A Close Examination of the 2016 Dallas and Baton Rouge Police Killers: Identifying Potential Risk Factors and Influences for Copycat Violence

Two of the worst targeted attacks on American police officers in recent history occurred within eleven days of each other. Although it seems clear their proximity was not merely attributable to chance, the connection between these incidents, and the implications for understanding copycat violence, have never been fully explored. This study analyzes the perpetrators of these attacks from a “thresholds of violence” perspective, which suggests the first actor in a sequence is more likely to be disturbed and violence prone, while subsequent actors are typically less disturbed but more socially influenced. Results suggest the thresholds model has both merits and limits. The first attacker did have more psychological problems and violence in his past, and the second did seem more influenced by violent role models. However, there were also many similarities between them, and both attacked due to a combination of internal and external factors. If this study's findings are generalizable, higher risks of becoming a copycat offender may exist for individuals who have (1) personal similarities with previous attackers, (2) a history of psychological problems, (3) a history of interest in violent actors, and (4) recent escalation in their online behavior. Recommendations are offered for future research, offender profiling, and violence prevention.


Introduction
On July 7, 2016, Micah Xavier Johnson shot 12 police officers in Dallas, Texas, killing five and wounding seven others. On July 17, 2016, Gavin Long shot six police officers in Baton Rouge, Louisiana, killing three and wounding three others. These incidents constitute two of the worst targeted attacks on police officers in the United States, but the connection between them, and the implications for understanding copycat violence, have never been fully explored.
It seems highly unlikely that the proximity of these attacks, which occurred within eleven days of each other, is attributable to chance. Although there have been other cases in which multiple police officers were killed in a single incident, violence of this type is extraordinarily rare. Data from the Federal Bureau of Investigation (2019) indicate that the odds of a U.S. law enforcement officer being feloniously killed in a given year are approximately 1 in 16,000, and the vast majority of those incidents do not involve attempts at mass murder. 1 It also seems unlikely that Johnson's and Long's attacks are fully explained by a shared precipitating cause: the police shootings of Alton Sterling and Philando Castile earlier that month. It is true that both perpetrators considered those shootings unjust and both cited them as a reason for attacking (Fernandez et al., 2016;State of Louisiana, 2017). However, if police shootings were extremely rare in the United States, it might be easier to believe they caused this extremely rare outcome (two targeted attacks in which multiple police officers were killed). Unfortunately, approximately 1,000 civilians are shot and killed by police in the United States each year (Washington Post, 2020), and controversial cases that provoke public outrage have been all too common. This suggests there is not a simple cause-and-effect relationship between police shootings of civilians and targeted attacks on police that would fully explain Johnson's and Long's behavior.
A better explanation for the proximity of these attacks may be that the former inspired the latter. If so, Johnson was the role model and Long was the copycat. This is consistent with research showing that mass murder, mass shootings, and school shootings sometimes have a copycat element, whereby susceptible individuals who have been influenced by previous attackers eventually commit an attack of their own (Follman, 2019;Helfgott, 2015;Langman, 2018;Lankford & Madfis, 2018). It also fits evidence from Long's (2016) social media posts that shows he was aware of Johnson's recent attack and considered it morally justified.
To analyze the factors that led to Johnson's and Long's behavior-including the copycat element -this study will closely examine both perpetrators' lives. Data from primary source documents, government reports, and other reliable sources will be used to make biographical comparisons, identify confirmed influences on each attacker, and examine other relevant information. Johnson and Long will be analyzed from a "thresholds of violence" perspective, which makes predictions about how actors in a behavioral sequence differ from each other.

