Abstract
Although research has explored the effects of protective factors on fostering resiliency within individuals, the same level of inquiry has not emerged investigating the causes of why high-risk organizations are resilient to serious violent delinquency. One type of organization that seems particular appropriate for research inquiry is the school. Using a sample of 307 school principals from the School Survey on Crime and Safety, this study investigates how protective factors are individually and cumulatively related to resiliency against serious violence within schools. The findings indicate schools may be more reactive than proactive in their efforts to remain resilient. The theoretical and policy implications of these findings are discussed.
Criminologists have produced a lengthy history of efforts to understand how risk factors increase the probability of involvement in a variety of problem behaviors (see Farrington & Coid, 2003; Farrington et al., 2006; LeBlanc & Loeber, 1998; Lösel & Bliesener, 1990). Much of this research has documented the importance of the ways in which the accumulation of risks substantially increases the probabilities of involvement in a variety of detrimental behaviors (Farrington et al., 2006; Farrington & Loeber, 1999; Rutter, 1990). Although an isolated risk explains a relatively small proportion of the variance in delinquency and other problem behaviors, the accumulation of risk has consistently explained significantly larger proportions of this variance (see Rutter, 1979). Despite the high probability of delinquency associated with multiple risk factors, a significant proportion of these individuals are able to withstand the pressures and strains they experience within high-risk environments and remain resilient (Farrington et al., 2006; Laub & Sampson, 2001; Rutter & Giller, 1983; Smith, Lizotte, Thornberry, & Krohn, 1995; Turner, Hartman, Exum, & Cullen, 2007; Werner, 1989). More specifically, the extant research has provided estimates that between 25% and 50% of these individuals refrain from involvement in delinquency and crime (Farrington et al., 2006; Smith et al., 1995).
To account for these so called “statistical anomalies,” scholars have focused on investigating the impact that protective factors have on resiliency (Garmezy, 1991; Masten, Best, & Garmezy, 1990). Protective factors are defined as those variables that function to moderate the effects of risk; their effects are minimal or absent in low-risk environments but become increasingly stronger in more risky or high-risk environments (Luthar, Cicchetti, & Becker, 2000). Despite efforts to understand how protective factors function to foster resiliency, there has been a noticeable omission in the types of outcomes associated with being resilient. That is, most resiliency-related research in the criminological literature has focused on resiliency at the individual level and its relationship to delinquent or criminal offending (Farrington et al., 2006; Turner et al., 2008), drug use (Smith et al., 1995), and psychological disorders (Rutter, 1990). Few studies, however, have focused on investigating the ways that protective factors function to moderate the effects of risk and foster resiliency at the organizational or institutional levels.
One type of institution that we believe is worthy of exploration is the school, due to the levels of violence that occur in educational institutions (Dinkes, Cataldi, Kena, & Baum, 2006). Although the victimization rate within schools has generally decreased since the mid-1990s, violence, theft, drugs, and weapons continue to pose problems for school administrators (National Center for Education Statistics, 2007). Moreover, the distribution of violent incidents within schools is not random. Some schools experience no violent incidents within a given academic year whereas others endure violence on a daily basis (Juvonen, 2001). These data suggest that certain schools are at an elevated risk of experiencing a violent incident, much like individuals. Despite the heightened risk of violence that schools experience, many schools are able to remain resilient and avoid serious forms of violence. Few studies, however, have examined the covariates that might increase the probability of fostering resiliency to violence within schools.
In this context, the present study seeks to fill this void in two important ways. First, we use data from the School Survey on Crime and Safety (SSOCS), a national study designed to assess the amount of delinquency and crime within schools, to investigate the effects that protective factors have on fostering resiliency against violent delinquency within schools. Second, we investigate the effects of the accumulation of protective factors on being resilient from violent delinquency. The findings give insight into the ways that high-risk schools are capable of avoiding violent delinquency and creating a healthy educational environment for students.
Individual Resiliency
According to Henderson and Milstein (2002), researchers have conceptualized resiliency in three ways: “(1) positive developmental outcomes among children who live in ‘high-risk’ contexts, such as chronic poverty or parental substance abuse, (2) sustained competence under prolonged stress such as the events surrounding the break up of their parents’ marriage, and (3) recovery from trauma, especially the horrors of civil wars and concentration camps” (p. viii). Research on individual resiliency has only recently drawn the attention of criminologists, though its foundation can be traced to developmental psychology and developmental psychopathology where it was identified that a majority of individuals within high-risk environments engaged in problem behaviors (Barocas, Seifer, & Sameroff, 1985; Coyne & Downey, 1991; Dubow & Luster, 1990; Farrington et al., 2006; Newcomb, Maddahian, & Bentler, 1986; Rutter, 1979, 1990; Sameroff & Seifer, 1990; Thomas & Chess, 1984; Werner, 1985).
