Transition to Adulthood and Collective Experiences Survey (TRACES)
April 2006 - December 2006
Former Yugoslavian countries: Slovenia, Croatia, Serbia, Montenegro, Bosnia-Herzegovina, Kosovo, FYR of Macedonia
Transition to Adulthood and Collective Experiences Survey (TRACES) is a research program focusing on war and economical victimization in ex-Yugoslavia and their impact on young adult's attitudes and values. The main quantitative survey was launched in 2006; Principal Investigator: Dario Spini; Project Coordinator: Guy Elcheroth. Its design included two partially embedded samples following a random sampling strategy stratified in 80 areas covering the entire ex-Yugoslavian territory. An individual-based questionnaire (Cohort sample, including individuals born between 1968 and 1974; N = 2'254) was coupled with a representative sample of the adult population (Random sample; N = 3'975). This second sample enabled to record, using life calendars, valid data on the experiences communities faced from 1990 to 2006 and during individuals' transition to adulthood (from 15 to 35 years old) across diverse regions of former Yugoslavia. As a result TRACES presents an innovative multilevel survey design enabling the location of respondents' experiences in time and space. This methodological and technical report provides details about survey design and instruments, about the context of the production of data and data quality. The purpose is to offer information as transparent as possible for current and future data users and to share this experience with the broad scientific community.
TRACES is a scientific project with the ambition to collect information on the collective experiences of young adults' vulnerability in the beginning of the nineties. The general hypothesis we follow is that collective experiences of vulnerability, be they due to armed conflicts or economic penury, shape social representations related to societal issues like rights, justice or intergroup relationships. In order to deal with these issues, we first decided to direct our investigations to a close region that had recently experienced such collective and dramatic events: former Yugoslavia. Then, a second important decision was to think about a survey design that would enable us to articulate individual expressions of attitudes to contexts in which individuals experienced different types of life experiences. In order to do so, we needed to gather data on collective experiences structured by time and space. However, precise information on what happened in former Yugoslavia during the wars is full of gaps. We thus had to imagine an innovative sampling strategy. We came to the conclusion that we needed to link the individual-based questionnaire on a specific cohort (individuals born between 1968 and 1974) to a representative sample of the adult population, which would enable us to record, using life calendars, valid data on the experiences communities faced during the nineties across diverse regions of former Yugoslavia. This resulted in what we could call a multilevel survey design, which we believe makes TRACES a very innovative survey within social sciences.
Method: Standardized face-to-face interviews
Population, unit of observation: TRACES design includes two partially embedded samples. Both follow a random sampling strategy stratified in 80 areas covering the entire ex-Yugoslavian territory. The first sample, called random sample, was a random selection of about 50 respondents per area belonging to the general adult population (born in 1981 or earlier). These respondents answered to a first part of the questionnaire composed of life calendars, in which they were asked to indicate whether, when and for how long a number of events related to war and daily life occurred to them. Responses from the random sample were used to construct second-level indicators. The second sample, named cohort sample, was a random selection of about 40 residents born between 1968 and 1974 within each area, who answered to attitudinal measures in addition to the first part of the questionnaire used for the random sample.
Size of sample/selections: Random Sample: N = 3,975 (Slovenia: n = 406; Croatia: n = 850; Bosnia-Herzegovina: n = 746; Serbia & Montenegro: n = 876; Kosovo: n = 551; FYR Macedonia: n = 546). Cohort Sample: N = 2,249 (Slovenia: n = 192; Croatia: n = 434; Bosnia-Herzegovina: n = 380; Serbia & Montenegro: n = 792; Kosovo = 253; FYR Macedonia: n = 198)
Participant selection or sampling method: Stratified random sample, with oversampling of the cohort 1968-1974
Survey design: Data were collected between April 2006 and Decembrer 2006. A stratification plan that divides the ex-Yugoslavian territory into 80 geographical areas was elaborated. A particular objective of the stratification procedure was to over-sample ethnic or national groups that represent between 5% and 10% of the total sample (Albanians, Bosniaks, Slovenes and Macedonians) in order to be able to compute statistical estimates for the subgroups separately. Concretely, we established the following guidelines for defining areas:
1. Areas are regional subdivisions within current state boundaries, as well as within boundaries of different political entities (i.e. Federation of Bosnia and Herzegovina/Republika Srpska; Serbia Proper/Montenegro/Vojvodina/Kosovo).
2. Each area is geographically continuous.
3. The total number of areas for the whole former Yugoslavia has been fixed to 80.
4. Each area is defined as a cluster of a certain number of municipalities, which respects the current boundaries of these municipalities.
5. When intermediate levels of political subdivisions exist (Croatia: counties; Federation of Bosnia and Herzegovina: counties; Serbia Proper and Vojvodina: districts; Kosovo: UNMIK districts; FYR Macedonia: historical communes), areas should as far as possible correspond either to one unit defined by this subdivision or to a cluster of several smaller units.
6. Six urban areas are defined by the boundaries of the major cities: Belgrade, Ljubljana, Pristina, Sarajevo, Skopje, and Zagreb.
7. Apart from the six urban areas, numbers of inhabitants by area should not vary dramatically from one area to another, especially within one political entity.
8. Smaller political entities are over-sampled compared to larger political entities, i.e. there should be smaller average numbers of inhabitants for areas within smaller countries than within larges countries.
9. Regions populated mainly by major ethnic groups that are less numerous within former Yugoslavia (Albanians, Bosniaks, Macedonians, and Slovenes) are over-sampled compared to regions populated mainly by the two most numerous ethnic groups (Croats and Serbs).
10. Regions which are highly heterogeneous with regard to geographical or historical factors - especially factors that affected their inhabitants' destiny during the 90s wars - should not be clustered within the same area.
11. Codes for non-urban areas are composed of two letters referring to the larger geographical region followed by one number which specifies the precise location: areas are ranked from West to East and from North to South within geographical regions. Codes for urban areas are composed of two letters only, referring to the city's name.
Data collected by: PRISM Research. Contracts with four partner agencies were established by PRISM in Slovenia (Mediana, Institute for Market and Media research), Croatia (Hendal market research), Serbia and Montenegro Argument, Research and Analytical Centre), and FYR Macedonia (BRIMA Gallup International, Public Opinion Poll Agency)
Other significant comments: For other remarks see the Methodological and Technical Report
|Financed by||Study type||
|Start - End date||01.01.2004 - 28.06.2006|