3
Mental, Emotional, and Behavioral Health Effects of School Active Shooter Drills

The committee was tasked with examining the potential short- and long-term mental, emotional, and behavioral health effects of school active shooter and lockdown drills, as well as other school security measures (e.g., metal detectors, police presence, locked doors/controlled access) on students and school staff. Given that both the nature and volume of empirical evidence on school active shooter drills differ significantly from what is available for other school security measures, this chapter focuses only on literature related to school active shooter drills. Specifically, the literature reviewed in this chapter includes all empirical studies identified by the committee that examine the responses of K–12 students and school staff to these drills.

The chapter begins by outlining the committee’s inclusion criteria for the literature reviewed. It then presents an overview of the available evidence, including the methodological challenges that have contributed to gaps in the research base. The chapter next examines the mental, emotional, and behavioral health outcomes associated with school active shooter drills, detailing the research designs of the identified studies and reviewing the empirical findings on key outcomes of interest as identified by the committee. These include perceptions of emergency preparedness, perceptions of safety or fear, negative mental health symptoms (including traumatic stress), and student behavior. Additionally, the committee highlights insights from qualitative studies, which offer critical

perspectives on student and staff reactions to school active shooter drills, complementing the findings from quantitative research.

In Chapter 4, the committee continues with a review of evidence regarding the mental, emotional, and behavioral health effects of other school security measures—specifically, metal detectors, law enforcement presence, and locked doors.

INCLUSION CRITERIA AND SEARCH STRATEGY

This chapter consists of a narrative literature review and assessment of the available research base regarding the health impacts of school active shooter drills. Studies included in this review met the following criteria: they contained empirical data; assessed the mental, emotional, or behavioral health impacts of drills; examined health impacts among K–12 students or school staff; and were published in English and accessible by the conclusion of the search period (January 15, 2025). The committee used standard academic search procedures to identify relevant studies, including keyword searches of research databases, consultation of key summative texts (e.g., Schildkraut et al., 2024b; Walsh et al., 2023), reverse-citation lookups within included studies, and input from experts in the field. Additionally, the committee reviewed an annotated bibliography created by the U.S. Department of Health and Human Services (2025) in January 2025 and confirmed that all studies referenced in that document had already been identified in the committee’s search.

The committee also acknowledges several adjacent categories of studies that were not included in this review. Dissertations, master’s theses, and other unpublished student work were excluded. Also excluded were studies conducted in school contexts outside of K–12 settings (e.g., college campuses), those focused on groups other than students and school staff (e.g., parents), and those examining types of drills not specified in the statement of task (e.g., fire drills). Studies on active shooter drills in nonschool contexts (e.g., hospitals) were excluded as well given the distinct developmental and contextual considerations pertinent to K–12 schools. Finally, the review

excluded studies assessing the effectiveness of drills from a security standpoint, specifically those evaluating whether drills reduce morbidity and mortality during a simulated or actual school shooting event.

The committee identified some studies aimed at investigating the effectiveness of drills using computer simulations or after-action analyses of real-world active shooter events. Such research is difficult to carry out with precision given the statistical rarity of school shootings and the many uncontrolled variables involved in such an event. Moreover, the committee emphasizes that (a) the security effectiveness and (b) the mental, emotional, and behavioral health outcomes are two distinctly different dimensions of drills that may be related or entirely unrelated; they must therefore be considered separately. The committee’s review does not take up the question of “security effectiveness” here; rather, the focus is limited to mental, emotional, and behavioral health outcomes for students and staff. School policymakers must balance information about potential harms of drills such as that presented in this report against the potential benefits of drills in light of their individual context and circumstances.

OVERVIEW OF THE AVAILABLE RESEARCH BASE

As noted in Chapter 1, empirical research on the mental, emotional, and behavioral health effects of active shooter drills is limited. Only a small number of empirical studies—fewer than two dozen in total—have investigated how drills affect K–12 students or school staff. In many ways, this limited research base is surprising—millions of students and school staff across the country participate in an intervention whose the health effects are largely unknown. On the other hand, the lack of robust research in this area is perhaps to be expected. Active shooter drills in schools appear not to have been a priority for federal research funding, as none of the papers included in the committee’s review list a federal funding source. This may be related to the 1996 Dickey Amendment, which banned

federal funding for research related to gun control from 1996—prior to the widespread implementation of active shooter drills in K–12 schools—until 2018. While the Dickey Amendment did not specifically ban research on gun violence or school active shooter drills, its interpretation resulted in significant restrictions on research addressing firearm-related injuries and deaths (Rostron, 2018). In 2019, Congress allocated $25 million to the Centers for Disease Control and Prevention and the National Institutes of Health to study firearm violence as a public health issue, which marked the first federal funding for such research in 2 decades. Furthermore, most research on school active shooter drills necessitates partnerships with a school or school district, and all research partnerships with schools entail practical, financial, and political considerations. Given these complexities, researchers may face barriers to engaging schools, as participation in such research can require significant time, energy, and resources and may raise concerns about undermining their current safety policies.

Within the small group of included studies, the mental, emotional, and behavioral health outcomes assessed vary and are limited, ranging from knowledge of drill procedures to perceptions of school safety, to feelings of fear and anxiety, to self-reported behavioral changes such as days of school missed. Many studies do not specify the precise nature of the drills assessed, making it difficult to distinguish whether different types or features of emergency preparedness procedures (e.g., didactic instruction vs. controlled rehearsal vs. realistic simulation) might have different health outcomes. Available studies used a range of methodologies, but virtually none employed the most rigorous research designs (e.g., random assignment to experimental vs. control groups), which would allow for the most confident conclusions about causation. Most studies had significant limitations—including small sample sizes, ad hoc outcome measures, and unmatched pre–post groups—that reduced the strength of conclusions that can be drawn.

Another striking characteristic of this body of research on active shooter drills is that more than one-third of all the empirical studies were produced by a single research team that assessed an

anonymized school district (or districts) in central New York using the same training and lockdown drill paradigm. Thus, when considering the strength of the entire research base on mental, emotional, and behavioral health outcomes of active shooter drills, it is important to highlight that many findings are drawn from the same geographic region, possibly from the same school district. This is a limitation of the research base, as these findings may not generalize to or be representative of the varied contexts of school systems across the country.

Unsurprisingly, given the small total number of empirical studies, there are many important issues regarding active shooter drills for which no studies exist. As one key example, aside from a single qualitative study of teachers’ perceptions, no peer-reviewed studies have assessed how students with functional or access needs (e.g., students with disabilities, students with sensory sensitivities, multilingual learners) react to drills.

Thus, while the following section presents a careful discussion of the available empirical research related to the committee’s statement of task (see Box 1-1 in Chapter 1), the committee emphasizes that the gaps in knowledge about the mental, emotional, and behavioral health effects of school active shooter drills overwhelm the limited research base. Considering that millions of students and school staff participate in these drills each year, more high-quality research in this area is urgently needed.