Thresholds of Violence
Research on copycat violence and contagion effects has primarily focused on a few key issues. For instance, scholars have established that high profile perpetrators can inspire and influence subsequent perpetrators (Follman, 2019;Helfgott, 2015;Langman, 2018); that high profile incidents may increase the likelihood of subsequent incidents (Kissner, 2016;Towers et al., 2015); and that outsized media coverage of perpetrators makes these effects more likely (Lankford & Madfis, 2018;Meindl & Ivy, 2018;Sidhu, 2017).
However, very little attention has been paid to how much violent role models and copycats resemble each other. Even if the behavior itself was copied, that does not clarify whether the role model and copycat were similar in other ways, or whether both appear likely to have attacked if their sequence had been reversed. In general, research on less extreme imitation (such as rudeness, smoking, binge-eating, and jaywalking) suggests that people are more likely to copy actors of the same sex and age (Meindl & Ivy, 2018). In addition, Langman (2018) cites several examples of mass shooters who said they were personally similar to the attackers who inspired their crimes. But it is unknown whether violent role models and copycats differ in other important ways, such as their propensity for violence. Gladwell (2015) offered his answer to this question in a widely read essay that extended  model of behavioral thresholds. The original model defines a "threshold" as the number of people who need to be doing something before another individual would act similarly.  hypothesized that people have different thresholds, and many people would only engage in certain behaviors if they witnessed a number of other people doing the same thing before them. For instance, a few people (labeled "threshold-zero") are "instigators" who would happily start a riot by throwing a brick through a window. But a person who would only throw a rock if he saw someone else throw one first would be labeled "threshold-one." In theory, the process continues all the way to the "threshold-100" person who would never throw a rock unless he witnessed one hundred people already doing that same thing.
Because people do not have to personally witness something to be influenced by it, Gladwell (2015) suggested this model could be extended to mass murder. He argued that years of school shootings across the United States are equivalent to a "slow-motion, ever-evolving riot" (Gladwell, 2015). From his perspective, the early perpetrators were deeply disturbed, violence prone, and did not require any role models, while subsequent perpetrators were less disturbed but more socially influenced by their predecessors. This hypothesis has never been tested. However, it is a frightening prospect, because if Gladwell (2015) is correct, the threshold for mass murder may have been changed by the presence of so many violent role models that "young men no longer need to be deeply disturbed to contemplate horrific acts." 2 If this "thresholds of violence" model applies to mass shooters of police, then Johnson, who was the first police killer in the 2016 sequence, would be the instigator and threshold-zero perpetrator. Although he is obviously not the first police killer in history and may have been influenced by prior killers (much like the instigator who starts a riot may have been aware of previous riots), this theory suggests that Johnson would have attacked regardless of whether anyone else had acted similarly, and he would have had more psychological problems and been more violence prone than Long. In turn, Long would be the threshold-one perpetrator who would only have attacked if someone else had already acted similarly. He would have had less psychological problems than Johnson and been significantly influenced by Johnson's actions.
Previous research on mass shooters has found significant variation in the severity of their mental health problems (Knoll & Annas, 2016;Lankford, 2018;Lankford & Cowan, 2020;National Council for Behavioral Health, 2019;Peterson & Densley, 2019), so it is certainly possible that Johnson and Long would differ in this way. Substantial variation appears in mass shooters' influences as well. Some perpetrators may not have needed any external sources of inspiration, as long as they had the motive, weapon, and opportunity (Gladwell, 2015). Other shooters have been profoundly influenced by previous attackers and worshipped them as gods, heroes, or kindred spirits (Follman, 2019;Langman, 2018).
Of course, it is impossible to know everything that may have influenced Johnson or Long, or whether Long would have attacked if Johnson had not attacked before him. However, this study will conduct a close examination of these two killers, their backgrounds, and the factors that appear to have influenced their attacks, in order to provide a more comprehensive assessment of their behavior.

