The following is an abridged version of an interview conducted by Václav Walach and Petr Kupka for publication in the Czech Journal of Criminology. The Czech version may be found here.

‘School mass shootings in Central and Eastern Europe are on the rise’

An interview with Dr. Alexei Anisin.


The mass shooting at Charles University in Prague on December 21, 2023 is one of the most tragic events of its kind in Europe. However, it is not an isolated event in Central and Eastern Europe (CEE). Research conducted by Alexei Anisin, Dean of the School of International Relations & Diplomacy at Anglo-American University, and his team shows there have been at least 76 mass shootings in the region in recent years.

In the following interview with Anisin, we compare crimes of this nature in the CEE with cases from the US, while touching on Anisin’s research methods. We likewise consider possible reasons and motivational factors for the perpetrators of mass shootings, including mental illness and the copycat phenomenon.

We also discuss Anisin’s testing of the influential cumulative strain theory, formulated by Jack Levine and Eric Madfis in 2009 to explain the emergence of mass shootings in schools. Finally, we move on to some possible preventive measures, in particular the regulation of firearms possession.

Anisin has published numerous studies on mass shootings in recent years, including the monograph Mass Shootings in Central and Eastern Europe. In addition to research on mass shootings, Anisin specializes in the study of various forms of political violence, protest social movements, and political mobilization, using both quantitative and qualitative methods.

The Interview

Václav Walach & Petr Kupka: You may be the only researcher who has written on mass shootings in the CEE. How did you come to this topic?

Alexei Anisin: I got into researching mass shootings through the study of rare atrocities. I had a background in the study of rare forms of violence and atrocities, namely in the state repression and human rights fields. Hence, stepping over to inquiry on mass shootings was not a big change. I was a PhD student in the Department of Government [at the] University of Essex, where I became interested [first] in human rights violations that were carried out by government against protesters and then in genocide.

In 2016, I first started researching mass shootings, collecting new data on US cases, and quickly learned that there is a complete gap in knowledge when it came to [such cases in] the CEE. I explored the American context and approached this topic from a social scientific perspective, broadly speaking, as I had trained in different qualitative and quantitative methodological approaches at the Essex Summer School in Social Science Data Analysis.

Regarding the CEE, I think we always gravitate towards familiar contexts. It would be difficult for me to research East Asia, for example, as I have never been there and do not know the local languages. But I lived and worked in Transylvania, Romania for a year as a visiting lecturer. Now, I am based in Prague. Then, there is the prospect of publications, grants, etc. With recent discussion about the concept of a “Global East” and its relative neglection in criminological literatures (Piacentini & Slade, 2023), I believe my research can offer a glimpse into regions that have indeed been neglected.

VW&PK: Can you tell us more about your research on mass shootings? What have you done so far? Which methods have you applied and to what results?

AA: I have published a range of studies pertaining to mass shootings in both the US and in the CEE. Perhaps the most intriguing of these studies are those in which comparisons between samples of cases from these regions have been made. I think there are interesting similarities between contexts, with the average age of mass shooters being around 34 years old. At the same time, there are interesting differences, with the US experiencing more ideologically motivated attacks and more fatalities and injuries overall.

In terms of the methodological approaches I have taken, I have drawn from standard statistical/quantitative methodologies, including regression analysis and hypothesis testing. I have also utilized fuzzy-set qualitative comparative analysis and case study methods to address cases.

Overall, I would say that my methodological approach varies and attempts to tackle research questions through both standard or traditional and non-standard techniques. I have used quantitative and qualitative approaches in my research on mass shootings as well as Fuzzy-Set Qualitative Comparative Analysis (QCA).

This latter method helps to untangle causal complexity and offers a nice middle ground between quantitative and qualitative methods. It aims at identifying different combinations of variables that can explain why an outcome occurred or did not. It is also based on different mathematical assumptions (Boolean Algebra) than regression analysis.

VW&PK: Your CEE dataset contains 76 cases that occurred in 15 countries from 1993 to 2021. How did you select them and what kind of difficulties have you faced during the data construction?

AA: Our dataset comprises both completed and attempted cases, with the former indicating cases with four or more fatalities and the latter less than four, zero included. As is common in mass shootings studies, we excluded cases such as drug- or gang-based homicides and domestic violence cases.

Eligible cases were identified using various databases and search engines to cover as wide a digital open access spectrum as possible. This procedure most commonly yielded media reports, but we also drew on the files of the police, courts and other official authorities that had made them public.

