National victims of crime survey
Overview of key findings
Occasional Paper 175 , December 2008
In October and November 2007, the Institute for Security Studies (ISS) conducted a national crime and victimisation survey (NCVS) to gather information on South Africans` perceptions and experiences of crime and the criminal justice system. The survey followed on from two earlier victim surveys, one commissioned by the Department of Safety and Security and the United Nations Interregional Crime and Justice Research Institute (UNICRI) and carried out by Statistics South Africa in March 1998 (Statistics South Africa, 1998), and an ISS study conducted in September 2003 (Burton et al, 2003). As with the 2003 study before it, the 2007 study had two central goals: to provide an accurate picture of crime levels in the country to complement that provided by the official crime statistics published annually by the South African Police Service (SAPS), and to contribute to the establishment of a longitudinal dataset that will make it possible to track trends over time.
Similar victim surveys have been developed and used in many countries over the last four decades to complement police statistics in formulating a holistic picture of crime. They provide an important addition to police statistics, which while essential for tracking crime trends, do not provide an entirely accurate picture of crime, as victims often do not report crime to the authorities.
Victim surveys have several advantages. By asking victims directly about their experience of crime, victim surveys avoid many of the problems relating to non-reporting that affect police data. The surveys also substantially improve the ability to understand the impact of crime on society, by providing more accurate estimates of the volume of crime and how it changes over time, as well as the nature of crime. Victim surveys also provide an opportunity to collect data on people`s perceptions and experiences of crime, as well as their views on police and court performance, and the treatment of victims. Finally, they allow for the collection of information on what communities are doing in response to crime, which is important for developing crime prevention and community-police partnerships (Burton et al, 2003).
Together, the 1998, 2003 and 2007 studies provide almost a decade`s worth of data on levels of victimisation; public perceptions about crime; the fear of crime; attitudes towards the police and courts; and attitudes towards non-state forms of policing and protection. This paper – the first in a series of papers on the survey results –Â provides an overview of the central findings of the 2007 survey. It examines the key victimisation and reporting trends since 1998, as well as changes in the public`s perception of crime, responses to crime and the performance of the criminal justice system. More in-depth analyses of these issues will be published in the forthcoming series of ISS papers. The survey results were analysed by race, province, gender, Living Standards Measure (LSM) groupings1, urban-rural location and community type. The paper presents only those findings that appear influenced by these variables.
The survey methodology
In order to ensure that the results of the 1998, 2003 and 2007 surveys could be compared, the research adopted the same methodology as used in the previous surveys, and used a consistent questionnaire. A nationally representative sample of 4 500 people was interviewed. In keeping with the previous surveys, the data was weighted to ensure representativity of all South Africans. The research was conducted in October and November 2007, and was carried out on behalf of the ISS by the private research company, Ipsos-Markinor.
The study used a probability-based, multi-stage cluster sampling approach, which involved:
Selecting a representative sample of Enumerator Areas (EAs) from Statistics South Africa`s 2007 updated 2001 Census records
Randomly selecting households within each EA
Randomly selecting individuals within each household
Respondents from a total of nine systematically selected households were interviewed from each of the 500 EAs chosen nationally. Each respondent had to be 16 years or older and was chosen using a Kish grid. The substitution of households where interviewers are unable to complete an interview negatively affects the probabilistic nature of the sample design. This is because the probability that a randomly selected respondent will refuse to respond cannot be calculated. Every effort was made to avoid such substitution. However, where household members could not be contacted after at least two return visits (on separate days and different times), or refused to be surveyed, other households were selected using statistically sound criteria.
The data was weighted to correct for any skews in gender, age, province and race in the realised sample. The weighting was done using the 2007 population estimates from the SAtoZ dataset, which generated demographic estimates for 2007 using benchmarking techniques and the mid-year estimates for 2007 published by Statistics South Africa. The SAtoZ dataset contains demographic estimatesÂ at various spatial levels including enumeration areas,Â municipalities and provinces, and was used because it provides a more up to date framework than the data currently available from Statistics South Africa. Weighting the data ensures that it is representative of all South Africans. In this paper the terms ‘respondents` and ‘South Africans` are used interchangeably.