MENTAL, EMOTIONAL, AND BEHAVIORAL HEALTH OUTCOMES OF DRILLS

Many sources (e.g., National Association of School Psychologists et al., 2021; Schonfeld et al., 2020) have expressed concerns that active shooter drills could increase fear and anxiety and decrease perceptions of school safety among students and staff. These sources hypothesize that the experience of the drills themselves—especially realistic or high-intensity drills that convincingly simulate a real active shooter event—may be distressing. Even low-

intensity drills may cause participants to perceive a genuine active shooter event as more likely, increasing fear or decreasing perceptions of school safety. Conversely, some have argued that active shooter drills may increase perceptions of school safety for some students and staff by making them feel more prepared in the event of an actual active shooter event (Cunningham, 2020; Moore-Petinak et al., 2020).

To shed light on these important questions, the committee reviewed the limited empirical research base, identifying around two dozen studies that assessed some aspect of the personal reactions of K–12 students or school staff to active shooter drills (e.g., emotions, perceptions, self-reported behaviors).¹⁷ Most of these studies assessed student reactions, with just a few assessing staff reactions. Most involved direct quantitative surveying of students or staff. One study involved analyzing a very large sample of social media posts to assess changes in language (e.g., words associated with anxiety, stress, or depression) from before and after a drill. The remaining handful of studies involved collecting qualitative responses via individual interviews, focus groups, or written responses to a text message poll. Studies differed in the constructs they measured (e.g., stress, anxiety, well-being, avoidance), as well as in the instruments they used to assess these constructs (e.g., unvalidated, ad hoc instruments created by the research team vs. more standardized, “brand name” instruments assessing well-studied constructs such as state anxiety).

Research Designs

One factor hampering the already limited research base on the mental, emotional, and behavioral outcomes of active shooter/lockdown drills is the relative weakness of many of the research designs. In general, the research design sets a ceiling for a study’s ability to make causal attributions between drills and any ensuing participant

___________________

¹⁷ This total count is approximate because some studies are described in two different manuscripts, and some manuscripts contain two separate studies (e.g., a qualitative and a quantitative study).

reactions. To be clear, ethical and practical constraints may make it challenging—or even impossible—to implement the most rigorous research designs (e.g., randomized controlled trials). But even if the limitations of the existing research base are understandable in light of these barriers, it does not change the fact that the committee’s confidence in the meaningfulness, precision, and generalizability of the findings of existing studies are limited as a result. Existing studies fall into four major categories of research design, which are described briefly in this section in rough descending order of rigor with respect to causal attribution (see Table 3-1).

TABLE 3-1 Overview of Studies About the Effects of Active Shooter Drills

Control Group Designs

Control group designs have the greatest potential to enable causal attributions as to any effects drills may have on participants’ emotions or perceptions. Random assignment to an experimental group versus a control group means that any observed differences between the groups are likely attributable to the drill rather than to unrelated events, unmeasured differences between the groups, or other confounding factors.

There are certainly challenges to conducting effective control group studies assessing reactions to active shooter drills. For example, states or school districts may have rules stipulating that all students must complete drills, drills may initiate a full-school response that must involve all students in the school, and most students will already have completed dozens of active shooter drills of some sort by the time researchers begin studying their reactions. However, these challenges can be overcome to some extent with thoughtful research designs (e.g., a modified waitlist-control procedure, in which schools within a district implement their drills on different schedules to facilitate comparisons).

Perhaps because of ethical and practical barriers, only one study (Zhe & Nickerson, 2007) employed a control group design. And that study had a notably small sample size of only 74 students.

Pre–Post Designs

Pre–post designs assess the same set of participants both before and after participation in a drill. Pre–post studies do not allow for the degree of causal attribution enabled by control group studies. For example, participants in a pre–post study may be aware of the intended outcome of the drill and respond in a manner intended to fulfill the researchers’ expectations. Nonetheless, pre–post designs successfully eliminate many potential sources of error by comparing participants against themselves. As a result, at least some potential confounding factors (e.g., students’ demographics, early life experiences, personality traits, and mental health conditions) are

eliminated. This advantage dissipates, however, when the pre- and post-drill groups are otherwise not identical (e.g., when participants drop out or new participants are added between the pre- and post-drill measurements).

Ten studies used pre–post designs to assess student or staff responses to active shooter drills. The studies tended to use very large samples (thousands of students and/or staff) across the school district; unfortunately, responses were anonymous because of school privacy requirements and so could not be linked pre- and post-drill. Furthermore, there were often large differences between the pre- and post-drill groups, across which all responses were aggregated. In many studies of student reactions, for example, the pre-drill surveys were placed in teacher mailboxes to administer to students and only a fraction of students completed them; similarly, only a fraction (presumably a different fraction) of students completed the post-drill surveys. As a result, statistical analyses had to treat the pre- and post-drill groups as completely separate (i.e., cross-sectional), significantly limiting confidence that any post-drill changes must be attributable to the drill rather than to changes in the composition of the groups being surveyed or unrelated events occurring in students’ lives when the baseline measures were administered approximately a week prior to the drill.

Cross-Sectional Designs

Cross-sectional designs collect data from participants at a single point in time. This design allows researchers to calculate statistical correlations between theorized predictors and outcomes of interest. However, given the maxim that “correlation does not imply causation,” cross-sectional designs cannot establish that a theorized predictor caused an outcome even if a statistical correlation is found. Causation may run in the opposite direction, or an unmeasured third variable may be driving both the theorized predictor and the outcome of interest. Therefore, cross-sectional designs are lower on the hierarchy of methodological rigor when it comes to establishing causal links between active shooter drills and mental, emotional, and behavioral health outcomes. Correlational designs can be

strengthened using statistical methods to control for other variables that may be driving an association between the theorized predictor and outcome.

Seven cross-sectional studies assessed student or staff perceptions and emotional reactions following active shooter drills. However, most of these cross-sectional studies did relatively little to control for potential confounds between groups, and thus the ability to draw a causal inference is particularly suspect. For example, Huskey and Connell (2021) surveyed 379 undergraduate college students and found that those who reported participating in at least one active shooter drill in high school also reported increased fear, increased perceptions of risk, and decreased perceptions of school safety compared with undergraduates who did not report participating in a drill. However, given the lack of statistical controls for confounds, it is possible that undergraduate students who were already more fearful or had greater concerns about school shootings were simply more likely to recall participating in active shooter drills (i.e., “reverse causation”).

Qualitative Designs

Qualitative research designs allow participants to provide open-ended responses about a subject of interest (e.g., via interviews, focus groups, or writing prompts). These kinds of open-ended responses cannot establish causation or statistically significant associations between variables. Indeed, qualitative research is not designed to furnish representative data on how frequently negative reactions to drills occur or how often they are related to other variables of interest. Instead, qualitative designs can yield rich, nuanced information about a problem, highlight individual stories that would be lost in aggregated quantitative data, and elicit ideas about solutions to problems. Commonly, qualitative research is used to help researchers gather preliminary information about understudied topics, generate new hypotheses about potential causal mechanisms, or elicit “raw material” that can be converted into quantitative survey content. Five qualitative studies assessed student or staff reactions

to active shooter drills, using individual interviews, focus groups, or open-ended text message polls.

