Eurostat Grant Agreement no 38400.2005.002-2006.052 with the European Institute for Crime Prevention and Control affiliated with the United Nations (HEUNI)
For action entitled ”Crime and Victimisation Surveys”
Final report
31.08.2007
Kauko Aromaa
Markku Heiskanen
Seppo Laaksonen
Minna Viuhko
Contents:
1. Description of the action
2. Background
3. Victim surveys conducted in the Member States
3.1. Introduction
3.2. International crime victimisation survey (ICVS)
3.3. Victimisations surveys during the last 10 years and future plans
3.4. Detailed responses of the surveys
3.4.1. Contents
3.4.2. Methodology
3.5. HEUNI-survey
3.6. Four victimisation surveys in a nutshell
3.7. Summary
4. Information exchange with the Eurostat during the project
5. Country visits
6. Implementation of the new European crime victimisation survey
6.1. Introduction
6.2. How often?
6.3. Sampling
6.3.1. Concepts in sampling
6.3.2. General schedule for sampling design
6.3.3. Fieldwork, sampling file and related data
6.3.4. Specific sampling questions in crime victim surveys
6.3.5. Conclusion and administrative views
6.4. Interviewing mode
6.5. Reference period
6.6. Timing of the fieldwork
6.7. Languages and translating the questionnaire
6.8. Stand-alone survey or module within an existing survey?
6.9. The structure of the questionnaire
6.10 Technical instructions for reading the questionnaire and for interviewing
7. European crime victim survey questionnaire draft
8. Interviewer training
9. Testing
10. Summary and discussion of the suggestions for the survey model
Appendices
Appendices
Appendix A. Agendas and participants of the country visits
Appendix B. HEUNI presentation in Rome 7.6.2007
Appendix C. Telephone coverage and different telephone types in the Member States
Appendix D. Presentation in Luxembourg 29.6.2007
Appendix E. Omitted section: Attitudes and opinions of the criminal justice system
Appendix F. Showcards
Appendix G. Advance letter
1. Description of the action
The proposal on crime and victimisation surveys was in the Grant Agreement (38400.2005.002-2006.052) defined in the following way:
1)a database of the surveys conducted in each MemberState of the European Union. The database will contain information of the results of the UNECE-UNODC Task Force on Victim Surveys conducted in the autumn of 2005. These results contain a comparative analysis (e.g. sample size, interviewing method, response rate, reference period, victimisation events), highlighting the common and distinctive features of the completed crime victim surveys in the countries.
A supplementary survey to the member countries will be completed in the summer of 2006. That survey is going to focus on the general possibilities of the countries to conduct different kinds of surveys (e.g. the possibility to cover minorities, all areas of the country, available survey methods, expertise in victimisation surveys in the relevant organisations, experience in decreasing the non-response). Also other sources of information are used to construct the database (e.g., the telephone coverage of the country).
2)a proposal for a model to be adopted at European level, which could be implemented either within existing surveys or as a stand alone survey. The proposal contains methodological recommendations, and recommends the information contents of the survey.
Methodological choices include sampling, interviewing mode, interviewing and field work instructions, estimation procedures. Information contents include the questions and instructions of the questionnaire, and instructions for interviewing.
The applicability of different interviewing methods will be thoroughly explored.
It is not the best policy to standardise the field work in every country too rigidly. It is, rather, essential to take into account the strengths and best practices in each country. This approach has been used in the European Social Survey (ESS), which is one of the best quality European surveys. Consequently, the sampling frames and fieldwork details should be tailored according to the circumstances in each country. A simple random sample of the population (persons) could be the ideal starting point in the base sample. However, this method is not possible to adopt in many European countries. Thus, each statistical authority should be asked of their best practises in sampling. If some features in the analysis are important, such as the rural/urban dimension, the sample can be stratified according to these variables. Weighting of the results should be done according to the population structure.
The questionnaire could be constructed to contain a permanent part, and a part for changing topics. The permanent part could include modules which are repeated in a comparable form in every sweep of the survey. These modules should include victimisation to property crimes, physical violence, threats, sexual violence, harassment, fear or insecurity, safety arrangements in personal life, attitudes towards crimes, corruption, other topics relevant to the work of the Commission and a sufficient amount of personal and behavioural variables.