Garmezy and others launched a scholarly effort to understand why a substantial percentage of individuals were able to cope with the pressures of residing in high-risk environments and never engage in serious delinquency and crime (Rutter, 1985; Werner, 1989; Werner & Smith, 1992). This literature suggests that individuals who are resilient rely on several different protective factors to cope with or manage the detriments associated with high-risk environments (Smith et al., 1995; Turner et al., 2007). Research has documented three general categories where protective factors are located: the individual, family, and external support systems (see Garmezy, 1985).
Within the individual context, scholars found that an individual’s temperament or disposition, particularly measured early in the life course, can foster resiliency in adolescence and adulthood (Kolvin, Miller, Fleeting, & Kolvin, 1988; Thomas & Chess, 1984; Werner & Smith, 1992). Resilient adolescents also score higher on levels of social maturity and sociability than those who were not resilient (Lewis & Looney, 1983). In perhaps one of the best documentations of resiliency, Werner (1993) found that resilient individuals have a higher likelihood of managing stressful situations through an active, rather than passive, problem-solving approach. Werner (1993) also found that significant active problem-solving skills at age 10 were one of the best predictors of successful adaptation to risk in the early adulthood developmental period. Finally, research suggests that individuals who score higher on measures of self-efficacy, self-confidence, and self-worth have a higher probability to abstain from involvement in delinquency in high-risk situations (Cicchetti, Rogosch, & Holt, 1993; Garmezy, 1985; Werner, 1990). In short, individual protective factors have been found to be some of the most impactful sources to foster resiliency.
Research has also explored the protective factors within familial environments that foster resiliency. This literature is much like the studies that identify important risk factors within the family that increase the probability of offending (see Loeber & Stouthamer-Loeber, 1986). For example, supportive and caring relationships between a parent and child have been found to increase resiliency among high-risk youths (Weinraub & Wolf, 1983; Werner, 1993). A supportive relationship with at least one parent in environments with severe familial discord has also resulted in a higher likelihood of resiliency (Egeland, Carlson, & Sroufe, 1993; Kimchi & Schaffner, 1990). In addition, Smith and her colleagues documented that parent–child relationship is significant in both directions. Resiliency emerged regardless of whether it was the parent or child who perceived the presence of a supportive and caring parent–child relationship (Smith et al., 1995). Finally, the extent to which familial environments offer cognitive stimulation to youth has been documented to enhance resiliency in high-risk youths (Turner et al., 2007).
Research suggests that those who are resilient are particularly skilled at identifying individuals external to the family to assist them in navigating the stress experienced within high-risk environments (Grossman & Garry, 1997; Kellam, Ensminger, & Turner, 1977; Smith et al., 1995; Werner & Smith, 1992). External environments, particularly within schools, community-level organizations, and religious institutions, have been found to be important mechanisms for adaptation for individuals residing in high-risk environments. For example, Herrenkohl, Herrenkohl, and Egolf (1994) found that higher-quality or more effective schools are particularly beneficial in increasing individual levels of self-worth among high-risk youth. Active involvement in religious institutions has also been found to insulate high-risk youths and provide them with stability and meaning to their lives (Anthony & Cohler, 1987; Werner & Smith, 1992). Thus, studies examining resiliency have identified individual, familial, and extra-familial sources of protective factors that distinguish “resilients” from “nonresilients” within high-risk environments.
Organizational Resiliency
Although resiliency research at the individual level is growing within the criminological literature, few have explored the impact of organizational resiliency research. As Youssef and Luthans point out, “Resiliency has been given considerable surface recognition, but has not yet been systematically understood, researched, or applied at the organizational level” (p. 304). Similarly, management scholars and practitioners have only recently turned their focus on conceptualizing and understanding the nature of resilient organizations, defined as “those able to survive, adapt, swiftly bounce back, and flourish despite uncertainty, change, adversity, or even failure” (Youssef & Luthans, 2005, pp. 303-304). Nevertheless, recent work in the organizational literature has given some clarity about the development of organizational resilience.