Committee’s Observations

Overall, the methodological rigor of the studies identified by the committee was low. Only one study (Zhe & Nickerson, 2007) used a control group design, and only two studies (Nickerson & Schildkraut, 2024; Zhe & Nickerson, 2007) used well-validated outcomes measures as opposed to ad hoc instruments with unknown reliability and validity. While the committee recognizes researchers’ efforts to undertake any empirical investigation on this important yet understudied topic, the small number of studies and relative weakness of the research designs limit the ability to draw strong causal inferences about the effects of drills. Additionally, many studies had additional methodological weaknesses beyond their research design, such as short follow-up periods, ad hoc outcome measures, unmatched pre–post groups, and lack of precision about what was considered a drill.

Review of Empirical Studies

With the above caveats about the quantity and strength of the research base in mind, the results of the available empirical studies differed depending on the study design, sample, type of drill, and outcome measures. Some studies showed positive reactions to drills (e.g., feelings of increased confidence about responding to emergencies); some showed no changes in emotions or perceptions; and some showed negative reactions to drills (e.g., increased depression or anxiety; however, only two studies assessed anxiety and depression using standardized measures with students in conjunction with actual drill participation [Nickerson & Schildkraut, 2024; Zhe & Nickerson, 2007]). In the discussion that follows, studies with quantitative measures are organized according to the type of outcomes assessed: (1) emergency preparedness; (2) perceptions of safety or fear; (3) mental health symptoms, including traumatic stress; and (4) self-reported behavior. Some studies assessed multiple types of

outcomes falling into more than one of these categories and were therefore discussed more than once. Qualitative studies are discussed separately, given that participants providing open-ended reflections on drills gave responses touching on many types of outcomes.

Drills and Emergency Preparedness

Implementation

Correct implementation of drill procedures is not in itself a health effect. However, basic data on how well a drill taught the intended knowledge and skills could be crucial backdrop for understanding the drill’s health effects. For example, if drills have harmful effects, conducting them at the minimum frequency required to teach the correct procedures adequately might mitigate those effects. But this kind of fine-tuned adjustment requires knowing how quickly students learn the correct procedures and how frequently drills must be repeated to maintain the desired level of skill.

Four studies assessed implementation issues regarding active shooter drills. Zhe and Nickerson (2007) randomly assigned 74 students in grades 4–6 in an upstate New York suburban school district to one of two groups: (1) students receiving a brief training on emergency procedures and participated in one “intruder crisis drill” (30 minutes total) or (2) a control group. This study found that students who received the training and participated in the drill demonstrated greater knowledge of correct drill procedures on a five-question multiple-choice questionnaire compared with students in the control group. Observation of the group that received the training showed that, during the drill, all students successfully moved their desks to a safe location in the room in under 2 minutes (the first compliance criterion), but no classroom remained entirely silent for 7 minutes during the drill (the second compliance criterion). These results suggest that a single brief training on drill procedures imparts some knowledge and skills but is insufficient to achieve the desired level of skill acquisition.

The remaining three studies on implementation issues used pre–post designs. Researchers assessed participants’ implementation

skills at baseline (i.e., before receiving any training), provided didactic training on correct drill procedures, and conducted a series of monitored drills to assess whether drill implementation skills improved following training and practice.

Dickson and Vargo (2017) used this pre–post design to assess drill implementation among 32 kindergartners at a single Texas elementary school in a low-income area. The full research protocol involved a series of 13–25 observed drills conducted at a frequency of two to three per week. Results indicated that prior to any training, the kindergarteners could complete zero correct lockdown steps and remained noisy during drills. The kindergarteners required five to seven training sessions to attain adequate skill acquisition (around 80% of correct lockdown steps). They were able to maintain this level of implementation fidelity for three additional drills conducted over the course of approximately 1 week, with no refresher trainings. This result demonstrates short-term retention of skills for even the youngest students, although long-term retention was not assessed.

Schildkraut and Nickerson (2020) and Schildkraut et al. (2023) conducted two studies on implementation issues. Both took place within a large urban school district in central New York that was transitioning to a standardized, district-wide emergency response program (SRP-X). Because these studies assessed compliance with correct drill procedures classroom by classroom, they can be considered an assessment of both student and staff skill acquisition. Both studies involved averaging large sample sizes (thousands of students and staff) at each timepoint.

The first of these implementation studies (Schildkraut & Nickerson, 2020) demonstrated that a single training session significantly increased correct implementation of lockdown procedures but was insufficient to reach full compliance. Classrooms demonstrating perfect compliance increased from 27.7% pre-training to slightly more than half (54.4%) post-training. After a single training session, each classroom demonstrated approximately 85% compliance for three of the four required behaviors (lock door, lights off, no responses to door knock) and somewhat lower compliance (around

70%) for staying out of sight. Although the researchers did not conduct statistical analyses comparing compliance across grade levels, it appears that dedicated middle schools (grades 6–8) had the lowest average compliance compared with elementary schools, pre-K¹⁸–8 schools, and high schools (Schildkraut & Nickerson, 2020).

The second of these implementation studies (Schildkraut et al., 2023) yielded similar results, although this study improved upon the first by observing nine drills spread across 4 years, rather than just a single post-training drill. While the other three implementation studies were conducted across just a few months, this was the only study assessing long-term retention of skills across years. As in the prior study, classroom compliance rose quickly to high levels (approximately 85% or higher) after just one training, for all steps aside from staying out of sight (71.3%), which apparently is the most difficult step for classrooms to complete successfully. Encouragingly, correct implementation of skills was maintained or even increased across subsequent drills conducted over the next 3 years, even without additional trainings. Classrooms with perfect compliance increased from 54.4% after the first post-training drill (year 1) to 81.1% after the fourth post-training drill (year 3). Perfect compliance then remained steady, hovering around the 80% mark, for the remaining five post-training drills through year 4 (Schildkraut et al., 2023).

Schildkraut et al. (2023) demonstrated that relatively high levels of implementation compliance can be achieved and maintained over a duration of years even without repeated training, perhaps because teachers and students provide informal instruction and feedback to each other following drills. However, although 80% of classrooms demonstrating perfect compliance is an encouraging result, it means that one of five classrooms (20%) were still failing to complete all four required lockdown steps after 4 years of practice.

___________________

¹⁸ The committee included pre-K students only when these students attended a school that also contained K–12 students. Some elementary school campuses include pre-K classrooms whose students may participate in school active shooter drills.

