Share this article.

Assessing the link between violent political rhetoric and mass shootings

ArticleDetailDownload PDF

The familiar course of blame that follows every mass shooting in the United States is focused upon the shooter and tends to examine the context of their upbringing. But what about trigger words – something someone said that might have justified the reprehensible actions in the shooter’s mind? Dr William Nugent, a measurement specialist in the field of social work at the University of Tennessee-Knoxville, and his colleagues Dr Thereasa Abrams, Dr Anne Conway and Dr Andrea Joseph, have hypothesised that there could be a link between the violent rhetoric increasingly employed by US politicians and social commentators, and mass shootings. Their recent findings shed light on an understudied dimension of contemporary violence.
After every mass shooting in the United States, a familiar narrative rolls out: shock, thoughts and prayers, calls for tighter gun control, and questions about why the shooter did it. The list of contributing factors is varied but familiar. Invariably, media and commentators will invoke mental health issues; the subtext is that no one in their right mind would perpetrate such an act. But what if someone provoked them? What if circumstances were such that, in their mind, their actions were justified? What if highly influential persons gave them warrant to squeeze the trigger?

It’s difficult to ignore the growing political polarisation in the US, evidenced by the increasingly vituperative language of politicians and social commentators emboldened by unrestricted social media. Examples of violent political rhetoric can be found from numerous US politicians and political commentators. Can such political rhetoric incite violence? History might say that’s self-evident, but could there now be a link between political violent rhetoric in the US and mass shootings? A senior researcher in social work with a penchant for crunching numbers has uncovered some sobering correlations.

“For every 15 articles reporting VPR per 100,000 articles, there was an expected increase of one mass shooting per year.”

Dr William R Nugent is a professor in the College of Social Work, at the University of Tennessee-Knoxville, in the US. He conducts research on measurement issues and the measurement foundations of meta-analysis, a statistical analysis methodology that combines the results of multiple scientific studies. He has also worked with colleagues on examining sleep disturbance and suicidal ideation in war veterans and possible bias in the assessment of persons with burn injuries for possible drug or alcohol abuse. In 2021, together with colleague Dr Anne Conway, he turned his attention to the violent political rhetoric (VPR) in his country. The researchers wondered if there was a connection between VPR and mass shootings. It’s a fair point for consideration, but Nugent and Conway had two initial challenges: how do they measure violent political rhetoric, and what constitutes a ‘mass shooting’?

Figure 1. Top figure shows actual numbers of mass shootings per month January 2014 through September 2019. Bottom figure shows model-predicted monthly numbers of mass shootings January 2014 through September 2019.

Calling people cockroaches

There are varying definitions of violent political rhetoric. Essentially it is provocative phrasing used by people with influence that defames, dehumanises, or is derogatory towards specific groups of people that could justify harming them. An extreme case in point is a 1992 speech by Leon Mugesera, a senior Hutu politician in Rwanda, comparing the Tutsi minority to cockroaches. Broadcasters repeated the slur in the lead up to the 1994 Rwandan genocide. Violent political rhetoric can also be more subtle, such as the use of a statement by a politician that a target group, such as LGBTQ+ people, ‘should be very afraid’. Numerous examples of violent political rhetoric by US politicians and commentators can be found, such as statements to the effect that a particular person should be assassinated, beheaded, or put to death.

The media typically document such statements. So it was to the media that Dr Thomas Zeitzoff turned. Zeitzoff is a professor in the School of Public Affairs at American University in Washington DC and is particularly interested in why politicians use violent political rhetoric. In 2019 he published, in an online draft of a book, research results tracking the rates of VPR in the US as covered in the New York Times from 1851 through 2019. His findings showed very high levels of VPR during the US Civil War years, followed by relatively low levels until about 2015, when levels surged to magnitudes comparable to US Civil War years. This surge was contemporaneous with the beginning of the 2016 US Presidential election. To ensure these findings on VPR were not unique to the New York Times, Zeitzoff supplemented his data with that from eight different national newspapers between January 2014 and September 2019. These new results were consistent with the data from the New York Times. This measurement of VPR in the US seems to be the first, and currently the only, study of its kind, and so these findings became an important part of Nugent’s research.

Figure 2. Top figure shows rates of mass shootings per year 1966 through September 2019. Bottom figure shows model-predicted rates of mass shootings 1966 through September 2019.

To compare the violent rhetoric covered in the media with mass shootings, Nugent turned to mass-shooting data compiled by two Washington Post graphics reporters, Bonnie Berkowitz and Chris Alcantara. In 2019 they published a graphics-rich report that tracked reported mass shootings in the US between 1 August 1966 and September 2019. This provided a time period for Nugent’s research. It also stuck a stake in the ground on the definition of a ‘public mass shooting’: ‘in which four or more people were killed, usually by a lone shooter’. Their report didn’t include shootings tied to robberies that went wrong, nor did it include domestic shootings that took place exclusively in private homes. By Berkowitz and Alcantara’s admission, ‘a broader definition would yield much higher numbers’.

Unsurprising and unexpected outcomes

Nugent and colleagues then considered other speculated influences on mass shootings: gun ownership, imitation of other mass shootings, income inequality, and changes in gross domestic product. They collected data on gun ownership, examined time lags between mass shootings, and tracked income inequality and changes in gross domestic product as possible explanations for mass shootings. Armed with this information, Nugent did what statisticians do – he started crunching all the data, using multiple analysis models, looking for possible relationships with mass shootings. What came out was both unsurprising and unexpected.