Although arguably less developed than the individual resiliency literature (see Masten & Reed, 2002), the organizational resiliency literature has provided insight into three valuable strategies for promoting resiliency development. First, resilient organizations have successfully implemented risk-focused strategies that emphasize the reduction or prevention of risks that are perceived to increase the probability of undesirable outcomes (Masten & Reed, 2002). Such strategies arguably focus on only half of the risk–protection equation and are generally viewed as limited in scope. Second, resilient organizations have successfully implemented asset-focused strategies that place emphasis on the development or enhancement of adaptive processes that can yield positive outcomes (Masten & Reed, 2002). Much like their risk-focus counterparts, such strategies are only focused on half of the risk–protection equation. The third set of organizational resiliency developmental strategies are process-focused strategies that rely on the energy of human adaptive systems within the organization to be mobilized in an effort to reduce risk and enhance protection (Masten & Reed, 2002). Process-focused strategies use the social capital of the participants within the organization to address the reduction of risk and enhance protection through a systematic set of strategies designed to result in positive change.
One of the potentially greatest contributions to organizational change and resiliency are the values of the organization. As Coutu (2002) states, “Strong values infuse an environment with meaning because they offer ways to interpret and shape events” (p. 52). In essence, organizational values provide a compass or direction in the decision-making processes to assist in navigating through high-risk periods. Typically the rules and regulations of the organization provide particular insight into the nature and depth of the organizational values. As Weick (1993) notes, “When people are put under pressure, they regress to their most habituated ways of responding” (pp. 638-639). Thus, organizational values provide a foundation of consistency and stability in times of crisis, whereas rules and regulations serve as a guide to positive and effective responses.
In terms of responding to adversity, Hamel and Valikangas indicate that resilient organizations generally deal with four challenges. First, resilient organizations attend to the “cognitive challenges,” such as the culture of denial that can emerge within institutions due to a belief that they are immune to changes or challenges outside of the organization (Hamel & Valikangas, 2003). Second, resilient organizations deal with “strategic challenges” by developing new options and alternatives to counteract existing challenges (Hamel & Valikangas, 2003). Third, organizations must attend to “political challenges,” by attempting to navigate the politics that inevitably develop while enacting potentially promising programs that counteract prevailing risk (Hamel & Valikangas, 2003). Finally, resilient organizations address “ideological challenges” where the current mentality of optimization is replaced with more innovative approaches to adversity and change (Hamel & Valikangas, 2003). The inevitable result of proactive problem solving is that resilient organizations are capable of success in the face of adversity.
School Violence
In light of the research on organizational resilience, it is important to consider how schools deal with threats to safety, and organizational risks more generally. The substantive literature on school characteristics and the likelihood of problem behavior have produced mixed results due in large part to the diverse measures used to operationalize concepts and outcomes (Gottfredson, Gottfredson, Payne, & Gottfredson, 2005). The grade level of the institutions examined also affects the potential for comparison, as factors that affect junior high schools may be different from those affecting senior high schools (Gottfredson et al., 2005). Despite potentially disparate findings, the research generally suggests that structural and institutional characteristics of schools play an important role in predicting the likelihood of violence and problematic student behavior (Gottfredson et al., 2005). For example, fair and consistently enforced disciplinary practices decrease the likelihood of school violence (Gottfredson & Gottfredson, 1985; Gottfredson et al., 2005). Reinforcing conforming behavior through the use of rewards for compliance with school standards is also correlated with lower levels of violence (Brezina, Piquero, & Mazzerole, 2001; Gottfredson & Gottfredson, 1985).
Institutions that foster a positive social climate also have lower rates of disorder and absenteeism (Bryk & Driscoll, 1988; Bryk, Lee, & Holland, 1993; Lee & Croninger, 1996; Wilcox & Clayton, 2001). Schools that clearly communicate expectations for student learning and achievement and foster caring social relationships between faculty, staff, and students appear to reduce the risk of disorder and violent behavior by bonding the students to the larger school community (Bryk & Driscoll, 1988; Bryk et al., 1993; Lee & Croninger, 1996; Wilcox & Clayton, 2001). In addition, resources for teachers and classroom strategies that incorporate students into the educational process may decrease the likelihood of violent incidents (Bryk & Driscoll, 1988; Bryk et al., 1993)
There is also some connection between school size and problematic behavior, such that larger classes have a greater likelihood of disorder and violence (Bryk & Driscoll, 1988; Gottfredson et al., 2005; Stewart, 2003). This finding is, however, contested in the literature, with some studies finding no association (Lee & Croninger, 1996; Welsh, Stokes, & Greene, 2000) or inverse relationships between size and behavior (Brezina et al., 2001). In light of this debate, the true impact of school size on student behavior is unclear, though there is an identifiable relationship present.
