Although commonly perceived as an American phenomenon, mass shootings have arisen in Canada, Norway, Germany, New Zealand, the Czech Republic, and Russia, among numerous other countries. Even though the incident rate of mass shootings is four to ten times lower in Europe or Asia when compared to the United States (Lankford, 2019), dozens of cases have occurred over the past three decades across Central and Eastern Europe (CEE). To date, researchers have only examined a limited sample of cases from these regions, which makes CEE contexts incredibly understudied. On one hand, there is a major gap in knowledge of mass shootings and their associated dynamics in these regions. On the other hand, more cases appear to be occurring across some CEE states in recent years. These regions of the world have endured incredible hardships over the course of the last century and many tumultuous processes of social change including two world wars, genocides, state disintegration, gulags, and industrial disasters.

When it comes to guns, the countries under investigation do not have any constitutional right to bear arms. Additionally, they have lower rates of gun ownership than the US and, on average, tend to experience fewer homicides per capita (Anisin, 2022a). At the time of writing this study, no comprehensive source of data on mass shootings in these regions exists. For example, in a single academic article (Malkki, 2014), one book (Hurka, 2017), and one policy report (Duquet et al., 2016), less than a handful of mass shootings that arose in CEE states have been identified. Silva’s recent (2022) global analysis of developed and developing countries probed several cases of CEE regions, including the countries of Croatia, Lithuania, Slovakia, the Czech Republic, and Russia. This study aims to put forward a reference point for scholars and policy makers who are interested in the cross-national study of mass shootings by presenting a new sample of attempted and completed mass shootings.

I ask whether there are salient differences or similarities between cases that arose in CEE states versus those in the commonly studied context of the United States. Cross-national research on mass murder warrants continuous development, scholarly interaction, and funding. Even though scholarly communities have produced salient policy recommendations about preventing this rare form of homicide, there is still much to be learned about mass shootings. One aspect that seems to be underemphasized in scholarship on mass shootings pertains to the fact that investigating a phenomenon in one single national context may impede our ability to generalize about it or fully understand its underlying nature. Aside from Silva’s (2022) recent inquiry, two studies by Lankford (2016a, 2016b) compared the frequency of mass shootings in the US to other states. Indeed, it is not the case that researchers have only wanted to investigate the United States because of intrinsic biases, but rather most attention has been placed on the American context probably because the incident frequency of mass shootings in the US has been greater than in other countries.

There is much to be learned from comparative inquiry on mass shootings if data from different countries are gathered and made available. This study puts forward the first sample of cases stemming to CEE regions of the world. It identifies 76 different attempted and completed mass shootings that occurred throughout some European Union member states (Bulgaria, Croatia, the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Romania, and Slovakia) and some non-EU states (Albania, Bosnia and Herzegovina, Serbia, Russian Federation, and Ukraine). These cases make up the Mass Shootings in Central and Eastern Europe (MSCEE) data set. After explaining how data were gathered and how cases were observed, this study presents descriptive statistics on case-level and offender-level characteristics. This is followed by a statistical analysis in which 48 of these cases are assessed in relation to a sample of cases (n = 101) drawn from U.S. mass shootings (1999-2020). Along with presenting new data, the study puts forward a codebook and an appendix with case summaries. The study concludes by presenting directions and recommendations for future social inquiry.

The Mass Shootings in Central and Eastern Europe Data Set

Core Concepts in MSCEE

As with any investigation of mass shootings, this phenomenon must be defined according to specific parameters. The interdisciplinary study of mass shootings has commonly differentiated mass shootings from other types of homicides, such as those related to gang or drug violence. Mass shootings are also events that must involve a certain threshold of victims that get attacked, and, importantly, the attack itself cannot be temporally elongated. This study defines a mass shooting according to criteria defined by Silva & Capellan (2019):

A mass public shooting is an incident of targeted violence where an offender has killed or attempted to kill four or more victims on a public stage. (1) the act can involve more than one offender and take place at multiple related locations within a 24-hour time period; (2) the main weapon has to be a firearm; and (3) the shooting is not related to state-sponsored or profit-driven criminal activity (e.g., drug trafficking or gang shootings). (p. 83)