As shown in Figures 1, 2 and 3, the weighted data differs slightly from the 2007 population estimates available data from Statistics South Africa with respect to the racial, gender and provincial breakdown of the sample. Given the use of the more current SAtoZ data to weight the data, it is believed that these distributions more closely reflect the demographic and geographic distribution of the South African population, and that the resulting estimations are more representative of the situation in South Africa.
A little less than one out of every five (22 per cent) of respondents had experienced a crime in the 12 months preceding the survey (Table 1). This is down from the 24,5 per cent who had experienced a crime in the year leading up to the 1998 survey, and is fractionally lower than the 23 per cent who had been victimised in 2003. This supports the argument by Burton et al in 2003 that crime levels are, if not declining to more acceptable levels, at least stabilising.
The victimisation profile remains largely consistent with the 1998 and 2003 surveys. As in these surveys, respondents were most likely to have been burgled, had personal property stolen or to have been robbed. While the rate of most crime types has decreased since 1998, car theft increased fractionally. Motorbike theft also increased slightly, although very few respondents reported having a motorbike stolen, suggesting the need for caution in interpreting these results. Stock left (3,1 per cent), assault (2,9 per cent) and fraud (2,8 percent) show the largest declines between 1998 and 2007. The greatest decreases between the 2003 and 2007 surveys occurred for corruption (2,7 percent) and assault (0,8 percent). On the other hand, car theft, murder and robbery increased slightly over this period.
With respect to corruption in the public sector, respondents were most likely to have been asked to pay bribes to the police, particularly to avoid traffic fines (Table 2). One out of every three South Africans was asked to pay a bribe to avoid a traffic fine. One in six had been asked to pay a bribe to obtain identity documents or a passport, while slightly fewer had been asked to pay a bribe to obtain a driver`s licence or a government job. In most cases bribes in the form of money were requested (2,7 per cent compared to 0.4 per cent each for a present or a favour). The number of people asked to pay a bribe increased most since 2003 in relation to traffic fines and prisons and obtaining a driver`s licence, and has decreased most in relation to obtaining a job, identity document or passport and water or electrical services.
In most cases, those who were asked to pay a bribe did so. Payment levels were highest in relation to traffic fines and customs procedures, where three out of four (76 per cent) and a little more than two out of every three (69 per cent) respondents respectively reported paying the bribe. They were least likely to pay bribes requested in relation to court services (17 per cent), schooling (31 per cent) or when visiting someone in prison (28 per cent). These findings suggest that South Africans are permissive when it comes to what may be perceived as ‘minor` transgressions.
This was borne out by research conducted in 2007 by Ipsos-Markinor. The bi-annual Government Performance Barometer Survey, conducted amongst a nationally representative sample of 3 500 South Africans, asked respondents to rank certain actions on a scale of one to ten according to whether they thought they were always justified (10), never justified (1) or something in between. As shown in Table 3, respondents were surprisingly accepting of tax evasion, avoiding paying for transport or illegitimately accepting government grants. They were also quite open to people accepting bribes and paying a traffic officer to quash a speeding fine.
In 2007, reporting rates for the six most common crimes – assault, housebreaking, theft out of a vehicle, robbery, property theft and corruption - were highest in the case of assault (76 per cent) and housebreaking (61 per cent), and lowest for corruption (6 per cent). With two exceptions, reporting rates for all crimes increased from 1998 (Figure 4). This is particularly so in the case of assault, where reporting rates almost doubled, from 38 per cent in 1998 to 76 per cent in 2007.
The reporting rate for theft of property out of a vehicle, however, declined from 59 per cent in 1998 to 51 per cent in 1998. Data on the reporting of corruption was not collected in 1998, but levels of reporting dropped from 11 per cent in 2003 to 6 per cent in 2007. The figures for corruption may, however, over-estimate levels of reporting compared to other crimes, as they also capture reporting to authorities other than the police, as well as calling 10111 without actually reporting a crime per se.
Public perceptions about changing crime levels
Despite the overall decline in the crime rate, almost three out of every five (57 per cent) of those interviewed in 2007 felt that crime levels increased in the four year interval between the 2003 and 2007 surveys. The 1998 survey did not gather data on respondents` perceptions of crime. The 2007 findings, however, show an increase from 2003, when just over half (53 per cent) of respondents felt that crime had risen over the three year interval between the 1998 and 2003 survey. One in four (25 per cent) felt that it remained the same (Figure 5).