Beyond social and institutional factors, technological and architectural strategies have been placed into schools as a means to protect and secure the institution. Metal detectors and closed-circuit television systems have been employed, as well as institutional designs that increase the capacity for administrative surveillance, along with gates and barriers to cordon off sections of a school from the student population (Gottfredson et al., 2004). These methods are commonly employed in urban institutions and are well implemented relative to other violence intervention programs (Gottfredson et al., 2004). However, the overall impact of technological devices and security barriers on school violence is not well documented. For example, a study of New York City high schools found that students were less likely to carry a weapon to school if their institution used metal detectors (Ginsberg & Loffredo, 1993). Thus, security devices may reduce the likelihood of school violence, though their overall impact has not been evaluated in a substantive fashion.
The previously described factors can all influence the likelihood of violence and disorder and are at the heart of multiple school-based delinquency prevention strategies (see Gottfredson & Gottfredson, 2001). Large-scale studies of school violence prevention practices, however, indicate that school administrators may pay little attention to institutional conditions and climate. Rather, individual-level programs such as counseling and student instructional programs are commonly employed to reduce problematic behavior (Gottfredson & Gottfredson, 2001). If and when violent incidents occur, it is possible that schools may respond and alter their prevention methods, though few researchers have considered the ways that institutions react to serious incidents.
Current Focus
Despite the recent proliferation of research on protective factors that foster individual resiliency, few have considered the influence of institutional-level protective factors on the development of organizational resiliency, especially within school environments. Exploring why high-risk schools evade experiencing serious violent incidents is necessary, given the seeming ubiquity of highly publicized, traumatic incidents in secondary and postsecondary schools across the country (National Center for Education Statistics, 2007). The present study isolates a high-risk cohort of middle and high schools and seeks to understand how protective factors insulate these institutions from experiencing serious violent incidents. Using a cohort of high-risk schools is appropriate because we should expect to observe the effects of protective factors to emerge most strongly within this cohort.
Thus, this investigation proceeds along two fronts. First, an investigation is made into whether institutional-level protective factors affect organizational resiliency in a similar fashion to efforts at the individual level. It is expected that an increase in the institutional-level protective factors should correspond with an increased likelihood that the school remains resilient. In light of research documenting the importance of the accumulation of protective factors in fostering resiliency at the individual level (see Turner et al., 2007), an investigation is made inquiring whether the accumulation of institutional-level protective factors assist in fostering organizational resiliency within the school. The combined results should prove useful in guiding the development of policies geared to resist violent incidents within high-risk schools.
Methods
Data for this study were provided by the 2003 School Survey on Crime and Safety (SSOCS), issued by the National Center for Education Statistics (NCES). The SSOCS is a stratified sample of 2,772 elementary, middle, and high schools throughout the United States. Principals from these schools were asked to report on a variety of topics related to the well-being of the school such as the frequency and type of crimes at school, disciplinary problems, school demographic characteristics, available technology (e.g., telephones in classrooms and security cameras), use of security or law enforcement personnel, drug possession and use, bullying, classroom disorder, and disciplinary actions taken in response to crime or misbehavior. The schools included in the sample were public elementary or secondary schools in the United States, excluding those in U.S. territories. The data were collected through the use of mail surveys and follow-up telephone calls.
Sample
Although the SSOCS collects data from 2,772 schools, the sample in the present study uses data from a subset of schools identified as “high-risk.” Two methods were used to isolate the high-risk sample. First, consistent with much of the prior resiliency research investigating individuals (see Smith et al., 1995; Turner et al., 2008), high-risk schools were differentiated from their low-risk counterparts by a simple count method. That is, nine risk factors that were found to increase the probability of a school to experience serious violent incidents were used to categorize schools into high and low risk (Miller, 2003). Risk factors included the number of disruptions (death threats, bomb threats, or chemical, biological, or radiological threats), the number of serious disruptions (daily or weekly racial tensions, bullying, sexual harassment of other students, verbal harassment of teachers, widespread disorder in classrooms, student acts of disrespect for teachers, and daily, weekly, or monthly gang or cult/extremist group activities), low percentage of students considering academics to be important, low percentage of students below the 15th percentile on standardized tests, high number of classroom changes, high or moderate crime levels in the areas in which students live, urban environment of the school, large school enrollment size, and large percentage of minority students. 1 The high-risk cohort contains all schools with at least seven risk factors (n = 307). A comparison of means t test showed the high-risk schools had a significantly greater proportion of schools (p = .4919) that experienced at least one serious violent incident than the low-risk cohort of schools (p = .2264, p < .001). Table 1 shows the proportion of schools experiencing at least one violent incident at each number of risk factors.
