You are hereHome ›
Illuminating the Dark Figure of Crime: Victimisation Surveys and Beyond
Before I say anything else, let me first thank the jury for its decision to bestow on me the European Criminology Award for lifelong contribution to criminology in Europe. I have, as mentioned, received other professional honours, including the Sellin-Glueck and the Stockholm Criminology Prize, but, indeed, for a European criminologist, the European Award is special.
In particular, I want to thank Michael Tonry, the chair of this year’s jury, for his eloquently-drafted and personalised jury report, highlighting some of my achievements. And I want, of course, to wholeheartedly thank my compatriot, Catrien Bijleveld, for her kind and flattering words on this happy occasion. Thank you all very much!
It seems a sort of tradition for the winners of this award to use this occasion to look back extensively at their professional careers. I hope you will agree with me that this would be inappropriate in my case. I am, as Michael just hinted, still far too young for such a retrospective. Rather than looking back, therefore, let me share with you some results of ongoing research, and, tell you about my plans for the future.
Estimating the dark figure of crime
A constant feature of my criminological work has been the effort to estimate the true volume of crime. This started with my 1974 supervision of the first full-fledged national victimisation survey in Europe, the Landelijke Slachtofferenquete Misdrijven (National Crime Victimization Survey) of the Netherlands. This survey subsequently evolved into the International Crime Victims Survey (ICVS), which has been carried out in over 80 countries at least once since 1988 (Van Dijk, Van Kesteren & Mayhew, 2014).
Why is it so important, one could ask, for governments to know how many crimes are actually committed against their populations? First, crimes often inflict harm on ordinary people, and that could and ought to be prevented. Many, if not most, of these crimes are never recorded by the police, and victims of these ‘dark numbers’ are unlikely to be offered any kind of services or specialised support. This, one could argue, is the basic victimological perspective on measuring the dark figure of criminal victimisation.
In addition, during my years at UNODC I have become convinced that organised crime, human trafficking, corruption and terrorism negatively impact the prosperity and welfare of national populations in a myriad ways. This was expressed in the subtitle of my book, informed by my seven years at the UN, The World of Crime, published by Sage in 2008 (Van Dijk, 2008). The subtitle was ‘Breaking the silence on the problems of security, justice and development across the world’. The message of the book – that rampant crime hampers sustainable development – has attracted less attention from my fellow criminologists than I had hoped for (Van Dijk, 2010). On a more positive note, it was picked up by experts at the World Bank and others analysing the macro factors impeding human development (Wenmann &. Muggay, 2010). In 2011, the overarching theme of the World Bank’s Development report focused on the relationships between conflict, security and development (World Bank, 2011). The report presented new evidence that, as argued in my book, high levels of serious crime impede economic growth, mainly by consuming scarce tax revenues and deterring foreign and domestic investment. In fact, the report shows that over the past 25 years, while poverty is declining globally, the countries most affected by organised crime and violence have been lagging behind in achieving development goals.
This summer, in August 2016, the United Nations’s Economic and Social Council (ECOSOC) adopted the 2030 Agenda for Sustainable Development, comprised of 17 Sustainable Development Goals (SDG). Goal 16 specifically addresses the need for effective crime prevention and peace-building, indicating the need to:
- promote peaceful and inclusive societies for sustainable development,
- provide access to justice for all and
- build effective, accountable and inclusive institutions at all levels.
The adoption of this SDG is, in my view, a belated but most welcome acknowledgement by the world community of the vital importance of effective and humane crime prevention for human development. Fifteen years ago such acknowledgment would have been almost unimaginable. Crime prevention was, much to the chagrin of senior UNODC staff at the time, not even footnoted in the Millennium Development Goals adopted in 2000.
