1. Grade Level Distinctions in Student Threats of Violence
Anna Grace Burnette, Tim Konold, and Dewey Cornell
University of Virginia
Virginia Youth
Violence Project
Abstract
Student threat assessment (TA) is a violence prevention practice that is expanding substantially with federal training funds from the STOP School Violence Act of Although student TA should take into consideration developmental differences across grades K-12, research in this area is limited. This study investigated grade-level differences in a statewide sample of 3,282 threat assessment cases from 1,021 Virginia schools. Student threats of violence significantly differed across grade level in 1) student demographics, 2) threat characteristics, and 3) outcome. Overall, these findings indicate that there are substantial differences in threats of violence across grades, and that schools should make grade-differentiated responses.
Rationale
TA has emerged as a specialized form of violence risk assessment that is conducted when a person threatens to carry out a specific targeted act within a relatively short timeframe. When used in school settings, TA teams must consider developmental factors such as the maturity and capability of students making a threat. Gender and race/ethnicity are also salient factors to consider.
Research Question
How do threats of violence differ in prevalence, characteristics, and outcome across grade?
Method
Virginia law mandated schools to report TA data in an annual school survey conducted by the Department of Criminal Justice Services. We examined statewide records for schools reporting threats across two school years ( and ). The analytic sample included 3,282 student threat assessment cases from 1,021 Virginia public schools: 548 (54%) elementary schools, 240 (24%) middle schools, and 233 (23%) high schools. The racial/ethnic breakdown was 51% White, 31% Black, 8% Hispanic, and 10% Other. Most threats (78%) were made by boys. Students making threats ranged from kindergarten to grade 12; the mean grade was 6 and the modal grade was 5.
Measures. Threat characteristics included whether the student threatened the use of a weapon (or was in possession of a weapon), student history of violence and/or prior disciplinary actions, threat to kill, threat to bomb the school, and threat to physically assault an individual without a weapon. Threat outcomes included attempts to complete the threat (attempted and carried out threats were combined due to low frequencies).
Analytic Strategy. Six logistic regression models demonstrated the association of threat characteristics or threat outcome (dependent variables) with student grade-level (continuous, independent variable). Each model controlled for student gender, race, and special education status. Clustered robust standard errors were used to account for student nesting in schools. The n for all charts is 3,268.
Results
Figures 1-6 depict distribution of threats by student demographics, selected characteristics, and outcome (threat attempted). Although 4th and 5th grade students made the most threats, they were less likely to attempt to carry out the threat compared to 9th grade students (Figure 6, Table 1). Linear and nonlinear logistic regression models examined the relations between predictors and student grade (Table 1). There was a linear increase in physical assaults across grades K-9 but a nonlinear pattern for attempted threats (Table 1).
As grade increased, girls were less likely compared to boys to threaten the use of a weapon (OR = 0.72, p < .01), to have a history of violence and/or disciplinary actions (OR = 0.59, p < .001), and to make a bomb threat (OR = 0.46, p < .001), but more likely to threaten physical assault (i.e., hit, fight, kick, strangle; OR = 1.50, p < .001). Black (OR = 0.59, p < .01) and Hispanic students (OR = 0.37, p < .001) were less likely to threaten to bomb the school compared to white students. Black students were also less likely to make a threat to kill (OR = 0.79, p < .05), but more likely to have a history of violence (OR = 1.52, p < .001) and to threaten physical assault (OR = 1.71, p < .001) compared to white students. Students in special education were two times more likely to have a history of violence (OR = 2.62, p < .001). Students were more likely to have a history of violence (OR = 1.03, p < .05) or threaten to physically assault someone (OR = 1.11, p < .001) as grade level increased. Student threats to bomb the school were not distinguishable by grade level.
Regarding threat outcome, the variability in students that attempted and/or carried out threats was attributable to curvilinear effects of grade level (OR = 1.02, p < .05). Specifically, threats were more likely to be attempted as students increased in grade level during early elementary and high school, but decreased after the 9th grade.
Discussion
An important limitation is that a retrospective and cross-sectional study cannot establish causal relationships. In addition, these data are based on student grade level rather than chronological age, which was not available for study.
Student threats of violence are distinguishable by grade level across threat characteristics and outcome. These findings can help TA teams recognize the role of developmental differences in the nature and frequency of threats.
Figure 1. Threats by Gender Across Grade Level
Figure 2. Threats by Race
OHI %
Regular Ed %
ED 38%
ED 38%
Figure 3. Threats by SPED
Figure 4. History of Violence
80%
Figure 5. Physical Assault
Figure 6. Attempted Threats
Table 1. Logistic Regression Odds Ratio for Threat Characteristics and Outcome
Notes
Further information and cited references are available from the author:
This project was supported in part by Grant #NIJ 2014-CK-BX-0004 awarded by the National Institute of Justice, Office of Justice Programs, U.S. Department of Justice. The opinions, findings, and conclusions or recommendations expressed in this report are those of the authors and do not necessarily reflect those of the U.S. Department of Justice or the Virginia Department of Criminal Justice Services
Predictors
Weapon Involvement
History of Violence
Threat to Kill
Bomb Threat
Physical Altercation
Attempted Threat OR SE
Student-level variables
Female1,2
0.72**
0.09
0.59***
0.06
1.14
0.13
0.46***
0.14
1.50**
0.18
1.32
0.31
Black3
0.94
0.11
1.52***
0.15
0.79*
0.59**
1.71***
0.21
1.52
0.34
Hispanic3
1.48
0.26
0.93
0.74
0.37***
0.17
1.15
0.24
0.79
0.35
Other3,4
1.06
0.19
0.67***
0.81
0.88
0.27
0.70*
0.64
0.28
SPED Status5,6
0.98
0.10
2.62***
0.23
1.13
0.22
1.22
1.62
Grade Level7
0.95**
0.02
1.03*
0.01
1.00
0.03
1.11***
Grade Level Squared --
1.02*
Note. *p < .05. **p < .01. ***p < Male is the reference group. 2Due to some schools failing to report the student’s gender, researchers controlled for unknown gender. 3White is the reference group. 4Other includes Asian, mixed race, other, or unknown. 5Students identified as non-special education is the reference group. 6Due to some schools failing to report the student’s special education status, researchers controlled for unknown special education status. 7Fourteen cases out of 3,282 (i.e., <1%) were missing. All results used cluster robust standard errors.