Roundtable 1
Abstracts and Handouts

Data for Comparative European Socio-Economic Research:
Challenges, Opportunities and Constraints
12-13 September 2002
Wivenhoe House, University of Essex

Abstracts are in the order of presentation

Marcia Freed Taylor
Setting the Scene

• About NESSIE – what it is, what it is intended to achieve
• NESSIE activities:

• Roundtables
• Guidelines of Good Practice
• Reviews of Data Availability
• Development of web-based tools
• Research and Development Projects

• This Roundtable

Kevin Schürer
Data from the Academic Sector
This short paper will compliment that of Bjørn Henrichsen in that it will discuss the work of the CESSDA network in making pan-international data available to the academic community. The paper will briefly explore the following issues:

• New developments within the UK;
• Data access agreements within CESSDA;
• Pan-international ‘academic’ data;
• Problems in obtaining comparative data;
• Building comparative data resources.

Bjørn Henrichsen
The CESSDA Experience
The first European social science data archives were founded in the sixties. Over the years most European countries have established data archives or data services as some of the countries call their institutions.

From the start there has been a close cooperation between the different national institutions and in 1976 the Council of European Social Science Data Archives (CESSDA) was established. Today, CESSDA includes data archives from almost all European countries. In 2002 there are 21 member institutions.

The main activity of CESSDA is to promote the acquisition, archiving and distribution of electronic data for social science teaching and research in Europe. It stresses exchange of data and fosters the development of new organizations that are sharing the same objectives.

In my presentation I will give an overview of CESSDA’s membership, the way the organization functions and briefly point to future directions for cooperation between the social science data archives.

Gaston Schaber
Bound to cooperate !
(Data –producers and Analysis – producers will have to work together in building a research infrastructure for the socio-economic sciences.)

Abstract

1. First a paraphrase of Immanuel KANT’s dictum: “Begriffe ohne Anschauungen sind leer. Anschauungen ohne Begriffe sind blind.” Which becomes here: Analyses without relevant data are vain. Date outside a conceptual frame for analysis are meaningless.
2. The uncertain and unsafe relationship between data production and analysis-production in the socio-economic sciences as compared to somewhat harder disciplines will be viewed in two perspectives:
   • functional : disconnected versus interactivestructural:
   • disconnected actors: office versus academia
3. This relationship is not fatal: it can be changed. Examples of bad and good practices.
4. Some texts related to the topic.
5. Outlook: some issues which, the socio-economic sciences better should not avoid while planning data infrastructures for research.

(Last-minute contribution in replacement of Pierre HAUSMAN’s planned presentation)

Ruud Bless
Creating comparative data from existing sources
Comparability of data is a requirement for studies on patterns and trends over time and space. But in the social sciences comparable data about a specific topic are hard to find, not only due to data and case definitions but also due to differences in sampling and data collection methods. In several fields of social science attempts have been made to improve comparability by harmonising definitions and methods. But even if in a lengthy process of consensus seeking the researchers in the field would agree on these issues, the benefits are only in the future, whereas the need for comparisons already exists today.

In recent years the European Monitoring Centre for Drugs and Drug Addiction (EMCDDA) has launched several initiatives to harmonise epidemiological data about drug use and drug related problems in the EU Member States. In most cases this has proven to be a complex process of consensus seeking about national researchers and other stakeholders. Although substantial progress has been made, the benefits will be in the future, whereas the need for comparisons already exists today. For this reason the harmonisation processes are paralleled, whenever possible, by attempts to maximise the comparability of existing data sources.

From my experience as coordinator for EMCDDA on the harmonisation of survey data on drug prevalence, I will first discuss some general methodological aspects of comparability, then continue with the principles applied to harmonise data items, and conclude with examples of making data comparable by applying these principles to existing datasets of different countries and different years. In between I will present recommendations for good practise in accountability of data collection and database construction.

European Monitoring Centre for Drugs and Drug Addiction (EMCDDA).
The EMCDDA was established by Council Regulation (1) in 1993 and became operational in 1995. The Centre's objective is to provide the Community and its Member States with objective, reliable and comparable information at European level concerning drugs, drug addiction and their consequences. The information is intended to help provide the Community and the Member States with an overall view of the drug phenomenon when, in their respective areas of competence, they take measures or decide on action.

To achieve this objective, the Centre is required to collect, analyse, compare and disseminate information at EU level, and in particular to ensure improved comparability, objectivity and reliability of data at European level by establishing indicators and common criteria. Since 1996, the EMCDDA has been developing and testing a number of epidemiological indicators to assess the extent and patterns of drug use, and its determinants and consequences. Some indicators have been more thoroughly developed and adopted as “Key Indicators” (2), although other key indicators may be eventually defined in the future.

EMCDDA collects data from all European Union countries on a regular basis through “National Focal Points”. These Focal Points provided each year a National Report, together with a set of “National Standardised Statistical Tables” with aggregated data. Specific projects collect more detail information on different indicators (“specialised databases”). EMCDDA is developing an integrated epidemiological database based on these sources; it is denominated EISDD (EMCDDA epidemiological info system on drug data).