Figure 3. Blue curve in figure shows actual monthly numbers of mass shootings January 2014 through September 2019. Red curve shows periodicity in mass shootings January 2014 through September 2019. Dotted line shows increase in mass shootings over time period January 2014 through September 2019.

It showed a positive link between income inequality and mass shootings and that, generally, periods of decreases in gross domestic product coincided with increases in mass shootings. In fact, income inequality accounted for 11% of the variation in mass shootings. These findings might support theories that mass shootings are often perpetrated by people under psychological stress and pressure. There seemed little connection between shootings and imitation effects. What was surprising was that the data failed to show any positive link between gun ownership and rates of mass shootings. This could have been because of limited available data on gun ownership, or, if accurate, it could suggest rate of gun ownership is uncorrelated with mass shootings – a finding that was not consistent with some prior research.

“Nugent and colleague’s work is remarkable in that it stands virtually alone in assessing a link between violent political rhetoric and mass shootings in the US.”

After controlling for these alternative explanations for mass shootings, as Nugent and colleagues expected, there was indeed a positive relationship between VPR and mass shootings. His research suggested that, after controlling for income inequality, changes in GDP, imitation effects, and gun ownership, as VPR increased mass shootings increased. The profile of the typical mass shooter is particularly interesting in this context: almost always male, generally young, and usually white. This would fit the description of those inspired by violent political rhetoric directed at ‘others’ who threaten their entitled positions: women, people of colour, immigrants, and LGBTQ+ persons.

To what extent might the incendiary rhetoric of politicians incite mass violence? studiostoks/

Redefining mass shootings

Intrigued by Berkowitz and Alcantara’s modest definition of a mass shooting, Nugent then wondered: what if the data included not only shootings where four or more people were killed but also where four or more people were killed or wounded by a shooter? Working with two other colleagues from the College of Social Work at the University of Tennessee-Knoxville, Dr Thereasa Abrams and Dr Andrea Joseph, Nugent repeated his statistical analyses, focusing on coverage of VPR from the eight national papers for the period January 2014 through September 2019, but with the expanded definition of ‘mass shooting’. The difference in numbers was significant. Considering only where four persons or more were killed, excluding the shooter, there were 39 mass shootings in the US during that period. Using the Gun Violence Archive definition of a mass shooting, four or more persons wounded or killed, that number skyrockets to 1,986.
The results largely echoed those of Nugent’s previous research, showing a clear correlation between the numbers of newspapers articles reporting on VPR and mass shootings. They found results consistent with a relationship between VPR and mass shootings: as VPR increased, mass shootings increased. His results also showed a periodicity, or cycles, in monthly mass shootings, with greater numbers in warmer months and lower numbers in colder months.

Nugent and his colleagues’ work is noteworthy in that it stands virtually alone in directly assessing a link between violent political rhetoric and mass shootings in the US. There is a major limitation to his studies, of which he and colleagues are fully aware: they focused purely on newspapers as sources of data on VPR and didn’t consider a broader, modern, media context that includes radio, TV, online talk shows, podcasts, and social media. It also doesn’t include smaller political rallies probably not covered by major media outlets. Nugent would like to see further research in this area and suggests study into grading the severity of violent political rhetoric – telling a crowd that certain others ‘had better tread lightly’ can’t be equated to telling the crowd those others ‘should be exterminated’. He is also urging those thinking about the implications of his research to not over-interpret their meaning: they do not show VPR causes mass shootings, but rather that there is a positive association: as VPR increases, mass shootings also appear to increase. The results are consistent with violent political rhetoric being a cause of mass shootings, but many alternate explanations need to be ruled out to establish causality. Much more research is called for and needed.

Victor Moussa/

But one thing is clear – when the next mass shooting occurs in the US, commentators should broaden the scope in their considerations for culpability.

You make recommendations for further studies based on your research; what study would you like to see first, and why?
I would like to see, most importantly, studies in which more data on violent political rhetoric are used. I would like to see studies in which the measurement of violent political rhetoric includes not just newspaper reports, but actual counts of violent political rhetoric heard on TV newscasts, in politicians’ speeches, on talk radio, on social media, and elsewhere. The severity of the violent political rhetoric also needs to be included. I would also like to see research covering longer periods of time than our research covered.




Research Objectives

Dr William Nugent and colleagues assess the link between violent political rhetoric and mass shootings.


  • Dr Thereasa (Teri) Abrams
  • Dr Anne Conway
  • Dr Andrea Joseph


Dr Nugent is Professor in the College of Social Work, University of Tennessee. He is former Associate Dean for Research and former Director of the PhD Program. He is a Fellow, Society for Social Work and Research. He does research on measurement and the measurement foundations of meta-analysis.

Dr William Nugent

College of Social Work
Henson Hall
University of Tennessee
Knoxville, Tennessee, USA

E: [email protected]
T: +1 865 974 6544

Creative Commons Licence

(CC BY-NC-ND 4.0) This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License. Creative Commons License

What does this mean?
Share: You can copy and redistribute the material in any medium or format
Related posts.