At the ECOSOC Summit, the United Nations not only adopted the new Sustainable Development Goals but also reached agreement on the set of indicators by which the progress of the SDGs will have to be monitored. At the advice of some of my former colleagues at UNODC, the indicators to measure SDG 16 include several key findings of victimisation surveys, eminently those of the ICVS:
- 16.1.3 Proportion of population subjected to physical, psychological or sexual violence in the previous 12 months
- 16.1.4 Proportion of population that feels safe walking alone around the area they live
- 16.3.1 Proportion of victims of violence in the previous 12 months who reported their victimisation to the police
- 16.5.1 Proportion of persons who had at least one contact with a public official and who paid a bribe to a public official, or were asked for a bribe by those public officials, during the previous 12 months
The adoption of these indicators is, therefore, most likely to increase global interest in subsequent rounds of the ICVS. After an extensive fifth round of the ICVS in 2005, the project has struggled to raise sufficient funding, largely due to the unfortunate decision of the European Parliament to block Eurostat’s well-prepared plans to carry out a adjusted version of the survey in all member states in 2013.
In the meantime the ICVS has since 2010 been duly conducted in 20 different countries from various world regions, namely in:
- Bulgaria ( 2014), Denmark (2010), Estonia (2010), Germany (2010), Luxembourg (2013), The Netherlands (2010), Sweden (2010), Switzerland (2015), and United Kingdom (2010)
- Georgia (2013), Kyrgyzstan (2015) and Kazachstan (2016)
- Canada (2010) and Brazil (2012)
- China (Beijing )(2014)
- Bahamas, Barbados, Jamaica, Suriname, Trinidad and Tobago (2014/2015)
The latest round of surveys conducted in developing nations has confirmed the potential of the ICVS to estimate not only the levels of frequently-occuring property crimes but also of various forms of serious violent crime as well. For example, Figure 1 presents the percentages of respondents victimized by assaults per world region according to the latest available ICVS results.
Figure 1 Percentages of the Public Victimized by Assaults, by World Region; source: ICVS 1995-2015/latest available/University of Lausanne
The results of the Caribbean surveys have also confirmed the survey’s potential to estimate rates of victimisation by sub-categories of violent crime such as weapons-related violence, as shown in Figure 2.
Figure 2 One year prevalence rates of robberies with knives, guns and overall and of assaults/threats with knives, guns and overall; Source Caribbean crime surveys/Interamerican Development Bank report 2016/forthcoming
An innovation in the latest ICVS round is a module on victimisation looking at the crime of bride-kidnapping in a USAID-funded survey in Kyrgyzstan. Through a questionnaire completed by the survey subjects themselves as the final part of face-to-face interviews, married female respondents were asked confidentially whether they had been kidnapped by their husband. Figure 3 presents the results.
Figure 3 Percentages of married women in Kyrgyzstan saying they have been kidnapped by their husbands for marriage; Source: Kyrgyzstan Public Safety Survey 2015
According to the survey’s findings, almost four percent of married women in the country reported having been kidnapped by their husbands without a prior understanding. A separate question posed to all respondents in the same survey showed that a large majority of the population is somewhat or very worried that ‘a daughter / grand-daughter / sister will be kidnapped by someone for marriage (bride stealing)’. These questions have also been included in a new ICVS-based survey in Kazakhstan, funded by the European Commission, to be conducted in 2017.
Estimating victims of human trafficking: alternative methods
One of the elements of Sustainable Development Goal 16 and related Goals, such as the one on Gender Equality, is to eliminate all forms of exploitation, human trafficking and modern slavery. The indicator chosen here is the ‘Number of victims of human trafficking per 100,000 population, by sex, age and form of exploitation’. The explanatory note accompanying the indicator in the report distinguishes between detected and undetected victims, and explains that methodologies for estimating undetected victims are under development.