This definition was used when compiling this data set. Along these lines, this study makes a differentiation between “attempted” mass shootings and “completed mass shootings.” The former are cases in which fewer than four fatalities arose (not including the offender), while the latter are cases in which four or more fatalities arose (not including the offender). For multiple reasons, it can be fruitful to investigate cases that feature less fatalities than commonly used thresholds, especially in contexts that have been understudied. Taking a comparative angle of this sort can enable researchers to identify potentially significant factors that can account for the lethality of cases across different countries. Another reason why attempted cases are included is that they constitute empirically relevant incidents of violence; often, the difference between two, three, or four fatalities is conceptually arbitrary. Finally, the inclusion of both attempted and completed cases into the data set will enable researchers to triangulate these data with cases from other contexts such as the United States. Potential inquiries of this sort probe why some cases result in more fatalities than others on a cross-national basis. They can also reveal patterns or specific tendencies that might be inherent to offenders’ attacks in different shooting location settings. In the MSCEE data set, 72% of cases resulted in four or more fatalities.

Relationship to Existing Data

At the time of when this data set was being compiled, scholars had only identified a handful of total cases that arose in CEE states. This meant that there was a systemic underrepresentation of mass shootings from CEE cases in scholarly research. The MSCEE data set was put together on a scholarly basis and hence may differ from the numerous public policy-oriented approaches to data on mass shootings, such as those put forward by Mother Jones, Everytown for Gun Safety, or the Mass Shooting Tracker. While there is nothing wrong or conceptually erroneous with these datasets, they tend to miss out on detailed offender-level characteristics and do not have as strict of a definition on what constitutes a mass shooting. As such, the present data set reflects methodological strategies that are akin to criminological inquiry on mass shootings, such as the Gun Violence Archive or the Stanford University Library’s Mass Shootings in America (MSA).

Another feature of this data set has to do with generalizability. The variables that were incorporated into the data gathering process were specified according to cultural independency. This means that there are no country-specific or culturally-specific characteristics inherent to any given variable in the data set. By cultural independence, I refer to characteristics that can be interpreted in the same manner regardless of national context. Culturally independent characteristics are not contingent on any particular qualitative or contextual condition. Indeed, it is important to consider that most variables can be influenced or affected by culture, however, a certain degree of abstraction has to be adopted in order to conduct comparative social inquiry. The inclusion of culturally independent variables in this data set will enable researchers to accomplish this. For example, if we take a culturally dependent variable such as corruption into consideration, one can observe that this factor manifests itself in qualitatively different forms in different contexts, such as the “Blat” system that developed in the late-USSR and expanded during the 1990s in Russia, or the Chinese Guanxi system (Ledeneva, 2008; Karhunen et al., 2018). Standard measures of corruption do not capture the culturally dependent characteristics that the phenomenon is reliant upon. As such, at this early point in the development of the MSCEE, the overarching aim is to place focus on culturally independent variables so they can be easily triangulated or combined with variables from other datasets.

Lastly, while the data set does feature the names of perpetrators of each of the 76 cases, this study will not directly mention or reveal offender names, which is in agreement with a recent proposition put forward by Lankford & Madfis (2018), who argue that media should refrain from naming perpetrators in order to lessen the propensity of unintended fame-attribution. Such proposals also reflect the “No Notoriety” and “Don’t Name Them” movements that are gaining prominence in public and scholarly discourses.

Creating MSCEE

The data collection process that went into identifying a sample of CEE attempted and completed mass shooting was rigorous and is a part of an ongoing collection effort based at the primary author’s home institution. Along with the help of three graduate research assistants, the primary author utilized numerous language sources, search indexes, and reports to identify cases across more than a dozen countries. In total, the research team that coded data on case observations was fluent in six languages, and where necessary, we consulted external assistants for interpretational help in languages we were not fluent in. The data gathering process took more than one year and is ongoing with the aim of keeping these data updated on an annual basis for the years to come.