For each case, we were interested in two sets of characteristics. We first gathered information on the offender: his sex, his motivation, what kind of events he had experienced in life, including different strains, mental health issues, etc. I say ‘he’ because there are no women in the CEE sample and just 2% of them in the US sample.

Then, we focused on the features of the actual attack, such as what and how many guns were used, the extent to which the attack was planned, etc. This procedure led us to exclude cases for which we could not find the observed information. There were about a dozen or more cases from Russia that featured probably ideologically motivated attacks against government institutions. But due to lack of information, we had to omit them. Overall, I think that the dataset is pretty close to what we may perceive as the entire universe of incidents commonly thought of as mass shootings. The data are being updated as time goes on.

VW&PK: How comparable are data from the CEE to those from the US? And how difficult is their study?

AA: They are certainly comparable, but there are limits. Above all, there are much fewer cases in the CEE than in the US, which reduces the possibilities of generalization. From a statistical point of view, the US data are better because there are so many more of them. The more cases, the easier it is to make inferences, confidence intervals, and so on. Since the US cases have been researched so heavily by so many people over such a long period of time, there is simply more information.

But otherwise, I do not see any significant differences. The research strategies that I described earlier were pretty much the same. We did not invent anything new in terms of data construction. We just tried to emulate what had already been done by other researchers.

VW&PK: There are, however, some contextual differences. Can you describe them?

AA: In addition to, say, income disparities and pronounced differences in civic association, which include populations in CEE regions spending less time volunteering and having less trust in their neighbors, there are other important factors, such as having fewer self-expressive or post-materialist values, including acceptance and tolerance of: immigration, LGBT+ preferences, same-sex marriage, gender equality, and environmental concerns.

There are also more easily observable and obvious differences in the number of firearms in society and the number of homicides committed with them. There are much fewer civilian-owned guns on average in the CEE. The only areas of Europe that have high rates (comparatively speaking) of gun ownership are the Balkan states. And this I think leads to the next point: except for Lithuania, Ukraine, and Russia, the US has a much higher homicide rate than countries in our dataset.

The number of gun-inflicted homicides in countries throughout most areas of the sample, including Czechia, have been historically very low. We can thus say that mass shootings in the CEE represent a much greater diversion from general criminal tendencies which are already globally low when it comes to violent crime.

Nevertheless, my data point to a change over time. In the 1990s, mass shootings were more localized and involved more elderly perpetrators. The perpetrator usually killed in his area of residence, be it a village or a smaller town. As time went on, there were more younger mass shooters in schools, including a university in Pécs, Hungary (2009); Kerch, Crimea (2018); schools in Brześć Kujawski, Poland (2019) and Kazan, Russia (2021); and in Serbia (Mladenovic and Smederevo, 2023), among numerous others.

VW&PK: How do mass shootings in CEE differ from those in the US?

AA: One of the first differences I have observed is that there are less ideologically driven cases in CEE than the US. I believe that this has something to do with a process of depoliticization observable in the post-communist period.

Obviously, there is a lot of variation within the region. Some countries have experienced a military conflict. There is populism or democratic backsliding, depending on what group of scholars we refer to. There are a lot of different political processes ongoing. But I would still stand by the claim that generally, there is far less societal and especially cultural polarization in CEE than in the US.

A second difference is that cases in the CEE tend to be more targeted towards an area that the shooter was familiar with, such as a residence, a workplace, a school, etc. In the US, there have been a greater frequency of attacks that aimed at the general public.

A third difference is that American mass shooters, on average, have been observed to have suffered from a higher degree of mental illness than those in the CEE. It is important to note that our ability to observe mental illness is limited, as many offenders were previously not diagnosed, and around half of all mass shooters either take their lives or get killed at the scene of their crime by authorities.

VW&PK: Your book Mass Shootings in Central and Eastern Europe explains this difference in the presence of mental health issue by the absence of the universal healthcare system in the US. But cannot this be also explained by other factors such as, perhaps, a greater willingness to share such information in the US? The openness about one’s health condition may differ depending on context.

AA: Building off of what I mentioned earlier, mental health is a thorny issue and, as researchers, we have to follow a careful methodological line when making claims about mental illness.

In my study with Joel Alfredo Capellan, we referred to mental illness via the concept of a “reported history of mental disturbance.” Here we noted that it is not only difficult to generalize about mental illness across different historical eras, but also that formal definitions of mental illness have varied and evolved.