For some years the Australia-based Walk Free organisation has released updates of its Global Slavery Index (GSI), providing estimates of the numbers of undetected victims per country. In response to reviews of its methodology, Walk Free has recently commissioned a comparative mini-survey on experiences with forced labour. The results of the first round of such surveys in 17 countries, carried out by Gallup International, have been incorporated into the 2016 GSI (Walk Free, 2016). Since the surveys have been confined to developing or middle-income countries, their usefulness to provide estimates of forced labour victims in high-income countries remains unproven. Considering that forced labour is likely to be less common in this latter category of countries and concentrated among special segments of the population , it is doubtful that representative surveys using sample sizes of 1,000 per country can reliably assess its prevalence. Constructing such an estimate requires special sampling methods, including respondent-driven sampling among high-risk populations, such as irregular migrants. Furthermore, the first batch of Walk Free surveys has uncovered very few cases of victimisation by sexual exploitation, suggesting that the measurement of victimisation by sexual exploitation may, like partner violence or bride-stealing, require special interview modes, e.g. self-completion questionnaires. The survey-based approach, then, however promising in many respects, faces methodological challenges of its own, and should not be seen as a panacea for all problems related to estimating the number of undetected victims of trafficking in persons.
An alternative method to estimate the dark figure of rare events like victimisation by modern slavery is Multiple Systems Estimation (MSE). This technique was originally designed to estimate the dark figure of animals in defined territories, such as salmon in a lake, and is also known as the Capture/Recapture Method. It is increasingly also applied in social research, e.g. to estimate casualties in armed conflicts. By comparing recorded cases on three or more different lists from, for example, police, health institutions, and mass media, and then extrapolating from the numbers appearing on more than one list, the dark figure can be estimated.
In the United Kingdom five different organisations maintain lists of presumed/detected victims, including the police, border police and NGO’s running shelter homes. Using the United Kingdom’s National Crime Agency’s (NCA) multi-source recording system on presumed victims of modern slavery, an analysis was made of the double or multiple counts. Of the total of 2,744 detected victims included in the Strategic Assessment, many appeared on only one of the five lists available, some appeared on two different lists and a few on three or four of the lists. Prof. B. Silverman, chief scientist of the Home Office and his team, applied MSE to estimate the figure of potential victims who do not appear on any of the lists, and hence to give an estimate of the total number of such victims (Silverman, 2014). The upshot of the MSE carried out on the NCA lists is that the true numbers of victims of modern slavery during 2013 was estimated to be between 7,000 and 10,000, or three to five times the numbers of detected victims. In other words, the ratio between detected victims and undetected victims was estimated as one to four. Calculations based on comparisons between recorded crime and the estimated rates of victimisation from the National Crime Surveys of England and Wales (formerly the British Crime Survey) indicate a ratio of one to four for total crime as well. Considering the relatively hidden nature of human trafficking, compared to, for example, household burglaries, which are—if only for insurance reasons—commonly reported to the police by victims, a ratio of one to four seems comparatively low.
In the Netherlands, the state-sponsored NGO CoMensha acts as clearinghouse of cases of presumed victimisation by human trafficking. Relevant governmental institutions, such as the 17 different police districts, the Border Police (Koninklijke Marechausee/Kmar), Regional Coordination Offices (decentralised units of CoMensha coordinating victim services) and Labour Inspectorates (Inspecties Sociale Zaken en Werkgelegenheid/ISZW) are instructed to report on all cases of possible victimisation by trafficking in persons coming to their knowledge. Specialised NGOs offering services to trafficking victims report their cases as well.
By applying MSE to the 1,560 cases recorded on one or more of the six available lists in 2014, Van Dijk and Van der Heijden (2016) arrived at an estimate of 17,800 victims. This estimate suggests that roughly ten percent of all victims are detected, or, in other words, that there are ten times more victims present in the Netherlands in the course of a year than those reported to CoMensha. The 95 percent confidence interval ranges from approximately 14,000 to 23,900. It seems worth noting that the numbers per capita and the ratio between detected and undetected victims are considerable larger than the ones estimated in the United Kingdom. One possible explanation is that, unlike its British counterpart, the Dutch database includes persons reported by the Border Agency who have been assisted in crossing the border in order to work in prostitution in either non-exploitative or exploitative conditions. Some of these persons may not qualify as victims of modern slavery in a comparative perspective.
The CoMensha databases include covariates such as age (minor or adult), type of exploitation experienced by the victim (sexual services, forced labour, forced criminality and unknown) and nationality (Dutch resident/foreigner). Using these variables, an ongoing study aims to produce both a more accurate total estimate as well as separate estimates for various subgroups of victims.