Method
Given the aforementioned research, we might expect copycat attackers to be similar to their role models in their demographics (Meindl & Ivy, 2018) and some other characteristics (Langman, 2018). However, if Gladwell (2015) is correct, there are also important differences: the first attacker in a sequence is more likely to be deeply disturbed and violence prone, while subsequent attackers are typically less disturbed but more socially influenced. This study will investigate these possibilities by exploring three main research questions. First, to what degree were Johnson and Long similar? Second, did Johnson have more psychological problems and violent tendencies than Long? And third, did Long appear more socially influenced than Johnson? To answer these questions, this study will compare both perpetrators across a series of biographical variables, identify confirmed influences on each attacker, and closely examine Long's social media posts, to assess how their content changed on his pathway to violence. (Johnson is not known to have made similar posts, so the same type of analysis could not be conducted for him. 3 ) For the biographical comparison, a wide range of variables were examined. These included (1) basic demographic factors (age, sex, race), (2) background factors (adverse childhood experiences, education, marital history, military deployment, military discharge, criminal record, and prior signs of violence), (3) psychological factors (substance use, paranoia, anger issues, mental health problems, and psychiatric medication), and (4) attack factors (targets, number of victims killed, number of victims wounded, and attack resolution).
For the analysis of confirmed influences that may have contributed to these perpetrators' decisions to attack, information was gathered on people, groups, and incidents that Johnson or Long are known to have been interested in. Only influences that appear directly related to Johnson and Long's expressed reason for attacking-unjust police shootings of Black civilians-were tracked. Naturally, it is impossible to document every source of influence on someone's thoughts or behavior. However, there is an important difference between confirmed influences and possible influences, so it is helpful to document the former. As examples, there is direct evidence that both Johnson and Long were influenced by the Nation of Islam, which is a designated hate group according to the Southern Poverty Law Center (2020). But there is only direct evidence that Long was influenced by police killer Christopher Dorner-not that Johnson was. Although it cannot be ruled out that Johnson was influenced by Dorner, no evidence could be found which supports that link.
In addition, Long's Twitter posts were closely examined to assess how their content changed on his pathway to violence. All 1,114 posts Long made between October 22, 2015 (when he joined Twitter) and July 17, 2016 (his attack date) were examined for content related to violence. This included Long's own statements, his public posts specifically directed at other Twitter users (which start with the "@" symbol), and his retweets and posts of content from other sources. When users retweet or recirculate content without offering additional commentary or correction that typically indicates the user's agreement with the content. As an example, because Long reposted the quote, "Yes The Govt is a hate Group, They Hate Black People" one week before his attack without additional commentary, it seems reasonable to interpret that as his agreement with the sentiment.
This content was analyzed using an inductive method (Thomas, 2006). The raw qualitative data (Twitter posts) were searched for a narrower set of data related to the specific research focus (posts related to violence), and then that narrower set of data was closely examined for recurring themes and patterns (Thomas, 2006). Because this was an inductive process, it did not constitute formal theory testing in which previously identified themes were sought among Long's posts. The advantage of an inductive approach is that it allows for both expected and unexpected patterns in Long's online posts to be identified.

Data Sources
Data were drawn from a wide variety of sources, including primary source documents, government reports, and news reports. The primary source documents were Long's (2016) Twitter posts and his YouTube page. The relevant government reports were the U.S. Department of the Army's (2014) investigation of concerns about Johnson while he was in the service, and the State of Louisiana's (2017) report from the district attorney about Long's behavior and motives. Finally, news reports were used to document additional biographical information about Johnson and Long. News reports were only used if they contained information from law enforcement, government sources, or people with direct knowledge of Johnson or Long.