Indians, and to a lesser extent whites, were significantly more negative about crime than Africans and coloureds. Well over four fifths (85 per cent) of Indians reported that crime has increased, compared to three fifths (64 per cent) of whites, just under three fifths (58 per cent) of coloureds and a little over half of Africans. Less than 2 per cent of Indians felt that crime levels have decreased. African respondents were the most positive about crime; one in five (22 per cent) believe that crime has decreased, while coloured and white respondents are more likely to feel that crime levels have stayed the same (26 per cent and 23 per cent respectively).
Such differences between perceptions of crime and reported victimisation levels are not unusual. Public perceptions of crime are not only influenced by first hand experiences of crime as victims or when friends and family fall victim to crime, but also by media reports, other documentary information about the crime situation, and word of mouth.
These differences between experience and perception are instructive, as they point to the public`s anxieties about the extent of crime, their feelings on the adequacy of the government`s response, and whether enough is being done to address their concerns. Widespread pessimism regarding the government`s handling of crime is strongly suggested by the finding that South Africans continue to perceive crime levels to be worse than is suggested by the victimisation data; and that more respondents said they believed that the crime situation is worsening. This is supported by data collected by the Institute for Democracy in South Africa`s (Idasa) Afrobarometer project, which shows that while South Africans were more positive about the government`s handling of crime in 2006 then 2003, they were still most likely to believe that government was handling the issue badly or very badly (Box 1).
Perceptions about safety
Feelings of safety have declined since 1998. While over half (56 per cent) of respondents reported feeling safe walking alone in their neighbourhood after dark in 1998, less than one in four (23 per cent) felt safe in 2003 - although only slightly fewer (21 per cent) report feeling safe in the 2007 survey. The differences between the 1998 and later surveys are less marked where respondents were asked to consider how safe they feel walking alone in their area during the day. Nevertheless, the results still suggest declining levels of perceived safety. While the same proportion (85 per cent) of people reported feeling safe walking around their neighbourhood during the day in the 1998 and 2003 studies, fewer people (75 per cent) report feeling safe in the 2007 survey.
Feelings of safety vary between race groups. Virtually all (97 per cent) Indian respondents reported feeling unsafe walking around their area at night, compared to four fifths of African (80 per cent) and coloured (79Â per cent) respondents and three quarters (73 per cent) of white respondents. Views also varied considerably between provinces. As shown in Figure 7, those living in Gauteng were most likely to report feeling unsafe, followed by those living in Mpumalanga province and North West. Despite the perception that crime has increased in the Western Cape, respondents in the province recorded the highest number of people feeling safe, after the Northern Cape. Unsurprisingly, while both men and women reported feeling generally insecure, women were more likely to unsafe (83 per cent compared to 74 per cent of men).
When asked which crimes they thought occurred most often in their neighbourhood, respondents ranked burglary as the most common, followed by robbery and assault (Table 4). This differs slightly from the 2003 survey, where theft of personal property came second to burglary, followed by robbery.
When respondents were asked what crimes they were most afraid of, burglary again ranked highest, followed by murder and sexual assault. This also differs somewhat from the earlier research, where respondents were most afraid of being murdered, followed by having their houses broken into and being sexually assaulted.
This indicates that while South Africans remain afraid of experiencing interpersonal violent crime, they are becoming increasingly worried about the loss of material possessions. This may be due to a perceived increase in burglars` use of violence, as well as the rising cost of replacing household items, particularly in the absence widespread access to private insurance.
Perceptions about who commits crime and why
There are marked differences between the 2003 and 2007 surveys with respect to respondents` views on who commits crime in their neighbourhood (FigureÂ 8). In 2003, almost two out of every three (64 per cent) South Africans felt that both property and violent crime were committed by people living in their area, with the remainder primarily of the opinion that crime was committed by people from outside their area. This dropped to roughly two in five and one in five respectively in 2007, with slightly fewer than one quarter of respondents reporting that property (23 per cent) and violent (22 per cent) crime were carried out by people both from within and outside their neighbourhood.
Views on perpetrators` birthplace have also shifted. While the majority of South Africans (79 per cent) still believed that property and violent crime are committed by South African nationals, they were more likely than in 2003 to feel that crime was committed by people born outside South Africa. In 2003 less than 5 per cent of respondents believed that crime was committed by foreigners; in the most recent survey, this doubled to one in every ten (10 per cent), with a similar proportion saying they were unsure of where the perpetrators of crime were born (10 per cent and 12 per cent for property and violent crime respectively).