If you look at the DSM scale published by the American Psychological Association, the scale of what constitutes mental illness expands every 10 years. It runs the risk of constituting a slippery slope, as more and more people are considered mentally ill.

With this being said, we coded this variable based on information that predated a given mass shooting. Such information must have either indicated a formal diagnosis or a suggested history of mental disturbance by family members and friends. The types of adversities that we considered to qualify included at least one of the following: schizophrenia, depression, anxiety disorders, addictive behavior, bipolar disorder, obsessive-compulsive disorder, and posttraumatic stress disorder.

VW&PK: What does this tell us about the exact role of mental health issues in terms of the causal mechanism behind mass shootings? Do only [so-called] “crazy people” commit mass murders, as we often hear?

AA: It is sometimes said that mental illness is a mechanism that leads to mass shooting. I do not think it is so simple. I see it more as a part of a larger causal process that is very complex. Looking at similar events such as gang-based homicides, those are not really driven by mental illness.

With mass shootings, it is very difficult to have any causal attribution to mental illness. About 12% of American adults, 30 million people, are on antidepressant prescription medications. They are technically mentally ill, but very few are at risk for committing a mass shooting. The only thing I really would stand by is that mental health issues are part of a causal process that can be influential in bringing about certain types of mass shootings (not all), rather than a single mechanism.

VW&PK: You also found that US mass shootings are characterized by more frequent injuries compared to CEE mass shootings. To what extent is this related to the greater availability of firearms in this context?

AA: I think this rather pertains to the more methodical nature of US attacks. There have been numerous cases where a great deal of weapons were used by attackers. On the other hand, one particular factor that can explain this was mentioned earlier—US mass shootings tend to be targeted at the general public to a higher frequency than those in the CEE

This inevitably means that mass shooters will go to a very public and populated place to carry out their attack. The deadliest mass shooting in US history happened in a hotel casino in Las Vegas, where the perpetrator opened fire on a music festival taking place below him. It was actually similar to the December 21 attack in Prague, where the shooter also brought many different weapons in a large bag.

So, the availability of firearms and ammunition is important. Regulations matter, preventing people from getting a gun easily via a permit would certainly help. In the US, no such permits exist and, in many states, the third-party marketplace for guns is completely unregulated.

However, I really think that the main difference lies in the degree of preparation: whether an offender created plans, whether he drew out the area he was going to attack, whether he scouted the area before the attack, whether he dressed in a certain way, etc. It is typically cases with large-scale preparation that result in more frequent injuries and even fatalities.

VW&PK: One of your studies examined six different pathways to mass shootings in the CEE. Can you explain what you mean by this and why is this significant?

AA: This was based on the qualitative comparative analysis (QCA) methodological approach. A pathway is a conjunctural combination of variables that are assumed to be sufficient with relation to one another’s presence in bringing about the outcome (dependent variable).

\What this means is that different cases are explainable by different pathways. Pathways are comprised of the presence and absence of different variables which are observed and coded across all cases in a given sample. Through a software application, data are analyzed in attempt to find the most salient pathways that can explain the highest proportion of cases in the data under attention.

In my study using this method on CEE vs. US cases, I identified six solutions for the CEE sample of cases and nine for the US sample. None of the CEE solutions featured the sole presence of mental illness. Mental illness is much more prevalent across the US sample of mass shootings as it is present in six out of nine solutions.

Another difference can be observed in the discovery of pathways featuring ideological motivation. In the CEE sample, only one solution featured the presence of ideological motivation, whereas in the US sample, five out of nine identified pathways feature the presence of this condition. The condition of group grievance was also found to be present in five out of nine pathways, which is clearly more important than it is in the CEE sample (two out of six).

These findings revealed that there is more complexity inherent to US mass shootings, as reflected by the total number of identified solutions. Conceptually, such an approach helps to capture variance across a sample of cases that traditional statistical methods are not suited to capture.

These findings are significant because they can enable us to gauge the extent to which specific pathways (e.g. ideologically motivated mass shootings) are prominent or not. They also can enable us to understand potential interactional effects between variables in ways that are not trivial.

In other words, QCA helps us to untangle the causal complexity behind certain pathways to mass shootings. Where regression analyses aspire to isolate one factor in a model and find significance (while holding others constant), QCA takes into account that there may be multiple combinations of factors.