In sum, the four existing studies on implementation of drills show that even a single training can quickly improve compliance with the required drill procedures. However, repeated training or drills (around four to seven implementations) are necessary to reach adequate—though not perfect—compliance. Some elements of lockdown procedures appear to be more difficult than others for classrooms to perform correctly in practice. Keeping out of sight and staying quiet appear to be most difficult for K–12 students and kindergartners, respectively.

Perceptions of Emergency Preparedness While just four studies measured participants’ implementation of correct emergency procedures during actual drills, other studies assessed perceptions of emergency preparedness.¹⁹ Six pre–post studies by Schildkraut and colleagues captured brief, ad hoc measures of student or staff perceptions of emergency preparedness (Schildkraut & Nickerson, 2020, 2022; Schildkraut et al., 2020, 2022, 2024a,c). Participants rated agreement with the statement “I know what to do” in response to each of the SRP-X training program’s scenarios: (1) lockout (now called secure), (2) lockdown, (3) evacuate, (4) shelter, and (5) hold.

Across these six studies, the researchers generally found that perceptions of emergency preparedness (“I know what to do” in different emergency scenarios) increased among both students and staff following training and participation in a drill (Schildkraut & Nickerson, 2020, 2022; Schildkraut et al., 2020, 2022, 2024a,c). In the subset of studies that reassessed perceptions following a second drill (Schildkraut & Nickerson, 2020; Schildkraut et al., 2020, 2022, 2024c), findings showed that student and staff perceptions of emergency preparedness continued to improve after the second drill. In other words, perception of students and staff that they “know what to do” did not always reach their zenith following just a single

___________________

¹⁹ These different research questions were at times contained within the same published article. For example, Schildkraut & Nickerson (2020) contains data both about classrooms’ implementation of emergency procedures during actual drills (i.e., how many classrooms closed their doors) and about student and staff’s perceptions that they “know what to do” in response to various emergency scenarios.

training or drill, and generally continued to rise following a second drill. This result again suggests that a single training or drill is insufficient.

Furthermore, Schildkraut et al. (2024c) demonstrated that perceptions of emergency preparedness did not change significantly following a single “drill as usual,” but did increase following a formal training on a Standard Response Protocol and then an additional drill. This result suggests that providing training on how to respond to emergency scenarios such as active shooter events—rather than just the act of completing a drill itself—may be important to improving feelings of emergency preparedness.

Moore et al. (2024) collected a much broader dataset to assess teachers’ perceptions of their own and their students’ emergency preparedness. They surveyed a randomly selected cross-sectional sample of 1,020 K–12 teachers and then weighted this sample to make it representative of the national population of K–12 public school teachers. Among the 90% of teachers who reported participating in at least one active shooter drill during the 2022–2023 school year, responses on emergency preparedness were divided: About half (47%) indicated that drills made them feel more prepared to respond to an active shooter incident, while the other half (50%) indicated that drills made them feel neither more nor less prepared. Rural teachers were more likely than their urban counterparts to indicate that drills made them feel more prepared to respond to an active shooter incident (Moore et al., 2024). The survey also asked teachers to comment on their perceptions of how drills affected students’ sense of emergency preparedness, although of course teachers’ perceptions of their students feelings may be inaccurate. About half (54%) of teachers indicated that they perceived that drills helped students feel more prepared, while one-third (32%) thought drills had no impact on students’ feelings of emergency preparedness. Furthermore, 11% of teachers said they did not know how drills affected students’ feelings of preparedness. Almost no teachers indicated that drills they made themselves or their students feel less prepared (Moore et al., 2024).

Taken together, the results of the Schildkraut et al. and Moore et al. studies suggest that on average, drills increase student and staff perceptions of emergency preparedness to some extent. However, averaging perceptions across all participants may mask the apparent finding that large portions of participants (e.g., fully half of all teachers in Moore et al., 2024) felt that drills had no impact on their perceptions of emergency preparedness. While these seven studies shed some light on the issue of perceptions of emergency preparedness, considerable uncertainty remains. For example, participants agreeing that they “know what to do” in the scenarios aligning with the SRP-X training may simply indicate that they effectively learned the responses they were taught, not that they perceive those prescribed responses will protect them in a real active shooter event. More detailed research is needed on student and staff perceptions of emergency preparedness and the specific practices that contribute to perceptions of preparedness.

Drills and Perceptions of Safety, Fear, or Risk

Some have argued that active shooter drills may make students feel more fearful or perceive their schools as less safe (Huskey & Connell, 2021; Moore-Petinak et al., 2020). Drills may cause students to imagine themselves in a life-threatening school shooting event or may create an impression that an active shooter event is likely to occur. Multiple studies with a variety of research designs assessed these kinds of outcome measures.

Just one previously discussed study in K–12 schools (Zhe & Nickerson, 2007) used a control group design to assess student reactions to active shooter drills. As a reminder, the researchers randomly assigned a small group of 74 of students in grades 4–6 to one of two groups: (1) a group in which students received a brief training on the correct procedure for an “intruder crisis drill” and participated in one drill (30 minutes total) or (2) a control group. The researchers assessed students’ perceptions of school safety (including anxiety about school violence) via a 10-item, ad hoc adaptation of two existing measures (the School Violence Anxiety Scale [Reynolds, 2003] and the School Crime Supplement to the National Crime

Victimization Survey [National Center for Education Statistics & Bureau of Justice Statistics, 2001]). The researchers also assessed students’ state anxiety, which is covered in the following section assessing mental health outcomes. Students in the drill group completed the questionnaires immediately after completing the training and drill.

Researchers found no statistically significant differences between the two groups regarding perceptions of school safety, which suggests that a carefully conducted active shooter drill did not make 4th through 6th graders more concerned about school violence when using a control group design. However, the small sample size of 74 would have precluded detection of anything smaller than medium-to-large effects (Cohen’s d = 0.66 at 80% power and alpha = 0.05 two-tailed), even though such large differences are relatively rare in research on educational interventions (Hill et al., 2008).

Six pre–post studies (also discussed above) assessed perceptions of school safety, fear, or perceived risk (Schildkraut & Nickerson, 2020, 2022; Schildkraut et al., 2020, 2022, 2024a,c). Perceptions of school safety were assessed in all six of these studies via a series of brief, ad hoc items. Students or staff rated agreement with the statement “I feel safe” for four different locations: (1) “at my school,” (2) “in my classroom,” (3) “in the hallway,” and (4) “outside on school grounds.” In three of the four studies assessing student samples, the researchers showed that students’ perceptions of school safety decreased following participation in a series of two lockdown drills (i.e., lower ratings of “I feel safe” in various school locations) (Schildkraut & Nickerson, 2022; Schildkraut et al., 2020, 2024c). In contrast, the fourth of these studies (Schildkraut et al., 2022) showed that student perceptions of school safety did not change following participation in one active shooter drill. Thus, although students reported feeling more prepared following participation in drills in these studies, they generally also reported feeling less safe (or at least not more safe). The only study directly assessing school staff’s perceptions of school safety found that these perceptions did not change following a series of two lockdown drills (Schildkraut et al., 2022).