If the time horizon of the survey is between 10-15 years, it is evident that the base questionnaire should be supplemented with special topics. Such topics could comprise, e.g., violence against women, male victimisation, ethnic minorities, at risk population, drug use, and fear related topics. If computer assisted interviewing (CAPI) could be used, the respondent could fill by him/herself parts of the survey (such as questions of his/her drug use). Otherwise, a supplementary mail questionnaire is possible.
Each country should also have the possibility to have a native module consisting of a set of questions dealing with actual topics in the criminal or legal debate in the country.
3)Co-operation and project organisation. Because a survey study consists of different parts, different expertise is needed in the project. The most important aspects are: international co-operation, survey methodology and sampling/estimation procedures.
Our team has three experts on these fields. Kauko Aromaa leads the international co-operation, Markku Heiskanen is responsible for the survey methodology and the questionnaire, and Seppo Laaksonen is responsible for statistical know-how in the project.
HEUNI has good contacts in most European countries. The team organises at least three international meetings, and experts of at least three countries are invited to participate in the survey team. The countries are selected to represent the Nordic countries, Middle- and South European countries and new member countries. The questionnaire and the field work are tentatively proven (tested) in this expert group.
Representatives of the relevant Commission services are invited to contribute to the work of the team with their expertise on topics relevant to the Commission.
4)The time table is based on the project outline in the call for proposals. Monthly e-mail uppdates of the status of the project are sent to the Commission. Also other ways of co-operation may be adopted to deliver the desired information.
The intermediate report describing the progress made is ready (t+6m). In this report, analysis of the database and the preliminary results of the new survey (modules) are presented. (First phase). This report also contains a detailed working plan of the project.
In the second phase draft list of topics and questions for EU survey (modules) on victimisation are prepared (t+9).
The final report in English describes the complete version of the database, and the detailed methodology of the survey modules (target population, sample, interviewing method, field work, survey instructions, questionnaire, and estimation recommendations). (t+12, the final report).
Also a summary presenting the progress of the work and the suggestions for the survey model will be presented.
5)Additional aspects. The survey instrument should be thoroughly studied in a survey laboratory setting to ensure the high quality of the data collection. Survey laboratories are institutions, which have been established e.g. in Finland, Sweden, Norway, Germany, the Netherlands and England to test scientifically the questionnaire and the field work.
The field testing module could include: translation of the English-language questionnaire to the native language in the countries, laboratory tests of how the respondents understand the questions (think aloud-method or other cognitive methods), and a small pilot study (how the questionnaire and the instructions work in the field; the full-fledged study in miniature).
The laboratory testing of the survey is not included in the costs of the application, but a careful testing e.g. in 3-5 countries could considerably improve the quality of the survey. Especially the cognitive testing of the questionnaire is very important. The final report will include a proposal for detailed laboratory testing of the survey.
The call for proposals does not discuss business victimisation surveys. Victim surveys could at a later stage also be directed at the business sector.
This report is the final report mentioned in section 4.
2. Background
The project started by finding out and gathering together information concerning the crime victimisation surveys conducted in EU Member States. After that, the main task was to provide a proposal for a crime victimisation survey questionnaire and forimplementation of the survey.The draft questionnaire isthe most extensive part of this report.
The project started in September 2006, and lasted for 12 months. The action has followed the time table presented in the Table 1.A progress report (intermediate report) was written and sent to Eurostat in the end of February 2007[1]. The intermediate report described the action during the first six months of the project.
This final report pulls together the outputs of the project and presents the key materials produced during the project.
Table 1. Timetable for the European Crime Victim Survey plan
The project teamconsisted of Director Kauko Aromaa (HEUNI), Dr. Markku Heiskanen (HEUNI), Professor Seppo Laaksonen (University of Helsinki) and Minna Viuhko (HEUNI).Kauko Aromaa led the international co-operation, Markku Heiskanen was the project manager and was responsible for the survey methodology and the questionnaire, and Seppo Laaksonen was responsible for the statistical expertise in the project. Minna Viuhko worked as a researcher in the project. The contact person in Eurostat was Mr. Geoffrey Thomas.