Proportion of Schools Experiencing Violence (Nonresilient Schools) by the Number of Risk Factors.
To verify the validity of the count method described above, propensity-score matching was also employed as a second method of isolating a high-risk cohort of schools. A logistic regression model was created to predict the log odds of a school experiencing at least one serious violent incident. 2 The nine risk factors identified above were used as the predictors in this model. Predicted log odds generated by this model ranged from a minimum of 0.07802 to a maximum of 0.58811. Table 2 shows the predicted log odds of experiencing serious violence at each number of risk factors. This propensity-score method showed a drastic increase in the predicted log odds of experiencing at least one serious violent incident at a level of seven risk factors. A comparison of means t test compared the mean-predicted log odds of the low-risk cohort (0 to 6 risk factors, M = 0.2299) to the mean-predicted log odds of the high-risk cohort (7 to 9 risk factors, M = 0.4643). The mean-predicted log odds of the high-risk cohort were significantly greater than the mean-predicted log odds of the low-risk cohort (p < .001). Thus, both the count method and propensity-score method demarcated the same boundary between the low- and high-risk cohorts. All subsequent analyses were performed on the high-risk cohort (≥7 risk factors).
Descriptive Statistics of Low- and High-Risk Schools.
Differences are significant at the 0.001 level.
Measures
Protective factors
School characteristics were then selected as being protective against violence. These characteristics were grouped into six categories of protective factors. The first category of factors was named “architectural factors” and included three items. These items were requiring visitors to the school to sign in at the school office, controlling access to school grounds, and controlling access to school buildings. Responses to each of these items were either “yes” or “no.” Reliability analysis of the architectural factors index yielded a Cronbach’s alpha of .322. 3
Eight items were categorized as “technological factors.” These factors included having students pass through metal detectors, having visitors pass through metal detectors, using security cameras to monitor the school, providing telephones in classrooms, random metal detector checks of students, random sweeps for contraband, random police dog sniffs for drugs, and providing the staff with two-way radios. Responses to these questions were also either “yes” or “no.” The Cronbach’s alpha of this index was .385.
“Security factors” were the third category and included seven items pertaining to law enforcement and security professionals. These items were the use of security personnel during school hours, when students arrive and leave campus, during selected school activities (e.g., athletic events), when school activities are not occurring, at any other time, the use of uniformed security guards, and security guards armed with firearms. Responses to these questions were either “yes” or “no.” The Cronbach’s alpha of this index was .329.
Questions pertaining to student and/or faculty policies were categorized as “identity factors.” These five items included requiring student uniforms, enforcing a strict dress code, allowing only clear or no book-bags or backpacks, requiring student ID cards, and requiring faculty ID cards. As above, responses to the questions were either “yes” or “no,” with a Cronbach’s alpha .408.
The largest category was called “training and planning factors” and included 12 items. These items consisted of having a written plan for school shootings; a written plan for bomb threats; formal programs intended to prevent or reduce violence 4 (reported as eight separate items); providing training to teachers on school-wide discipline policies and practices related to violence, alcohol, and/or drug use; and training teachers to recognize early warning signs of violent behavior. Responses to these questions were either “yes” or “no.” Reliability analysis produced a Cronbach’s alpha of .613.
The final category, “community involvement factors,” included eight items pertaining to cooperation between the school and various groups to maintain safe, disciplined, and drug-free schools. These items were having involved parent groups, social service agencies, juvenile justice agencies, law enforcement agencies, mental health agencies, civic organizations/service clubs, private corporations and businesses, and religious organizations. Responses to the questions were either “yes” or “no,” with a Cronbach’s alpha of .735.