Results
Overall, this study's results indicate that police killers Micah Johnson and Gavin Long had a tremendous amount in common. As shown in Table 1, demographically, they were both in their twenties, male, and Black. Both of their backgrounds included parental divorce, college struggles that ended in them dropping out, military deployments of less than one year, and clean criminal records. Psychologically, both appeared to have substance use problems, paranoia, anger issues, and mental health problems, and both were prescribed psychiatric medication.
Some of the details are stunning in their similarities. For instance, Johnson told a doctor "I feel like I can't trust all of these strangers around me" (Burke, 2016), while Long told a doctor he avoided crowds and experienced "suspiciousness of others" (Associated Press, 2016b). As another example, Johnson told a doctor "It's hard for me to be around other people and I am so angry and irritable" (Burke, 2016), while Long's family recounted that he became angry easily and it caused tension with his former wife (Associated Press, 2016b). Additionally, both perpetrators believed they had post-traumatic stress disorder, and both were prescribed anti-depressants (Associated Press, 2016b; Burke, 2016;CBS News, 2017). Finally, both seem to have intended to die following their attacks on police, and both were killed by law enforcement (Fernandez et al., 2016;State of Louisiana, 2017).
On the other hand, Johnson did show more signs of being disturbed and violence prone than Long. For instance, Long was honorably discharged from the military, but Johnson was expected to receive an "other than honorable discharge" due to his acts of sexual harassment and other concerning behaviors. As it turned out, Johnson actually received an honorable discharge, but this was apparently due to an administrative error (Department of the Army, 2014; Starr, 2016). Furthermore, whereas no evidence of violence could be found in Long's background, Johnson had several concerning instances. He received medical treatment after punching out a car window with his hand; while in the military, he stole ammunition for a grenade launcher and kept it in his belongings; he regularly practiced combat techniques in his backyard; and he was asked to leave the New Black Panthers Party after pressuring them to acquire more guns and ammunition and take the fight to the enemy (Department of the Army, 2014; Emily, 2016;Hlavaty, 2016;Vanden Brook & Locker, 2016).
Among these instances, the most details are available from Johnson's time in the military. At one point, an officer investigating Johnson was so concerned that he "asked if for safety reasons we should relieve [PFC] Johnson of his firearm and any bladed weapons in his possession," and was told by his sergeant that it was "a good idea" (Department of the Army, 2014, p. 4). In turn, after the officer found a "M169, 40 mm High Explosive Dual Purpose, Mk19 grenade" in Johnson's sleeping bag, leaders expressed concern that "PFC Johnson was storing an explosive article in a barracks facility where soldiers lived" even though he "did not have any reason to be in possession of the explosive device" (Department of the Army, 2014, p. 10, 13).
This study also found many confirmed influences for both Johnson and Long. As shown in Table 2, they had several in common: both expressed interest in Malcolm X, the Nation of Islam, and the police shootings of Alton Sterling and Philando Castile.
However, Johnson's other confirmed influences that appear related to his attack were either cases of Black victimization-such as the Rodney King beating and the Martin Luther King Jr. assassination-or Black militant groups-such as the New Black Panthers Party, the Black Riders Liberation Party, and the African American Defense League. Notably, the African American Defense League encouraged violence against police officers on its Facebook page in the days He told a doctor "I feel like I can't trust all of these strangers around me" (Burke, 2016) He told a doctor he avoided crowds and experienced "suspiciousness of others" (Associated Press, 2016b) Anger issues He told a doctor "It's hard for me to be around other people and I am so angry and irritable" (Burke, 2016) His family recounted that he became angry easily and it caused tension with his former wife (Associated Press, 2016b) Mental health problems He showed symptoms of post-traumatic stress disorder, and sought treatment for anxiety, depression and hallucinations (Burke, 2016) He told doctors he had post-traumatic stress disorder, but he was diagnosed as having an adjustment disorder with depressed mood (CBS News, 2017) Psychiatric medication He was prescribed an antidepressant, anti-anxiety medication, and sleep medication (Burke, 2016) He was prescribed an antidepressant (Associated Press, 2016b) before Johnson's attack-and Johnson was known to have liked and followed their page (Griffin et al., 2016). Specifically, the group posted "The Pig has shot and killed Alton Sterling in Baton Rouge, Louisiana! You and I know what we must do and I don't mean marching, making a lot of noise, or attending conventions. We must 'Rally The Troops!' It is time to visit Louisiana and hold a barbeque. The highlight of our occasion will be to sprinkle Pigs Blood!" (Griffin et al., 2016). Although it is impossible to know for sure, this may have been one of the most powerful influences that led Johnson to act. By contrast, Long's other confirmed influences were qualitatively different from Johnson's. Although Long did express interest in an obsolete African-American self-defense group from the 1960s, in the George Zimmerman/Trayvon Martin incident from 2012, and in another police shooting of a Black civilian from July 2016, unlike Johnson, Long is known to have been interested in other Black perpetrators who used lethal violence. In particular, his confirmed influences include Nat Turner, who led a slave rebellion in which more than fifty people were killed, Christopher Dorner, who murdered police officers and their family members, Vester Flanagan, who killed a reporter and cameraman on live television, and of course, Johnson, who committed the mass shooting of police officers in Dallas. 4 Findings from the third part of this study-the close examination of Long's Twitter posts-show a clear escalation in frequency and severity of concerning content over time (see Table 3 and Figure 1). Over seven and a half months-from October 22, 2015 to July 4, 2016-only ten of Long's posts were related to violence, and many of those implied he did not approve of it. But in less than 2 weeks after Alton Sterling was killed, he made 24 posts related to violence, and many of those endorsed it. Overall, the content of these posts had two major, repeated themes. First, they show that Long (2016) increasingly felt Black people face an existential threat, so much that he warned they would "become the Native Americans … EXTINCT" unless they did something drastic. Second, he increasingly believed violence was the best response to this threat-better than negotiations or peaceful protest.
Long's Twitter posts also show that Johnson was a role model for him. The most times Long ever posted about violence was on July 8, 2016, which was when Johnson's attack ended and he was widely covered by the media (see Figure 1). For example, Long tweeted a news story about Johnson's mass murder of police officers in Dallas, along with his excited discovery: "The Shooter was NOT WHITE, He was one of us! # My religion is Justice" (Long, 2016). This was accompanied the same day by him posting a link to an article entitled "Micah Xavier Johnson: 5 Fast Facts You Need to Know," which discussed Johnson's anger about recent police shootings, his military background, his online connections to Black separatist hate groups, and more. That  Confirmed (Long, 2016) Micah Johnson, who shot 12 police officers (and killed 5 of them) July 7, 2016 n/a Confirmed (Long, 2016) Alva Braziel's shooting by police July 9, 2016 n/a Confirmed (Long, 2016) Note: these are confirmed influences for which there is direct evidence that Johnson or Long expressed interest in the person, group, or incident. Only influences that appear directly related to Johnson and Long's expressed motive for attacking-unjust police shootings of Black civilians-were tracked. This is not an exhaustive list; it is impossible to document every source of influence on someone's thoughts or behavior. a SPLC = Southern Poverty Law Center (2020), which provides a list of designated hate groups. b The first paragraph of Long's suicide note was adapted from Christopher Dorner's manifesto.  Repeated theme: violence is the best response to the threat Black people face (and is better than nonviolent methods).
(continued)  same day, Long also posted a retweet indicating his support for Johnson: "Blk ppl, please stop apologizing 4 the actions of other ppl who happen 2 be Black. Bcuz white people arent apologizing 4 these race soldiers." Here Long implied that Johnson's crimes should be a source of celebration, not regret.
To that end, this study explored three main research questions. First, to what degree were Johnson and Long-who committed two of the worst targeted attacks on American police officers in recent history-similar to each other? Second, did Johnson have more psychological problems and violent tendencies than Long? And third, did Long appear more socially influenced than Johnson? Although this study's results were presented in the prior section, they warrant further discussion.