In terms of the time period under attention, all cases were analyzed from 1990 through 2021. In CEE states, data predating 1990 is nearly nonexistent because every single country in the sample had a communist system of governance that included state-controlled media; censorship was rampant both internally and for foreign audiences. It is also worth considering that this overall 31-year period contains variance within it. In the 1990s, largely homogeneous societies experienced a significant decline in socio-economic status where some of the largest privatization processes in the history of economics took place; state-owned resources were privatized, and entire institutions were either transformed or collapsed into an abyss of corrupt marketization. Mortality rates rose in nearly all post-communist European countries, and unemployment and inflation sky-rocketed (Stuckler et al., 2009). Citizens across post-communist societies were twice as likely to feel unsafe walking outside at night than in Western Europe (Holmes, 2009). Most contexts saw crime and corruption rise steeply, which led to widespread social strife and informal systems of exchange that were dominated by criminal groups.

Into the 2000s, however, an upsurge in economic prosperity, increases in longevity and lifespan (an increase of nearly 10 years) and decreases in corruption and crime occurred in all Central European states (McNamara, 2021), and in some Eastern post-communist states. Recent events in Ukraine and Russia have disrupted this trend and are likely to lead to adverse outcomes across both the Central and Eastern regions of Europe for the foreseeable future. Yet, even as recently as 2009, only 5 out of the 24 post-communist states had crime rates that were still trending upwards (Holmes, 2009). Interestingly enough, throughout the data gathering process, there was a clear relation between the number of sources identified and the total fatality rate. Higher fatality incidents received more media coverage and public attention, especially in cases that occurred in recent years (compared to those that took place in the 1990s). Such a dynamic has also been observed in the study of U.S. mass shootings in that major national news outlets are likelier to report on mass shootings that are very deadly (Silva & Capellan, 2019).

Identifying Cases

In total, 76 cases were discovered. Open-source methodology was implemented which reflects common approaches used in data gathering on active shooter events, terrorism, and school shootings (Silver et al., 2019). In identifying cases, the primary author as well as three graduate research assistants first over-viewed scholarly articles and books on topics related to mass murder in CEE. To our surprise, only a few articles fit the inclusion criteria, and these studies featured very few cases from CEE. Policy papers were then searched, and only one relevant study that included two cases from the regions under attention were identified. When an incident that fit our criteria was identified in results provided by search engines, the specific name of the offender along with the city and country of the shooting was searched as a follow up to obtain more information. After running these searches, more references on each shooting and greater details about characteristics and factors related to the shooting were obtained such as post-shooting investigations and trials. After searching through scholarly sources, the LexisNexis archive was explored. Here, eight different cases were identified based on Euronews articles that were in English.

Most cases were identified via Google and Yandex (which is the Russian equivalent of Google). The following terms and phrases either on their own or in combination with one another were searched: “public,” “shooting,” “European shooting,” “mass shooting in Europe”; these led to hundreds of different articles and reports about different cases that arose across the European continent. These two search engines provided us with information for nearly all cases. In detail, search terms in English, Russian and Ukrainian were used in both Google and Yandex. Where necessary, country-specific Google search indexes (e.g., Google.pl for Poland or Google.ro for Romania) were relied on. In the latter, several lower profile cases that arose in nations that do not have large populations (such as Romania) were identified and here, the utilization of country specific search engines was needed because international media did not report heavily on these incidents.

Coding Incident Characteristics

The most common type of source that was encountered were media reports. For each of the 76 cases in this data set, at least two sources were used, and each source was cross-checked by two or more different research assistants as well as the principal investigator. In instances where disagreements about coding case characteristics arose, in-depth reviews of each case were carried out in a collaborative manner. The extent of independent assessments and evaluations of characteristics was representative of what research methodologists refer to as equivalent conclusions (Lombard et al., 2002). Importantly, no variables in this data set have missing values. We purposefully collected all information relevant to the 24 different variables because we are cognizant that some methodological approaches, such as Qualitative Comparative Analysis and its variants (csQCA, fsQCA), cannot work and do not function with missing values. Thus, for any variable in which a value of 0 was coded, this means that the characteristic of the variable was not present for the incident to which the variable is associated with. Table 1 lists a summary of the frequencies of cases across each country in the MSCEE data along with the average fatality count.