The most violent people rarely use violence, and under the most violent circumstances, violence rarely takes place. How come? My work focuses on understanding and explaining how people prevent violence and crime in everyday face-to-face encounters, and under what circumstances de-escalation is successful. I use video analysis and ethnographic fieldwork for my research, and work on human-aligned-video AI methodologies to analyze complex human behavior in detail in natural settings.

How ordinary interactions in public space turn violent is a common thread running through my work. I was robbed at knife point in 2001 when I lived in Cape Town as a student, which inspired me to study how behaviour during such incidences—my refusal to hand over my phone and my lack of screaming—played a role in the fact that I was not stabbed. As an anthropologist by training, I did an ethnography for my PhD dissertation focusing on how behaviour during conflicts in the life of people who are involved in gangs played a role in whether they would use violence. During 14 months of fieldwork in Cape Town, I managed to observe four conflicts that involved violence, and a lot that did not. I published a monograph on my results, in which I explained violence as shaped by interactional dynamics rather than personal characteristics of the people engaging in it. Rather than attributing violence to certain people, cultures, or contexts, in my monography on the topic – Surviving Gangs, Violence and Racism in Cape Town- Ghetto Chameleons - I proposed to understand violence as related to interactions between people in time and space (Scholar google profile). What I still found lacking, though, was the empirical basis for my interactional model. 

As a postdoc, I used my theoretical understanding of violence as driven by interactional dynamics to develop a case-control interview method to study the same persons in different situations to determine the extent to which interactional situations influence violent outcomes. I used this method to interview people convicted for robberies in the Netherlands and found that they were more likely to inflict physical harm on victims if they resisted the robbery. However, this finding contrasted with results from victim-based studies, which suggested that violence in robberies was unrelated to the victims’ behaviour and instead dependent on personal characteristics of the perpetrators. When I presented my results on the Dutch national news, some argued that I had listened too much to people who might have an interest in blaming violence on the victims. This made me realize that I needed more objective measurements of interactions in robberies.

Access to CCTV footage of armed robberies enabled me to scale up my studies and to objectively measure in what way victim behaviour influenced violence during robberies. My video-based analysis showed that victim resistance occurred prior to the violence, offering sequential evidence that victim resistance is followed by victimization. More importantly, thanks to the ability to watch how robberies were unfolding in real time, I discovered that bystanders played a key role in the interactions. Going back to my field notes from Cape Town and the robbery interviews, I realized that my participants had mentioned bystanders abundantly. I had just not paid attention to this third-party role, because I thought they would be irrelevant, as I—like other people influenced by the public myth about bystander apathy—was taught about their passivity.

The CCTV footage, I realized that bystanders were key to understanding what was going on in violent incidences. My first study focused on consolation behaviour in the aftermath of robberies, which was a type of behaviour I—despite being a robbery scholar—was entirely unfamiliar with before watching the footage. Unexpectedly, this study brought me in contract with the field of primatology, with its rich observational methods and extensive theory-building on conflict management, which became a major inspiration for the video-observational methodology I developed. Funded by the Danish and Dutch research councils ([DFF – 6109-00210; VI.Vidi.195.083), I studied bystander actions in violent situations with two postdocs and three PhDs. I also recruited a postdoc primatologist through a MSCA-IF grant for her to come and work with my group. We published our work in American Psychologist and Social Psychological and Personality Science, as well as in other leading journals in the field. Our work gained massive media attention worldwide, including by BBC World News, The Washington Post, The Wall Street Journal, Le Monde, Daily Mail, and Volkskrant, and has impacted national violence prevention strategies in the Netherlands.

By the Covid-19 pandemic, I extended my attention from violence to emergencies and rule-compliance in public space more broadly. In the first week of the pandemic, I set up a collaboration with the City of Amsterdam and the Dutch police to monitor public space behaviour with the use of CCTV footage, and became the leader of a 20-person Covid-19 research team. With funding from the Dutch medical research council (ZonMw: 1043002201001), this collaboration resulted in numerous studies, including a natural experimental study of the impact of face mask wearing on social distancing in public settings. This study showed that the laboratory-generated hypothesis about ‘risk compensation’ did not generalize to real life behaviour—i.e., face masks do not create a false sense of security that makes them non-comply with social distancing measures. Our work was mentioned twice by the Dutch prime minister in his weekly TV briefings to the public, it was directly used by the government for their decision to install face masks as a mitigating measure in semi-public places, and was published in Nature’s Scientific Reports. Further, together with computer vision scientists, I developed an algorithm to automatically and reliably measure social distancing behaviour in public settings. This collaboration resulted in a publication on the use of computer vision in video analysis in Sociological Methods and Research. It also served as a stepping stone in, as a co-PI, securing a lighthouse grant for establishing a human-aligned video AI laboratory with colleagues at the University of Amsterdam, involving seven different faculties and employing ten PhDs.

I am a research group leader of a group of 20 scholars working in the group "Crime in Context" at the NWO-I criminology institute NSCR, and a management team member since 2022. I also formed the NSCR Video Group together with Lasse Liebst, Peter-Ejbye-Ernst and Virginia Pallante.