Personal Similarities
This study found a remarkable number of similarities between Johnson and Long. This included their basic demographics; some of their childhood, educational, and professional experiences; some of their psychological problems and treatment; and the nature of their attacks. They also shared some confirmed influences, such as Malcolm X, the Nation of Islam, and the police shootings of Alton Sterling and Philando Castile.
Some of these similarities may be a coincidence, but others may have contributed to Long's decision to copy Johnson. As noted earlier, if the "thresholds of violence" model applies, then Long may be the threshold-one perpetrator who would only have attacked if someone like Johnson had already acted similarly. Long posted an article on Twitter that included details about Johnson's background, so he may have been excited that someone like him committed such a high profile attack. This would fit with research suggesting that in general, people are more likely to imitate behavior if they consider the actor similar to themselves (especially according to age and sex) and if they believe the actor was rewarded for that behavior (Meindl & Ivy, 2018).
Another possibility is that Long was actually a threshold-four perpetrator who would not have attacked unless he had been influenced by several role models he considered similar to himself. As this study found, Long was interested in Nat Turner, Christopher Dorner, Vester Flanagan, and Micah Johnson-all of whom killed for reasons they believed were righteous. Therefore, Long may have viewed himself as not merely proceeded by Johnson, but instead as following a wellestablished path of people who responded to injustice with violence.
Like Long, all four of these predecessors were Black men. Their average age was 33 years, while Long was age 29 years. And from the perspective of those who approved of their violence, these four predecessors could be perceived as having been rewarded with fame and legacy. Notably, fameseeking is a common goal of mass shooters (Langman, 2018;Lankford, 2016;Silva & Greene-Colozzi, 2019), and it seemed very important to Long (2016), given his constant efforts at self-promotion on social media as a self-described "alpha preneur." If Long was similar to other mass shooters, he would not have been deterred by the fact that he would receive negative attention or that his "legacy" would not be fully forged until after his death. After all, that is what he had already observed with Turner, Dorner, Flanagan, and Johnson.

Psychological Problems and Violent Tendencies
Despite the similarities between these attackers, the "thresholds of violence" model seems to have correctly predicted that Johnson would have more psychological problems and violent tendencies than Long. Although both perpetrators struggled with mental illness, anger, and paranoia-and both were prescribed psychiatric medications-Johnson also had hallucinations, his alcohol abuse seemed more consistent, and he needed anti-anxiety and sleep medications, so his problems appeared more severe.
In addition, Johnson had multiple signs of violence in his past and seemed more like the proverbial "ticking time bomb" for whom the question was not "if" he would explode in violence, but "where" and "when." Post-attack investigations suggest he was stockpiling weapons and practicing combat techniques well before the police shootings of Sterling and Castile (Emily, 2016). By contrast, Long had no signs of past violence that this study could find.
However, a few comments on Gladwell's (2015) "thresholds of violence" model are warranted. As a reminder, he warned that because the threshold for mass murder in the United States has changed with so many recent incidents, "young men no longer need to be deeply disturbed to contemplate horrific acts" (Gladwell, 2015). If he was suggesting that normal, psychologically healthy individuals are now apt to commit mass murder because the "slow-motion riot" is in full effect, this study's findings do not support that claim. Maybe Gladwell is simply wrong and was overstating his case. Or maybe he is right, but Johnson and Long are both low threshold attackers, and more high profile attacks against police would need to occur before copycats would adhere to the pattern Gladwell believes is present among school shooters. More broadly, research on public mass shooters suggests they almost always have mental health problems or suicidal intent (Lankford & Cowan, 2020).