VW&PK: In your book, you tested the Cumulative Strain Theory (CST). What is it and how did the test go?

AA: This theory stems back to a study by Jack Levin and Eric Madfis (2009) who attempted to explain school shootings. It has deeper roots in Robert Agnew’s (1992) General Strain Theory. Strain encompasses different negative experiences that arise in a person’s life. If they are not adequately addressed and dealt with, they can turn into chronic strain.

CST features five different stages, where (1) chronic strain is the first one. Next comes (2) uncontrolled strain which has origin in Hirschi’s social control theory from the late 1960s and is probably the most significant factor that can lead a person to carry out harm either on himself or others. This is followed by (3) acute strain which tends to serve as a catalyst or precipitant to not only mass shootings, but all kinds of mass murder. Then the final two stages are the (4) planning stage, in which an individual plans an attack, and the (5) event stage which comprises the actual attack.

This theoretical framework informed four different hypotheses which were tested via logistic regression. The outcome was whether a case resulted in four or more fatalities (it compared attempted versus successful mass shootings in CEE contexts). Results revealed that the theory fell short in explaining the outcome.

Two variables were found to be statistically significant—suicide (whether an offender killed themselves at the scene of the attack) and group grievance. The odds of a mass shooting arising in cases where offenders take their own lives at the scene of attack are nearly four-fold in comparison to those where suicide did not occur.

In accounting for these findings, I considered the following:

“That the coefficients and odds ratios for suicide were the strongest of any of the variables tested in the analyses…is a surprising result and one that needs to be carefully considered alongside CST. Could it be the case that a potential stage is missing in the framework of CST? Why did cases in which offenders committed suicide at the scene of the attack result in more fatalities than those in which they did not?

One potential explanation is that there is a stage that is missing in CST that could plausibly be fitted somewhere between the third and fourth stages—a “motivation” stage may be needed to compliment the framework considering that it does not enable us to understand the motivations behind why individuals carry out mass shootings” (Anisin, 2022b: 124).

VW&PK: The motivation stage seems to infuse CST with a more subjective element. What do we know about motivation from your research on mass shootings?

AA: The motivation factor is definitely the most subjective, which is probably why it does not form an actual stage in the model. In my book, I considered different motivating factors, including ideological motivation, mental illness, group grievance, fame-seeking, and suicide.

From an observational point of view, when coding data and conceptualizing how empirical occurrences can be included in research design, motivation(s) of mass shooters are quite difficult to deal with. The reason for this is twofold: First, we can only find out about motivation after an attack, and around half of mass shooters either take their own life or get killed by authorities, which tends to leave us with a number of cases in which motivation is unattributable. Second, there are also a number of cases that have not been driven by any particular motivating factor. In the Las Vegas mass shooting [, for example,] authorities still have no idea of a motive.

Similarly, when it comes to mental illness, psychiatric diagnosis is rather difficult to determine. In contrast, ideological motivation is not difficult to determine it if the perpetrator left behind a manifesto, a letter, or even social media posts explaining his motivation. It may be more difficult to distinguish between, for example, membership in a terrorist organization versus sympathy for it.

But otherwise, ideological motivation is straightforward. It typically includes right-wing ideologies, religious fundamentalism, racism, and other hate-based factors. Lastly, suicide can be a motivating factor because some people choose to kill themselves first and then take some others with them.

As has already been said, the results of my comparative inquiries revealed that suicide was prevalent in the CEE sample of cases, more so than in the US sample. CEE mass shooters also shared a greater degree familiarity with the area they attacked than American mass shooters.

Likewise, concerning group grievances, in the US there have been many shootings in which attackers held a grievance against a given societal group or member of people, without necessarily being ideologically motivated. For example, there have been way more workplace shootings in which a worker got fired or experienced something adverse on the job, then came in the next day and shot everybody.

VW&PK: How does the shooting at the Charles University fit into all this? Based on what we know so far, is it rather a typical or atypical case?

AA: This seems to be a typical case of a school shooting based on previous instances of violence, many of which stem from the 1999 Columbine attacks in Colorado. The only difference is that the shooter also killed other people before attacking the university. This is not necessarily typical to the school shooting phenomenon. Although I have not seen any post-forensic investigations, it will be interesting to observe whether or not there were major stressors experienced by the offender in the months leading up to the attack.

VW&PK: Can fame-seeking present a good account of what he did, including killings before the university attack?