The findings differ according to province. Those living in the melting pot of Gauteng were most likely to feel that crime was committed by people born outside South Africa (18 per cent and 19 per cent for property and violent crime respectively), followed, perhaps unsurprisingly, by those living in the border provinces of the Free State (12 per cent and 10 per cent), Limpopo (10 per cent and 9 per cent), North West (both 9 per cent) and Mpumalanga (9 per cent and 6 per cent) – although KwaZulu-Natal and the Northern Cape come in behind the Western and Eastern Cape respectively (Figure 9). The fact that more people in the border provinces than the other provinces think that non-South Africans perpetrate most crime may simply be indicative of their proximity to Lesotho, Zimbabwe, Botswana and Mozambique. But in light of the recent xenophobic attacks against foreigners, these views may also point to a general growing hostility towards non-South Africans.Â
When asked about what they think motivates criminals in their neighbourhood, respondents most often believed that those committing violent and property crime were motivated by greed (both 39 per cent) and to a lesser extent non-financial reasons (28 per cent and 25 per cent for violent and property crime respectively). This represents a slight hardening of views since 2003, when respondents were more likely to attribute violent crime to non-financial motives (41 per cent) than greed (37 per cent) and almost equally likely to see property crime as the product of greed (34 per cent) and non-financial motives (33 per cent). Respondents were also less likely to believe that violent and property crime was motivated by real need than in 2003 (15 per cent and 23 per cent compared to 17 per cent and 30 per cent for 2007 and 2003 respectively).
The public`s response to crime
The 2003 and 2007 surveys asked respondents what measures they had taken to protect themselves or their households from crime and violence. In 2007, just under half (46 per cent) of South Africans reported having taken measures to protect themselves or their family from crime. This is more than in 2003, when only 39 per cent reported having done something, but still shows that the majority of South Africans have not taken steps to protect themselves.
Preliminary analysis suggests that the likelihood of respondents having taken measures to protect themselves and their family increases with LSM grouping. Analysis by LSM grouping suggests that better-off individuals may be more likely to institute security measures. Those who had been a victim of crime in the four years preceding the survey were also more likely to have taken steps to protect themselves, indicating that victimisation may also influence whether people implement security measures.
The protection measures taken by South Africans most often took the form of measures to increase the security of their home (72 per cent), hiring commercial private security companies (15 per cent), and amongst car owners, increasing the security of their vehicle (15 per cent). The overall increase in the number of people adopting security measures appears primarily due to higher levels of involvement in self-help groups, such as neighbourhood watches and street committees, and greater investment in securing vehicles, rather than home security, which has declined since 2003 (Table 5). Yet, the proportion of South Africans who are involved in self-help groups remains very low.
Community Policing Forums (CPFs) are one mechanism for enabling members of the public to actively engage with the SAPS to prevent crime. CPFs are the interface between public and state responses to crime, and aim to strengthen and better coordinate both sectors` responses. As shown in Figure 10, the public`s knowledge of what a CPF is has improved slightly between the two most recent surveys, rising from 45 per cent in 2003 to 48 per cent in 2007. Participation levels, however, have remained similar, with only one in four South Africans reporting that they have attended a CPF meeting.
Participation in CPFs appears influenced by race. In 2007, approximately one out of every five African (22 per cent), Asian (22 per cent) and coloured (19 per cent) South Africans had attended CFP meetings, compared to slightly fewer than one out of every six (13 per cent) whites. Participation levels are highest in the Northern Cape (39 per cent), followed by the Free State and the Eastern Cape (both 25 per cent). Despite generally feeling safer, those living in rural formal areas are most likely (28 per cent) to participate in their local CPF, while those living in urban informal settlements are more likely than those living in formal areas to attend meetings (24 per cent compared to 19 per cent respectively).
Views on how government should respond to crime
In addition to collecting information on what the public do to prevent crime, the 2003 and 2007 surveys asked respondents what they think government should spend money on in order to reduce crime levels.
The 2007 findings suggest that South Africans are becoming less confident that social development can address crime. In 2003, three out every five (62 per cent) South Africans felt that government should concentrate on social development to address property crime, and half (49 per cent) felt this should be the focus in order to address violent crime (Figure 11). In 2007, only half (52 per cent) of the respondents felt that social development should be prioritised to address property crime, while only two fifths (37 per cent) felt that this would be most effective in addressing violent crime.