A follow-up study commissioned by UNODC and Walk Free will try to carry out MSE on the multi-source databases of selected other countries in Europe. The results of these studies will, together with the results of the next round of Gallup sample surveys on experiences with modern slavery, form the basis of new national, regional and global estimates of the numbers of undetected victims in the coming years.
Over the past four decades, the empirical foundations of criminology have been strengthened by the application of quantitative research methods, perhaps most notably in dark figure studies, such as self-reported delinquency and victimisation surveys. Such surveys are now carried out regularly across the globe and have become a staple of modern criminology. The results of these studies have opened the doors for a renaissance of comparative international criminology (Van Dijk, 2015) , once pioneered in the 19th century by Quetelet, Lacassagne, Von Mayr and Bonger inter alia, when administrative data on crime had first become widely available.
More recently, victimisation surveys have also attracted the attention of international organisations such as UNODC, UNDP and the ILO and of international non-governmental organisations (INGOs) such as Transparency International, Walk Free, the World Justice Program and the World Internal Security and Police Index. Several organisations now regularly commission comparative surveys on people's or business executives’ experiences with crime and the police. More and more survey-based crime data are becoming available. On the downside, few of these studies meet the quality standards developed by criminologists over the past four decades of surveying crime. Definitions of criminal victimisations tend to be unspecific and /or insufficiently fixed in time. For example, a question might ask: have you or any one else in your household ever been the victim of a theft? And the sample designs often leave much to be desired, too.
Obviously, the freshly agreed-upon UN indicator of victimisation by physical, psychological sexual violence poses major methodological challenges in terms of both the formulation of questions and the appropriate interview mode. If the statistical authorities of countries and concerned INGOs are truly committed to measuring progress in implementing SDG 16, they will need to hire criminological expertise. Criminological knowledge, then, is likely to be more in demand in the coming years than before. I, for one, regard that as good news for our profession. We have our work carved out for us.
Jan Van Dijk is emeritus professor victimology, INTERVICT, Tilburg University, the Netherlands
Silverman, B. (2014) Modern Slavery: an application of Multiple Systems Estimation, Home Office. Available at: https://www.gov.uk/government/publications/modern-slavery-an-application-of-multiple-systems-estimation
Dijk, J.J.M. van (2008), The World of Crime; breaking the silence on problems of justice, security, and development across the world, Thousand Oaks: Sage
Dijk, J.J.M. van (2014). ‘Crime.’ In: D. Rowe (ed), Achieving Sustainability: Visions, Principles, and Practices, Detroit: Macmillan Reference USA
Dijk, J.J.M. van (2015). The case for survey-based comparative measures of crime, In: European Journal of Criminology. Vol 12, 4, p. 437-456 19 p
Dijk, J. J. M. van, van Kesteren, J. N. & Mayhew, P. (2014). The international crime victims surveys: A retrospective, In: International review of victimology. 20, 1, p. 49-69 21 p.
Dijk, J.J.M van & P. G.M. van der Heijden (2016), Multiple Systems Estimation for estimating the number of victims of human trafficking across the world, Research Brief, Vienna: UNODC https://www.unodc.org/documents/data-and-analysis/tip/TiPMSE.pdf
Jansson, K. (2006). British Crime Survey – Measuring crime for 25 years, London: Home Office Kyrgyzstan Public Safety Survey 2015 (2015), http://wp.unil.ch/icvs/news/information-and-data-from-kyrgyzstan
Wenmann, A. and Muggah,R. (2010), More Violence, Less Development. Geneva: Geneva Declaration Secretariat. Available at http://www.genevadeclaration.org/fileadmin/docs/MDG_Process/MoreViolenceLessDevelopment.pdf
Walk Free (2016), Global Slavery Index 2016, www.globalslaveryindex.org/
World Bank (2011). World Development Report 2011: Conflict, Security and Development. Washington, DC: World Bank. Available from http://publications.worldbank.org/index.php?main_page=product_info&products_id=23888.