EMCDDA presents results mainly through its Annual Report that includes a comprehensive set of summarised statistical tables (data library), based on the National Reports, National Tables and project’s results (SEE LINKS BELOW). In the medium term, the EISDD will be accessible to Focal Points, researchers and others. Conditions and terms of access to it have to be defined. It is likely that different parties have different accession rights.

http://www.emcdda.org/

http:// annualreport.emcdda.org/

http:// annualreport.emcdda.org/en/sources/index.html

(1) Council Regulation (EEC) No 302/93 of 8 February 1993 on the establishment of a European Monitoring Centre for Drugs and Drug Addiction

(2) The indicators are: 1) Extent and pattern of drug use in the general population, 2) prevalence of problem drug use, 3) demand for treatment by drug users, 4) drug-related deaths and mortality of drug users and 5) drug-related infectious diseases (HIV, hepatitis)

Ekkehard Mochmann
Communications Research:  Data Production, Data Access and Data Needs from a European Perspective
Media and world spanning communication systems define the daily agendas, contribute to shaping national identities and support the development of transnational orientations. The dynamic developments of communication technologies add to the reasons, why communications research should open to a wider perspective, going beyond national orientations.

Communications research in the emerging Europe is facing a multitude of sociolinguistic communities and a highly fragmented media system. While media companies have entered a concentration process and fight for market shares, communications research is still in the starting wholes to develop a European perspective. Persistent organisational efforts have succeeded to create transnational networks like the European Consortium for Communication Research (ECCR). There are even European Journals devoted to this discipline like “Communications: The European Journal of Communication Research”.

In spite of these successful attempts at infrastructure building, academic empirical communication research with a true European perspective is seriously handicapped in a situation where the contrary could have been expected: Although European societies  with their cultural diversity basically are an ideal laboratory for comparative research. Like in other disciplines of socio- economic research this situation has been rarely used for comparative research. Consequently there is a serious lack of coherent European data bases, which would meet even the minimal requirements for functionally equivalent measurement of central variables over space and time.

Nevertheless communication research can build on a remarkable number of data for the individual member states. Making use of these available data requires better access to available national holdings and better integration of the available data. This would contribute to create at least a rudimentary historic data base for empirical analyses. The emerging Europe, however, needs new and additional initiatives which help to open new horizons for the  nationally  focussed communications research of the past. Among these with high priority:

1. Creating inventories and opening access to data bases which are not accessible via the CESSDA network so far.
2. Defining the needs and coordinate the collection of comparative data in the academic, commercial and statistical sector.
3. Establish codes of best practice for generating, documenting, distributing and using the data bases.
4. Provide support for competence centres with relevant data bases and networks of researchers.
5. Provide training in data collection, documentation and analysis.

The European Consortium for Communication Research (ECCR) has agreed to support the NESSIE initiative by a survey on available data and user needs of its membership.

Roger Jowell
Creating a new Comparative Dataset in Europe
The European Social Survey (ESS) is a new comparative time series funded by a combination of European Commission, European Science Foundation and national academic funding sources.  Its purpose is to fill a gap in time series designed to monitor changing underlying attitudes and values over time.  While surveys of this sort exist in certain countries, and in some forms at a European level too, the basis they provide for rigorous cross-national analysis remains patchy and fragile.

The ESS, which will span 24 European countries, is an attempt to pioneer and ‘prove’ a new standard of methodology for cross-national attitudinal surveys.  It will attempt rigorously to monitor and interpret the interaction between Europe’s changing institutions and the attitudes, beliefs and behaviour patterns of its diverse populations.

The paper will describe the key features of the new project and the motivation behind them.

Franz Kraus
Official Statistics from International Organisations and National Statistical Offices

The presentation consists of 4 parts.

In the introductory part the need for an efficient and adequate European data infrastructure and some basic components is shortly discussed and related to official statistics as a major source for cross-national research.

The second part deals with the internationalisation of official statistics. It starts with a short glance back to the evolution of modern statistics and the growing interaction between national statistics production and international requirements for cross-nationally comparable data. Then major areas and types of statistical co-operation in today’s multi-level statistics system are summarized (statistical harmonisation machinery, production and reporting systems).

The third part then describes at length the current European Statistical System: its organisational and legal characteristics, the expansion of broad subject matter areas in EU statistics, and the impact of increasing EU information needs on national statistics production. Here, emphasis is placed on the diffusion of major socio-economic surveys with European orientation.

The fourth part deals with data availability: tabular databases, micro-data and meta-data available from international organisations, Eurostat, and national statistical offices. Emphasis is placed on access modes and access conditions to microdata as well as availability of adequate meta-information on survey content and comparability. This part draws on the results of a recently completed project (EuReporting, subproject ‘Access to Comparative Official Microdata’), co-financed by the European Union. It shows the need for substantial improvements in data documentation, more differentiated and increased access to microdata, and a strengthening of co-operation between official statistics and academic infrastructure activities at European level.