Schildkraut et al. (2024c) merits noting because this study sought to determine how a key student variable—prior exposure to violence—may influence the relationship between perceptions of school safety and participation in repeated drills. The researchers found that, across their entire pooled sample of three cross-sections comprising 8,627 students in grades 6–12, higher exposure to violence was associated with lower perceptions of school safety. The researchers also found that this relationship weakened over time (i.e., after students participated in two drills and one training on a Standard Response Protocol). The researchers speculated that the effect of prior exposure to violence on perceptions of school safety may matter less once students have been trained and rehearsed on lockdown procedures (Schildkraut et al., 2024c).

Two of additional pre–post studies by Schildkraut and colleagues measured students’ feelings of fear and perceived risk of victimization, although in slightly different ways. These fear and risk variables are distinct from the items inquiring about a more generalized sense of school safety (i.e., “I feel safe” in various school locations) described above, and they returned contradictory results. First, for about 10,000 students in grades 6–12, Schildkraut and Nickerson (2022) assessed their (1) fear that they would be harmed at school or traveling to/from school and (2) perceived risk that “I think a school shooting could happen at my school.” This study found that both fear and perceived risk decreased significantly after students participated in a series of two lockdown drills (Schildkraut et al., 2022). Second, Schildkraut et al. (2024a) assessed approximately 300 high schoolers’ perceptions in 2022, including (1) fear of a school shooting happening at their school during that year and (2) perceived fear of being a victim of a school shooting. In contrast to the study by Schildkraut and Nickerson (2022), which found reduced fear and perceived risk post-drill, this study showed that students’ perceived risk of victimization increased significantly after participating in one drill, although fear did not change post-drill (Schildkraut et al., 2024a).

In sum, evidence from the control group and pre–post studies is mixed and limited regarding the commonly raised concern that

participating in active shooter drills may make students feel more afraid or perceive a school shooting as more likely. Some studies showed post-drill perceptions of fear and/or risk increasing, some decreasing, and some remaining unchanged. Given that many of these studies used similar student populations within the same central New York area, it is unclear why students reacted differently to drills in different studies.

Several cross-sectional studies also assessed outcome measures related to perceptions of school safety, fear, or perceived risk. Again, the committee emphasizes the limited ability of cross-sectional studies to establish causality, as well as typically limited or missing information about what constituted a “drill” across these large, in some cases national, cross-sectional surveys. In other words, students and teachers may be responding to meaningfully different drill procedures (e.g., high-sensorial vs. not) when responding to survey items. For example, Moore et al. (2024) surveyed about 1,000 K–12 teachers and did inquire about realistic, sensorial elements in drills. Findings indicated that few teachers reported high-sensorial elements during the 2022–2023 school year: 6% reported that drills included sounds of real gunfire; 5% reported police or security officers firing blanks; 4% reported prop or toy firearms; 1% reported the sounds of explosions; and 1% reported the use of fake blood. Most teachers (55%) reported no realistic elements during drills, which included no presence of police or other uniformed security personnel inside the school (the most common “realistic” element reported, at 44%). Unfortunately, Moore et al. (2024) did not analyze teachers’ perceptions of emergency preparedness or school safety according to the number or type of sensorial elements, so we cannot know the effect of these realistic drill components on teachers. With these caveats, we describe the results of cross-sectional studies about school safety, fear, or perceived risk:

• Moore et al. (2024) surveyed about 1,000 youth12 teachers (as previously discussed) to inquire about sense of school safety. A majority of teachers (69%) indicated that drills had no impact on their sense of school safety. Of the remainder, 20% of teachers indicated that drills made them feel safer at

• school, while 12% indicated that drills made them feel less safe at school. Rural teachers were more likely than their urban counterparts to report that drills made them feel safer at school. The split was roughly similar when teachers were asked about their perceptions of how drills affected their students’ sense of safety, although of course teachers’ perceptions of student responses may be inaccurate: 38% of teachers thought drills had no impact on students’ sense of safety; 24% thought drills made students feel safer, and 20% thought drills made students feel less safe. The remainder of teachers said they did not know how drills affected students’ sense of safety.
• Huskey and Connell (2021) surveyed 379 undergraduate students and found that those who reported participating in at least one active shooter drill in high school also reported increased fear, increased perceptions of risk, and decreased perceptions of school safety compared with undergraduates who did not report participating in a drill.
• Rauk et al. (2023) used nationally representative data collected from more than 2,000 adolescents to investigate how students’ race (students of color compared with White students) influenced associations between reported school safety policies and perceptions of school safety. They found that students of color who reported a greater number of total safety policies (including active shooter drills) at their school also reported lower perceptions of school safety, but White students showed no association between these variables. When only active shooter drills were considered, however, there was no moderating effect of students’ race on perceptions of school safety.
• Jonson et al. (2020) surveyed approximately 1,250 students in grades 4–12 from a single Midwestern school district following the implementation of discussion-based exercises used to teach the ALICE multi-option protocol (see Chapter 1 for a description of the ALICE protocol). Using two surveys—one for elementary school (grades 4–5) and one for

• junior high and high school students (grades 6–12)—the researchers found that 10% of students reported increased negative emotions post-drill, including fear, worry, and confusion after teacher-led discussions of ways they may respond to a dangerous person entering the school using the ALICE protocol. In comparison, 85% of students reported either no emotional changes or increased positive emotions, including feelings of safety, confidence, and emergency preparedness after discussions about the ALICE protocol.²⁰

Overall, these cross-sectional studies yielded mixed results regarding student and staff perceptions of school safety, fear, or perceived risk following active shooter drills. In very rough terms, it appears that the most common outcome is no change to perceptions of safety, fear, or risk due to drills.

Drills and Mental Health Symptoms, Including Traumatic Stress

Few studies have investigated whether drills contributed to more significant or long-lasting mental health symptoms such as anxiety, depression, or traumatic stress. Virtually all used brief, but previously published and well-validated self-report questionnaires about anxiety, depression, traumatic stress, or well-being.

The single control group study was Zhe and Nickerson (2007) (discussed previously), assessing 74 students grades 4–6. The researchers assessed changes in students’ anxiety post-drill via the 20-item A-State Scale from the well-validated State-Trait Anxiety Inventory for Children (Spielberger, 1973). They found no statistically significant differences between the group who participated in a brief training and drill and the control group. The use of a control group design and a well-validated measure of state anxiety are significant and unusual strengths for this study, which suggests that a carefully

___________________

²⁰ For questions related to fear, respondents could choose between “less scared,” “no change,” or “more scared.” Similarly, for questions about school safety, respondents indicated whether training made them feel “less safe,” whether there was “no change,” or whether they felt “more safe.” Responses were collapsed into two categories: 0 = no harm (i.e., no change, more safe, less fear) and 1 = harm (less safe, more fear).

conducted active shooter drill did not make 4th- through 6th-graders more anxious in the moment (Zhe & Nickerson, 2007). Aside from the small sample size of 74, which precludes detection of anything less than medium-to-large effect sizes, we also note that this study was conducted almost 20 years ago. It is possible that the greater number and cultural salience of school shootings since 2007 may have changed students’ perceptions of drills in recent years.