These results show that South Africans are becoming increasingly more punitive in their feelings towards criminals, particularly when it comes to violent crime. Respondents in the most recent survey were considerably more in favour of increased government spending on punitive measures, particularly improvements to the judicial system and prisons, and to a much lesser extent, policing. While only one in six (15 per cent) South Africans felt that property crime could be addressed through greater spending on punishment in 2003, one in four (26 per cent) favoured this option in 2007. The number of people advocating such spending to address violent crime increased from 21 per cent in 2003 to 35Â per cent in 2007.
There are differences between race groups, although the findings suggest a hardening of attitudes amongst all race groups when it comes to violent crime. African (58Â per cent) and coloured (52 per cent) respondents were more likely to advocate social development to address property crime than whites and Indians (23,5 and 16,5Â per cent respectively), who were most likely to advocate policing and punishment respectively (Figure 12). The differences between race groups are much smaller when it comes to violent crime. While African (40 per cent) and coloured (40,4 per cent) respondents were still most likely to advocate social development to address violent crime, they were only marginally less likely to advocate punitive actions, with policing further behind (Figure 13). Whites, however, were more likely than other race groups to advocate spending on courts and prisons to address violent crime (52 per cent, compared to 40 per cent, 37,3 per cent and 32,7 per cent of Indians, Coloureds and Africans, respectively).
The public`s perception of the criminal justice system
Comparison of the 2003 and 2007 data on public perceptions of police performance shows that South Africans are slightly less satisfied with the quality of policing now than they were in 2003. The results also suggest growing ambivalence regarding police performance. While the number of people reporting that the police are doing a good job declined from 52 per cent in 2003 to 49 per cent in 2007, there are also less people who believe that the police are performing poorly; the bulk of the change appears due to a substantial increase in the share of people who are unsure about well the police are performing (Figure 14).
Virtually all of those participating in the 2007 survey (96 per cent) had interacted with the police, either at a police station or while they were on duty outside of the police station. In most cases, this contact improved people`s perception of the police, with slightly more people reporting an improvement in their opinion of the police in 2007 (58 per cent) than in 2003 (58 per cent) (See ISS Paper 176 by Michael O`Donovan for a more in-depth look at public perception of the police and courts). However, one in five (21 per cent) reported that their dealings with the police had worsened their opinion of them. This is substantially lower than in 2003, when a little over one third (35 per cent) became more negative. Nevertheless, the 2007 results suggest that more needs to be done to improve the quality and public face of the service provided by frontline police personnel.Â
In contrast to South African`s perceptions of crime almost all race groups were similarly positive about the police`s performance. Half (50 per cent) of African respondents reported that the police were doing a good job, as did only slightly fewer coloured (48 per cent) and white (46 per cent) South Africans. Indians by contrast were the most negative; almost three out of every five Indians felt that the police are performing poorly, compared to 41 per cent of coloureds and 38 per cent of African and white respondents.
The most common reasons given for the police performing well was that they:
Arrest criminals (37 per cent)
Always respond on time (36 per cent)
Are committed to fighting crime (32 per cent)
Are trustworthy (19 per cent)
The primary reason for feeling that the police were performing poorly was that they:
Do not respond timeously (56 per cent)
Are lazy (26 per cent)
Do not come into the respondent`s neighbourhood (23 per cent)
Are corrupt (22 per cent).
South Africans` perception of the judiciary has improved slightly since 2003, with 75,5 per cent of respondents reporting that the courts were performing adequately in 2007, compared to 71,4 per cent in 2003 (FigureÂ 15). Coloureds were most positive about the courts` performance, followed by Africans, Asians and whites (55 per cent, 50 per cent, 44 per cent and 37 per cent respectively). Those who had engaged with the judicial system – in this case by attending court as an observer, claimant or witness – were more positive about the courts; performance than those who had not had any contact with the justice system. Two out of very three (66 per cent) of those who had been to court felt that the court system was performing satisfactorily, compared to one in four (26 per cent) of those who had not.