Mark Smith
Labour Markets and Employment
This presentation considers the successes and problems relating to the availability and comparability of data relating to labour markets and employment. The presentation will also provide an overview of the state of data provision for comparative research in labour markets and employment. There is a wealth of data available relating to employment and the labour market reflecting the use of such information by policy makers, research and international bodies. Although there are often common definitions in use and core elements to national surveys, comparability is not necessarily straightforward. These comparability problems often reflect the impact of national institutions and policies that impact upon both the collection of data and the responses of individuals. Furthermore different national regulations shape the availability and method of publication of both national and international datasets. These can have a significant impact on comparability and the potential uses that researchers can put the data to. This presentation draws on the speaker's experience in using comparative datasets on a number of European research projects including in work for the European Expert Group on Gender and Employment and two TSER networks.

Eric Marlier
Data access issues in an EU context
The paper will discuss the implications for European comparative research of three key new European  regulations relating to statistical confidentiality: Council Regulation (EURATOM, EEC) No 1588/90  of 11 June 1990 on the transmission of data subject to statistical confidentiality to the Statistical Office of the European Communities; Council Regulation (EC) No 322/97 of 17 February 1997 on Community Statistics; and Commission Regulation (EC) No 831/2002 of 17 May 2002 implementing Council Regulation (EC) No 322/97 on Community Statistics, concerning access to confidential data for scientific purposes.

Dimiter Philipov
Demographic data - state of the art

1. Sources of demographic data:

• Population census
• Vital statistics
• Population register
• Civil registration system
• Surveys

Surveys are the main source of scientific data. In CEEC (Central and Eastern European Countries) there have been very few surveys, if any, during the last 10 years.

2. Fields of research and their data requirements - population size, fertility, family, morbidity, mortality, migration.

Problems exist with data concerning family and migration, because their definitions are fuzzy. In the case of family formation, it is cohabitation that causes some problems. When does it begin? In the case of migration, the definition should clearly distinguish any moves, such as visits or business trips, from real migration.

3. Data comparability. Most of the data in censuses and vital statistics are crudely comparable. Few data are finely comparable though. The Human Mortality Database developed at the Max-Planck Institute for Demographic Research is an example of a data base that uses vital statistics and population census data for finer comparability among countries.

The Fertility and Family Surveys (FFS) are an example of a harmonised and coordinated data collection. It was initiated by the Population Activities Unit (PAU) at UN ECE, Geneva, and carried out during the 90s. One and the same core questionnaire was used for surveys in 23 countries. The data were gathered at PAU and distributed with the one and the same file format and structure.

4. Access to data sources.
Access to population registers is diverse. In some countries these data are strictly confidential; in other counties access is free provided the data are used in the statistical office where they are stored. Some countries may sell register data but the price is high.

Survey data are subject to conventional problems such as confidentiality. Seemingly problems of accessibility are not as acute as in other social fields. Once anonymised the data are usually available for scientific research.

5. Funding is a significant problem particularly for the CEEC. It is the reason for not having good quality surveys in these countries.

6. Non-response in demographic surveys is usually high in Western European countries but low in CEEC. Confidentiality fears exist. They are significant obstacles for census and surveys. Other fears: respondents do not want to let the interviewers enter their house.

7. Potential for data improvement
There is little potential in improving population census and vital statistics. Changes in these systems of data gathering are very expensive.

There is a significant potential in the improvement of surveys. The examples of GGS and PPA are relevant. GGS was initiated by PAU at UN ECE. It is based on the coordinated work of 3 scientific institutions. A joint team of researchers works for the creation of a core questionnaire; the creation of a unified contextual data base. It is supposed that the survey design will be analogous in the countries where the survey will be carried out. PPA is based on the co-ordination of a larger group of institutions and other scientific groups. They design a core questionnaire. Both GGS and PPA do not consider financial involvement. Thus the problem of funding remains.

Demographic questionnaires include a number of issues that are the topic of other social and economic surveys. Examples: income, economic status, values. For this purpose demographers include relevant questions from proper specialized surveys. For example, values are studied by incorporating questions from EVS. The problem is in that the specialized surveys rely on a detailed study of the topic of interest, while demographers must restrict that topic to a small number of questions. Hence the quality could be low. One way to ease the problem is to dispose with a bank of questionnaires where one could be able to find diverse approaches to one and the same topic. The creation of such a bank is easy and inexpensive.

Erwin K. Scheuch
Half of my task is summarising and commenting on the various contributions during the first day of our meeting. In addition, I like to suggest three topics for methodological undertakings:

• The high level of processes of diffusion between countries in Western Europe calls in to question the suitability of nation as a sampling frame, and as a casual factor in explaining between-country variations.

• In cross-national research, within-country variations should be compared with between-country variations. This is particularly advisable in the case of Germany, where correspondence analyses suggest that for some factors East and West Germany are further apart than West Germany and its Western neighbours. In such analyses Northern Italy was closer to Western European neighbours than to Southern Italy.

• A systematic evaluation of multilevel methods for purposes of international comparisons should be performed with already available studies – such as the LAMMP Project or the study of fertility by Entwisle and Mason. Can one add census tract data to cross national data sets to permit cross-level analyses?

We require a more detailed inventory of data sets permitting cross-national comparisons that are available in IFDO data centres. To prepare this, we need to develop a format that is acceptable to all IFDO organizations.