Just one of the pre–post studies by Schildkraut and colleagues included mental health variables (Nickerson & Schildkraut, 2024). After participating in an active shooter drill, 763 out of 925 students in a rural high school completed a brief measure of state anxiety (although only 610 students completed the pre-drill questionnaire, and there was no analytical linkage of individual students pre- and post-drill). State anxiety was measured with the 6-item short-form adaptation of the well-validated Spielberger State-Trait Anxiety Inventory (Marteau & Bekker, 1992). The researchers used exploratory factor analysis to divide the six items into two factors of three items each, which they labeled anxiety-present (“I am tense,” “I am worried,” “I feel upset”) and anxiety-absent (“I feel calm,” “I am relaxed,” “I feel content”). Post-drill, students showed small decreases in state anxiety (anxiety-present items) and small increases in well-being (anxiety-absent items), although students tended to show low anxiety and moderate well-being at both time points (Nickerson & Schildkraut, 2024). These results appear to be at odds with the consistent findings across Schildkraut and colleagues’ studies that students felt less safe following participation in active shooter drills. Idiosyncrasies in the changing samples, contexts, and metrics may explain this apparent contradiction.

Just three cross-sectional studies assessed student mental health variables. Riggs et al. (2023) surveyed 108 youth–caregiver pairs presenting to a single emergency department in a large, Midwestern metropolitan area. They found that approximately one-third of the youths reported increased depression and anxiety symptoms post-drill. The questions used for the assessment appear to have been drawn from the well-validated National Institutes of Health’s Patient-Reported Outcomes Measurement Information System

measures (HealthMeasures, 2023). Interestingly, the caregivers consistently perceived the youths’ post-drill anxiety and depression as less severe than was reported by the youths themselves (Riggs et al., 2023). This result suggests that parents may not always be able to detect when their children are bothered by drills. It should be noted, however, that this study was conducted during the COVID-19 pandemic when schools were closed, and survey participants were asked to reflect on their experiences with drills over the prior 2 years.

Moore et al. (2024) (described previously) asked teachers to comment on how they believe active shooter drills affected students’ mental health. In a national survey of around 1,000 K–12 teachers, about one-third of respondents (29%) indicated that “they were aware of students at their schools who experienced trauma or heightened anxiety and stress following their participation in a drill designed to prepare them for school shootings” (p. 21). Perceptions of student trauma, anxiety, or stress were higher among elementary and middle school teachers compared with high school teachers; higher among teachers who reported multiple drills in a school year compared with just one drill; and higher among teachers who reported more realistic elements during drills (although the use of any realistic elements beyond police presence was quite rare). However, just 1 out of 10 teachers reported that their school provided mental health supports to students after the completion of drills. Furthermore, only a minority of teachers reported that drills were designed to meet the needs of diverse students. Only 12%–26% of teachers believed that drills were designed to meet the needs of students with physical disabilities, developmental disabilities, emotional challenges, or prior traumatic experiences, although a somewhat greater percentage (40%) believed that drills were designed to address the needs of students in different age groups (Moore et al., 2024).

A smaller survey of preservice teachers (N = 97), most of whom were interning in schools at the time, also asked about their perceptions of the impact of active shooter drills on students’ mental health (Laguardia et al., 2024). Almost all the preservice teachers (89.7%) believed that drills can have an impact on students’ psychological

well-being, while a lesser percentage (58.8%) believed that drills can have an impact on students’ academic performance. The survey results indicate that preservice teachers “generally felt underprepared to address the psychological and emotional needs of students related to active shooter response training” and desired better training in this area (Laguardia et al., 2024, p. 194).

Another cross-sectional study assessed trauma symptoms among teachers. Stevens et al. (2019) surveyed approximately 300 teachers (preschool through high school) regarding associations between exposure to drills and secondary traumatic stress. The researchers prompted teachers to describe the elements of the drills in which they had participated (e.g., shelter-in-place vs. simulation components) and then created a sum score in which a higher score indicated a greater number of different drill components. There was no significant relationship between the number of different drill components reported by teachers and their reported degree of secondary traumatic stress (Stevens et al., 2019).

Finally, ElSherief et al. (2021) assessed post-drill mental health using a very different methodology (see also Everytown for Gun Safety, 2020). Rather than asking participants to complete questionnaires about mental health variables, this study used a complex webscraping scheme to assess how the language used in social media posts changed following school active shooter drills. According to the researchers, collecting voluntarily shared, publicly available social media data has many benefits over traditional survey methods. Researchers explain that this method “constitutes a promising opportunity to study psychological states unobtrusively and passively” when existing survey methods have shown little ability to capture rich, real-world data about mental health changes following crises (ElSherief et al., 2021, p. 3). The researchers also argue that this method can establish causation between an intervention and outcomes of interest, particularly when randomized control group designs may be “infeasible or unethical” (ElSherief et al., 2021, p. 3). This social media study is a clear outlier within the limited research base on school safety drills in terms of approach, sample size, and statistical attempts to control for confounds.

In their report, research team explains the complex methodology at length. Briefly, they first surveyed stakeholders involved in a gun violence advocacy group to identify the precise dates when recent active shooter drills had occurred in their children’s schools (114 schools spanning 33 states; ElSherief et al., 2021). They then used web-scraping procedures to compile more than 27 million Twitter posts and 1,400 Reddit posts that could be tied to one of the identified schools (e.g., via geography or following the school’s official Twitter account). These social media posts could be authored by school staff, parents, students, or community members and spanned 90 days pre-drill and 90 days post-drill. This is the only research the committee found that attempted to measure longer-term post-drill reactions. Finally, the researchers used lexical models to analyze changes in the language of social media posts pre- to post-drill (e.g., changes in the frequency of words statistically associated with stress, anxiety, depression, and concerns about death). The researchers used a variety of statistical comparison techniques to approximate causal inference, asserting that observed changes in social media posts are best attributed to the drills themselves, rather than unrelated seasonal or temporal changes.

Results indicated that active shooter drills are followed by social media posts showing a 39% increase in words associated with depression, a 42% increase in words associated with stress and anxiety, and a 23% increase in words associated with physiological health problems (e.g., blood, pain, clinics, and pills). Likewise, words associated with concerns about death increased 22% following active shooter drills. Other post-drill trends included increases in expressions of social and personal concerns, fewer positive statements, more requests for help, and increased profanity. These trends were sustained throughout the 90-day post-drill period, suggesting that the negative impacts of active shooter drills are not fleeting but last at least 3 months. Although all school levels (elementary, middle, and high school) demonstrated negative post-drill effects, social media posts associated with high schools showed the largest changes. The researchers theorized that, because most school shootings occur

at high schools, the perceived impact of drills may be greater within those schools (ElSherief et al., 2021).