Resiliency to Violence
An outcome variable was created reflecting whether a school was resilient to serious violence. If a high-risk school did not experience a serious violent incident (as defined above), the school was classified as resilient “1” and if a high-risk school experienced one or more serious violent incidents, it was classified as nonresilient “0.” Resilient schools (n = 156) comprised 50.8% of the high-risk cohort. Individual t tests compared the mean levels of each protective factor of the resilient schools to those of the nonresilient schools. Interestingly, the nonresilient schools had higher mean levels of all protective factors, except for identity factors. Admittedly, however, several of the mean differences were not statistically significant. Table 3 shows the mean levels of the protective factors of the resilient and nonresilient schools.
Mean Levels of Protective Factors of Resilient and Nonresilient Schools.
p < 0.5, ***p < .001
Analytic Strategy
The subsequent analysis proceeds in two stages. The first set of analyses investigate the impact of the individual effects of each of the protective factors on the probability of the school remaining resilient. The second set of analyses investigate the cumulative effects of the protective factors via a protective factor index (PFI), to be explained below. Combined, these analyses provide insight into the individual and cumulative effects of protective factors on resiliency against serious violent incidents.
Findings
To begin, the individual protective factors were entered into a multivariate logistic regression model predicting resiliency to violence. Results from this model are shown in Table 4. This logistic model was a significant predictor of resiliency (model χ2 = 18.375, df = 6, p < .01). Consistent with results from the t tests, increases in all but one of the categories of the protective factors decreased a school’s odds of being resilient to violence. Identity factors were the only category of protective factors whose increase would result in an increase in a school’s odds of being resilient. For each additional identity factor found in a particular school, the odds of that school being resilient to violence increased by a factor of 1.260 (p < .05).
Logistic Regression Using Protective Factors to Predict Resiliency.
p < .05. **p < .01. ***p < .001.
Several of the risk factors were employed as controls, and therefore, not included again as controls to avoid redundancy. However, three items not used as risk factors were identified as indicators of instability in a school and were used as controls in subsequent analyses. These items were measures for average daily attendance, limited English proficiency, and the number of students transferred to the school. Average daily attendance was reported as a percentage of the student population and ranged from a minimum reported value of 65% to a maximum value of 99%. Limited English proficiency was also reported as a percentage of the student population and ranged from a minimum reported value of 0% to a maximum reported value of 87%. The number of students transferred to the school was a simple count and ranged from a minimum of 0 students transferred to a maximum of 1,000 students transferred.
Protective Factor Index
A protective factor index (PFI) was then created for the high-risk cohort of schools as a summation of how many protective factors a particular school exhibited. The protective factors were divided at their respective means. Scores above the mean were given a value of 1 in the index, and scores at or below the mean were given a value of 0. Creating this index allowed the protective factors to be included in a regression model as a singular variable. Thus, it was possible to assess the combined impact of the protective factors on resiliency to violence. Because they consistently operated in the reverse direction at statistically significant levels, identity factors were excluded from this protective factor index. Leaving identity factors in the PFI caused the impact of the protective factors on resiliency to be drastically underestimated. The PFI values thus ranged from a minimum of 0 to a maximum of 5.
The PFI was entered into a logistic regression model predicting the log odds of a high-risk school being resilient to violence. As mentioned above, identity factors were included apart from the PFI. The model also included the three control variables mentioned above. Results indicated that the model was able to significantly predict resiliency to violence, as indicated by the model chi-square (χ2 = 30.584, df = 5, p < .001). Furthermore, the PFI was a significant predictor of resiliency (p = .001), net of influence by the above controls. The exp b of the PFI, 0.662, indicated that increases in the index resulted in decreases in a school’s odds of being resilient to violence. In other words, higher scores on the PFI were associated with smaller predicted odds of resiliency. The identity factors remained a significant predictor of resiliency (p < .05) and continued to operate in the opposite direction from the remainder of the protective factors. Higher numbers of identity factors were associated with higher odds of resiliency. Table 5 shows the results of this model.
Logistic Regression Using the Protective Factor Index (PFI) to Predict Resiliency.
p < .05. **p < .01. ***p < .001.
Discussion
Although a considerable amount of research has focused on individual adaptations to high-risk environments, much less scholarly attention has centered on how organizations remain resilient and withstand the pressures associated with vulnerable environments. Research has yet to demarcate the importance of how protective factors function to prevent high-risk schools from experiencing serious violent incidents. The present research addresses these deficiencies using a national probability sample of middle and high school principals to investigate the individual and cumulative effects that a variety of protective factors have on a school’s probability of being resilient against serious violent incidents. Four key findings have emerged from these efforts.
