctoral training regarding the scientific and academic process from the start, especially how scientists use policy to formulate research questions and learn policy-oriented methodologies. Including direct instruction on using social media, writing op-eds and policy briefs, and speaking with media and educators would be important. However, this should be done in the context of scholarship rather than dissemination or implications. Graduate students (and instructors) in social work, public health, education, and psychology should be educated about policy analyses and how to write about and conduct research for policy change.
Most policies are tied to time-sensitive matters. Researchers need to prioritize using research for policy change even before submitting it for peer review. Very often, findings are published after several years and by then, they are not relevant to policymakers. This requires quicker analyses and the need to integrate pending laws and issues into data collection time frames and processes and prioritize immediate impact with time-relevant findings. Credit should be given to publicly facing scholars with products that go beyond scientific journals and include policy briefs and popular media. Some medical journals prioritize findings that have policy relevance and publish them first. Journals in social work and social sciences should consider this practice as well. Due to the reward systems in academia, researchers often prioritize the longer process of peer-reviewed research over the immediacy of policy research. Policy efforts are not always rewarded in academia as scholarship and instead are seen more as a service, even though these efforts (at least in part) may lead to new federal and state laws and resources for schools. These reward systems in universities should be looked at carefully so that policy-oriented research and efforts based on empirical results are rewarded and encouraged as central to the research endeavor.
Main Takeaways and Recommendations for Research
Universities and researchers are part of the theoretical model of school safety, perhaps for the first time. Universities may have a major impact on school safety policies and practices. There are immediate ways that university researchers can integrate policy in school safety theories and research designs and methods.
Making Policies Part of the Contextual Factors in School Safety Studies
The impact of policies needs to be empirically assessed, either directly or by examining whether variations in policy context are associated with variations in key variables (e.g., level of violence, presence of weapons in school). Policy analysis and impact are developed areas of research with distinct methodologies, yet rarely are these methods used to carefully examine the direct impact of specific school safety policies—for example, laws surrounding suspension and expulsion, special education policies that respond to extreme behaviors, laws focused on identifying bullies, and corporal punishment laws in response to misbehaviors.
Studying the Effects of the Implementation of Specific Policies
Some policies may have research supporting them, but their implementation on the ground is not successful due to contextual factors. For example, recent policies that attempt to “de-police” schools due to calls from Black Lives Matter by removing security guards, SROs, or school police may be dealt with differently by region and community, resulting in divergent outcomes and mixed results. Similarly, restorative justice programs may not work well in specific contexts but work very well in others. Related resources, training, and other systems designed to support victims of violence at school may influence the success of policies in different contexts. The implementation or removal of policies needs to be explored more in future research.
Identifying What Processes are Responsible for Policy Formation and Changes
As previously discussed, watershed events, political changes, media, and time anchors need to become part of school safety research so that findings are linked to events and actual policy. This is particularly important for researchers interested in influencing policy through empirical studies. A good example of this is the massive decrease in most if not all types of violent events in U.S. schools starting in the early 2000s. Media coverage was intense and never subsided on issues of school shootings. Policymakers started passing laws and making funds available for programs and interventions, including law enforcement. However, this narrative is not captured by research or studies. Watershed moments and the processes that follow need to be included in the theoretical foundation, research, methods, and analyses of studies. Because these events have such a potentially massive impact on all other systems, it is important to include these ecological layers in studies.
Anticipating the Need for New or Modified Policies and Conducting Informative Studies
COVID-19, for example, has been a theoretical and empirical challenge because researchers did not integrate the possibility of pandemics, which affect entire countries and the world at all socioecological levels, as part of school safety research. Very quickly, it became evident that the role of the school was central to the functioning of families (e.g., work) and society (e.g., economy and student mental health). Social battles over masks, vaccines, and online or hybrid instruction became issues of safety, not only for the students but also for school staffs and families. Safety will be an issue when the pandemic subsides. Questions about the school context may be critical for both policymakers and the validity of school safety research studies. Research tools and methods that capture these outer ecological layers would link what is being researched in schools with what is happening in the outside world, making it relevant to policy. Similarly, policies that perhaps made teachers or families feel unsafe could be integrated and captured so research is more relevant to policy. Other potential historic changes in the horizon may pertain to school safety yet are currently not in the school safety research literature. For example, climate change, mass migration, and climate refugees could all affect school safety. Can researchers alter existing theories and methods to anticipate and capture these historic context changes? This is possible, but how society thinks about schools—especially how schools are nested in global society—needs to be altered first.
Designing Studies That Examine Cross-Level Interactions
Although called for by socioecological theory, it is rare to find studies employing cross-level interactions that aim to identify, for instance, how certain groups of students are affected differently by schools with different characteristics. Similarly, it is difficult to find school-focused mixed-methods research, case studies, or qualitative work on safety in schools with different characteristics, especially how schools or school systems have grappled with and overcome issues. Many studies have focused on students and their voices and narrative, which is always informative but doesn't address the school organization. This would be very valuable work that would inform policy.
Studying the Training and Professional Development of Educators Regarding School Safety
Universities train school professionals and are engaged in the lifelong professional development of all school personnel—teachers, psychologists, social workers, counselors, and others. As such, they have the responsibility and opportunity to provide information and guidance on existing school safety policies. As an example, recent studies exploring military-connected civilian schools described a process that educates undergraduate, master's, and doctoral students about existing policies affecting these schools so that they could help districts find ways to take advantage of resources made available through these policies. Furthermore, part of the training included acquiring knowledge and skills in advocacy and community practice to change policies, so graduate students could better address the safety and other needs of students, teachers, and parents (Berkowitz et al., 2014; Cederbaum et al., 2014; Esqueda et al., 2014). The data collected from these studies were published in dozens of op-eds, research studies, and mass media and helped orchestrate two White House meetings and multiple congressional briefings advocating for new resources, laws, and supports for military students in civilian schools (Astor et al., 2021). The university and its students played a major role with national partners such as Congress, the Obama Administration, the Military Child Educational Coalition, and Department of Defense Educational Activity to accomplish the policy outcomes. Currently, the military student identifier in the Every Student Succeeds Act and the Military Family Compact is present in all 50 states, offering an array of resources for schools that serve military families (Esqueda et al., 2012).
The activities described in this article, which are intimately tied to research, can be framed as service or they can be documented, measured, described, and included in scholarly agendas as both independent and dependent variables. The more policy is theoretically, empirically, and educationally considered in social work and social sciences school safety scholarship, the more this kind of research can influence policy at the national, state, and regional levels.