In sum, ElSherief et al. (2021) purport to demonstrate consistent, pervasive negative emotional effects of drills on students, parents, and school staff via changes in social media posts. These findings are more consistently and strongly negative regarding the results of drills than the more mixed pre–post survey results generated by Schildkraut and colleagues. Numerous differences between the social media and survey methodologies may explain these differences (e.g., many states vs. a single school or school district; social media data vs. brief self-report measures; a 3-month follow-up period vs. immediate post-drill measurement). While ElSherief et al. (2021) argue that their methodology produces a more accurate, nuanced understanding of reactions to drills compared with traditional surveying, others may question the indirect way in which the researchers culled social media posts and associated the words in those posts with emotional outcomes. The results may have been skewed by selecting only schools identified by gun violence advocates, since these schools could have had exaggerated negative post-drill social media activity compared with the average American school. Furthermore, because the methodology captured all social media posts associated with a specific school, the researchers could not distinguish among posts by students, teachers, parents, activists, community members, or others. In other words, the social media analysis does not necessarily reflect the responses (or only the responses) of students and school staff who were exposed to drills. In addition, the wide variation in drill practices and protocols across K–12 schools, along with the inconsistent use of terminology to describe these activities, means the types of drills that occurred in the sampled schools are unknown. The researchers did ask the initial respondents who identified dates of local drills about high-sensorial elements (i.e., “Did the drill involve simulations that mimic an actual active shooter incident?”, see Table S-1 in the Summary). However, researchers did not gather any further details about what was meant by “simulations,” nor did they publish analyses that statistically associated changes in social media posts at a particular school with the

presence versus absence of simulation elements at that school. Finally, since the data collection period encompassed multiple mass violence events and heightened activism following the 2018 school shooting at Marjory Stoneman Douglas High School in Parkland, Florida, the results need to be replicated during other time periods to rule out period effects.

Drills and Student Behavior

Within the limited research base on the mental, emotional, and behavioral health outcomes of active shooter drills, studies have almost exclusively assessed internalizing variables (e.g., emotions, perceptions) rather than behaviors. Only three studies (Riggs et al., 2023; Schildkraut & Nickerson, 2022; Schildkraut et al., 2024a) assessed any kind of behavioral outcome. All three employed self-report questionnaires about participants’ recollections and perceptions of their own behavior rather than any direct metric of behavior (e.g., disciplinary incidents or absences as documented in school records). Of course, students’ recollections of their own past behaviors may be inaccurate for any number of reasons (e.g., tending to reflect personality traits or present emotional state).

Two of the pre–post studies by Schildkraut and colleagues queried students about their self-reported avoidance behaviors. First, Schildkraut & Nickerson (2022) found that, while students in grades 6–12 reported less fear and perceived risk following a series of two lockdown drills, they also reported more avoidance behaviors (i.e., staying home or avoiding class because of fears about victimization) after the two drills. Second, Schildkraut et al. (2024a) found that high schoolers reported more total self-protective behaviors following training on a Standard Response Protocol and participation in a single lockdown drill. The researched used a self-protective behaviors scale consisting of subscales for avoidant and proactive behaviors. In contrast to Schildkraut & Nickerson (2022), which found an increase in self-reported behaviors post-drill, Schildkraut et al. (2024a) did not find a change in avoidant behaviors post-drill. This study did find a significant post-drill increase in both the proactive behaviors subscale and the broader self-protective behaviors scale.

However, these findings of increased proactive and general self-protective behaviors may be experimental artifacts, since both measures involved items assessing “attended trainings on emergency plans” and “participated in a lockdown drill.” Endorsement of both items would be expected to increase after students were required to participate in a training and drill as part of the study, so these increased scale scores may not reflect students’ volitional self-protective behavior changes post-drill.

The third and final study to assess self-reported behavior changes following school active shooter drills was authored by Riggs et al. (2023). As previously discussed, the researchers surveyed 108 youth–caregiver pairs presenting to a single emergency department and found that approximately one-third of the youth reported increased depression and anxiety symptoms post-drill, when asked to reflect on any active shooter drills experienced over the past 2 years. Furthermore, youth that reported feeling anxious about drills were more likely to report being bullied, skipping school because of concerns about victimization, and participating in an unannounced drill (i.e., a drill they did not know was a drill at the time; Riggs et al., 2023). This study provides some support for the idea that anxiety about drills may contribute to absenteeism and lost learning time, at least for some proportion of students, although all the limitations about cross-sectional research designs continue to apply here.

Insights from Qualitative Studies

The committee identified five qualitative studies assessing participants’ reactions to active shooter drills. Two involved only students, two involved only staff, and one involved both students and staff. As discussed, qualitative studies are not intended to be representative and cannot establish causality. Instead, these qualitative studies furnish rich, nuanced examples of the kinds of responses that some participants experience to drills. Because participants’ responses are open-ended, the wide-ranging information they shared spans many of the outcome categories discussed earlier in this

chapter. These five qualitative studies varied in the data collection method and population assessed. They tended to suggest mixed responses to drills, but with negative responses outweighing positive or neutral ones. Of course, negative responses to drills may be more memorable or salient than positive or neutral responses, and participants who experienced or observed negative responses to drills may have been particularly motivated to share these perceptions in a voluntary qualitative study. Box 3-1 presents additional qualitative information about student responses to school active shooter drills, as documented in the popular media by an investigative reporter, rather than in a peer-reviewed scientific article.

Student Participants Only

• Moore-Petinak et al. (2020) collected written responses to five open-ended probes from 815 youth (aged 14–24) who responded to a national text message poll. About 50% of youth reported that drills made them feel unsafe or scared (e.g., “scared because we thought they were real”). Another 15% of the youth reported other negative emotions such as sadness, helplessness, or frustration. In contrast, 20% of the youth reported feeling more prepared (e.g., “safer to know what to do in an active threat”). Approximately one-quarter of youth felt that drills had no emotional effect on them or failed to prepare them effectively for a real shooter (e.g., because drills also informed potential school shooters about emergency response procedures). We note that this age range (14–24) certainly includes students who are no longer in high school, who may have been recollecting their reactions to secondary school drills from many years ago.
• Bonanno et al. (2021) conducted individual, semi-structured interviews with 11 White students aged 8–11 living in a New York City suburb. Most of the students described feeling fearful of drills (e.g., “I worry if something bad is happening, if the lockdown drill is real, and this is a real emergency”; “It just freaks me out; I don’t like it at all”). Many described the need to engage in coping strategies to help them address

• their fear or anxiety (e.g., reminding themselves that they are safe, talking to adults). A couple of students described the drills as fun or exciting, but most described them as frustrating and disruptive to the school day. The students expressed general sentiments about the need for emergency preparedness, but many were unclear about precisely what kind of threats the drills were meant to guard against.