Social Influences
The "thresholds of violence" thesis also suggests Long would be more socially influenced than Johnson. Overall, this study found mixed evidence for that hypothesis. More of Long's confirmed influences were violent role models, and Long did seem profoundly affected by Johnson's behavior. In addition, the escalation in Long's Twitter posts over the final two weeks of his life suggests he was highly susceptible to the influence of recent events and the discourse surrounding them. His repeated themes about Black people being persecuted to the point of extinction, and violence being the best response, may have been exacerbated by the social media environment where he often communicated.
However, it is not clear that Johnson was less socially influenced than Long, even if his influences were qualitatively different. Johnson did not appear to be copying or imitating previous attackers, but as reviewed earlier, he was influenced by historical acts of Black victimization and by Black militant groups. Much like it is impossible to know for certain whether Long would have attacked if Johnson had not preceded him, it is impossible to know whether Johnson would have attacked police officers if not for the African American Defense League's encouragement to spill "Pigs Blood." Although the "thresholds of violence" model has merits, it is overly simplistic to define someone's propensity for violence by only counting the number of people who previously engaged in the same behavior. Most crime is driven by the interaction of internal and external factors, and the presence of role models is only one of many external variables that could have an important effect. As this study demonstrated, Johnson and Long were both influenced by historical figures, militant groups, police shootings of civilians, and more-all of which may have contributed to their decisions to attack. Furthermore, although some criminal phenomena, such as public riots, may involve all actors being in the same physical location and therefore witnessing each other's behavior, in many other cases, people can be profoundly influenced by things from other locations or eras. As just one example, both Johnson and Long were influenced by Malcolm X, even though he died more than twenty years before either of them were born.

Conclusion
Although Micah Johnson and Gavin Long never met or communicated with each other, they committed two of the worst targeted attacks on American police officers in recent history-within the span of eleven days. This study provided a close examination of these police killers, their backgrounds, and the factors that appeared to influence their attacks.
It remains to be seen whether this study's findings apply to copycat attackers in general, such as mass shooters, school shooters, lone wolf terrorists, and other targeted killers. However, if other copycat offenders do have a similar psychological and behavioral profile to Long, it suggests that higher risks may exist for individuals who have (1) personal similarities with previous attackers, (2) a history of psychological problems, (3) a history of interest in violent actors, and (4) recent escalation in their online behavior. Research with much larger samples would be needed to determine whether this study's findings are generalizable, and hopefully, this study's design can help guide those efforts.
For instance, the Columbine school shooters have inspired more than one hundred imitators (Follman, 2019), and there is anecdotal evidence that they are often school-aged White males who feel marginalized, just like the original killers (Langman, 2018;Lankford, 2016;Silva & Greene-Colozzi, 2019). However, this could be studied empirically to determine whether similarities appear significantly more often than would be expected due to chance. If so, this could have powerful implications for law enforcement and security officials, who would benefit from knowing whether copycat killers are likely to have similar profiles to their role models.
As another recommendation, given this study's findings regarding Long's escalation in frequency and severity of concerning Twitter posts prior to his attack, future research could test this relationship between social media behavior and dangerous behavior more broadly. It may be that social media posts provide an extremely valuable window into the minds of some people who are becoming agitated and unstable. That possibility could be tested in large samples across different types of dangerous behavior. The key would be to use both test and control groups, so that cases in which escalation in social media posts leads to nothing bad would also be examined, and the risk of false positives would be considered. If a valid predictive relationship were found, this could also be extremely valuable for law enforcement and security officials.
Finally, researchers should continue to investigate the complex nature of social influences on violent behavior. Contagion and copycat effects have become subjects of great interest, but less attention has been paid to the influence of historical influences than contemporary ones. As this study has shown, today's mass shooters and police killers can be influenced by other perpetrators of the same type and era, but also by people whose entire lives occurred before they were born. A better understanding of how internal and external factors interact to produce violence could greatly enhance future efforts at prevention.

Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Ethical Approval
This study analyzed publicly available information and did not contain any interaction with human participants or animals.