AA: I think it is very difficult to come up with good accounts now when we know almost nothing about his motivations, his strains, his health condition. Unless there is some evidence from his email or other sources, it is open to interpretation.

As I said, it seems to be a classical example of a school mass shooting. Some of the mass shooters that had carried out violence recently in Russia left statements such as how they were inspired by Columbine or other events. On the other hand, it is not so typical for fame seekers to shoot relatives or other people such as the young father and his baby in the forest. We will have to wait for the police or the state prosecutor to provide more information.

Nevertheless, your question makes me think of an interesting research question about whether fame-seeking mass shooters are likely to kill family members or other people before the attack or not. I would think that they would probably not, that they would focus on their one attack only.

VW&PK: Before the shooting, you claimed in the media that there is an increasing potential for mass shootings occurring in the CEE because of the copy-cat mechanism, with some mass shooters in the region inspired by US cases. Can you tell us more about this concept and how it relates to the cultural script of school shooting, another concept used in understanding mass shootings?

AA: Yes, a sub-set of mass shooters in the CEE, particularly young offenders, have carried out what are referred to as ‘copy-cat’ attacks based on previous successful mass shooters in the US; these typically feature school shooters and university campus shooters. They learn and study such cases through the internet, through forums, and similar outlets.

For example, the Kerch Polytechnic college mass shooting, which arose on October 17, 2018, featured a clear-cut case in which the offender, an 18-year-old student at the college, killed 20 people and injured 70. He extensively planned the attack and emulated the 1999 Columbine school massacre through the way he dressed, the types of weapons he brought, and the specific usage of explosive devices. The Columbine attackers made homemade bombs that they placed in their high school’s cafeteria. The bombs fortunately did not get successfully detonated, but offenders still ended up shooting many students. A similar situation occurred in the Kerch case, but here, the bomb brought by the perpetrator did explode.

If we are to theorize the conceptual roots of this phenomenon, it is, in my view, representative of a self-led process of online radicalization. As long as there are internet outlets and information on these acts of violence, there will be a finite number of individuals in a given country that may be attracted to carrying out such an act themselves.

With this being said, there is a lot of vagueness inherent to the copy-cat phenomenon and whether or not this relates to the concept of a cultural script. There currently are no major theoretical breakthroughs here, which is significant, as it appears that these terms are being used with little theoretical underpinning. There probably are concepts in literature on terrorism and conflict studies, such as diffusion and emulation, which can inform theory on the mass shooting copy-cat phenomenon.

As in the case of mental health, it is also true here that only a very small sub-set of the youth population or demographic is prone to this form of planning and violence. But nevertheless, it only takes one person to carry out a horrific act of murder that can impact an entire country and society.

VW&PK: How can we adjust gun legislation to decrease the amount of mass shootings and/or fatalities, if at all? And how do you view research into averted attacks and the prevention of mass shootings in general?

AA: Ammunition capacity regulations, fewer bullets, and less powerful guns would be a potential route to take. However, in states (US states) that have stringent regulations of this sort, such as California or New York, there still are mass shootings with very high fatalities. There is no easy answer or solution to this question because in contexts that experience fewer total cases than others, there still are very lethal incidents.

In the US context, there have been several inquiries on averted or failed attacks, but the issue here is that such inquiries are based on cases that were already in the making or already in the plan of getting carried out. Thus, the aspect of prevention here is only applicable to cases that were already starting.

Ultimately, mass shootings are incredibly difficult to prevent because they are also comparatively very rare forms of homicide (versus those that are brought about by domestic violence or by gang or drug-based motivations).

VW&PK: What direction do you think future research will take?

AA: There are a lot of different trajectories for future inquiry on mass shootings. As mentioned earlier, sub-regional variance in the CEE is important to address. Also mentioned was the question pertaining to cases that featured offenders which killed some of their own family members and then proceeded to carry out an attack on members of the public. Comparative social inquiry remains to be done in terms of cases that were foiled versus those that were not.

There are current attempts at launching global analyses of mass shootings. And as more data are gathered, the general nature of this phenomenon will be better understood. Whether or not this will lead to effective prevention policies remains to be seen.

Lastly, I think studying mass shootings alongside terrorism can be a fruitful approach to take. There is much to be uncovered about severe forms of violence aimed at the public, and scholars should provide theoretical development in the future.

This interview was conducted by Václav Walach and Petr Kupka from the Department of Social Work, Faculty of Social Studies, University of Ostrava.