3. Victim surveys conducted in the Member States
One of the tasks of the project was to produce a report of the survey that the UNECE (United Nations Economic Commission for Europe) and the UNODC (United Nations Office on Drugs and Crime) had sent to the UNECE Member States. The aim was to give an overview on how crime victimisation surveys are conducted in the EU Member States, in the candidate countries and in the EFTA countries.
3.1. Introduction
The purpose of this part of the report is to provide background information for developing victimisation surveys in Europe. The information reported here comes from a questionnaire, which was sent by UNECE and UNODC to countries in the UNECE region in 2005. The intention in this report is to study what kind of crime victimisation surveys the EU Member States (and candidate countries and some other UNECE member counties) have conducted in the last ten years. The countries were also asked whether they have plans to collect new crime victimisation data.
A statistical database of the survey has been compiled by the UNECE. It contains all the information of the results of the UNECE-UNODC Task Force on Victim Surveys. Our intention here is to focus on the general possibilities of the countries to conduct different kinds of victimisation surveys.
The questionnaire was sent to 33 European countries[2]. Of these, 25 replied. The countries were: Belgium, Bulgaria, the Czech Republic, Denmark, Estonia, Finland, France, Germany, Hungary, Iceland, Ireland, Italy, Lithuania, the Netherlands, Norway, Poland, Portugal, Republic of Moldova, Romania, Slovenia, Sweden, Switzerland, Turkey, Ukraine and the United Kingdom.
Seven countries didnot answer:Austria, Greece, Latvia, Luxembourg, Malta, Slovakia and Spain. Cyprus had not done and had no plans to do any victimisation surveys. So Cyprus was excluded from UNECE-UNODC dataset. Thus one quarter of the countries did not respond. No non-response work was done by UNECE or HEUNI.
This chapter begins with a short description of the ICVS (International Crime Victimisation Survey), because it has been completed in all Member States except Cyprus (although only eight countries reported of the ICVS in their country; therefore we do not use the UNECE-UNODC data in the description of the ICVS). Second, we describe what kind of victim surveys the countries have conducted, and what their future plans in this respect are. Then, we give a short overview of the contents and the methodology of the surveys. We also report briefly the results of the HEUNI-survey. Last, we describe studies of four countries in some more detail.
3.2. International crime victimisation survey (ICVS)
The International Crime Victimisation Survey (ICVS) is a standardised sample survey that describes household members’ experiences of different crimes. The ICVS has been carried out at least once in all EU Member Statesduring 1989-2005. The ICVS has been conducted five times, andthe United Kingdom, the Netherlands and Finland have participated in all five ICVS sweeps. These countries have both acquired knowledge of the ICVS and experience of large national victimisation surveys.
On the other hand Greece and Luxemburg have participated only once. It is likely that these countries did not gain much cumulative knowledge of the 2005 sweep, either, in particular because this most recent sweep was carried out without local participation.
Table 2. The ICVS in the old EU Member States in different years. [3]
15 old Member States / 1989 / 1992 / 1996 / 2000 / 2005 / TotalAUSTRIA / 1 / 1 / 2
BELGIUM / 1 / 1 / 1 / 1 / 4
DENMARK / 1 / 1 / 2
FINLAND / 1 / 1 / 1 / 1 / 1 / 5
FRANCE / 1 / 1 / 2
GERMANY / 1 / 1 / 2
GREECE / 1 / 1
ITALY / 1 / 1 / 2
IRELAND / 1 / 1 / 1 / 1 / 4
LUXEMBOURG / 1 / 1
THE NETHERLANDS / 1 / 1 / 1 / 1 / 1 / 5
PORTUGAL / 1 / 1 / 2
SPAIN (CATALONIA) / 1 / 1 / 1 / 3
SWEDEN / 1 / 1 / 1 / 1 / 4
UNITED KINGDOM / 1 / 1 / 1 / 1 / 1 / 5
Total / 8 / 6 / 5 / 9 / 15 / 43
The fieldwork of the surveys during 1989-2000 was completed in most countries by an international market organization group (Interview-NSS). In Finland the 1992-2000 sweeps were completed by Statistics Finland. In Catalonia (2000) a national interview organization was used. These surveys were completed in each participating country in co-operation with national experts in criminology, who also contributed to the financing of the survey. This means that at least some researchers in each country were acquainted with the ICVS. In 2005, a consortium chaired by Gallup International, organised the field work; the fieldwork was not completed in most countries in close co-operation with national experts. This means that no scientific cumulative knowledge emerged of the survey content, analysis and use of the results in the respective countries.