I have served as an evaluation committee memember of the Veni board of the NWO, the Swedish research council, and the South African research council. I also took part in numerous PhD and supervision committees of the Dutch police and the WODC.  

Fundamental projects:

2024-2025: Research Priority Area Real Emotion project, University of Amsterdam, co-PI. The project integrates video analysis and a scenario-based approach to understand the situational characteristics eliciting bystanders’ attention when observing conflicts.

2023-2028: Lighthouse grant, Data Science Centre, University of Amsterdam, co-PI. Funding an interdisciplinary Human-Aligned-Video-AI Laboratory to study complex behaviour in videos.

2021-2025: Host of a Marie Skłodowska-Curie Individual fellowships (H2020-MSCA-IF-EF-CAR 2020). Funded primatologist postdoc Virginia Pallante to explore and explain differences and similarities between human and primate bystander interventions.

2020-2024 : ZonMw Covid-19 program, Dutch Research Council; Health, Healthcare & Wellbeing. Consortium grant (50-56300-98-603). PI. Towards Evidence-based Social Distancing Policy: Tracking Behavioral Responses to Pandemic Mitigation Measures and Implementation Strategies. Video analysis of rule compliance behaviour with Covid-19 measures in public space during the pandemic.

2019-2024: Vidi award, Dutch Research Council (NWO VI.Vidi.195.083), PI. Conflicts, Violence, and Bystanders in Action. Video analysis of bystander behavior in violent conflicts in Amsterdam.

2018-2023: Danish labor environment grant. (20185100161), PI. Reduction of threats and violence against ticket inspectors and police officers. Video analysis of interactions between citizens and ticket inspectors and police officers on the street.  

2016-2021: Research Leader 2 award, Danish Independent Research Council (FSE. 6109-00210B), PI. Violence, Bystanders, and…Action! Video analysis of intervention in street fights.

2016: Aspasia award, Dutch Research Council (015.012.043). Career talent award for women.

2005: WOTRO, Dutch Research Council (W-52-1085). PhD grant for ethnographic study of conflict and violence among young men in Cape Town.

Applied projects:

2024: Dutch Ministry of Justice, PI. Study of employee-passenger aggression on Dutch trains.

2023: Dutch Ministry of Justice, PI. Study of employee-customer aggression in Dutch shops.

2023: Dutch Retail Industry, PI. Study of employee-customer aggression in Dutch shops.

2022: City of Amsterdam, PI. Video-based evaluation of drug dealing intervention in Amsterdam.

2022: Police and Science, PI. Video-based study of bystander and police collaborations.

2021: City of Amsterdam, PI. Video-based study of bystander interventions in street riots.

2021: City of Amsterdam, PI. Video-based evaluation of the Covid-19 curfew.

2021: Safety Regions of Amsterdam and Rotterdam, PI. Video-based evaluation of the Dutch   face mask implementation in public space.

2020: Dutch Public Health Institute (RIVM), PI. Video-based study of face mask and proximity compliance on the street.

2018: Police and Science, PI. Video-based study of bystanders interventions in street riots.

2018: Danish Crime Prevention Council, PI. Dissemination of work aggression project.

2017: Public transport company (Movia), PI. Video-based study of aggression on between ticket inspectors and passengers.

2013: Police and Science, PI. Video based study of violence during shop robberies.

2010: Police and Science, co-PI. Interview based study of situational dynamics of robberies.

Peter Ejbye-Ernst (2022). See dissertation here: https://dare.uva.nl/search?identifier=ee93ff49-0e0d-435e-89b6-3bce46009f90

Camilla Bank Friis (2022). See dissertation here: https://static-uris.ku.dk/portal/files/376291818/Camilla_Bank_Friis_PhD_Dissertation_2022.pdf

Floris Mosselman (ongoing)

Carlijn van Baak (ongoing)

Hans Myhre Sunde (ongoing)

Marly van Bruchem (ongoing)

Lenneke van Lith (ongoing)

Mara van Daalen (ongoing)

Laura Pighini (ongoing)

Vuyolwethu Ncube (ongoing)

Rutha Rooparaghunath (ongoing)

Lasse Liebst: https://www.sociology.ku.dk/staff/professor-and-associate-professor/?pure=en/persons/265894

Peter Ejbye-Ernst: https://nscr.nl/medewerker/peter-ejbye-ernst-msc/

Virginia Pallante: https://nscr.nl/medewerker/virginia-pallante/

Cees Snoek: https://www.uva.nl/profiel/s/n/c.g.m.snoek/c.g.m.snoek.html

Don Weenink: https://www.uva.nl/en/profile/w/e/d.weenink/d.weenink.html

Camilla Bank Friis: https://www.sociology.ku.dk/staff/assistant-professor-and-postdoc/?pure=en/persons/436817

Brett Bowman: https://www.wits.ac.za/people/academic-a-z-listing/b/brettbowmanwitsacza/

Lisa van Reemst: https://research.vu.nl/en/persons/lisa-van-reemst

Vana Hutter: https://nscr.nl/medewerker/vana-hutter/

Karin Proost: https://www.kuleuven.be/wieiswie/nl/person/00003503