First, much like the research on individuals, a significant number of schools were capable of being resilient regardless of their high-risk classification. Specifically, just over half (50.8%) of the high-risk schools did not experience a serious violent delinquent event and were, therefore, classified as resilient. Second, contrary to the resiliency research on individuals, schools generally experienced lower levels of the protective factors used in the analysis. In fact, mean levels of five of the six protective factors measured were lower for resilient schools versus nonresilient schools. It should be noted, however, that only two of these differences were significant at conventional levels. Nevertheless, this trend is antithetical to what is typically found at the individual level where resilient individuals typically have higher mean levels of protective factors. These differences might suggest that organizations, unlike individuals, build or develop protective mechanisms after they experience violent incidents. Unfortunately, these data do not permit an analysis of the changes in protective factors over time as a result of experiencing violence.
Third, the multivariate analyses investigating the individual effects of the protective factors on resilience indicated that schools with policies that generally standardized the environment within the school (i.e., dress code, uniforms, ID cards) resulted in an increased probability of being resilient. Schools with higher levels of architectural measures, such as requiring visitors to sign in at the school office, controlling access to school grounds, and controlling access to school buildings, corresponded with a decreased likelihood of resilience. Thus, these findings suggest that the impact of protective factors vary depending on the type of protective factor. Schools that develop more restrictive policies focused on security and technology may not necessarily increase the resilience of their institution; thus, careful attention must be paid to the responses organizations take to deal with violence.
Fourth, the final analysis focused on investigating the cumulative effects of protective factors on a school’s ability to remain resilient. The results indicated that the accumulation of protection decreased a school’s ability to foster resiliency. This finding could be the result of schools reacting to what they might have experienced in the past. Changes to increase protection, however, do not appear to result in significant changes in the outcome of serious violent incidents. Future research is needed to explore the impact of protective factors on schools over time to assist in disentangling this complex relationship.
Although the present research has uncovered several key contributions to the impact of protective factors on schools and their ability to remain resilient, it is important to acknowledge the limitations of these efforts. First, much like resiliency research at the individual level, the present study takes a cross-sectional approach and only measures resiliency at one point in time. The present study cannot disentangle the temporal ordering of the protective factors and the measure of resiliency. Future research should attend to this limitation and examine schools over time. Such a strategy would permit an assessment of changes in protection and its impact on school resilience over time.
Second, the present study only measured six different areas of protection, most of which directly fall within the structural nature of scholastic institutions. There was no assessment of a school’s cultural aspects, particularly as they relate to protective mechanisms. Documenting the protective impact that cultural values or norms have on the likelihood of resiliency to serious violent incidents within the school could significantly improve our knowledge of the risk of violence within institutions. Related, the organizational resiliency research literature is rather sparse; thus, the present study lacks any theoretical direction into the selection and assessment of the protective factors. The analyses in this study were shaped by data that were “available” rather than data that were “necessary” based on some theoretical articulation. Therefore, a fruitful line of resiliency research should explore the development of a coherent theory articulating how structural factors shape the cultural protective mechanisms within organizations.
Fourth, there was no attempt made in the present study to assess whether the importance of protective factors varied by different characteristics of schools. Drawing on the resiliency research at the individual level, it is instructive to observe that recent studies have begun to assess the differential impact across individual characteristics like gender (see Christiansen & Evans, 2005; Fagan, Van Horn, Hawkins, & Arthur, 2007; Hartman, Turner, Daigle, Exum, & Cullen, 2009) and race (Bennett, 2007; Zimmerman, Bingenheimer, & Notaro, 2002). Similar efforts at the organizational level should prove to be instructive in the development of policies to impact serious violence and theoretical articulations to explain organizational resiliency.
Taken as a whole, we are optimistic that research efforts will emerge seeking to understand why high-risk organizations, such as schools, are capable of creating or maintaining environments free of serious violent incidents. Arguably, the knowledge gained by this research is potentially critical in the development of policies and programs designed to prevent delinquency, crime, and other problem behaviors from occurring within vulnerable institutions. We caution readers to not oversimplify the results from the present study, which generally point to protective mechanisms that contradict resiliency research at the individual level. It would be prudent to explore longitudinal data at the organizational level to identify the complexities surrounding protective factors and organizational resiliency. Such clarification will ultimately be invaluable to scholars seeking to articulate theoretical attempts at explaining organizational resiliency.
Footnotes
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.