The situation in the new EU Member States and candidate countries is less unique compared to the old 15 Member States, and only Poland, Hungary, Estonia and Turkey participated in the 2005 sweep. In the new Member States city-sampling has been common procedure probably because of undeveloped sampling frames.
Table 3. The ICVS in the new EU Member States and in the candidate countries in 1989-2005.
10 New Member States / 1989 / 1992 / 1996 / 2000 / 2005 / TotalCYPRUS / -
CZECH REPUBLIC / N / M+R / C / 3
ESTONIA / R / M+R / C / M+R / 4
HUNGARY / C / C / M+R / 3
LATVIA / M+R(95) / M(98) / 2
LITHUANIA / M+R / C / 2
MALTA / C / 1
POLAND / C / N / N / N / M+R / 5
SLOVAKIA / N / C / 2
SLOVENIA / C / R / C / 3
Total / 1 / 5 / 9 / 7 / 3 / 25
New Member States from 1.1.2007
BULGARIA / C / C
ROMANIA / R / C
Candidates
CROATIA / C / C
TURKEY
C = city / R = rural sample included
M = multiple cities / N = from a national survey
In the ICVS-questionnaire, victimisation is asked for the last five years, which probably yield an underestimate of the incidents.On the other hand, a severe problem in the ICVS is the small sample size (the goal has been approximately 2000 interviews/country). Victimisation to specific crimes is usually rare, and therefore the number of victims in national samples is very small and prone to heavy random variation.
Because the ICVS has been conducted in all Member States (except in Cyprus), it is despite of its defects the most important crime victim survey for international comparisons, and if time series in history are required, comparability to the ICVS results is essential.
3.3. Victimisations surveys during the last 10 years and future plans
Because the aim of our project is to develop guidelines for a general victimisation survey, violence against women surveys are excluded from the analysis. Many countries have conducted violence against women surveys. The main idea in violence against women surveys is to measure violence especially in partner relationships, which often has not succeeded sufficiently well in general victimisation surveys, including the ICVS. One solution could be to add a separate violence against women section to a general victimisation survey (like in the British Crime Survey), another could be to change the general victimisation survey in the direction of the violence against women surveys. To include the violence against women survey approach in general victimisation surveys would require improvements in the fieldworkand questionnaire design. The fieldwork implementation would require interviewer training taking into account the sensitive issues of the survey for both sexes. The screening questions should be more concrete with illustrative examples of different everyday victimisation situations.
Of violence against women a separate UNECE survey was launched[4]. The results of this study are not reported here.
The countries were asked, what kind of household surveys they haveconducted to collect data on crime in the last ten years and do they have plans to conduct new surveys, and if they do, what kind of surveys.
Table 4. The survey types different countries have conducted and planned.
Survey type / Completed / Plannedsurveys / surveys
1. Ad hoc victimisation survey / 11 / 4
2. Periodic victimisation survey / 15 / 14
3. Continuous victimisation survey / 4 / 4
4. Multipurpose survey / 16 / 8
5. Other / 3 / 1
The five row variables in Table 4 describe the survey type. A multipurpose survey was the most common type. A multipurpose survey means that the victimisation section is only part of the survey. Statistical offices often conduct multipurpose surveys into which a short set of victimisation question is included, e.g. living conditions surveys, health surveys or general household surveys. E.g. living condition surveys may contain a safety section in which questions of victimisation, fear and precautionary measures are asked. Although this allows a wide description of the connection between victimisation and fear, and other areas of living conditions, the contents of the victimisation question set is often too narrow to allow a reliable and detailed analysis of victimisation.