Victim surveys are generally utilised to gather information that complements police crime statistics. Their findings may be of value for cities in the planning of crime prevention strategies.

The Dar es Salaam victim survey was undertaken to gain an understanding of the levels of crime and violence in different communities within the City Council’s jurisdiction. This entailed:

• the collection, collation and synthesis of the available information on the types of communities and residential areas located in the city area;
  
  
• the development of broad profiles of the identified ward areas in Dar es Salaam; and
  
  
• the development of the necessary sampling formulae and the refinement and adaptation of the questionnaire to suit the Dar es Salaam environment and reality.

The survey included:

• a street scan of 2 400 interviews
• a household survey of 1 000 interviews; and
• 42 in-depth interviews with abused women.

The street scan was undertaken to identify the ratio of victims to non-victims of crime by residential area in randomly selected ward areas. A household victim survey followed. The survey among a limited sample of female survivors of violence was administered to cover issues relating to violence against women in more depth, as these would not necessarily emerge in the responses to the general victims of crime survey.

Stage 1: Refinement and adaptation of the questionnaire and the selection of areas

A reference group was formed consisting of a:

• field manager;
• research advisor;
• gender representative from the Tanzania Media Women Association (TAMWA);
• representative from the Police Crime Prevention Unit;
• Dar es Salaam City solicitor;
• co-ordinator of Safer Cities Dar es Salaam; and
• deputy co-ordinator of Safer Cities Dar es Salaam.

The questionnaires, refined through the experiences and results of the victim surveys conducted in South Africa by the Institute for Security Studies and DRADevelopment, were adapted to suit the Dar es Salaam environment and reality.

The Institute for Security Studies has conducted four city victim surveys, one survey of crime in rural areas, two victim surveys in small towns and one national survey of violence against women in South Africa. These surveys were conducted with the aim of providing city, regional and national government officials with data to utilise in the design and improvement of crime reduction strategies. The questionnaires for the victim surveys were developed on the basis of those used by the International Crime Victims Survey (ICVS) conducted by United Nations Interregional Crime and Justice Research Institute (UNICRI).

In the Dar es Salaam survey, it was decided that the research would represent the different area types within the city. Since there are very few areas in Dar es Salaam that do not have a residential population, all geographic localities were included in the sample framework. The reference group divided Dar es Salaam’s 73 wards according to whether they were in the city centre, established suburbs, emerging or new suburbs or predominantly rural settlement areas. A quota was assigned to each and a sample of 20 was randomly chosen. Interviews were only held with respondents over 18 years of age.

Stage 2: Training and piloting

Fieldworkers were selected from unemployed youth who had applied for admission to university courses and were awaiting a response. A two-day inhouse training course was held with fieldworkers, followed by one day of field training. Debriefing was held after each pilot. Although the questionnaires were in English, all interviews were conducted in Kiswahili. Considerable time was spent during the inhouse training to ensure uniformity across all team members. These sessions were led by the Safer Cities Dar es Salaam project team.

The pilot study of the street scan and household survey functioned as a training tool to refine interview techniques and to measure the length of the questionnaire. During the household pilot study, 22 questionnaires were administered (one per fieldworker). The length of the survey was approximately 45 minutes, but the quantity of different crimes that the respondent had experienced could lengthen the interview to one and a half hours. Fieldworkers often had difficulty terminating the interviews.

In South Africa, it was shown that the success rate, both in terms of selecting respondents and the time taken to execute the subsample, was much higher among enumerators dressed in an identifiable ‘uniform’. Fieldworkers were therefore dressed in T-shirts bearing the logo of Safer Cities Dar es Salaam.

In the scan survey, 2 400 people were approached in the 20 selected wards in Dar es Salaam. At each sampling node, 120 respondents were approached in two age categories (18-35, 35+), equally divided between female and male respondents. Prior to approaching the respondents, no distinction was made between victims and non-victims. The only screening criterion was age. In each ward, a sample point was chosen:

• markets (fish markets, second-hand clothing markets, fruit and vegetable markets);
• transport nodes (bus terminals and bus stands);
• hospitals, health and welfare centres;
• industrial centres; and
• education centres.

In the more detailed household survey on the experience and perception of crime, 1 000 respondents were approached. In this survey, the same 20 ward areas were selected, with the exception of two areas which were replaced in the household survey by other areas as they were not residential areas (Kivukoni was replaced by Kariakoo and Gerezani was replaced by Kisutu). The ward executive officer supplied each fieldworker with a list of the plot and flat numbers in their respective areas. The fieldworker was then able to choose randomly which households would be approached and respondents interviewed in their homes. In a block of flats, only two households were to be interviewed.

In each ward area, 50 respondents were interviewed in three age categories (table 4).

A respondent selection technique to ensure that the selection process was both random and rigorously implemented was constructed. During the fieldwork stage, visits were made to 117 (11%) of the interviewed respondents to ensure that standards were being adhered to.

Household interviews were conducted during the day and in the evening, including weekends, to ensure that the views of a broad cross-section of people were collected, including those who worked during the day.