The most common reason given for being satisfied with the courts` performance was that they pass appropriate sentences (68 per cent). Other frequently mentioned reasons included that they have a high conviction rate (40 per cent) and are not corrupt (19 per cent). Those unhappy with the judiciary most often felt that the sentences meted out were too lenient (54 per cent) and that they release criminals unconditionally (42 per cent), and further behind, that court processes often drag on for a long time (26 per cent).Â Â Â
The 2007 victimisation survey provides an invaluable addition to the data collected in the 1998 and 2003 surveys. Together, the three surveys provide an unprecedented snapshot of both changing victimisation levels and perceptions of crime and the criminal justice system over the last decade. Key trends suggested by the data include:
A stabilisation in the rates of most types of crime
Reporting rates have generally increased, in some cases exponentially, perhaps suggesting greater confidence in the police
Despite the apparent stabilisation of victimisation levels, the majority of South Africans feel that crime has increased since 1998. Those living in Gauteng and the Free State are most likely to feel that crime has increased
Feelings of safety are lower than now than in 1998, and are influenced by factors such as race, where in South Africa one lives, and gender
South Africans are most worried about burglary, murder and assault, with burglary and assault believed to be amongst the most common types of crimes
While most South Africans believe that crime is primarily committed by people living in their area, there is a growing perception that crime is also committed by outsiders
While South Africans are held responsible for most crime, it is also increasingly attributed to people born outside South Africa, particularly among people living in Gauteng and provinces neighbouring other countries
Despite an upward trend in the number of people taking measures to protect themselves against crime, the majority of South Africans have not taken steps to protect themselves
South Africans increasingly feel that government should prioritise spending on prisons and courts in order to address crime
They are also more positive about the performance of the police and judiciary, although poor response times by the police and the perception that courts are too lenient on criminals remain a challenge for the authorities
In conclusion, the findings highlight the value of victim surveys conducted over a period of time. While public perception reflects the growing unease and insecurity that South Africans feel in relation to their safety, these perceptions are contextualised by alternative and reliable data that suggest that most crime types have in fact not increased as perceived. This suggests the need for a concerted, visible and high level commitment to tackling the issue of crime that will start to address the apparent lack of confidence that fuels these public perceptions.
The LSM divides the population into 10 LSM groups, 10 (highest) to 1 (lowest).Â The LSM measure assigns a score on the basis of several variables, including possession of a range of durable household articles; access to reticulated water and electricity supplies, use of the various media, and a wide range of demographics including population group, income, education and others. For more information on the LSM measure, see The South African Advertising Research Foundation (SAARF), ‘The SAARF AMPS Living Standards Measure (LSM)`, available at www.saarf.co.za/AMPS/technicalreport2006/datafiles/Technical/Tech per cent202006 per cent20~ per cent20Pages per cent2098-104 per cent20Com.pdf, [accessed 1 April 2008].
For more information on the Afrobarometer survey, as well as copies of the questionnaires and data, see www.afrobarometer.org [last accessed April 2008].
This graph was compiled from the 2003 (round 2) and 2006 (round 4) datasets available on the Afrobarometer website (www.afrobarometer.org/data.html)
Burton, P, Du Plessis, A, Louw, A, Mistry, D and Van Vuuren, H 2003. National Victims of Crime Survey: South Africa 2003, ISS Monograph Series No. 101. Pretoria: Institute for Security Studies.
Household Surveys Unit 1998. Victims of Crime Survey. Pretoria: Statistics South Africa.
Ipsos-Markinor 2007. The Crime Statistics: Are we our own worst enemy? Available at http://ipsos-markinor.co.za/news/the-crime-statistics-are-we-our-own-worst-enemy, [accessed July 2008].
ABOUT THIS PAPER
In 2007 the Institute for Security Studies conducted a national crime and victimisation survey. Th e survey followed on from two earlier victim surveys conducted by the ISS and Statistics South Africa in 1998 and 2003 respectively. Together, these three surveys provide an unprecedented snapshot of both changing victimisation levels and perceptions of crime and the criminal justice system over the last decade. This paper, the first of several on the research results, provides an overview of the central findings of the 2007 survey, key victimisation and reporting trends since 1998, as well as changes in the public`s perception of crime, responses to crime and the performance of the criminal justice system.
ABOUT THE AUTHOR
Robyn Pharoah is an independent researcher working in Cape Town, in South Africa. She worked previously for the ISS Crime and Justice Programme in Pretoria and oversaw the running of the 2007 national victimisation survey.