Student and Staff Participants

ElSherief et al. (2021), who conducted the social media webscraping study previously discussed, also conducted a series of six virtual focus groups with a total of 34 participants (students, parents, teachers) from across the country who had participated in at least one drill or had a child who had done so. Many of the focus group participants described negative emotional reactions to drills, such as students “texting their parents, praying, crying” because they “thought they were going to die”; students becoming hypersensitive to signs of potential danger at school (e.g., a door handle jiggling); or teachers “break[ing] down at recess” because of stress following a drill. Other participants reported that students were desensitized to active shooter drills and unbothered by them.

Staff Participants Only

Goodman-Scott and Eckhoff conducted a series of four focus groups with a total of 26 school counselors from the southeastern United States (Eckhoff & Goodman-Scott, 2023; Goodman-Scott & Eckhoff, 2020). Numerous school counselors mentioned unintentional emotional fallout from participation in drills, including students’ fear and anxiety; parents’ worry and protectiveness; and school counselors’ fear, anxiety, and uncertainty. For example, one school counselor shared, “[Even when you tell the students], ‘This is just practice,’ it still gets them upset. They end up having a mini meltdown.” Others described simulation aspects of drills (e.g., a school staff member jiggling the door handle to check if it is locked) as “traumatic,” noting that students with past traumas are

particularly sensitive (e.g., “I had a student with a past trauma. . . . When there was a lockdown drill, his anxiety kicked in. . . . He would relive it [the trauma] during the lockdown drills”). School counselors also described the drills as stressful and anxiety provoking for themselves, not just the students (e.g., “It’s a lot of wear and tear mentally, going through the drills”; “Your anxiety’s up, your fear is heightened, and you do go into a fight-or-flight response”). Some school counselors mentioned how drills evoke troubling uncertainty about what would really happen in the case of a genuine school shooting, including fears about their own safety and perceptions of the weight of responsibility to protect students (e.g., “We’re responsible for all those babies and I don’t want one of them to get hurt. . . . It’s a huge responsibility”).

Jackson & Golini (2024) interviewed 10 teachers or paraprofessionals who were employed in a classroom with at least one 4- to 8-year-old child with autism spectrum disorder (ASD) during the 2018–2019 school year. Staff described numerous difficulties when attempting to teach and practice drills with students with ASD. For example, because children with ASD tend to thrive on routine, the abruptness of drills created particular disruption for these students. Furthermore, drills often involved atypical sensory experiences that could be markedly stressful for children with ASD (e.g., being crowded in close proximity to other students within a small space). Staff also reported that the challenging behaviors of children with ASD (e.g., tantrums, aggression, elopement) made it difficult for those students to learn and perform the drills correctly. Regarding mental, emotional, and behavioral responses to drills, participants described the drills as sometimes “traumatizing” for both students and staff, particularly if the drills involved realistic simulations (e.g., actors shooting blanks). One participant described the perspective of a student with ASD as follows: “Why would I want to gather with all of you in a small space in the dark? Like what is happening? Why are you telling me to be quiet? It’s just overwhelming . . . definitely fight or flight.” Another participant emphasized that even the staff could become “dysregulated” during drills because of sensory overload. Many participants reported feeling stressed by the weight of

their responsibility for protecting students with ASD or other special needs. For example, one participant described physically carrying a student with ASD to safety when a drill occurred while the student was on the playground: “I had to yell, ‘Grab him!’ because he was breaking down. . . . He didn’t know why he had to leave recess. So I just scooped him up and ran inside.” In general, participants stated that they wished there were more resources for teaching students with ASD about drills.

Committee Observations

Overall, the qualitative research demonstrates that some proportion of students and school staff feel afraid, uncertain, or overwhelmed by active shooter drills, while others feel desensitized and unbothered. A consistent theme was that students and staff participating in drills often feel unsure whether a drill is a simulation or a real emergency, which increases feelings of anxiety and uncertainty. Furthermore, school staff reported feeling burdened by the weight of physically protecting students from harm, which is a role they do not normally enact. However, in the largest qualitative study (Moore-Petinak et al., 2020), a sizeable minority of students reported feeling more prepared or experiencing no emotional changes post-drill.

CONCLUSION

The available research base on the mental, emotional, and behavioral health outcomes of active shooter drills in K–12 schools is mixed and limited. Overall, the published literature tends to suggest that these practices increase perceptions of emergency preparedness post-drill, coupled with increases in negative responses such as fear, anxiety, and avoidance. However, these are only inferences grounded in average trends; multiple studies suggest that students and staff have varied responses to drills. Studies indicate that

somewhere between 10% and 65% of students demonstrated negative emotional reactions post-drill, reactions that may be more common among students with preexisting vulnerabilities (e.g., self-reported bullying victimization). Nonetheless, the results still indicate that the proportion of students reporting no changes or positive changes (e.g., greater confidence in their ability to respond to an emergency) post-drill ranges from a substantial minority to a majority of students. Studies that allowed richer, more nuanced observations of post-drill responses (i.e., the social media webscraping study and qualitative studies) tended to document more negative reactions compared with studies that surveyed participants directly using discrete, limited outcome measurements (i.e., circling responses on Likert scales). This observation suggests that researchers need to give more attention to how they are assessing potential negative reactions, as brief or unvalidated outcome measures may not be effective at capturing subtle mental or emotional changes post-drill, particularly if those changes affect only a minority of students and may be washed out by calculating only average trends across a large sample (see Chapter 7 for an in-depth discussion of future research needs).

In sum, the limited available research on student and staff responses to active shooter drills in K–12 schools raises concerns that these drills may be associated with negative mental, emotional, and behavioral health outcomes for many (though not all, or maybe even not most) people exposed to them. These potential risks must be weighed against the potential benefits accompanying any emergency preparedness procedure. However, existing data paint with a broad brush because of the lack of high-quality research investigating so many important dimensions of drills—most notably the effects of different drill components, the importance of drill frequency, the effects on staff, the occurrence of changes in behaviors of interest, and the needs of students from vulnerable groups. This lack of relevant research using rigorous research designs is striking given that active shooter drills are an intervention that has been applied to millions of students and staff across the country for many years. One might reasonably ask whether the existing scientific

evidence base is sufficient to guide those making decisions about active shooter drills, or to settle on default procedures. But because drills are already happening in some fashion, and will continue to happen, waiting for more data is not an option. Schools and state legislatures will need to continue to draw on the limited existing research on drills; perceived security considerations; basic scientific knowledge on stress, learning, and child development; applicable laws and regulations; and their personal understanding of their own students and school contexts. While we wait for more high-quality and detailed scientific research, decision-makers will need to take all these considerations into account and make a reasoned choice about whether, how, and under what circumstances to implement school active shooter drills.

REFERENCES

This page intentionally left blank.